Inorganic fullerene-like nanoparticles and inorganic nanotubes

Tenne, Reshef (Ed.)
Enyashin, Andrey N. (Ed.)

December 2015

212

MDPI - Multidisciplinary Digital Publishing Institute

The subjects of the presented papers cover a wide range of challenges in the area of inorganic fullerene-like nanoparticles and nanotubes. However, it can include only a few comprehensive experimental and theoretical efforts, stepwise evaluating the rationalization of the synthesis, and elucidation of the stability, mechanical, electronic and adhesive properties of these nanostructures. We believe that this thematic issue can be helpful, not only for an advanced researcher to grasp the latest developments in this field, but also to permit a beginner to gain a deeper insight into the field of inorganic fullerene-like nanoparticles and nanotubes.

Inorganic Chemistry

English

9783038421504; 9783038421498

10.3390/books978-3-03842-149-8

Inorganic Fullerene-like Nanoparticles
and Inorganic Nanotubes
Edited by

Reshef Tenne and Andrey N. Enyashin
Printed Edition of the Special Issue Published in Inorganics

www.mdpi.com/journal/inorganics

Reshef Tenne and Andrey N. Enyashin (Eds.)

Inorganic Fullerene-like Nanoparticles
and Inorganic Nanotubes

This book is a reprint of the special issue that appeared in the online open access journal
Inorganics (ISSN 2304-6740) in 2014 (available at:
http://www.mdpi.com/journal/inorganics/special_issues/inorganic_fullerenes).

Guest Editors
Reshef Tenne
Weizmann Institute
Israel
Andrey N. Enyashin
Institute of Solid State Chemistry UB RAS
Russia

Editorial Office
MDPI AG
Klybeckstrasse 64
Basel, Switzerland

Publisher
Shu-Kun Lin

Managing Editor
Min Su

1. Edition 2015
MDPI • Basel • Beijing • Wuhan

ISBN 978-3-03842-149-8 (PDF)
ISBN 978-3-03842-150-4 (Hbk)

© 2015 by the authors; licensee MDPI, Basel, Switzerland. All articles in this volume are
Open Access distributed under the Creative Commons Attribution 4.0 license
(http://creativecommons.org/licenses/by/4.0/), which allows users to download, copy and
build upon published articles even for commercial purposes, as long as the author and
publisher are properly credited, which ensures maximum dissemination and a wider impact of
our publications. However, the dissemination and distribution of physical copies of this book
as a whole is restricted to MDPI, Basel, Switzerland.

III

Table of Contents
List of Contributors ............................................................................................................... V
About the Guest Editors..................................................................................................... VIII
Preface .................................................................................................................................IX

Nourdine Zibouche, Mahdi Ghorbani-Asl, Thomas Heine and Agnieszka Kuc
Electromechanical Properties of Small Transition-Metal Dichalcogenide Nanotubes
Reprinted from: Inorganics 2014, 2(2), 155-167
http://www.mdpi.com/2304-6740/2/2/155 .............................................................................. 1

Volker Brüser, Ronit Popovitz-Biro, Ana Albu-Yaron, Tommy Lorenz,
Gotthard Seifert, Reshef Tenne and Alla Zak
Single- to Triple-Wall WS2 Nanotubes Obtained by High-Power Plasma Ablation of WS2
Multiwall Nanotubes
Reprinted from: Inorganics 2014, 2(2), 177-190
http://www.mdpi.com/2304-6740/2/2/177 ............................................................................ 14

Claudia C. Luhrs, Michael Moberg, Ashley Maxson, Luke Brewer and Sarath Menon
IF-WS2/Nanostructured Carbon Hybrids Generation and Their Characterization
Reprinted from: Inorganics 2014, 2(2), 211-232
http://www.mdpi.com/2304-6740/2/2/211 ............................................................................ 28

Matthew R. Farrow, John Buckeridge, C. Richard A. Catlow, Andrew J. Logsdail,
David O. Scanlon, Alexey A. Sokol and Scott M. Woodley
From Stable ZnO and GaN Clusters to Novel Double Bubbles and Frameworks
Reprinted from: Inorganics 2014, 2(2), 248-263
http://www.mdpi.com/2304-6740/2/2/248 ............................................................................ 50

Mohammed Naffakh and Ana M. Díez-Pascual
Thermoplastic Polymer Nanocomposites Based on Inorganic Fullerene-like Nanoparticles and
Inorganic Nanotubes
Reprinted from: Inorganics 2014, 2(2), 291-312
http://www.mdpi.com/2304-6740/2/2/291 ............................................................................ 65

IV
Fang Xu, Nannan Wang, Hong Chang, Yongde Xia and Yanqiu Zhu
Continuous Production of IF-WS2 Nanoparticles by a Rotary Process
Reprinted from: Inorganics 2014, 2(2), 313-333
http://www.mdpi.com/2304-6740/2/2/313 ............................................................................ 87

Albert Rimola and Mariona Sodupe
Gas-Phase and Microsolvated Glycine Interacting with Boron Nitride Nanotubes. A B3LYPD2* Periodic Study
Reprinted from: Inorganics 2014, 2(2), 334-350
http://www.mdpi.com/2304-6740/2/2/334 .......................................................................... 113

Olga Brontvein, Reshef Tenne and Andrey Enyashin
The Role of Lead (Pb) in the High Temperature Formation of MoS 2 Nanotubes
Reprinted from: Inorganics 2014, 2(2), 363-376
http://www.mdpi.com/2304-6740/2/2/363 .......................................................................... 130

Jamie Cook, Steven Rhyans, Lou Roncase, Garth Hobson and Claudia C. Luhrs
Microstructural Study of IF-WS2 Failure Modes
Reprinted from: Inorganics 2014, 2(3), 377-395
http://www.mdpi.com/2304-6740/2/3/377 .......................................................................... 143

Georges Moussa, Chrystelle Salameh, Alina Bruma, Sylvie Malo, Umit B. Demirci,
Samuel Bernard and Philippe Miele
Nanostructured Boron Nitride: From Molecular Design to Hydrogen Storage Application
Reprinted from: Inorganics 2014, 2(3), 396-409
http://www.mdpi.com/2304-6740/2/3/396 .......................................................................... 162

Daniel Raichman, David Strawser and Jean-Paul Lellouche
Design of Experiments: Optimizing the Polycarboxylation/Functionalization of Tungsten
Disulfide Nanotubes
Reprinted from: Inorganics 2014, 2(3), 455-467
http://www.mdpi.com/2304-6740/2/3/455 .......................................................................... 176

Nourdine Zibouche, Agnieszka Kuc, Pere Miró and Thomas Heine
Noble-Metal Chalcogenide Nanotubes
Reprinted from: Inorganics 2014, 2(4), 556-564
http://www.mdpi.com/2304-6740/2/4/556 .......................................................................... 190

V

List of Contributors
Ana Albu-Yaron: Department of Materials and Interfaces, Weizmann Institute of Science,
P.O. Box 26, Rehovot 76100, Israel.
Samuel Bernard: IEM (Institut Europeen des Membranes), UMR 5635 (CNRS-ENSCMUM2), Universite Montpellier 2, Place E. Bataillon, F-34095 Montpellier, France.
Luke Brewer: Mechanical and Aerospace Engineering Department. Naval Postgraduate
School 700 Dyer Rd. Watkins Hall. Monterey, CA 9394, USA.
Olga Brontvein: Department of Materials and Interfaces, Weizmann Institute of Science,
Rehovot 7610001, Israel.
Alina Bruma: Laboratoire CRISMAT, UMR 6508 CNRS/ENSICAEN/UCBN, 6 boulevard
du Maréchal Juin, 14050 Caen, France.
Volker Brüser: Leibnitz Institute for Plasma Science and Technology (INP),
Felix-Hausdorff-Straße 2, 17489 Greifswald, Germany.
John Buckeridge: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
C. Richard A. Catlow: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
Hong Chang: College of Engineering, Mathematics and Physical Sciences, University of
Exeter, Exeter EX4 4QF, UK.
Jamie Cook: Mechanical and Aerospace Engineering Department, Naval Postgraduate
School, 700 Dryer Rd., Watkins Hall Rm. 305, Monterey, CA 93943, USA.
Umit B. Demirci: IEM (Institut Europeen des Membranes), UMR 5635 (CNRS-ENSCMUM2), Universite Montpellier 2, Place E. Bataillon, F-34095 Montpellier, France.
Ana M. Díez-Pascual: Instituto de Ciencia y Tecnología de Polímeros (ICTP-CSIC), Juan de
la Cierva 3, 28006 Madrid, Spain.
Andrey Enyashin: Institute of Solid State Chemistry RAS Pervomayskaya Str., 91, 620990
Ekaterinburg, Russia.
Matthew R. Farrow: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
Mahdi Ghorbani-Asl: School of Engineering and Science, Jacobs University Bremen,
Campus Ring 1, 28759 Bremen, Germany.
Thomas Heine: School of Engineering and Science, Jacobs University Bremen, Campus
Ring 1, 28759 Bremen, Germany.
Garth Hobson: Mechanical and Aerospace Engineering Department, Naval Postgraduate
School, 700 Dryer Rd., Watkins Hall Rm. 305, Monterey, CA 93943, USA.
Agnieszka Kuc: School of Engineering and Science, Jacobs University Bremen, Campus
Ring 1, 28759 Bremen, Germany.
Jean-Paul Lellouche: Department of Chemistry, Nanomaterials Research Center, Institute of
Nanotechnology & Advanced Materials, Bar-Ilan University, Ramat-Gan 52900, Israel.

VI
Andrew J. Logsdail: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
Tommy Lorenz: Physikalische Chemie, Technische Universität Dresden, Bergstrasse, 66b,
01062 Dresden, Germany.
Claudia C. Luhrs: Mechanical and Aerospace Engineering Department, Naval Postgraduate
School, 700 Dryer Rd., Watkins Hall Rm. 305, Monterey, CA 93943, USA.
Sylvie Malo: Laboratoire CRISMAT, UMR 6508 CNRS/ENSICAEN/UCBN, 6 boulevard du
Maréchal Juin, 14050 Caen, France.
Ashley Maxson: Mechanical and Aerospace Engineering Department. Naval Postgraduate
School 700 Dyer Rd. Watkins Hall. Monterey, CA 9394, USA.
Sarath Menon: Mechanical and Aerospace Engineering Department. Naval Postgraduate
School 700 Dyer Rd. Watkins Hall. Monterey, CA 9394, USA.
Philippe Miele: IEM (Institut Europeen des Membranes), UMR 5635 (CNRS-ENSCMUM2), Universite Montpellier 2, Place E. Bataillon, F-34095 Montpellier, France.
Pere Miró: School of Engineering and Science, Jacobs University Bremen, Campus Ring 1,
28759 Bremen, Germany.
Michael Moberg: Mechanical and Aerospace Engineering Department. Naval Postgraduate
School 700 Dyer Rd. Watkins Hall. Monterey, CA 9394, USA.
Georges Moussa: IEM (Institut Europeen des Membranes), UMR 5635 (CNRS-ENSCMUM2), Universite Montpellier 2, Place E. Bataillon, F-34095 Montpellier, France.
Mohammed Naffakh: Departamento de Ingeniería y Ciencia de los Materiales, Escuela
Técnica Superior de Ingenieros Industriales, Universidad Politécnica de Madrid, José
Gutiérrez Abascal 2, 28006 Madrid, Spain.
Ronit Popovitz-Biro: Department of Chemical Research Support, Weizmann Institute of
Science, P.O. Box 26, Rehovot 76100, Israel.
Daniel Raichman: Department of Chemistry, Nanomaterials Research Center, Institute of
Nanotechnology & Advanced Materials, Bar-Ilan University, Ramat-Gan 52900, Israel.
Steven Rhyans: Mechanical and Aerospace Engineering Department, Naval Postgraduate
School, 700 Dryer Rd., Watkins Hall Rm. 305, Monterey, CA 93943, USA; Hartnell College,
Salinas, CA 93901, USA.
Albert Rimola: Departament de Química, Universitat Autònoma de Barcelona, 08193
Bellaterra, Catalonia, Spain.
Lou Roncase: Navy's Weapons Survivability Laboratory (WSL), Naval Air Warfare Center
(NAWC), China Lake, CA 93555, USA.
Chrystelle Salameh: IEM (Institut Europeen des Membranes), UMR 5635 (CNRS-ENSCMUM2), Universite Montpellier 2, Place E. Bataillon, F-34095 Montpellier, France.
David O. Scanlon: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK; Diamond Light
Source Ltd., Diamond House, Harwell Science and Innovation Campus, Didcot, Oxfordshire
OX11 0DE, UK.
Gotthard Seifert Seifert: Physikalische Chemie, Technische Universität Dresden,
Bergstrasse, 66b, 01062 Dresden, Germany.

VII
Mariona Sodupe: Departament de Química, Universitat Autònoma de Barcelona, 08193
Bellaterra, Catalonia, Spain.
Alexey A. Sokol: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
David Strawser: Department of Chemistry, Nanomaterials Research Center, Institute of
Nanotechnology & Advanced Materials, Bar-Ilan University, Ramat-Gan 52900, Israel.
Reshef Tenne: Department of Materials and Interfaces, Weizmann Institute of Science, P.O.
Box 26, Rehovot 76100, Israel.
Nannan Wang: College of Engineering, Mathematics and Physical Sciences, University of
Exeter, Exeter EX4 4QF, UK.
Scott M. Woodley: Department of Chemistry, Kathleen Lonsdale Materials Chemistry,
University College London, 20 Gordon Street, London WC1H 0AJ, UK.
Yongde Xia: College of Engineering, Mathematics and Physical Sciences, University of
Exeter, Exeter EX4 4QF, UK.
Fang Xu: College of Engineering, Mathematics and Physical Sciences, University of Exeter,
Exeter EX4 4QF, UK.
Alla Zak: Faculty of Science, Holon Institute of Technology, P.O. Box 305, Holon 58102,
Israel.
Yanqiu Zhu: College of Engineering, Mathematics and Physical Sciences, University of
Exeter, Exeter EX4 4QF, UK.
Nourdine Zibouche: School of Engineering and Science, Jacobs University Bremen, Campus
Ring 1, 28759 Bremen, Germany.

VIII

About the Guest Editors
Reshef Tenne earned his Ph.D. in 1976 in the Hebrew University
of Jerusalem. He joined the Weizmann Institute in 1979, where
he was promoted to a professor in 1995. He headed the
Department of Materials and Interfaces and was the director of the
G. Schmidt Minerva Center for Supramolecular Chemistry
(2000–2007) and the Helen and Martin Kimmel Center for
Nanoscale Science (2003–2014).
Prof. Tenne is recognized for the discovery of the inorganic
fullerene-like (IF) nanostructures and inorganic nanotubes (INT)
of WS2 and MoS2 (in 1992); for their detailed study and for their
commercialization as superior solid lubricants. He holds the Drake Family Chair in
Nanotechnology (2004–); received the Materials Research Society Medal (2005); The
Kolthoff Prize in Chemistry of the Technion, Israel (2005); The Israel Vacuum Society
Excellence in Science Prize (2006); The Landau Prize in Nanotechnology (2006); was
nominated as MRS Fellow in 2008; receiving the Israel Chemical Society Prize (2008) and
the European Research Society (ERC) Advanced Research Grant (2008). In 2011, he was
elected to the Israel Academy of Sciences and Academia Europaea and became a Fellow of
the Royal Society of Chemistry. He was chosen to deliver the CNR Award Lecture (Indian
Chemical Res. Soc.) in 2012 and received the ChinaNano plenary lecture prize (2011). His
paper in Nanomaterials and Energy Journal received the best paper award of the year (in
2014) from the Institute of Civil Engineering (UK). Prof. Tenne is recipient of the Rothschild
Prize in Physics and Chemistry (2015).

Andrey N. Enyashin received his Ph.D. in chemistry at Ural
Technical State University (Yekaterinburg) in 2005. After a postdoc in
the theoretical chemistry group of Prof. Gotthard Seifert at the
Dresden University of Technology, Germany, he was appointed senior
researcher at the Institute of Solid State Chemistry of UB RAS
(Russian Federation). He was also a guest scientist at the
Donostia International Physics Center, Spain. He is a recipient of the
Samsung Electro-Mechanics Medal (2007). Currently, his field of
interest is computational materials science of inorganic and carbon
nanostructures.

IX

Preface
Inorganic Fullerene-Like Nanoparticles and
Inorganic Nanotubes
Fullerene-like nanoparticles (inorganic fullerenes; IF) and nanotubes of inorganic layered
compounds (inorganic nanotubes; INT) combine low dimensionality and nanosize, enhancing
the performance of corresponding bulk counterparts in their already known applications, as
well as opening new fields of their own. This issue gathers articles from the diverse area of
materials science and is devoted to fullerene-like nanoparticles and nanotubes of layered
sulfides and boron nitride, and collating the most current results obtained at the interface
between fundamental research and engineering.
Arising from a fortuitous lab discovery, the commercial production of inorganic hollow
nanoparticles was focused on molybdenum and tungsten disulfides. Their superior solid
lubrication effects have engendered intense industrial scale-up and commercialization, with sales
of thousands of tons of formulated lubricants per year. Yet, the search and evaluation of more
cost-effective and environmentally friendly manufacturing technologies continues. The paper by
Xu et al. published recently recent “Continuous Production of IF-WS2 Nanoparticles by a
Rotary Process” describes an attempt for further rationalization and scale-up of the
manufacturing of WS2 nanoparticles after gas–solid reductive sulfurization of WO3 nanoparticles
in a rotary furnace. This systematic study included the investigation of many reaction
parameters, such as precursor type, reaction temperature and time, and the reducing
atmosphere. This new technique could, in the future, become a successful alternative for
increasing the yield of IF production compared to the current fluidized-bed reactor.
The fullerene-like morphology of MoS2 and WS2 considerably improves the tribological
properties of these compounds, pushing ahead the large-scale use of layered sulfides in
machinery, aerospace and, in the future, also in medical industries as dry and oil-based
lubricants as well as wear-resistant surface coatings. Such applications require deep
understanding of the factors determining the mechanical and structural stability of inorganic
nanoparticles under extreme conditions of high pressure or intense irradiation. The study of
Cook et al. “Microstructural Study of IF-WS2 Failure Modes” explores the failure
mechanisms found in WS2 Ifs, treated with diverse pressure loading methods. The authors
uncover at least two distinct fracture modes, i.e., the collapse of quasi-spherical morphology
into agglomerated plate-like sheets and the delamination and exfoliation of the IF-WS2
nanoparticles. The latter process is accomplished by inductively-coupled radio-frequency
plasma irradiation of multiwall WS2 nanotubes, which is discussed in another paper “Single-to
Triple-Wall WS2 Nanotubes Obtained by High-Power Plasma Ablation of WS2 Multiwall
Nanotubes” in this issue. The authors were able to control the process of layer-by-layer
“undressing” of multilayered INTs and, in this manner, fabricate WS 2 nanotubes with ultrathin (one to three layers) walls.

X
The significant stability of hollow sulfide nanoparticles (IF/INT) under shock-wave
propagation or irradiation suggests their potential use as fillers for impact resilient polymer or
ceramic composites. Apparently, such applications of WS2 or MoS2 fullerene-like particles or
nanotubes may yield polymer composites with a high degree of crystallinity and,
consequently, with improved thermoplastic and mechanical properties, as demonstrated by
Naffakh et al. in the paper “Thermoplastic Polymer Nanocomposites Based on Inorganic
Fullerene-Like Nanoparticles and Inorganic Nanotubes”. However, in many cases, the
adhesion between a nanoparticle and the polymer matrix is limited to the weak van der Waals
interaction and could be enhanced by means of covalent bonding at the nanoparticle–polymer
interface. Surface functionalization of IFs or INTs, as reported by Raichmann et al. in “Design
of Experiments: Optimizing the Polycarboxylation/Functionalization of Tungsten Disulfide
Nanotubes”, could provide stronger adhesion of nanoparticles with the polymer matrix. The
emphasis in this direction was given to the non-trivial functionalization of the hydrophobic
WS2 nanotubes by hydrophilic carboxyl groups, which could further stimulate the fabrication
of hydrophilic polymer composites or ceramics.
Substantial progress in safe production and pioneering use of IFs and INTs has been made
possible due to the comprehensive experimental research of their formation conditions,
chemical activity, mechanical and electronic characteristics. However, novel and modified
nanoparticles of sulfides and other compounds, such as boron nitride, can provide a much
larger diversity of new materials in catalysis, electronics and electrochemistry, and their
detailed characterization is still required. In the paper “Nanostructured Boron Nitride: from
Molecular Design to Hydrogen Storage Application”, high-temperature spray-pyrolysis
synthesis of hollow-core BN nanoparticles was demonstrated. The synthesized nanoparticles
were carefully characterized and studied as a host for hydrogen storage applications.
Computational materials science can be a valuable tool for a preliminary study of this kind
of nanoparticles and this issue contains examples of theoretical papers describing
investigations of this type. For example, the paper “Gas-Phase and Microsolvated Glycine
Interacting with Boron Nitride Nanotubes: A B3LYP-D2* Periodic Study” examines the
adsorption of the amino-acid glycine on the surface of zig-zag BN nanotubes. Pure and
solvated glycine moieties have been investigated. In several cases, chemisorption was found
to be important, while in others ʌ-ʌ stacking, or through water molecules, was found to be
more relevant. In another study, nanotubes of noble-metal dichalcogenides were designed and
described as stable semiconductors in theoretical work by Zibouche et al.: “Noble-Metal
Chalcogenide Nanotubes”. It can guide experimental groups in researching fullerene-like
nanoparticles and nanotubes of other compounds. In another theoretical paper: “From Stable
ZnO and GaN Clusters to Novel Double Bubbles and Frameworks”, bottom–up construction
of hollow clusters (“bubbles”) of high-symmetry were systematically studied.
The subjects of the presented papers cover a wide range of challenges in the area of
inorganic fullerene-like nanoparticles and nanotubes. However, it can include only a few
comprehensive experimental and theoretical efforts, stepwise evaluating the rationalization of
the synthesis, and elucidation of the stability, mechanical, electronic and adhesive properties
of these nanostructures. We believe that this thematic issue can be helpful, not only for an

XI
advanced researcher to grasp the latest developments in this field, but also to permit a
beginner to gain a deeper insight into the field of inorganic fullerene-like nanoparticles and
nanotubes.

Reshef Tenne and Andrey N. Enyashin
Guest Editors

1

Electromechanical Properties of Small Transition-Metal
Dichalcogenide Nanotubes
Nourdine Zibouche, Mahdi Ghorbani-Asl, Thomas Heine and Agnieszka Kuc
Abstract: Transition-metal dichalcogenide nanotubes (TMC-NTs) are investigated for their
electromechanical properties under applied tensile strain using density functional-based methods.
For small elongations, linear strain-stress relations according to Hooke’s law have been obtained,
while for larger strains, plastic behavior is observed. Similar to their 2D counterparts, TMC-NTs
show nearly a linear change of band gaps with applied strain. This change is, however, nearly
diameter-independent in case of armchair forms. The semiconductor-metal transition occurs for
much larger deformations compared to the layered tube equivalents. This transition is faster for
heavier chalcogen elements, due to their smaller intrinsic band gaps. Unlike in the 2D forms, the
top of valence and the bottom of conduction bands stay unchanged with strain, and the zigzag
NTs are direct band gap materials until the semiconductor-metal transition. Meanwhile, the applied
strain causes modification in band curvature, affecting the effective masses of electrons and holes.
The quantum conductance of TMC-NTs starts to occur close to the Fermi level when tensile strain
is applied.
Reprinted from Inorganics. Cite as: Zibouche, N.; Ghorbani-Asl, M.; Heine, T.; Kuc, A.
Electromechanical Properties of Small Transition-Metal Dichalcogenide Nanotubes. Inorganics
2014, 2, 155–167.
1. Introduction
In the past few years, transition-metal dichalcogenides (TMCs) have become a class of materials
most widely investigated in the fields of physics, materials science or nanotechnology. Especially,
two-dimensional (2D) layered forms of TMCs are of great interest, as they can be easily manufactured
to monolayers using chemical or mechanical exfoliation and chemical deposition techniques [1–3].
They possess desirable intrinsic band gaps ranging from about 1.0 to 2.0 eV, and they were
utilized in nanoelectronic applications to produce field-effect transistors, logical circuits, amplifiers
and photodetectors [4–7]. The electronic properties of 2D TMCs can be tuned by various means,
including quantum confinement [8–11], mechanical deformations [12–14], electric fields [15,16] or
local defects [17–19].
Similar to carbon, tubular and fullerene-like nanostructures can be formed from other inorganic
materials, including sulfo-carbides [20], boron-carbon-nitrides [21] or TMCs [22,23]. Though less
than their carbon counterparts, in particular MoS2 and WS2 nano-onions and nanotubes have been
investigated both theoretically and experimentally [24–30]. TMC nanotubes (TMCs-NTs) behave
as exceptional lubricants [31,32], and it has been shown that when the MoS2 NTs or nano-onions
are added to base grease, the friction coefficient remains low, even at very high loads [33]. The
mechanical properties of TMC-NTs have been investigated experimentally, where tubes were subject

2
to tensile strain using atomic force microscopy [34–37]. Elastic deformations were predicted from
linear strain-stress relation up to the fracture point (at 13 GPa and 12% strain for WS2 TMC-NTs),
and fracture was directly related to the formation of local defects [35]. The mechanical properties
of WS2 NTs under axial tension and compression [36] shows that they are ultra-strong and elastic,
which distinguishes them from other known materials. Quantum-mechanical simulations showed
that under squeezing, MoS2 NTs start to form platelets, partially attached to the grips, which provide
good lubrication at the position of closest contact. This is interpreted as ‘nano-coating’.
Single-walled TMC-NTs have interesting electronic properties that depend on their diameter and
chirality. Zigzag (n,0) NTs are direct band gap semiconductors, resembling 1H TMC forms, while
armchair (n,n) NTs are indirect band gap materials, similar to the 2H TMC structures [24,29]. Zigzag
tubes are, therefore, suggested for luminescent devices, an application that would not be possible for
carbon NTs. With increasing tube diameter, the band gaps increase and eventually approach the
single-layer limit.
Doping inorganic semiconducting NTs may lead to new optoelectronic nanomaterials.
Ivanovskaya et al. [38] have investigated the effect of Mo to Nb doping on the electronic structure
of MoS2 NTs using the density functional based tight-binding (DFTB) method. It has been found
that composite Mo1−x Nbx S2 NTs are more stable than the corresponding mixture of pure tubes. This
effect was even stronger for larger tube diameters. The authors reported that all doped NTs were
metallic, independent of their chirality, diameters or the substitutional patterns. The density of states
close to the Fermi level of Nb-substituted MoS2 NTs can be tuned in a wide range by the degree
of doping.
Electromechanical properties have been widely investigated theoretically for TMC
monolayers [14,39,40], but they remain to be explored for the associated tubular structures. Because
of their excellent lubricating properties, the application of tensile stress on 2D TMC systems is rather
difficult in experiments. The experimental setup for direct tensile tests of TMC-NTs is, however,
state-of-the-art [35,37]. We have recently shown that the electronic properties of large-diameter
TMC-NTs can be tuned by an external tensile strain for nanoelectromechanical applications [41] and
that Raman spectroscopy is an ideal tool to monitor the strain of the individual tubes due to a linear
correlation between the Raman shift and the strain. These results hold, however, for large diameter
nanotubes. For small diameter tubes, finite size effects are expected.
In this work, we have investigated the electromechanical properties of small diameter TMC
nanotubes by applying axial tensile strain. Stress-strain relations, the electronic structure and
quantum conductance response to the mechanical deformations were compared between Mo- and
W-based NTs with different chalcogen atoms. The results were compared to the available
experimental and theoretical works. Our calculations show that up to 2%–5% elongations, the
stress-strain relations scale linearly, and we obtained Young’s moduli of about 200 GPa for armchair
and zigzag tubes, with the notable exception of much softer WTe2 . The shape of the band structures is
strongly affected, the conduction bands have a loose dispersion for zigzag tubes, while the dispersion
deepens for armchair materials. We find nearly a linear decrease in the band gap for all types of
nanotubes, and eventually, the semiconductor-metal transition occurs. This is, however, observed at

3
larger elongations than for the corresponding layered forms. Tensile strain enhances conductance
closer to the Fermi level.
2. Computational Details
We have investigated the (n, n) armchair and the (n, 0) zigzag TX2 nanotubes (T = Mo, W;
X = S, Se, Te) with n = 21 and 24 (see Figure 1). All structures were fully optimized (atomic positions
and lattice constants) employing helical boundary conditions as implemented in the Crystal09
software package [42].
Figure 1. Front and sided views of zigzag and armchair TX2 NT structures at equilibrium
and under tensile strain (ε).

For the strained structures, only the atomic coordinates have been re-optimized, while the unit cell
parameter along the tube axis was kept fixed, as reported in our previous works [14,41]. The tensile
strain is defined as ε = (L − L0 )/L0 , where L0 and L are equilibrium and strained lattice values,
respectively (cf. Figure 1). The elastic properties of the tubes under tensile stress were calculated as
force, F, acting on the area, A. The area can be calculated as follows:
A = 2πR0 δ

(1)

where R0 is the tube radius, defined as the distance between the center of the tube and the metal
atom, and δ is the thickness of tube wall, taken as the interlayer distance of the bulk material. The
Young’s modulus, Y , is obtained from the second derivative of the total energy with respect to the
applied strain at the equilibrium volume, V0 :
Y =

1 ∂ 2E
V0 ∂ε2

(2)

where V0 = AL0 .
Structural and electromechanical properties have been calculated using density functional theory
(DFT) in the representation by Perdew, Burke and Ernzerhof (PBE) [43], a method that was validated
for the TMC systems earlier [11,29,41]. The all-electron 86-311G* basis was chosen for sulfur
atoms, while for the heavier elements, the effective core potential (ECP) approach with large cores
was employed, accounting for scalar relativistic effects [44,45]. The shrinking factor was set to

4
eight, resulting in 5 k points in the irreducible Brillouin zone according to the Monkhorst–Pack
sampling [46]. Band structures were calculated along the high symmetry points using the Γ−X path.
The coherent electronic transport calculations were carried out using the density functional based
tight-binding (DFTB) [47–49] method in conjunction with the non-equilibrium Green’s function
technique [50,51] and the Landauer–Büttiker approach. The present approach was already validated
and described in detail in our previous works on various TMC materials [14,41,52].
3. Results and Discussion
We have calculated the electromechanical properties of TX2 nanotubes by applying tensile strain
(ε) to the tubes along their axis. Tensile strain causes changes in the geometry and results in the
elongation of the T–X bonds (see Figure 2). These bond lengths increase nearly linearly with ε and
are more sever for zigzag NTs. This trend is similar to the corresponding mechanical deformations
in the 2D TMC structures [14,41]. In nanotubes, one needs to distinguish between T–X bond lengths
in the outer and inner walls, the latter being slightly shorter. While for armchair NTs, outer and
inner bonds change in the same way, this is not the case for zigzag NTs. At larger ε, the inner bonds
undergo stronger elongations, eventually approaching the same values as for the outer bonds. We
obtain elongation of 0.8–1.0 pm and 0.5–0.6 pm per 1% of strain for the zigzag and armchair NTs,
respectively. Stronger elongations of bonds in the zigzag NTs can be understood, such that along the
tube, where the tensile strain is applied, there are many bonds oriented exactly parallel to the axis,
while this is not the case in armchair tubes. These bonds can be easily stretched, resulting also in a
reduction of the X–T–X angles.
Once the tubes are subject to ε, also the tube diameters change, namely they have to shrink to
compensate for the elongations along the tube axis. On average, the tube diameters shrink by 0.6 Å
and 0.2 Å per 1% of strain for zigzag and armchair NTs, respectively.
The stress-strain relations for all the studied tubes are shown in Figure 3. If the curves are fitted
to the harmonic approximation for small deformations according to Hooke’s law, the stress-strain
plots are linear, and the plastic deformations for larger strain values could not be observed. From
this fitting, however, we have obtained the Young’s moduli for all the tubes (see Table 1). Our
results are in agreement with the available experimental and theoretical values. For example, the
experimental Young’s modulus of multi-wall WS2 NTs is found to be 152 GPa [36], 171 GPa [34]
and 223 GPa [37]. For single-wall MoS2 NT ropes, the lowest measured Young’s modulus was
120 GPa [26], whereas theoretical values estimated from DFTB calculations for MoS2 NTs are 200
GPa [53] and 230 GPa [30,54]. Moreover, Li et al. [55] have reported 150 and 127 GPa for (6, 6)
and (10, 0) MoS2 NTs, respectively.
If the curves are fitted to a higher order polynomial (here, the fourth order polynomial was
chosen), we observe that already, for the ε of 3%–5% (for sulfides and selenides) or 2%–3% (for
tellurides), the curves deviate from linearity.

5
Figure 2. The metal-chalcogen bond length (T–X) change with the applied tensile
strain of exemplary MoS2 and WSe2 NTs. Similar linear changes are obtained for other
transition-metal dichalcogenide-NTs.
(24,0) MoS2
(24,24) MoS2

T-X / Å

2.7

Inner Wall
Outer Wall

2.6

2.5
(24,0) WSe2
(24,24) WSe2

2.4
0

0.1

0.2 0
ε = ΔL/L0

0.1

0.2

Figure 3. The calculated strain-stress relation of TX2 NTs under applied tensile strain
along the tube axis. Note the different scale on x- and y-axes of TTe2 NTs.

Table 1. The calculated Young’s moduli of all the studied NTs. The numbers are
obtained from the harmonic approximation following Hooke’s law for small values of
tensile strain.

System
MoS2
MoSe2
MoTe2
WS2
WSe2
WTe2

(21,0)
191
184
110
160
184
44

Chirality
(24,0) (21,21)
259
188
132
177
165
36

235
165
119
256
177
59

(24,24)
232
174
150
203
211
57

6
Changes in the geometry of TMC NTs under mechanical deformations also affect the electronic
structure of these materials. The band structure responses to the tensile strain are shown in Figures 4
and 5 for zigzag and armchair NTs, respectively. In the equilibrium, zigzag NTs are direct band
gap semiconductors at Γ, while armchair NTs are indirect band gap materials with the valence band
maximum (VBM) at Γ and the conduction band minimum (CBM) at 2/3 between Γ and X. This is
in close agreement with the DFTB calculations of Seifert et al. [24]. These features are unaffected
by ε; however, the CBM of armchair tubes shifts slightly towards the X point.
Figure 4. The calculated band structure response to the applied tensile strain of zigzag
TX2 NTs. (a) MoX2 and (b) WX2 .
ε = 10%

E-EF / eV

1

ε = 17%

(24,0) MoS2

2
Δ = 1.4 eV

0

Δ = 0.7 eV

ε = 10%

Δ = 0.4 eV

1

ε = 17%

(24,0) WS2

2

-1

Δ = 1.5 eV

EF

Δ = 0.7 eV

0

Δ = 0.4 eV

-1
(24,0) MoSe2

2
1
Δ = 1.1 eV
0
-1
Γ

Δ = 0.4 eV

X|Γ

(24,0) WSe2

2
Δ = 0.1 eV

E-EF / eV

E-EF / eV

ε = 0%

b)

EF

E-EF / eV

ε = 0%

a)

X|Γ

Δ = 0.1 eV
1
Δ = 0.9 eV

-1
Γ

X

Δ = 0.4 eV

0

X|Γ

X|Γ

X

Figure 5. The calculated band structure response to the applied tensile strain of armchair
TX2 NTs. (a) MoX2 and (b) WX2 .
ε = 10%

E-EF / eV

1

ε = 17%

Δ = 1.7 eV

Δ = 0.9 eV

Δ = 0.5 eV

-1

EF

Δ = 1.8 eV
Δ = 1.0 eV

0

Δ = 0.0 eV

Δ = 1.4 eV

Δ = 0.6 eV

Δ = 0.3 eV

0

X|Γ

(24,24) WSe2

2
E-EF / eV

E-EF / eV

1

ε = 17%

-1
(24,24) MoSe2

2

-1
Γ

ε = 10%
(24,24) WS2

2

0

1

ε = 0%

b)

EF

(24,24) MoS2

2

E-EF / eV

ε = 0%

a)

X|Γ

X

1
Δ = 1.5 eV

Δ = 0.4 eV

0
-1
Γ

X|Γ

Δ = 0.0 eV

X|Γ

X

Both the valence and the conduction bands are affected by the mechanical deformations. While
in the zigzag NTs, the CBM looses its dispersion with the applied strain, it is opposite in the case
of armchair tubes. In the latter, the dispersion deepens, and the CBM position shifts towards the X
point. The valence bands get more dispersed along the Γ − X paths for larger deformations, and this
is chirality-independent.
The band gap evolution with the tensile strain is shown in Figure 6. Nearly linear scaling is found
for ε of 10%–12%. The semiconductor-metal transition occurs for elongations much larger than in

7
the case of layered 2D forms [14,41], but it is faster for the NTs with heavier chalcogen atoms, as
they have a smaller intrinsic band gap. We have noticed that for armchair NTs, there is almost no
band gap dependency on the tube diameter for the whole range of ε, as it is in the zigzag forms.
Figure 6. The calculated band gap evolution with the applied tensile strain of zigzag and
armchair TX2 NTs.
2.0

Δ / eV

1.0

0.0
2.0

1.0

0.0

0.0

0.1

(24,0) MoS2

(24,24) MoS2

(21,0) MoS2

(21,21) MoS2

(24,0) MoSe2

(24,24) MoSe2

(21,0) MoSe2

(21,21) MoSe2

(24,0) MoTe2

(24,24) MoTe2

(21,0) MoTe2

(21,21) MoTe2

(24,0) WS2

(24,24) WS2

(21,0) WS2

(21,21) WS2

(24,0) WSe2

(24,24) WSe2

(21,0) WSe2

(21,21) WSe2

(24,0) WTe2

(24,24) WTe2

(21,0) WTe2

(21,21) WTe2

0.2

0.0
ε = ΔL/L0

0.1

0.2

2.0

1.0

0.0
2.0

1.0

0.0

Figure 7 shows the intrinsic quantum conductance (G) calculated along the TX2 NTs with respect
to the applied tensile strain. As the materials are stretched along the tube axis, G starts to appear
closer to the Fermi level, and eventually, the transport channel opens. For all NTs, the conductance
below the Fermi level reduces with strain; however, above the EF , it stays unchanged (it increases) for
zigzag (armchair) NTs. Our quantum transport calculations aim to describe the intrinsic conductance
of the entire tubes along their principal symmetry axis. Note that the quantum transport calculations
are carried out using the DFTB method, which tends to overestimate the electronic band gap.
We have calculated the effective masses of electrons and holes at the CBM and VBM, respectively
(see Table 2). The effective masses of holes are reduced with the tensile strain, which is consistent
with stronger dispersion in the VBM. The masses of electrons at the equilibriums are similar for
zigzag and armchair NTs of the same type. These numbers are larger for heavier chalcogen atoms,
similar to the electron effective masses. We expect the effective masses of electrons to increase
(decrease) for zigzag (armchair) NTs with ε as the dispersion of bands decreases (increases). The
effective masses are calculated from the harmonic approximation and fitting to the energy point close
to the VBM and CBM. This, therefore, strongly depends on the number of k-points along the path
in the Brillouin zone. We have chosen very fine k-point sampling of 150 points between Γ and X.
Thus, we do not observe as clear trends as expected in the effective masses of electrons. For the (6,6)
and (10,0) MoS2 NTs, Li et al. [55] have obtained effective masses of electrons and holes of 0.53,
0.51, 0.83 and 1.55, for armchair and zigzag forms, respectively.

8
Figure 7. The electron quantum conductance of TX2 NTs under applied tensile strain
along the tube axis.
MoS2
(21,0)

0.0 %
9.1 %
15.2 %

(21,21)

9

18

12

6

12

6

6

3

6

3

0

0
(24,24)

24

(24,0)

12

9

0

0
(24,0)

(24,24)

24

12

9

18

9

12

6

12

6

6

3

6

-1

0

1

2 -2
-1
Energy / eV

0

1

0
2

0
-2

3
-1

0

1

WS2

2 -2
-1
Energy / eV

0

1

0
2

WSe2
32

12

24

16

8

16

8

8

4

8

4

0
32

0
16

24

(24,24)

(24,0)

2 -1

16

(21,0)

0.0 % (21,21)
9.5 %
17.1 %

G / 2e h

32

2 -1

12

18

0
-2

G / 2e h

(21,0)

24

2 -1

2 -1

MoSe 2

(21,21)

12

18

G / 2e h

0.0 %
10.1 %
17.6 %

G / 2e h

24

0.0 %
9.1 %
15.2 %

0
32

(21,21)

(21,0)

16
12

(24,24)

(24,0)

0
16

24

12

24

12

16

8

16

8

8

4

8

4

0
-2

-1

0

1

2 -2
-1
Energy / eV

0

1

0
2

0
-2

-1

0

1

2 -2
-1
Energy / eV

0

1

0
2

Table 2. The calculated effective electron and hole masses (in m0 units) of TX2 NTs
with respect to the applied tensile strain (ε). For zigzag NTs, both effective masses are
calculated at the Γ point; for armchair NTs, effective masses of holes are calculated at Γ
and of electrons between Γ and X. Note, for the latter, we do not specify the exact k-point
as the conduction band minimum (CBM) shifts with ε. The negative values of hole
effective masses come from the band curvature at the valence band maximum (VBM).

System

Chirality

Electron masses

Hole masses

0%

5%

10%

0%

5%

10%

MoS2

(24,24)
(24,0)

0.621
0.655

0.631
0.621

0.707
0.660

−6.209
−10.03

−1.684
−1.990

−0.974
−1.033

MoSe2

(24,24)
(24,0)

0.868
1.092

0.938
0.998

0.884
0.739

−15.366
−0.485

−2.233
−3.079

−1.140
−3.461

WS2

(24,24)
(24,0)

0.434
0.475

0.457
0.434

0.524
0.467

−11.498
−12.348

−1.934
−3.171

−0.498
−1.270

WSe2

(24,24)
(24,0)

0.621
0.585

0.715
0.714

0.659
0.916

−15.147
−1.362

−2.285
−2.846

−7.922
−1.396

9
4. Conclusions
We have investigated the electromechanical properties of inorganic nanotubes of the TX2 -type
under applied tensile strain. The tubes undergo changes in geometry, namely the T–X bond lengths
are elongated. More pronounced changes are obtained for the zigzag NTs, in which some of the
bonds are oriented parallel to the tube axis, which means along the acting deformation force. The
stress-strain relation fitted to the harmonic approximation for small deformations gives Young’s
moduli of about 200 GPa, with the exception of WTe2 , which produces notably smaller values around
50 GPa.
The electronic properties and the quantum transport are particularly affected by mechanical
deformations. Nearly a linear change in the band gap is observed for elongations up to 12%. The
semiconductor-metal transition is eventually obtained for all type of tubes; however, it is much faster
for heavier chalcogen atoms. Nanotubes require larger tensile strain to become metallic than the
corresponding 2D materials.
The dispersion of the valence and conduction bands changes strongly with the applied strain.
Notably, the VBM deepens the dispersion, which results in the lowering of the hole effective masses.
The CBM is chirality dependent, and the dispersion is lost (enhanced) for zigzag (armchair) NTs.
The transport channels start to open closer to the Fermi level for larger ε.
The electronic properties and the possibility to tune by tensile strain suggest that inorganic
NTs, such as TMC materials, could be considered in nanoelectronic applications, for example as
switching materials.
Acknowledgments
This work was supported by the German Research Council (Deutsche Forschungsgemeinschaft,
HE 3543/18-1), the European Commission (FP7-PEOPLE-2009-IAPP QUASINANO, GA 251149
and FP7-PEOPLE-2012-ITN MoWSeS, GA 317451).
Author Contributions
N. Zibouche, M. Ghorbani-Asl, A. Kuc and T. Heine generated, analyzed and discussed the
results. T. Heine conceived of this project. All authors contributed in writing this paper.
Conflicts of Interest
The authors declare no conflict of interest.
References
1. Coleman, J.N.; Lotya, M.; O’Neill, A.; Bergin, S.D.; King, P.J.; Khan, U.; Young, K.;
Gaucher, A.; De, S.; Smith, R.J.; et al. Two-Dimensional Nanosheets Produced by Liquid
Exfoliation of Layered Materials. Science 2011, 331, 568–571.

10
2. Novoselov, K.S.; Jiang, D.; Schedin, F.; Booth, T.J.; Khotkevich, V.V.; Morozov, S.V.; Geim,
A.K. Two-dimensional atomic crystals. Proc. Nat. Acad. Sci. USA 2005, 102, 10451–10453.
3. Lee, Y.H.; Zhang, X.Q.; Zhang, W.; Chang, M.T.; Lin, C.T.; Chang, K.D.; Yu, Y.C.;
Wang, J.T.S.W.; Chang, C.S.; Li, L.J.; et al. Synthesis of Large-Area MoS2 Atomic Layers
with Chemical Vapor Deposition. Adv. Mater. 2012, 24, 2320–2325.
4. Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A. Single-layer MoS2
transistors. Nat. Nanotechnol. 2011, 6, 147–150.
5. Radisavljevic, B.; Whitwick, M.B.; Kis, A. Integrated Circuits and Logic Operations Based on
Single-Layer MoS2 . ACS Nano 2011, 5, 9934–9938.
6. Radisavljevic, B.; Whitwick, M.B.; Kis, A. Small-signal amplifier based on single-layer MoS2 .
Appl. Phys. Lett. 2012, 101, 043103:1–043103:4.
7. Lopez-Sanchez, O.; Lembke, D.; Kayci, M.; Radenovic, A.; Kis, A. Ultrasensitive
photodetectors based on monolayer MoS2 . Nat. Nanotechnol. 2013, 8, 497–501.
8. Li, T.S.; Galli, G.L. Electronic properties of MoS2 nanoparticles. J. Phys. Chem. C 2007, 111,
16192–16196.
9. Splendiani, A.; Sun, L.; Zhang, Y.B.; Li, T.S.; Kim, J.; Chim, C.Y.; Galli, G.; Wang, F.
Emerging Photoluminescence in Monolayer MoS2 . Nano Lett. 2010, 10, 1271–1275.
10. Ramakrishna Matte, H.S.S.; Gomathi, A.; Manna, A.K.; Late, D.J.; Datta, R.; Pati, S.K.;
Rao, C.N.R. MoS2 and WS2 Analogues of Graphene. Angew. Chem. Int. Ed. 2010, 49,
4059–4062.
11. Kuc, A.; Zibouche, N.; Heine, T. Influence of quantum confinement on the electronic structure
of the transition metal sulfide TS2 . Phys. Rev. B 2011, 83, 245213:1–245213:4.
12. Bertolazzi, S.; Brivio, J.; Kis, A. Stretching and Breaking of Ultrathin MoS2 . ACS Nano 2011,
5, 9703–9709.
13. Castellanos-Gomez, A.; Poot, M.; Steele, G.A.; van der Zant, H.S.J.; Agrait, N.; Rubio-Bollinger, G.
Elastic Properties of Freely Suspended MoS2 Nanosheets. Adv. Mater. 2012, 24, 772–775.
14. Ghorbani-Asl, M.; Borini, S.; Kuc, A.; Heine, T. Strain-dependent modulation of conductivity
in single layer tranisition metal dichalcogenides. Phys. Rev. B 2013, 87, 235434:1–235434:6.
15. Ramasubramaniam, A.; Naveh, D.; Towe, E. Tunable band gaps in bilayer transition-metal
dichalcogenides. Phys. Rev. B 2011, 84, 205325:1–205325:10.
16. Liu, Q.; Li, L.; Li, Y.; Gao, Z.; Chen, Z.; Lu, J. Tuning Electronic Structure of Bilayer
MoS2 by Vertical Electric Field: A First-Principles Investigation. J. Phys. Chem. C 2012,
116, 21556–21562.
17. Enyashin, A.N.; Bar-Sadan, M.; Houben, L.; Seifert, G. Line Defects in Molybdenum Disulfide
Layers. J. Phys. Chem. C 2013, 117, 10842–10848.
18. Van der Zande, A.M.; Huang, P.Y.; Chenet, D.A.; Berkelbach, T.C.; You, Y.; Lee, G.H.;
Heinz, T.F.; Reichman, D.R.; Muller, D.A.; Hone, J.C. Grains and grain boundaries in highly
crystalline monolayer molybdenum disulphide. Nat. Mater. 2013, 12, 554–561.
19. Ghorbani-Asl, M.; Enyashin, A.N.; Kuc, A.; Seifert, G.; Heine, T. Defect-induced conductivity
anisotropy in MoS2 monolayers. Phys. Rev. B 2013, 88, 245440:1–245440:7.

11
20. Goyenola, C.; Gueorguiev, G.K.; Stafström, S. Fullerene-like CSx : A first-principles study of
synthetic growth. Chem. Phys. Lett. 2011, 506, 86–91.
21. Hellgren, N.; Berlind, T.; Gueorguiev, G.K.; Johansson, M.P.; Stafström, S.; Hultman, L.
Fullerene-like B-C-N thin films: A computational and experimental study. Mater. Sci. Eng. B
2004, 113, 242–247.
22. Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and cylindrical structures of tungsten
disulfide. Nature 1992, 360, 444–446.
23. Margulis, L.; Salitra, G.; Tenne, R.; Talianker, M. Nested Fullerene-like Structures. Nature
1993, 365, 113–114.
24. Seifert, G.; Terrones, H.; Terrones, M.; Jungnickel, G.; Frauenheim, T. Structure and electronic
properties of MoS2 nanotubes. Phys. Rev. Lett. 2000, 85, 146–149.
25. Seifert, G.; Kohler, T.; Tenne, R. Stability of Metal Chalcogenide Nanotubes. J. Phys. Chem.
B 2002, 106, 2497–2501.
26. Kis, A.; Mihailovic, D.; Remskar, M.; Mrzel, A.; Jesih, A.; Piwonski, I.; Kulik, A.J.;
Benoit, W.; Forro, L. Shear and Young’s moduli of MoS2 nanotube ropes. Adv. Mater. 2003,
15, 733–736.
27. Milosevic, I.; Nikolic, B.; Dobardzic, E.; Damnjanovic, M.; Popov, I.; Seifert, G. Electronic
properties and optical spectra of MoS2 and WS2 nanotubes. Phys. Rev. B 2007, 76, 233414.
28. Tenne, R.; Redlich, M. Recent progress in the research of inorganic fullerene-like nanoparticles
and inorganic nanotubes. Chem. Soc. Rev. 2010, 39, 1423–1434.
29. Zibouche, N.; Kuc, A.; Heine, T. From layers to nanotubes: Transition metal disulfides TMS2 .
Eur. Phys. J. B 2012, 85, 1–7.
30. Lorenz, T.; Teich, D.; Joswig, J.O.; Seifert, G. Theoretical Study of the Mechanical Behavior
of Individual TiS2 and MoS2 Nanotubes. J. Phys. Chem. C 2012, 116, 11714–11721.
31. Mattheis, L.F. Energy-Bands For 2H-NbSe2 And 2H-MoS2 . Phys. Rev. Lett. 1973,
30, 784–787.
32. Drummond, C.; Alcantar, N.; Israelachvili, J.; Tenne, R.; Golan, Y. Microtribology and
friction-induced material transfer in WS2 nanoparticle additives. Adv. Funct. Mater. 2001,
11, 348–354.
33. Stefanov, M.; Enyashin, A.N.; Heine, T.; Seifert, G. Nanolubrication: How Do MoS2 -Based
Nanostructures Lubricate? J. Phys. Chem. C 2008, 112, 17764–17767.
34. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Seifert, G.; Tenne, R.
Mechanical behavior of individual WS2 nanotubes. J. Mater. Res. 2004, 19, 454–459.
35. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Ivanovskaya, V.; Heine, T.;
Seifert, G.; Wagner, H.D.; Tenne, R. Direct tensile tests of individual WS2 nanotubes. In
Proceedings of the Pricm 5: The Fifth Pacific Rim International Conference on Advanced
Materials, Pts 1–5, Beijing, China, 2–5 November 2004; Volume 475–479, pp. 4097–4102.

12
36. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Ivanovskaya, V.V.; Heine, T.;
Seifert, G.; Wagner, H.D.; Tenne, R. On the Mechanical Behavior of Tungsten Disulfide
Nanotubes under Axial Tension and Compression. Proc. Natl. Acad. Sci. USA 2006,
103, 523–528.
37. Tang, D.M.; Wei, X.; Wang, M.S.; Kawamoto, N.; Bando, Y.; Zhi, C.; Mitome, M.; Zak, A.;
Tenne, R.; Golberg, D. Revealing the Anomalous Tensile Properties of WS2 Nanotubes by in
Situ Transmission Electron Microscopy. Nano Lett. 2013, 13, 1034–1040.
38. Ivanovskaya, V.V.; Heine, T.; Gemming, S.; Seifert, G. Structure, stability and electronic
properties of composite Mo1−x Nbx S2 nanotubes. Phys. Stat. Solidi B 2006, 243, 1757–1764.
39. Yun, W.S.; Han, S.W.; Hong, S.C.; Kim, I.G.; Lee, J.D. Thickness and strain effects on
electronic structures of transition metal dichalcogenides: 2H-MX2 semiconductors (M = Mo,
W; X = S, Se, Te). Phys. Rev. B 2012, 85, 033305.
40. Scalise, E.; Houssa, M.; Pourtois, G.; Afanas’ev, V.; Stesmans, A. Strain-induced
semiconductor to metal transition in the two-dimensional honeycomb structure of MoS2 .
Nano Res. 2012, 5, 43–48.
41. Ghorbani-Asl, M.; Zibouche, N.; Wahiduzzaman, M.; Oliveira, A.F.; Kuc, A.; Heine,
T. Electromechanics in MoS2 and WS2 : Nanotubes vs. monolayers.
Sci. Rep. 2013, 3,
2961:1–2961:8.
42. Dovesi, R.; Saunders, V.R.; Roetti, R.; Orlando, R.; Zicovich-Wilson, C.M.; Pascale, F.;
Civalleri, B.; Doll, K.; Harrison, N.M.; Bush, I.J.; et al. CRYSTAL09 User’s Manual; University
of Torino: Torino, Italy, 2009.
43. Perdew, J.P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple.
Phys. Rev. Lett. 1996, 77, 3865–3868.
44. Cora, F.; Patel, A.; Harrison, N.M.; Roetti, C.; Catlow, C.R.A. An ab initio Hartree-Fock study
of alpha-MoO3 . J. Mater. Chem. 1997, 7, 959.
45. Cora, F.; Patel, A.; Harrison, N.M.; Dovesi, R.; Catlow, C.R.A. An ab initio Hartree-Fock
study of the cubic and tetragonal phases of bulk tungsten trioxide. J. Am. Chem. Soc. 1996,
118, 12174–12182.
46. Monkhorst, H.J.; Pack, J.D. Special Points For Brillouin-Zone Integrations. Phys. Rev. B
1976, 13, doi:10.1103/PhysRevB.13.5188.
47. Seifert, G.; Porezag, D.; Frauenheim, T. Calculations of molecules, clusters, and solids with
a simplified LCAO-DFT-LDA scheme. Int. J. Quantum Chem. 1996, 58, 185–192.
48. Oliveira, A.F.; Seifert, G.; Heine, T.; Duarte, H.A. Density-Functional Based Tight-Binding:
An Approximate DFT Method. J. Br. Chem. Soc. 2009, 20, 1193–1205.
49. Wahiduzzaman, M.; Oliveira, A.F.; Philipsen, P.; Zhechkov, L.; van Lenthe, E.; Witek, A.H.;
Heine, T. DFTB Parameters for the Periodic Table: Part 1, Electronic Structure. J. Chem.
Theory Comput. 2013, 9, 4006–4017.
50. Di Carlo, A.; Gheorghe, M.; Bolognesi, A.; Lugli, P.; Sternberg, M.; Seifert, G.;
Frauenheim, T. Molecular Devices Simulations Based on Density Functional Tight-Binding.
J. Comput. Electron. 2002, 1, 109–112.

13
51. Datta, S. Quantum Transport: Atom to Transistor, 2 ed.; Cambridge University Press:
Cambridge, UK; New York, NY, USA, 2005.
52. Ghorbani-Asl, M.; Juarez-Mosqueda, R.; Kuc, A.; Heine, T. Efficient Quantum Simulations
of Metals within the Γ-Point Approximation: Application to Carbon and Inorganic 1D and 2D
Materials. J. Chem. Theory Comput. 2012, 8, 2888–2895.
53. Li, T. Ideal strength and phonon instability in single-layer MoS2 . Phys. Rev. B 2012,
85, 235407.
54. Zhang, D.B.; Dumitrica, T.; Seifert, G. Helical Nanotube Structures of MoS2 with Intrinsic
Twisting: An Objective Molecular Dynamics Study.
Phys. Rev. Lett. 2011, 107,
doi:10.1103/PhysRevLett.107.109902.
55. Li, W.; Zhang, G.; Guo, M.; Zhang, Y.W. Strain-tunable electronic and transport properties of
MoS2 nanotubes. Nano Res. 2014, accepted.

14

Single- to Triple-Wall WS2 Nanotubes Obtained by
High-Power Plasma Ablation of WS2 Multiwall Nanotubes
Volker Brüser, Ronit Popovitz-Biro, Ana Albu-Yaron, Tommy Lorenz, Gotthard Seifert,
Reshef Tenne and Alla Zak
Abstract: The synthesis of inorganic nanotubes (INT) from layered compounds of a small size
(<10 nm in diameter) and number of layers (<4) is not a trivial task. Calculations based on density
functional tight-binding theory (DFTB) predict that under highly exergonic conditions, the reaction
could be driven into a “window” of (meta-) stability, where 1–3-layer nanotubes will be formed.
Indeed, in this study, single- to triple-wall WS2 nanotubes with a diameter of 3–7 nm and a length of
20–100 nm were produced by high-power plasma irradiation of multiwall WS 2 nanotubes. As target
materials, plane crystals (2H), quasi spherical nanoparticles (IF) and multiwall, 20–30 layers, WS2
nanotubes were assessed. Surprisingly, only INT-WS2 treated by plasma resulted in very small, and
of a few layers, “daughter” nanotubules. The daughter nanotubes occur mostly attached to the outer
surface of the predecessor, i.e., the multiwall “mother” nanotubes. They appear having either a
common growth axis with the multiwall nanotube or tilted by approximately 30° or 60° with respect
to its axis. This suggests that the daughter nanotubes are generated by exfoliation along specific
crystallographic directions. A growth mechanism for the daughter nanotubes is proposed. High
resolution transmission and scanning electron microscopy (HRTEM/HRSEM) analyses revealed the
distinctive nanoscale structures and helped elucidating their growth mechanism.
Reprinted from Inorganics. Cite as: Brüser, V.; Popovitz-Biro, R.; Albu-Yaron, A.; Lorenz, T.;
Seifert, G.; Tenne, R.; Zak, A. Single- to Triple-Wall WS2 Nanotubes Obtained by High-Power
Plasma Ablation of WS2 Multiwall Nanotubes. Inorganics 2014, 2, 177–190.
1. Introduction
Multiwall inorganic nanotubes of WS2 (INT-WS2) were discovered in 1992 [1], and the route for
their scaled-up synthesis was developed in 2009 [2]. Together with BN [3] and MoS2 [4,5], they
probably constitute the most investigated kind of inorganic nanotubes from layered compounds. The
crystalline and electronic structure of INT has been studied in great detail [6–8]. In particular,
calculations have shown that multiwall WS2 (MoS2) nanotubes become more stable than the
respective nanosheets at a threshold outer diameter of about 15 to 20 nm and being made up of at
least 5–10 layers [9]. Indeed, many of the high-temperature (above 700 °C) synthetic strategies ended
up in multiwall nanotubes exhibiting a high-crystalline order, which agree quite well with the
predicted sizes [2,10,11].
Nonetheless, these conditions are not sufficiently exergonic to drive the reaction into windows of
(meta-) stability far enough from equilibrium, where 1–3-layer nanotubes could be formed. It was
shown in the past that reactions carried out under highly exergonic conditions, like laser ablation [12],
for example, can yield closed-cage MoS2 nanoparticles having a small size and number of layers.
Calculations based on density functional tight-binding theory (DFTB) [9] (see Figure 1) present the

15
energy-per-atom of nanotubes as a function of the number of atoms in the unit length (unit cell), Ntot,
and for different number of layers (k = 1–4). They are compared with nanostripes (nanoribbons) of
the same number of atoms. For the sake of simplicity, the calculations were carried out for MoS2,
which is structurally analogous to WS2. It is noticed that the energy-per-atom increases with a
decreasing number of atoms for both the nanostripes and the nanotubes, but for different reasons.
The energy-per-atom for the nanostripes increases, due to edge effects, i.e., the abundance of rim
atoms with dangling bonds. On the other hand, the nanotubes become less stable at a smaller radius
of curvature, due to the increasing elastic energy of folding. In addition, the folding energy increases
more steeply for the nanotubes than the energy of the nanoribbons as the number of atoms shrinks.
Consequently, smaller diameter nanotubes become less stable than the straight nanostripes to the left
of the cross-over point (stability threshold) of the two curves. While the cross-over point itself moves
to the left as the number of layers decreases, the corresponding threshold energy-per-atom rapidly
increases (becomes less negative), particularly for nanotubes with three layers and below. It is
therefore clear that the generation of nanotubes of a small size and number of layers (k < 4) requires
highly exergonic conditions, which is the subject of the present work.
Figure 1. The calculated energy-per-atom for MoS2 nanotubes and nanostripes with
1–4 walls as a function of the number of atoms in the tube unit cell, Ntot.

Interestingly, in the range of ~390 < Ntot < 670, which corresponds to nanotubes with outer
diameters of 5.1 nm < D3 < 8.0 nm, the triple layer nanotubes are more stable than nanotubes with
k = 2 and k = 4 (see Figure 1). The diameter (Dk) represents here the outer diameters of the nanotubes
with k shells. This theoretical prediction is in agreement with experimental results presented in this
work: the majority of the daughter nanotubes were triple-walled. Note that nanotubes with the same
(outer) diameters, but different number of shells, have consequently a different (total) number of

16
atoms. Thus, a single-wall tube with a larger diameter may have less atoms than triple-walled tubes
of a smaller diameter.
A similar situation has been encountered with the stability window of MoS2 nanotetrahedra and
nanooctahedra consisting of 2–4 layers. These nanostructures were proposed first in [13,14] and
realized in [15,16]. Indeed, MoS2 nanooctahedra/nanotetrahedra were obtained by rapid quenching
of laser- [15–18] or solar- [19] ablated MoS2 soot or by an arc-discharge process [20]. It can,
therefore, be concluded that highly exergonic reaction conditions and rapid quenching of the
nanoclusters can access (meta-) stability windows, which favor new nanotubes that are not reachable
by the conventional thermally-driven synthesis at <1000 °C.
2. Results and Discussion
In the present work, 1–3-layer WS2 nanotubes with a diameter of 3–7 nm and a length of
20–100 nm were produced by applying inductively coupled radio-frequency plasma irradiation on
multiwall INT-WS2.
2.1. Scanning and Transmission Electron Microscopy Analysis
Typical scanning and transmission electron microscopy images of a pristine (untreated) multiwall
WS2 (“mother”) nanotube are presented in Figure 2a,b, respectively. The majority of the predecessor
INT was 5–20 microns in length and 30–120 nm in diameter.
The HRTEM images in Figure 3a,b display the range of daughter nanotubes obtained by plasma-treating
of the multiwall WS2 nanotubes at 600 W for 40 min: tiny daughter nanotubes adjacent to the outer
surface (Figure 3a) of the mother nanotube and a few isolated daughter nanotubes (Figure 3b). The
amount of such daughter nanotubes increased with the treatment time from 10 to 40 min at 400 W
plasma power. The extension of the plasma treatment time to 80 min did not reveal any additional
improvement; however, the increase of plasma power from 400 to 600 W resulted in a sharp increase
in the amount of the daughter nanotubes.
At 600 W and 40 min of treatment, a rough statistical estimate shows that daughter nanotubes
were attached to about 80% of the plasma treated multiwall nanotubes. In comparison, only 10–20%
of the multiwall WS2 nanotubes were covered with daughter nanotubes by a 400-W plasma treatment.
Some nanostructures could be better described as nanoscrolls. However, the majority of the daughter
nanostructures are nanotubes, having at least one perfectly closed layer. Future work will be focused
on devising this technique to increase the yield of a single- to a few-layer nanotubes of WS2 or other
INT, as well. Indeed, by irradiating MoS2 powder with a focused solar beam, single-wall MoS2 was
rarely observed in the processed powder [21], which confirms that highly exergonic conditions
produced by focused solar (laser) ablation may lead to the production of single-wall nanotubes of
this kind.

17
Figure 2. (a) SEM and (b) TEM micrograph of a pristine multiwall WS2 nanotube.

Figure 3. TEM images of daughter WS2 nanotubes obtained by plasma ablation of
multiwall inorganic nanotube (INT)-WS2 at 600 W for 40 min: (a) A large number of
daughter nanotubes next to a treated multiwall nanotube; (b) A group of daughter
nanotubes isolated from plasma-treated multiwall WS2 nanotubes by sonication.

18
Moreover, many daughter nanotubes were found attached and being tilted at ca. 30 or 60° with
respect to the mother nanotube axis (see the white arrows in Figure 4a,b). However, some daughter
nanotubes were found to be attached, having a common growth axis with the multiwall nanotube
(see the black arrow in Figure 4b).
Figure 4. (a) TEM images of the daughter nanotubes tethered to the surface and tilted at
approximately 30 or 60° with respect to the mother nanotube growth axis (white arrows);
(b) TEM image of a daughter nanotube with growth axis parallel to the mother nanotube
(black arrows).

These observations suggest that the small nanotubes were exfoliated by unzipping the outer
walls of the mother nanotubes along specific crystallographic directions. In addition to
nanotubes/nanoscrolls, a few layers-thick WS2 nanoplatelets of typical sizes in the range of the
nanotubes’ length, i.e., 50–100 nm, were also observed.
In an attempt to separate the daughter nanotubes from the mother nanotubes, the plasma-ablated
WS2 nanotube powder was ultrasonically treated in ethanol for 10 min. The high-resolution TEM
(HRTEM) images in Figure 5 (see also Figure 3b) clearly depict the daughter nanotubes being more
easily observed after detachment from the predecessor nanotubes.
Rough statistical analysis revealed that the interlayer distance in most daughter nanotubes varied
between 6.3–6.5 Å (see Figure 5a), which is larger than the interlayer spacing of 2H-WS2
(6.23 Å/2-Theta = 14.32°) and multiwall nanotubes (6.31 Å/2-Theta = 14.13°) [2]. This observation
suggests that the daughter nanotubes were not fully relaxed during the growth process and that the
annealing of the sample could possibly lead to further structural relaxation. The Fourier (FFT)
analysis (see the inset in Figure 5b) of the area framed by the square shows that the nanotube is chiral
with a helical angle of six degrees. Once daughter nanotubes are observed in larger yields, techniques
like ultracentrifugation could be used to separate them according to the number of layers and
the length.

19
Figure 5. (a) HRTEM images of a two–three-layer nanotube with a non-uniform
diameter after detachment from the large WS2 multiwall nanotube; (b) Another
three-layer daughter nanotube after detachment. The Fourier (FFT) analysis (see the
inset) of the area framed by the square shows that the nanotube is chiral with a helical
angle of six degrees.

Energy dispersive X-ray analysis (EDS) within the TEM (not shown) confirmed that the
nanotubes are made solely of tungsten and sulfur. Negligible traces of oxygen were found, which
could be mainly attributed to surface impurities.
In a separate series of experiments, several powders of different microcrystalline layered
materials, including 2H-WS2, 2H-MoS2 and 2H-NbS2, and also the respective diselenides, received
a similar plasma treatment. No daughter nanotubes were found in these treated samples, whatsoever.
A few layers-thick WS2 nanoplatelets with typical sizes in the range of the nanotubes (50–100 nm)
were nevertheless abundant in plasma-treated 2H-WS2 powder. Furthermore, the NbSe2 powder
turned out to be unstable under the plasma treatment conditions. Plasma treatment (400 W) of
fullerene-like WS2 nanoparticles with hollow cage structure (inorganic fullerene-like (IF-WS2)
nanoparticles) resulted in a few layers exfoliation and a few small-sized (“daughter”) fullerene-like
nanoparticles or nanotubes (see Figure 6). The daughter IF-WS2 nanoparticles are reminiscent of the
arc-discharge produced IF-MoS2 nanoparticles [20]. In concluding this large series of experiments,
it is possible to state that only plasma irradiation of multiwall WS2 nanotubes yielded daughter
nanotubes in a reproducible fashion.

20
Figure 6. TEM image of the attached daughter single wall fullerene-like nanoparticles
generated by plasma treatment of multiwall fullerene-like WS2 nanoparticles.

Single-wall carbon nanotubes can be obtained in large quantities, e.g., via the arc-discharge
technique [22]. Given the interlayer distance of 3.4 Å in graphite, the monoatomic graphene plane
can be closed into nanotubes of a diameter smaller than 0.5 nm [23,24]. On the other hand, the WS2
(MoS2) layer consists of six-fold bonded tungsten (molybdenum) atoms sandwiched between two
sub-layers of three-fold bonded sulfur atoms. This makes the WS2 layers pretty rigid, with interlayer
spacing of 6.23 Å; it is no wonder that the elastic energy for WS2 (MoS2) nanotube formation is
appreciably larger than that of graphitic carbon. If one takes the elastic energy threshold for folding
to be 0.05 eV/atom, the calculated diameter of a single-wall carbon nanotube is between one and
1.2 nm [25], and that of a single-wall MoS2 should be 6.2 nm [9,26]. Therefore, the diameters of the
daughter WS2 nanotubes observed in the current series of experiments reconcile very well with the
previous calculations.
2.2. Growth Mechanism
It is hypothesized that the formation of the daughter nanotube occurs through a strong interaction
of the highly energetic plasma, used in this work, with a point or line defect on the outer surface of
the mother nanotube, leading to rapid unzipping and exfoliation of 1–3 layers-thick WS2 fragments.
One way for the exfoliated nanosheets to release the large elastic strain and fold into a nanotube is
through an “inverted umbrella” reaction, which is the manifestation of the “Walden inversion”
typical of a nucleophilic attack of a stereoisomer by an electron-rich moiety [27].
In an effort to understand the mechanism of formation of these daughter nanotubes, additional
HRSEM analysis of the plasma-treated nanotubes was undertaken. The HRSEM in Figure 7a reveals
a reversely revolved nanoscroll of ~20 nm in diameter attached to the surface of the mother nanotube.
Furthermore, a clearly observed step defect or dark contrast on the mother nanotube beneath the
daughter nanoscroll is reminiscent of the exfoliation process of the WS2 patch. Unfortunately, the
resolution of the SEM did not permit viewing the smaller (3–7 nm) daughter nanotubes.

21
Figure 7. (a) HRSEM micrograph of a daughter nanoscroll attached to a large nanotube;
(b) (I) HRTEM of a two-layer daughter nanoscroll viewed head-on along its axis and
attached to a large multiwall nanotube; (II) an initial scrolling stage of an exfoliated
single layer and three (III) layers before the formation of the nanotube; (c) HRSEM
image of a daughter nanotube attached to a mother INT; and (d) HRTEM images
revealing the folding of WS2 nanosheets to form a daughter nanotube.

Nonetheless, this analysis suggests very strongly that the elastic strain of the exfoliated WS2 sheet
produces oppositely revolved daughter nanotubes. Moreover, at better resolution, the HRTEM image
in Figure 7b depicts a two-layer daughter nanoscroll viewed head-on along its axis marked by “I”.
Furthermore, an initial scrolling stage of an exfoliated single layer (“II”) and three layers (“III”)
before the formation of the nanotube was also observed. In addition, Figure 7c,d shows an HRSEM
image of a daughter nanotube attached to a mother nanotube and HRTEM images of a WS2 nanosheet
in the process of folding to form a daughter nanotube.
A schematic model for the growth mechanism of the daughter nanotubes is depicted in Figure 8.
This growth mechanism proposes that the fragments of the outermost (1–3) layers of the predecessor
nanotubes were unzipped by the plasma treatment, exfoliated and folded into daughter nanotubes.
The large excitation energies of the plasma together with the mechanical strain lead to a nanoscopic
“Walden-type inversion” [27]. The reactive edges of the inverted layers induce further folding into
daughter nanotubes with a smaller radius of curvature than the predecessor (mother) multiwall WS2

22
nanotube. Detachment of the 1–3 layers from the mother nanotube may also be followed by rotation
and inclination, in this case, the axes of the mother and daughter nanotubes do not necessarily
coincide or form a specific angle between them. In other cases the rapid quenching of the excess
energy of the nanosheets does not permit them to fully close, which leads to nanoscrolls or to a
nanotube with one closed wall and the others remaining unclosed. Nanoscrolls may also occur due
to steric hindrance, where the plasma-induced exfoliated nanosheets released their energy without
being able to undergo timely inversion.
Figure 8. Schematics of the proposed growth mechanism of the daughter nanotubes by
plasma treatment of the multiwall mother nanotubes.

In a few cases, WSx nanoclusters were observed adjacent to the daughter nanotubes (see Figure 9).
These non-stoichiometric nanoclusters could be obtained by the condensation of tungsten and sulfur
atoms or WS2 molecules from the vapor phase. In turn, the condensation of clusters onto the tube
edges could lead to further elongation or even the growth of an extra layer on the daughter nanotube
surface. Another plausible event is the condensation of the vapors into separate nanosheets, which,
upon quenching, form isolated nanotubes.
The proposed mechanism is consistent with the data presented in this work. In order to shed light
on the detailed growth mechanism of the (daughter) nanotubes and to control their length, diameter
and the number of layers, future experiments will focus on the variation of the plasma treatment
process, including the substrate temperature, pressure in the chamber, etc.
Since highly excited clusters of WS2 (MoS2) can be formed using arc-discharge and a variety of
other techniques, high-power plasma ablation would possibly allow synthesizing a few-wall
nanotubes under controlled conditions in higher yields.

23
Figure 9. TEM images of nanoclusters surrounding daughter nanotubes. Presumably,
the clusters were generated by the condensation of tungsten and sulfur atoms or WS2
molecular clusters from the vapor phase created by the plasma treatment of the multiwall
WS2 nanotubes.

3. Experimental Section
3.1. Plasma Treatment
The schematic drawing and photograph of the experimental set-up for the inductively coupled
radio-frequency plasma irradiation (27.12 MHz) [28] of the multiwall WS2 nanotubes is depicted in
Figure 10.
In these experiments, non-thermal plasma with electrons, atoms and ions, having different
temperatures each, was used to irradiate powders of multiwall WS2 nanotubes (INT-WS2), inorganic
fullerene-like (IF) quasi spherical nanoparticles and different transition metal dichalcogenides
microcrystalline 2H-platelets. A plasma power in the range of 400–600 W was applied for 5, 10,
20, 40 and 80 min. The electron temperature in these experiments was in the range of
1.7 × 104–2.3 × 104 °K (1.5–2 eV), and the electron density was in the range of ~1012/cm3 [29].
The argon gas pressure was 10 Pa, and the flow speed of the Ar gas was 30–35 cm3/s, while the
base pressure before the Ar gas was 10í4 Pa. The temperature of the neutral Ar atoms and Ar+ ions
is approximately two orders of magnitude smaller than that of the electrons. The plasma energy
impact on the substrate surfaces was 2.3 W/cm2 at 400 W and 3.1 W/cm2 at 600 W [30]. The plasma
parameters, pressure and energy impact were constant over the treatment time. The temperature of
the substrate increased with time and depended on the heat conductivity and the quality of the thermal
contact between the powder and the susceptor. The nanoparticles temperature was different from that
of the gas. It was influenced by a number of factors, including the electron and ion bombardment,
electron-ion recombination, reaction enthalpy from the chemical surface reaction, energy loss by heat
radiation and conduction. The temperature of the nanotubes could be estimated to be in the range of
a few hundred degrees centigrade [31].

24
Figure 10. (a) Schematic representation and (b) Photograph of the experimental set-up
for the plasma treatment of the multiwall WS2 nanotubes.

(a)

(b)

3.2. Electron Microscopy
The resulting samples were examined by transmission electron microscopy (TEM) (Philips
CM120 operating at 120 kV, equipped with an energy dispersive X-ray spectroscopy (EDS) detector
(EDAX Phoenix Microanalyzer) for chemical analysis). High-resolution TEM (HRTEM) (FEI
Technai F30-UT, with a field-emission gun operating at 300 kV) and scanning transmission electron
microscopy (STEM) (FEI Technai F20 operating at 200 kV equipped with a high-angle annular dark
field (HAADF) detector and EDS detector (EDAX-Phoenix Microanalyzer)) were also used.
Complementary information was obtained by high-resolution scanning electron microscopy
(HRSEM) (Zeiss Ultra model V55 and LEO model Supra 55VP equipped with an EDS detector
(Oxford model INCA) and backscattering electron (BSE) detector).
4. Conclusions
In conclusion, WS2 nanotubes of 1–3 layers (“daughter nanotubes”), 20–100 nm-long and with
diameters varying between 3–7 nm, were obtained by plasma treatment of multiwall nanotubes in
inert atmosphere. The proposed growth mechanism of the daughter nanotubes involves the strong
interaction of the plasma with point or line defects, causing unzipping and exfoliation of the
outermost layers of the multiwall nanotube, the release of the elastic strain, followed by scrolling and
closure into small nanoscrolls or nanotubes. Sublimation of W and S atoms, WS2 molecules and
cluster formation could serve as an additional building material for daughter tube formation and
extension. These few-layered WS2 nanotubes represent a locally stable, highly excited state of this
solid. Being of such small dimensions, they should reveal a quantum confinement effect, as well as
new optical, electrical and mechanical properties. Furthermore, these nanotubes can be suspended in
different solvents and could possibly be of particular interest, e.g., for drug delivery.

25
Acknowledgments
We gratefully acknowledge Gal Radovski for assisting with the HRSEM microscopy analysis.
We gratefully acknowledge the support of the ERC project INTIF 226639, ITN MoWSeS 317451,
the Israel Science Foundation, the COST action COINAPO MP0902, the FTA project of the Isr. Natl.
NanoIntiative; the G.M.J. Schmidt Minerva Center for supramolecular chemistry; the Harold
Perlman and the Irving and Azelle Waltcher Foundations, and the Irving and Cherna Moskowitz
Center for Nano and Bio-Nano Imaging. R.T. is the Drake Family Chair in Nanotechnology and
director of the Helen and Martin Kimmel Center for Nanoscale Science.
Author Contributions
V.B.: Plasma treatment; R.P.-B.: TEM analyses and discussion of the results; A.A.-Y.: TEM
analysis and discussion of the results; T.L.: Theoretical calculations; G.S.: Theoretical calculations;
R.T.: Analysis of the nanotubes, discussion and the dissemination of the results; A.Z.: Synthesis of
inorganic multiwall WS2 nanotubes, TEM/SEM analyses, discussion of the plasma experiment,
discussion the dissemination of the results.
Conflicts of Interests
The authors declare no conflict of interests.
References
1.
2.
3.
4.
5.
6.
7.
8.

9.

Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and cylindrical structures of tungsten
disulphide. Nature 1992, 360, 444–446.
Zak, A.; Sallacan-Ecker, L.; Margolin, A.; Genut, M.; Tenne, R. Insight into the growth
mechanism of WS2 nanotubes in the scaled-up fluidized-bed reactor. Nano 2009, 4, 91–98.
Chopra, N.G.; Luyken, J.; Cherry, K.; Crespi, V.H.; Cohen, M.L.; Louie, S.G.; Zettl, A. Boron
Nitride Nanotubes. Science 1995, 269, 966–967.
Feldman, Y.; Wasserman, E.; Srolovitz, D.J.; Tenne, R. High-rate, gas-phase growth of MoS2
nested inorganic fullerenes and nanotubes. Science 1995, 267, 222–225.
Remskar, M.; Skraba, Z.; Cleton, F.; Sanjines, R.; Levy, F. MoS2 as microtubes. Appl. Phys. Lett.
1996, 69, 351–353.
Rubio, A.; Corkill, J.F.; Cohen, M.L. Theory of graphitic boron nitride nanotubes. Phys. Rev. B
1994, 49, 5081–5084.
Seifert, G.; Terrones, H.; Terrones, M.; Jungnickel, G.; Frauenheim, T. Structure and electronic
properties of MoS2 nanotubes. Phys. Rev. Lett. 2000, 85, 146–149.
Linnolahti, M.; Pakkanen, T.A. Quantum chemical treatment of large nanotubes via use of line
group symmetry: Structural preferences of magnesium dichloride nanotubes. J. Phys. Chem. B
2006, 110, 4675–4678.
Seifert, G.; Köhler, T.; Tenne, R. Stability of metal chalcogenide nanotubes. J. Phys. Chem. B
2002, 106, 2497–2501.

26
10. Therese, H.A.; Li, J.; Kolb, U.; Tremel, W. Facile large scale synthesis of WS2 nanotubes from
WO3 nanorods prepared by a hydrothermal route. Solid State Sci. 2005, 7, 67–72.
11. Li, Y.D.; Li, X.L.; He, R.R.; Zhu, J.; Deng, Z.X. Artificial lamellar mesostructures to WS2
nanotubes. J. Am. Chem. Soc. 2002, 124, 1411–1416.
12. Sen, R.; Govindaraj, A.; Suenaga, K.; Suzuki, S.; Kataura, H.; Iijima, S.; Achiba, Y.
Encapsulated and hollow closed-cage structures of WS2 and MoS2 prepared by laser ablation at
450–1050 °C. Chem. Phys. Lett. 2001, 340, 242–248.
13. Margulis, L.; Salitra, G.; Tenne, R.; Talianker, M. Nested fullerene-like structures. Nature 1993,
365, 113–114.
14. Tenne, R. Doped and Heteroatom-Containing Fullerene-like Structures and Nanotubes.
Adv. Mater. 1995, 7, 965–995.
15. Singh, D.M.D.J.; Pradeep, T.; Bhattacharjee, J.; Waghmare, U.V. Novel cage clusters of MoS(2)
in the gas phase. J. Phys. Chem. A 2005, 109, 7339–7342.
16. Parilla, P.A.; Dillon, A.C.; Jones, K.M.; Riker, G.; Schulz, D.L.; Ginley, D.S.; Heben, M.J. The
first true inorganic fullerenes? Nature 1999, 397, 114.
17. Enyashin, A.N.; Gemming, S.; Bar-Sadan, M.; Popovitz-Biro, R.; Hong, S.Y.; Prior, Y.; Tenne, R.;
Seifert, G. Structure and stability of molybdenum sulfide fullerenes. Angew. Chem. Int. Ed.
2007, 46, 623–627.
18. Enyashin, A.N.; Bar-Sadan, M.; Sloan, J.; Houben, L.; Seifert, G. Nanoseashells and
nanooctahedra of MoS2: Routes to inorganic fullerenes. Chem. Mater. 2009, 21, 5627–5636.
19. Albu-Yaron, A.; Levy, M.; Tenne, R.; Popovitz-Biro, R.; Weidenbach, M.; Bar-Sadan, M.;
Houben, L.; Enyashin, A.N.; Seifert, G.; Feuermann, D.; et al. MoS2 hybrid nanostructures:
from octahedral to quasi-spherical shells within individual nanoparticles. Angew. Chem. Int. Ed.
2011, 123, 1850–1854.
20. Alexandrou, I.; Sano, N.; Burrows, A.; Meyer, R.R.; Wang, H.; Kirkland, A.I.; Kiely, C.J.;
Amaratunga, G.A.J. Structural investigation of MoS2 core–shell nanoparticles formed by an arc
discharge in water. Nanotechnology 2003, 14, 913–917.
21. Gordon, J.M.; Katz, E.A.; Feuermann, D.; Albu-Yaron, A.; Levy, M.; Tenne, R. Singular MoS2,
SiO2 and Si nanostructuresüsynthesis by solar ablation. J. Mater. Chem. 2003, 18, 458–462.
22. Journet, C.; Maser, W.K.; Bernier, P.; Loiseau, A.; de-la Chapelle, M.L.; Lefrant, S.; Deniard,
P.; Lee, R.; Fischer, J.E. Large-scale production of single-walled carbon nanotubes by the
electric-arc technique. Nature 1997, 388, 756–758.
23. Qin, L.-C.; Zhao, X.; Hirahara, K.; Miyamoto, Y.; Ando, Y.; Iijima, S. Materials science: The
smallest carbon nanotube. Nature 2000, 408, 50.
24. Wang, N.; Tang, Z.K.; Li, G.D.; Chen, J.S. Materials science: Single-walled 4 Å carbon
nanotube arrays. Nature 2000, 408, 50–51.
25. Hernandez, E.; Goze, C.; Bernier, P.; Rubio, A. Elastic properties of C and BxCyNz
composite nanotubes. Phys. Rev. Lett. 1998, 80, 4502–4505.
26. Lorenz, T.; Teich, D.; Joswig, J.-O.; Seifert, G. Theoretical study of the mechanical behavior of
individual TiS2 and MoS2 nanotubes. J. Phys. Chem. C 2012, 116, 11714–11721.

27
27. Koshland, D.E. Stereochemistry and the mechanism of enzymatic reactions. Biolog. Rev. 1953,
28, 416–436.
28. Herrmann, I.; Brüser, V.; Fiechter, S.; Kersten, H.; Bogdanoff, P. Electrocatalysts for oxygen
reduction prepared by plasma treatment of carbon-supported cobalt tetramethoxyphenylporphyrin.
J. Electrochem. Soc. 2005, 152, A2179–A2185.
29. Meichsner, J.; Schmidt, M.; Schneider, R.; Wagner, H.E. Nonthermal Plasma Chemistry and
Physics; CRC Press: Boca Raton, FL, USA, 2013; Section 3.8.7, pp. 120–121.
30. Savastenko, N.A.; Müller, S.; Anklam, K.; Brüser, M.; Quade, A.; Walter, C.; Brüser, V. Effect
of plasma treatment on the properties of Fe-based electrocatalysts. Surf. Coat. Technol. 2011,
205, 439–442.
31. Kortshagen, U. Nonthermal plasma synthesis of semiconductor nanocrystals. J. Phys. D
Appl. Phys. 2009, 42, 113001.

28

IF-WS2/Nanostructured Carbon Hybrids Generation and
Their Characterization
Claudia C. Luhrs, Michael Moberg, Ashley Maxson, Luke Brewer and Sarath Menon
Abstract: With the aim to develop a new generation of materials that combine either the known
energy absorbing properties of carbon nanofibers (CNF), or the carbon-carbon bond strength of
graphene sheets (G), with the shock resistance properties reported for Inorganic Fullerene type WS2
structures (IF-WS2), hybrid CNF/IF-WS2 and G/IF-WS2 were generated, characterized and tested.
Experimentation revealed that in situ growth of carbon nanostructures with inorganic fullerene
tungsten disulfide particulates had to be performed from particular precursors and fabrication
conditions to avoid undesirable byproducts that hinder fiber growth or deter graphene generation.
The novel protocols that allowed us to integrate the IF-WS2 with the carbon nanostructures,
producing dispersions at the nanoscale, are reported. Resulting hybrid CNF/IF-WS2 and G/IF-WS2
products were analyzed by X-ray Diffraction (XRD), Scanning Electron Microscope (SEM) and
TEM (Transmission Electron Microscopy). The thermal stability of samples in air was evaluated by
Thermogravimetric Analysis (TGA). CNF/IF-WS2 and G/IF-WS2 hybrids were introduced into
epoxy matrices, and the mechanical properties of the resulting composites were analyzed using
nanoindentation. Epoxy composite samples showed drastic improvements in the Young’s modulus
and hardness values by the use of only 1% hybrid weight loadings. The carbon nanofiber inclusions
seem to have a much greater impact on the mechanical properties of the composite than the graphene
based counterparts.
Reprinted from Inorganics. Cite as: Luhrs, C.C.; Moberg, M.; Maxson, A.; Brewer, L.; Menon, S.
IF-WS2/Nanostructured Carbon Hybrids Generation and Their Characterization Inorganics 2014, 2,
211–232.
1. Introduction
Tungsten disulfide inorganic fullerene-like particulates (IF-WS2), a dichalcogenide with distinct
physical and chemical properties [1,2] that presents a hollow cage structure with potential uses as
lubricant, component in batteries, supercapacitors or catalyst, among others [3–9], was introduced a
few years ago as a material with shock resistance properties [10–13]. The shock absorbing ability of
IF-WS2 particulates allows them to endure pressures up to 25 GPa, with concomitant temperatures of
up to 1000 °C, without structural degradation or phase change [12], a characteristic that opens an
exciting window of possibilities for protective systems applications. Recent reports have further
explored the features of such WS2 nanostructures and some studies of their inclusion in polymeric
matrices, along with the composites mechanical properties, have been published [14–24]. Despite
those outstanding characteristics, the WS2 density might be considered a drawback if it is to be used
as component in protective gear that should be lightweight, able to withstand high temperatures and
be flame resistant. A possible compromise between all those requirements might be made by joining
the sought-after shock absorbing attribute of IF-WS2 with lighter materials.

29
Due to their strength and light weight, tridimensional carbon assemblies and porous carbon
structures such as nanotubes, foams or intertwined nanofibers [25–37], along with two-dimensional
structures, such as Graphene [38–40], have been the focus of attention for energy absorbing
applications. However, despite the advances in the field, the performance of personal protective
equipment or sporting gear composed of carbon nanostructures can still be improved by the use of
materials that could divert, distribute or dissipate the energy of impacts and the shock waves
associated with them, in a more efficient manner.
Given the shock resistance characteristics of IF-WS2 and the energy absorption of carbon
nanostructures mentioned above, the combination of those two types of materials seems as a natural
next step in the design of protective systems. Recent work that has successfully explored the
possibility of merging the carbon component with IF-WS2 using polymeric matrices [16,18,41–45].
However, the energy dissipation characteristics of those products could be attributed, to an extent, to
the role of the viscoelastic polymeric matrices. Conversely, new reports in the generation of carbon
fiber tridimensional structures have proven that CNF can be grown to form a macroscopic foam with
viscoelastic properties without the need of a polymeric component [46]. Adding the known shock
absorbing characteristics of IF-WS2 to such type of carbon structures will be highly desirable and is
the focus of this investigation.
The development of a protocol to combine the IF-WS2 characteristics with those of the carbon
structures in the absence of a polymer, to generate a random distribution of those two phases at the
nanoscale (ca. a hybrid made of solely inorganic components, thus, avoiding some of the polymer
drawbacks such as aging but gaining in terms of lightweight and thermal stability), remains as one
of the major challenges.
Experimentation by our group has shown that adding IF-WS2 to an already existing 3D carbon
structure, even when using solvents to achieve the mixture, renders an inhomogeneous solid.
The present work aims to produce a well-dispersed hybrid system composed of a carbon solid
(Carbon Nanofiber or Graphene) and low loading levels of IF-WS2. We found that to produce a well
dispersed 3D structure of carbon nanofibers and IF-WS2 it is necessary to employ the two stage in
situ protocol described in the next section. In contrast, to produce a hybrid of Graphene (2D layered
structure) with IF-WS2, the use of the in situ protocol did not show significant improvement in the
phase distribution when compared to physically mixing the components.
As an extension of the primary goals, we studied the mechanical properties of epoxy composites
based on hybrid CNF/IF-WS2 and G/IF-WS2 made by in situ routes, and contrast the latter to those
created from physical mixtures of the components.
2. Results and Discussion
With the goal of generating hybrid Carbon nanostructure/IF-WS2 we conducted the steps
summarized in the experimental section and produced samples whose qualities are described below.
We have divided the results and discussion segment into three parts; the first one presenting the
microstructural characteristics and stability of samples based on CNF/IF-WS2, the second one
including the study of Graphene/IF-WS2 and the third showing the mechanical properties of both
CNF/IF-WS2 and Graphene/IF-WS2 epoxy composites. The samples made in the absence of polymer

30
were denominated hybrids and the ones with nanostructures embedded in epoxy are referred as
composites as a way to distinguish them.
2.1. Carbon Nanofiber/IF-WS2 Hybrids
The combination of existing 3D CNF structures with IF-WS2 in the absence of a polymeric matrix,
as mentioned in the introduction section, results in inhomogeneous solids. For example, the addition
of IF-WS2 to already existing CNF generates samples where diffusion paths dominate the final
structure: the surface of the carbon fibers tends to contain larger amounts of the IF-WS2 phase than
sections of the sample not exposed to the surface. The use of solvents and sonication to disperse the
IF or its precursors, is not enough to produce a homogeneous product. A second example; The
addition of metal catalyst particles (to promote the carbon fiber growth) to existing IF-WS2 followed
by thermal treatments to proceed with the carbon nanofibers growth in the presence of a carbon
source, produces a thin layer of metal sulfides in the surface of the metal particles. This phenomenon
is observed also in the absence of carbon sources: at moderate temperatures (below 350 °C according
to our findings) the metal reacts with the tungsten sulfide producing thin films of secondary phases
in the metal particle surface, which in practice poison the catalyst, capping it and hindering the
growth of the carbon fibers. Hence, this method also fails to produce the expected product. Variations
in the experimental parameters, such as gas flows, placement of precursors, size and nature of metal
catalyst, amid others, generates, in the best case scenario, amorphous carbonaceous byproducts of
different sizes and morphologies, far from the desired fiber or porous structures [47].
Diverse precursor options could be used to generate IF-WS2 structures; (NH4)2WS4 thermal
decomposition and WO3 reaction with sulfur-containing compounds are among the most common
routes employed [1,48–56]. The generation of IF-WS2 carried out in our laboratories using such
precursors resulted in samples with different levels of long range order and particle size, as can be
inferred from its XRD pattern and SEM analysis. Figure 1 presents the WS2 patterns obtained when
decomposing (NH4)2WS4 and from the reaction of commercial WO3 nanoparticles with H2S
atmospheres. It is worth noting that the temperatures needed for such reactions to occur differ,
the former was performed at moderate temperatures (ca. 500 °C) while the latter required higher
temperatures to be carried out to completion (ca. 800 °C). The diffraction peaks for IF-WS2 generated
from WO3 are narrower and the reflections (00l) with l = 2n more evident and well defined than the
ones obtained from (NH4)2WS4. SEM observation of both products confirmed the presence of the
hollow cage structures, the so-called 3R phase. Moreover, the former tetrathiotungstate precursor
reacts with metal particles when it decomposes, which makes WO3 the best choice to be combined
with metal catalyst for fiber growth.
The two stage in situ synthesis of homogeneous CNF/IF-WS2 required diverse atmospheres and
temperature steps to be accomplished successfully. The first stage had the objective to grow CNF
from a metal catalyst intermixed with WO3 and the second stage consisted on sulfurizing the mixture
of CNF/WO3 to transform the tungsten oxide into IF-WS2.
The dispersion of the metal particle (nickel in this case) with the WO3 nanoparticles was
performed using a solvent and sonication, followed by evaporating the mixture until dry since it
rendered a more homogenous precursor than simply grinding the solids. A first temperature step was

31
carried out at 350 °C in reducing atmospheres as a precautionary measure to assure that the metal
catalyst surface was free of oxides. A second step at the same temperature including ethylene (as
carbon source) and small amounts of oxygen diluted in inert gas (as reaction initiator and radical
originator) was performed to cover the metal particle with an initial thin layer of carbon (which
prevents the metal gross agglomeration and sintering at higher temperatures). The growth step was
made at 550 °C, following protocols previously developed [57]. A final temperature stage at 900 °C
in H2S containing environment was included to transform the WO3, now dispersed in between
carbon nanofibers, into IF-WS2. The sample was then allowed to cool to room temperature using an
inert atmosphere.
Figure 1. XRD patterns of IF-WS2 particles generated from different precursors.
IF-WS2 generated from (a) commercial WO3 particles; (b) from ammonium
tetrathiotungstate. All peaks were identified as IF-WS2 [12].

The electron micrographs of the nickel nanoparticles employed, the carbon nanotube intertwined
nanofibers generated and the hybrid CNF/IF-WS2 are presented in Figure 2. The ratios of catalyst to
WO3 used were designed to have nominal loading values of 0.5, 1.5, and 5% of IF-WS2 with respect
to the total weight of the CNF. The distribution of the IF particles was studied using SEM in
Secondary Electron (SE) and Backscattered Electron (BSE) modes along with Energy Dispersive
Spectroscopy (EDS) mapping and can be described as discrete pockets of cage like particulates of
approximately 1–2 microns dispersed into the CNF threads (Figure 2c). This finding is consistent
across different sample locations.
The catalyst used for the carbon nanofibers was not removed and represents close to 3% of the
total weight of the sample. Its effects on the properties of the epoxy composites are discussed in
Section 2.3.
The last stage of the in situ protocol, the sulfurization, transforms not only the WO3 into WS2 but
the nickel catalyst into nickel sulfide. Indeed, the hybrid CNF/IF-WS2 samples are composed by the
solids mentioned in Table 1 (experimental section), where nominal values and elemental analysis by
ICP methods are contrasted. The inductive plasma emission spectroscopic data shows that the final
composition attained is within 0.2% of the targeted values.

32
As a means to understand the characteristics of the samples and the different phases formed at the
diverse synthesis steps, some of the runs where halted before completion and the products at that
point, were analyzed by diverse techniques. As part of those trials, experiments of fiber growth in
the absence of WO3 with and without sulfurization treatment were performed. The X-ray diffraction
analysis of the latter, containing only CNF samples, is presented in Figure 3. The main reflection,
close to 26°, corresponds to the (002) peak of graphite, characteristic of many carbon products that
include a crystalline component with various degrees of basal plane alignment [58]. In the present
case the peak is associated to the CNF presence. The diffraction pattern of samples after the carbon
fiber growth step but before sulfurization (red line in Figure 3), show only the peaks of graphite and
a couple of reflections, close to 45 and 52°, that correspond to the nickel particles used as catalyst.
The phases encountered after sulfurization (blue line) corroborate the presence of two extra
crystalline structures, nickel sulfide: Ni3S2 [59,60] and Ni17S18 [61].
Figure 2. Microstructural Analysis of CNF and WS2 precursors and CNF/IF-WS2 hybrid.
Scanning Electron Microscopy secondary electron images of (a) starting nickel
nanoparticles; (b) carbon nanofibers grown from Ni catalyst; (c) nanoparticle dispersion
and cluster of WS2 nanoparticles within carbon nanofibers (CNF/IF-WS2 hybrid).

Figure 3. X-ray diffraction pattern of carbon nanofibers. (a) Typical peaks of graphite
and nickel metal are identified in samples as prepared; (b) XRD pattern after
sulfurization step shows graphite and nickel sulfide reflections.

33
The XRD patterns of CNF/IF-WS2 hybrids at the three different levels of loadings prepared are
shown in Figure 4. The same phases recognized in Figure 3 after sulfurization; graphitic peaks,
Ni3S2 and Ni17S18, along with a comparatively weak peak corresponding to IF-WS2, can be identified.
Figure 4. X-ray diffraction patterns for hybrid CNF/IF-WS2 samples. The main
difference between XRD patterns in samples at diverse loadings is the relative intensity
for the nickel sulfide vs. graphitic peaks.

SEM in BSE mode studies along elemental mapping by EDX confirmed the presence of IF-WS2
as highly dispersed phase. Figure 5 shows the metal distribution as bright spots (top left), including
both nickel and tungsten phases. The IF-WS2 distribution can be inferred as the positions where W
elemental mapping is found (top right). The W mapping shows the presence of mainly small (nm)
particulates homogeneously distributed in the sample. In contrast, the Ni particles used as catalyst,
now converted into sulfides, appear both as nanometer particulates and micron size agglomerates;
with dimensions that could be correlated with the different fiber diameters (bottom right). The fact
that the nickel elemental distribution shows micron size clusters is a sign that some agglomeration
of some of the initial particles might have occurred during the fiber growth, since the original catalyst
had a size distribution in the nm range, as shown in Figure 2a. The sulfur map (bottom left)
encompasses both, the IF-WS2 and Ni sulfide components.
The samples’ thermal analysis under oxygen containing atmospheres show that all specimens;
CNF (catalyst included), CNF sulfurized and hybrid CNF/IF-WS2 are stable at least up to a
temperature of 500 °C (Figure 6). At higher temperatures the carbonaceous component burns off to
produce CO2, the tungsten sulfide reverts to its oxidized form and the metal catalyst oxidizes.
Experiments conducted in inert atmosphere showed that hybrid CNF/IF-WS2 samples do not present
weight changes in the window of conditions used (RT to 950 °C) and can be considered thermally
stable. Indeed, the hybrid samples stability observed surpasses the ones observed for polymer
products without WS2, which in either atmosphere start decomposing at much lower temperatures,
and are in agreement with stability observed in air for IF-WS2 based composites [19,20,62]. The

34
reported thermal stability of WS2 fullerene-like particles alone under oxidizing atmospheres indicates
that, depending on particle size, oxidation might begin close to 290 °C (100 nm) or delay up to close
to 440 °C (3 m), while exposed to inert gases the phase is stable up to 1200 °C [63,64]. In our case,
the use of CNF without polymeric component seems to increase the thermal stability of the hybrid,
given that the average IF-WS2 particle size produced was in the nanometer scale (expected to oxidize
close to 300 °C) and changes in weight were not evident until above ~500 °C. The thermal stability
observed for IF-WS2 might be related to the particulates being embedded in the CNF intertwined
fibers, which might delay their reaction. Further studies to fully explain these phenomena are currently
under way.
Figure 5. Phase distribution studied by SEM/EDX analysis. Top left: Backscattered
electron image showing the carbon fibers in grey and the metal containing phases as
bright spots in the sample CNF/IF-WS2 with the lowest IF-WS2 loading. Elemental
mapping by EDX analysis confirmed the presence of IF-WS2 mainly as highly dispersed
phase (top right) with small (nm) particle size. The elemental Ni map shows the presence
of nm and micron size particulates (bottom right). The sulfur map (bottom left) includes
both, the IF-WS2 and Ni sulfide components.

35
Figure 6. Thermogravimetric Analysis. Temperature programmed oxidation (burn off
process) for (a) carbon nanofibers-including nickel catalyst; (b) sulfurized carbon
nanofibers -including nickel sulfide and (c) CNF/IF-WS2 hybrid with 3% tungsten sulfide
loading. All of the samples are stable in oxygen containing environment up to 500 °C.

2.2. IF-WS2/Graphene Hybrids
The complications introduced by the use of metal catalyst for nanofibers growth were not
encountered during the synthesis of Graphene/IF-WS2 hybrids, since no extra metal components
were needed to generate them. Only small amounts of urea were used as an expansion agent. The
production of graphene was accomplished by the reduction-expansion of graphite oxide (GO), which
was generated from graphite flakes by the process described in the experimental section. The GO
mediated process generates disordered graphene; where graphene sheets tend to entangle with each
other but remain separated enough to maintain a relative high surface area (ca. 600 m2/g as per BET
analysis). It is worth noting that the oxygen content in the synthesis environment when using GO as
a carbon precursor is sufficient to oxidize some of the IF-WS2 if the latter is used since the beginning
of the protocol, forcing the use of WO3 as the tungsten source. Thus, WO3 nanoparticles were subject
to H2S treatments to sulfurize them after its mixtures with graphene were obtained.
The two variations on the synthesis conditions consisted of: (i) dispersing the WO3 nanoparticles
with GO and then performing the exfoliation at high temperature (to generate Graphene/WO3)
followed by sulfurization (to obtain Graphene/IF-WS2) or (ii) perform the GO exfoliation first and
then directly mixing the resulting graphene with WO3 and then sulfurizing the product (rendered
Graphene/IF-WS2). The microstructures created are presented in Figure 7.
The first synthetic approach, GO plus WO3 exfoliation followed by sulfurization, resulted in
hybrid Graphene/IF-WS2 where the IF structures were located on the surface of disordered graphene
sheets (marked by arrows, Figure 7a). Due to the thermal exfoliation and urea reduction process,
when volatile groups leave the graphite oxide structure at temperatures close to 200 °C, they carry
along the WO3, which ends in top of the layers, where they remain through the sulfurization step that
converts them into IF-WS2 cage structures. The exfoliation process separates the graphene sheets but
also promotes the separation of IF particles, which were found in all cases as individual particles.

36
The position of the IF particulates following the second synthetic approach, is quite distinct. In
this protocol, GO is first thermally exfoliated and reduced and the subsequent graphene is mixed
with WO3 (with the aid of solvents and sonication, followed by solvent evaporation until dry) and
the mixture is then sulfurized. The IF particulates are found between the graphene layers, intimately
embedded into the sheets structure (Figure 7b). Small (few nm diameter) IF particulate agglomerates,
marked in the image with a blue arrow, were observed along individual particles, indicated by red
arrows. The IF-WS2 observed by Transmission Electron Microscopy presents the characteristic
hollow cage, partially faceted structure and an interlayer spacing of 0.62 nm.
For comparison, samples in which Graphene and IF-WS2 were formed as individual phases and
then added as a physical mixture using solvents and sonication (not grown in situ), present similar
structures to the ones in Figure 7b, demonstrating that the protocols developed to introduce IF-WS2
into carbon structures are required only when tridimensional architectures of the carbonaceous
component are needed but not if a layered, two dimensional structure is used.
Figure 7. Microstructural Analysis of Graphene/IF-WS2 hybrids and IF-WS2. Scanning
Electron Microscopy images of (a) Graphene/IF-WS2 hybrid generated from GO and
WO3; (b) Graphene/IF-WS2 hybrid made directly from Graphene and WO3; and (c)
Transmission electron micrograph showing the characteristic hollow core and interlayer
spacing for IF-WS2 particles.

2.3. Epoxy Composites
The use of epoxy as polymeric matrix was intended (independent of its properties as a composite
for use in protection systems) to determine the mechanical properties observed when using carbon
nanomaterials with 3D vs. 2D structures (CNF vs. Graphene) that included low loadings of IF-WS2.
Other instances of WS2 nanostructures embedded in epoxy can be found in [14,23], where WS2
nanotubes were used as filler instead of the IF particulates used in this study. To the best of our
knowledge, the only existing references including IF-WS2 particles in epoxy systems do not involve
a carbonaceous component [21–24].
Experimental testing in our laboratory and previous epoxy resin composite research data [47,65]
were used to select 1% as targeted loading of filler material into epoxy matrix. Bare epoxy resin,
produced in identical conditions to the ones containing diverse fillers, was used as reference.
Table 1 presents the values for the different filler components used to prepare the composites, all

37
cases include only 1% of total filler and 99% of epoxy, by weight. On the table, the first column
contains the nominal values of IF-WS2 component targeted during synthesis, the rest of the columns
are the values encountered when analyzing the filler samples by ICP Emission Spectroscopy.
Two types of nanoindentation measurements showed clear increases in the Young’s modulus of
the epoxy-CNF composites. As described in the experimental section, dynamic mechanical analysis
was performed on epoxy-CNF composites to ascertain the degree of viscoelastic deformation of these
materials at room temperature. At loading frequencies of both 1 Hz and 45 Hz, the elastic modulus
response was dominated by the storage (elastic) modulus with a very minor contribution, less than
5%, from the loss (viscous) modulus. As such, quasi-static indentation was deemed appropriate for
these materials. We used quasi-static nanoindention to measure the Young’s modulus and hardness
for each of the composites using the conditions listed in Table 2. The inclusion of carbon nanofibers
into the epoxy matrix increases the epoxy modulus by 29%. CNF sulfurized, with no IF-WS2 but
subject to sulfurization process with H2S, improves it nearly by 52% (Figure 8a). A possible
explanation to the higher modulus values for the later might be related to the presence of sulfur and
its effects on the epoxy-CNF interface adhesion. The three hybrid samples tested, CNF containing
IF-WS2 particulates with 0.5, 1.5, and 3%, show much higher modulus than the rest, with values that
almost double the one for bare epoxy (for hybrid CNF/IF-WS2 0.5%). While we observed a clear
variation in sample performance for hybrids containing CNF and diverse IF-WS2 contents, the data
does not show a clear correlation with the actual amount of IF added.
For all samples that included fillers we observed a more dramatic change in hardness than for the
elastic modulus; the initial addition of carbon nanofibers had a large effect and increased the hardness
by more than 114% (Figure 8b). The addition of sulfurized fibers further raised the hardness by
another 38%. The sulfurization of the fibers appears to improve the ability of the fibers to bond with
the epoxy matrix, further improving the interfacial strength. The increase in hardness could also be
partially due to the inclusion of nickel sulfide particles. As shown in Figures 3 and 4, after the
sulfurization step is carried out all nickel used as catalyst for the CNF growth transforms into Ni3S2
or Ni17S18. While these particles contribute to an increase in performance in this case, it is possible
that other effects could appear under different situations.
Nickel sulfide contamination is an ongoing concern for the tempered glass industry. High
temperature structures of nickel sulfide can develop in the glass manufacturing process and become
included in tempered glass. This inclusion will result in a stress concentration and if the particle is of a
certain size, the pane of glass will shatter under loading far below what is expected of the material [66].
Depending on the future inclusions of these materials into other matrixes, the possible impact of
sulfide particles as stress concentrators should be considered and further studied. In the case of
polymeric matrices, like the ones studied herein, no failure modes related to the existence of nickel
sulfide are expected, since the working and processing temperatures do not reach those where the
nickel sulfide high temperature phase appears (715 °C). In the case of hybrids in the absence of
polymers, the porous carbon fiber structure is presumed to accommodate the volume expansion of
the phase change and does not constitute a concern.
All the hybrid CNF/IF-WS2 samples showed a significant increment in hardness values, being
CNF/IF-WS2 0.5% the most remarkable, with a 247% improvement over the pure epoxy hardness.

38
Figure 8. Mechanical properties of epoxy composites. All composites contained 99%
epoxy and 1% loading of filler nanostructures: CNF, sulfurized CNF or hybrid
CNF/IF-WS2. For the later, the hybrids contained mostly carbon fibers, with only 0.5,
1.5, and 3% of IF-WS2. (a) Modulus data and (b) Hardness values.

In order to fully understand the CNF/IF-WS2 epoxy composite mechanical properties observed,
further study of their interfaces by IR or other spectroscopic techniques is recommended. Such data
will help characterize the changes introduced by the inorganic components in the oxirane ring bands,
the overall epoxy resin structure and its degree of polymerization.
The inclusion of two-dimensional graphene into the epoxy matrix in equal loadings than the ones
described above, 1wt% filler in epoxy, by itself or as Graphene/IF-WS2, resulted in increased values
of modulus and hardness (Figure 9). However, the improvement when compared to the 3D nanofiber
structures is modest; the hybrid Graphene/IF-WS2 with 0.5% of IF-WS2 showed a 9% modulus
increment over bare epoxy and a 47% increase in hardness. Histograms for the composites that
contain CNF have been highlighted in blue and the ones containing graphene in red. From the Figure
is clear that the samples based in 3D CNF architectures containing IF-WS2 present the highest values.
During the composite fabrication, the graphene based samples seemed to be more difficult to disperse
into the epoxy matrices than the CNF ones and inconsistencies in the sample distribution into the
epoxy puck might have a detrimental effect on the mechanical properties values. Moreover, the fact
that the graphene sheets were distributed randomly in the sample and not oriented in the (002)
direction, where the strong covalent bonds are located, might be another point to be considered to
explain the observed mechanical behavior.

39
As mentioned in Section 2.2, the use of Graphene/IF-WS2 prepared in situ vs. the components
physical mixtures do not seem to present an advantage in terms of phases distribution or their
microstructural characteristics. In terms of mechanical properties, the in situ route and the physical
mixtures of Graphene/IF-WS2 are comparable; the values for modulus are almost the same and the
hardness is higher for the former. This result might be related to the fact that IF particulates might
easily disperse in between the graphene sheets without the need of a multistep process. It is worth
noting that in the sample created in situ, the graphene structure has been exposed to H2S atmospheres,
while in the physical mixture the graphene was pristine and never in contact with sulfurizing
environments. Given that the improvement in modulus and hardness over neat epoxy observed in
those samples was minimal, the present study did not further investigate the interfacial effects created
by diverse functionalities in the graphene surface.
Figure 9. Mechanical properties of epoxy composites. with 1% loading of CNF,
sulfurized CNF, Graphene, IF-WS2, CNT/IF-WS2 hybrids or Graphene/IF-WS2 hybrids.
(a) Normalized Modulus data; (b) Normalized Hardness values.

Overall, the epoxy composite samples containing 3D CNF structures along IF-WS2 created
in situ, showed drastic improvements in the Young’s modulus and hardness values by the use of
only 1% hybrid weight loadings. The carbon nanofiber inclusions seem to have a much greater
impact in the mechanical properties of the composite than the graphene based counterparts for similar
IF loadings.

40
The values for modulus and hardness improvements over the bare polymer mentioned above for
CNF/IF-WS2 composites surpass the ones encountered for other hybrid polymeric nanocomposites
incorporating IF-WS2 nanoparticles and a carbon phase [43–45]. However, in order to strictly
compare mechanical properties for diverse fillers and polymeric matrices, data gathering should be
performed under the same conditions and similar experimental setup to be valid. In particular, the
use of nanoindentation instead of DMA tends to produce different values (usually larger for the
former technique), as recently corroborated by Flores et al. [67]. Thus, comparison between the two
techniques outcomes is not adequate given that no correlation between them exist to date.
3. Experimental Section
3.1. Carbon Nanofiber Hybrids/IF-WS2
The amounts of IF-WS2 included into CNF were determined by the knowledge gained in previous
studies of composites by different groups [14,15,19], and consistent with the intent of generate
lightweight materials, making the less dense carbon nanostructure component the most abundant
phase. The multistep process employed to generate in situ structures is depicted in Figure 10.
Figure 10. CNF/IF-WS2 preparation steps. Ni nanoparticles and WO3 precursors were
combined in solution using sonication in ethanol, dried and placed in a crucible inside a
quartz tube of a tubular furnace. The mixture was then exposed to diverse environments
and thermal treatments to generate Carbon Nanofiber/tungsten disulfide hybrids
(CNF/IF-WS2).

41
The gas flow rates employed for each step were: 300 sccm N2 for air removal, 37 sccm of Ar/H2
(93%/7%) for reduction, a mixture of 44 sccm N2/15 sccm Ethylene/2 sccm O2 for fiber growth
(both low temperature and high), a mixture of 150 sccm N2/15 sccm H2S for sulfurization and
300 sccm N2 during the cool down process.
3.2. IF-WS2/Graphene Hybrids
The generation of graphene was performed in all cases starting from graphite flakes, which were
oxidized to form graphite oxide (GO) following a modification of the process developed by
Marcano et al. [68]. The GO was then used either as (i) dispersion with WO3 nanoparticles and then
exfoliated at high temperature with urea as expansion-reducing agent (to generate Graphene/WO3)
followed by a sulfurization step using H2S atmosphere (to obtain Graphene/IF-WS2) or (ii) exfoliated
at high temperature with urea to form graphene, which was then mixed with WO3 and the mixture
sulfurized (this step rendered Graphene/IF-WS2). The processes described are depicted in Figure 11.
Figure 11. Synthesis protocol followed to generate Graphene/IF-WS2 hybrids from GO.
Initial precursors (graphite flakes, acids and oxidant) will render graphite oxide.
Protocols followed after GO generation (i) mixture of GO with WO3 followed by
exfoliation and sulfurization and (ii) Production of graphene through GO thermal
exfoliation, mixture with WO3 and sulfurization of mixture.

42
3.3. Epoxy Composites
We employed commercially available SpeciFix-20 (Struers, Ballerup, Denmark) as the epoxy
resin, which consists of a resin and a hardening agent (26:5 ratio). Unlike some epoxies that require
heating or pressure, this epoxy is designed to cure under atmospheric conditions. Further, it has a
working period of 60 min, allowing time for the sample to be handled and nanocomposite added and
mixed prior to the onset of curing. A loading of 1% nanoparticles to 99% epoxy by weight was
selected for testing. The composition of the filler materials, the hybrids and individual phases
described in previous sections, is included in Table 1 below. The filler components were added to
the uncured resin using sonication and the composite was left for a period of at least 24 hours to cure
into a hard puck. After the nanoparticles were successfully embedded into the epoxy matrix, the
surface of the pucks was mechanically polished to remove scratches and imperfections and then
transferred to the nanoindentor.
Table 1. Filler compositions used for epoxy composites (first column).
Sample ID (IF-WS2 nominal

ICP value

ICP value

ICP value

composition – filler only)

CNF

Ni

IF-WS2

CNF

96.8%

3.2%

CNF/IF-WS2 (3%)

94%

3%

3%

CNF/IF-WS2 (1.5%)

94.6%

3.2%

1.6%

CNF/IF-WS2 (0.5%)

96.4%

2.9%

0.7%

IF-WS2 (100%)
Sample ID (IFWS2 nominal
composition-filler only)

100%
Graphene

IF-WS2

Graphene

100%

G/IF-WS2 0.5%

99.5%

0.5%

Physical Mix G/IF-WS2 1%

99%

1%

3.4. Characterization Methods
In order to examine the microstructure of the hybrid specimens the samples were analyzed using
a Zeiss Neon 40 High Resolution Scanning Electron Microscope (SEM). Images were acquired at
diverse magnifications while the microscope was operated at 10 or 20 kV. Energy Dispersive

43
Spectroscopy (EDS) experiments were conducted in conjunction with the SEM using the EDAX
equipment with an Apollo 10 silicon drift detector (SDD). Data was collected and analyzed using
Genesis Spectrum software.
A Netzsch STA 449 FE Jupiter, operated in a Temperature Programmed Oxidation (TPO) mode,
was used to study the thermal stability of the samples. The samples were exposed to an Ar/O2,
80%/20% atmosphere, with a total flow of 120 mL minuteí1, from RT to 1000 °C at a heating rate
of 10 °C minuteí1.
The XRD utilized was a Philips 1830 PAnalytical X-ray Diffractometer. The X-ray tube contained
a copper source and the X-rays utilized had a primary wavelength, or K-Alpha, of 1.54 Å. The
samples were placed into a silicon low background sample holder and the diffraction patterns
recorded between 5–70° (2 theta) with 0.020 degrees step size and one second per step.
A JEOL 2010F FASTEM field emission gun scanning transmission electron microscope
(STEM/TEM) equipped with Gatan GIF image filtering system was employed. Samples were
prepared by dispersing the powders in a few ml of ethanol and a drop of the dispersion was placed
in a copper holey-carbon TEM grid where the ethanol was allowed to evaporate.
A Perkin Elmer ICP 5300 DV-AES Inductive Coupled Plasma Emission Spectrometer, was used
to determine the elemental composition of the carbon nanofiber base filler materials before their
addition to the epoxy matrix (hybrids CNF/IF-WS2).
Brunauer Emmet Teller (BET) surface area analysis was performed employing a Quantachrome
Nova 4200. A 300 °C degas step was conducted prior to the analysis; samples were then allowed to
cool down to room temperature and then transferred to the analysis station. The measurements were
done using nitrogen atmosphere.
Nanoindentation was used to measure the composite mechanical properties (elastic modulus and
hardness) of epoxy composites filled with mixtures of CNFs and IF-WS2 particles. The samples were
prepared by mixing the specified amounts of nanophase material with Struers Speci-Fix 20 two-part
epoxy in a 28 mm diameter mold and then allowing the mixture to cure for 24 h. After curing,
the surface of the epoxy composite was polished using standard metallographic techniques, including
diamond suspension polishing using suspended aluminum oxide particles of 1 ȝm and 0.05 ȝm
diameters. The indentations were performed using an Agilent G200 nanoindenter. We performed two
types of experiments with this instrument.
The first experiment was a quasi-static indentation to a set depth, 2 ȝm for all samples.
Other indentation parameters can be found in Table 2. This experiment used a diamond, Berkovich
indenter tip with a nominal tip radius of 150 nm, calibrated using a fused silica standard. A grid of
20 indentation points spaced by 50 ȝm was measured for each epoxy nanocomposite. The Young’s
modulus and hardness were calculated using the approach of Oliver and Pharr [69,70].
The second experiment was dynamic mechanical analysis using a 50 ȝm diameter flat punch.
This experiment allowed the measurement of both the storage and loss moduli of the epoxy
composites. These measurements were performed for five frequency values between 1 Hz and 45 Hz
on the neat epoxy, CNF and IF-WS2 samples. Other parameters for the measurement can be found
in Table 3. A grid of 20 measurements with a 100 ȝm separation between indentations was used for

44
each specimen. The storage modulus, loss modulus, and tan į properties for each specimen were
calculated using the measurement parameters in Table 3 and the methods of Hay and Herbert [71].
Table 2. Parameters used for quasi-static indentation measurements.
Depth limit
Strain rate during loading
Maximum allowable drift rate
Peak hold time
Assumed Poisson’s ratio
% to unload
% Unload in stiffness calculation

2000 nm
0.08/s
0.05 nm/s
10 s
0.40
90
50

Table 3. Parameters used for dynamic mechanical analysis.
Flat punch diameter
Assumed Poisson’s ratio
Pre-compression depth

50 —m
0.40
2 —m

Oscillation amplitude

50 nm

4. Conclusions
Novel hybrid CNF/IF-WS2 with diverse IF loadings were generated using an in situ protocol that
allowed the integration of the two phases into a tridimensional architecture, producing homogeneous
dispersions at the nanoscale in the absence of a polymeric matrix. CNF 3D structures loaded with
IF-WS2 could only be fabricated using a two stage process that involved: (a) the carbon nanofiber
growth from a mixture of metal catalyst with tungsten oxide nanoparticles, using ethylene as carbon
source and moderate temperatures to render CNF/WO3, followed by (b) the sulfurization of the
sample to convert the tungsten precursor into IF-WS2.
In contrast, Graphene/IF-WS2 hybrids were easily obtained either by mixing graphene and
tungsten oxide followed by a sulfurization step or by direct dispersion of the layered graphene
structure with existing IF particles using solvents.
The thermal stability of the CNF/IF-WS2 hybrid samples is much higher than those observed for
IF-WS2 by itself or mixed with polymeric components.
Epoxy composites with 1% weight loadings of hybrid CNF/IF-WS2 showed drastic improvements
in the Young’s modulus and hardness values, with approximately 100 and 250% increase
respectively, over the bare epoxy values. The CNF/IF-WS2 inclusions seem to have a much greater
impact in the mechanical properties of the composite than the Graphene/IF-WS2 based counterparts.
Acknowledgments
The work depicted in this manuscript has been possible with the support of the Office of Naval
Research, Force Protection Thrust, Code 30. We appreciate the help of Dr. Abdul-Mehdi S. Ali from
University of New Mexico, who conducted the analysis of the filler components elemental analysis

45
by ICP. Our team is thankful for the HRTEM analysis of IF-WS2 particles conducted by JEOL,
USA Inc.
Conflicts of Interest
The authors declare no conflict of interest.
References
Margulis, L.; Tenne, R.; Iijima, S. Nucleation of WS2 fullerenes at room temperature. Microsc.
Microanal. Microstruct. 1996, 7, 87–89.
2. Rapoport, L.; Bilik, Y.; Feldman, Y.; Homyonfer, M.; Cohen, S.; Tenne, R. Hollow
Nanoparticles of WS2 as potential solid-state lubricants. Nature 1997, 387, 791–793.
3. Eidelman, O.; Friedman, H.; Tenne, R. Metallic films with fullerene-Like WS2 (MoS2)
nanoparticles: Self-lubricating coatings with potential applications fullerene-like nanoparticles.
In Technological Innovations in Sensing and Detection of Chemical, Biological, Radiological,
Nuclear Threats and Ecological Terrorism; Springer: Dordrecht, Netherlands, 2012; pp. 59–67.
4. Rapoport, L.; Fleischer, N.; Tenne, R. Applications of WS2 (MoS2) inorganic nanotubes
and fullerene-like nanoparticles for solid lubrication and for structural nanocomposites.
J. Mater. Chem. 2005, 15, 1782–1788.
5. Drummond, C.; Alcantar, N.; Israelachvili, J.; Tenne, R.; Golan, Y. Microtribology and
friction-induced material transfer in WS2 nanoparticle additives. Adv. Funct. Mater. 2001, 11,
348–354.
6. Rothschild, A.; Cohen, S.; Tenne, R. WS2 nanotubes as tips in scanning probe microscopy.
Appl. Phys. Lett. 1999, 75, 4025–4027.
7. Dominko, R.; Arþon, D.; Mrzel, A.; Zorko, A.; Cevc, P.; Venturini, P.; Gaberscek, M.;
Remskar, M.; Mihailovic, D. Dichalcogenide nanotube electrodes for li-ion batteries.
Adv. Mater. 2002, 14, 1531–1534.
8. Cheng, F.Y.; Chen, J.; Gou, X.L. MoS2/Ni Nanocomposites as catalysts for
hydrodesulfurization of thiophene and thiophene derivatives. Adv. Mater. 2006, 18, 2561–2564.
9. Wang, Y.; Shi, Z.; Huang, Y.; Ma, Y.; Wang, C.; Chen, M.; Chen, Y. Supercapacitor devices
based on graphene materials. J. Phys. Chem. 2009, 113, 13103–13107.
10. Zhu, Y.; Sekine, T.; Brigatti, K.; Firth, S.; Tenne, R.; Rosentsveig, R.; Kroto, H.; Walton, D.
Shock-wave resistance of WS2 nanotubes. J. Am. Chem. Soc. 2003, 125, 1329–1333.
11. Zhu, Y.; Sekine, T.; Li, Y.; Wang, W.; Fay, M.; Edwards, H.; Brown, P.; Fleischer, N.; Tenne, R.
WS2 and MoS2 Inorganic fullerenes—super shock absorbers at very high pressures. Adv. Mater.
2005, 17, 1500–1503.
12. Zhu, Y.; Sekine, T.; Li, Y.; Fay, M.; Zhao, Y.; Poa, C.; Wang, W.; Roe, M.; Brown, P.;
Fleischer, N.; et al. Shock-absorbing and failure mechanisms of WS2 and MoS2 nanoparticles
with fullerene-like structures under shock wave pressure. J. Am. Chem. Soc. 2005, 127,
16263–16272.
1.

46
13. Volkova, E.I.; Jones, I.A.; Brooks, R.; Zhu, Y.; Bichoutskaia, E. Meso-scale Modelling of Shock
Wave Propagation in a SiC/Al nanocomposite reinforced with WS2-inorganic fullerene
nanoparticles. Compos. Struct. 2013, 96, 601–605.
14. Tehrani, M.; Luhrs, C.C.; Al-Haik, M.S.; Trevino, J.; Zea, H. Synthesis of WS2 nanostructures
from the reaction of WO3 with CS2 and mechanical characterization of WS2 nanotube
composites. Nanotechnology 2011, 22, 285714.
15. Naffakh, M.; Diez-Pascual, A.M.; Marco, C.; Ellis, G.J.; Gomez-Fatou, M.A. Opportunities and
challenges in the use of inorganic fullerene-like Nanoparticles to Produce Advanced Polymer
Nanocomposites. Prog. Polym. Sci. 2013, 38, 1163–1231.
16. Naffakh, M.; Diez-Pascual, A.M.; Gomez-Fatou, M.A. New Hybrid nanocomposites containing
carbon nanotubes, inorganic fullerene-like WS2 nanoparticles and poly(ether ether ketone)
(PEEK). J. Mater. Chem. 2011, 21, 7425–7433.
17. Naffakh, M.; Diez-Pascual, A.M.; Marco, C.; Gomez, M.A.; Jimenez, I. Novel melt-processable
poly (ether ether ketone) (PEEK)/inorganic fullerene-like WS2 nanoparticles for critical
applications. J. Phys. Chem. B 2010, 114, 11444–11453.
18. Diez-Pascual, A.M.; Naffakh, M.; Marco, C.; Ellis, G. Rheological and tribological properties
of carbon nanotube/thermoplastic nanocomposites incorporating inorganic fullerene-Like WS2
Nanoparticles. J. Phys. Chem. 2012, 116, 7959–7969.
19. Diez-Pascual, A.M.; Naffakh, M. Mechanical and thermal behaviour of isotactic polypropylene
reinforced with inorganic fullerene-like WS2 nanoparticles: Effect of Filler Loading and
Temperature. Mater. Chem. Phys. 2013, 141, 979–989.
20. Ma, T.; Zhang, T.; Gao, P.G.; Zhang, J.C. Synthesis and properties of ultrahigh molecular
weight polyethylene/WS2 nanoparticle fiber for bullet-proof materials. Chin. Sci. Bull. 2013, 58,
945–948.
21. Shneider, M.; Dodiuk, H.; Tenne, R.; Kenig, S. Nanoinduced morphology and enhanced
properties of epoxy containing tungsten disulfide nanoparticles. Polym. Eng. Sci. 2013, 53,
2624–2632.
22. Shneider, M.; Rapoport, L.; Moshkovich, A.; Dodiuk, H.; Kenig, S.; Tenne, R.; Zak, A.
Tribological performance of the epoxy-based composite reinforced by WS2 fullerene-like
nanoparticles and nanotubes. Phys. Status Solidi A 2013, 210, 2298–2306.
23. Zohar, E.; Baruch, S.; Shneider, M.; Dodiuk, H.; Kenig, S.; Tenne, R.; Wagner, H.D. The Effect
of WS2 nanotubes on the properties of epoxy-based nanocomposites. J. Adhes. Sci. Technol.
2011, 25, 1603–1617.
24. Shneider, M.; Dodiuk, H.; Kenig, S.; Tenne, R. The Effect of Tungsten Sulfide Fullerene-Like
Nanoparticles on the toughness of epoxy adhesives. J. Adhes. Sci. Technol. 2010, 24, 1083–1095.
25. Lalwani, G.; Kwaczala, A.T.; Kanakia, S.; Patel, S.C.; Judex, S.; Sitharaman, B. Fabrication
and characterization of three-dimensional macroscopic all-carbon scaffolds. Carbon 2013, 53,
90–100.
26. Franklin, N.; Dai, H. An enhanced CVD approach to extensive nanotube networks with
directionality. Adv. Mater. 2000, 12, 890–894.

47
27. Calvo, M.; Garcia, R.; Arenillas, A.; Suarez, I.; Moinelo, S. Carbon foams from coals.
A preliminary study. Fuel 2005, 84, 2184–2189.
28. Calvo, M.; Garcia, R.; Moinelo, S.R. Carbon foams from different coals. Energy Fuels
2008, 22, 3376–3383.
29. Chen, C.; Kennel, E.; Stiller, A.; Stansberry, P.; Zondlo, J. Carbon foam derived from various
precursors. Carbon 2006, 44, 1535–1543.
30. Chakrapani, N.; Wei, B.; Carrillo, A.; Ajayan, P.; Kane, R. Capillarity-driven assembly of
two-dimensional cellular carbon nanotube foams. Proc. Natl. Acad. Sci. USA 2004, 101,
4009–4012.
31. Chakrapani, N.; Wei, B.; Carrillo, A.; Ajayan, P.; Kane, R. Capillarity-driven assembly of
cellular carbon nanotube foams. Abstr. Pap. Am. Chem. Soc. 2004, 228, U481–U481.
32. Gallego, N.; Klett, J. Carbon Foams for Thermal Management. Carbon 2003, 41, 1461–1466.
33. Kaur, S.; Ajayan, P.M.; Kane, R.S. Design and characterization of three-dimensional carbon
nanotube foams. J. Phys. Chem. B 2006, 110, 21377–21380.
34. Dai, L.; Patil, A.; Gong, X.; Guo, Z.; Liu, L.; Liu, Y.; Zhu, D. Aligned nanotubes.
Chemphyschem 2003, 4, 1150–1169.
35. Atwater, M.A.; Mousavi, A.K.; Leseman, Z.C.; Phillips, J. Direct synthesis and characterization
of a nonwoven structure comprised of carbon nanofibers. Carbon 2013, 57, 363–370.
36. Safdari, M.; Al-Haik, M.S. Optimization of stress wave propagation in a multilayered
elastic/viscoelastic hybrid composite based on carbon fibers/carbon nanotubes. Polym. Compos.
2012, 33, 196–206.
37. Khalid, W.; Ali, M.S.M.; Dahmardeh, M.; Choi, Y.; Yaghoobi, P.; Nojeh, A.; Takahata, K.
High-aspect-ratio, free-form patterning of carbon nanotube forests using micro-electro-discharge
machining. Diamond Relat. Mater. 2010, 19, 1405–1410.
38. Qiu, L.; Liu, J.Z.; Chang, S.L.Y.; Wu, Y.; Li, D. Biomimetic superelastic graphene-based
cellular monoliths. Nat. Commun. 2012, 3, 1241.
39. Geim, A.K.; Novoselov, K.S. The rise of graphene. Nat. Mater. 2007, 6, 183–191.
40. Stankovich, S.; Dikin, D.A.; Dommett, G.H.B.; Kohlhaas, K.M.; Zimney, E.J.; Stach, E.A.;
Piner, R.D.; Nguyen, S.T.; Ruoff, R.S. Graphene-based composite materials. Nature 2006, 442,
282–286.
41. Naffakh, M.; Diez-Pascual, A.M.; Marco, C.; Ellis, G. Morphology and thermal properties of
novel poly (phenylene sulfide) hybrid nanocomposites based on single-walled carbon nanotubes
and inorganic fullerene-like WS2 nanoparticles. J. Mater. Chem. 2012, 22, 1418–1425.
42. Diez-Pascual, A.M.; Naffakh, M.; Gomez-Fatou, M.A. Mechanical and electrical properties of
novel poly(ether ether ketone)/carbon nanotube/inorganic fullerene-like WS2 hybrid
nanocomposites: Experimental measurements and theoretical predictions. Mater. Chem. Phys.
2011, 130, 126–133.
43. Diez-Pascual, A.M.; Naffakh, M. Inorganic Nanoparticle-Modified Poly(Phenylene
Sulphide)/Carbon Fiber Laminates: Thermomechanical Behaviour. Materials 2013, 6, 3171–3193.

48
44. Diez-Pascual, A.M.; Naffakh, M. Tuning the properties of carbon fiber-reinforced poly
(phenylene sulphide) laminates via incorporation of inorganic nanoparticles. Polymer 2012, 53,
2369–2378.
45. Diez-Pascual, A.M.; Naffakh, M.; Marco, C.; Ellis, G. Mechanical and electrical properties of
carbon nanotube/poly(phenylene sulphide) composites incorporating polyetherimide and
inorganic fullerene-like nanoparticles. Compos. Part A 2012, 43, 603–612.
46. Luhrs, C.C.; Daskam, C.D.; Gonzalez, E.; Phillips, J. Fabrication of a Low Density Carbon Fiber
Foam and Its Characterization as a Strain Gauge. Materials, submitted for publication, 2014.
47. Moberg, M. Carbon Fiber and Tungsten Disulfide Nanoscale Architectures for Armor
Applications. MS Thesis, Naval Postgraduate School, June 2012.
48. Feldman, Y.; Frey, G.; Homyonfer, M.; Lyakhovitskaya, V.; Margulis, L.; Cohen, H.; Hodes, G.;
Hutchison, J.; Tenne, R. Bulk synthesis of inorganic fullerene-like MS2 (M = Mo, W) from the
respective trioxides and the reaction mechanism. J. Am. Chem. Soc. 1996, 118, 5362–5367.
49. Tenne, R.; Homyonfer, M.; Feldman, Y. Nanoparticles of layered compounds with hollow cage
structures (inorganic fullerene-like structures). Chem. Mater. 1998, 3225–3238.
50. Tenne, R.; Remskar, M.; Enyashin, A.; Seifert, G. Inorganic nanotubes and fullerene-like
structures (IF). Carbon Nanotubes 2008, 111, 631–671.
51. Bar-Sadan, M.; Kaplan-Ashiri, I.; Tenne, R. Inorganic fullerenes and nanotubes: Wealth of
materials and morphologies. Eur. Phys. J.-Spec. Top. 2007, 149, 71–101.
52. Wiesel, I.; Arbel, H.; Albu-Yaron, A.; Popovitz-Biro, R.; Gordon, J.M.; Feuermann, D.; Tenne, R.
Synthesis of WS2 and MoS2 Fullerene-Like Nanoparticles from Solid Precursors. Nano Res.
2009, 2, 416–424.
53. Bastide, S.; Borra, J.; Duphil, D.; Levy-Clement, C. Synthesis of inorganic fullerenes and
nanoboxes of MoS2 and WS2 by spray pyrolysis. Abstr. Pap. Am. Chem. Soc. 2004, 228,
U482–U482.
54. Zink, N.; Pansiot, J.; Kieffer, J.; Therese, H.A.; Panthofer, M.; Rocker, F.; Kolb, U.; Tremel, W.
Selective synthesis of hollow and filled fullerene-like (IF) WS2 nanoparticles via metal-organic
chemical vapor deposition. Chem. Mater. 2007, 6391–6400.
55. Brooks, D.; Douthwaite, R.; Brydson, R.; Calvert, C.; Measures, M.; Watson, A. Synthesis of
inorganic fullerene (MS2, M = Zr, Hf and W) phases using H2S and N2/H2 microwave-induced
plasmas. Nanotechnology 2006, 17, 1245–1250.
56. Tenne, R. Inorganic nanotubes and fullerene-like nanoparticles. Nat. Nanotechnol. 2006, 1,
103–111.
57. Luhrs, C.C.; Garcia, D.; Tehrani, M.; Al-Haik, M.; Taha, M.R.; Phillips, J. Generation of carbon
nanofilaments on carbon fibers at 550 degrees C. Carbon 2009, 47, 3071–3078.
58. Pierson, H.O. Handbook of Carbon, Graphite, Diamond, and Fullerenes; Noyes Publications:
Park Ridge, NJ, USA, 1993.
59. Fleet, M.E. Crystal-Structure of Heazlewoodite, and Metallic Bonds in Sulfide Minerals.
Am. Mineral. 1977, 62, 341–345.
60. Parise, J.B. Structure of Hazelwoodite (Ni3S2). Acta Crystallogr. Sect. B 1980, 36, 1179–1180.

49
61. Villars, P.; Cenzual, K. Landolt-Bornstein—Group III Condensed Matter; Springer: Berlin,
Germany, 2010.
62. Xu, F.; Almeida, T.P.; Chang, H.; Xia, Y.; Wears, M.L.; Zhu, Y. Multi-walled carbon/IF-WS2
nanoparticles with improved thermal properties. Nanoscale 2013, 5, 10504–10510.
63. Schuffenhauer, C.; Wildermuth, G.; Felsche, J.; Tenne, R. How stable are inorganic
fullerene-like particles. Thermal analysis (STA) of inorganic fullerene-like NbS2, MoS2, and
WS2 in oxidizing and inert atmospheres in comparison with the bulk material. Phys. Chem.
Chem. Phys. 2004, 6, 3991–4002.
64. Chang, L.; Yang, H.; Fu, W.; Yang, N.; Chen, J.; Li, M.; Zou, G.; Li, J. Synthesis and thermal
stability of W/WS2 inorganic fullerene-like nanoparticles with core-shell structure. Mater. Res. Bull.
2006, 41, 1242–1248.
65. Dutta, A.; Penumadu, D.; Files, B. Nanoindentation testing for evaluating modulus and hardness
of single-walled carbon nanotube-reinforced epoxy composites. J. Mater. Res. 2004, 19,
158–164.
66. Gromowski, K. Glass Breakage—Nickel Sulfide Inclusions; Penn State: State College, PA, USA,
2010.
67. Flores, A.; Naffakh, M.; Diez-Pascual, A.M.; Ania, F.; Gomez-Fatou, M.A. Evaluating the
Reinforcement of Inorganic Fullerene-like Nanoparticles in Thermoplastic Matrices by
Depth-Sensing Indentation. J. Phys. Chem. C 2013, 117, 20936–20943.
68. Marcano, D.C.; Kosynkin, D.V.; Berlin, J.M.; Sinitskii, A.; Sun, Z.; Slesarev, A.; Alemany, L.B.;
Lu, W.; Tour, J.M. Improved Synthesis of Graphene Oxide. ACS Nano 2010, 4, 4806–4814.
69. Oliver, W.; Pharr, G. An Improved Technique for Determining Hardness and Elastic-Modulus
using Load and Displacement Sensing Indentation Experiments. J. Mater. Res. 1992, 7,
1564–1583.
70. Pharr, G.; Oliver, W.; Brotzen, F. On the Generality of the Relationship among Contact
Stiffness, Contact Area, and Elastic-Modulus during Indentation. J. Mater. Res. 1992, 7, 613–617.
71. Hay, J.; Herbert, E. Measuring the Complex Modulus of Polymers by Instrumented Indentation
Testing. Exp. Tech. 2013, 37, 55–61.

50

From Stable ZnO and GaN Clusters to Novel Double Bubbles
and Frameworks
Matthew R. Farrow, John Buckeridge, C. Richard A. Catlow, Andrew J. Logsdail,
David O. Scanlon, Alexey A. Sokol and Scott M. Woodley
Abstract: A bottom up approach is employed in the design of novel materials: first, gas-phase
“double bubble” clusters are constructed from high symmetry, Th, 24 and 96 atom, single
bubbles of ZnO and GaN. These are used to construct bulk frameworks. Upon geometry
optimization—minimisation of energies and forces computed using density functional theory—the
symmetry of the double bubble clusters is reduced to either C1 or C2, and the average bond lengths
for the outer bubbles are 1.9 Å, whereas the average bonds for the inner bubble are larger for ZnO
than for GaN; 2.0 Å and 1.9 Å, respectively. A careful analysis of the bond distributions reveals that
the inter-bubble bonds are bi-modal, and that there is a greater distortion for ZnO. Similar bond
distributions are found for the corresponding frameworks. The distortion of the ZnO double bubble
is found to be related to the increased flexibility of the outer bubble when composed of ZnO rather
than GaN, which is reflected in their bulk moduli. The energetics suggest that (ZnO)12@(GaN)48 is
more stable both in gas phase and bulk frameworks than (ZnO)12@(ZnO)48 and (GaN)12@(GaN)48.
Formation enthalpies are similar to those found for carbon fullerenes.
Reprinted from Inorganics. Cite as: Farrow, M.R.; Buckeridge, J.; Catlow, C.R.A.; Logsdail, A.J.;
Scanlon, D.O.; Sokol, A.A.; Woodley, S.M. From Stable ZnO and GaN Clusters to Novel Double
Bubbles and Frameworks. Inorganics 2014, 2, 248–263.
1. Introduction
A new class of materials is sought that will support the separation of electrons and holes typically
generated during photo-excitation by solar radiation. In this context, heterostructures of ZnO/GaN
attract particular interest, as such materials have great potential in a wide range of applications from
semiconductor optoelectronics to photo-catalysis [1–5].
Previous computational simulations [6–8] have predicted that both ZnO and GaN, at the nanoscale,
form clusters with a bubble architecture that are dramatically different from models cut from their
wurtzite bulk structures. Using ZnO and SiC as two simple examples, we have shown [9,10] how
individual bubbles can combine to form extended framework materials; alternative constructions and
the viability (or stability) of similar frameworks using bubbles as building blocks have also been
reported [6,9–11]. For framework structures, an increase in density is typically correlated with an
enhanced stability, which can be achieved by connecting appropriate building units. In our approach
to framework construction, we choose a new type of unit, the so-called “double bubble”, that are
themselves denser than single-shell bubbles and which are a preferred motif for larger sized
clusters [12].
For binary oxide and semiconductor II-VI and III-V materials with a 1-1 stoichiometry, fullerene
type structures have been the focus of materials modelling at the nanoscale in the last decade. This

51
interest has partly been spurred by reports of synthesis of (MX)n clusters of these materials, where
M denotes metals, or cations, and X represents anions, with the mass spectra of such systems showing
unexpected preference for certain sizes n. The preferred values of n are widely known as “magic
numbers” [8]. The stability of such clusters has been explained on thermodynamic grounds: the
binding energy per formula unit as a function of size having a minimum (i.e., the energy released on
cluster formation has a maximum). Alternative explanations have been proposed using: (i) a kinetic
argument based on whether the cluster growth or shrinkage is an energetically favourable process;
and (ii) a statistical argument: a particular cluster size may be realized in a greater number of atomic
configurations compared to others, and therefore is favoured entropically. For any particular
experiment, one or a combination of these factors may in fact be relevant.
Considering the atomic structure of stable clusters as a function of increasing size, we recognise
an evolution of basic structural units with increasing dimensionality: from 1D—sticks; to 2D—rings
and patchworks of rings; and then to 3D units that are initially composed of one layer—a tube or
bubble—and then multiple layers, and finally bulk like phases (those that could be stable or
metastable on the macroscopic scale). The bubble structures are also denoted in the literature as
cages, spheroids, or fullerenes. The above classification is also based on the atomic connectivity or
bonding, which increases with increased dimensionality. Using additional definitions given in
reference [13], perfect closed bubbles are an important subclass of single walled fullerene-like
clusters, in which each atom has three nearest neighbours, and the surface of such fullerenes is
composed of a patchwork of hexagonal faces that is wrapped in three dimensions by the introduction
of six “defects”, or tetragonal faces. The existence and stability of the fullerene-like inorganic
structures have been known from both theory and experiment for BN, ZnO, and MoS2 [14–16].
Perfect bubbles can also include larger patches with an even number of sides complemented by an
appropriate number of tetragonal faces. In contrast to carbon fullerenes, pentagonal faces are not
realised for the heterogeneous semiconductor class of compound, as they would require formation of
M–M and/or X–X bonds that are electrostatically unfavourable. Due to the ionic nature of bonding in
these materials, the charge disproportionation is not compatible with electron localisation, required
for metal-metal bond formation, or hole localisation, which stabilises di-oxygen or di-nitrogen species.
On further size increase (cluster growth), the appearance of layered structures becomes a possibility,
in which a smaller sized cluster unit is contained within a larger bubble structure. Indeed, such
structures have been discovered in molecular dynamical studies of ZnS, where the smallest is found
for n = 60: an n = 48 bubble forms a concentric shell around an n = 12 sodalite cage [12]. Although
both single bubbles have the same high symmetry Th point group, the double bubble can relax into
a lower symmetry form depending on the composition. The Th symmetry unit, however, can be
stabilised when this unit is used in frameworks that were constructed previously from the individual
single layered components.
In this paper, we investigate the different possible atomic structural relaxations of the double
bubble and the effect of mixing components of different compositions, for both the individual clusters
and the constructed frameworks.

52
2. Construction of Double Bubbles
High Symmetry Double Bubble Clusters as Secondary Building Units
We consider the 1-1 compounds that are predicted to have stable (lowest energy for a particular
size) and metastable bubble, or fullerene like, structures. Perfect versions of these structures are
composed of only three-coordinated atoms, sets of which create rings with an even number of sides
that can be visualised as one of the faces of the bubble; an odd number of sides is unlikely as this
would require at least one neighbouring pair of vertices of only cations or anions. Except for
the smallest sized clusters, in which the curvature of the layer is important, cluster configurations
containing one- or two-coordinated atoms are less stable than the perfect bubbles. Another
characteristic of the stable bubbles is that, typically, the number of tetragonal rings is minimised
(and, to a lesser extent, the distance between these should be maximised), while the number of
hexagonal rings is maximised. A layer consisting of only hexagonal rings has no curvature, and
therefore a perfect bubble of hexagons would require an infinite number of atoms (i.e., a 2D infinite
hexagonal sheet). To obtain a perfect bubble with a finite number of atoms, the sheet requires the
introduction of six tetragonal rings (Euler’s rule) as each tetragonal ring increases the curvature of
the sheet. Increasing the number of tetragonal rings results in open (as opposed to closed) perfect
bubbles, which contain much larger rings, e.g., octagonal, assuming the chemistry of the compound
does not favour the bonding and coordination required for the formation of cuboids, i.e., cuts from
rock salt.
The higher symmetry configurations of the perfect bubbles are typically found to be more stable.
As high symmetry cluster structures are only possible for certain sizes, they are not only the stable
state for their particular size, but usually have a greater stability than clusters of neighbouring sizes.
In our previous studies [6,9,10] we have, therefore, focused our attention on families of high
symmetry structures, and, in particular, those with symmetry Th, Td and T. These (MX)n structures
are found if n = 4 (Td), 12 (Th), 16 (Td), 28 (T), 36 (Td), 48 (Th), 64 (Td), … Larger Th and Td clusters
include n = 108 and 192 and n =100, 144 and 196, respectively; examples are shown in Figure 1.
These clusters can be visualised as truncated octahedra, where there is one tetragonal ring of the
bubble at each of the six truncated corners, and the hexagonal patchworks form the octahedron’s
faces. In this morphology, the distance between all tetragonal rings is maximised for a given size n,
and the separation increases monotonically with n.
Figure 1. Models of high symmetry (MX)n bubbles with: (a) n = 12 with symmetry Th;
(b) n = 48 with symmetry Th; (c) n = 28, 36 and 48 with octahedra superimposed.

53
Smaller bubble clusters can be readily generated using an appropriate global optimiser (e.g., one
based on Monte Carlo basin hopping [17] or genetic algorithm [18] routines), and once one has
determined the relationship between them, the larger bubble clusters can be constructed by simply
increasing the number of rows of hexagonal rings in each face or edge of the octahedron. For
example, each octahedron’s edge of a Td bubble links a side of two tetragonal rings via a “ladder” of
m hexagonal rings (one hexagon wide)üsee the highlighted ladder in Figure 2aüwith the remaining
hexagonal rings completing the faces of the octahedron. Note that the line of the octahedron edge
bisects the rings of this ladder and that the tetragonal ring is out of phase with the tetragonal face
created by truncating the octahedron. Constructions with m = 0, 1, 2, 3, 4 and 5 corresponds to perfect
bubbles at n = 4, 16, 36, 64, 100, 144 and 196, respectively. In contrast, each octahedron edge in a
Th bubble links the corner of two tetragonal rings via m + 1 M–X sticks that are separated by m
hexagonal rings, forming an alternating linear pattern. Each stick is actually a shared side of two
hexagonal rings, with each ring part of a hexagonal patchwork that covers a face of the octahedron.
These sticks form a line segment of the octahedron edge (see highlighting in Figure 2b), and this line
bisects opposite angles of the hexagonal rings, rather than opposite sides, and the tetragonal rings are
in-phase with the tetragonal face created by truncating the octahedron. The smallest Th bubble,
m = 0, or n = 12, has just one stick between two neighbouring tetragonal rings. The next smallest
size Th bubble, m = 1, or n = 48, is constructed using two sticks and one hexagonal ring; then, for
m = 2, or n = 108, there are three sticks and two hexagonal rings. Comparing the growth of the
octahedron edges, it is evident why there are more bubbles with Td rather than Th symmetry.
Figure 2. Models of high symmetry (MX)n bubbles with: (a) n = 64, symmetry Td and
the ladder of hexagonal rings, highlighted in yellow, that corresponds to one of the
twelve edges of an octahedron; (b) n = 48, symmetry Th with a fragment that corresponds
to one of twelve edges highlighted in yellow; (c) n = 48, symmetry Th with one of
twelve patchworks that correspond to the octahedron side highlighted in purple.

As discussed in the Introduction, framework structures with an increased density are typically
more stable [19,20]. We therefore choose to investigate double bubbles that are composed of high
symmetry perfect bubbles as these are more dense. Two bubbles are combined by inserting the
smaller bubble inside the larger; aligned with the same centre of mass and identical direction of
orthogonal axes, with each axis passing through the centre of mass and the centre of opposite
truncated corners, or tetragonal faces. The rotation of these tetragonal faces about the octahedral axes

54
is dependent upon the symmetry of the cluster. When clusters have Th and Td symmetry this rotation
is 45° out-of-phase, and, if one of the bubbles has T symmetry, between 0° and 45°. For stability, the
best match is obtained when the inner and outer bubbles are taken from the set of Th bubbles and the
highest density obtained by combining the smallest two of these: n = 12 (a sodalite cage) and
n = 48 [10].
In the MX bulk phase considered here, the atoms are four-coordinated tetrahedra, so the stability
of the double bubbles will improve if M–X linkages between layers are found. These linkages can
be expected to be located between aligned pairs of hexagonal patchworks that form the faces of the
octahedra rather than between the truncated corners. For the n = 60 double bubble, the inner n = 12
bubble has one hexagonal ring on each face, whereas the outer n = 48 bubble is composed of a
patchwork of five and a half hexagonal rings; a central hexagonal ring that is linked via one
hexagonal ring to each of three nearest tetragonal rings and three hexagonal rings that are shared
with neighbouring faces of the octahedron (see Figure 2c). Importantly, the central hexagonal ring
can bond with the hexagonal ring of the inner bubble; see Figure 3. Analogous to our structures,
experimentally observed cages of boron nitride (BN), [15,16] and molybdenum sulphide (MoS2) [21,22]
have been reported to be constructed from four and six (hexagonal) membered ring building units.
CdSe cage structures have been experimentally observed to be stable and formed from truncatedoctahedra [23]. DFT calculations on cage structures of CdSe that are similar to our structures have
also been reported [24].
Figure 3. Models of the n = 60 Th double bubble, with inter-layer links between the inner
n = 12 sodalite cage and the eight hexagonal rings that are in the centres of the
octahedron faces of the outer n = 48 bubble, highlighted using ball-and-sticks rather than
line representation for: (a) no bridging links; (b) four bridging links; and (c) all eight
bridging links.

The first framework is constructed from Th bubble (sodalite cage) secondary building units
(SBUs) of (ZnO)12 and (GaN)12, see Figure 1a. As the typical Zn–O and Ga–N bond lengths are
similar (1.98 Å and 1.95 Å in their ground state wurtzite form), their respective SBUs are also similar
in size. Consider each SBU as an octahedron. By corner sharing the octahedra, and assuming an
equal number of SBUs for each compound, we construct an fcc, rock-salt like lattice, as shown in
Figure 4c. The second framework is constructed from the n = 60 Th double bubbles; see Figures 1c
and 4b. Again, imagining each SBU as an octahedron, but rather than corner sharing they are now
stacked so that they share edges, each double bubble is surrounded by twelve others (see Figure 4d),

55
and each edge of the outer bubble is one bond length from an edge of a neighbouring bubble forming
an n = 6 double ring (a drum) and two n = 2 rings. Each tetragonal ring of an outer n = 48 bubble
combines with five others to form an n = 12 Th bubble, i.e., the void is a sodalite cage. The inner
sodalite cage of each double bubble is formed from (i) the same compound and (ii) two compounds,
which we alternate.
We start the double bubble construction from two relaxed single bubbles. If the distance between
each inner hexagonal ring and its corresponding central hexagonal ring of the outer bubble is
approximately a typical M–X bond length, then we shall refer to this as an ideal match, and the
relaxed double bubble is expected to maintain Th symmetry. Whether there is an ideal match depends
on the composition: if the two layers are of the same composition and there is not an ideal match
then the inner bubble is too small. The outer eight planes of hexagonal patchworks, or octahedron
faces, have more flexibility than the corners. During a geometry relaxation of the double bubble, the
central hexagonal ring of each outer patchwork can move inward, maintaining the Th symmetry or,
due to the repulsion between neighbouring patchworks, only the central hexagons from alternate
patchworks, i.e., four of the eight, move inwards reducing the symmetry to T; see Figure 3.
For all ZnO/GaN compositions investigated here, n = 60 double bubble structures of high
symmetry (Th and T) were constructed and then geometry optimised. Low symmetry structural
distortions were allowed in the optimisation process in order to find the lowest energy double
bubble configuration.
Figure 4. Ball and stick models of two framework structures. (a) Constructed from Th
bubbles of (GaN)12 and (ZnO)12; (b) Constructed from Th double bubbles of (ZnO)48 and
(GaN)12; (c) the same structure as (a) but with each (GaN)12 coloured red and each
(ZnO)12 coloured blue (lighter/darker shades used in the front/back row); (d) the same
structure as (b) but with each (GaN) 12 hidden and each (ZnO)48 uniquely coloured.

56
3. Results and Discussion
3.1. Double Bubble Clusters
With each layer (single bubble) composed of either ZnO or GaN, there are four possible n = 60
double bubble structures that can be constructed using the procedure discussed in Section 2.
The cluster structures are relaxed so as to minimise the energy, which is initially defined using a
semi-empirical potentials and then, during a final refinement stage, using density functional theory
(DFT); see Sections 4.1 and 4.2 for details. The four double bubble clusters consist of: (a) only zinc
oxide, denoted (ZnO)12@(ZnO)48; (b) only gallium nitride, denoted (GaN)12@(GaN)48; (c) a gallium
nitride sodalite cage inside a zinc oxide bubble, denoted (GaN)12@(ZnO)48; and the inverse (d) a zinc
oxide sodalite cage inside a gallium nitride bubble, denoted (ZnO)12@(GaN)48. During geometry
optimization, although the high Th and T symmetry that is maintained when semi-empirical
calculations are employed, there is a reduction of symmetry for all four systems to Cn, where n = 1
or 2. As reported in Table 1, double bubble clusters with internal (ZnO)12 sodalite cages adopt C2
symmetry, whereas those that had gallium nitride sodalite cages adopt C1 symmetry—i.e., there is
no symmetry in those structures. The average relaxed bond lengths, separated into inner-bubble
bonds, outer-bubble bonds, and inter-layer bubble bonds (M–X bonds connecting the inner to the
outer bubbles) are also reported in Table 1. The average bond lengths of zinc oxide and gallium nitride
are similar; although the average bond length for zinc oxide inner bubbles are slightly greater than
the average bond lengths of gallium nitride inner bubbles.
Now, consider the distribution of bond lengths, G(x), using a Gaussian broadening function for
each bond length, which is normalised to the number of linkages between the inner and outer bubble
(N = 48 for our n = 60 double bubbles):
ே

‫ܩ‬ሺ‫ݔ‬ሻ ൌ ‫ ܥ‬෍ expሺܾ௜ െ ‫ݔ‬ሻଶ Ȁʹߪ ଶ

(1)

௜ୀଵ

C is a normalising constant, bi is the length of bond i, and ı is the dispersion (width) of the Gaussian
function. This function is plotted in Figure 5 for two values of ı: 0.02 Å (red line) and
0.10 Å (blue line). The greater value of ı allows the resolution of two distinct peaks for the systems
of interest. These two peaks are reported in Table 1, labelled as A and B inter-bubble bond distances.
Table 1. Structural parameters of double bubble clusters, where Douter is the mean distance
between M–X atoms in the outer bubble, Dinner is the mean distance between M–X atoms
in the inner bubble, and Dinter is the distance between the inner and outer bubbles.
(Number in parentheses indicates standard error.)
System

Symmetry

Douter (Å)

Dinner (Å)

(GaN)12@(ZnO)48
(ZnO)12@(GaN)48
(ZnO)12@(ZnO)48
(GaN)12@(GaN)48

C1
C2
C2
C1

1.92
1.89
1.93
1.93

1.93
1.96
1.95
1.92

Dinter (Å)
A
2.05 (0.1)
2.13 (0.1)
2.10 (0.0)
2.10 (0.1)

B
3.08 (0.1)
2.94 (0.1)
2.94 (0.2)
2.94 (0.1)

57
Figure 5. Bond distribution plots for the double bubble cluster systems. Red line:
Dispersion of Gaussian = 0.02, Blue line: Dispersion of Gaussian = 0.1.

We observe that the pure double-bubble clusters have similar bond distributions, and notice only
a difference of a small peak at 2 Å for the pure GaN system, which appears as a shoulder on the
2.3 Å peak in the pure ZnO system. We mark this shoulder (at approximately 2.25 Å) as the split of
the distribution into bonded and non-bonded linkages. The number of bonded linkages, in fact, is
constant for all the systems except for (ZnO)12@(GaN)48 and has a value of twenty-four, which is
related to the ideal T symmetry octahedral shape. In this type of linking, two extremes can be
possible: four of the eight hexagonal rings form drums with the outer bubble, or only half of the
possible bonds are formed in such drums—see Figure 3b. The (ZnO)12@(GaN)48 double bubble, in
contrast, has only twenty-two bonded linkages, which is not due to an inner bubble displacement
from the centre of the outer bubble but is caused by a distortion in the outer bubble. To relate these
observations to macroscopic properties of the systems, we considered the deformation as seen from
the displacement of the centre of mass (COM) of the inner bubbles with respect to the outer bubbles,
and their normalized second moments of atom distribution, as given in Table 2.
Table 2. Centre of mass (COM) differences and normalised second moments of atom
distributions for the double bubble clusters (x, y, z coordinates).
System
(GaN)12@(ZnO)48
(ZnO)12@(GaN)48
(ZnO)12@(ZnO)48
(GaN)12@(GaN)48

COM difference
(COMOuter–COMInner)
0.00, 0.11, 0.05
0.00, 0.00, 0.01
0.00, 0.00, 0.04
0.00, 0.01, í0.06

Normalised second moments of atom distribution
Inner
Outer
1.05, 1.01, 0.94
1.05, 1.01, 0.94
1.02, 1.00, 0.98
1.01, 1.00, 0.99
1.09, 1.00, 0.91
1.05, 1.00, 0.95
1.04, 1.01, 0.96
1.04, 1.01, 0.95

The largest COM displacement is seen in the (GaN)12@(ZnO)48 system and smallest in the inverse
(ZnO)12@(GaN)48 system. The deformation is also lowest in the latter system, but has the largest

58
values in pure ZnO. We explain this behaviour by considering the relative sizes of the inner and outer
bubbles: the larger ZnO inner bubble fills in the space offered by the smaller GaN outer bubble better
than the GaN counterpart. An additional point to take into account is the greater flexibility of the
ZnO bubbles as compared with GaN: the size mismatch between the inner and outer bubble is
accommodated easier by ZnO, the bubbles of which show the greater deformations. This flexibility
is also seen in the bulk framework systems as discussed in Section 3.2 below. We show in Table 3
the energy of association, EAssoc, calculated as the difference in total energy of the double bubble
cluster from their moieties, i.e., the n1 = 12 and n2 = 48 bubbles, and formation enthalpy, Hf,:
‫ܪ‬௙ ൌ 

‫ܧ‬஽஻ െ ሾ݊ଵ ሺ‫ܧ‬௔ ሻ ൅ ݊ଶ ሺ‫ܧ‬௕ ሻሿ
ǡ
݊ଵ ൅ ݊ଶ

(1)

where EDB is the total energy of the double bubble cluster, Ea and Eb are the total energies of the pure
bulk wurtzite structures, where a and b can be ZnO or GaN. We find that the formation of the double
bubble systems is most favourable for the (GaN)12@(ZnO)48 system and least favourable for the
inverse system, and that the pure double bubbles have equal formation energies.
Table 3. Energy of association, EAssoc of single-shell cages and enthalpy of formation,
Hf per atom for double bubble clusters as defined in Equation (2).
System
(GaN)12@(ZnO)48
(ZnO)12@(GaN)48
(ZnO)12@(ZnO)48
(GaN)12@(GaN)48

EAssoc (kJ/mol)
í11.27
í8.17
í9.38
í9.16

Hf (kJ/mol)
78.32
104.55
68.50
116.18

We find that the formation of the homogeneous (ZnO)12@(ZnO)48 system is the most favourable
closely followed by the heterogeneous (GaN)12@(ZnO)48 system compared to the homogeneous bulk
wurtzite phases. Systems that have an outer-bubble of GaN are less likely to form when compared
with bulk (at zero temperature). If we consider the mixing energies per atom:
‫ܧ‬௠௜௫ ൌ

‫ܧ‬஽஻ െ ሾͲǤͺሺ‫ܧ‬௔ ሻ ൅ ͲǤʹሺ‫ܧ‬௕ ሻሿ
ͳʹͲ

(2)

where Ea and Eb are the energies of the pure double bubbles that make up the mixed system, we find
that the energy of mixing for (GaN)12@(ZnO)48 and (ZnO)12@(GaN)48 are 0.07 kJ/mol and
í0.96 kJ/mol respectively.
3.2. Double Bubble Frameworks
We took the double bubble frameworks that were constructed using the procedure discussed in
Section 2, and also corresponding systems of pure ZnO and GaN, and optimised their geometry (see
Section 4.2 for details). The structural analysis performed in Section 3.1 was repeated for these
frameworks. The calculated average bond lengths are presented in Table 4, again separated into
inner-bubble bonds, outer-bubble bonds and, inter-layer bubble bonds (bonds connecting the inner
to the outer bubbles). The graphs of the corresponding bond-length distribution analysis can be seen

59
in Figure 6. Table 4 also has two additional pieces of information—the lattice parameter and the bulk
modulus, which are available for these extended crystalline frameworks. Similar to the double bubble
clusters, we find that the bonds in the ZnO inner bubble are slightly larger than the equivalent GaN
bonds. In the framework systems this has a noticeable effect on the bond distribution: when the inner
bubble is composed of ZnO, the bond length distribution is no longer bi-modal but has a single peak
at 2.3 Å (Figure 6), which, similar to the double bubble clusters, is due to the larger ZnO bubble
occupying the space inside the outer bubble. In this case, however, as the outer bubble is in a
framework, it is unable to deform to the same degree as the gas-phase cluster, and only a single peak
forms in the bond length distribution.
We also see that the lattice parameters for the double bubble frameworks are, as expected, related
to the composition of the outer bubble, and that when the outer bubble is composed of ZnO the lattice
parameters are larger. Comparing the bulk moduli of the systems, we find that the pure GaN system
is the least compressible, whereas the pure ZnO system is the most. This agrees with the double
bubble cluster findings, where the ZnO systems exhibit the greatest deformations. Table 5 shows the
corresponding structural parameters for the wurtzite systems used in the framework analysis, and the
bulk modulus of the GaN system is much larger than that of the ZnO.
Table 4. Structural parameters of double bubble frameworks. (Number in parentheses
indicates standard error).
System
(GaN)12@(ZnO)48
(ZnO)12@(GaN)48
(ZnO)12@(ZnO)48
(GaN)12@(GaN)48

Lattice
parameter (Å)
19.26
18.84
19.26
18.94

Bulk modulus
(GPa)
78.84
77.88
69.78
101.92

Douter (Å)

Dinner (Å)

1.96
1.90
1.94
1.93

1.94
2.01
2.00
1.94

Dinter (Å)
A
B
2.08 (0.1)
2.96 (0.2)
2.27 (0.1)
2.26 (0.0)
2.18 (0.2)
3.04 (0.1)

Figure 6. Bond distribution plots for the double bubble frameworks. Red line: Dispersion
of Gaussian = 0.02, Blue line: Dispersion of Gaussian = 0.1.

60
Table 5. Structural parameters of wurtzite phases.
System
ZnO
GaN

Lattice parameter, a (Å)
3.251
3.187

Lattice parameter, c (Å)
5.204
2.760

Bulk modulus (GPa)
146.136
188.367

u
0.382
0.378

We observe that the framework system of (GaN)12@(ZnO)48 has a similar inter-bond length
distribution to that found in the double bubble systems which is again due to the fact that the smaller
GaN cage has more freedom to move inside the larger ZnO bubble. The (GaN)12@(ZnO)48 system
has a more clearly defined bi-modal distribution for the framework systems than observed for the
double bubble cluster systems, and is likely due to reduced degrees of structural freedom with the
extended bulk framework. Table 6 gives the formation enthalpies for the framework systems, and
although these energies are positive i.e., unfavourable with respect to the pure bulk wurtzite phases,
they are small enough to be accessible at experimental temperatures, and are comparable to the
formation of fullerene (C60) with respect to bulk carbon (ca. 40 kJ/mol) [25,26]. The pure GaN double
bubble framework was found to be the least likely to form, whereas the (GaN)12@(ZnO)48 framework
was found to be most favourable—again agreeing with the formation energy double bubble
cluster findings.
Table 6. Enthalpy of formation per atom of double bubble frameworks as defined in
Equation (2).
System
(GaN)12@(ZnO)48
(ZnO)12@(GaN)48
(ZnO)12@(ZnO)48
(GaN)12@(GaN)48

HF/atom (kJ/mol)
13.17
21.46
18.54
27.71

When we compare the energies of mixing (using Equation (3)) we find that the energies per atom
for (GaN)12@(ZnO)48 and (ZnO)12@(GaN)48 are 1.98 kJ/mol and í2.61 kJ/mol, respectively.
4. Computational Detail
4.1. Interatomic Potentials Calculations
We have used the semi-classical GULP code [27] to construct and optimise ZnO structures
prior to refining them with DFT. We employed polarisable shell inter-atomic potentials
parameterised for bulk ZnO [7,28] in the double bubble cluster and framework calculations. The
resulting atomic structures were used not only for ZnO, but also GaN and mixed ZnO/GaN structures;
note that the bond lengths in GaN are very similar those in ZnO (see Tables 1–3 in Section 3), and
we only required approximate initial atomic coordinates for input into the DFT calculations, as
outlined below.

61
4.2. Density Functional Theory Calculations
In all of the ab initio calculations, we have used the solids-corrected Perdew-Burke-Ernzerhof
(PBEsol) GGA exchange-correlation functional [29,30], and all structural optimisations were
deemed converged when the atomic forces were less than 0.01 eV/Å.
A natural choice for the calculations on the double bubble clusters, due to its computational
efficiency, is the DFT code FHI-aims [31] as it uses numeric atom-centred basis sets. These
calculations were performed with the species defaults for the “tight” basis sets for accuracy (energies
converged to 1 meV/atom) and with scalar ZORA relativistic treatment [32]. We have used the
plane-wave DFT code VASP [33–36] to determined the equilibrium structures of the double bubble
based framework (extended crystal systems—see Section 2), and, for comparison, wurtzite bulk ZnO
and GaN. Within VASP, we employed the projector augmented wave (PAW) method [37] to describe
the interactions between the cores (Zn:[Ar], Ga:[Ar], O:[He] and N:[He]) and the valence electrons.
To determine the equilibrium bulk structures avoiding the problem of Pulay stress, we have
optimised the atomic coordinates at a series of different volumes, and fitted the resulting energy
versus volume data to the Murnaghan equation of state.
We have found that for the framework systems, an energy cut-off of 500 eV, and Monkhorst-Pack
k-point meshes of 8 × 8 × 6 and 1 × 1 × 1 for, respectively, the pure bulk wurtzite systems, and
the (A)12@(B)48 systems, where A and B stand for either ZnO or GaN, provide convergence in
total energy up to 10í5 eV for the framework systems, which is comparable with our double bubble
cluster calculations.
5. Conclusions
We have constructed double-bubble clusters and frameworks of ZnO and GaN from a bottom
up approach from cage structures analogous to fullerenes formed from hexagonal building
units [15,16,22]. The four systems we have considered, (GaN)12@(ZnO)48, (ZnO)12@(GaN)48,
(ZnO)12@(ZnO)48 and (GaN)12@(GaN)48, were first geometry optimized using a semi-empirical
potential within the GULP code and then refined using FHI-aims (for the double bubble clusters) or
VASP (for the frameworks) at the DFT level of theory using the PBEsol exchange-correlation
functional. We found that although the average bond lengths of both ZnO and GaN are similar, the
average bond lengths for ZnO inner bubbles were larger than the GaN inner bubbles of both the
double bubble cluster systems and the frameworks. This relative size difference, we believe, means
that the larger ZnO inner bubble fills in the space offered by the smaller GaN outer bubble better
than the GaN counterpart. In addition, we found that the greater flexibility of the ZnO bubbles from
calculations of bulk moduli, as compared with that of GaN bubbles, means that the size mismatch
between the inner bubble and outer bubble is more readily accommodated by ZnO. Furthermore, the
structural analysis of the pure ZnO double bubbles also showed the greater deformations. The
average M-X inter-bubble bonds were found to exhibit a bi-modal distribution for both clusters and
frameworks, except for the pure ZnO and (ZnO)12@(GaN)48 framework systems. These single-peak
distributions were due to the larger ZnO inner bubble that has less freedom to move than in the
inverse systems. The association energies of the double bubble clusters show that the systems

62
investigated here are favourable when compared to individual bubbles, although when compared to
bulk wurtzite phases, the clusters are less favourable.
The standard formation enthalpies for the framework systems are lower than those of carbon
fullerenes. Therefore, we suggest that these double bubble systems should be thermodynamically
accessible and could provide valuable material properties in the future.
Acknowledgments
We thank kindly our former collaborators Said Hamad, Eleonora Spano, Stefan T. Bromley,
Stephen A. Shevlin, Matthew B. Watkins, and Abdullah A. Al-Sunaidi, discussions with who have
been inspiring and instrumental for us undertaking this research. We also thank EPSRC for providing
the funding for Matthew Farrow and Scott Woodley on grant numbers EP/I03014X9 and
EP/K038958; John Buckeridge and Alexey Sokol on grant number EP/IO1330X; and Andrew
Logsdail on grant numbers EP/I030662/1 and EP/K038419/1. The authors also acknowledge the use
of the UCL Legion High Performance Computing Facility (Legion@UCL) and associated support
services; the IRIDIS cluster provided by the EPSRC funded Centre for Innovation (EP/K000144 and
EP/K000136); this work made use of the facilities of HECToR and ARCHER, the UK’s national
high-performance computing service through membership of the UK’s HPC Materials Chemistry
Consortium, which is funded by EPSRC (EP/L000202).
Author Contributions
The structures were constructed by Scott M. Woodley. The interatomic potential calculations were
performed by Scott M. Woodley and Alexey A. Sokol. The double bubble cluster calculations were
performed by Matthew R. Farrow. The double bubble framework calculations were performed by
John Buckeridge. The literature was researched by Andrew J. Logsdail and David O. Scanlon.
Expertise both in relevant materials science and methodology was provided by C. Richard A. Catlow
and Scott M. Woodley. Vital contributions to simulations design, the data analysis and preparation
of the manuscript were made by all of the authors.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.
2.

3.

Fujishima, A.; Honda, K. Electrochemical photolysis of water at a semiconductor electrode.
Nature 1972, 238, 37–38.
Vispute, R.D.; Talyansky, V.; Choopun, S.; Sharma, R.P.; Venkatesan, T.; He, M.; Tang, X.;
Halpern, J.B.; Spencer, M.G.; Li, Y.X.; et al. Heteroepitaxy of ZnO on GaN and its implications
for fabrication of hybrid optoelectronic devices. Appl. Phys. Lett. 1998, 73, 348–350.
Nakamura, S.; Pearton, S.; Fasol, G. The Blue Laser Diode: The Complete Story;
Springer-Verlag: Berlin, Germany, 2000.

63
4.

5.

6.

7.

8.
9.

10.

11.
12.
13.

14.
15.
16.
17.
18.

19.

Yu, Q.-X.; Xu, B.; Wu, Q.-H.; Liao, Y.; Wang, G.-Z.; Fang, R.-C.; Lee, H.-Y.; Lee, C.-T.
Optical properties of ZnO/GaN heterostructure and its near-ultraviolet light-emitting diode.
Appl. Phys. Lett. 2003, 83, 4713.
Zhu, H.; Shan, C.-X.; Yao, B.; Li, B.-H.; Zhang, J.-Y.; Zhang, Z.-Z.; Zhao, D.-X.; Shen, D.-Z.;
Fan, X.-W.; Lu, Y.-M.; et al. Ultralow-threshold laser realized in zinc oxide. Adv. Mater. 2009,
21, 1613–1617.
Shevlin, S.A.; Guo, Z.X.; van Dam, H.J.J.; Sherwood, P.; Catlow, C.R.A.; Sokol, A.A.;
Woodley, S.M. Structure, optical properties and defects in nitride (III–V) nanoscale cage
clusters. Phys. Chem. Chem. Phys. 2008, 10, 1944–1959.
Catlow, C.R.A.; French, S.A.; Sokol, A.A.; Al-Sunaidi, A.A.; Woodley, S.M. Zinc oxide: A
case study in contemporary computational solid state chemistry. J. Comput. Chem. 2008, 29,
2234–2249.
Catlow, C.R.A.; Bromley, S.T.; Hamad, S.; Mora-Fonz, M.; Sokol, A.A.; Woodley, S.M.
Modelling nano-clusters and nucleation. Phys. Chem. Chem. Phys. 2010, 12, 786–811.
Watkins, M.B.; Shevlin, S.A.; Sokol, A.A.; Slater, B.; Catlow, C.R.A.; Woodley, S.M. Bubbles
and microporous frameworks of silicon carbide. Phys. Chem. Chem. Phys. 2009, 11,
3186–3200.
Woodley, S.M.; Watkins, M.B.; Sokol, A.A.; Shevlin, S.A.; Catlow, C.R.A. Construction of
nano- and microporous frameworks from octahedral bubble clusters. Phys. Chem. Chem. Phys.
2009, 11, 3176–3185.
Carrasco, J.; Illas, F.; Bromley, S.T. Ultralow-density nanocage-based metal-oxide polymorphs.
Phys. Rev. Lett. 2007, 99, 235502.
Hamad, S.; Catlow, C.R.A.; Spanó, E.; Matxain, J.M.; Ugalde, J.M. Structure and properties of
ZnS nanoclusters. J. Phys. Chem. B 2005, 109, 2703–2709.
Al-Sunaidi, A.A.; Sokol, A.A.; Catlow, C.R.A.; Woodley, S.M. Structures of zinc oxide
nanoclusters: As found by revolutionary algorithm techniques. J. Phys. Chem. C 2008, 112,
18860–18875.
Behrman, E.C.; Foehrweiser, R.K.; Myers, J.R.; French, B.R.; Zandler, M.E. Possibility of
stable spheroid molecules of ZnO. Phys. Rev. A 1994, 49, R1543–R1546.
Jensen, F.; Toftlund, H. Structure and stability of C24 and B12N12 isomers. Chem. Phys. Lett.
1993, 201, 89–96.
Golberg, D.; Bando, Y.; Stephan, O.; Kurashima, K. Octahedral boron nitride fullerenes formed
by electron beam irradiation. Appl. Phys. Lett. 1998, 73, 2441–2443.
Woodley, S.M. Applications of Evolutionary Computation in Chemistry; Springer: Berlin,
Germany, 2004; Volume 110.
Woodley, S.M. Prediction of crystal structures using evolutionary algorithms and related
techniques. In Applications of Evolutionary Computation in Chemistry; Springer-Verlag: Berlin,
Germany, 2004; Volume 110, pp. 95–132.
Foster, M.D.; Friedrichs, O.D.; Bell, R.G.; Paz, F.A.A.; Klinowski, J. Structural evaluation of
systematically enumerated hypothetical uninodal zeolites. Angew. Chem. 2003, 115,
4026–4029.

64
20. Zwijnenburg, M.A.; Illas, F.; Bromley, S.T. Apparent scarcity of low-density polymorphs of
inorganic solids. Phys. Rev. Lett. 2010, 104, 175503.
21. Enyashin, A.N.; Gemming, S.; Bar-Sadan, M.; Popovitz-Biro, R.; Hong, S.Y.; Prior, Y.; Tenne, R.;
Seifert, G. Structure and stability of molybdenum sulfide fullerenes. Angew. Chem. Int. Ed.
2007, 46, 623–627.
22. Parilla, P.A.; Dillon, A.C.; Jones, K.M.; Riker, G.; Schulz, D.L.; Ginley, D.S.; Heben, M.J.
The first true inorganic fullerenes? Nature 1999, 397, 114.
23. Kasuya, A.; Sivamohan, R.; Barnakov, Y.A.; Dmitruk, I.M.; Nirasawa, T.; Romanyuk, V.R.;
Kumar, V.; Mamykin, S.V.; Tohji, K.; Jeyadevan, B.; et al. Ultra-stable nanoparticles of CdSe
revealed from mass spectrometry. Nat. Mater. 2004, 3, 99–102.
24. Botti, S.; Marques, M.A.L. Identification of fullerene-like cdse nanoparticles from optical
spectroscopy calculations. Phys. Rev. B 2007, 75, 035311.
25. Diogo, H.P.; da Piedade, M.E.M.; Dennis, T.J.S.; Hare, J.P.; Kroto, H.W.; Taylor, R.;
Walton, D.R.M. Enthalpies of formation of buckminsterfullerene (C60) and of the parent ions
C60+, C602+, C603+ and C60í. J. Chem. Soc. Faraday Trans. 1993, 89, 3541–3544.
26. Curl, R.F.; Haddon, R.C. On the formation of the fullerenes. Philos. Trans. 1993, 343, 19–32.
27. Gale, J.D.; Rohl, A.L. The general utility lattice program (gulp). Mol. Simul. 2003, 29,
291–341.
28. Whitmore, L.; Sokol, A.A.; Catlow, C.R.A. Surface structure of zinc oxide (1010), using an
atomistic, semi-infinite treatment. Surf. Sci. 2002, 498, 135–146.
29. Perdew, J.P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple.
Phys. Rev. Lett. 1996, 77, 3865–3868.
30. Perdew, J.P.; Ruzsinszky, A.; Csonka, G.I.; Vydrov, O.A.; Scuseria, G.E.; Constantin, L.A.;
Zhou, X.; Burke, K. Restoring the density-gradient expansion for exchange in solids and
surfaces. Phys. Rev. Lett. 2008, 100, 136406.
31. Blum, V.; Gehrke, R.; Hanke, F.; Havu, P.; Havu, V.; Ren, X.; Reuter, K.; Scheffler, M.
Ab initio molecular simulations with numeric atom-centered orbitals. Comput. Phys. Commun.
2009, 180, 2175–2196.
32. Van Lenthe, E.; Baerends, E.J.; Snijders, J.G. Relativistic total energy using regular approximations.
J. Chem. Phys. 1994, 101, 9783.
33. Kresse, G.; Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Rev. B 1993, 47,
558–561.
34. Kresse, G.; Hafner, J. Ab initio molecular-dynamics simulation of the liquid-metal–amorphoussemiconductor transition in germanium. Phys. Rev. B 1994, 49, 14251–14269.
35. Kresse, G.; Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and
semiconductors using a plane-wave basis set. Comput. Mater. Sci. 1996, 6, 15–50.
36. Kresse, G.; Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using
a plane-wave basis set. Phys. Rev. B 1996, 54, 11169–11186.
37. Blöchl, P.E. Projector augmented-wave method. Phys. Rev. B 1994, 50, 17953–17979.

65

Thermoplastic Polymer Nanocomposites Based on Inorganic
Fullerene-like Nanoparticles and Inorganic Nanotubes
Mohammed Naffakh and Ana M. Díez-Pascual
Abstract: Using inorganic fullerene-like (IF) nanoparticles and inorganic nanotubes (INT) in
organic-inorganic hybrid composite, materials provide the potential for improving thermal, mechanical,
and tribological properties of conventional composites. The processing of such high-performance
hybrid thermoplastic polymer nanocomposites is achieved via melt-blending without the aid of any
modifier or compatibilizing agent. The incorporation of small quantities (0.1–4 wt.%) of IF/INTs
(tungsten disulfide, IF-WS2 or molybdenum disulfide, MoS2) generates notable performance
enhancements through reinforcement effects and excellent lubricating ability in comparison with
promising carbon nanotubes or other inorganic nanoscale fillers. It was shown that these IF/INT
nanocomposites can provide an effective balance between performance, cost effectiveness, and
processability, which is of significant importance for extending the practical applications of diverse
hierarchical thermoplastic-based composites.
Reprinted from Inorganics. Cite as: Naffakh, M.; Díez-Pascual, A.M. Thermoplastic Polymer
Nanocomposites Based on Inorganic Fullerene-like Nanoparticles and Inorganic Nanotubes.
Inorganics 2014, 2, 291–312.
1. Introduction
Over the past few years, research interest in the field of thermoplastic composites has changed
from “high-performance” advanced materials towards the development of “cost-performance”
engineering composites. Especially, carbon fiber (CF) or glass fiber (GF) reinforced,
thermoplastic-based composites have shown to offer design, processing, performance, and cost
advantages compared to metals for manufacturing structural parts. Among the advantages provided
by fiber-reinforced thermoplastics over metals and ceramics, that have been recognized for years,
are improved fracture toughness, impact resistance, strength to weight ratio, as well as high resistance
to corrosion and enhanced thermal and fatigue properties that have often been put in good use for
practical applications in the aeronautic, automotive, and energy sectors [1–3]. Nevertheless, these
applications require new properties and functionalities, especially superior mechanical performance,
flame and chemical resistance, magnetic field and UV resistance, high electrical conductivity,
environmental stability, low water absorption, and so forth. To address these issues, the integration
of inorganic nanoparticles into a polymer matrix allows both properties from inorganic nanoparticles
and polymer to be combined, thus, resulting in advanced polymer nanocomposites (PNCs) [4]. In
particular, additional nanoscale fillers, such as carbon nanotubes (CNTs) [5] or inorganic
nanoparticles [6], have been mixed with CFs to reinforce polymer matrices. Their high specific
surface area enables the formation of a large interphase in the composite and strong filler-matrix
interactions. In the same way, the addition of nanoclays to fiber-reinforced thermoplastic composites
has been reported to improve damping properties, fatigue life, toughness, and wear resistance [7,8].

66
The synergetic effect of CFs with the inorganic nanoparticles is believed to be the major cause for
the mechanical improvement achieved.
Recently, inorganic fullerenes (IFs) and nanotubes (INTs), based on layered metal dichalcogenides,
such as WS2 and MoS2, have emerged as one of the most promising developments in the area of
nanomaterials. These types of nanoparticles are currently the subject of intense research, summarized
in these reviews that include synthetic methodologies, diverse properties of these new nanomaterials
and their potential applications [9,10]. The first synthesis of such nanoparticles was reported by
Tenne et al., in 1992 and 1993 [11,12]. Since then, the synthetic technology has advanced
considerably and almost pure materials (>99%) are currently synthesized in large amounts by
ApNano Materials, Inc. (NanoMaterials, Ltd., Yavne, Israel) and employed in a wide variety of
fields, such as aerospace, automotive, naval, defense, medical, energy, electronics, and various other
industries. The physical properties of WS2 and MoS2 nanostructures (IF/INTs) have been studied in
detail, both experimentally and by theoretical modeling. These properties are interesting, not only
academically, but also because these kinds of nanostructures show substantial potential for becoming
part of the ultrahigh-strength nanocomposite technology [13].
The objective of this article is to emphasize the most recent findings about the influence of IF
nanoparticles and INTs on the structure, morphology and properties of thermoplastic polymer
nanocomposites, in comparison with PNCs incorporating other nanofillers. Particular interest
has been devoted to analyze the thermal, mechanical, and tribolological property enhancements
attained in multiscale fiber-reinforced thermoplastic composites containing inorganic fullerene-like
WS2 nanoparticles.
2. Preparation and Dispersion of IF/INT into Thermoplastic Polymers
The mixing of polymers and nanoparticles is opening new avenues of research and development
of advanced engineering flexible composites that exhibit advantageous magnetic, electrical, optical,
or mechanical properties. The main challenge in fabrication of these polymer nanocomposites for
structural and functional applications is uniform dispersion of nanoparticles in the polymer matrix.
However, good dispersion of nanoparticles in polymer composite materials is extremely difficult to
achieve since nanoparticles have a strong tendency to aggregate due to their nano-size and high
surface energy. In the case of organic–inorganic nanocomposites, the strength or level of interaction
between the organic and inorganic phases is another important factor in improving the overall
properties of the composites. Physical or simple mechanical mixing usually lead to a weak interaction
between the phases via hydrogen bonding or van der Waals forces. In order to minimize interface
energies between particles and polymer matrices, several surface modification/functionalization and
stabilization techniques have been developed that are mainly used in chemical methods, such as sol-gel,
in situ polymerization, etc. Owing to numerous papers published on polymer organic–inorganic
composite materials, it is impossible to completely review this field. The reader is referred to the
literature cited for a more detailed description of synthetic methods used for the processing of PNCs
reinforced with different types of inorganic nanofillers [13–15].

67
Inorganic layered materials, such as transition metal dichalcogenides MS2 (M = Mo, W), are one
of the most modern and the most promising development areas in the field of nanomaterials.
Inorganic fullerene-like (IF) nanoparticles can provide significant advantages over other spherical
nanoparticles for the preparation of advanced PNCs [13]. In particular, the incorporation of
environmentally-friendly IF-WS2 nanoparticles has been shown to improve thermal, mechanical, and
tribological properties of a series of thermoplastic polymers, including isotactic polypropylene (iPP) [16],
polyphenylene sulfide (PPS) [17], poly(ether ether ketone) (PEEK) [18], and nylon-6 [19]. The efficient
dispersion of IF-WS2 was achieved through simple melt-blending without using modifiers or surfactants.
Moreover, the combination of inorganic fullerenes with other organic micro-particles (nucleating agents),
micro-fibers (CFs) or nanofillers (CNTs) allows tailoring of more sophisticated hybrid materials with
complex architectures, interactions, morphology, and functionality [20–24]. In the same way, the use
of INT-WS2 (MoS2) offers the opportunity to produce novel advanced polymer nanocomposite
materials with excellent nanoparticle dispersion. More specifically, since the beginning of 2011, we
have successfully developed a new family of nanocomposites, which integrated MoS2 nanotubes into
an isotactic polypropylene (iPP) matrix, one of the most widely investigated polymers in the preparation
and application of nanocomposites, employing a simple and cost effective melt-processing route [25].
This strategy yields finer dispersion, with INT-MoS2 almost fully debundled into individual tubes or
small clusters, which are randomly oriented in the iPP matrix. Additionally, well-dispersed WS2
inorganic nanotubes were efficiently incorporated into epoxy matrix, poly(methyl methacrylate)
(PMMA), poly(propylene fumarate) (PPF), and poly(3-hydroxybutyrate) (PHB), using various
processing techniques [26–29]. Figure 1 shows, as an example, typical SEM images of the fracture
surfaces of composites containing inorganic fullerene-like nanoparticles or inorganic nanotubes
obtained under optimal processing conditions. It has been demonstrated by statistical analysis of the
surface density of IF-WS2 nanoparticles in the iPP nanocomposites, that the degree of dispersion
strongly depends on the duration of melt blending [16]. For 1.0 wt.% IF-WS2 (Figure 1a), it can be
seen that these nanoparticles are almost spherical, with an average diameter of around 80 nm, similar
to that observed for the raw nanofiller, and are individually dispersed for mixing times between 5
and 20 min. However, for IF-WS2 contents • 4.0 wt.%, 5 min is not enough time to attain single
particle distribution, and for the highest concentration incorporated of 8.0 wt.% (not shown here),
the influence of the mixing time on the degree of dispersion is even stronger. With increasing loading,
the interparticle distance decreases, hence, flocculation of these nanoparticles can occur after
the mixing is stopped. Thus, the crystallization rate, as well as the modulus of iPP, initially rise
with increasing filler content and finally level-off at filler loadings of around 1.0 wt.% [16].
In the case of multiscale fiber-reinforced thermoplastic composites, the laminates were prepared by
the film-stacking process. Four layers of GF or CF were alternatively stacked within five
iPP/IF-WS2 (PPS/IF-WS2) films in a closed mold. Consolidation of the material was made at 210 °C
in a hot-press (320 °C in the case of PPS matrix) [22,23]. The results obtained are very promising
and suggest that the use of IF/INT can provide an effective balance between cost effectiveness and
processability, making the resulting polymer nanocomposites highly suitable for a wide range of
applications at a large scale.

68
Figure 1. SEM micrographs of novel polymer/IF(INT) nanocomposites. (a) iPP/IF-WS2
(1.0 wt.%); (b) PPS/IF-WS2 (1.0 wt.%); (c) iPP/INT-MoS2 (1.0 wt.%); (d) iPP/IF-WS2
(2.0 wt.%)/GF and (e) PPS/IF-WS2 (2.0 wt.%)/CF.

3. Thermal Properties
It is well known that the crystalline morphology and structure obtained during the thermoplastic
processing plays an important role on the physico-mechanical behavior of the resulting polymeric
material, conditioning its potential uses. In this way, the control of the crystallization process can be
seen as a successful approach for improving physico-mechanical properties of polymers. Therefore,
it is of great interest to investigate the nucleation, crystallization, and structural development of the
matrix in IF/INT reinforced polymer nanocomposites [13]. This would help to optimize the
manufacturing conditions in order to obtain high-performance nanocomposites and to fully exploit
their potential in practical applications.

69
Figure 2. TGA thermograms under a nitrogen atmosphere for neat iPP, PPS and some
hierarchical laminates. The inset shows the initial degradation temperature (Ti) vs.
nanoparticle loading.
100
90

PPS/CF
PPS/IF-WS 2(1.0 wt.%)/CF

70

PPS/IF-WS 2(2.0 wt.%)/CF

60

iPP
PPS

500

50
40

iPP/GF
iPP/IF-WS 2(1.0 wt.%)/GF

540

Ti (ºC)

Weight
weight (%)
(%)

80

30
20
300

iPP/IF-WS 2(2.0 wt.%)/GF

460
420
380
340
0

1

2

3

4

IF-WS2 (wt.%)

400

500

600

700

T (ㇺ )(ºC)
Temperature
The thermal stability of several polymer matrices reinforced with IF-WS2 nanoparticles was
compared with that observed for other spherical inorganic nanofillers, organized by the nature of the
matrix [13]. It was found that the incorporation of nanometer-sized k particles into a polymer
enhances the thermal stability of the matrix inhibiting the formation and escape of volatile byproducts
generated during the decomposition process. In the case of the hierarchical thermoplastic-based
composites, the thermal stability of IF-WS2 reinforced iPP [22] and PPS [23] laminates has been
investigated using TGA, and typical thermograms under a nitrogen atmosphere for the neat matrices,
and composites reinforced with 1.0 and 2.0 wt.% IF-WS2 are shown in Figure 2. It is found that all
the composites exhibit a single decomposition stage in a nitrogen environment, similar to that found
for the neat polymers, indicating that the random scission of the polymeric chains is the predominant
degradation process. The incorporation of increasing nanoparticle contents induces a progressive
thermal stabilization of both matrices (see inset of Figure 2), the effect being more significant in the
case of iPP, probably related to the lower thermal stability of this commodity plastic compared to
high-performance PPS. Thus, an increase in the initial degradation temperature (Ti) of 12 °C and 47 °C
is attained at 2.0 wt.% loading in comparison to the reference PPS and iPP laminate, respectively. A
similar trend is found for the temperature of 10% weight loss (T10) and maximum rate of weight loss
(Tmax). This thermal stability enhancement has been ascribed to the barrier effect of the nanoparticles
that effectively obstruct the diffusion of volatile products from the bulk of the polymer to the gas
phase, therefore slowing down the decomposition process. Upon increasing IF-WS2 loading, the
barrier effect becomes stronger, which is reflected in higher degradation temperatures. An analogous
effect of thermal stability increase has been reported for PP/GF composites reinforced with other
inorganic nanoparticles such as clays [30]. Nevertheless, for the same nanofiller loading, the
improvements in thermal stability are larger in the case of IF-WS2, indicative of a more effective

70
heat barrier effect of the IF nanoparticles likely arising from their more homogenous dispersion and
spherical shape, thus, larger specific surface area.
In the same way, the incorporation of INTs can also lead to an improvement in the thermal
stability of polymer/INTs [25,27]. As an example, the characteristic weight loss temperatures for
PP nanocomposites, filled with different nanoreinforcements in nitrogen, are summarized in
Table 1 [31–42]. The data reveal that the concentration of non-modified INT-MoS2 has a dramatic
effect on the thermal stability of the iPP nanocomposites. T10 of iPP/INT-MoS2 (1.0 wt.%) was
almost 60 °C higher than that of neat iPP, suggesting that INT-MoS2 have outstanding properties for
improving the thermal stability at low nanofiller content [31]. As a comparison, approximately the
same increment was observed for iPP nanocomposites filled with 10 wt.% of silane-modified
halloysite nanotubes (HNTs). In the case of iPP/HNTs, the thermal stability and flame-retardant
effects are believed to result from the hollow tubular structure of HNTs, the barriers for heat and
mass transport and the presence of iron in the HNTs [32–34]. Layered silicates, such as
montmorillonite (MMT), also have important effects on the thermal stability of the PP matrix
(Table 1). The dramatic improvement in thermal stability of around 90 °C was related to the
confinement of the single nanoparticles in approximately 1 nm3 volume using sophisticated methods
of modification/exfoliation [39–41].
The flammability behavior of PPS/IF-WS2/CF has been investigated by pyrolysis combustion
flow calorimetry, in order to determine the heat release rate (HRR) at different nanoparticle
contents [24]. The addition of IF-WS2 leads to a progressive drop in the average peak HRR, the
reduction being about 17% for the laminate with 1.0 wt.% loading. Further, the onset temperature at
which begins the release of heat and the temperature at peak HRR increase gradually with the
nanoparticle loading, with maximum increments of 19 and 23 °C, respectively, at 2.0 wt.% IF-WS2.
These improvements are probably related to the low degree of porosity and enhanced thermal
stability of the hybrids. Moreover, there seems to be a synergistic effect of both micro- and
nano-fillers on increasing the polymer resistance to fire. The coexistence of CFs and IF-WS2 in the
laminates results in a more effective confined geometry that increases the barrier resistance to the
evolution of flammable volatiles. Similar synergistic behavior has been described for different
polymer/clay/carbon nanotube hybrids [43,44].
The degree of crystallinity is a key parameter in thermoplastic polymers because it has strong
influence on both the chemical and mechanical properties. The crystalline phase improves the
stiffness and tensile strength whilst the amorphous phase helps to absorb the impact energy. The
influence of IF-WS2 on the crystallization behavior of PPS/CF [23] and iPP/GF [22] has been
analyzed by DSC, and typical cooling thermograms for composites with 1.0 and 2.0 wt.% loading
are shown in Figure 3. Moreover, the crystallization temperature (Tp) as a function of IF-WS2
concentration is plotted in the inset of this Figure. Noticeable differences are detected depending on
the thermoplastic polymer. In the case of PPS based composites, the addition of low nanoparticle
contents (i.e., 0.1 or 0.5 wt.%) results in a decrease in Tp and the degree of crystallinity (Xc), indicating
the absence of a nucleating effect of the IF-WS2 on the polymer crystallization, and that the transport
of macromolecular segments to the growing surface of PPS in the composite is hindered. However,
the incorporation of higher nanoparticle contents leads to an increase in both Tp and Xc, by up to

71
9 °C and 14%, respectively demonstrating that higher nanoparticle contents act as nucleating agents
for PPS. On the other hand, these nanoparticles effectively nucleate the iPP matrix in the
concentration range of 0–4.0 wt.%, with increases up to 22 °C and 6% in Tp and Xc, respectively, at
the highest loading tested. These improvements are greater than those reported for binary
iPP/IF-WS2 nanocomposites [16], pointing towards a synergistic effect of both fillers on promoting
the crystallization of iPP. This behavior is in agreement with the reported for PP/ZnO/GF [45] and
PP/SiO2/GF hybrids [46], where the combination of nano- and micro-fillers additionally increased
the Tp of the matrix, albeit the increments found in those hybrids (~7 and 6 °C at 2.0 wt.% ZnO and
1.0 wt.% SiO2 content, respectively) are smaller than the increases found for the same amount of
IF-WS2. Further, Xc of PP dropped upon incorporation of ZnO or SiO2 and GF, while the combined
nucleating effect of IF-WS2/GF provoked a slight increase in crystallinity.
Table 1. Thermal stability, crystallization, and mechanical data for isotactic polypropylene
(iPP) nanocomposites using nanoreinforcing fillers with different morphologies (e.g.,
tubular, spherical and laminar-like particles) taken from literature. ¨T10 = increment of
degradation temperature for 10% weight loss, ¨Tp = increment of crystallization peak
temperature, GE = Percentage variations of Young’s modulus, G ıy = Percentage variations
of tensile strength and Gİb = Percentage variations of strain at yield.
Filler

INT-MoS2 [31]

HNTs [32–34]

CNTs [35–37]

rod-Si3 N4 [38]
Nanoclay (MMT)
[39–41]
IF-WS2 [16,42]

Filler
content
(wt.%)
0.1
0.5
1
1
2
5
10
20
30
0.1
0.25
0.5
1
2
1
2
3

¨T10
(ºC)

¨Tp
(ºC)

G(
(GPa)

Gıy
(MPa

Gİb
(%)

54
59
59
60
46
50
90

3.9
10
10.1
3.9
8.9
10
12.8
13.8
7.6
8.4
10.7
10
2
3
5

15%
28%
40%
32%
23%
722%
152%

13%
34%
41%
22%
15%
292%
95%

í9%
í18%
í52%
í15%
í30%
0%

0.1

11

9.8

-

-

-

1
2
4
8

14
15
27
44

13
19
20.5
22.1

39%
-

41%
-

í59%
-

72
Figure 3. DSC crystallization thermograms for neat iPP, PPS and some IF-WS2
reinforced multiscale laminates. The inset shows the crystallization peak temperature Tp
vs. IF-WS2 content.

PPS/IF(1.0 wt.%)/CF
PPS/IF(2.0 wt.%)/CF
PP/CF

PP/IF(1.0 wt.%)/GF
240

Tp (ºC)

Heat flow (a.u.)

Endo >

PPS/CF

iPP
PPS

200
160
120
80

PP/IF(2.0 wt.%)/GF

0

1

2

3

4

IF-WS2 (wt.%)

40

80

120

160

200

240

280

320

T(ㇺ ) (ºC)
Temperature
In this way, the control of the crystallization behavior has been shown to be a successful approach
for improving physico-mechanical properties of polymer/INT nanocomposites. Table 1 summarizes
the findings of several studies on the nucleating efficiency (NE) of nanoreinforcing fillers, and data
can be compared by analyzing the difference between the crystallization peak temperature (Tp) of
each nanocomposite and that of the neat matrix (ÄTp). Clearly, the ÄTp value for INT-MoS2 far
exceeds the values observed for montmorillonite nanoclay [39] and rod-Si3N4 [38], and is comparable to
that observed for MWCNTs [35]. However, the nucleation efficiency of INT-MoS2 is significantly
lower in comparison to the value of 40% observed for inorganic fullerene-like WS2 nanoparticles at
1.0 wt.% [16]. The results obtained clearly show that the addition of INT-WS2 plays a remarkable
role in accelerating the crystallization rate of iPP. In these systems, the crystallinity of iPP was found
to rise up to 14% with increasing the INT-MoS2 content, from a value of 50% for iPP, to values of
54, 57 and 56% for the nanocomposites with 0.1 wt.%, 0.5 wt.% and 1 wt.%, respectively [25].
Furthermore, a new study on the crystallization behavior of biopolymer/INTs suggests that INT-WS2
exhibits much more prominent nucleation activity on the crystallization of PHB than other specific
nucleating agents or nano-sized fillers [29]. An increment of 35 °C in the crystallization temperature
of PHB was observed for as little as 0.1 wt.% INT-WS2. This corresponds to the highest value
observed hitherto for PHB formulations using specific nucleating agents (e.g., talc, boron nitride
lignin) or nano-sized fillers (e.g., CNTs, graphene oxide) [29].

73
Figure 4. Room temperature thermal conductivity of iPP and PPS-based laminates as a
function of IF-WS2 concentration.
PPS/IF-WS2/CF

iPP/IF-WS2/GF

-1

-1

Thermal Conductivity, (W m K )

0.26
0.24
0.22
0.20
0.18
0.16
0.0

0.1

0.5

1.0

2.0

4.0

IF-WS2 (wt.%)

The addition of thermally conductive organic or inorganic nanofillers typically enhances the
thermal conductivity (Ȝ) of polymers, which is interesting for applications that require effective
dissipation of accumulated heat like connectors or thermal interface materials. It depends on several
factors, namely the filler size, aspect ratio, concentration and state of dispersion, the nature,
molecular weight and degree of crystallinity of the polymer, as well as the porosity of the material.
The room temperature thermal conductivity of iPP- [22] and PPS- [24] based laminates has been
measured in the transverse directions, and the results are shown in Figure 4. The incorporation of
IF-WS2, which exhibit about twice the thermal conductivity of the neat matrices [47], results in significant
Ȝ improvements in the case of iPP/GF laminates, up to 21% at 2.0 wt.% loading, whilst for PPS/CF
composites the increments are smaller, about 9% for the same loading. This discrepancy is ascribed to
the low thermal conductivity of the GF fabric (~0.05 W mí1 Kí1) compared to that of
CF (>200 W mí1 Kí1). It seems that the CFs play a dominating role in the thermal conductivity
properties and mask the effect of the IF-WS2, as can be deduced from the comparison with the results
of binary PPS/IF-WS2 nanocomposites [48], where Ȝ rose by up to ~45% upon addition of 2.0 wt.%
IF-WS2. However, for iPP-based samples, the improvements in the hierarchical laminates are
comparable to those reported for the corresponding binary composites [49], indicating that effect of
the nanoparticles predominates. An analogous behavior has been reported for other hierarchical
laminates based on thermoplastic polymers, such as PEEK/CNT/GF laminates [50], where Ȝ increased
by ~48% at 1.0 wt.% CNT, similarly to the enhancements found in the binary composites [51]. It is
worthy to note that for the same nanofiller concentration, the increases in Ȝ upon addition of CNTs
are only about double those achieved with the incorporation of the IF-WS2, while much higher
differences would be expected considering the extraordinary high thermal conductivity of CNTs. The
strong agglomerating tendency of CNTs, the small thermal conductance of the nanotube-polymer

74
interface and the high interfacial thermal resistance between nanotubes within a bundle probably
limits the property enhancement, whereas for composites incorporating IF-WS2 the large
nanofiller-matrix interfacial contact area and the very homogeneous dispersion lead to experimental
Ȝ values even higher than the theoretical predictions.
4. Mechanical Properties
The dynamic mechanical properties of the multiscale composites were explored by DMA,
technique that provides information about the viscoelastic behavior of the matrix, indicating changes
in the stiffness and the relaxation processes that occur as a function of temperature. The influence of
the IF-WS2 on the dynamic mechanical behavior of polymer/IF-WS2 nanocomposites has also been
investigated [16–18]. In particular, it was observed that the improvements in the storage modulus
values of PPS/IF-WS2 nanocomposites are noticeably higher than those achieved in other
thermoplastic nanocomposites based on IFs (e.g., iPP, nylon-6, PEEK), suggesting the presence of
specific polymer-filler interactions in the case of PPS. The molecular nature of these interactions are
still not understood, but they may be associated with the presence of outer S atoms on the IF
nanoparticles, and more work is required in order to explain this phenomenon. Figure 5 presents the
storage modulus (E') and loss tangent (tan į) at the frequency of 1 Hz for PPS- and iPP-based
composites incorporating 1.0 and 2.0 wt.% IF-WS2, and the glass transition temperature (Tg) vs.
nanoparticle content is shown in the inset of the Figure. Different behavior is also observed
depending on the polymer matrix. Regarding PPS/CF laminates, the addition of very low IF-WS2
loadings (i.e., 0.1 wt.%) leads to a slight drop in E' (~7% at 25 °C), probably related to the decrease
in the crystallinity found for this sample, as revealed by DSC analysis, since the crystalline regions
enhance the modulus of semicrystalline polymers. The laminate incorporating 0.5 wt.% IF-WS2
exhibits similar E' to that of PPS/CF, since the reinforcement effect of the IF-WS2 should compensate
for the slight decrease in crystallinity. In contrast, the incorporation of nanoparticle contents > 0.5 wt.%
leads to significant E' increments, by up to 22% for 2.0 wt.% nanoparticle content at 25 °C. On the
other hand, the gradual addition of IF-WS2 to iPP/GF results in progressive E' increases, by about
27% at 2.0 wt.% loading. This behavior is associated with the increase in crystallinity caused by
heterogeneous nucleation, combined with an effective reinforcement effect arising from a very
homogeneous nanoparticle dispersion. For both types of composites, the reinforcement effect is more
pronounced at temperatures below Tg, in agreement with the behavior reported for PP/nanoclay/GF
composites [30], where significant E' enhancements were found at low temperatures whereas
the differences in modulus among the samples became insignificant at temperatures above the
glass transition.

75
Figure 5. Evolution of the (a) storage modulus E' and (b) tan į as a function of
temperature for the neat polymers and some multiscale laminates. The inset shows the
glass transition temperature Tg vs. IF-WS2 content.

The evolution of tan į (ratio of the loss to storage modulus, a measure of the damping within the
system) as a function of temperature (Figure 5b) exhibits an intense peak, named Į relaxation that
corresponds to the Tg. Further, the iPP/GF laminates show a peak at about 88 °C related to the
relaxation of the crystalline phase (Įc). In an unfilled system, the polymer chain segments are free
from restraints. The incorporation of fillers decreases the free volume and restricts the mobility of
the matrix chains, which is reflected in higher Tg values (see inset of Figure 5). Once again, different
trend is found depending on the nature of the matrix. Thus, in the case of PPS/CF, the incorporation
of low IF-WS2 contents (” 0.5 wt.%) led to a downshift in Tg, while the addition of higher
concentrations resulted in an upshift. As mentioned above, the addition of low nanoparticle loadings
slows down the crystallization rate of PPS, leading to the formation of a more amorphous phase that
provokes a slight drop in Tg. However, the incorporation of higher contents has a nucleation effect,

76
thereby raising the crystallinity of the polymer, which combined with a larger IF-WS2-matrix
interfacial contact area results in an effective immobilization of the polymer chains, consequently an
increase in Tg of up to 18 °C at 2.0 wt.% IF-WS2. In contrast, the Tg progressively increases upon addition
of these nanoparticles to iPP/GF, the increment being about 6 °C for the same nanoparticle loading. In
the same way, the presence of IF-WS2 causes an increase in the crystalline relaxation temperature Įc of
iPP, since the strong nucleation effect of these nanoparticles accelerates the crystallization of iPP in
the nanocomposites.
The magnitude of the tan į peak is indicative of the filler-matrix interactions. For both types of
composites, the height of the tan į peak decreases with increasing IF-WS2 content, indicative of a
strong nanofiller-matrix interfacial adhesion. Moreover, this reduction probably arises from a
synergistic effect between the micro- and nano-fillers on restricting chain mobility, in agreement
with the behavior reported for other GF-reinforced hierarchical composites [52]. The incorporation
of both reinforcements has a strengthening effect, leading to a lower degree of molecular motion,
hence, lower damping characteristics. It also noteworthy that the width of the tan į peak becomes
broader with increasing nanoparticle loading, phenomenon that can be interpreted as improved
nanofiller-matrix interactions, and is another indication of the larger nanoparticle-matrix interfacial
area. The IF-WS2 and microscale fibers disturb the relaxation of the neighbour polymer chains, which
would behave differently from those situated in the bulk matrix, resulting in a wider maximum. This
behavior was also observed in IF-WS2 reinforced iPP [16] and PEEK [18] nanocomposites, attributed
to a more inhomogeneous amorphous phase in the composites in relation to the pure matrix.
The static mechanical properties of iPP and PPS based hybrid laminates have been investigated
by tensile tests [22,23], and the Young’s modulus (E), tensile strength (ıy), elongation at break (İb),
and toughness (T) as a function of nanofiller loading are plotted in Figure 6. The trends observed are
similar to those described previously for the storage modulus. E and ıy rise progressive with
increasing nanoparticle loading in the case of iPP/GF composites, while they decrease slightly at low
loadings and then grow in PPS/CF laminates, behavior that is directly related to the crystallinity of the
samples, as discussed previously. Interestingly, both parameters only rise marginally upon addition of
the IF-WS2, the maximum increments being ~14% and 11% at 2.0 wt.% nanoparticle content,
respectively, in the case of PPS/CF, and even smaller for iPP/GF composites (Figure 6). However,
considerably larger increases were observed for the binary iPP/IF-WS2 nanocomposites [49], where E
and ıy improved by around 42 and 31%, respectively, for the indicated loading. For multiscale
composites, it is expected that the nanofillers predominantly influence the properties that are
matrix-dominated; consequently, only small increases are observed in the Young’s modulus and
tensile strength of the hybrids, since the tensile properties are more fiber-dominated. These results
are consistent with the behavior reported for other thermoplastic-based hybrids [53], where E and ıy
of the fiber reinforced polymer only improved marginally upon incorporation of the nanoscale fillers
due to the dominating role of the fibers.
With regard to the strain at break (İb), the trend found is very similar for both composite series. A
moderate increase is found at low nanoparticle loadings, followed by a sharp reduction at higher
concentrations. This indicates that higher amounts of IF-WS2 hinder the ductile flow of the matrix.
This tendency is in contrast to that typically reported for CNT-reinforced multiscale laminates [53],

77
where İb systematically decreases upon addition of the carbon nanofillers, attributed to the presence
of aggregates that produces stress concentrations at the filler-matrix interface, leading to premature
failure. Similarly, Rahman et al. [30] found around 50% reduction in tensile strain upon incorporation
of 6.0 wt.% nanoclay to PP/GF (30 wt.%), also ascribed to the poor nanoclay dispersion that strongly
limits the plastic deformation of the matrix. The surprising behavior observed for the composites
filled with IF-WS2 is probably related to the lubricant character and more uniform dispersion of these
inorganic nanoparticles combined with their spherical shape that reduce the stress concentration sites,
thereby improving the matrix ductility. However, for IF-WS2 concentrations higher than 1.0 wt.%, a
stiff hybrid network of micro- and nano-fillers could be formed that acts very effectively as a barrier
for the mobility of the polymer chains, thus limiting the ductile deformation. A qualitatively similar
behavior is found for the toughness, measured as the area under the tensile curve, that increases
considerably at low IF-WS2 loadings (i.e., by 35% at 0.1 wt.% content compared to iPP/GF) while
drops moderately at concentrations higher than 1.0 wt.% (around 20% decrease at 2.0 wt.% loading
compared to PPS/CF). The small aggregates contribute to increase the brittleness under high strain
rates, since they nucleate secondary cracks and favour the formation of dimples.
Figure 6. (a) Young’s modulus (E), (b) tensile strength (ıy), (c) elongation at break (İb)
and (d) toughness (T) as a function of IF-WS2 loading. Solid and open symbols
correspond to PPS/IF-WS2/CF and iPP/IF-WS2/GF systems, respectively.
210

(a)

Tensile Strength, Vy (MPa)

Young's Modulus, E (GPa)

16
15
14
13
3.5
3.0

200
190
180

2.5
0.5
1.0
1.5
IF-WS2 content (wt.%)

70

2.0

0.0

0.5

1.0

1.5

2.0

IF-WS2 content (wt.%)

8.0

20

7.5
7.0

18
3

Toughness, 7 (MJ/m )

Elongation at Break, Hb (%)

75

65
0.0

6.5
6.0
5.5
3.0
2.5

(b)

(c)

2.0

16
14
12
4
3

(d)

2
0.0

0.5

1.0

1.5

IF-WS2 content (wt.%)

2.0

0.0

0.5

1.0

1.5

IF-WS2 content (wt. %)

2.0

78
The influence of the IF-WS2 on the flexural properties of iPP/GF and PPS/CF has also been
investigated [22,24]. In this case, maximum increments in the flexural modulus Ef and flexural
strength ıfM of iPP/GF up to 26 and 22%, respectively, have been attained at 2.0 wt.% loading.
Similarly, enhancements of 25 and 15% have been found in PPS/CF composites for the same nanofiller
loading. The comparison of the results with those obtained for the corresponding binary
nanocomposites [22,24] reveals a synergistic effect of both fillers on enhancing the flexural properties
of the matrix.
Table 2. Comparison of the increment in static mechanical properties (in %) for different
polypropylene (PP) and polyphenylene sulfide-(PPS) based hierarchical laminates.
MWCNT: multi-walled carbon nanotubes; MMT: montmorillonite; Woll: Wollastonite;
E: Young’s modulus; ıy: tensile strength at yield; G: impact strength; Ef: flexural
modulus; ıfM: flexural strength.
Matrix
PP
PP
PP

Fiber
(wt.%)
GF
(5)
CF
(5)
GF
(30)

CNT
¨E (%)
(wt.%)
MWCNTs
40

¨ıy
(MPa)
39

¨T
(%)
24

¨Ef
(%)
36

¨ıfM
(%)
43

Ref.
[18]

MWCNTs

57

37

34

51

35

[18]

MMT
(6)

6

6

-

9

10

[3]

PP

GF
(30)

Woll
(10)

í6

í6

í31

í2

í3

[19]

PP

GF
(40)
GF
(30)
GF
(40)

SiO2
(1)
IF-WS2
(2)
CaCO3
(3 wt.%)

22

3

í5

2

12

[9]

10

8

-

26

22

[1]

27

9

14

-

-

[20]

GF
(40)
CF

CaCO3
(3)
IF-WS2
(2)

-

-

20

0

3

[21]

14

11

í20

25

15

[2,4]

PP
PPS
PPS
PPS

Table 2 compares the improvements in static mechanical properties reported for various PP and
PPS-based hierarchical composites [30,46,54–57]. Clearly, the highest improvements are attained
upon addition of multi-walled carbon nanotubes (MWCNTs) to fiber-reinforced PP composites [54],
which is reasonable taking into account the very high modulus of these carbon nanofillers.
Nevertheless, among the various inorganic fillers, IF-WS2 lead to larger stiffness and strength
improvements than montmorillonite [30], wollastonite [55], or nanosilica [46], and comparable to
those of CaCO3 [56,57].
In the same way, the incorporation of INTs can also lead to improvement in the mechanical
properties of polymer/INTs [25,27]. As an example, the characteristic mechanical data (e.g., Young’s
modulus, E, tensile strength, ıy and strain at yield, İy) for the PP nanocomposites incorporating

79
nanoreinforcing fillers with different morphologies are summarized in Table 1 [31–42]. It can be
observed that the addition of INT-MoS2 progressively enhances the Young’s modulus of the matrix,
with increments of 15, 28, and 40% for loading fractions of 0.1, 0.5, and 1.0 wt.%, respectively. The
improved E obtained in this work is ascribed to the very uniform dispersion of the INT-MoS2 and
their high aspect ratio, which results in larger nanofiller-polymer interfacial area. Qualitatively
similar trends were found for the tensile strength, where the increments were around 13, 34, and 41%
for the abovementioned nanofiller contents. On the other hand, the incorporation of the inorganic
nanotubes leads to a slight decrease in İy. This is a typical behavior of nanofiller-reinforced polymer
composites, since the nanofillers restrict the ductile flow of the matrix, and is in agreement with the
results reported by Lopez-Gaxiola et al. [58] for carbon filler-reinforced PP composites. Table 1 also
shows the percentage variations in the mechanical properties of iPP nanocomposites containing
similar amounts (a1.0 wt.%) of various nanofillers. Remarkable improvements in the mechanical
properties are observed for iPP/INT-MoS2, where the non-modified nanofillers were dispersed
uniformly in the iPP matrix for all the compositions prepared [31]. The magnitude of increase in the
modulus and strength is similar to that obtained for IF-WS2 nanoparticles [42] and far exceeds that
reported for both modified HNTs [34] and CNTs [37]. However, silicon nitrides clearly provide the
best reinforcement for PP matrix, which has been related to the alignment and exfoliation of rod-shaped
Si3N4 particles [38]. These phenomena were also mainly responsible for the 95% enhancement in the
tensile strength and 152% increase in the tensile modulus of PP using p-aminobenzoic acid
modified-clay with PP-g-MA as a compatibilizer [40]. On the other hand, Reddy et al. have reported
that the high rigidity of INT-WS2 and the effective load transfer from the matrix to the INT-WS2
were responsible for the improved mechanical properties of PMMA/INT-WS2 nanocomposites [27].
In particular, it was observed that the elastic modulus of PMMA fiber meshes was increased by
10 and 22 times upon incorporation along the fiber axis of 1.0 and 2.0 wt.% INT-WS2, respectively.
Analogously, the tensile strength of the composite fibers increased by 35 and 32% for the indicated
nanoparticle loadings. However, the toughness of the sample with 2.0 wt.% INT-WS2 was lower than
that of the neat PMMA fiber, since nanofiller aggregation started to take place. Overall, experimental
results point out the advantages of using these environmentally friendly and cheap inorganic fullerenes
and nanotubes instead of conventional nanoparticles for improving the mechanical performance of
thermoplastic composites.
5. Tribological Properties
Inorganic nanoparticles are frequently incorporated into thermoplastic polymers with the aim to
improve the tribological properties. The nanoparticles exhibit some advantages compared to conventional
microfillers, such as higher specific surface area, lower abrasiveness due to a reduced angularity,
enhanced strength, modulus and toughness. In addition, IF-WS2 possess a lubricant character, and
have been shown to be effective for improving the tribological properties of thermoplastic polymers
such as PPS or PEEK [59,60]. Figure 7 displays the change in the coefficient of friction (—) and wear
rate of PPS/CF upon addition of IF-WS2 [23]. The incorporation of 0.1 wt.% IF-WS2 leads to a slight
increase in — (~5%) compared to the reference laminate, probably related to the decrease in stiffness
and strength found for this sample that prevails over the lubricant effect of the IF-WS2. Further

80
increasing the nanoparticle loading, — drops strongly, reaching the lowest value at 2.0 wt.% IF-WS2
(about 32% drop compared to the reference laminate). Rapoport et al. [61] proposed a rolling
mechanism for these nanoparticles, in which they act as a ball-bearing component, implying that they
roll instead of sliding between the surfaces, hence, decreasing the shear stress, contact temperature
and coefficient of friction. Likewise, the abovementioned behavior can be attributed to a synergistic
effect between the CFs and the inorganic nanoparticles, as reported previously for CF-reinforced
PEEK incorporating ZnS or TiO2 nanoparticles [62].

P

10

3

K

0.25

8

0.20

6

Coefficient of friction, P

0.30

6
0.15
4
0.10
2

0.05
0.00

Specific wear rate, K x 10 (mm / Nm)

Figure 7. Coefficient of friction and wear rate of PPS/IF-WS2/CF laminates as a function
of IF-WS2 content.

0.0

0.1

0.5

1.0

2.0

0

IF-WS2 (wt.%)
With regard to the wear rate, a progressive reduction in this parameter is found upon increasing
IF-WS2 concentration, which decreases by nine-fold for the composite with 2.0 wt.% loading
compared to the reference laminate. This increase in wear resistance has been attributed to the
formation of a thin, continuous, and smooth transfer film on the counterface during sliding combined
with the reinforcing effect, and it is enhanced by the presence of the two fillers. The adhesion of the
transfer film would be stronger since a homogeneous mixture of the debris is formed, and the
resistance to cracking and fatigue failure would also increase in the presence of the nanoparticles. An
analogous trend was reported for the wear behavior of PEEK/ZrO2/CF composites [62], where a
synergistic effect of CFs with ZrO2 nanoparticles on enhancing the matrix wear resistance was
proposed. Overall, the combination of conventional CF-reinforced thermoplastics with lubricant
nanoparticles like IF-WS2 is a promising approach to develop multiscale hybrids with superior
tribological performance.

81
Table 3. Wear rate (K) data of PP nanocomposites nanocomposites using nanoreinforcing
fillers with different morphologies.
Filler

Filler
content
(wt.%)

Wear rate
(K)×104
(mm3/ Nm)

Percentage variation
of K (%)

INT-MoS2 [31]

0

6.27

-

Nanoclay [65]
IF-WS2 [42]

0.1
0.5
1
1
1

5.97
4.35
2.98
-

5
31
53
38.5
63

Table 3 collects the wear rate of melt-procesable iPP/INT-MoS2 nanocomposites [31]. With the
incorporation of INT-MoS2 the wear resistance of the polymer is considerably enhanced and the
nanocomposite with 1.0 wt.% loading shows a reduction of about 53%. These inorganic nanotubes
dispersed in the polymer matrix can act as a barrier and prevent large-scale fragmentation of the iPP. It
has been reported that nanofillers of similar dimensions as the segments of the surrounding polymer
chains enable a milder material removal and aid the formation of uniform tenacious transfer film [63,64].
Table 3 also compares the percentage of variations in the wear rate of PP nanocomposites containing
1.0 wt.% of nanoclay [65], IF-WS2 [42], and INT-MoS2 [31]. In particular, PP/INT-MoS2 showed
higher wear property improvement than that of PP/nanoclay without the need for an exfoliation
process. The highest percentage of improvement in wear rate is found for IF-WS2 solid lubricant
nanoparticles, which have recently been identified as ideal candidates for improving the tribological
performance of polymers like epoxy [61], nylon-6 [19], and PEEK [18].
6. Conclusions and Future Developments
The addition of IF/INTs has been demonstrated to be a very efficient strategy to improve the
thermal, mechanical and tribological properties of thermoplastic polymers like iPP, PPS, or PEEK
and their fiber-reinforced composites. These materials can be fabricated by simple melt-processing
and compression molding without the need for modifiers or surfactants, leading to a very homogenous
dispersion of the nanofillers within the matrix. More importantly, they exhibit similar or enhanced
performance when compared with composites that incorporate CNTs, nanoclays or other inorganic
spherical nanoparticles, but are substantially more cost-effective, efficient and environmentally-friendly.
Results demonstrate the existence of synergistic effects of both micro-and nanoscale fillers on
enhancing the stiffness, strength, thermal conductivity, thermal stability, flammability, and wear
resistance of hierarchical thermoplastic-based composites. This new family of materials has a wide
range of potential applications ranging from medicine to the aerospace, automotive, and electronics
industries. Some of these applications are still at an early stage of research and development.
However, for optimal control of the properties of these new materials, it is highly important to tailor
the fabrication process from the viewpoint of the final product. In particular, the improvement and
application of these nanocomposites in comparison with other organic-inorganic hybrid
nanomaterials (silica, metal oxides, clays, etc.) depend on how effectively we optimize and scale-up

82
their fabrication method. For specific applications, these nanoparticles should be surface
functionalized in order to confer more selectivity, specificity and reactivity with the polymer chains.
An additional demanding area is the potential of these nanoparticles in the field of biocompatible
and/or biodegradable polymeric composites for packaging and medical applications and their
eventual toxicological effects, if any, need to be investigated. Research and progress in these areas
will not only benefit the current applications but would also lead to new markets as well as to future
development of diverse hierarchical thermoplastic-based composites.
Acknowledgments
This work was supported by the Spanish Ministry Economy and Competitivity (MINECO),
Project MAT-2010-21070-C02-01. Dr. M. Naffakh would like to acknowledge the Ministerio de
Economía y Competitividad (MINECO) for a “Ramón y Cajal” Senior Research Fellowship and Ana
Diez-Pascual wishes to acknowledge the CSIC for a JAE Postdoctoral Fellowship cofinanced by
the EU.
Author Contributions
This project was conceived and designed by MN. AD characterized and discussed the mechanical
and tribological properties. MN analyzed and discussed the morphology and thermal properties. Both
authors contributed in writing this paper.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.

2.
3.
4.
5.
6.
7.

Offringa, A.R. Thermoplastic composites in aerospace-proven through cost-effective
processing. In Proceedings of the 41th International Conference on Automated Composites
(ICAC), Nottingham, UK, 6–7 September 1995.
Henshaw, J.M.; Han, W.J.; Owens, A.D. An overview of recycling issues for composite
materials. J. Thermoplast. Compos. Mater. 1996, 1, 4–20.
Marsh, G. Next step for automotive materials. Mater. Today 2003, 6, 36–43.
Sanchez, C.; Belleville, P.; Polpall, M.; Nicole, L. Applications of advanced hybrid
organic-inorganic nanomaterials: From laboratory to market. Chem. Soc. Rev. 2011, 40, 696–753.
Qian, H.; Greenhalgh, E.S.; Shaffer, M.S.P.; Bismarck, A. Carbon nanotube-based hierarchical
composites: A review. J. Mater. Chem. 2010, 20, 4751–4762.
Lin, G.; Xie, G.; Sui, G.; Yang, R. Hybrid effect of nanoparticles with carbon fibers on the
mechanical and wear properties of polymer composites. Compos. Part B 2012, 43, 44–49.
Hussain, M.; Nakahira, A.; Nishijima, S.; Niihara, K. Evaluation of machanical behaviour of
CFRC transverse to the fiber direction at room and cryogenic temperature. Compos. Part A
2000, 31, 173–179.

83
8.
9.
10.
11.
12.
13.

14.
15.
16.

17.

18.

19.

20.

21.

22.

23.

Timmerman, J.F.; Hayes, B.S.; Seferis, J.C. Nanoclay reinforcement effects on the cryogenic
microcraking of carbon fiber/epoxy composites. Compos. Sci. Technol. 2002, 62, 1249–1258.
Tenne, R. Inorganic nanotubes and fullerene-like nanoparticles. Nat. Nanotechnol. 2006, 1,
103–111.
Tenne, R.; Redlich, M. Recent progress in the research of inorganic fullerene-like nanoparticles
and inorganic nanotubes. Chem. Soc. Rev. 2010, 39, 1423–1434.
Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and cylindrical structures of WS2.
Nature 1992, 360, 444–445.
Margulis, L.; Salitra, G.; Tenne, R.; Talianker, M. Nested fullerene-like structures. Nature 1993,
365,113–114.
Naffakh, M.; Díez-Pascual, A.M.; Marco, C.; Ellis, G.; Gómez-Fatou, M.A. Opportunities and
challenges in the use of inorganic fullerene-like nanoparticles to produce advanced polymer
nanocomposites. Prog. Polym. Sci. 2013, 38, 1163–1231.
Kickelbick, G. Concepts for the incorporation of inorganic building blocks into organic
polymers on a nanoscale. Prog. Polym. Sci. 2003, 28, 83–114.
Rozenberg, B.A.; Tenne, R. Polymer-assisted fabrication of nanoparticles and nanocomposites.
Prog. Polym. Sci. 2008, 33, 40–112.
Naffakh, M.; Martin, Z.; Fanegas, N.; Marco, C.; Gómez, M.A.; Jiménez, I. Influence of
inorganic fullerene-like WS2 nanoparticles on the thermal behavior of isotactic polypropylene.
J. Polym. Sci. B 2007, 45, 2309–2321.
Naffakh, M.; Marco, C.; Gómez, M.A.; Gómez-Herrero, J.; Jiménez, I. Use of inorganic
fullerene-like WS2 to produce new high-performance polyphenylene sulfide nanocomposites:
Role of the nanoparticle concentration. J. Phys. Chem. B 2009, 113, 10104–10111.
Naffakh, M.; Díez-Pascual, A.M.; Marco, C.; Gómez, M.A.; Jiménez, I. Novel melt-processable
poly(ether ether ketone) (PEEK)/inorganic fullerene-like WS2 nanoparticles for critical
applications. J. Phys. Chem. B 2010, 114, 11444–11453.
Naffakh, M.; Marco, C.; Gomez, M.A.; Jimenez, I. Novel melt-processable nylon-6/inorganic
fullerene-like WS2 nanocomposites for critical applications. Mater. Chem. Phys. 2011, 129,
641–648.
Naffakh, M.; Marco, C.; Ellis, G. Novel polypropylene/inorganic fullerene-like WS2
nanocomposites containing a ȕ-nucleating agent: Dynamic crystallization and melting
behaviour. J. Phys. Chem. B 2011, 115, 10836–10843.
Naffakh, M.; Díez-Pascual, A.M.; Gómez-Fatou, M.A. New hybrid nanocomposites containing
carbon nanotubes, inorganic fullerene-like WS2 nanoparticles and poly(ether ether ketone)
(PEEK). J. Mater. Chem. 2011, 21, 7425–7433.
Díez-Pascual, A.M.; Naffakh, M. Tuning the properties of carbon fiber-reinforced
poly(phenylene sulphide) laminates via incorporation of inorganic nanoparticles. Polymer 2012,
53, 2369–2378.
Díez-Pascual, A.M.; Naffakh, M. Polypropylene/glass fiber hierarchical composites
incorporating inorganic fullerene-like nanoparticles for advanced technological applications.
ACS Appl. Mater. Interfaces 2013, 5, 9691–9700.

84
24. Díez-Pascual, A.M.; Naffakh, M. Inorganic nanoparticle-modified carbon fiber
fabric/poly(phenylene sulphide) laminates with enhanced thermomechanical behaviour.
Materials 2013, 6, 3171–3193.
25. Naffakh, M.; Remškar, M.; Marco, C.; Gómez, M.A.; Jiménez, I. Towards a new generation of
polymer nanocomposites based on inorganic nanotubes. J. Mater. Chem. 2011, 21, 3574–3578.
26. Zohar, E.; Baruch, S.; Shneider, M.; Dodiu, H.; Kenig, S.; Wagner, D.H.; Zak, A.; Moshkovith, A.;
Rapoport, L.; Tenne, R. The Mechanical and Tribological Properties of Epoxy Nanocomposites
with WS2 Nanotubes. Sens. Transducers J. 2011, 12, 53–65.
27. Reddy, C.S.; Zak, A.; Zussman, E. WS2 nanotubes embedded in PMMA nanofibers as energy
absorptive material. J. Mater. Chem. 2011, 21, 16086–16093.
28. Lalwani, G.; Henslee, A.M.; Farshid, B.; Parmar, P.; Lin, L.; Qin, Y.X.; Kasper, F.K.;
Mikos, A.G.; Sitharaman, B. Tungsten disulfide nanotubes reinforced biodegradable polymers
for bone tissue engineering. Acta Biomater. 2013, 9, 8365–8373.
29. Naffakh, M.; Marco, C.; Ellis, G. Inorganic WS2 Nanotubes that improve the crystallization
behavior of biodegradable poly(3-hydroxybutyrate). Cryst. Eng. Comm. 2013, 16, 1126–1135.
30. Rahman, N.A.; Hassan, A.; Yahya, R.; Lafia-Araga, R.A.; Hornsby, P.R. Polypropylene/glass
fiber/nanoclay hybrid composites: morphological, thermal, dynamic mechanical and impact
behaviors. J. Reinf. Plast. Compos. 2012, 31, 269–281.
31. Naffakh, M.; Díez-Pascual, A.M.; Remškar, M.; Marco, C. New inorganic nanotube polymer
nanocomposites: Improved thermal, mechanical and tribological properties in isotactic
polypropylene incorporating INT-MoS2. J. Mater. Chem. 2012, 22, 17002–17010.
32. Liu, M.; Guo, B.; Du, M.; Chen, F.; Jia, D. Halloysite nanotubes as a novel ȕ- nucleating agent
for isotactic polypropylene. Polymer 2009, 50, 3022–3030.
33. Du, M.; Guo, B.; Jia, D. Thermal stability and flame retardant effects of halloysite nanotubes on
polypropylene. Eur. Polym. J. 2006, 42, 1362–1369.
34. Prashantha, K.; Lacrampe, M.F.; Krawczak, P. Processing and characterization of alloysite
nanotubes filled polypropylene nanocomposites based on a masterbatch route: Effect of
halloysites treatment on structural and mechanical properties. Express Polym. Lett. 2011, 5,
295–307.
35. Marco, C.; Naffakh, M.; Gómez, M.A.; Santoro, G.; Ellis, G. The crystallization of
polypropylene in multiwall carbon nanotube based composites. Polym. Compos. 2011, 32,
324–333.
36. Jin, S.H.; Kang, C.H.; Yoon, K.H.; Bang, D.S.; Park, Y.B. Effect of compatibilizer on
morphology, thermal and rheological properties of polypropylene/functionalized multi-walled
carbon nanotubes composite. Appl. Polym. Sci. 2009, 111, 1028–1033.
37. Lui, Y.; Gao, J. Mechanical properties and wear behavior of polypropylene/carbon nanotube
nanocomposites. Adv. Mater. Res. 2011, 798, 299–300.
38. Rangari, V.K.; Shaik, Y.M.; Mohammad, G.M.; Jeelani, S. Reinforcement of Si3N4
nanoparticles in Polypropylene single fibers through melt extrusion process and their properties.
J. Appl. Polym. Sci. 2011, 121, 1512–1520.

85
39. Xu, W.; Ge, M.; He, P. Nonisothermal crystallization kinetics of polypropylene/montmorillonite
nanocomposites. J. Polym. Sci. Polym. Phys. 2002, 40, 408–414.
40. Sharma, S.K.; Nayak, S.K. Surface modified clay/polypropylene (PP) nanocomposites: Effect
on physico-mechanical, thermal and morphological properties. Polym. Degrad. Stab. 2009, 94,
132–138.
41. Sharma, S.K.; Nema, A.K.; Nayak, S.K. Polypropylene nanocomposite film: A critical
evaluation on the effect of nanoclay on the mechanical, thermal, and morphological behavior.
J. Appl. Polym. Sci. 2010, 115, 3463–3473.
42. Naffakh, M.; Díez-Pascual, A.M.; Marco, C.; Ellis, G. Novel polypropylene/inorganic fullerene-like
WS2 nanocomposites containing a ȕ-nucleating agent: Mechanical, tribological and rheological
properties. Mater. Chem. Phys. 2014, 144, 98–106.
43. Gao, F.G.; Beyer, G.; Yuan, Q.C. Mechanistic study of fire retardancy of carbon
nanotube/ethylene vinyl acetate copolymers and their clay composites. Polym. Deg. Stab. 2005,
89, 559–564.
44. Haiyun, M.; Lifang, T.; Zhongbin, X.; Zhengping, F. Synergistic effect of carbon nanotube and clay
for improving the flame retardancy of ABS resin. Nanotechnology 2007, 18, 375602–375610.
45. Cui, Y.H.; Wang, X.X.; Li, Z.Q.; Tao, J. Fabrication and properties of Nano-ZnO/Glass fiber
reinforced polypropylene composites. J. Vinyl. Addit. Technol. 2010, 16, 189–194.
46. Jacob, S.; Suma, K.K.; Mendaz, J.M.; George, A.; George, K.E. Modification of
polypropylene/glass fiber composites with nanosilica. Macromol. Symp. 2009, 277, 138–143.
47. Díez-Pascual, A.M.; Naffakh, M.; Gómez-Fatou, M.A. Mechanical and electrical properties of
novel poly(ether ether ketone)/carbon nanotube/inorganic fullerene-like WS2 hybrid
nanocomposites: experimental measurements and theoretical predictions. Mater. Chem. Phys.
2011, 130, 126–133.
48. Naffakh, M.; Díez-Pascual, A.M.; Marco, C.; Ellis, G. Morphology and thermal properties of
novel poly(phenylene sulfide) hybrid nanocomposites based on single-walled carbon nanotubes
and inorganic fullerene-like WS2 nanoparticles. J. Mater. Chem. 2012, 22, 1418–1425.
49. Díez-Pascual, A.M.; Naffakh, M. Mechanical and thermal behaviour of isotatic polypropylene
reinforced with inorganic fullerene-like IF-WS2 nanoparticles: Effect of filler loading and
temperature. Mater. Chem. Phys. 2013, 141, 979–989.
50. Ashrafi, B.; Diez-Pasual, A.M.; Johnson, L.; Genest, M.; Hind, S.; Martínez-Rubi, M.;
González-Domínguez, J.M.; Martínez, M.T.; Simard, B.; Gómez, M.A.; et al. Influence of
interfacial characteristics on properties of PEEK/Glass fiber laminates modified with single-walled
carbon nanotubes. Compos. Part A 2012, 43, 1267–1279.
51. Díez-Pascual, A.M.; Naffakh, M.; Gómez, M.A.; Marco, C.; Ellis, G.; Martinez, M.T.; Anson, A.;
Gónzaelz-Dominguez, J.M.; Martinez-Rubi, Y.; Simard, B. The influence of a compatibilizer
on the thermal and dynamic mechanical properties of PEEK/carbon nanotube composites.
Nanotechnology 2009, 20, doi:10.1088/0957-4484/20/31/315707.

86
52. Díez-Pascual, A.M.; Ashrafi, B.; Naffakh, M.; González-Domínguez, J.M.; Johnston, A.;
Simard, B.; Martínez, M.T.; Gómez, M.A. Influence of carbon nanotubes on the thermal,
electrical and mechanical properties of PEEK/glass fiber laminates. Carbon 2011, 49,
2817–2833.
53. Díez-Pascual, A.M.; Naffakh, M.; Marco, C.; Gómez, M.A.; Ellis, G. Multiscale fiber-reinforced
thermoplastic composites incorporating carbon nanotubes: A review. Curr. Opin. Mater. Sci.,
in press.
54. Rahmanian, S.; Thean, K.S.; Suraya, A.R.; Shazed, M.A.; Salleh, M.A.M.; Yusoff, H.M. Carbon
and glass hierarchical fibers: Influence of carbon nanotubes on tensile, flexural and impact
properties of short fiber reinforced composites. Mater. Des. 2013, 43, 10–16.
55. Joshi, H.; Purnima, J. Development of glass fiber, wollastonite reinforced polypropylene hybrid
composite: Mechanical properties and morphology. Mater. Sci. Eng. A 2010, 527,1946–1951
56. Haribabu, J.; Sir, K.A.; Ravikiran, B. Fabrication and characterization of PPS/40%GF/nano-CaCO3
hybrid composites. Int. J. Mod. Eng. Res. 2013, 3, 2262–2266.
57. Narayana, S.; Suman, K.N.S.; Babu, S.R.; Ramesh, N. Effect of nano-CaCO3 content on flexural
and impact properties of PPS/GF ternary composites. Int. J. Eng. Res. Technol. 2012, 1, 1–7.
58. Lopez-Gaxiola, D.L.; Jubinski, M.M.; Keith, J.M.; King, J.A.; Miskioglu, I. Effects of carbon
fillers on tensile and flexural properties in polypropylene-based resins. J. Appl. Polym. Sci.
2010, 118, 1620–1633.
59. Díez-Pascual, A.M.; Naffakh, M.; Marco, C.; Ellis, G. Rheological and tribological properties
of carbon nanotube/thermoplastic nanocomposites incorporating inorganic fullerene-like WS2
nanoparticles. J. Phys. Chem. B 2012, 116, 7959–7969.
60. Díez-Pascual, A.M.; Naffakh, M.; Marco, C.; Ellis, G. Mechanical and electrical properties of
carbon nanotube/polyphenylene sulphide composites incorporating polyetherimide and
inorganic fullerene-like WS2 nanoparticles. Compos. Part A 2012, 43, 603–612.
61. Rapoport, L.; Nepomnyashcy, O.; Verdyan, A.; Popovitz-Biro, R.; Volovik, Y.; Ittah, B.; Tenne, R.
Polymer nanocomposites with fullerene-like solid lubricant. Adv. Eng. Mater. 2004, 6, 44–48.
62. Friedrich, K.; Schlarb, A.K. Tribology of Polymeric Nanocomposites; Briscoe, B.J., Ed.;
Elsevier: Amsterdam, The Netherlands, 2008.
63. Wang, Q.H.; Xu, J.; Shen, W.; Xue, Q. The effect of particle size of nanometer ZrO2 on the
tribological behaviour of PEEK. Wear 1997, 209, 316–321.
64. Shwartz, C.J.; Bahadur, S. Studies on the tribological behavior and transfer film-counterface
bond strength for polyphenylene sulfide filled with nanoscale alumina particles. Wear 2000,
237, 261–273.
65. Kanny, K.; Jawahar, P.; Moodley, V.K. Mechanical and tribological behavior of clay
polypropylene nanocomposites. J. Mater. Sci. 2008, 43, 7230–7238.

87

Continuous Production of IF-WS2 Nanoparticles by
a Rotary Process
Fang Xu, Nannan Wang, Hong Chang, Yongde Xia and Yanqiu Zhu
Abstract: This manuscript demonstrates the design, modification and initial investigation of a rotary
furnace for the manufacturing of inorganic fullerene WS2 nanoparticles. Different preparation
methods starting with various precursors have been investigated, of which the gas-solid reaction
starting with WO3 nanoparticles was the most efficient technique. Furthermore, the influence of
temperature, reaction time, and reaction gases etc. on the synthesis of inorganic fullerene WS2
nanomaterials was investigated, and these parameters were optimised based on combined
characterisations using XRD, SEM and TEM. In addition, the furnace was further modified to include
a baffled tube, a continuous gas-blow feeding system, and a collection system, in order to improve
the batch yield and realise continuous production. This technique has improved the production from
less than 1 g/batch in a traditional tube furnace to a few tens of g/batch, and could be easily scaled
up to industry level production.
Reprinted from Inorganics. Cite as: Xu, F.; Wang, N.; Chang, H.; Xia, Y.; Zhu, Y. Continuous
Production of IF-WS2 Nanoparticles by a Rotary Process. Inorganics 2014, 2, 313–333.
1. Introduction
Inorganic nanotubes and inorganic fullerene-like (IF) structures of WS2 were first discovered by
Tenne et al. in 1992 [1], which has opened a challenging field for the synthesis and applications of
numerous such layered structures, such as WS2 [1], MoS2 [2,3], BN [4], NiCl2 [5], and etc. Various
synthesis methods have been reported, such as the microwave treatment of W(CO)6 (Tungsten
carbonyl) reacting with H2S (hydrogen sulphide) [6], ultrasonic irradiation of W(CO)6 solution
mixed with diphenylmethane and sulphur followed by heating at 800 °C [6], commercial WS2
activation [7], iodine transport method [8], direct pyrolysis of WS42í and CTAB (Cetyltrimethyl
Ammonium Bromide) [9], and chemical vapour deposition [10]. Recently, IF-WS2 nanoparticles and
nanotubes have also been obtained from WCln and WOxCly reacting with H2S [11].
These IF-WS2 and IF-MoS2 nanomaterials, in addition to their significant mechanical, biocompatible
and electronic properties, are excellent solid lubricants [12–20]. Accordingly, the incorporation of
these nanomaterials into a proper matrix in composites will lead to new products with hugely
improved physical and mechanical properties. Another extraordinary property of WS2 nanostructures
is their superb shock absorbing performance [21–23], which suggests an important field of
application in lightweight and high performance protective composites [24]. Such applications will
obviously demand large amounts of IF-WS2 supply, however their synthesis was only obtained in
gram level at the early stage, which was far too less for any practical work. More recently, Tenne’s
group has produced such IFs in large quantities by using new high tower reactors [25], hence realising
a great industrial level success in nanomaterials. Nevertheless, an alternative, simple, versatile and
yet effective process for the synthesis of such novel nanomaterials remains highly desirable.

88
Therefore, to develop an innovative, simple and scalable technique that is suitable for the
continuous manufacture of IF-WS2 nanomaterials becomes the primary aim of this work. From
previous studies [25–27], where a deeper understanding of the formation mechanism of IF-WS2
nanoparticles has been gained, it is found that the key technical barriers for scaling up of the IF-WS2
lie in the powder agglomeration and superficial reaction, which occurs inevitably in a static gas–solid
reaction. In order to achieve a large quantity manufacturing, effective measures should be taken to
overcome these challenges.
This manuscript describes the design and modification of a rotary furnace, and the initial
investigations in scaling up manufacturing of IF-WS2 nanoparticles using the rotary furnace. Several
processes starting with different precursors have been investigated, of which the gas–solid reaction
using WO3 nanoparticles as the precursor was the most efficient technique. The influence of
temperature, reaction time, precursor types and reaction gases etc. on the synthesis of IF-WS2
nanomaterials will be optimised. A significantly improved batch yield and a continuous process have
been achieved.
2. Results and Discussion
2.1. Design and Modification of the Rotary Furnace
A novel rotary reactor has been designed for the manufacturing of IF-WS2 nanomaterials. The
rotary reactor was designed based on a conventional tube furnace. The furnace is 1 m long and has a
working temperature up to 1200 °C, which ensures a long enough hot zone for complete reactions at
the required temperatures.
Figure 1. Sketch of the rotary furnace. In a traditional static furnace, the WOx particles
stay still in the quartz tube at high temperature (800–900 °C), which leads to the dominance
of 2H-WS2 in the final products; whilst in the present furnace, the quartz working tubes
are rotating, forcing the WO3 and formed WS2 nanoparticles to rotate and move during
the process, resulting in better separated IF-WS2 products at minimal agglomeration.

As shown in Figure 1, the rotary furnace consists of several essential parts: a basic tube furnace,
a motorised driving and rolling system, an inclining system, a dynamic rotary seal system, a

89
continuous feeding system and a collection system. The quartz working tube can be further modified
to improve the batch yield. The working tube has inner diameter of 36 mm and outer diameter of 40
mm. Further modification was realised by adding two small oppositely positioned quartz rod blades
of Ø5 mm to the inner wall (Figure S1), to provide extra forces driving the movement of powders
inside the reactor. This would further eliminate particle agglomerations.
In order to extend the batch process to a continuous production, a continuous feeding system is
required. In the current laboratory scale trial, after considering the existing screw feeding and piston
feeding mechanisms [28–30], we adopted a simple pump piston feeder, which is more cost-effective.
Actually, the current piston feeder under gravitation and gas blow works well in the current set up,
since the slight vibration of the reactor helped avoid system blocking and keep a constant feeding. In
industrial level, a proper, accurate screw feeder could be used to replace the current feeder. As shown
in Figures 1 and S2, the feeder consists of a 20 mL push pump connected to a “T” junction, acting
as the combined inlet blow gas and nanoparticle passage channel. The conceptual rotary furnace
enables us to experiment with various parameters for optimal IF-WS2 manufacture. A simple
collection system has also been designed, as shown in the dashed box in Figure 1, which makes the
process continuous. The detailed schematic drawing of the collection system is shown in Figure S3.
The key features of the practically assembled furnace are listed in Table 1.
Table 1. Key parameters of the designed furnace.
Temperature Gas supply Feeding level Speed range Inclination angle
0–1200 °C
Ar, H2S, H2 Up to 100 g*
0–350 rpm
í5 to 5°
*: depending on numbers of loaded pump container.

2.2. IF-WS2 Synthesis by Different Methods
Different raw materials, including WO3 and S mixture under H2, ammonium paratungstate (APT)
reacting with H2S gas, and WO3 reacting with H2S, will be tested in this section.
2.2.1. WO3 and S Reaction under H2
The WO3 precursor exhibits an average particle size of around 60 nm, having a monoclinic WO3
structure (JCPDS No. 43-1035), as shown in Figures S4 and S5.
For the initial trial (experiment S1, as listed in Table 2), a composite powder of WO3 and S at a
molar ratio of a 1:10 was used as the precursor, and reacted under Ar/H2 atmosphere at 800 °C,
following the procedure described in Experimental Section 3.3.1. SEM images in Figure 2 show that
very small nanoparticles and agglomerates were observed, with some nanoplatelets. However, the
EDX spectrum shows that although WS2 has formed, WOx still dominates the products. This is also
confirmed by XRD study as shown in Figure 3. In Figure 3, the peaks labelled with a triangle matched
well with WS2 (JCPDS No. 84-1398) [26], and all the other peaks labelled with a star could be
assigned to W18O49 and W20O58. The highest WS2 peak is at 2-theta 14.3°, corresponding to (002).
The relatively low intensity of (002) plane for WS2 reveals that only a few WS2 layers have formed
from the outside and leaving behind a WOx core which has been reduced from WO3 to W18O49 and

90
W20O58. The presence of W18O49 and W20O58 is in agreement with previous studies [26,31–33],
in which partially reduced W20O58 and thoroughly reduced W18O49 were formed during the
oxide-to-sulphide conversion from WO3 to IF- WS2. There is no detectable S left in the products.
Figure 2. SEM images (a–c) and EDX spectrum (d) for products from experiment S1.
(a)

(c)

(b)

(d)

Figure 3. XRD pattern for WS2 from experiment S1 (reaction of WO3, S and H2).

Although a molar ratio of 10:1 for S and W had been initially used, it seems that there was not
enough S to react with the reduced WOx core, because the S could not stay long enough in the hot
zone. As soon as the powders reached the high temperature zone, the S would be sublimed and blown

91
out of the high temperature zone, even under an optimised Ar/H2 flow rate of 80 mL/min [34].
Therefore, the oxide-to-sulphide conversion could not be completed.
There are some similar works that used the same molar ratio of S:WO3 = 10:1 [34], but only
achieved a yield of less than 0.1 g per batch. It is obvious that this method is unsuitable for scaling
up, as the product quality depends upon the quantity of WO3 nanoparticles used and it is difficult to
increase this for traditional static furnaces. To avoid the S loss, another batch process has also been
reported recently [35], using solid NaBH4 or LiAlH4 as the H2 releasing agent in a sealed ampoules.
Again, these processes are not suitable for continuous production.
In this context, a continuous feeding and evaporation of S was tested, to compensate for the S loss
during the reaction, as shown in Figure 1. The WO3 and S mixed powder at a molar ratio of 1:10 was
fed directly into the hot zone gradually using the pump, at elevated temperature, rather than preplaced inside the quartz tube (experiment S2). The XRD analyses of the powders collected from both
the inlet and hot zones have showed that only a few WS2 layers formed and the majority remained
as WOx, and S peaks were also present (peaks labelled with circle); although samples from the hot
zone exhibited a more complete oxide-to-sulphide conversion (Figure S6).
Morphologically, the blue powders collected from the inlet zone consist of both very small
particles and big aggregates, with some bright S residues, as shown in Figure 4a,b. At the relatively
low temperature inlet zone, S could not completely react with the reduced WOx, thus only a few layer
WS2 formed. For the hot zone samples (Figure 4c,d), some big agglomerates are visible, and they are
indeed composed of nanoparticles after ultrasonic treatment.
Figure 4. SEM images of products collected from the inlet zone (a and b), and hot-zone
(c and d) of experiment S2.
(a)

(b)

(c)

(d)

92
Therefore, although this technique has the potential for scaling up based on the continuous feeding
system, the quality of the products is not high enough. Furthermore, when the temperature dropped
below its boiling point, the S vapour could block the outlet gas pipes, which could be a practical
issue. The large amounts of S consumption, although can be recycled, makes this process not
cost effective.
2.2.2. APT as Precursor and H2S as Reaction Gas
Since WO3 nanoparticles were fabricated by the decomposition of APT, as described in the SI
(Supplementary Information 2.2.2, Figures S7 and S8), and were a very valuable precursor for
IF-WS2 nanomaterials in previous two-step studies [1,25,26,36], i.e., first decomposition of APT to
form WO3 then via sulphidisation to create IF-WS2, it is thus interesting to combine the two
steps together in our present set-up. This would be an advantage in terms of energy saving and
process efficiency.
The as-received APT particles are crystals of several tens of —m in size, with nano-sized particles
attached to the surface of these big crystals. Under high magnification, cracks and sub —m sized
particles could be observed at the corners of some damaged crystals, indicating that the APT crystals
might be agglomerates of small pieces. Thus, prior to experiment (experiment AHS1), the APT
particles were ultrasonic treated, to break the agglomerates into small pieces. The IF-WS2 particles
collected from hot zone are shown in Figure 5. On average, the agglomerates are smaller (Figure 5a),
although some bigger aggregates which kept the original APT crystal shape were also observed, as shown
Figure 5a,b. Higher magnification study shown in Figure 5c and d reveals the nanostructural feature
within the agglomerates, and they are composed of both IF-WS2 nanoparticles and nanoplatelets, roughly
at the same proportion. XRD study (Figure 6) reveals that the hot zone products have very high intensity
of WS2 peaks, with only minor peaks of WOx. In contrary, the outlet zone products contain a high portion
of WOx, owing to the shorter reaction time than those from the hot zone.
To form WOx by decomposing APT, the temperature required will be above the melting point of
tungsten oxide, e.g., 1200 °C to form WO3 nanoparticles and 1350 °C for micro particles [37,38].
This allows for the tungsten oxide vapour to be brought to and deposited in the low temperature area.
In the present process, the formed nanoparticles would then play the template role during subsequent
sulphidisation, to form the IF-WS2. Because of the different temperature requirements for the two
steps, the direct IF-WS2 synthesis using APT to react with H2S seems to be an unsuitable choice for
the present furnace.

93
Figure 5. SEM images of samples from experiment AHS1, exhibiting the original shape
and size of the APT (a and b), and the nanostructural feature within the agglomerates
(c and d).
(a)

(b)

(c)

(d)

Figure 6. XRD patterns of samples collected from different areas (a, outlet and b, hot
zone) in experiment AHS1.

2.2.3. WO3 and H2S Synthesis of WS2 Nanomaterials
Figure 7 shows the SEM results of experiment W1. A feature of nanoparticle domination is
visible, with big agglomerates (Figure 7a). The semi-spherical IF-WS2 nanoparticles exhibit different
sizes (Figure 7b): the tiny ones have diameters of <50 nm which are the same as the WO3 precursor;
whilst the big ones are about 100–200 nm, possibly merging from two or more nanoparticles. The
presence of nanosheets or nanoplatelets amongst nanoparticles can be seen from Figure 7c (arrowed).

94
XRD investigation confirms that the majority of the products are IF-WS2 nanoparticles, with
left-shifted (002) peak and broadened (103) and (105) peaks [3], Figure 7d. Tiny peaks at
23–25 degrees (labelled with a star) can be assigned to WOx core residue. XRD comparison between
the produced IF-WS2 and the commercial 2H-WS2 has been presented in the Supplementary
Information, as Figure S9, where the differences between XRD pattern of the resulting IF-WS2 and
2H-WS2 are discussed in detail.
Figure 7. SEM images (a–c) and XRD pattern (d) of IF-WS2 synthesised using the
rotary process.
(a)

(c)

(b)

(d)

Further TEM examination shows that the sample contains both nanoparticles and
nanosheets/nanoplatelets (Figure 8a). Indeed, the IFs exhibit the multi-layered, hollow core
characteristics, being the dominant phase. Some particles followed the original shape of their oxide
precursor, appearing in a spherical, seamless, and close-caged structure (arrows A); whilst some
displayed a peanut-like structure (arrows B) or a long elliptical shape (arrow C). These unusual
particles were possibly co-formed from adjacent WO3 nanoparticles that fused together during the
heating. The continuous contour of WS2 layers suggests that these particles must have fused together
first, then the oxide-to-sulphide conversion occurred. This observation can also explain the different
particle sizes observed under SEM, as shown in Figure 8b,c. The products, regardless of their
different shapes, possess a hollow core and a generally equal d(002) spacing of 0.62 nm for IF-WS2.

95
Figure 8. TEM images of the WS2 synthesised using the rotary process (a–d; b is a
zoomed-in image of framed area in a).

These characterisations of the products have confirmed that IF-WS2 is the dominant phase in the
product with high quality. This shows the great potential for the production of IF-WS2 at large
quantities and high quality, using the present rotary process. Further investigations will focus on
quality assessment and quantity improvement.
2.3. Optimization of IF-WS2 Synthesis from WO3 Precursor and H2S Gas
2.3.1. Reaction Time
After applying fixed rotary speed and temperatures, the influence of different reaction times were
assessed, as listed in Table 3. Figure 9 shows the resulting products exhibiting similar features as

96
described earlier. For the 10 min product (experiment A, Figure 9a), the oxide was just about to be
reduced. Only the outer layer finished the oxide-to-sulphide conversion, whilst the inner core remained
intact, which is in line with Tenne et al.’s TEM observation [26]. For the 50 min sample (Experiment
B, Figure 9b), by analysing the intensity changes of the diffraction peaks, it is clear that the oxide
particles have mostly converted to IF-WS2 and there is much less suboxide left in the core. Similarly,
after an 80 min reaction, the WOx peaks at around 23–25 degrees continued to be reduced, and more
and more IF-WS2 layers formed, suggesting an almost complete conversion. Further increase of the
reaction time leads to no significant differences in the XRD profiles, and there is almost no peak
detected for any tungsten oxide. Thus, a 110 min reaction time is believed to be sufficient for a
thorough sulphidisation.
Figure 9. XRD profiles of samples from experiments A–F, demonstrating the effect of
different reaction time at 800 °C from 10–170 min.

2.3.2. Reaction Temperatures
Of the experiments listed in Table 3, batches C and G used similar parameters, except for the
temperature which was 900 °C for batch G but 800 °C for batch C. From 900 °C, the intensity of the
resulting IF-WS2 (Figure 10) is much higher than that from 800 °C (Figure 10), confirming more
WS2 layers formed at higher temperatures, under the same reaction time of 80 min.
2H-WS2 flakes were presented in both samples (Figure 11), however there was much more 2HWS2 formed in batch G, possibly due to the higher temperature which made agglomeration more
severe. The IF-WS2 particles are also considerably larger in batch G, some of which exhibited a
diameter of 300–500 nm, in contrast to 100–200 nm in experiment C. This observation accounted
for the extremely high relative intensity of the (002) peak for sample G.

97
Figure 10. XRD profiles of samples from batch C (800 °C, red curve) and G (900 °C,
black curve) and their SEM images.

Figure 11. SEM pictures of samples from batch C (a–c) and batch G (d–f).
(a)

(b)

(c)

(d)

(e)

(f)

Actually, when compared with other experiments from A–G, the quality of products from batch
G (the highest temperature used, 900 °C) are the worst of all, regardless of different reaction times.
Thus, temperature is considered to be one of the most significant parameters directly linked with
agglomeration in the present process.
2.3.3. Influence of H2
H2 was introduced to some experiments (e.g., experiment H1), to promote the WOx and H2S reaction.
In these samples, the majority of the product is IF-WS2 of less than 100 nm in size (Figure 12a,b), with

98
very few 2H-WS2 (Figure 12c). Rod shaped WS2 were also observed occasionally, as shown in
Figure 12d. Indeed, the formation of IF-WS2 nanotubes was previously reported in the presence of
H2 in the reaction gases [27,39,40]. The XRD profile looks identical to those without H2. Overall,
the results show a positive effect when H2 was added to the reaction gas.
Similarly, many previous studies have indicated that, during the conversion from WO3 to WS2,
the WO3 was first reduced to tungsten suboxide such as W20O58, W18O49 etc. From a localized view,
the S replaced the O as soon as the reduction of tungsten oxide initiated. From an overall point of
view, the reduction and sulfurization processes must have taken place in parallel [26,31–33]. Thus,
the addition of H2 into the reaction gases would accelerate the reduction of tungsten oxides and speed
up the subsequent sulphidization, thereby leading to prompt formation of the WS2 layers on the oxide
surface. Since the early formation of an inert WS2 layer would prohibit the diffusion and
agglomeration of neighbouring nanoparticles, H2 introduction helped prevent particles from
agglomeration, hence effectively reducing the chance for the formation of 2H-WS2.
Figure 12. SEM images of samples involving H2 in reaction, showing particles less than
100 nm (a–c), and the presence of nanotubes (d).
(a)

(b)

(c)

(d)

2.4. Further Refinement and Modification
2.4.1. IF-WS2 Synthesis Using a New Baffled Quartz Tube
To improve the batch yield, a new baffled quartz tube was designed and adopted for the
experiment (Figure S1). Based on previous optimal parameters, changes were made to reflect the

99
significantly increased precursor from 6–18 g. For example, the gas rate of H2 and H2S, the reaction
time and inclination angle etc. were finely adjusted in experiments B1–B3.
For the new baffled tube, a bigger inclination angle was required at the beginning of the
experiment in order to drive the particles towards the hot zone, otherwise they would only move
around rather than move forward. As soon as all the particles marched into the hot zone, the angle
could be decreased to zero for better reaction. In a typical experiment, 18 g WO3 nanoparticles were
used with an initial inclination angle of 5 degrees, and H2: H2S: Ar ratio of 10:30:160 for the whole
process. After the experiment, 15 g out of the 18 g were collected from the hot zone. By reducing the
H2 gas content in the flow in the later stage of the reaction (experiment B3), the amount of WS2
nanotubes has been dramatically decreased, and they were much shorter (compared to experiments B1
and B2), Figure 13a. The dominant IF-WS2 nanoparticles appeared to be uniform, fine and spherical,
with diameters <100 nm (Figure 13b). Their XRD pattern is very promising, with strong WS2 peaks
and very tiny WOx signals, which is again indicative of a good sulphidisation.
Figure 13. SEM images for samples collected from the hot zone of experiment B3,
revealing the existence of a small amount of nanotubes (a), and the fine and uniform
IF-WS2 (b).
(a)

(b)

TEM characterisation for sample B3 (Figure 14) further confirm the dominant nature of the IFWS2, with sizes ranging from below 50 nm to up to 100 nm. Figure 14b also demonstrates that some
particles coalesced from two or three nanoparticles, exhibiting a peanut shape (arrowed). Those
single particles are always <50 nm. Figure 14b–d show high resolution images of well-crystallised
IF-WS2 particles, which again reveal their typical hollow core and seamless shell layer features. Some
particles still possess a residue WOx core (particle A in Figure 14d). Nevertheless, the new baffled
working quartz tube was most promising, enabled a yield improvement from 5 g–15 g per batch whilst
successfully maintaining the products quality. However, as the mixing efficiency drops with the filling
degree [41], a too high amount of WO3 precursor input would lead to less effective mixing and thus
compromise the quality of the final product. Thus, the batch yield and product quality are limited by the
quantity of the precursors loaded at the beginning of each batch. Further improvement is still necessary.

100
Figure 14. TEM images for samples collected from the hot zone of experiment B3,
demonstrating the overall distribution of particle (a) and the multi-layered characteristics
of different particles (b–d).

2.4.2. A Continuous Feeding System
In this feeding system (Figure 1), the WO3 precursor was first stored in a pump and then
introduced to the system by gravity and gas-blow. A single long feeding tube was used to act as the
extended pathway for WOx particles directly blown into the system (as shown in Figures 1 and S2),
while the reaction gases were fed in by a separated tube. The new design was tested (FB1), and the
result showed a complete conversion of WOx into IF-WS2 for particles from the hot zone (Figure
S10). The products were quite uniform, with sizes no more than 100 nm (Figure S11).
The process was further modified to simulate a real, continuous one, by immediately replacing the
empty pump with a full one (FB2). All other experimental parameters made no alteration to FB1, except
for the longer feeding time. In this case, around 50 g samples were collected from the hot zone. TEM
images of the hot zone sample from FB2 were shown in Figure 15. Majority nanoparticles of around 50
nm in size are displayed in Figure 15a,c, some from 20 nm to almost 100 nm are exhibited in Figure 15b.
A typical IF-WS2 particle with a hollow core and around 15 seamless layers is shown in Figure 15d.
To summarise, the continuous feeding system has been proven a success and has improved the
yield of IF-WS2 to several tens of grams per batch, without having an obvious compromise on
quality. Technically, the feeding of precursor could be simply continued by reloading; however, a
longer reaction time would be required. In an industry environment, where a proper sample collection
system is easily available, an automated reaction would not be a limitation. This reactor is easily

101
adaptable as a whole continuous rotary process for the scaling up production of IF-WS2
nanoparticles, using a proper metallic working tube.
Figure 15. TEM images for particles collected from the hot zone from experiment FB2,
showing overall uniformity of particles (a and c) and detailed closed-cage feature under
high resolution (b and d).

3. Experimental Section
3.1. Materials
Commercial tungsten trioxide powder used as precursor for IF-WS2 nanomaterials was
monoclinic WO3 powder (light yellow, with size < 100 nm, purity 99%, Sigma-Aldrich, UK).
Ammonium paratungstate (APT, (NH4)10H2(W2O7)6·xH2O, 99.99%) particles used to produce WOx
nanoparticles were bought from Sigma-Aldrich, UK. Sulphur (S, sublimed, 100 mesh, 99.5%) was
purchased from Cole-Parmer, UK. For comparison with the produced IF-WS2 nanoparticles,
2H-WS2 platelets (powder, 2 —m, 99%) were also bought from Sigma-Aldrich, UK.
3.2. Experimental Set-Up
Based on the furnace shown in Figure 1, the parameters for the experiments were optimized
through a series of experiments, such as different reaction temperatures, reaction time, reaction gas
atmosphere, modification of working tube, and a new feeding system. The optimization experiments
were carried out without the feeding system, unless specifically stated. The products were collected
from different areas separately, where the central 80 cm part of the quartz tube in the heated zone
was defined as hot zone, the 60 cm part close to the gas inlet end was named as the inlet zone, and

102
similarly the outlet zone, as also shown in Figure 1. Note that the reaction time could be varied mostly
through the inclination of the quartz reactor, and the motor could drive the working quartz tube both
clockwise and anti-clockwise. In such a way, the particles could be moved either backwards or
forwards as required, to achieve sufficiently long reaction time.
The morphology and chemical composition of the as-produced IF-WS2 nanoparticles were
analysed by XRD, SEM and TEM. The TEM samples were prepared via an ultrasonic treatment of
the nanoparticles (randomly collected from hot zone) for 5 min in acetone, and the resultant uniform
suspension was pipetted onto a holey carbon coated copper grid (300 mesh, Agar).
3.3. IF-WS2 Synthesis by Different Methods
3.3.1. Synthesis of IF-WS2 Nanoparticles from Mixed WO3 and S Powder with H2
Two typical types of experiment were carried out to synthesize the IF-WS2, starting from mixed
WO3 and S powder with H2 (experiments S1 and S2). In experiment S1, around 6 g WO3 and S
composite powder at a molar ratio of 1:10 was placed into the inlet area. Ar gas was used to purge
the quartz tube at 50 mL/min for 30 min, after the system was properly assembled and sealed. The
working tube started rotating from 600 °C, at which point H2 (80 mL/min) and Ar (30 mL/min) gases
were also introduced. The reaction at 800 °C lasted for 1 h. Similarly, experiment S2 was carried out
by adopting a continuous S feeding system.
3.3.2. Synthesis of IF-WS2 Nanoparticles from APT Precursor with H2S Gas
Six g APT were ultrasonically treated and dried before being placed into the centre of the quartz
tube for the experiment (AHS1). Ar (50 mL/min) was flushed for 30 min before the furnace was
switched on. H2S flowing was started from 550 °C till the end of the process, at a rate of 12 mL/min.
The whole reaction lasted for 2 h at 800 °C, at a fixed Ar flow rate of 100 mL/min.
Except for experiments on various reaction times (summarised separately), all other experimental
parameters are summarised in Table 2.
Table 2. Example parameters used for the synthesis of IF-WS2.

S1
S2
AHS1
W1
H1
B1
B2

Precursor put in
(g)
6
12
6
6
6
18
18

B3

18

FB1

50

FB2

100

Gas feed Rate
(H2:H2S:Ar) (mL/min)
80:0:30
80:0:30
0:12:100
0:12:100
8:12:100
10:30:160
20:30:100
20:30:100
10:30:100
50:50:100
0: 25:100
50:50:100
0:25:100

Reaction time
(min)
60
120
120
90
90
120
120

Reaction
temperature (°C)
800
800
800
800
800
800
800

120

800

180

800

300

800

103
3.3.3. Synthesis of IF-WS2 Nanoparticles WO3 and H2S Gas
The synthesis of WS2 nanomaterials from reaction of WO3 and H2S gas was similar to that from
WO3 and S (Section 3.3.1), except for the H2 gas being replaced by H2S. The first trial was named
experiment W1 (Table 2).
Table 3. Example parameters used for the synthesis of IF-WS2 with varying reaction times.
WO3 Precursor (g)
A
B
C
D
E
F
G

6
6
6
6
6
6
6

Gas feed rate (H2S:Ar) Reaction time Reaction temperature
(mL/min)
(min)
(°C)
12:100
10
800
12:100
50
800
12:100
80
800
12:100
110
800
12:100
140
800
12:100
170
800
12:100
80
900

Table 3 summarises the parameters used for a series of experiment at a fixed rotation speed of
140 rpm, with reaction time varying from 10–170 min.
4. Conclusions
Based on the built rotary system, the large scale manufacture of IF-WS2 nanoparticles has been
realised. Several synthesis methods have been studied, and the process starting with WO3 precursor
and H2S reaction gas is most successful. Systematic studies have been carried out to optimise the
parameters for IF-WS2 nanoparticle production, including precursor types, reaction temperatures,
reaction time and reaction gases. Further refinements by modifying the working quartz tube and the
feeding system have made the continuous manufacturing possible. This new technique, as a simple
alternative to the fluidised bed method, has improved the yield of IF production from less than
1 g/batch using a traditional tube furnace to a few tens of g/batch. This process is easily scalable to
industry production level, by incorporating existing equipment.
Supporting Information
1. Results and Discussion
1.1. Design and Modification of a Rotary Furnace
The quartz tube modification is shown in Figure S1. Two small quartz rod blades of Ø5 mm were
oppositely attached to quartz tube.

104
Figure S1. Image (left) and sketch (right) of a new tube with baffles inside, created by
adding two quartz rods inside the normal quartz tube.

In order to move the batch process to a continuous production, a continuous feeding system was
designed using a 20 mL syringe piston feeder which was connected to a push type ‘T’ junction, as
shown in Figure S2.
Figure S2. Sketch (a) and picture (b) of a new gas blow feeding system, to realise
continuous feeding.
(a)

(b)
WO3 carried by Ar
H2S

Figure S3. Sketch of the collection system, where (1) quartz tube and (2) housing are
rotating while all other parts are kept still during action.

105
The collection system makes the process a real continuous one, as shown in Figure S3. During
the process, the quartz tube (1) and housing were driven to rotate by the motor, while the shaft (3)
was kept static, as well as the gas outlet (4) and collection outlet (5 and 6) attached to it. Note that
the product discharge outlet allows for easy isolation to collect samples for quality control during the
manufacturing.
2.2. IF-WS2 Synthesis by Different Methods
2.2.1. Mixed WO3 and S with H2
The SEM images of the as-received WO3 precursor have been displayed in Figure S4; and the
average particles size was about 60 nm, estimated using Derby-Scherrer Equation. The XRD pattern
shown in Figure S5 matches very well with the monoclinic WO3 (JCPDS No. 43-1035).
Figure S4. SEM images of the as-received WO3 nanoparticles (~60 nm in size).

Figure S5. XRD pattern of the as-received WO3 precursor.

106
2.2.1. WO3 and S reaction under H2
The XRD patterns of Sample S2 from both the inlet and the hot zones are shown in Figure S6.
Only a few WS2 layers may have formed, as the majority was WOx, along with S peaks (labelled
with circle), although the hot zone samples exhibited a more complete oxide-to-sulphide conversion.
Figure S6. XRD patterns of Sample S2. (a) from the inlet zone and (b) from the central
hot zone.

2.2.2. APT as Precursor and H2S as Reaction Gas—Production of WOx Nanoparticles from APT
WOx nanoparticles, as an important precursor for IF-WS2 nanoparticles, have been produced by
the decomposition of APT at a high temperature furnace. At high temperature, the following reaction
took place: (NH4)10[H2W12O42] 4H2O ĺ 12 WO3 + 10NH3 + 11H2O.
In fact, the APT started to decompose into WO3 at 500 °C. The WO3 then started to melt and
vaporise in the central zone where temperature reached 1200 °C and 1250 °C respectively, and the
WOx vapour would be brought downstream by the Ar flow and deposit at the inner wall of the
quartz tube.
After optimization of the parameters, very fine nanoparticles with high batch yield were produced
at 1350 °C, with long quartz tube (1.5 m) and high gas flow rate (6 L/min). The results were shown
in Figure S7a–c. Both spherical and polyhedral particles are presented in all images, with uniform
size less than 100 nm, mostly around 50 nm. XRD pattern in Figure S7d matches very well with
monoclinic WO3 (JCPDS No. 43-1035), with the strongest peaks appear at 23–25 degree.

107
Figure S7. SEM images (a–c), and XRD pattern (d) of tungsten oxide nanoparticle
(1350°C, 6 L/min, 30 min, 1.5 m).
(d)

Figure S8. TEM images of WO3 from decomposition of APT (1350, 1.5 m, 6 L/min,
30 min).
(a)

(b)

Figure S8 shows the TEM images of WO3 nanoparticles from the decomposition of APT. The
majority of nanoparticles are of a polyhedral shape, around 50 nm.
2.2.3. WO3 and H2S Synthesis of WS2 Nanomaterials
A comparison of the XRD profiles of the commercial 2H-WS2 and the present IF-WS2 samples is
shown in Figure S9. Both patterns showed peaks at very similar positions. The peaks of IF-WS2 were
assigned according to 2H-WS2 (JCPDS No. 84-1398), as no standard XRD pattern is available for
IF-WS2 [1]. These are the typical peaks at 2ș angles at: 14.364 (002), 28.959 (004), 32.769 (100),
33.587 (101), 35.943 (102), 39.599 (103), 44.055 (006), 44.289 (104), 49.798 (105), 55.977 (106),
57.495 (110), 60.010 (008), 60.573 (112), 62.746 (107), 69.169 (201), 70.080 (108), 72.957(203),

108
and 76.040(116). For the 2H-WS2, all peaks are very sharp, indicating a well-crystallized and
standard 2H structure. Its (002) is the strongest peak, followed by (103) as the second strongest.
Compared with 2H-WS2, the (002) peak of IF-WS2 is left-shifted, indicating a lattice expansion of
the (002) layers due to stains in the curved closed-cage layers [2]. The (103) and (105) peaks are
broadened, attributing to the ultra-low dimensions. Whilst the (002) remains as the strongest peak,
two peaks representing (100) and (101) have merged into one exhibiting the second highest intensity.
The (006) and (104) peaks also merged, at around 44.1 degree. The (102) peak at around 35.9 degree
was not detected in the IF-WS2 profile. The small peaks appeared at 23-25 degree in the IF-WS2
pattern were assigned to WOx, which must exist as the residue core of some IF-WS2 particles.
Figure S9. XRD profiles of commercial 2H-WS2 and currently synthesised IF-WS2.

2.3. Further Refinement and Modification
2.3.1. IF-WS2 Synthesis Using a Continuous Feeding System—Experiments Based on Both New
Feeding System and Baffled Tubes are Named FBn
The XRD patterns of samples collected from Sample FB1 (Figure S10) demonstrate a completely
conversion of WOx into WS2. No WOx peaks for the hot zone samples and extremely minor peaks of
WOx for the outlet zone samples.

109
Figure S10. XRD profiles of samples collected from different reaction zones in
experiment FB1.

The SEM images in Figure S11 have revealed the morphology of particles collected from the hot
zone in FB1. The particles were quite uniform, with sizes no more than 100 nm. The very bright
pieces at the right bottom corner of Figure S11b could be residue S, owing to the excessive amount
of H2S gas used.
Figure S11. SEM images for particles collected from hot zone of experiment FB1.
(a)

(b)

References
1.

2.

Feldman, Y.; Frey, G.L.; Homyonfer, M.; Lyakhovitskaya, V.; Margulis, L.; Cohen, H.; Hodes, G.;
Hutchison, J.L.; Tenne, R. Bulk synthesis of inorganic fullerene-like MS2 (M = Mo, W) from
the respective trioxides and the reaction mechanism. J. Am. Chem. Soc. 1996, 118, 5362–5367.
Feldman, Y.; Wasserman, E.; Srolovitz, D.J.; Tenne, R. High rate, gas phase growth of MoS2
nested inorganic fullerenes and nanotubes. Science 1995, 267, 222–225.

110
Acknowledgments
The authors thank EPSRC (Engineering and Physical Sciences Research Council, UK) for
financial support.
Author Contributions
Fang Xu designed, tested and modified the rotary furnace, conducted the experiments and drafted
the manuscript; Nannan Wang and Hong Chang helped with some experiments and particle
characterisations; Yongde Xia contributed to the data analyses; Yanqiu Zhu oversaw the entire
project progress, and was the project leader.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.

Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and cylindrical structures of tungsten
disulphide. Nature 1992, 360, 444–446.
2. Margulis, L.; Salitra, G.; Tenne, R.; Talianker, M. Nested fullerene-like structures. Nature 1993,
365, 113–114.
3. Feldman, Y.; Wasserman, E.; Srolovitz, D.J.; Tenne, R. High rate, gas phase growth of MoS2
nested inorganic fullerenes and nanotubes. Science 1995, 267, 222–225.
4. Chopra, N.G.; Luyken, R.J.; Cherrey, K.; Crespi, V.H.; Cohen, M.L.; Louie, S.G.; Zettl, A.
Boron nitride nanotubes. Science 1995, 269, 966–967.
5. Rosenfeld, H.Y.; Grunbaum, E.; Tenne, R.; Sloan, J.; Hutchison, J.L. Cage structures and
nanotubes of NiCl2. Nature 1998, 395, 336.
6. Vollath, D.; Szabo, D.V. Synthesis of nanocrystalline MoS2 and WS2 in a microwave plasma.
Mater. Lett. 1998, 35, 236–244.
7. Mackie, E.B.; Galvan, D.H.; Adem, E.; Talapatra, S.; Yang, G.L.; Migone, A.D. Production of
WS2 nanotubes by an activation method. Adv. Mater. 2000, 12, 495–498.
8. Remskar, M.; Skraba, Z.; Regula, M.; Ballif, C.; Sanjine’s, R.; Levy, F. New crystal structures
of WS2: microtubes, ribbons, and ropes. Adv. Mater. 1998, 10, 246–249.
9. Li, Y.D.; Li, X.L.; He, R.R.; Zhu, J.; Deng, Z.X. Artificial lamellar mesostructures to WS2
nanotubes. J. Am. Chem. Soc. 2002, 124, 1411–1416.
10. Li, X.L.; Ge, J.P.; Li, Y.D. Atmospheric pressure chemical vapor deposition: An alternative
route to large-scale MoS2 and WS2 inorganic fullerene-like nanostructures and nanoflowers.
Chem. A Eur. J. 2004, 10, 6163–6171.
11. Margolin, A.; Deepak, F.L.; Popovitz-Biro, R.; Bar-Sada, M.; Feldman, Y.; Tenne, R.
Fullerene-like WS2 nanoparticles and nanotubes by the vapor-phase synthesis of WCln and H2S.
Nanotechnology 2008, 19, doi:10.1088/0957-4484/19/9/095601.

111
12. Seifert, G.; Terrones, H.; Terrones, M.; Jungnickel, G.; Frauenheim, T. Structure and electronic
properties of MoS2 nanotubes. Phys. Rev. Lett. 2000, 85, 146–149.
13. Dominko, R.; Arcon, D.; Mrzel, A.; Zorko, A.; Cevc, P.; Venturini, P.; Gaberscek, M.;
Remskar, M.; Mihailovic, D. Dichalcogenide nanotube electrodes for Li-Ion batteries.
Adv. Mater. 2002, 14, 1531–1534.
14. Kaplan-Ashiri, I.; Tenne, R. Mechanical properties of WS2 nanotubes. J. Cluster Sci. 2007, 18,
549–563.
15. Rapoport, L.; Feldman, Y.; Homyonfer, M.; Cohen, H.; Sloan, J.; Hutchison, J.L.; Tenne, R.
Inorganic fullerene-like material as additives to lubricants: Structure–function relationship.
Wear 1999, 225–229, 975–982.
16. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Ivanovskaya, V.; Heine, T.; Seifert, G.; Wiesel, I.;
Wagner, H.D.; Tenne, R. On the mechanical behavior of WS2 nanotubes under axial tension and
compression. Proc. Natl. Acad. Sci. 2006, 103, 523–528.
17. Rapoport, L.; Bilik, Y.; Feldman, Y.; Homyonfer, M.; Cohen, S.R.; Tenne, R. Hollow
nanoparticles of WS2 as potential solid-state lubricants. Nature 1997, 387, 791–793.
18. Hou, X.H.; Shan, C.X.; Choy, K.L. Microstructures and tribological properties of PEEK-based
nanocomposite coatings incorporating inorganic fullerene-like nanoparticles. Surf. Coat. Technol.
2008, 202, 2287–2291.
19. Brown, S.; Musfeldt, J.L.; Mihut, I.; Betts, J.B.; Migliori, A.; Zak, A.; Tenne, R. Bulk vs.
nanoscale WS2: Finite size effects and solid-state lubrication. Nano Lett. 2007, 7, 2365–2369.
20. Frey, G.L.; Elani, S.; Homyonfer, M.; Feldman, Y.; Tenne, R. Optical-absorption spectra of
inorganic fullerene like MS2 (M = Mo, W). Phys. Rev. B 1998, 57, 6666–6671.
21. Zhu, Y.Q.; Sekine, T.; Li, Y.H.; Wang, W.X.; Fay, M.W.; Edwards, H.; Brown, P.D.; Fleischer, N.;
Tenne, R. WS2 and MoS2 inorganic fullerenes—super shock absorbers at very high pressures.
Adv. Mater. 2005, 17, 1500–1503.
22. Zhu, Y.Q.; Sekine, T.; Li, Y.H.; Fay, M.W.; Zhao, Y.M.; Poa, C.H.P.; Wang, W.X.; Roe, M.J.;
Brown, P.D.; Fleischer, N.; et al. Shock-absorbing and failure mechanisms of WS2 and MoS2
nanoparticles with fullerene-like structures under shock wave pressure. J. Am. Chem. Soc. 2005,
127, 16263–16272.
23. Zhu, Y.Q.; Sekine, T.; Brigatti, K.S.; Firth, S.; Tenne, R.; Rosentsveig, R.; Kroto, H.W.;
Walton, D.R.M. Shock-wave resistance of WS2 nanotubes. J. Am. Chem. Soc. 2003, 125,
1329–1333.
24. Hogg, P.J. Composites in Armor. Science 2006, 314, 1100–1101.
25. Feldman, Y.; Zak, A.; Popovitz-Biro, R.; Tenne, R. New reactor for production of tungsten
disulfide hollow onion-like (inorganic fullerene-like) nanoparticles. Solid State Sci. 2000, 2,
663–672.
26. Feldman, Y.; Frey, G.L.; Homyonfer, M.; Lyakhovitskaya, V.; Margulis, L.; Cohen, H.;
Hodes, G.; Hutchison, J.L.; Tenne, R. Bulk synthesis of inorganic fullerene-like MS2 (M = Mo,
W) from the respective trioxides and the reaction mechanism. J. Am. Chem. Soc. 1996, 118,
5362–5367.

112
27. Rothschild, A.; Sloan, J.; Tenne, R. Growth of WS2 nanotubes phases. J. Am. Chem. Soc. 2000,
122, 5169–5179.
28. Dai, J.J.; Cui, H.P.; Grace, J.R. Biomass feeding for thermochemical reactors. Prog. Energy
Combust. Sci. 2012, 38, 716–736.
29. Wilson, D.H.; Dunnington, D.L. Plan ahead to avoid feeding problems. Chem. Eng. N. Y. 1991,
98, 72–81.
30. Hao, X.H.; Guo, L.J.; Mao, X.; Zhang, X.M.; Chen, X.J. Hydrogen production from glucose
used as a model compound of biomass gasified in supercritical water. Int. J. Hydrogen Energy
2003, 28, 55–64.
31. Deepak, F.L.; Tenne, R. Gas-phase synthesis of inorganic fullerene-like, structures and
inorganic nanotubes. Cent. Eur. J. Chem. 2008, 6, 373–389.
32. Rosentsveig, R.; Margolin, A.; Feldman, Y.; Popovitz-Biro, R.; Tenne, R. WS2 nanotube
bundles and foils. Chem. Mater. 2002, 14, 471–474.
33. Margolin, A.; Rosentsveig, R.; Albu-Yaron, A.; Popovitz-Biro, R.; Tenne, R. Study of the
growth mechanism of WS2 nanotubes produced by a fluidized bed reactor. J. Mater. Chem.
2004, 14, 617–624.
34. Alaei, M.; Rashidi, A.; Mahjoub, A. Two suitable methods for the preparation of inorganic
fullerene-like (IF) WS2 nanoparticles. Iran. J. Chem. Chem. Eng. 2009, 28, 91–98.
35. Wiesel, I.; Arbel, H.; Albu-Yaron, A.; Popovitz-Biro, R.; Gordon, J.M.; Feuermann, D.; Tenne, R.
Synthesis of WS2 and MoS2 fullerene-like nanoparticles from solid precursors. Nano Res. 2009,
2, 416–424.
36. Zhu, Y.Q.; Hsu, W.K.; Terrones, H.; Grobert, N.; Chang, B.H.; Terrones, M.; Wei, B.Q.;
Kroto, H.W.; Walton, D.R.M.; Boothroyd, C.B.; et al. Morphology, structure and growth of
WS2 nanotubes. J. Mater. Chem. 2000, 10, 2570–2577.
37. Li, Y.H.; Zhao, Y.M.; Ma, R.Z.; Zhu, Y.Q.; Fisher, N.; Jin, Y.Z.; Zhang, X.P. Novel route to
WOx nanorods and WS2 nanotubes from WS2 inorganic fullerenes. J. Phys. Chem. B 2006, 110,
18191–18195.
38. Zhu, Y.Q.; Hu, W.B.; Hsu, W.K.; Terrones, M.; Grobert, N.; Hare, J.P.; Kroto, H.W.;
Walton, D.R.M.; Terrones, H. Tungsten oxide tree-like structures. Chem. Phys. Lett. 1999, 309,
327–334.
39. Zak, A.; Sallacan-Ecker, L.; Margolin, A.; Feldman, Y.; Popovitz-Biro, R.; Albu-Yaron, A.;
Genut, M.; Tenne, R. Scaling up of the WS2 nanotubes synthesis. Fullerenes Nanotubes
Carbon Nanostruct. 2011, 19, 18–26.
40. Zak, A.; Sallacan-Ecker, L.; Margolin, A.; Genut, M.; Tenne, R. Insight into the growth
mechanism of WS2 nanotubes in the scaled-up fluidized-bed reactor. Nano 2009, 4, 91–98.
41. Finnie, G.J.; Kruyt, N.P.; Ye, M.; Zeilstra, C.; Kuipers, J.A.M. Longitudinal and transverse
mixing in rotary kilns: A discrete element method approach. Chem. Eng. Sci. 2005, 60,
4083–4091.

113

Gas-Phase and Microsolvated Glycine Interacting with Boron
Nitride Nanotubes. A B3LYP-D2* Periodic Study
Albert Rimola and Mariona Sodupe
Abstract: The adsorption of glycine (Gly) both in gas-phase conditions and in a microsolvated
state on a series of zig-zag (n,0) single-walled boron nitride nanotubes (BNNTs, n = 4, 6, 9 and 15)
has been studied by means of B3LYP-D2* periodic calculations. Gas-phase Gly is found to be
chemisorbed on the (4,0), (6,0) and (9,0) BNNTs by means of a dative interaction between the NH2
group of Gly and a B atom of the BNNTs, whose computed adsorption energies are gradually
decreased by increasing the tube radius. On the (15,0) BNNT, Gly is found to be physisorbed with an
adsorption driving force mainly dictated by S-stacking dispersion interactions. Gly adsorption in a
microsolvated environment has been studied in the presence of seven water molecules by
progressively microsolvating the dry Gly/BNNT interface. The most stable structures on the (6,0),
(9,0) and (15,0) BNNTs present the Gly/BNNT interface fully bridged by the water solvent molecules;
i.e., no direct contact between Gly and the BNNTs takes place, whereas on the (4,0) BNNT the most
stable structure presents a unique direct interaction between the COOí Gly group and a B atom of the
nanotube. Further energetic analyses indicate that the (6,0), (9,0) and (15,0) BNNTs exhibit a low
water affinity, which favors the Gly/water interactions upon BNNT coadsorption. In contrast, the (4,0)
BNNT has been found to show a large water affinity, bringing the replacement of adsorbed water by
a microsolvated glycine molecule as an unfavorable process.
Reprinted from Inorganics. Cite as: Rimola, A.; Sodupe, M. Gas-Phase and Microsolvated Glycine
Interacting with Boron Nitride Nanotubes. A B3LYP-D2* Periodic Study. Inorganics 2014, 2,
334–350.
1. Introduction
Bioconjugated nanostructured materials resulting from the coupling of biomolecules with
inorganic nanomaterials including nanotubes, nanowires, nanoparticles and nanosheets have attracted
much attention during the last years as they exhibit unique features derived from combining
synergistically the properties of the interacting components. These exclusive physico-chemical
properties render these materials as suitable substrates with potential applications in diverse
biological- [1,2] and material-related [3] areas such as biocatalysis,[4,5], drug delivery [6–8],
biosensing [9–13] and medical diagnostics [14,15]. The functionalities resulting from these biohybrid
materials are largely mediated by the biomolecule/inorganic surface interactions, which in turn are
dictated by the structure-specific binding properties of the two partners. Accordingly, precise
knowledge on the interactions between the biomolecule and the inorganic components is of
fundamental relevance.
Among the different nanostructured materials, boron nitride nanotubes (BNNTs) have been
proposed to be suitable candidates to be combined with biomolecules [16]. BNNTs are isosteres and
structurally similar to carbon nanotubes (CNTs), in which alternating B and N atoms substitute for C

114
atoms. However, these two type of nanotubes exhibit different physico-chemical properties. Whereas
CNTs exhibit metallic or semiconducting behavior, which moreover is strongly dependent on the tube
diameter, helicity and concentric layers, BNNTs are electrical insulators with a band gap of ca. 5.5 eV
regardless of the tube geometry features [17]. Moreover, at variance with the non-polar C-C bonds in
CNTs, the B-N bonds of BNNTs exhibit a certain polar character, the degree of which depends on the
curvature of the nanotube. That is, the increase of the tube curvature induces the transformation of the
sp2 hybrid character of the B and N atoms in large diameter BNNTs into a sp3 one in small diameter
BNNTs. As recently shown by us [18], this has important consequences for the nature of interaction
between functional molecules and the BNNTs walls; i.e., polar molecules strongly chemisorb on small
radius BNNTs, whereas interaction of non-polar molecules are energetically more favourable when
physisorbed on large radius BNNTs. Furthermore, unlike CNTs, which present an inherent
cytotoxicity [19], BNNTs have been found to be nontoxic [20] due to their high chemical and
structural stability and high oxidation resistance, which alongside their uniformity and stability in
dispersion in solution [21] make them suitable for biomedical applications.
Different experimental studies have focused on the interaction of peptides and proteins with
BNNTs, showing a natural affinity between the two conjugates, which allows a direct immobilization
of proteins on the BNNTs [22] as well as the isolation of individual BNNTs through a novel pathway
based on peptide wrapping [23]. Moreover, biofunctionalized BNNTs via glycine interaction are good
reactant substrates to obtain polysaccharide-coated BNNTs under mild conditions, in which the
role of glycine is crucial during the interfacial process. The interactions of DNA and RNA with
BNNTs have also been addressed and exploited to obtain nematic ordered ensembles of BNNT [24].
Other works have been devoted to assess the cytotoxicity of BNNTs when in contact with cells.
Chen et al. [20] concluded that pristine BNNTs are inherently non-cytotoxic in view of the non-altered
growth of human embryonic kidney (HEK) cells when cultured with BNNTs. Similar results were
found by Ciofani and coworkers, in which coated-BNNTs presented a good cytocompatibility with
human cells [25–28]. However, Goldberg and coworkers more recently found that BNNTs are
actually cytotoxic for cells present in the lung alveoli and for HEK, in which the discrepancies with
the other works were discussed and suggested to be due to the different morphology and size
distribution of the BNNTs tested and the different assay techniques [29].
Theoretical works, mainly based on density functional theory (DFT) methods, have also studied
the interaction of biomolecules with boron nitride nanostructures, most studies being limited to
biomolecule building blocks, (amino acids and DNA and RNA nucleobases) due to the demanding
computational cost of these calculations. Works on the gas-phase interaction of nucleobases using the
local density approximation (LDA) and generalized gradient approximation (GGA) DFT levels of
theory showed that this depends on the individual polarizations of the nucleobases [30–32]. The
interaction of BNNTs with glycine (Gly) among other different amines has been studied in the
gas-phase revealing an affinity of the BNNT with the NH2Gly group [33]. Study on the gas-phase
interaction of the arginine (Arg), aspartic acid (Asp) and tryptophane (Trp) amino acids, with basic,
acidic and aromatic side chain functionalities, respectively, at the LDA DFT level revealed that the
binding is accompanied by charge transfer following the trend of Arg > Asp > Trp [34]. The binding
of different biomolecules inside the cavity of BNNTs has also been studied at the LDA level [35].

115
The calculated weak interactions led the authors to suggest BNNTs to be suitable biological carriers
due to the limited delivery kinetic barrier.
All these works focus on the intrinsic adsorption properties; i.e., they are limited to the gas-phase
and, accordingly, solvation effects were not accounted for. Moreover, each work addresses the
interaction of biomolecules with a particular BNNT. Since it has been shown that the tubular radius
can modulate the adsorption properties of BNNTs [18], which is also applicable for biomolecules, the
nature of interaction can significantly be different depending on the radius of the BNNT. Moreover,
for the particular works addressing the interaction of amino acids, no conformational exploration to
find out the most stable amino acid/BNNT adduct was performed (i.e., the initial amino acid
conformation guesses were the most stable gas-phase structure), which is an important drawback due
to the large conformational mobility of these molecules. In order to provide a more complete
atomic-scale description of the interaction of amino acids with BNNTs, the present work reports a
systematic periodic B3LYP-D2* study, using a hybrid functional and including dispersion corrections,
on the interaction of Gly with different zig-zag (n,0) single-walled BNNTs (n = 4, 6, 9 and 15)
rendering nanotubes of different radius. Note that dispersive effects, not included in previous works,
are expected to play a role in these systems. Moreover, with the aim to study in a more realistic way
the interactions between biological systems and BNNTs, the very same Gly/BNNTs interaction study
has been addressed considering a microsolvated environment modeled by the presence of seven water
molecules. The effect of water has been analyzed from a structural and energetic point of view, with
particular attention paid to the Gly/BNNT interface to determine whether the interaction is direct or
bridged by the water molecules.
2. Results and Discussion
2.1. Gas-Phase Interaction
An exhaustive description relative to the modeling of the different BNNTs used in this work in
their pristine state is available in a very recent work by us [18]. The BNNTs are automatically
generated by the CRYSTAL code through the wrapping of a periodic boron nitride monolayer
(hexagonal P3m1 layer symmetry group) into cylinders of different radius and fully exploiting the
symmetry operators of the nanotubes [36–40]. In the interest of the present work, it is worth
mentioning that the calculated electrostatic potential maps indicate a prominent positive/negative
valued potential region for the (4,0) BNNT, which become progressively less pronounced for BNNT
with increasing radius, until obtaining a practically shallow electrostatic potential for the (15,0)
BNNT. For the interaction with Gly, the geometry optimizations were carried out as 1D polymers
within the P1 space group, in which the unit cell parameters have been enlarged twice to avoid lateral
interactions between molecules of adjacent unit cells.
The different optimized adducts for the adsorption of Gly on the BNNTs are shown in Figure 1,
whereas the calculated adsorption energies alongside the pure electronic and dispersion contributions
are shown in Table 1. The adsorption energies (¨Eads) per unit cell of the probe molecules with the
BNNTs are computed as:

116
¨Eads = E(Gly/BNNT) í E(BNNT) í Em(Gly)

(1)

where E(Gly/BNNT) is the energy of a fully relaxed unitary cell containing the BNNT in interaction
with Gly, E(BNNT) is the energy of a fully relaxed unitary cell of the BNNTs alone, and Em(Gly) is
the molecular energy of the free Gly.
Figure 1. (a) B3LYP-D2* optimized structures of the different calculated adducts for
glycine interacting with the considered boron nitride nanotubes (BNNTs) in the gasphase; (b) B3LYP-D2* optimized structures for those complexes in which a spontaneous
proton transfer occurs during the geometry optimization (see text). Distances in Å: bare
values for the (4,0) BNNT; values in parenthesis for the (6,0) one; values in brackets for
the (9,0) one; and italic underlined values for the (15,0) one.

For all the BNNT systems, six initial structural guesses were considered (see Figure 1a): (i) pure
interaction between the NH2 group of Gly and one nanotube B atom (hereafter referred to as
BN/NH2); (ii) BN/NH2 interaction plus H-bonding between the Gly OH group and one nanotube N
atoms (hereafter referred as BN/NH2–OH); (iii) and (iv) interaction between the Gly CO group and
one nanotube B atom plus H-bonding between the Gly OH group and one nanotube N atom, with the
difference that the nanotube B and N atoms are chemically bonded or not (hereafter referred to as
BN/COOH-1 and BN/COOH-2, respectively); (v) interaction of Gly in its zwitterionic form
(hereafter referred to as BN/zwitt), in which the Gly COOí group interacts with one B atom and the

117
Gly NH3+ is H-bonded to one nanotube N atom; and (vi) interaction between Gly and the BNNTs
purely through the S system of the Gly COOH group (hereafter referred to as BN/S).
Data reported in Table 1 clearly indicate that the most stable adduct for the (4,0) BNNT is the
BN/NH2-OH whereas for the (6,0) BNNTs both BN/NH2 and BN/NH2–OH are nearly degenerate.
These complexes result from dative covalent interactions between the Gly NH2 group and the B atoms,
which act as Lewis acid sites. These findings are consistent with the data reported for the interaction
of these BNNTs with probe molecules, in which the interaction of NH3 with the (4,0) and (6,0) BNNTs
was found to be the strongest one among all tested molecules [18]. The fact that the BN/NH2-OH
adduct becomes the most stable one for the Gly/BNNT(4,0) system is consistent with the large polar
character of this nanotube, the N atoms acting as H-bonding acceptor groups. This is not in line with
the most stable BN/NH2 adduct found for the (6,0) BNNT and is due to the weaker H-bonding acceptor
character of the N nanotube atom when increasing the nanotube radius, which is reflected by an
increase of the H-bond distance in the BN/NH2–OH adducts (1.686 and 1.748 Å for the (4,0) and (6,0)
BNNTs, see Figure 1a). In the same way, the B-NGly bond lengths of the dative interactions in the
BN/NH2 and BN/NH2–OH complexes also increase with increases in the nanotube radius due to the
progressive decrease of the Lewis acceptor character of the B atom. Because of that, the calculated
adsorption energy (¨Eads) is more negative and larger for the most stable BN/NH2–OH adduct of
Gly/BNNT(4,0) than the most stable BN/NH2 adduct of Gly/BNNT(6,0) (i.e., í33.2 and í18.9 kcal
molí1, respectively). The calculated energetic contributions; i.e., purely electronic and dispersion (¨Eel
and ¨ED values of Table 1, respectively) indicate that Gly adsorption on the (4,0) BNNT is largely
dictated by the covalent dative interaction, whereas on the (6,0) BNNT ¨Eel decreases in favor of
dispersion. An analysis of the Mulliken charges (Q values of Table 2, only limited to the most stable
Gly/BNNT complexes) confirms the formation of charge transfer complexes for both the
Gly/BNNT(4,0) and Gly/BNNT(6,0) systems, the computed charge transfer values from Gly to the
BNNTs being 0.30 e and 0.22 e, respectively.
The most stable Gly/BNNT(15,0) adduct has been found to be the BN/S one. This is in perfect
agreement with the EPM results, which point out the (15,0) BNNT as a practically non-polar
nanomaterial. Data reported in Table 1 clearly indicate that the binding mechanism involved in this
adduct is mainly based on dispersive forces (calculated ¨Eads is practically equal to ¨ED; i.e., í10.2
and í11.7 kcal molí1, respectively) dictated by S-stacking interactions between the S systems of the
COOH group and the B-N hexagon rings of the (15,0) BNNT. Because of the presence of only
non-covalent interactions, ¨Eads is less negative compared to Gly interaction on (4,0) and (6,0)
BNNTs. It is worth mentioning that the very same BN/S complexes have also been calculated for the
Gly/BNNT (4,0) and (6,0) systems, meaning that, in the former case, the structure collapses onto the
BN/NH2 complex, whereas for the latter case the calculated ¨Eads is found to be 9.0 kcal molí1 above
the most stable one, due to the lower propensity of this BNNT to establish S-stacking interactions.

118
Table 1. Calculated adsorption energies (¨Eads), including the pure electronic energy
contribution (¨Eel) and the contribution of dispersion (¨ED). The relative electronic energies
(¨Erel) for a given Gly/BNNT system are also included. Values in units of kcal molí1.
System
Gly/BNNT(4,0)

Gly/BNNT(6,0)

Gly/BNNT(9,0)

Gly/BNNT(15,0)

Adduct
BN/NH2
BN/NH2-OH
BN/S
BN/zwitt
BN/COOH-1
BN/COOH-2
BN/NH2
BN/NH2–OH
BN/S
BN/zwitt
BN/COOH-1
BN/COOH-2
BN/NH2
BN/NH2–OH
BN/S
BN/zwitt
BN/COOH-1
BN/COOH-2
BN/NH2
BN/NH2–OH
BN/S
BN/zwitt
BN/COOH-1
BN/COOH-2

¨Eel
í20.8
í22.7
í20.7
í24.6
í9.8
í6.3
í6.8
+0.5
í1.8
í3.3
í3.3
+1.5
+4.3
+1.0
+14.0
í2.6
í2.2
+7.2
+9.0
+1.5
+20.7
í1.4
í1.4

¨ED
í11.2
í10.5
í10.4
í7.1
í7.8
í12.6
í11.4
í10.4
í11.7
í6.3
í6.2
í13.5
í12.0
í11.1
í12.5
í6.0
í6.1
í14.3
í12.6
í11.7
í13.1
í6.1
í6.2

¨Eads
í31.9
í33.2
í31.2
í31.7
í17.6
í18.9
í18.2
í9.9
í13.5
í9.6
í9.5
í12.0
í7.7
í10.1
+1.5
í8.6
í8.3
í7.1
í3.6
í10.2
+7.6
í7.5
í7.6

¨Erel
1.3
0.0
2.0
2.2
15.6
0.0
0.6
9.0
5.4
8.0
8.1
0.0
4.3
1.9
13.5
3.4
3.7
3.1
6.6
0.0
17.8
2.7
2.6

Table 2. Muliken charge (Q) of Gly adsorbed on the BNNTs in the most stable adducts
and respective calculated direct band gaps (Eg).
System
Gly/BNNT(4,0)
Gly/BNNT(6,0)
Gly/BNNT(9,0)
Gly/BNNT(15,0)

Adduct
BN/NH2
BN/NH2-OH
BN/NH2
BN/NH2-OH
BN/NH2
BN/S
BN/S

Q (e)
0.30
0.22
0.23
0.16
0.19
í0.02
í0.02

Eg (eV) a
3.68
3.69
4.46
4.45
5.29
5.42
6.06

a

calculated direct band gaps for the pristine BNNTs: 3.67, 4.42, 5.42 and 5.99 eV for (4,0), (6,0), (9,0)
and (15,0), respectively.

The interaction of Gly with the (9,0) BNNT is a frontier case between small radius (i.e., (4,0) and
(6,0)) and large radius (i.e., (15,0)) BNNTs. Although the BN/NH2 adduct has been found to be the
most stable one, the BN/S complex is the second most stable one lying 1.9 kcal molí1 above.
Calculated ¨Eads values, however, can suffer from the basis set superposition error (BSSE). Indeed,

119
upon correction, results indicate that these two complexes are nearly degenerate (BSSE-corrected
¨Eads values being í9.4 and í8.9 kcal molí1 for the BN/NH2 and BN/S, respectively).
The interaction of Gly through the COOH group by means of a simultaneous CO–B dative bond
and a OH···N (BNNT) H-bond has also been considered (see BN/COOH-1 and BN/COOH-2 adducts).
Although none of the calculated complexes are the most stable ones, important structural and energetic
features deserve to be mentioned. For the BN/COOH-1 adduct, Gly adsorption on the (4,0) BNNT
results in a spontaneous proton transfer from the Gly COOH group to the N nanotube atom, hence
forming a COO-/BNNT-H+ ion pair (see Figure 1b, structure of right). Such a proton transfer was
already observed for the adsorption of HCOOH on the very same (4,0) BNNT and is attributed to the
net charge transfer occurring from Gly to the BNNT, which induces an increase of the COOH acidity
and the nanotube basicity up to the point of promoting the proton transfer to a nearby N atom of the
nanotube. Moreover, for this adduct the CO–B distance is significantly shorter than for the other
BN/COOH-1 adducts (1.487 Å versus 2.514–2.921 Å, respectively), which results in a stronger
interaction (¨Eads = í31.7 kcal molí1 and | í9.6–í7.5 kcal molí1, respectively). For the BN/COOH-2
complex on the (4,0) BNNT, no proton transfer has been found; although the OH···N (BNNT) H-bond
and the CO–B dative bond are actually shorter than those present in the other BNNTs, in line with what
has been described for the BN/COOH-1 cases. Interestingly, the difference between BN/COOH-1 and
BN/COOH-2 is that in the former the COOH interaction occurs on B and N atoms that are chemically
bonded to each other, whereas in the latter this is not the case. Accordingly, the fact that the
spontaneous proton transfer only occurs in the former system seems to indicate that the charge transfer
in enhanced by a cooperative effect between the OH···N (BNNT) and the CO–B interactions when the B
and the N atoms are chemically bonded, which is in agreement with the larger and more negative
calculated ¨Eads values (í31.7 and í17.6 kcal molí1 for BN/COOH-1 and BN/COOH-2, respectively).
Finally, it is worth mentioning that the interaction of Gly in its zwitterionic state has also been
computed (BN/zwitt). On the (4,0) and (6,0) BNNTs, a spontaneous proton transfer from the NH3+
group to the N nanotube atom has been found, whereas for the (9,0) and (15,0) BNNTs the zwitterionic
form is maintained. Consistently, calculated ¨Eads values are negative for the two former adducts
(í31.2 and í13.5 kcal molí1, respectively), whereas for the two latter ones they have been found to
be positive (+1.5 and +7.6 kcal molí1, respectively) and, accordingly, are not stable complexes.
2.2. Microsolvated Interaction
Here, results on the interaction of Gly with the BNNTs in the presence of seven water molecules
are reported. We have introduced seven water molecules since this is the minimum number of water
molecules to have a relatively complete first-solvation shell of Gly upon adsorption; i.e., three water
molecules interacting with the NH3+ group, two water molecules interacting with the COOí group and
two more water molecules to complete the solvation shell. For these cases, the unit cell parameters of
the BNNTs have been enlarged thrice to avoid lateral interactions between water molecules of
adjacent unit cells. It is worth mentioning that a statistical sampling of the hypersurface of these
systems can be carried out adopting either the molecular dynamics or the Monte Carlo
approaches [41]. However, these calculations are extremely expensive at the ab-initio level adopting
realistic models for BNNTs like the (9,0) and the (15,0) ones, which contain 129 and 201 atoms,

120
respectively. For the present work, we have followed a different approach consisting of a progressive
microsolvation of the dry interface at the Gly/BNNT structures. This microsolvation procedure
consists of adding water molecules at the dry Gly/BNNT interface in such a way that Gly
progressively loses direct contact with the BNNTs up to a point in which the interaction is fully
bridged by water. This procedure has already been performed by some of us in other works for the
interaction of Gly with silica [42] and hydroxyapatite [43] surfaces. Since the most stable state of Gly
in water is the zwitterionic one, we considered the BN/zwitt adducts as initial guesses for the
progressive microsolvation. For the (4,0), (6,0) and (9,0) BNNTs, the resulting structures are shown
in Figure 2a. In the BN/CONH adduct, Gly directly interacts with the BNNTs in a similar fashion as
in the gas-phase (i.e., COOí···B BNNT dative bond and NH3+···N BNNT H-bond), while the seven
microsolvating water molecules are simple spectators interacting with available points of the COOí
and NH3+ groups through H-bonding. It is worth remarking that now for the (4,0) and (6,0) BNNTs
no H transfer from Gly to the BNNT occurs (at variance with the gas-phase adsorption, vide supra)
due to the screening effect of water. The BN/CO-w/NH and BN/NH-w/CO adducts result from
moving one spectator water molecule from the outer shell to the inner shell so that the following water
mediated interactions NH3+···H 2O···N BNNT and COOí···H 2O···B BNNT occur respectively. Finally,
From these two adducts, a second water displacement to remove the remaining direct Gly/BNNT
interaction gives the w/CONH adduct, in which water fully mediates the Gly/BNNT contact. It is
worth mentioning that each of the seven H2O molecules can in principle be displaced from their
positions to lead to a water mediated contact between Gly and BNNT. We choose the one exhibiting
the weakest interaction energy with the other water molecules by computing the cost to remove one
water molecule from the BN/CONH adduct by a single point energy evaluation for each H2O.
For the (15,0) BNNT, all the optimization calculations collapsed to structures with no direct contact
between Gly and BNNT, the most stable one being presented in Figure 2b. In this structure, water
fully solvates the Gly molecule and, at variance with the other BNNT cases, no charge transfer
between water and BNNT takes place, due to the highly apolar character of this nanotube.
The relative stabilities between the different calculated adducts for a given microsolvated complex
are shown in Table 3. As one can observe, for the (6,0), (9,0) and (15,0) BNNTs the most stable
systems are the w/CONH adducts; i.e., those in which no direct Gly/BNNT contact occurs, whereas
for the (4,0) one, the BN/CO-w/NH adduct (direct interaction only through the COOí) was found to
be the most energetically stable one. It is worth mentioning, however, that the energy difference
between the BN-CO-w/NH and w/CONH adducts for the (4,0) BNNT (the first and second most
stable ones) is relatively small (2.6 kcal molí1) and, accordingly, it might be inverted due to entropic
effects associated with water rearrangement, as it is shown for peptide adsorption on hydrophobic
and polar surfaces [44]. To further analyze this point, finite temperature molecular dynamics
simulations would be desirable.

121
Figure 2. B3LYP-D2* optimized structures of the different calculated complexes for
glycine interacting with the (4,0), (6,0) and (9,0) BNNTs (a) and with the (15,0) BNNT
(b) in the presence of seven water molecules. Distances in Å: bare values for the (4,0)
BNNT; values in parenthesis for the (6,0) one; values in brackets for the (9,0) one; and
italic underscored values for the (15,0) one. For this latter case, the distance is that
between the C atom and the plane defined by the closest B-N hexagon ring.

Besides these results, three different processes have moreover been considered to study the stability
of the structures shown in Figure 2. The first one involves Gly solvated by 7 H2O molecules being
adsorbed on the clean walls of the BNNTs, whose reaction energy was computed as (reported by ¨ER1
of Table 2)
Gly/7w + BNNT ĺ Gly/7w/BNNT

(2)

where Gly/7w is glycine solvated by the seven H2O molecules and Gly/7w/BNNT represents the
microsolvated complexes. The ¨ER1 column shows that for all BNNTs the process is exoenergetic,
meaning that the structural rearrangement of the seven H2O molecules around Gly is compensated
by the interaction with the BNNTs. Remarkably, limited to the most favorable adducts per BNNT,

122
¨ER1 values are less negative with increases in the nanotube radius, consistent with the less polar
behavior of the BNNTs.
Table 3. Reaction energies (¨ER1, ¨ER2 and ¨ER3) and relative energies (¨Erel) of the
formation of the Gly/7w/BNNT complexes. Values in units of kcal molí1.
System
Gly/7w/BNNT(4,0)

Gly/7w/BNNT(6,0)

Gly/7w/BNNT(9,0)

Gly/7w/BNNT(15,0)

Adduct
BN/CONH
BN/CO-w/NH
BN/NH-w/CO
w/CONH
BN/CONH
BN/CO-w/NH
BN/NH-w/CO
w/CONH
BN/CONH
BN/CO-w/NH
BN/NH-w/CO
w/CONH
BN/CONH
BN/CO-w/NH
BN/NH-w/CO
w/CONH

¨Erel
5.4
0.0
4.3
2.6
5.3
2.9
4.2
0.0
12.4
7.7
10.5
0.0
0.0

¨ER1
í28.9
í34.3
í30.0
í31.7
í13.2
í15.7
í14.4
í18.6
í5.1
í9.8
í7.0
í17.5
í12.8

¨ER2
í11.5
í16.9
í12.6
í14.3
í20.5
í23.0
í21.7
í25.9
í20.7
í25.4
í22.6
í33.1
í31.4

¨ER3
11.9
6.5
10.8
9.1
2.9
0.4
1.7
í2.5
2.7
í2.0
0.8
í9.7
í8.1

The second process envisages gas phase Gly adsorbed on the already microsolvated BNNTs by the
seven water molecules, whereas the third one envisages the solvated Gly being adsorbed on the seven
water hydrated BNNTs giving rise to the shown adducts with expulsion of seven water molecules
(here considered as a H-bonded cluster), whose reaction energies were computed as (reported by ¨ER2
and ¨ER3 of Table 2)
Gly + 7w/BNNT ĺ Gly/7w/BNNT

(3)

Gly/7w + 7w/BNNT ĺ Gly/7w/BNNT + 7w

(4)

where 7w/BNNT is the BNNT solvated by the seven H2O molecules and 7w the H-bonded cluster
made up by seven water molecules. These last two processes require the BNNTs in the presence of
seven water molecules, whose optimized structures are given in Figure 3. The initial guess of these
systems were the corresponding w/CONH adducts (no direct interaction between Gly and the BNNTs
and accordingly the interaction between water molecules and BNNTs is maximum) in which the Gly
molecule was removed. Consistent with the polar character of the BNNTs, several interactions
between the water molecules and the (4,0) BNNT via covalent dative and H-bond interactions take
place, whereas by increasing the BNNT radius these interactions are progressively missed, up to the
point in which for the (15,0) no apparent interaction is observed.

123
Figure 3. B3LYP-D2* optimized structures of the different calculated complexes for the
considered BNNTs in the presence of seven water molecules. Distances in Å: bare values
for the (4,0) BNNT; values in parenthesis for the (6,0) one; values in brackets for the
(9,0) one; and italic underscored values for the (15,0) one. For this latter case, the distance
is that between the closest water O to the plane defined by the B-N hexagon ring.

The energetics of R2 is a tradeoff between the water affinity of the BNNTs and Gly (R2 is in
essence the capture of the adsorbed water by Gly). The trend provided by the calculated ¨ER2 values
indicates that R2 is more favorable when increasing the radius. That is, since the (15,0) BNNT does
not exhibit water affinity, the calculated ¨ER2 values are large and negative due to the strong
interaction of water with Gly. In contrast, both the (4,0) BNNT and Gly exhibit large water affinity and
accordingly the calculated ¨ER2 values are the less negative ones of the series. R3 is probably the most
physically sound process as it involves the replacement of adsorbed water solvent by an already solvated
Gly. Calculated ¨ER3 values indicate that such a water replacement is energetically favorable on the
(15,0) and (9,0) BNNTs probably due to the low water affinity of these BNNTs. On the (6,0) the process
is still favorable by some amount but less than on the other two nanotubes (¨ER3 = í2.5 kcal molí1),
whereas on the (4,0) calculated ¨ER1 was found to be positive, indicating that the overall interactions
between Gly, water solvent and the nanotube are not as stable as the interaction between this nanotube
and water.
3. Computational Details
All calculations were carried out using the periodic ab-initio code CRYSTAL09 [45]. All the SCF
calculations and geometry optimizations were performed using the B3LYP-D* density functional
method, which includes an empirical a posteriori correction term proposed by Grimme [46] to account
for dispersion forces (missed in the pure B3LYP [47,48] method), but whose initial parameterization
(D) was modified for extended systems (D*) [49], to provide accurate results for the calculations of

124
cohesive energies of molecular crystals and of adsorption processes within a periodic treatment [49–51].
The adopted Gaussian functions consisted of an all electron triple-ȗ 6-311G* standard basis set for
the B and N atoms of the BNNTs and a TZP basis set from Ahlrichs and coworkers [52] for the atoms
of Gly. This basis set combination has been proved to exhibit small basis set superposition error
interaction energies [18,50]. The shrinking factor of the reciprocal space net defining the mesh of k
points in the irreducible Brillouin zone was set to 5, which requires diagonalizing the Hamiltonian
matrix in 3 k points [53]. The accuracy of both Coulomb and exchange series was set to values of
overlap integrals of 10í7 and 10í16, respectively, which ensure a very good numerical accuracy. A
pruned (75, 974) grid has been used for the Gauss–Legendre and Lebedev quadrature schemes in the
evaluation of functionals. The condition to achieve SCF convergence between two subsequent cycles
was set to 10í7 Eh. Full relaxations of both lattice parameters and internal atomic coordinates by means
of analytical energy gradients [54–56] have been carried out. The geometry optimization was
performed by means of a quasi-Newton algorithm in which the quadratic step (BFGS Hessian
updating scheme) is combined with a linear one (parabolic fit) [57].
4. Conclusions
Periodic quantum mechanical calculations have been used to simulate the adsorption of glycine
(Gly) on different zig-zag (n,0) single-walled boron-nitride nanotubes (BNNTs, n = 4, 6, 9 and 15)
both in the gas-phase and in a microsolvated state (i.e., modeled by the presence of seven explicit
water molecules) with the aim of determining the adsorption properties and the effect exerted by water
as a function of surface curvature. These calculations are based on the B3LYP-D2* method, which
includes the B3LYP hybrid functional with a revised version of the empirical a posteriori correction
term (D2*) to account for dispersion interactions.
Gas-phase results clearly indicate that the most stable interaction between Gly and the (4,0), (6,0)
and (9,0) BNNTs takes place through a covalent dative interaction between the NH2 group of Gly and
the B atom of the BNNTs, which produce charge transfers from Gly to the BNNTs. In contrast, the
interaction between Gly and the (15,0) BNNT is mainly governed by non-covalent dispersive forces
based on a S-stacking between the S systems of the Gly COOH group and the B-N hexagon rings of
the nanotube. Remarkably, the energy difference between these two adducts decreases when
increasing the BNNT radius, in line with the polar/apolar character of the considered nanotubes.
The adsorption of Gly on the BNNTs in the presence of seven water molecules has been studied
adopting a progressive microsolvation procedure, in which water solvent molecules are added at the
dry Gly/BNNT interfaces, hence progressively removing the direct interaction between Gly and the
BNNTs. The obtained results indicate that for the (6,0), (9,0) and (15,0) BNNTs, the most stable
microsolvated systems were found to exhibit no direct contact between the two partners; that is, the
Gly/BNNT interfaces are fully bridged by the water solvent molecules. In contrast, for the (4,0) one
of the most stable systems shows direct contact between Gly and the BNNT through an interaction
between the Gly COOí group and a nanotube B atom, although entropic effects (not accounted for in
this work) might favor water mediated interface. Further energetic results point out that the larger the
BNNT radius, the less water affinity. Accordingly, for larger radius BNNTs, the interaction between
water and Gly was found to be predominant, in detriment to their interaction with the BNNT.

125
However, it is found that the (4,0) BNNT exhibits a large water affinity, which is reflected by the fact
that the replacement of seven adsorbed water molecules by a microsolvated Gly has been found to be
an unfavorable process.
The results presented here provide evidence that the adsorption properties of the BNNTs
as well as their water affinity can significantly be modulated by controlling the tube diameter,
as they are expected to exhibit different physico-chemical features, which may be of interest for the
design of bioconjugated systems based on boron-nitride nanostructures and their potential
bionanotechnological applications.
Acknowledgments
Albert Rimola is indebted to Ministry of Economy (MINECO) of the Spanish Government for a
Juan de la Cierva contract. Financial support from MINECO (projects CTQ2013-40347-ERC and
CTQ2011-24847/BQU) and from Department of Economy of the Catalan Government (project
2009SGR-638), and the use of the Catalonia Supercomputer Centre (CESCA) are gratefully
acknowledged. Mariona Sodupe also acknowledges support through the 2011 ICREA Academia award.
Author Contributions
Simulations were carried out by Albert Rimola. Both Albert Rimola and Mariona Sodupe
contributed to the discussion and writing of the paper.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.
2.

3.
4.

5.

6.

Kujawa, P.; Winnik, F.M. Innovation in Nanomedicine through Materials Nanoarchitectonics.
Langmuir 2013, 29, 7354–7361.
Wong, S.S.; Joselevich, E.; Woolley, A.T.; Cheung, C.L.; Lieber, C.M. Covalently
functionalized nanotubes as nanometre- sized probes in chemistry and biology. Nature 1998,
394, 52–55.
Aono, M.; Bando, Y.; Ariga, K. Nanoarchitectonics: Pioneering a New Paradigm for
Nanotechnology in Materials Development. Adv. Mater. 2012, 24, 150–151.
Hong, R.; Fischer, N.O.; Verma, A.; Goodman, C.M.; Emrick, T.; Rotello, V.M. Control of
Protein Structure and Function through Surface Recognition by Tailored Nanoparticle
Scaffolds. J. Am. Chem. Soc. 2004, 126, 739–743.
You, C.-C.; Agasti, S.S.; De, M.; Knapp, M.J.; Rotello, V.M. Modulation of the Catalytic
Behavior of Į-Chymotrypsin at Monolayer-Protected Nanoparticle Surfaces. J. Am. Chem. Soc.
2006, 128, 14612–14618.
Kang, Y.; Liu, Y.-C.; Wang, Q.; Shen, J.-W.; Wu, T.; Guan, W.-J. On the spontaneous
encapsulation of proteins in carbon nanotubes. Biomaterials 2009, 30, 2807–2815.

126
7.
8.
9.

10.
11.
12.
13.

14.

15.
16.
17.
18.
19.
20.

21.
22.
23.
24.

Gao, H.; Kong, Y. Simulation of DNA-nanotube interactions. Annu. Rev. Mater. Res. 2004, 34,
123–150.
Chen, X.; Kis, A.; Zettl, A.; Bertozzi, C.R. A cell nanoinjector based on carbon nanotubes.
Proc. Natl. Acad. Sci. USA 2007, 104, 8218–8222.
Chen, R.J.; Bangsaruntip, S.; Drouvalakis, K.A.; Wong Shi Kam, N.; Shim, M.; Li, Y.; Kim, W.;
Utz, P.J.; Dai, H. Noncovalent functionalization of carbon nanotubes for highly specific
electronic biosensors. Proc. Natl. Acad. Sci. USA 2003, 100, 4984–4989.
Chen, R.J.; Zhang, Y.; Wang, D.; Dai, H. Noncovalent Sidewall Functionalization of Single-Walled
Carbon Nanotubes for Protein Immobilization. J. Am. Chem. Soc. 2001, 123, 3838–3839.
Cui, Y.; Wei, Q.; Park, H.; Lieber, C.M. Nanowire Nanosensors for Highly Sensitive and
Selective Detection of Biological and Chemical Species. Science 2001, 293, 1289–1292.
Staii, C.; Johnson, A.T.; Chen, M.; Gelperin, A. DNA-Decorated Carbon Nanotubes for
Chemical Sensing. Nano Lett. 2005, 5, 1774–1778.
Heller, D.A.; Jeng, E.S.; Yeung, T.-K.; Martinez, B.M.; Moll, A.E.; Gastala, J.B.; Strano, M.S.
Optical Detection of DNA Conformational Polymorphism on Single-Walled Carbon Nanotubes.
Science 2006, 311, 508–511.
Pantarotto, D.; Partidos, C.D.; Graff, R.; Hoebeke, J.; Briand, J.-P.; Prato, M.; Bianco, A.
Synthesis, Structural Characterization, and Immunological Properties of Carbon Nanotubes
Functionalized with Peptides. J. Am. Chem. Soc. 2003, 125, 6160–6164.
Tu, X.; Manohar, S.; Jagota, A.; Zheng, M. DNA sequence motifs for structure-specific
recognition and separation of carbon nanotubes. Nature 2009, 460, 250–253.
Golberg, D.; Bando, Y.; Huang, Y.; Terao, T.; Mitome, M.; Tang, C.; Zhi, C. Boron Nitride
Nanotubes and Nanosheets. ACS Nano 2010, 4, 2979–2993.
Golberg, D.; Bando, Y.; Tang, C.C.; Zhi, C.Y. Boron Nitride Nanotubes. Adv. Mater. 2007, 19,
2413–2432.
Rimola, A.; Sodupe, M. Physisorption vs. chemisorption of probe molecules on boron nitride
nanomaterials: The effect of surface curvature. Phys. Chem. Chem. Phys. 2013, 15, 13190–13198.
Magrez, A.; Kasas, S.; Salicio, V.; Pasquier, N.; Seo, J.W.; Celio, M.; Catsicas, S.; Schwaller, B.;
Forró, L. Cellular Toxicity of Carbon-Based Nanomaterials. Nano Lett. 2006, 6, 1121–1125.
Chen, X.; Wu, P.; Rousseas, M.; Okawa, D.; Gartner, Z.; Zettl, A.; Bertozzi, C.R. Boron Nitride
Nanotubes Are Noncytotoxic and Can Be Functionalized for Interaction with Proteins and Cells.
J. Am. Chem. Soc. 2009, 131, 890–891.
Yu, J.; Chen, Y.; Cheng, B.M. Dispersion of boron nitride nanotubes in aqueous solution with
the help of ionic surfactants. Solid State Commun. 2009, 149, 763–766.
Zhi, C.; Bando, Y.; Tang, C.; Golberg, D. Immobilization of Proteins on Boron Nitride
Nanotubes. J. Am. Chem. Soc. 2005, 127, 17144–17145.
Gao, Z.; Zhi, C.; Bando, Y.; Golberg, D.; Serizawa, T. Isolation of Individual Boron Nitride
Nanotubes via Peptide Wrapping. J. Am. Chem. Soc. 2010, 132, 4976–4977.
Zhi, C.; Bando, Y.; Wang, W.; Tang, C.; Kuwahara, H.; Golberg, D. DNA-Mediated Assembly
of Boron Nitride Nanotubes. Chem. Asian J. 2007, 2, 1581–1585.

127
25. Ciofani, G.; Raffa, V.; Menciassi, A.; Dario, P. Preparation of Boron Nitride Nanotubes Aqueous
Dispersions for Biological Applications. J. Nanosci. Nanotechnol. 2008, 8, 6223–6231.
26. Ciofani, G.; Raffa, V.; Menciassi, A.; Cuschieri, A. Cytocompatibility, interactions, and uptake
of polyethyleneimine-coated boron nitride nanotubes by living cells: Confirmation of their
potential for biomedical applications. Biotechnol. Bioeng. 2008, 101, 850–858.
27. Ciofani, G.; Danti, S.; D’Alessandro, D.; Moscato, S.; Menciassi, A. Assessing cytotoxicity of
boron nitride nanotubes: Interference with the MTT assay. Biochem. Biophys. Res. Commun.
2010, 394, 405–411.
28. Ciofani, G.; Danti, S.; D’Alessandro, D.; Ricotti, L.; Moscato, S.; Bertoni, G.; Falqui, A.;
Berrettini, S.; Petrini, M.; Mattoli, V.; et al. Enhancement of Neurite Outgrowth in Neuronal-Like
Cells following Boron Nitride Nanotube-Mediated Stimulation. ACS Nano 2010, 4, 6267–6277.
29. Horváth, L.; Magrez, A.; Golberg, D.; Zhi, C.; Bando, Y.; Smajda, R.; Horváth, E.; Forró, L.;
Schwaller, B. In Vitro Investigation of the Cellular Toxicity of Boron Nitride Nanotubes.
ACS Nano 2011, 5, 3800–3810.
30. Akdim, B.; Kim, S.N.; Naik, R.R.; Maruyama, B.; Pender, M.J.; Pachter, R. Understanding
effects of molecular adsorption at a single-wall boron nitride nanotube interface from density
functional theory calculations. Nanotechnology 2009, 20, 355705.
31. Saikat, M.; Gowtham, S.; Ralph, H.S.; Ravindra, P.; Shashi, P.K. Theoretical study of physisorption
of nucleobases on boron nitride nanotubes: A new class of hybrid nano-biomaterials.
Nanotechnology 2010, 21, 165703.
32. Lin, Q.; Zou, X.; Zhou, G.; Liu, R.; Wu, J.; Li, J.; Duan, W. Adsorption of DNA/RNA
nucleobases on hexagonal boron nitride sheet: An ab initio study. Phys. Chem. Chem. Phys.
2011, 13, 12225–12230.
33. Wu, X.; An, W.; Zeng, X.C. Chemical Functionalization of BoroníNitride Nanotubes with NH3
and Amino Functional Groups. J. Am. Chem. Soc. 2006, 128, 12001–12006.
34. Mukhopadhyay, S.; Scheicher, R.H.; Pandey, R.; Karna, S.P. Sensitivity of Boron Nitride
Nanotubes toward Biomolecules of Different Polarities. J. Phys. Chem. Lett. 2011, 2, 2442–2447.
35. Yang, C.-K. Exploring the interaction between the boron nitride nanotube and biological
molecules. Comput. Phys. Commun. 2011, 182, 39–42.
36. Noel, Y.; D’Arco, P.; Demichelis, R.; Zicovich-Wilson, C.M.; Dovesi, R. On the Use of
Symmetry in the Ab Initio Quantum Mechanical Simulation of Nanotubes and Related
Materials. J. Comput. Chem. 2010, 31, 855–862.
37. Demichelis, R.; Noel, Y.; D’Arco, P.; Rerat, M.; Zicovich-Wilson, C.M.; Dovesi, R. Properties
of Carbon Nanotubes: An ab Initio Study Using Large Gaussian Basis Sets and Various DFT
Functionals. J. Phys. Chem. C 2011, 115, 8876–8885.
38. D’Arco, P.; Noel, Y.; Demichelis, R.; Dovesi, R. Single-layered chrysotile nanotubes:
A quantum mechanical ab initio simulation. J. Chem. Phys. 2009, 131, doi:10.1063/1.3251791.
39. Demichelis, R.; Noel, Y.; D’Arco, P.; Maschio, L.; Orlando, R.; Dovesi, R. Structure and
energetics of imogolite: A quantum mechanical ab initio study with B3LYP hybrid functional.
J. Mater. Chem. 2010, 20, 10417–10425.

128
40. Ferrari, A.M.; Szieberth, D.; Zicovich-Wilson, C.M.; Demichelis, R. Anatase(001) 3 ML
Nanotubes, The First TiO2 Nanotube With Negative Strain Energies: A DFT Prediction. J. Phys.
Chem. Lett. 2010, 1, 2854–2857.
41. Marx, D.; Hutter, J. Ab Initio Molcular Dynamics: Theory and Implementation. In Modern
Methods and Algorithms of Quantum Chemistry; Grotendorst, J., Ed.; Forschungszentrum
Jülich: Jülich, Germany, 2000.
42. Rimola, A.; Civalleri, B.; Ugliengo, P. Neutral vs. Zwitterionic Glycine Forms at the
Water/Silica Interface: Structure, Energies, and Vibrational Features from B3LYP Periodic
Simulations. Langmuir 2008, 24, 14027–14034.
43. Rimola, A.; Corno, M.; Zicovich-Wilson, C.M.; Ugliengo, P. Ab initio modeling of
protein/biomaterial interactions: Competitive adsorption between glycine and water onto
hydroxyapatite surfaces. Phys. Chem. Chem. Phys. 2009, 11, 9005–9007.
44. Schwierz, N.; Horinek, D.; Liese, S.; Pirzer, T.; Balzer, B.N.; Hugel, T.; Netz, R.R.
On the Relationship between Peptide Adsorption Resistance and Surface Contact Angle: A
Combined Experimental and Simulation Single-Molecule Study. J. Am. Chem. Soc. 2012, 134,
19628–19638.
45. Dovesi, R.; Saunders, V.R.; Roetti, C.; Orlando, R.; Zicovich-Wilson, C.M.; Pascale, F.;
Civalleri, B.; Doll, K.; Harrison, N.M.; Bush, I.J.; et al. CRYSTAL09 User’s Manual; University
of Torino: Torino, Italy, 2009; p. 307.
46. Grimme, S. Semiempirical GGA-type density functional constructed with a long-range
dispersion correction. J. Comput. Chem. 2006, 27, 1787–1799.
47. Becke, A.D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys.
1993, 98, 5648–5652.
48. Lee, C.; Yang, W.; Parr, R.G. Development of the Colle-Salvetti correlation-energy formula
into a functional of the electron density. Phys. Rev. B 1988, 37, 785–789.
49. Civalleri, B.; Zicovich-Wilson, C.M.; Valenzano, L.; Ugliengo, P. B3LYP augmented with an
empirical dispersion term (B3LYP-D*) as applied to molecular crystals. CrystEngComm 2008,
10, 405–410.
50. Rimola, A.; Civalleri, B.; Ugliengo, P. Physisorption of aromatic organic contaminants at
the surface of hydrophobic/hydrophilic silica geosorbents: A B3LYP-D modeling study.
Phys. Chem. Chem. Phys. 2010, 12, 6357–6366.
51. Civalleri, B.; Maschio, L.; Ugliengo, P.; Zicovich-Wilson, C.M. Role of dispersive interactions
in the CO adsorption on MgO (001): Periodic B3LYP calculations augmented with an empirical
dispersion term. Phys. Chem. Chem. Phys. 2010, 12, 6382–6386.
52. Schafer, A.; Horn, H.; Ahlrichs, R. Fully optimized contracted Gaussian basis sets for atoms Li
to Kr. J. Chem. Phys. 1992, 97, 2571–2577.
53. Monkhorst, H.J.; Pack, J.D. Special points for Brillouin-zone integrations. Phys. Rev. B 1976,
13, 5188–5192.
54. Doll, K.; Saunders, V.R.; Harrison, N.M. Analytical Hartree–Fock gradients for periodic
systems. Int. J. Quantum Chem. 2001, 82, 1–13.

129
55. Doll, K. Implementation of analytical Hartree–Fock gradients for periodic systems.
Comput. Phys. Commun. 2001, 137, 74–88.
56. Doll, K.; Dovesi, R.; Orlando, R. Analytical Hartree–Fock gradients with respect to the cell
parameter for systems periodic in three dimensions. Theor. Chem. Acc. 2004, 112, 394–402.
57. Civalleri, B.; D’Arco, P.; Orlando, R.; Saunders, V.R.; Dovesi, R. Hartree-Fock geometry
optimisation of periodic systems with the CRYSTAL code. Chem. Phys. Lett. 2001, 348,
131–138.

130

The Role of Lead (Pb) in the High Temperature Formation of
MoS2 Nanotubes
Olga Brontvein, Reshef Tenne and Andrey Enyashin
Abstract: Recent studies have clearly indicated the favorable effect of lead as a growth promoter for
MX2 (M = Mo, W; X = S, Se) nanotubes using MX2 powder as a precursor material. The experimental
work indicated that the lead atoms are not stable in the molybdenum oxide lattice ion high
concentration. The initial lead concentration in the oxide nanowhiskers (Pb:Mo ratio = 0.28) is
reduced by one order of magnitude after one year in the drawer. The initial Pb concentration in the
MoS2 nanotubes lattice (produced by solar ablation) is appreciably smaller (Pb:Mo ratio for the
primary samples is 0.12) and is further reduced with time and annealing at 810 °C, without
consuming the nanotubes. In order to elucidate the composition of these nanotubes in greater detail;
the Pb-“modified” MX2 compounds were studied by means of DFT calculations and additional
experimental work. The calculations indicate that Pb doping as well as Pb intercalation of MoS2 lead
to the destabilization of the system; and therefore a high Pb content within the MoS2 lattice cannot
be expected in the final products. Furthermore; substitutional doping (PbMo) leads to p-type
semiconducting character; while intercalation of MoS2 by Pb atoms (Pby/MoS2) should cause n-type
semiconducting behavior. This study not only sheds light on the role of added lead to the growth of
the nanotubes and their role as electron donors; but furthermore could pave the way to a large scale
synthesis of the MoS2 nanotubes.
Reprinted from Inorganics. Cite as: Brontvein, O.; Tenne, R.; Enyashin, A. The Role of Lead (Pb)
in the High Temperature Formation of MoS2 Nanotubes. Inorganics 2014, 2, 363–375.
1. Introduction
Inorganic nanotubes (INT) have become a prominent research subject in recent years. Their unique
mechanical, optical and electrical properties [1–8] prompted extensive investigation. A number of
synthetic routes have been developed, for example, chemical vapor transport using bromine for
INT-MoS2 [9,10], bismuth catalyzed vapor-liquid-solid method for SnS2 nanotubes [11], misfit
compounds superstructures of SnS-SnS2 and NbS-PbS2 nanotubes [12,13] and sulfidization of
tungsten suboxide nanowhiskers leading to the formation of INT-WS2 [14].
Recent studies have shown a new strategy for successful synthesis of MX2 (M = Mo, W; X = S, Se)
nanotubes using Pb as a growth promoter [15,16]. Figure 1 presents a SEM image of MoS2 nanotubes
(Figure 1a) and high resolution TEM image of single MoS2 nanotube (Figure 1b) formed by irradiating
MoS2 powder by a focused (×15,000) sunlight for 10 min [15]. The inset in Figure 1b shows the
distance profile of nanotube layers. The interlayer spacing of 0.634 nm is somewhat larger than that
of bulk 2H-MoS2 (0.62 nm). This 2% expansion of the lattice spacing in the nanotubes is well
documented [17] and is attributed to strain relaxation. Figure 1c shows the EDS spectrum of single
MoS2 nanotube after the irradiation, the table in the inset in Figure 1c shows the atomic % of the
elements. In both cases the formation mechanism can be described as a Pb-promoted MX2 conversion

131
into Mo(Pb)O3íx nanowhiskers at high temperature (>2500 °C). Once formed, the nanowhiskers react
back with the X-vapor which leads to the formation of MX2 nanotubes. Numerous attempts to
synthesize MoS2 nanotubes in a conventional oven (up to 1000 °C) or induction furnace (up to 1600 °C),
with the same precursor materials, or generation of Mo(Pb)O3íx nanowhiskers and their subsequent
sulfidization did not succeed. It appears, therefore, that the high temperatures (>2500 °C) and the
presence of lead are critical for the conversion of MoS2 powder into Mo(Pb)O3íx nanowhiskers which
serve as template for the synthesis of MX2 nanotubes according to this process. In addition, the use of
Pb as a growth promoter is environmentally unfavorable. Therefore, stabilization of molybdenum
suboxide phases could be done using different metals, for example V, Ta or W [18–20].
Figure 1. MoS2 nanotubes formed using solar ablation system (a) SEM image of MoS2
nanotubes; (b) high resolution TEM image of a single MoS2 nanotube (adapted after [15]).
The inset shows the profile of the interlayer distances of the nanotube wall. The interlayer
spacing is somewhat larger than that of bulk 2H-MoS2 (0.62 nm); (c) EDS spectrum of a
single MoS2 nanotube; the inset table shows atomic % of the nanotube’s elements.

132
The main goal of the current work is to understand the role of lead (Pb) as a growth promoter of
INT-MoS2; evaluate its solubility limits in the two lattices (oxide and sulfide) and their electronic state
in the INT-MoS2 lattice. In particular, the stability of this heavy metal in the oxide precursor and the
formed MoS2 nanotubes was investigated through both theory and experiment. Due to the fast kinetics
of the process, the samples (oxide nanowhiskers and nanotubes) formed in such extreme conditions
contain a certain amount of Pb. However, the content of Pb is found to be time-dependent and can
considerably decrease upon thermal annealing. DFT calculations give preliminary information about
the coordination and the electronic state of Pb atoms within the MoS2 lattice. They verify a low affinity
of the Pb atoms as dopants or intercalants in the sulfide matrices and indicate that the experimentally
observed high amount of lead (Pb) is too far from the thermodynamic equilibrium conditions.
2. Results and Discussion
2.1. Annealing
To distinguish between the different samples of oxide nanowhiskers and the molybdenum
disulfide nanotubes studied in this work a labeling system was used (see Table 1). For this work, two
different precursors were investigated: MoO3íx (2) nanowhiskers and MoS2 (2) nanotubes, which were
left one year in the drawer after solar ablation synthesis. The detailed description of their synthesis
can be seen in the experimental part. Due to the small quantities of the MoS2 nanotubes and MoO3íx
nanowhiskers that are obtained in the synthesis (~1%–2% yield) only single particle EDS
measurement can be suitable for their chemical analysis.
Table 1. The samples labeling that were used during the experimental work.
Precursor
MoO3íx
MoS2

After the Solar
Ablation Synthesis
MoO3íx (1)
MoS2 (1)

After a Year in
Ambient Conditions
MoO3íx (2)
MoS2 (2)

30 min
Annealing in H2S
MoS2 (3a)
MoS2 (4a)

1 h Annealing
in H2S
MoS2 (3b)
MoS2 (4b)

2 h Annealing
in H2S
MoS2 (3c)
MoS2 (4c)

The experimental work which has recently been carried-out indicated that the Pb atoms are not
stable in high concentrations in the MoO3íx or MoS2 lattice. According to the EDS measurements the
initial lead concentration in the molybdenum suboxide nanowhiskers MoO3íx (1) was reduced by one
order of magnitude after one year in the drawer (Pb:Mo ratio reduced from ~0.28 to 0.03 in MoO3íx(2))
(Figure 2). The same analysis showed that in the case of MoS2 (1) nanotubes the Pb concentration
decreased from ~0.12 to 0.03 in MoS2; (2). Subsequent annealing of both types of nanostructures, i.e.,
MoO3íx (2) nanowhiskers and MoS2 (2) nanotubes, leads to additional reduction of the Pb
concentration. The Pb:Mo ratio before and after annealing calculated from EDS measurements can be
seen in Figure 2. It should be emphasized that the accuracy of the EDS is greatly compromised at the
lower concentration limit of Pb.
The outdiffusion of the lead from the molybdenum oxide nanowhiskers MoO3íx (2) did influence
their high-temperature stability and their conversion into MoS2 (3) nanotubes upon annealing in H2S
atmosphere at 810 °C. In all cases, MoS2 (3,4) nanotubes were observed after the sulfidization and
annealing of the lead-depleted MoO3íx (2) nanowhiskers and MoS2 (2) nanotubes. TEM images of

133
the precursors and products can be seen in Figure 3. Detailed description of the annealing conditions
can be seen in the experimental section.
Figure 2. Pb:Mo ratio of the precursors and products according to the EDS
measurements, (a) MoO3íx (2) nanowhiskers as a precursor material; (b) MoS2 (2)
nanotubes as a precursor material.

Due to the overlap between the Mo(L) and Pb(M) peaks (2.3 KeV) in the EDS spectra,
the calculated atomic concentration of Pb is based on the Pb(L) (10.5 KeV) and Mo(K) (17.4 KeV)
peaks, which provide lower accuracy. Therefore, the detection of Pb atoms after a year in the drawer
is limited to the detected range (>0.2 at%). On the other hand, it can be seen from the EDS spectra of
the samples before and after annealing that in the case of molybdenum suboxide nanowhiskers MoO3íx
(2) as a precursor material, the Pb peak disappears after two hours annealing in H2S atmosphere, while
the same peak disappears after 30 min annealing at 810 °C in the case of MoS2 (2) nanotubes
(Figure 4).
Figure 3. TEM images, (a) MoO3íx (1) nanowhisker obtained after the synthesis
(solar ablation for 30 s); (b) MoO3íx (2) nanowhisker a year after the synthesis;
(c–e) MoS2 (3) nanotubes after 30 min, 1 and 2 h of H2S annealing, respectively, with
MoO3íx nanowhiskers as a precursor; (f) MoS2 (1) nanotubes after the synthesis (solar
ablation for 10 min); (g) MoS2 (2) nanotube a year after the synthesis, (h–j) MoS2 (4)
nanotubes after 30 min, 1 and 2 h (H2S) annealing, respectively, with MoS2 (2) nanotubes
as a precursor.

134
Figure 4. EDS spectra of the samples before and after annealing at 810 °C, (a) MoO3íx
(2) nanowhiskers as a precursor; (b) MoS2 (2) nanotubes as a precursor.

2.2. DFT Calculations of Mo1íxPbxS2 Solid Solutions and Pby/MoS2 Intercalates
The experimental work has indicated a time-dependent content of Pb in the samples of MoS2 nanotubes
after Pb-promoted synthesis, which suggests that the MoS2 lattice forms an unfavorable environment
for the Pb atoms in these nanostructures. In order to elucidate preliminary the accommodation type,
chemical bonding and the highest possible (at equilibrium) concentration of Pb atoms in the samples
of MoS2 nanotubes a set of DFT calculations were performed. Two cases were considered: (1) the

135
lead atoms substitute the Mo in the lattice and serve as dopants (Mo1íxPbxS2 solid solutions); (2) the
Pb atoms intercalate into the interlayer space of the MoS2 host lattice (Pby/MoS2 intercalates).
2.2.1. Electronic Structure
Test DFT calculations at LDA level were performed on example of the bulk 2H-MoS2 and fcc-PbS
crystals (see Section 3). 2H-MoS2 was found to be a semiconductor with direct band gap 2.18 eV
and indirect Ƚĺ½KȽ gap of 0.59 eV. The valence band is composed mainly of S3p-states,
about ~3.5–6.0 eV below the Fermi level. The top of the valence band and the bottom of the conduction
band are dominated by Mo4d-states. In the electronic structure of fcc-PbS the occupied Pb6s band is
present between 6.5 and 9.5 eV below the Fermi level. The three predominately S3p bands are located
between í6 and 0 eV. The fundamental band gap of 0.11 eV is direct at the L point with the conduction
band composed predominately of Pb6p states. While the values of the band gaps are also typically
underestimated in LDA approach, the band structures are in full accordance with other theoretical and
experimental results [21,22].
In general, the band structure of MoS2 and the corresponding picture of the densities of states
(DOS) are essentially not perturbed by substitutional doping of the Mo sublattice by a single Pb atom
(Figure 5a). A single level separates from the bottom of the conduction band into the band gap of
pristine MoS2, while the Fermi energy shifts downwards the top of valence band. In addition, a new
level emerges at 1 eV below the bottom of the valence band of doped MoS2. It is composed mainly
of Pb6s states, which do not overlap with the S3p band and do not contribute to the chemical bonding
of the system. Thus, like in the case of PbS compound, the chemical bonding of Pb atom with S atoms
is released due to the depopulation of only Pb6p states, which can be found at 5–6 eV above the Fermi
energy. All these features are evidence for a p-type semiconducting character of the Pb-doped MoS2
and nominal 2+ oxidation state of Pb. Such oxidation state and the arising electron deficiency of MoS2
electronic structure should lead to destabilization of the chemical environment of the Pb-doped MoS2
lattice. The map of the differential electron densities for this system supports this deduction and
demonstrates a vanishing electron density and weakening of the bonds near the Pb atom compared to
the electron density nearby the Mo atoms (Figure 5a).
Another possible process, which could occur during the Pb-promoted growth of MoS2 nanotubes,
is an intercalation of MoS2 lattice by Pb atoms. The intercalation by a single Pb atom also does not
affect the band structure essentially (Figure 5b). Yet, in this case the Fermi level shifts upwards
into the conduction band. The localized Pb6s states occur deeper at 2 eV below the bottom of valence
band. Pb6p states of the intercalating Pb atom are more delocalized within the conduction band, than
in the aforementioned case of Pb doping. Thus, a Pby/MoS2 intercalate should behave as an n-type
semiconductor. The map of differential electron density for this system does not reveal any essential
covalent bond formation with S atoms and serves as an evidence for the non-amicable environment
of the intercalating Pb atoms.

136
Figure 5. (a) Band structure; electronic partial densities of states (DOS) for Pb and
differential electron density map for 2H-MoS2 crystal doped by single Pb atom; (b) Band
structure, DOS for Pb and differential electron density map for 2H-MoS2 crystal
intercalated by single Pb atom. LDA DFT calculations.

2.2.2. Estimation of the Stability Limit for Lead Atoms in the MoS2 Lattices
Both doping and intercalation by single Pb atoms lead to a destabilization of the MoS2 electronic
system. A weakening of the chemical bonding within the MoS2 lattice due to electron deficiency of
the valence band or occupation of anti-bonding Mo4d-states, are observed, respectively. A
stabilization of the lattice is not favored also by the sterical strain induced after the difference in the
radii between Mo and Pb atoms. e.g., optimized lattice parameters for substitutionally Pb-doped MoS2
lattice reveal a slight increase of the lattice parameter a from 3.12 Å to 3.15 Å due to the single Pb
impurity atom and up to 3.19 Å due to the “cluster” of four Pb atoms. The interlayer distance is not
considerably affected in both cases and is ~0.05 Å smaller, than in the pure 2H-MoS2. The
intercalation of MoS2 lattice by individual Pb atoms has opposite effect: the a lattice parameter is
increased by ~0.01 Å, while the interlayer distance is getting larger by ~0.4 Å. These trends agree
well with the large atomic radius of Pb atoms. Indeed, the calculated metal-sulfur bond lengths in the
bulk of PbS and MoS2 are: 2.94 Å for Pb-S and 2.40 Å for Mo-S, while the length of the covalent
Pb-S bond within Pb-doped MoS2 lattice is 2.65 Å.
Furthermore, the influence of the concentration and ordering of the impurity atoms on the
thermodynamic stability of doped Mo1íxPbxS2 and intercalated Pby/MoS2 phases can be considered.
To characterize the stability of Mo1íxPbxS2 and the Pby/MoS2 phases, the cohesion energies Ecoh
were calculated for a set of 4 × 4 × 1 2H-MoS2 supercells modified by 1–4 Mo atoms (Table 1).
In agreement with the picture of the electronic structure, the absolute values of Ecoh for all modified
systems vanish with the growing content of Pb, which is an evidence for the weakening of chemical
bonding in both doped and intercalated MoS2 compared to the pristine MoS2. Noteworthy,
the cohesion energies of the solid solutions as well as intercalates vary almost in the same order of
magnitude, and the competing formation of both phases during the Pb-promoted growth of MoS2
nanotubes can be anticipated.
The exact growth mechanism and the main compounds participating in the growth of MoS2
nanotubes have still to be found. Yet, a first insight in this process is possible by the consideration of
some model reactions. The formation energies ǻE for Mo1íxPbxS2 and Pby/MoS2 phases from MoS2

137
and the corresponding compounds were estimated using calculated change in the total energies of the
next formal reactions:
(1 í x) MoS2 + xPbS + xS = Mo1íxPbxS2

(1)

yPb + MoS2 = Pby/MoS2

(2)

Both reactions have been found to be highly endothermic (Table 2). Concerning the calculated
values of ǻE, a strong tendency for the phase separation of Pb-modified MoS2 lattice into a mixture
of simple binary sulfides and simple elements may be contemplated. These theoretical observations
are in agreement with the experimental finding of the time-depended Pb content in fabricated
MoS2 samples.
Table 2. Cohesion energies Ecoh and energies of formation ǻE of Mo1íxPbxS2 solid solutions
and Pby/MoS2 intercalates concerning reactions (1) and (2), as a function of the content
and the arrangement of Pb atoms within model supercells. LDA DFT calculations.
Isomer

Pb atoms Arrangement

Supercell Composition

Ecoh, eV/atom

ǻE, eV/Pb

no Pb atoms

Mo32S64, x,y = 0.00

í1.58

-

Pb1Mo31S64, x = 0.031

í1.50

4.65

Pb2Mo30S64, x = 0.063

í1.45

3.75

Pb2Mo30S64, x = 0.063

í1.43

4.63

Pb2Mo30S64, x = 0.063

í1.43

4.42

Pb2Mo30S64, x = 0.063

í1.43

4.46

Pb3Mo29S64, x = 0.094

í1.39

3.34

Pb4Mo28S64 , x = 0.125

í1.33

3.18

Pb1/Mo32S64, y = 0.031

í1.51

5.72

Pb2/Mo32S64, y = 0.063

í1.47

3.79

Pb2/Mo32S64, y = 0.063

í1.46

4.28

Pb2/Mo32S64, y = 0.063

í1.44

5.66

Pb2/Mo32S64, y = 0.063

í1.43

5.73

Mo1íxPbxS2 Doped by
D1
D2a
D2b
D2c
D2d
D3a
D4a

single Pb atom
two neighbor Pb atoms
within the same layer
two distant Pb atoms
within the same layer
two neighbor Pb atoms in
different layers
two distant Pb atoms in
different layers
three neighbor Pb atoms
within the same layer
four neighbor Pb atoms
within the same layer

Pby/MoS2 Intercalated by
I1
I2a
I2b
I2c
I2d

single Pb atom
two neighbor Pb atoms
within the same vdW gap
two distant Pb atoms
within the same vdW gap
two neighbor Pb atoms in
different vdW gaps
two distant Pb atoms in
different vdW gaps

Noteworthy, the formation energies for the model supercells with “associated” (neighboring)
Pb atoms in both Pb-doped and Pb-intercalated MoS2 are considerably lower, than for those
supercells, where all Pb atoms are separated. e.g., the interaction between two Pb atoms separated by

138
the distance ~6.3 Å in Mo1íxPbxS2 solid solutions is almost absent (ǻE for D2b and D2c isomers are
close to that of D1, Table 1). The same tendency can be obtained in Pby/MoS2 intercalates, yet, with
a higher range of interaction between Pb atoms (ǻE for I2b isomer is still less, than for I1, Table 1).
Thus, the solid solutions of Mo1íxPbxS2 should be more stable than Pby/MoS2 intercalates, since the
coalescence of intercalated Pb atoms is more favorable and Pby/MoS2 intercalates might exist in the
narrower part of the phase diagram at a lower Pb content, than Mo1íxPbxS2 solid solutions.
Concerning the analysis of calculated formation energies for the case of single doping or
intercalating Pb atoms, the formation of Mo1íxPbxS2 solid solutions is more likely, than the formation
of Pby/MoS2 intercalates. As well, the values of formation energies allow an estimation of the limit of
the lead solubility at certain temperature T at thermodynamic equilibrium. As example, we consider
roughly the possible content within Mo1íxPbxS2 compounds after reaction (1). The preference of one
of two states in a chemical reaction is determined by the free energy change ǻF = ǻU í TǻS, where
ǻU is the change of internal energy, ǻS is the change of entropy. The condition corresponding to the
phase separation is ǻF ” 0.
In the first approximation, ǻS can be defined as the configurational entropy of randomly distributed
Pb atoms in the metal sublattice of Mo1íxPbxS2. From the theory of ideal solutions it follows, that
'S

1  x  ln 1  x ·
§
R ¨ ln x 
¸
x
©
¹

(3)

where ǻS is expressed per mole of Pb. ǻU can be approximated as the energy of reaction (1)
(ǻU = ǻE). The estimation of the free energy change for the phase separation within Mo1íxPbxS2
solid solutions with a fortiori low x imply the use of the energy for the formation of single Pb atom
within MoS2 lattice, i.e., ǻE = 4.65 eV or 448.7 kJ/Pb-mol (isomer D1, Table 1).
The results of this approach based on the formal reaction (1) are visualized in Figure 6. They
reveal that the substitution of Pb atom instead of the Mo atom within the MoS2 lattice is a quite
rare event and the content of Pb at thermodynamic equilibrium would be around x = 10í7 at the
temperatures ~3000 K. In this manner, the experimentally fabricated MoS2 nanotube samples with
primary Pb content of x = 0.12, which were formed during the extreme reaction conditions of the
sun-light driven evaporation, are thermodynamically unstable and should show a high urge towards
a phase separation.
Figure 6. Calculated free energy ǻF of the phase separation within Mo1íxPbxS2 solid
solutions depending on the temperature and Pb content (x).

139
3. Experimental Section
MoO3íx nanowhiskers and MoS2 nanotubes were prepared using solar ablation system [15]. The
Pb and MoS2 mixture was sealed in a quartz ampoule and irradiated using solar ablation system for
30 s in the case of MoO3íx nanowhiskers, and for 10 min in the case of MoS2 nanotubes. Both
nanostructures contain a few atomic percent of Pb. After spending a year in ambient conditions
these samples were annealed at 810 °C in the presence of H 2S and forming gas (5% H2 and 95% N2).
The annealing time varied from 30 min to 2 h. The sample labeling can be seen in Table 2.
At each stage the samples were examined with a transmission electron microscope (TEM) operating
at 120 kV, equipped with an energy-dispersive X-ray spectroscopy (EDS) detector for chemical
analysis. The SEM image was generated using a scanning electron microscope (SEM), and
high-resolution imaging was achieved with a field-emission gun TEM operating at 300 kV.
The calculations were performed using the SIESTA 2.0 (Madrid, Spain) implementation [23,24]
within the framework of density-functional theory (DFT) [25]. The exchange-correlation potential
within the local-density approximation (LDA) with the Perdew-Zunger parametrization was used [26].
The core electrons were treated within the frozen core approximation, applying norm-conserving
Troullier-Martins pseudopotentials [27]. The valence electrons were taken as 3s23p4 for S, 4d55s15p0
for Mo, 6s26p2 for Pb, while Pb5d10 electrons were included as semi-core state. The pseudopotential
core radii were chosen, as suggested by Martins, and are equal to 1.70 aB for all S states, 2.45 aB for
Mo4d and Mo5s, 2.65 aB for Mo5p states, 3.20 aB for Pb6s and Pb6p, 3.00 aB for Pb5d states. In all
calculations double-ȗ basis set is used for all atoms. For k-point sampling, a cutoff of 10 Å was
used [28]. The k-point mesh was generated by the method of Monkhorst and Pack [29]. The real-space
grid used for the numeric integrations was set to correspond to the energy cutoff of 300 Ry. For the
study of substitutional doping and intercalation of MoS2 by Pb atoms the supercell of 4 × 4 × 1 of
2H-MoS2 unit cells was considered. The energies of chemical reactions were estimated from the
calculations of total energies for 2H-MoS2, bcc-Mo, fcc-Pb, fcc-PbS and molecular S8. All
calculations were performed using variable-cell and atomic position relaxation, with convergence
criteria set to correspond to a maximum residual stress of 0.1 GPa for each component of the stress
tensor, and maximum residual force component of 0.01 eV/Å.
4. Conclusions
The experimental work shows that MoO3íx (1) nanowhiskers and MoS2 (1) nanotubes synthesized
using solar ablation system, are stable for more than a year in the ambient conditions. However,
Pb atoms that were observed in these nanostructures, are not stable in high concentration and tend
to diffuse out of the lattice. The Pb outdiffusion from these nanostructures did not influence their
high-temperature stability and their conversion into MoS2 (3,4) nanotubes. In all cases MoS2 nanotubes
were observed after the sulfidization and annealing of MoO3íx (2) nanowhiskers and MoS2 (2) nanotubes.
According to the EDS measurements the initial Pb concentration in both nanostructures was reduced
after one year in the drawer. While subsequent annealing of these nanostructures lead to additional
reduction of their Pb concentration. It can be seen from the EDS spectra that in the case of molybdenum
suboxide nanowhiskers MoO3íx (2) as a precursor material, the Pb peak disappears after two hours

140
(H2S) annealing, while the same peak disappears after 30 min annealing at 810 °C in the case of
MoS2 (2) nanotubes.
DFT calculations have verified a very low affinity of Pb atoms as either substitutional or intercalating
agents to the host MoS2 lattice. Both types of modifications should lead to the destabilization of the
electronic structure of the pristine MoS2 and cause a weak chemical bonding between Pb and S atoms
with the oxidation state of Pb2+. The calculations have also demonstrated that the interaction between
Pb atoms as intercalates is slightly stronger, than the interaction between Pb atoms as substitutional
dopants in the MoS2 lattice. Thus, Pby/MoS2 intercalates might exist only at lower Pb content
compared to a solid solution of Mo1íxPbxS2. The latter phase seems to be a more favorite form of Pb
within the fabricated MoS2 nanotubes. The estimations of the free energies for the phase separation of
Mo1íxPbxS2 solid solutions explain the experimentally observed time- and high-temperature reduction
of the Pb content in the nanotube lattice. The experimentally reached Pb concentrations (x = 0.3)
may be attributed to a thermodynamically unstable system which was obtained by a fast kinetic of
the chemical reaction between lead atoms and the primary 2H-MoS2 lattice during the solar
ablation experiments.
Acknowledgments
We wish to thank Jeffrey M. Gordon and Daniel Feuerman from the Jacob Blaustein Institutes for
Desert Research in Sede Boqer of the Ben-Gurion University for their assistance in the synthesis of
the oxide nanowhisekrs and nanotubes. This research was supported by the ERC grant INTIF 226639;
the EU ITN 317451 grant and a grant of the Israel Science Foundation. Reshef Tenne acknowledges
the support of the Harold Perlman Foundation and the Irving and Azelle Waltcher Foundation in honor
of Moshe Levy. We are grateful also to the Irving and Cherna Moskowitz Center for Nano and BioNano Imaging. Reshef Tenne holds the Drake Family chair in Nanotechnology and is the director of
the Helen and Martin Kimmel Center for Nanoscale Science.
Author Contributions
Olga Brontvein and Reshef Tenne designed the research. Olga Brontvein recorded and analyzed
TEM, SEM and EDS data. Andrey Enyashin carried out DFT calculations. The manuscript was
written through contributions of all authors. All authors have given approval to the final version of
the manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.

Kaplan-Ashiri, I.; Cohen, S.; Gartsman, K.; Rosentsveig, R.; Seifert, G.; Tenne, R. Mechanical
behavior of individual WS2 nanotubes. J. Mater. Res. 2004, 19, 454–459.

141
2.

3.
4.

5.
6.
7.
8.
9.
10.

11.
12.
13.

14.
15.

16.

17.
18.

Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Ivanovskaya, V.; Heine, T.; Seifert, G.; Wiesel, I.;
Wagner, H.; Tenne, R. On the mechanical behavior of WS2 nanotubes under axial tension and
compression. Proc. Natl. Acad. Sci. USA 2006, 103, 523.
Nagapriya, K.S.; Goldbart, O.; Kaplan-Ashiri, I.; Seifert, G.; Tenne, R.; Joselevich, E. Torsional
stick-slip behavior in WS2 nanotubes. Phys. Rev. Lett. 2008, 101, 195501.
Frey, G.L.; Tenne, R.; Matthews, M.J.; Dresselhaus, M.S.; Dresselhaus, G. Optical properties
of MS2 (M = Mo, W) inorganic fullerenelike and nanotube material optical absorption and
resonance raman measurements. J. Mater. Res. 1998, 13, 2412–2417.
Staiger, M.; Rafailov, P.; Gartsman, K.; Telg, H.; Krause, M.; Radovsky, G.; Zak, A.; Thomsen, C.
Excitonic resonances in WS2 nanotubes. Phys. Rev. B 2012, 86, doi:10.1103/PhysRevB.86.165423.
Virsek, M.; Jesih, A.; Milosevic, I.; Damnjanovic, M.; Remskar, M. Raman scattering of the
MoS2 and WS2 single nanotubes. Surf. Sci. 2007, 601, 2868–2872.
Enyashin, A.; Gemming, S.; Seifert, G. Nanosized allotropes of molybdenum disulfide.
Eur. Phys. J.-Spec. Top. 2007, 149, 103–125.
Levi, R.; Bitton, O.; Leitus, G.; Tenne, R.; Joselevich, E. Field-effect transistors based on WS2
nanotubes with high current-carrying capacity. Nano Lett. 2013, 13, 3736–3741.
Remskar, M.; Skraba, Z.; Cleton, F.; Sanjines, R.; Levy, F. MoS2 as microtubes. Appl. Phys. Lett.
1996, 69, 351–353.
Remskar, M.; Mrzel, A.; Virsek, M.; Godec, M.; Krause, M.; Kolitsch, A.; Singh, A.; Seabaugh, A.
The MoS2 nanotubes with defect-controlled electric properties. Nanoscale Res. Lett. 2010,
6, 26.
Yella, A.; Mugnaioli, E.; Panthöfer, M.; Therese, H.A.; Kolb, U.; Tremel, W. Bismuth-catalyzed
growth of SnS2 nanotubes and their stability. Angew. Chem. Int. Ed. 2009, 48, 6426–6430.
Radovsky, G.; Popovitz-Biro, R.; Stroppa, D.G.; Houben, L.; Tenne, R. Nanotubes from
chalcogenide misfit compounds: Sn–S and Nb–Pb–S. Acc. Chem. Res. 2013, 47, 406–416.
Radovsky, G.; Popovitz-Biro, R.; Staiger, M.; Gartsman, K.; Thomsen, C.; Lorenz, T.; Seifert, G.;
Tenne, R. Synthesis of copious amounts of SnS2 and SnS2/SnS nanotubes with ordered
superstructures. Angew. Chem. Int. Ed. 2011, 50, 12316–12320.
Zak, A.; Sallacan-Ecker, L.; Margolin, A.; Genut, M.; Tenne, R. Insight into the growth
mechanism of WS2 nanotubes in the scaled-up fluidized bed reactor. Nano 2009, 4, 91–98.
Brontvein, O.; Stroppa, D.G.; Popovitz-Biro, R.; Albu-Yaron, A.; Levy, M.; Feuerman, D.;
Houben, L.; Tenne, R.; Gordon, J.M. New high-temperature Pb-catalyzed synthesis of inorganic
nanotubes. J. Am. Chem. Soc. 2012, 134, 16379–16386.
Brontvein, O.; Jayaram, V.; Reddy, K.P.J.; Gordon, J.M.; Tenne, R. Two-step synthesis of MoS2
nanotubes using shock waves with lead as growth promoter. Z. Anorg. Allg. Chem. 2014, 640,
1152–1158.
Feldman, Y.; Wasserman, E.; Srolovitz, D.; Tenne, R. High-rate, gas-phase growth of MoS2
nested inorganic fullerenes and nanotubes. Science 1995, 267, 222–225.
Yamazoe, N.; Kihlborg, L. Mo5O14—Twinning and three-dimensional structure, determined
from a partly tantalum-substituted crystal. Acta Crystallogr. Sect. B 1975, 31, 1666–1672.

142
19. Yamazoe, N.; Ekstrom, T.; Kihlborg, L. Structural effects of vanadium substitution in
molybdenum oxide (Mo17O47). Acta Chem. Scand. Ser. A 1975, 29, 404–408.
20. Kihlborg, L. Stabilization of the tunnel structure of molybdenum oxide (Mo5O14) by partial
metal atom substitution. Acta Chem. Scand. 1969, 23, 1834–1835.
21. Boker, T.; Severin, R.; Muller, A.; Janowitz, C.; Manzke, R.; Voss, D.; Kruger, P.; Mazur, A.;
Pollmann, J. Band structure of MoS2, MoSe2, and Į-MoTe2: Angle-resolved photoelectron
spectroscopy and ab initio calculations. Phys. Rev. B 2001, 64, 235305.
22. Walsh, A. Effects of reduced dimensionality on the electronic structure and defect chemistry of
semiconducting hybrid organic-inorganic PbS solids. Proc. R. Soc. A 2011, 467, 1970–1985.
23. Ordejon, P.; Artacho, E.; Soler, J.M. Self-consistent order-N density-functional calculations for
very large systems. Phys. Rev. B 1996, 53, 10441–10444.
24. Soler, J.M.; Artacho, E.; Gale, J.D.; Garcia, A.; Junquera, J.; Ordejon, P.; Sanchez-Portal, D.
The SIESTA method for ab initio order-N materials simulation. J. Phys.-Condens. Matter 2002,
14, 2745–2779.
25. Hohenberg, P.; Kohn, W. Inhomogeneous electron gas. Phys. Rev. 1964, 136, B864–B871.
26. Perdew, J.P.; Zunger, A. Self-interaction correction to density-functional approximations for
many-electron systems. Phys. Rev. B 1981, 23, 5048–5079.
27. Troullier, N.; Martins, J.L. Efficient pseudopotentials for plane-wave calculations. Phys. Rev. B
1991, 43, 1993–2006.
28. Moreno, J.; Soler, J.M. Optimal meshes for integrals in real-space and reciprocal-space unit
cells. Phys. Rev. B 1992, 45, 13891–13898.
29. Monkhorst, H.J.; Pack, J.D. Special points for Brillouin-zone integrations. Phys. Rev. B 1976,
13, 5188–5192.

143

Microstructural Study of IF-WS2 Failure Modes
Jamie Cook, Steven Rhyans, Lou Roncase, Garth Hobson and Claudia C. Luhrs
Abstract: This manuscript summarizes the failure mechanisms found in inorganic fullerene-type
tungsten disulfide (IF-WS2) nanoparticles treated with diverse pressure loading methods. The
approaches utilized to induce failure included: the use of an ultrasonic horn, the buildup of high
pressures inside a shock tube which created a shock wave that propagated and impinged in the
sample, and impact with military rounds. After treatment, samples were characterized using electron
microscopy, powder X-ray diffraction, energy dispersive X-ray spectroscopy, and surface area
analysis. The microstructural changes observed in the IF-WS2 particulates as a consequence of the
treatments could be categorized in two distinct fracture modes. The most commonly observed was
the formation of a crack at the particles surface followed by a phase transformation from the 3D
cage-like structures into the 2D layered polymorphs, with subsequent agglomeration of the plate-like
sheets to produce larger particle sizes. The secondary mechanism identified was the incipient
delamination of IF-WS2. We encountered evidence that the IF-WS2 structure collapse initiated in all
cases at the edges and vertices of the polyhedral particles, which acted as stress concentrators,
independent of the load application mode or its duration.
Reprinted from Inorganics. Cite as: Cook, J.; Rhyans, S.; Roncase, L.; Hobson, G.; Luhrs, C.C.
Microstructural Study of IF-WS2 Failure Modes. Inorganics 2014, 2, 377–395.
1. Introduction
The first inorganic fullerene-like nanoparticles of WS2 were discovered by Tenne et al., working
at the Weizmann Institute of Science in 1992 [1]. Using the diffusion-controlled sulfurization of metal
oxides, they were able to empirically prove the existence of inorganic compounds based on WS2 with
structures that were believed to only exist for carbon-based materials. Shortly thereafter, the discovery
of WS2 nanotubes and fullerene-like structures led to the establishment of a new field of inorganic
chemistry; one dealing with closed-hollow nanomaterials [2–5].
Originally, carbon fullerenes, made of concentric layers of carbon, were thought to provide
outstanding tribological properties, and under low loads and high velocities they do [6,7]. However,
due to their phase transition from graphite to diamond at high temperatures and pressures, they cannot
entirely live up to the expectations that their structures suggest [8]. IF-WS2 particles have numerous
properties similar to the carbon fullerenes, making them excellent solid lubricants; their extremely
small size (in the nanometer scale) gives them the ability to fill imperfections in the lubricated material
to effectively smooth the surface and prevent degradation. Due to their spherical shape, IFs are also
said to act like nano-ball bearings, feature that allows them to roll rather than slide, performing better
than other solid lubricants [4].
Currently, IF-WS2 particulates are recognized for their potential not only as lubricants but also as
structural nanocomposites and shock absorbers [9–15].

144
The impressive shock absorbing performance of IF-WS2 and IF-MoS2 structures were first
reported by Zhu et al. [16,17]. Using uniaxial shockwave pressures up to 30 GPa and studying
the microstructural features of samples treated at diverse pressure settings, it was found that these
cage-like particulates have superior performance than their carbon counterparts. Those reports also
identified a lattice reduction for the samples treated at the higher pressures and provided a glimpse of
the material breakage mechanisms. Moreover, those journal articles introduced the concept that
smaller, more spherical IF particulates might be less prone to sustain damage than larger particles, a
key feature to have in mind when designing highly resistant nanoparticles.
By studying the structural characteristics of a material that has fractured one can understand how
materials fail and then make changes to the design and prevent, to a certain extent, the encountered
failure modes. Fracture mechanics principles allow us to predict maximum working stress for a given
material, establish relationships between materials properties, stress levels, crack producing flaws and
conditions for the cracks propagation that ultimately will result in catastrophic breakdown of structures.
Here, we induced the breakage of IF-WS2 particulate structures employing three different
setups: (i) the use of an ultrasonic horn operated at diverse amplitudes and periods of time with the
sample immersed in a solvent; (ii) pressure waves created inside a shock tube; and (iii) impact with
military rounds. The ultrasonic horn was used at diverse amplitudes and treatment times and produced
cyclic loading conditions similar to the ones used for fatigue studies of macroscopic objects. The
shock tube and the military rounds produced a single uniaxial impact event over only fractions of
a second.
The objective of the experiments was to study the microstructural characteristics of the materials
postmortem and based on those observations identify failure modes—fracture mechanisms. Then
establish relationships between those and the type and duration of each treatment method employed,
identifying the most important variables for the particulates failure.
Our results demonstrated that the effects observed when applying pressure loads to the IF-WS2
nanometric particles are a particular case of the more general principles established by fracture
mechanics to explain macroscopic crack propagation: stress concentrators at the particle surface play
a much larger role in the material failure than the directionality or duration of methods used to apply
the loads.
2. Results and Discussion
2.1. Sample Characteristics before Treatment
The IF-WS2 particulates utilized for this study were analyzed by diverse techniques before
treatment. TEM analysis showed that the IF nanostructures presented the typical hollow cores and the
particles exhibit polyhedral shapes, as seen in Figure 1. The value for the lattice spacing observed
varied depending on where the measurement was performed; the faces of the polyhedrons presented
an average value of 0.62 nm, while the vertices had an average separation of 0.63 nm. Thus, when the
IF particulates do not have a perfect spherical shape the lattice distance seem not to be uniform.
Observation by SEM determined that the particle size distribution was between 30 and 420 nm, with

145
an average diameter of 143 nm and only a few particles of more than 500 nm. BET surface area of
the sample was determined to be 35 m2/g.
Figure 1. TEM image of an IF-WS2 particle before treatment.

2.2. Sample Postmortem Analysis
Before presenting the post treatment sample characteristics, it is worth noting that the diverse
methods used to impinge in the sample (each described in the experimental methods section) impart
very different levels of energy at very different time scales:
It has been estimated that the use of military rounds of the size used at the velocities recorded
provided approximately 2070 Joules in a non-isotropic single event that lasted only fractions of a
second. The compressed-gas driven shock tube experiment produced a wave that first impacted in the
Kevlar layer located in front of the sample (which was used solely to contain the powder IF-WS2)
impacting next the sample in a non-isotropic manner. This single incident also lasted fractions of a
second. A pressure sensor located after the Kevlar layer recorded an average pressure of 1742 KPa
when using He as the gas, as such, the estimated pressure of the wave that actually reached
the sample. In contrast, during the ultrasonic treatment, the particles were completely immersed
in a liquid, thus, being subject to isotropic shocks for much longer periods of time (between 3,600 and
10,800 s). The estimated energy conveyed when using a 1200 W ultrasound horn during such periods
of time was from 4032 to 12,096 kilojoules respectively. In all cases, only IF-WS2 particulates were
tested, no composites are included in this study.
Some research groups have studied the variations in the d002 interlayer spacing in the samples
before and after pressure loads using XRD powder data or TEM observations to relate such values to
unit cell expansions or contractions [17,18]. Given the dispersion in lattice values noted by TEM
studies mentioned above in the untreated samples, the extraction of useful postmortem TEM data
could result very complicated. The change in lattice constant for IF-WS2 seen before treatment arises
from the differences in curvature between different locations of the polyhedral cage-like spherical or
semispherical IF-WS2 nanoparticles. To determine if XRD patterns could provide more reliable data,
given that they include information from all the crystal locations and orientations, we performed such
analysis for the samples treated by the three loading methods.

146
The XRD patterns of samples subjected to the diverse treatments are shown in Figure 2. The graph
in the upper section of the image shows the peaks encountered between 5° and 70° (2ș), while the
bottom image presents only the (002) peak. Specimens treated by ultrasound waves in wet conditions
(dispersed in a solvent) and then dried for the analysis, present a (002) reflection that has shifted to
higher angles. This alteration in the X-ray pattern indicates a reduction in the cell spacing when
compared to the untreated sample. A similar observation was made in references above [17,18]. We
believed that such shift could be related not only to a compressed IF cell but to a phase transition from
the IF cage structure to a more dense 2D plate-like WS2 arrangement with slightly smaller 00l
distance. The width of the (002) XRD peak for ultrasound treated powders in Figure 2 (red line) might
suggest the presence of both, IF-WS2 and two-dimensional WS2 in the same powder sample.
The peaks identified with the * symbol do not belong to WS2 but to WO3. The existence of small
amounts of WO3 could not be verified by EDS analysis of large portions of the sample by SEM but
was clearly identified by TEM of selected locations (uploaded as Figure S1).
By comparison, the samples collected after shock tube tests and military rounds insult show either
no change or a slight shift to lower angles. Such result could be interpreted as no modification of the
structure or a small expansion in the lattice parameter. We believe, however, that disparity in those
results, when compared to the ultrasonic treated sample, might be more related to the experimental
constrains faced when collecting the powders than a clear representation of the specimen features.
The collection of the section of the sample that was directly exposed to the shockwave or to the impact
of the penetrator was difficult; the particles had the tendency to get expelled from the impact site and
only some of them remained in the center of the tested specimen or close to the hole left by the round
and got recovered. That is, the XRD signal represents the average of all the particles gathered in the
postmortem, some of which were not as affected by the event as the ones at the center of the impact
location. In the case of ultrasonic treatment, the full sample was contained at all times and all particles
were equally affected by the treatment, therefore rendering a powder that presented more uniform
features and more reliable data, thus confirming a shift in the d002.
The precise effects of applying diverse loading conditions in the specimens’ microstructures were
much more evident once the samples were studied by electron microscopy. We were able to conduct
the analysis with smaller amounts of sample and use only particulates from affected locations. Figure 3
presents a SEM image of the IF-WS2 particles before (a) and after (b) ultrasonic treatment. The
existence of extended layered structures (identified with arrows) along with the original IF polyhedral,
semispherical shapes is the most noticeable change. According to particle size distribution
measurements (Figure 3c) the average diameter of the particles in the sample after treatment has
shifted to larger sizes, 252 nm (blue histogram), and clearly shows a bimodal distribution. Particles
between 800 nm and 1 micron, not seen in the untreated sample, are present in the ultrasonic treated
specimen. The larger plate-like particulates correspond to the 2D polymorph and no longer present
the hollow cage structure of IF-WS2. The appearance of the 2D phase supports the XRD data collected
regarding both, the XRD peak shift and its width, and agrees with the initial reports suggesting a
lattice compression [17,18]. Notably, the particles did not only collapsed and lose their hollow cores;
they also transformed into 2D structures and then suffered an agglomeration mechanism that turned
them into much larger, irregular shape units.

147
Figure 2. X-ray diffraction analysis of particulates recovered after treatments.
Top: reflections observed between 5° and 70°; Bottom: (002) peak for all samples. Only
the sample gathered after ultrasonic treatment showed a shift to higher angles. The
symbol * identifies a WO3 peak.

Given those observations, we have denominated this transformation or fracture mode as
“agglomeration”, since the collapsed particles seem to have larger dimensions after the experiment,
implying the agglomeration of many of the original IF spheroids. However, agglomeration is a general
term that only refers to the enlargement of the particles and does not completely account for all the
complex steps that the particles suffer during this failure mode. Some of the stages we believe are
necessary for the 2D polymorphs to acquire the sizes observed imply: (i) an initial crack formed in
the surface of the IF particles; (ii) crack propagation towards the core, leading to a complete fracture
of the walls; (iii) a structure collapse and disappearance of the hollow cores; (iv) sheets re-arrangement;
and (v) new bonds being formed. Moreover, the large sheet-like particles in Figure 3c show little
evidence of grain boundaries, which suggests the involvement of a sintering or crystal growth

148
mechanism, rather than simple agglomeration. Some of the flattened sections (Figure 3c) reached even
micrometer lengths, far from the 143 nm average diameter in the untreated IF-WS2.
Figure 3. SEM micrographs of (a) untreated particles, (b) samples treated with ultrasonic
horn during 3 h at 20% amplitude (17 —pp). Note the appearance of layered particles of
much larger dimensions than original spheroids; (c) Particle diameter comparison
between the two samples.

The temperature during the sonication experiments was controlled by an ice bath to prevent solvent
evaporation. Nevertheless, given that the particles were exposed to the ultrasonic waves during
extended periods of time (up to 10,800 s in some cases) the process seems to have provided enough
energy to allow all the steps mentioned above to happen (i–v), followed by crystal growth. The fact
that the (002) peak for ultrasound sample, in Figure 2, is much more intense than the one from the
untreated specimen supports the idea that more particles with such orientation exist, as is typical of
layered structures, fact now verified also by the SEM data.
The study of the ultrasonic sample postmortem by transmission electron microscopy corroborates
the existence of both the 2D and 3D (IF) polymorphs. The TEM study of the IF particles that did not
collapse presented evidence of crack initiation in the vertices of the polyhedrons. Figure 4 (left) shows
one of those cracks, which has not yet propagated through the whole structure. In a similar fashion
that surface defects can promote the fracture of macroscopic objects at smaller levels of stress than
the ones predicted by theoretical calculations, in IF-WS2 nanostructures the imperfections in the

149
particles’ surfaces act as stress concentrators. No WO3 was identified in the specific particle structure
where such observations were made, ruling out its effect in the particle fracture mechanisms.
Figure 4. Left: TEM micrograph of a particle treated with ultrasonic horn that presents
a surface crack. The inset is a magnified version of the area pointed by the arrow;
Right: TEM studies corroborated the damage of both, internal and external layers of the
IF structures.

As pointed out in multiple references [19–22], materials do not fracture while in compression, they
fracture in tension. The failure modes observed in samples post ultrasonic treatment present a similar
signature as the ones expected from fatigue experiments; they resemble the fracture of materials that
have been subjected to a cyclic load which involves the contraction and expansion of the structure.
We found evidence of breakage in the surface (crack in Figure 4 left) presumably created during one
of the expansion cycles while in tension. However, there is also evidence of some of the sheets
breaking inside the particles’ core (Figure 4 right), indicating a similar mechanism occurring but in
the opposite direction, pointing to a cyclic load.
The complete breakage of the outer layers in the crack area (in Figure 4 left) seem to happen
without extensive plastic deformation, resembling the typical brittle failure. Such feature is observed
in many particles and diverse sections of the sample. However, we also identified particles that present
a slightly larger separation between the first and second layers of the lattice than the one found for the
inner layers. Such a fact can be interpreted as evidence of a very small amount of plastic deformation
occurring in the direction perpendicular to the lattice fringes.
Apart from the multiple particles suffering the agglomeration type failure (fracture-agglomerationphase transformation steps described before), a second mechanism was identified by the observation
of the samples by TEM: incipient “delamination”. Indeed, small sections (ca. 10–15 nm) of a few of
the IF particles seem to partially separate from the main body of the structure. Figure 4 (right)
exemplifies that case, the arrows pointing to sites of partial delamination. A similar exfoliating
mechanism has been observed by other groups [17,23]. We did not find evidence of larger sections or
complete layers totally separating from the IF spheroids in the ultrasonic treated specimens. The
amount of particles presenting partial delamination, when compared to the amount that presented
agglomeration, was minimal.
The main difference found between the two failure modes is that agglomeration mechanism destroys
the hollow cage structures, generating 2D extended structures, and delamination does not. The main

150
parallelism between them is that both, incipient delamination in the outer layers of the IF structures,
and crack formation with propagation towards the particle core are observed in the vicinity of vertices,
edges or surface defects.
The fact that the vertices in untreated particulates presented a larger average interlayer distance
(0.63 nm) than the faceted sides or walls (0.62 nm) is suspected to have an influence over which
failure more (agglomeration vs. delamination) is predominant. However, the difference between such
values is so small that more data and a higher resolution instrument might be required to be able to
make a definitive correlation.
There is no evidence that suggest that one failure mode represents a first step towards the other.
The transformation between the IF particles and the 2D structures in the agglomeration mode seem to
start with a crack that advances towards the particle core, feature observed in samples treated in
different conditions. The few cases of delamination seen do not exhibit such crack directionality but a
separation of only the outer layers. The failure modes do not seem to be associated to particle sizes
either. Perfect spherical particles might present a correlation between such since the size will determine
the radius of curvature, which is a factor that affects the crack propagation in macroscopic objects.
One of the main limitations we encountered during this study was the lack of a method to quantify
the level of damage that the samples suffered. From the SEM images we could estimate the percentage
of the sample surface that transformed into larger particles due to the agglomeration mechanism (as
in Figure 3). However, multiple images need to be taken to assure that the features observed actually
represent the bulk of the sample and the existence of the flat particles underneath the surface layer of
the sample could be overlooked. Since the most common failure mechanism observed by large was
agglomeration, and only very small sections of the sample delaminated, we contemplated the use of
surface area measurements to evaluate the level of damage. Indeed, most particles suffer an
enlargement that will produce a surface area reduction (due to the agglomeration mechanism) and
only a few of them will have more area exposed (due to delamination).
The surface area data from samples treated with the ultrasonic horn at diverse amplitudes was
calculated using the BET methodology and is presented in Figure 5. Samples treated at higher
amplitudes or for longer periods of time (not shown) suffer an evident surface area reduction. Both
amplitude and time imply more energy being imparted to the particles. The former because the
energy transported by a wave is directly proportional to the square of the amplitude of the wave:
A high-energy wave is characterized by a high amplitude; a low-energy wave is characterized by a
low amplitude. The latter, because energy is calculated directly from the product of the Watts provided
by the source and the time of the treatment. The trend towards smaller surface areas in Figure 5 could
be correlated to particle agglomeration and the presence of extended 2D layered structures, as the ones
seen in the example presented in Figure 3.
The BET surface area measurement does not provide a definitive way to evaluate damage when
both agglomeration and delamination occur in large extents. However, it provided a way to compare
the extent of damage in bulk samples when agglomeration was the dominant mechanism (as in the
ultrasound treated samples) without the need for lengthy SEM/TEM analysis.

151
Figure 5. Surface area analysis (m2/g) vs. amplitude found for samples subjected to
ultrasonic treatments.

As previously stated, the study of samples by electron microscopy provided a more definitive proof
of the changes that the IF-WS2 structure suffered under load than the shifts observed in the XRD
analysis alone. Such was particularly true for the samples subjected to shockwave tests and impacted
with military rounds, since smaller amounts of powder could be analyzed and only particulates from
insulted regions were collected. See Figure 6.
Figure 6. Cracks in the vertices of the particles are the only evidence of damage in
samples subjected to shock tube tests. No crack propagation or structure collapse was
observed. No evidence of delamination was found either.

The evaluation of the samples treated by shockwaves revealed that the damage sustained was less
than the one observed for ultrasonic methods, as expected from the low pressures registered by the
sensors. The SEM observation of the sample from shock tube revealed that only very small sections
of the sample showed agglomeration. TEM images showed crack initiation and Figure 6 exemplifies
this feature. The arrows on the image point to two different sites where the crystal structure fractured.
The pressures used by shockwaves produced by N2 and by He were not enough to cause the sample to
suffer a complete collapse. We did not find evidence of delamination in any of the multiple particles
analyzed. Thus, we found no evidence that the agglomeration failure mode is preceded by delamination.
Samples impinged with military rounds suffered a larger level of damage than that observed for
shock tube tests. Unfortunately, as stated before, it was difficult to collect large amounts of sample in
the circumference left by the penetrator, reason for the XRD and BET data not to be as reliable as the

152
one seen for ultrasonic treatments. Nevertheless, the SEM analysis of the sample provided proof of
the same mechanisms described earlier in this article. We found layered structures similar to the ones
in Figure 3b; however, in this case the agglomerates reached not only micrometer dimensions in length
and width, but also in height. There are regions in the postmortem sample that the IF particles can still
be encountered but the agglomerates are much more abundant.
Figure 7 shows both, the more frequently observed agglomeration (left) and an image of the less
common delamination (right). Regarding the agglomeration case, once the IF particles fracture at the
polyhedron vertices or edges, they collapse, their sheets re-arrange to diminish any spaces in between
layers and are added to the larger crystals in a fashion similar to the one observed during crystal growth
(Figure 7 center). In Figure 7 left, grain boundaries mark the regions where IF particles have joined
the main particle. The letters in the figure represent: (a) matter has completely been incorporated onto
the agglomerate and is now part of the surface to which new particles will connect; (b) the grain
boundaries of the initial IF particle are still visible within the main body of the agglomerate; (c) a new
layer is being formed, some sections of the same present layered structure with sharp edges and others
partially conserve the spherical shape; and (d) an IF-WS2 semispherical particle still lays in the surface.
Figure 7. Sample recovered from impact with military rounds. Note the similarities
between regions where IF have been added to an existent layered structure, such as a, b,
c, and d (left) with sites proposed by crystal growth theories for addition matter in a
surface (center). Particle showing layers that delaminated (right).

The angles in the vertices in the polyhedral particles seem to play a similar role than the radius of
curvature of the surface or internal cracks have in traditional fracture mechanics analysis.
Furthermore, the larger crystal to which other IF are being added present imperfections such as kinks,

153
ledges, steps, etc. Each of those sites represent regions in the sample where dangling bonds exist;
those locations are more prone to accept the addition of new matter to reduce the overall surface
energy as is usually described by crystal growth theories [24,25]. According to the latter, the
thermodynamics of crystal surface formation and transformation is based on the idea that the energy
that each atom’s position on a crystal surface is determined by the number of neighboring atoms and
the amount of unsatisfied bonds. The numbers in Figure 7 (center) represent: surface atoms, surface
vacancies, adatoms, kinks, and steps. Figure 7 (right): this image represents the only instance that we
found of what it seems as a complete layer delamination; the outer sheets of the material seem to come
apart to reveal an inner particle still intact.
The effect of temperature in the material agglomeration during the military round impact certainly
plays a significant role to transform the IF into extended sheet-like WS2 structures. However, given
the complexity of the test, no temperature values were recorded and the heat effect will only be subject
of further study and discussion in future reports.
Figure 8 summarizes our findings. In the agglomeration mechanism: IF particles are more
susceptible to breakage in areas where defects or discontinuities, such as edges and vertices, are
located. Those act as stress raisers and are the primary sites where crack initiates, in the outer layers
of the cage structures. Once the crack propagates it will reach the hollow cores, which will promote
the particles collapse and the transition between the 3D and the 2D WS2 polymorphs. The layered
structures will agglomerate to reach micron scale dimensions. The collapse and agglomeration of the
IF structures was the main effect identified as a consequence of the application of ultrasonic treatment
in diverse conditions and the use of military rounds despite the diverse time scale of each of those
methods. Shock waves in the conditions employed only produced surface cracks without the full
structure collapse. Agglomeration signature is identified as the predominant mechanism in the
experimental conditions used.
Figure 8. Representation of the two failure modes identified in the IF-WS2 structures.

154
We also identified evidence that suggests that the particles can suffer delamination. The areas in
the IF particles surface where this failure mode was identified also correspond to locations where
defects or discontinuities are observed. During this failure mode the Van der Waals attraction keeping
the layers in the together is broken as a consequence of the shock and small portions of the outer layer
come apart, leaving behind sections with a smaller diameter particle but still IF characteristics. The
separation between surface layers do not seem to propagate into internal layers. Remarkably, during
our experiments we observed incipient delamination in only very few particles treated by sonication.
No correlation was found between the extent of this failure mode and times of treatment. We observed
only one instance of an almost complete layer separation (Figure 7c).
In sum, all the electron microscopy observations conducted in the postmortem products point to
the existence of surface flaws as starting point where agglomeration or delamination take place,
independent of the amount energy imparted by the insult, the time of the event or the isotropic or
non-isotropic nature of the experiment.
The structure fails by two modes: either as a consequence of the creation of a surface crack and its
propagation into the hollow core, followed by the structure collapse and resulting in the transformation
into a 2D sheet-like polymorph and the agglomeration of particles into larger bodies, or by delamination
of small sections of the sample, which does not completely destroy the IF structure but removes
superficial sections, leaving the hollow cores and inner walls intact.
What conditions favor agglomeration over delamination is still subject of study. No link has been
found here between the two fracture mechanisms and levels of stress. Crack formation, usually
followed by propagation and structure collapse has been identified without evidence of delamination
at diverse levels of stress and over a wide range of times of treatment. The interlayer separation and
the radius of curvature (given by particle size and existence of faceted structures) are suspected to
play a role on which mode dominates however, such has not been clearly demonstrated/quantified.
Studies of brittle fracture demonstrate that the experimental fracture strengths of most materials
are lower than the theoretical ones (based on calculations related to atomic bond energies) due to the
existence of microscopic flaws that act as stress raisers, amplifying the stress at a given point. The
postmortem TEM analyses indicate that the IF-WS2 particles deform and break where structural
discontinuities appear: defects and edges in the polyhedral shapes, along the degree of curvature of
the particulates, seem to play a critical role on the material resistance to breakage, following a pattern
similar to the ones observed in more general fracture mechanics principles. These findings provide
evidence that without the stress concentrators in the particle surface, as would be the case if more
spherical particles were used, the material should be able to support higher pressure loadings,
supporting previous reports [17]. The bottom of Figure 8 shows a schematic representation of how
such particles will look like.
No evidence of crack formation or delamination was seen in the more uniform sections of the
particle surfaces. These results suggest that IF-WS2 is prone to show a phase transformation into a 2D
polymorph and only near spherical particles might show remarkable shock absorption characteristics.
Despite finding that the structures have similar failure modes than their carbon counterparts,
which can diminish their shock resistance, our findings also point to a solution to the failure: use of

155
defect free (more spherical, not faceted) particulates, supporting previous statements made by
Tenne et al. [17].
It is worth noting that the creation of defect free IF-WS2 particles might prove challenging; the use
of completely spherical WO3 produced by microwave plasma methods in our labs and reacted with
H2S, produced highly faceted IF-WS2, see Figure 9. Different treatment conditions should be developed
in order to achieve completely spherical nanoparticles.
Figure 9. IF-WS2 generated at 900 °C from the WO3 produced using a microwave
atmospheric plasma system. Despite the highly spherical nature of the precursor, the
IF-WS2 particulates present facets and polyhedral shapes. A few atoms thick 2D WS2
were also observed.

With failure modes now clearly characterized future research could (a) correlate the cracks
observed, the angles at the vertices and the levels of stress used, to generate mathematical expressions
that could predict mechanical properties such as fracture toughness for the particular size of IF-WS2
and (b) design nanoparticles against fracture.
3. Experimental Section
Commercially manufactured inorganic fullerene-type tungsten disulfide (IF-WS2) was obtained in
the form of a commercial lubricant Nanolub (NanoMaterials Ltd., Ness Ziona, Israel). The sample
was washed with ethanol to remove additives; the solvent was separated using an Eppendorf (New
York, NY, USA) Model 5418 centrifuge and the IF-WS2 particles dried at room temperature in a
desiccator. The sample was studied by XRD and SEM before being subject to treatments and
as postmortem.
3.1. Methods Used to Induce Failure
The ultrasonic tester used for this procedure was a Sonomechanics (New York, NY, USA) 1200 W
ultrasonic liquid processor shown in Figure 10. The equipment consists of a generator that provides
electric signal input to the piezoelectric transducer, a water-cooled barbell horn and a protective
enclosure. The samples were placed in a solvent (water or ethanol) and the beaker containing such
dispersion was placed in a second beaker, which was filled with iced water and a copper coil that was

156
also circulating cold water to maintain the bath at a constant temperature. The generator produced a
constant frequency of 20 kHz at the horn tip; however diverse samples were exposed to the ultrasonic
waves during different conditions: the amplitude was varied between 20% and 100% (which
correspond to 17 to 81 —pp, respectively) and times of exposure extended from 3600 to 10,800 s.
Figure 10. Ultrasonic horn setup. Note that a cooling bath was used to maintain a
constant temperature, however the bath that included the cooling coil and the sample in
the liquid media were not in contact. The latter were placed in a second beaker. The blue
foam was used to keep the inner beaker in place.

Ice Water Tube to
Transducer
Water Cooled
Transducer

Bell
Horn

Ice Water
Tube to
Copper
Sample +
liquid media

Copper
cooling
coil

A shock tube was used for creating a shock wave that propagated and impinged on the sample.
A Kevlar layer was positioned on the surface of the particles to prevent their dispersion inside the
shock tube. The system uses a pressurized driver section and evacuated driven section separated by a
pair of diaphragms to create a supersonic wave that propagates along the tube and impinges on a test
holder at the end of the device. Pressure transducers are located along the tube allow for monitoring
of the shock and measuring its speed and strength (pressure sensors 1, 2, and 3 in Figure 11). Detailed
theory can be found in [26].
The material to be tested was attached to the holder and bolted onto the end section where pressure
transducer cable was connected. The shock tube was then loaded with two heat-treated copper
diaphragms 0.025 inches thick scored to a depth of 0.013 inches. Vacuum was then drawn to an
absolute pressure of approximately 2 mm Hg in the driven section through the use of a roughing pump.
When this was complete, gas was added in a controlled manner until the driven section reached
25 mm Hg. The vacuum gauge was isolated and then disconnected to prevent damage to it during
creation of the pressure event. The driver section was then loaded with either Nitrogen or Helium
(depending on the desired Mach number) to 720 psig. When the pressure gauge reached 350 psig, the

157
mid-section was isolated. The shock tube was then fired by opening the firing valve. As pressurized
gas entered the diaphragm section the pressure differential across the second diaphragm caused it to
burst creating a wave through the driven section. The subsequent drop in pressure in the diaphragm
section caused the first diaphragm to burst and the formation of a second, faster wave. The second
wave overtook the first and the two coalesced into a single wave that traveled along the driven section
and to the sample holder.
Figure 11. Shock tube setup (top) and image of the sample contained with Kevlar (bottom).

For the impact with military rounds the IF powder sample was dispersed in between Kevlar layers
and the layers introduced into a nylon pouch. The later was then positioned in front of a clay bed and
a 7.62 mm NATO round was then fired at it as indicated in Figure 12. The IF-WS2 particles located
close to the hole left by the penetrator were then collected and analyzed by the methods described in
the next section.
Figure 12. Sample testing setup (left) and orifice left by penetrator in a nylon bag that
contained the sample in between Kevlar layers (right).

158
3.2. Methods Used to Characterize Samples
A Philips 1830 PAnalytical X-Ray Diffractometer (Almelo, The Netherlands), using a copper
source, KĮ of 1.54 Å was employed for the analysis of the crystalline components of the samples. The
powders were positioned into a silicon zero background sample holder and the diffraction patterns
recorded between 5° and 70° (2ș) with 0.020° step size and one second per step.
A JEOL (Tokyo, Japan) 2010F FASTEM field emission gun scanning transmission electron
microscope (STEM/TEM) equipped with Gatan (Pleasanton, CA, USA) GIF image filtering system
was used to collect images and electron diffraction patterns. The samples were dispersed in ethanol
and a drop of the mixture was allowed to dry in a Holey carbon copper grid. The JEOL 2010F lattice
resolution is 0.1 nm.
The microstructural features of the specimens post treatment were analyzed using a Zeiss Neon
(Oberkochen, Germany) 40 High Resolution Scanning Electron Microscope (SEM). Images were
acquired at diverse magnifications while the microscope was operated between 5 and 20 kV. Energy
Dispersive Spectroscopy (EDS) experiments were conducted in conjunction with the SEM using an
Apollo 10 silicon drift detector (SDD). Data was collected and analyzed using Genesis Spectrum
software (EDAX, Mahwah, NJ, USA).
Brunauer Emmet Teller (BET) surface area analysis was performed employing a Quantachrome
(Boynton Beach, FL, USA) Nova 4200 Physisorption analyzer. A degas step was conducted prior to
the analysis. The measurements were done using nitrogen atmosphere.
3.3. Attempts to Produce Highly Spherical IF-WS2
Highly spherical WO3 was generated in an atmospheric plasma torch microwave system following
procedures previously reported [27,28]. The WO3 sample was then transformed into IF-WS2 by a
thermal treatment in a tubular furnace operated at 900 °C for a period of 2 h using H2S atmosphere.
4. Conclusions
The microstructural changes observed in the IF-WS2 particulates as a consequence of treatments
with ultrasonic horn, shock tube and military rounds could be categorized in two distinct fracture modes.
The most commonly observed was the phase transformation of the 3D cage-like structures to the
2D layered polymorphs, which involved a crack forming at the particulate surface leading to the
structure collapse and subsequent agglomeration of the plate-like sheets, to produce larger particle
sizes. Such mechanism dominated the samples’ microstructure for experiments performed using
ultrasonic waves and the ones exposed to military rounds. Incipient cracks were identified in shock
wave postmortem samples.
A less common secondary mechanism of particle breakage was also identified; the delamination
of IF-WS2. However, the later was mainly observed as an incipient process, with only very small sections
of the shells separating from the IF main body and the survival of the IF hollow cage characteristics.
The steps required for particle failure by the first mechanism included: (i) an initial crack formed
in the surface of the IF particles; (ii) crack propagation towards the core, leading to a complete fracture
of the walls; (iii) structure collapse and disappearance of the hollow cores; (iv) sheet re-arrangement;

159
and (v) new bonds being formed, which resulted in an overall agglomeration. We encountered
evidence that the IF-WS2 structure collapse initiated at the edges of the polyhedral particles, which
acted as stress concentrators, demonstrating that general fracture mechanics concepts can be applied to
materials at the nanoscale. Based on those findings and in agreement with previous reports,
defect-free perfectly spherical IF-WS2 surfaces are expected to present improved shock absorbing
performance than that observed for polyhedral shape IFs.
We found that surface area measurements could be correlated to the extent of particle damage when
agglomeration could be recognized as the main effect of the pressure load.
We demonstrated that more irregular particles (faceted) tend to fail at defect sites that act as
stress concentrators independently of how energy is delivered: shock being applied in fractions of a
second, or over long periods of time, as an isotropic or non-isotropic event, as a single occurrence or
by cyclic treatment.
Supplementary Information
Figure S1. TEM image of IF-WS2 nanoparticles (left) along EDS spectra (right) showing
that in some locations WO3 can be identified.

Acknowledgments
This work was conducted with support of the Office of Naval research, Code 30. We appreciate
the help that Chris Clay and John Gibson provided to conduct the shock tube tests. We thank Jing
Bing from University of New Mexico and JEOL, USA Inc. for facilitating the collection of TEM data.
We also thank Alexey Peshkovsky, from Industrial Sonomechanics for guidance using the transducer.

160
Author Contributions
The findings in this manuscript are part of Jamie Cook’s Master degree thesis work.
Claudia C. Luhrs advised the thesis and directed the research. Garth Hobson guided the shock tube
tests. Lou Roncase conducted the military round impacts and Steven Rhyans performed a summer
internship aiding the team with some of the ultrasonic experiments.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.

Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and cylindrical structures of tungsten
disulfide. Nature 1992, 360, 444–446.
2. Margulis, L.; Tenne, R.; Iijima, S. Nucleation of WS2 fullerenes at room temperature.
Microsc. Microanal. Microstruct. 1996, 7, 87–89.
3. Feldman, Y.; Frey, G.; Homyonfer, M.; Lyakhovitskaya, V.; Margulis, L.; Cohen, H.; Hodes, G.;
Hutchison, J.; Tenne, R. Bulk synthesis of inorganic fullerene-like MS2 (M = Mo, W) from the
respective trioxides and the reaction mechanism. J. Am. Chem. Soc. 1996, 118, 5362–5367.
4. Rapoport, L.; Bilik, Y.; Feldman, Y.; Homyonfer, M.; Cohen, S.; Tenne, R. Hollow nanoparticles
of WS2 as potential solid-state lubricants. Nature 1997, 387, 791–793.
5. Tenne, R.; Homyonfer, M.; Feldman, Y. Nanoparticles of layered compounds with hollow cage
structures (inorganic fullerene-like structures). Chem. Mater. 1998, 10, 3225–3238.
6. Kroto, H.; Heath, J.; Obrien, S.; Curl, R.; Smalley, R. C-60—Buckminsterfullerene. Nature
1985, 318, 162–163.
7. Kroto, H. The stability of the fullerenes C-24, C-28, C-32, C-36, C-50, C-60 and C-70. Nature
1987, 329, 529–531.
8. Vul’, A.; Davidenko, V.; Kidalov, S.; Ordan’yan, S.; Yashin, V. Fullerenes catalyze the
graphite-diamond phase transition. Tech. Phys. Lett. 2001, 27, 384–386.
9. Rapoport, L.; Fleischer, N.; Tenne, R. Applications of WS2 (MoS2) inorganic nanotubes and
fullerene-like nanoparticles for solid lubrication and for structural nanocomposites. J. Mater. Chem.
2005, 15, 1782–1788.
10. Diez-Pascual, A.M.; Naffakh, M. Mechanical and thermal behaviour of isotactic polypropylene
reinforced with inorganic fullerene-like WS2 nanoparticles: Effect of filler loading and
temperature. Mater. Chem. Phys. 2013, 141, 979–989.
11. Naffakh, M.; Diez-Pascual, A.M.; Marco, C.; Ellis, G.J.; Gomez-Fatou, M.A. Opportunities and
challenges in the use of inorganic fullerene-like nanoparticles to produce advanced polymer
nanocomposites. Prog. Polym. Sci. 2013, 38, 1163–1231.
12. Diez-Pascual, A.M.; Naffakh, M.; Marco, C.; Ellis, G. Rheological and tribological properties
of carbon nanotube/thermoplastic nanocomposites incorporating inorganic fullerene-like WS2
nanoparticles. J. Phys. Chem. B 2012, 116, 7959–7969.

161
13. Tehrani, M.; Luhrs, C.C.; Al-Haik, M.S.; Trevino, J.; Zea, H. Synthesis of WS2 nanostructures
from the reaction of WO3 with CS2 and mechanical characterization of WS2 nanotube
composites. Nanotechnology 2011, 22, 285714.
14. Zohar, E.; Baruch, S.; Shneider, M.; Dodiuk, H.; Kenig, S.; Tenne, R.; Wagner, H.D. The Effect
of WS2 Nanotubes on the Properties of Epoxy-Based Nanocomposites. J. Adhes. Sci. Technol.
2011, 25, 1603–1617.
15. Naffakh, M.; Diez-Pascual, A.M.; Marco, C.; Gomez, M.A.; Jimenez, I. Novel melt-processable
poly(ether ether ketone)(PEEK)/inorganic fullerene-like WS2 nanoparticles for critical applications.
J. Phys. Chem. B 2010, 114, 11444–11453.
16. Zhu, Y.; Sekine, T.; Li, Y.; Wang, W.; Fay, M.; Edwards, H.; Brown, P.; Fleischer, N.; Tenne, R.
WS2 and MoS2 inorganic fullerenes—Super shock absorbers at very high pressures. Adv. Mater.
2005, 17, 1500–1503.
17. Zhu, Y.; Sekine, T.; Li, Y.; Fay, M.; Zhao, Y.; Poa, C.; Wang, W.; Roe, M.; Brown, P.; Fleischer, N.
Shock-absorbing and failure mechanisms of WS2 and MoS2 nanoparticles with fullerene-like
structures under shock wave pressure. J. Am. Chem. Soc. 2005, 127, 16263–16272.
18. Vasu, K.; Matte, H.S.S.R.; Shirodkar, S.N.; Jayaram, V.; Reddy, K.P.J.; Waghmare, U.V.;
Rao, C.N.R. Effect of high-temperature shock-wave compression on few-layer MoS2, WS2 and
MoSe2. Chem. Phys. Lett. 2013, 582, 105–109.
19. Meyers, M.; Chawla, K. Mechanical Behavior of Materials, 2nd ed.; Cambridge University
Press: Cambridge, UK, 2009.
20. Griffith, A.A. The Phenomena of Rupture and Flow in Solids. Philos. Trans. Ser. A 1920, 221,
163–198.
21. Irwin, G.R. Fracture Dynamics; American Society for Metals: Cleveland, OH, USA, 1948.
22. Orowan, E. Fracture and Strength of Solids. Rep. Prog. Phys. 1949, 12, doi:10.1088/00344885/12/1/309.
23. Joly-Pottuz, L.; Martin, J.; Dassenoy, F.; Belin, M.; Montagnac, G.; Reynard, B.; Fleischer, N.
Pressure-induced exfoliation of inorganic fullerene-like WS2 particles in a Hertzian contact.
J. Appl. Phys. 2006, 99, 023524.
24. Kossel, W. Extenoling the law of bravais. Nach. Ges. Wiss. Gottingen. 1927, 143.
25. Stranski, I.N. Zur Theorie des Kristallwachstums. Z. Phys. Chem. 1928, 136, 259–278 (In German).
26. Crinklaw, J.D. Acoustic Shock Testing on Re-entry Vehicle Insulation Material. Master’s Thesis,
Naval Postgraduate School, Monterey, CA, USA, March 2005.
27. Wakeland, S.; Cui, Y.; Knapp, A.; Richard, M.; Phillips, J.; Luhrs, C. Multilayered nanoparticles
generated by plasma methods for energy storage applications. Nanosci. Nanotechnol. Lett. 2012,
4, 316–322.
28. Phillips, J.; Luhrs, C.C.; Richard, M. Review: Engineering particles using the aerosol-through-plasma
method. IEEE Trans. Plasma Sci. 2009, 37, 726–739.

162

Nanostructured Boron Nitride: From Molecular Design to
Hydrogen Storage Application
Georges Moussa, Chrystelle Salameh, Alina Bruma, Sylvie Malo, Umit B. Demirci, Samuel
Bernard and Philippe Miele
Abstract: The spray-pyrolysis of borazine at 1400 °C under nitrogen generates boron nitride (BN)
nanoparticles (NPs). The as-prepared samples form elementary blocks containing slightly
agglomerated NPs with sizes ranging from 55 to 120 nm, a Brunauer-Emmett-Teller (BET)-specific
surface area of 34.6 m2 gí1 and a helium density of 1.95 g cmí3. They are relatively stable in air
below 850 °C in which only oxidation of the NP surface proceeds, whereas under nitrogen, their
lower size affects their high temperature thermal behavior in the temperature range of 1450–2000 °C.
Nitrogen heat-treated nanostructures have been carefully analyzed using X-ray diffraction, electron
microscopy and energy-dispersive X-ray spectroscopy. The high temperature treatment (2000 °C)
gives hollow-cored BN-NPs that are strongly facetted, and after ball-milling, hollow core-mesoporous
shell NPs displaying a BET-specific surface area of 200.5 m2·g í1 and a total pore volume of
0.287 cm3·g í1 were produced. They have been used as host material to confine, then destabilize
ammonia borane (AB), thus improving its dehydrogenation properties. The as-formed AB@BN
nanocomposites liberated H2 at 40 °C, and H2 is pure in the temperature range 40–80 °C, leading to
a safe and practical hydrogen storage composite material.
Reprinted from Inorganics. Cite as: Moussa, G.; Salameh, C.; Bruma, A.; Malo, S.; Demirci, U.B.;
Bernard, S.; Miele, P. Nanostructured Boron Nitride: From Molecular Design to Hydrogen Storage
Application. Inorganics 2014, 2, 396–409.
1. Introduction
Advanced nanostructured materials may be defined as materials having one dimension in the 1 to
100 nm range. The massive academic and industrial research efforts concerning these materials over
the past decade arose from the remarkable variations in their physical and chemical properties when
their dimension shrinks to the nanometric scale. In this category of materials, the interest for
hexagonal-boron nitride (h-BN, but expressed here as BN) grew during the past few decades in relation
to their unique combination of key properties.
BN is a synthetic chemical compound containing boron (B) and nitrogen (N) atoms in a one-to-one
ratio. The in-plane atoms are linked through covalent bonds, while the out-of-plane layers are bonded
by weak interactions (van der Waals forces) between B and N atoms, alternatively, providing
anisotropic properties. BN displays a large band gap (~5.5 eV) and offers the lowest density
(d = 2.26 g·cm í3) among non-oxide ceramics. It proposes relatively good thermal stability in air and
vacuum, high thermal conductivity, good thermal shock resistance, high electrical resistance, a low
dielectric constant and loss tangent, microwave transparency, non-toxicity and easy machinability.
Furthermore, it is non-abrasive, lubricating and non-reactive towards molten metals [1–6].

163
BN was obtained for the first time by Balmain [7] in 1842 through the reaction between boric oxide
and potassium cyanide. It is nowadays produced by conventional powder technology, requiring
nitridation or carbothermal reaction of boric acid/boric oxide with melamine or urea and the use of
additives during the further sintering process [8]. It is used in various fields of chemistry, metallurgy,
high temperature technology, electronics and in thermal management applications. However, beside
the fact that the use of boric oxide inherently induces the presence of oxygen-containing phases, BN
is only produced as powders with a plate-like morphology and workpieces. This inherently limits the
development of BN.
Recently, interest at the academic level has arisen in both the synthesis of nanostructured BN and
their applications for energy and the environment [9–13]. The important industrial challenges in line
with nanostructured BN production requires the development of materials in which topologies, shapes
and morphologies are tuned on demand. Inherent difficulties of traditional techniques to manufacture
such materials can be addressed by the development of synthetic pathways where molecular/inorganic
chemistry, processing and material chemistry/science are combined rationally to process BN with
tailor-made properties [14]. The key step in nanostructured BN preparation is the selection of the BN
precursors. Precursors with a good B:N ratio where/for which hydrogen (H) is the only element added
to B and N are preferred. Borazine and derived polyborazylene are the most appropriate
candidates [15,16]. Within this context, in this review article, we discuss the use of borazine (BZ) as
a single-source molecular precursor used for the design of BN nanoparticles (NPs), hollow-cored
BN-NPs that are strongly facetted and hollow core-mesoporous shell NPs. The latter have been used
as host materials to encapsulate and store ammonia borane (AB).
2. Results and Discussion
2.1. Borazine-Derived BN Nanoparticles
Borazine (BZ) had been originally discovered by Alfred Stock in 1926 [17]. It displays a chemical
formula H3B3N3H3. It is a preformed B-N-like ring structure and has the correct B-to-N ratio.
Furthermore, it is economically competitive and attractive from a technical point of view, based on
its reaction starting from cheap compounds, such as (NH4)2SO4 and (NaBH4), reacting in tetraglyme
at low temperature (120–140 °C) [18]. Borazine offers the advantage of being liquid with an adequate
vapor pressure to be applied in gas phase pyrolysis processes to prepare nanostructured BN. As an
illustration, we have demonstrated the interest of BZ to produce BN nanoparticles by
spray-pyrolysis [19–22]. In our process, BZ is nebulized into an aerosol, and the stream consisting of
tiny BZ droplets suspended in the carrier gas is transported by the carrier gas to be passed through the
preheated tubular furnace at 1400 °C under nitrogen. In the hot-zone, the conversion of the nebulized
precursor occurs through molecular condensation and ring-opening mechanisms involving the
evolution of dihydrogen and producing vapors of BN ring-based species. The latter, reacting to form
the consolidated boron nitride network, are swept by the nitrogen carrier-gas flow and, then,
condensed into a white product getting collected into the cooling traps near the outlet of the furnace.
The as-obtained product is stored inside an argon-filled glove-box. The scanning electron microscopy
(SEM) images in Figure 1a show that the sample consists of particles with a relatively homogeneous

164
size. This indicates that the most important operating factors, including the properties of the starting
precursor, the pyrolysis temperature, the nitrogen flow rate, the residence time and heating rate of the
droplet particles, are controlled during processing.
Figure 1. SEM (a), TEM (b) and HRTEM (c) images of samples obtained by spray-pyrolysis
of borazine (BZ).

The low-magnification transmission electron microscope (TEM) bright field image of the sample
(Figure 1b) show elementary blocks that are composed of slightly agglomerated nanoparticles (NPs).
The particle size ranges from 55 to 120 nm. The high resolution TEM (HRTEM) image (Figure 1c)
of the particle core demonstrates that the specimen consists of very fine BN crystallites in which sp2
layers are significantly buckled in a disordered stacking sequence, exhibiting a size corresponding to
less than six atomic basal planes, whereas their length does not exceed 50 Å. This points to the fact
that BN is poorly crystallized similarly to a turbostratic structure. The TEM data are reinforced by the
X-ray diffraction (XRD) experiments (Figure 2). The corresponding XRD patterns show very broad
peaks at the h-BN (002), (100)/(101)/(004) and (110) positions. In particular, the (002) peak slightly
shifts to lower angles in such samples, and the (100), (101) and (004) peaks merge into a single broad
peak. Finally, the samples displayed a chemical formula of B1.0N0.9. Their specific surface area is
34.6 m2 gí1, and the helium density is 1.95 g cmí3, as measured by Brunauer-Emmett-Teller (BET)
and helium pycnometry, respectively.
B1.0N0.9-NPs are stable in air below 850 °C in which only surface oxidation proceeds [21]. Here,
we report the evolution of the nanostructural organization of B1.0N0.9-NPs in the temperature range of
1450–2000 °C under nitrogen. The XRD patterns in Figure 2 range from 10° to 90° for heat-treated
B1.0N0.9-NPs.

165
Figure 2. XRD patterns of borazine-derived B1.0N0.9-NPs and annealed at a temperature
ranging from 1450 to 2000 °C.
(002)

As-prepared B1.0N0.9 NPs

Count (a. u.)

Heat-treated at 1450°C
Heat-treated at 1600°C
Heat-treated at 1700°C
Heat-treated at 1800°C
Heat-treated at 2000°C
(100)/(101)
(004)

10

20

30

40

50

(110)

60

70

80

90

2 theta (°)

The XRD patterns of samples heat-treated in the temperature range of 1450–1600 °C display
features similar to the ones recorded for B1.0N0.9-NPs, indicating a turbostratic structure. For the
sample annealed at 1700 °C, the (002) peak at 25.30° is sharpened, suggesting that the crystallite size
became larger in the c-axis direction, although the shoulder-shaped broad feature remained on the
low-angle side of the peak. This is also shown for the sharper (100)/(101)/(004) peak, which tends to
be separated into the (004) peak and the (100)/(101) peak. The increase of the heat-treatment
temperature to 1800 °C and 2000 °C resulted in an increased resolution of the XRD patterns. We can
clearly distinguish the (002), (100)/(101), (004) and (110) peak positions. According to the sharpening
of the (002) and (100)/(101) peaks, we suggest that the crystallite size continuously increased in the
c- and a-axes directions from 1400 °C to 2000 °C. However, no clear peaks corresponding to the (102)
and (112) planes were observed. These findings tend to demonstrate that B1.0N0.9-NPs annealed at
2000 °C exhibit a turbostratic structure. The variation of the average crystallite size in the c-axis from
the (002) peak ( L c ) and the interlayer d002 spacing of the samples during heat-treatment is shown in
Figure 3. The dimension d002 is calculated from Bragg’s law using the diffraction angle of the (002)
peak. Lc represents the average number of stacked layers in the crystallites. The average stack height
L c is calculated from the Scherrer relation ( L c = 0.9Ȝ/(B2 í B'2)1/2cosșwhere Ȝ is the CuKĮ1
wavelength (Ȝ = nm), ș the Bragg angle of the (002) diffraction peak, B the full width at half
maximum intensity (FWHM) of the peak and B' the instrumental contribution).
In the range of 1450 °C ( L c = 1.10 nm; d002 = 0.367 nm)–1600 °C ( L c = 1.42 nm; d002 = 0.363 nm),
there is no major modification in both the apparent average grain size ( L c (002)) and the value of
the interlayer d-spacing d002. Values are close to those calculated for as-prepared B1.0N0.9-NPs
( L c = 1.06 nm; d002 = 0.376 nm). This indicates a relatively high amount of disorder in the structure
of the corresponding samples. At 1700 °C, the apparent average grain size increases slightly
( L c = 2.23 nm). Although the crystallization state in NPs heat-treated at 1700 °C is slightly improved,
the BN phase remains poorly ordered as confirmed by the value of d002 (d002 = 0.351 nm), higher than

166
that in a h-BN crystal (0.3327 nm). At 1800 °C, L c increases to 4.63 nm and the interlayer d002 spacing
is found to be 0.345 nm, which are values characteristic of a turbostratic phase. The minor changes in
the XRD patterns of samples heat-treated at 2000 °C is reflected in the values of L c (4.65 nm) and
d002 (0.346 nm). In addition to XRD studies, we investigated TEM (Figure 4) and HRTEM (Figure 5)
experiments to follow the evolution of the nanostructural organization in the temperature range of
1450–2000 °C.
Figure 3. Evolution of L c (002) and d002 vs. annealing temperature.
5
0,380
0.380
0,375
0.375

Interlayer spacing (d002)
Average crystallite size (c-axis)

0,365
0.365
3
0,360
0.360
0,355
0.355
2
0,350
0.350

Crystallite size (nm)

Interlayer spacing (nm)

4
0,370
0.370

0,345
0.345
1
0.340
0,340
1400

1500

1600

1700

1800

1900

2000

Temperature (°C)

Figure 4. TEM images of the samples annealed at (a) 1450 °C; (b) 1600 °C; (c) 1700 °C;
(d) 1800 °C and (e) 2000 °C.

167
Figure 5. HRTEM images of the samples annealed at (a) 1450 °C; (b) 1600 °C;
(c) 1700 °C; (d) 1800 °C; (e) 2000 °C; (f) evidence of a core-shell structure generated at
2000 °C.

The annealed samples form elementary blocks composed of nanosized particles that are round in
shape and slightly agglomerated. Both the average size of annealed particles and the agglomeration
seem to increase with the temperature of the annealing, which is in good agreement. This is clear for
the samples annealed at 1800 °C (Figure 4d) and 2000 °C (Figure 4e), respectively. We therefore
extended our analysis, by performing high resolution TEM (HRTEM), in order to refine/emphasize
the structural information.
Figure 5 reports HRTEM images of the same samples.
Clear differences appear between the samples annealed in the range of 1450–2000 °C. After
heat-treatment to 1450 °C (Figure 5a), the sample displays a turbostratic BN structure with more
distinct (002) layers in comparison to the nanostructure observed in pristine B1.0N0.9-NPs (Figure 1c).
In the sample annealed at 1600 °C (Figure 5b) and 1700 °C (Figure 5c), we can also observe the
formation of nanodomains made of BN layers surrounding voids. The HRTEM image reveals the
formation of concentric shelled nanodomains. The lattices of these BN nanostructures have a local
interlayer spacing of 3.51 Å. Annealing at a temperature of 1800 °C (Figure 5d) and 2000 °C
(Figure 5e,f) leads to hollow-cored BN-NPs that are strongly facetted, forming polygonal particles
with an interlayer spacing of 3.34 Å. We investigated the potential of samples heat-treated at 2000 °C
to confine H2 storage materials.
2.2. Hydrogen Storage Applications
Ammonia borane (AB) is a white crystalline solid that was first prepared by Shore and Parry
in 1955 [23]. Over the past decade, this compound has attracted considerable attention as portable

168
hydrogen storage materials, according to its high gravimetric hydrogen contents (ca. 20% by
weight) [24–29]. A very pertinent review dedicated to this compound and related derivatives as
dihydrogen sources was proposed by Staubitz et al. in 2010 [29].
In the pristine state, AB is almost stable under inert conditions up to about 100 °C and decomposes
within the range 100–200 °C through a two-step exothermic process where two equivalent H2, as well
as undesired by-products, such as borazine B3N3H6 and NH3, are evolved [24,25]. This decomposition
suffers from three important problems: (1) the process is exothermic, which means that the
storage reversibility is thermodynamically impossible in acceptable operating conditions; (2) the
dehydrogenation temperature is too high for the portable/mobile application prospects; (3) the
emission of undesired by-products is detrimental, as they are incompatible with the use of proton
exchange membrane fuel cell (PEMFC) [27].
A promising solution seems to be the decrease of the particle size at the nanoscale (<10 nm)
via confinement of the borane in a porous compound (i.e., scaffold) [30]. As an illustration,
Gutowska et al. showed that AB confined in the mesoporosity of silica SBA (Santa Barbara
Amorphous)-15 has improved dehydrogenation behavior in comparison to the pristine hydride, with an
onset at 70 °C and the liberation of borazine-free H2 [31]. The destabilization of AB is generally
explained by two phenomena. The first one is the nanosizing of the hydride particle. At the nanoscale,
both kinetics and thermodynamics might be altered by both size and interface effects. In fact, the
surface energy value could be different as a result of the interactions between the active confined
material and the scaffold. The second phenomenon is associated with Hį+···H įí surface interactions,
with Hįí of the BH3 moiety of AB and Hį+ belonging to surface/terminal hydroxyl groups (OH)
generally found on carbonaceous or oxide nano-scaffolds. Such acid-base interactions enhance H2
release, but usually lead to an unstable material at room conditions [25].
An improved strategy we recently demonstrated is to use nano-scaffolds free of reactive surface
groups [22]. For that purpose, we used the B1.0N0.9-NPs annealed at 2000 °C, which we labeled
B1.0N0.9-NP2000. As measured by energy dispersive X-ray spectrometry (EDX), boron, nitrogen
and oxygen contents are 43.55, 55.7 and 0.75 wt%, respectively. Unfortunately, they exhibit a
Brunauer-Emmett-Teller (BET)-specific surface area of 21.8 m2·g í1, which is low to achieve the
nanoconfinement of AB. Therefore, we applied a ball-milling process of this sample to tentatively
increase the specific surface area, leading to the sample labeled B1.0N0.9-NP2000BM. In comparison to
B1.0N0.9-NP2000, the sample B1.0N0.9-NP2000BM shows a considerably increased BET-specific
surface area with 200.5 m2·g í1 and a total pore volume of 0.424 cm3·g í1 as measured by the
Barrett-Joyner-Halenda (BJH) analysis. As a result of the ball-milling, the HRTEM images
(Figure 6a,b) of the sample showed that cleavage of the walls occurred through the basal planes.
In addition, Figure 6c show that the stacking sequence can in some cases be disordered similarly to
those of t-BN after ball-milling.
As a result of the BET and TEM investigations, we successfully demonstrated that the walls of the
hollow-cored BN-NPs could be opened to provide porosity after ball-milling.

169
Figure 6. HRTEM images of the sample B1.0N0.9-NP2000BM evidencing in (a) and (b),
a cleavage of the walls in the area delimited by the white arrows, and in (c), a disordering
of the stacking sequence.

(a)

(b)

(c)
Hydrogen storage materials can be confined within porous scaffolds by melt infiltration (if the
active hydrogen storage material melts and do not decompose) or solution infiltration. In our
procedure, a solution of AB in tetrahydrofuran (THF) was infiltrated into the framework of the sample
B1.0N0.9-NP2000BM according to an optimized procedure described elsewhere [25]. A nanocomposite
labeled AB@B1.0N0.9-NP2000BM was formed. It was stored at 3–4 °C. The successful impregnation
of AB in B1.0N0.9-NP2000BM was followed by N2 adsorption/desorption analysis of the nanocomposite.
A BET-specific surface area of 6.7 m2·g í1 and a total pore volume of 0.023 cm3·g í1 are measured,
which demonstrates that AB was inserted into the hollow core and blocked the pores of the
nano-scaffolds. More interesting is that the decomposition of AB is down to 81 °C (compared to
110 °C for the pristine AB in our conditions) and that a major evolution of H2 is identified by MS. In
our experimental conditions, the only by-product was identified to be NH3 above 80 °C. At 80 °C, a
weight loss of 1.7 wt% was measured, which means an effective gravimetric hydrogen storage
capacity of 3.4 wt% by considering a weight ratio equal to 1:1 in AB@B1.0N0.9-NP2000BM.
Our results confirmed the remarkable benefit of hollow-cored BN-NPs on the dehydrogenation
behavior of AB. The performance is comparable to the dehydrogenation results of AB confined into a
magnesium metal organic framework (MOF) [32,33] or nickel MOF [34], whereas only nanoconfinement

170
is considered here. Most interesting, the MS results suggest that there is no detectable trace of borazine
as a gaseous by-product. Another important observation standing from the thermogravimetric analysis
coupled mass spectrometry (TGA-MS) result is that AB@B1.0N0.9-NP2000BM is stable at room
conditions. Accordingly, the stability of AB@B1.0N0.9-NP2000BM at <40 °C is clearly attributed to
the absence of surface Hį+, and the improvement of the dehydrogenation properties of AB in
AB@B1.0N0.9-NP2000BM can be exclusively ascribed to the effect of nanoconfinement.
3. Experimental Section
The synthesis of borazine was carried out in an argon atmosphere, using argon/vacuum lines
and Schlenk-type flasks. Argon (>99.995%) was purified by passing through successive columns of
phosphorus pentoxide (Sigma-Aldrich, Saint Quentin, France), sicapent• (Millipore S.A.S,
Molsheim, France) and copper oxide-based catalysts (Sigma-Aldrich, Saint Quentin, France). The
Schlenk flasks were dried at 120 °C overnight before pumping under vacuum and before filling with
argon for the synthesis. Sodium borohydride (NaBH4, •98.5%, powder from Sigma-Aldrich, Saint
Quentin, France), ammonium sulfate ((NH4)2SO4, •99.0% from Sigma-Aldrich (Saint Quentin,
France) and tetraethylene glycol dimethyl ether(CH3O(CH2CH2O)4CH3, 99.0%, from Sigma-Aldrich
(Saint Quentin, France) were used as-received. It should be mentioned that ammonium sulfate was dried
at 120 °C inside an oven for three days, then put under vacuum during cooling for 1 h. Manipulation
of the chemical products was made inside an argon-filled glove box (Jacomex BS521; Dagneux,
France) dried with phosphorus pentoxide.
Borazine Synthesis: The operating procedure, adapted from the literature [18], was previously
reported by our group [19]. FTIR (Caesium Iodide (CsI) windows/cmí1): (N–H) = 3451 medium;
(B–H) = 2509 medium; (B–N) = 1454 small; (B–N–B) = 897 medium. 1H NMR
(300 MHz/CDCl3/ppm): = 3.30–5.35 (quadruplet, 3H, BH), 5.35–6.05 (triplet, 3H, NH). 11B NMR
(96.29 MHz/C6D6/ppm): = 30.1 (br).
Nanoparticle Preparation: The experimental set-up is composed of a nebulized spray generator
(RBI, Meylan, France), in which the spray is generated by a piezoelectric device (barium titanate).
Frequency (800 kHz) and power (100 W) alimentations are adjusted to obtain the aerosol. The aerosol
temperature is first held at 15 °C by a regulated water circulation to avoid borazine evaporation and/or
condensation. The piezoelectric device generates an ultrasound beam, which is directed to the liquidgas interface; a fountain formed at the surface followed by the generation of the spray, resulting from
vibrations at the liquid surface and cavitations at the gas-liquid interface.
The borazine was directly introduced in the aerosol generating chamber under nitrogen, then
aerosolized and carried to the pyrolysis furnace with a 0.5 mL·min í1 nitrogen flow rate. The thermal
decomposition of borazine was performed in a hot alumina tube containing an isothermal zone of
0.1 m in length. The fast heating rate implies gaseous species generation leading to powder formation
by a chemical vapor condensation route. The particles were finally trapped into two collectors placed
before the vacuum pump and containing filter-barriers made of microporous alumina (pore size of
1 —m). Yield was estimated to be 0.22 g·min í1. After synthesis, the particles are stored inside an
argon-filled glove-box. In a typical experiment, 27 mL (21.9 g) of borazine is used to produce 6.5 g
of B1.0N0.9-NPs. However, the exact yield is difficult to estimate, because of the design of the

171
spray-pyrolysis system. A non-negligible/considerable quantity of powders, deposited in the furnace
tube, cannot be recovered. To study the evolution of their crystallization degree, 2 g of the B1.0N0.9-NPs
are placed into boron nitride boats and then introduced in a graphite furnace (Gero 5 Model HTK 8).
The furnace chamber is subsequently suctioned with a pump charged with nitrogen before heating. A
cycle of ramping at 10 °C·min í1 is used to heat the sample to the desired temperature (in the range
1400–2000 °C) with a holding time of 1 h, before cooling down to RT at 10 °C·min í1. Chemical analysis
found (wt%): B, 50.0; N, 49.4; O, 0.6. The milling process of B1.0N0.9-NP2000 is performed under
inert condition (argon) with a planetary ball-miller (Retsch PM100; Haan, Germany). The described
process has been optimized (in terms of mass, ratio balls/BN, time, rotation) to our conditions.
Typically, degassed B1.0N0.9-NP2000 (at 150 °C under dynamic vacuum for 24 h) is introduced into a
stainless steel reactor (25 mL). Balls in stainless steel are added (weight ratio balls: B1.0N0.9-NP2000
of 20). The milling process is performed at 600 rpm for 1 h. The as-obtained B1.0N0.9-NP2000BM is
finally sieved.
The infiltration of ammonia borane is performed as follows: the host material B1.0N0.9-NP2000BM
(100 mg) is degassed at 150 °C under dynamic vacuum for 24 h in a Schlenk tube and then cooled to
0 °C. In an argon-filled glove box, a concentrated solution of ammonia borane (100 mg, 97%; Sigma
Aldrich, Saint Quentin, France) is prepared using 0.5 mL of anhydrous THF (Sigma Aldrich, Saint
Quentin, France). The ammonia borane solution is injected into the Schlenk tube containing
B1.0N0.9-NP2000BM kept under static vacuum and at 0 °C. By capillary action, the ammonia borane
solution fills the channels of the host rapidly, which is evidenced by vigorous effervescence. When
the effervescence stops, the sample is put under ultrasonic treatment for 20 min at 0 °C. Finally, the
as-obtained sample AB@B1.0N0.9-NP2000BM (weight ratio B1.0N0.9-NP2000BM:AB of 1) is dried
under dynamic vacuum for 48 h at 0 °C. The composite samples obtained are denoted
AB@B1.0N0.9-NP2000BM. Samples are transferred in an argon-filled vial and then stored in a fridge
at 3–4 °C.
Characterizations: The B1.0N0.9-NPs and annealed samples are first mounted on carbon film-covered
stainless pads for scanning electron microscopy (SEM, Hitachi S4800, Tokyo, Japan) including
Energy Dispersive X-ray Spectroscopy (EDX, EDAX/TSL Genesis 4000, Tokyo, Japan). Due to the
insulating properties of BN, the samples are sputtered with 10 nm of a Pd/Au mixture to prevent charging
during SEM observations. In parallel, the same samples are ultrasonicated in ethanol, and the resulting
solution is afterwards deposited on a collection of hollow carbon-film-covered copper grids for
transmission electron microscopy (TEM, TOPCON 002B working at 200 kV, Tokyo, Japan)
observation. Samples were characterized using a Philips PW 3040/60 X’Pert PRO X-ray diffraction
system (Eindhoven, The Netherland). Powder samples are prepared by placing ~100 mg on the XRD
sample holder (PVC), and the sintered pieces were put down on the XRD sample holder for data
collection. Cu KĮ (Ȝ = 1.54 Å) radiation with a Ni filter was used with a working voltage and a current
of 40 kV and 30 mA, respectively. Scans were continuous from 2ș = 10°–90° with a time per step of
0.85 s in increments of 0.017°. Peak positions and relative intensities were characterized by
comparison with JCPDS (Joint Committee on Powder Diffraction Standards) files of the standard
material (JCPDS card No 34-0421). Debye-Scherrer line broadening was used to calculate the average
crystallite sizes from each XRD pattern. The transmission electron microscopy (TEM) studies of

172
B1.0N0.9-NP2000BM samples were carried out with a JEOL (Tokyo, Japan) GmbH 2010F
transmission electron microscope (Cs = 1 mm) operating at 200 kV. The characterization of the
samples was performed by N2 adsorption/desorption (Sorptomatic 1990 Series, Thermo Fisher
Scientific Inc, Waltham, MA, USA). Thermogravimetric analysis (TGA) measurements (repeated at
least three times to ensure the reproducibility of the results) were performed with a Mettler Toledo
TGA/SDTA 851e (Schwerzenbach, Switzerland) under the following conditions: sample mass
9–10 mg, aluminum crucible of 100 —L with a pinhole, heating rate of 5 °C·min 1, temperature range
25–200 °C and atmosphere of N2 (60 mL·min 1). The purity of H2 was analyzed by mass spectrometry
(Canon Anelva Corporation MQA200TS, Tokyo, Japan) coupled to the TGA apparatus.
4. Conclusions
This article reviews our recent advancements in the synthesis and energy application of
nanostructured boron nitride. Such materials exhibit chemical and physical properties that are
significantly different from those of bulk and microsized materials. In our approach, we discussed our
recent strategy based on borazine, which has been used as a vapor phase pyrolysis precursor for the
synthesis and fabrication of BN nanoparticles. In particular, we demonstrated the possibility of
tailoring the nanostructure of these nanoparticles by a further annealing process at high temperature
(2000 °C), leading to hollow-cored BN-NPs that are strongly facetted, forming polygonal particles
with an interlayer spacing of 3.34 Å. The ball-milling of these nanostructures allowed developing the
specific surface area of the material while hollow-cored BN-NPs porous shell structures were obtained.
They show a BET-specific surface area of 200.5 m2·g í1, a total pore volume of 0.287 cm3·g í1 and
a narrow pore size distribution centered at 3.5 nm. They were used as nano-scaffolds of ammonia
borane in order to improve its dehydrogenation properties to form a nanocomposite able to liberate
pure H2 in the temperature range 40–80 °C in our conditions. The only trace of by-product being
detected at >80 °C is ammonia. Considering the regenerability of ammonia borane [35], our results
suggest that our composite material is a safe and practical hydrogen storage material. This
improvement is exclusively ascribed to the nanoconfinement effect.
Acknowledgments
The authors acknowledge Vincent Salles for spray-pyrolysis and Arnaud Brioude for the
TEM observation of nanoparticles and samples annealed in the temperature range of 1450–2000 °C
before ball-milling.
Author Contributions
The findings in this manuscript are part of Georges Moussa thesis work. Chrystelle Salameh
performed borazine synthesis. Philippe Miele, Umit B. Demirci and Samuel Bernard advised the
thesis and directed the research. Alina Bruma and Sylvie Malo guided the TEM experiments of the
sample B1.0N0.9-NP2000BM. The manuscript was written through contributions of all authors. All
authors have given approval to the final version of the manuscript.

173
Conflicts of Interest
The authors declare no conflict of interest.
References
1.
2.

3.
4.
5.

6.

7.
8.
9.
10.
11.

12.
13.

14.
15.

Paine, R.T.; Narula, C.K. Synthetic route to boron nitride. Chem. Rev. 1990, 90, 73–91.
Wu, J.; Han, W.-Q.; Walukiewicz, W.; Ager, J.W., III; Shan, W.; Haller, E.E.; Zettl, A. Raman
spectroscopy and time-resolved photoluminescence of BN and BxCyNz nanotubes. Nano Lett.
2004, 4, 647–650.
Watanabe, K.; Tanigushi, T.; Kanda, H. Direct-bandgap properties and evidence for ultraviolet
lasing of hexagonal boron nitride single crystal. Nat. Mater. 2004, 3, 404–409.
Kubota, Y.; Watanabe, K.; Tsuda, O.; Tanigushi, T. Deep ultraviolet light-emitting hexagonal
boron nitride synthesized at atmospheric pressure. Science 2007, 317, 932–934.
Macnaughton, J.B.; Moewes, A.; Wilks, R.G.; Zhou, X.T.; Sham, T.K.; Tanigushi, T.; Watanabe, K.;
Chan, C.Y.; Zhang, W.J.; Bello, I.; et al. Electronic structure of boron nitride single crystals and
films. Phys. Rev. B 2005, 72, 195113:1–195113:8.
Zhong, W.; Wang, S.; Li, J.; Bechelany, M.C.; Ghisleni, R.; Rossignol, F.; Balan, C.; Chartier, T.;
Bernard, S.; Miele, P.; et al. Design of carbon fibre reinforced boron nitride matrix composites
by vacuum-assisted polyborazylene transfer moulding and pyrolysis. J. Eur. Ceram. Soc. 2013,
33, 2979–2992.
Balmain, W.H. Bemerkungen über die Bildung von Verbindungen des Bors und Siliciums mit
Stickstoff und gewissen Metallen. J. Prakt. Chem. 1842, 27, 422–430 (In German).
Lipp, A.; Schwetz, K.A.; Hunold, K. Hexagonal boron nitride: Fabrication, properties and
applications. J. Eur. Ceram. Soc. 1989, 5, 3–9.
Lei, W.; Portehault, D.; Liu, D.; Qin, S.; Chen, Y. Porous boron nitride nanosheets for effective
water cleaning. Nat. Commun. 2013, 4, 1777–1783.
Rousseas, M.; Goldstein, A.P.; Mickelson, W.; Worsley, M.A.; Woo, L.; Zettl, A. Synthesis of
highly crystalline sp2-bonded boron nitride aerogels. ACS Nano 2013, 7, 8540–8546.
Li, J.; Xiao, X.; Xu, X.; Lin, J.; Huang, Y.; Xue, Y.; Jin, P.; Zou, J.; Tang, C. Activated
boron nitride as an effective adsorbent for metal ions and organic polluants. Nature 2013, 3,
3208–3214.
Weng, Q.; Wang, X.; Zhi, C.; Bando, Y.; Golberg, D. Boron nitride porous microbelts for
hydrogen storage. ACS Nano 2013, 7, 1558–1565.
Siria, A.; Poncharal, P.; Biance, A.-L.; Fulcrand, R.; Blasé, X.; Purcell, S.T.; Bocquet, L. Giant
osmotic energy conversion measured in a single transmembrane boron nitride nanotube. Nature
2013, 494, 455–458.
Bernard, S. Design, Processing and Properties of Ceramic Materials from Preceramic
Precursors in Materials Science and Technologies; Nova Publishers: New York, NY, USA, 2012.
Alauzun, J.G.; Ungureanu, S.; Brun, N.; Bernard, S.; Miele, P.; Backov, R.; Sanchez, C. Novel
Monolith-type Boron Nitride Hierarchical Foams Obtained through Integrative Chemistry.
J. Mater. Chem. 2011, 21, 14025–14030.

174
16. Termoss, H.; Toury, B.; Payan, S.; Brioude, A.; Bernard, S.; Cornu, D.; Vallette, S.; Benayoun, S.;
Miele, P. Preparation of boron nitride-based coatings on metallic substrates via infrared
irradiation of dip-coated polyborazylene. J. Mater. Chem. 2009, 19, 2671–2674.
17. Stock, A.; Pohland, E. Borwasserstoffe, IX.: B3N3H6. Ber. Dtsch. Chem. Ges. 1926, 59(B),
2215–2223 (In German).
18. Wideman, T.; Sneddon, L.G. Convenient procedures for the laboratory preparation of borazine.
Inorg. Chem. 1995, 34, 1002–1003.
19. Salles, V.; Bernard, S.; Li, J.; Brioude, A.; Chehaidi, S.; Foucaud, S.; Miele, P. Design of Highly
Dense Boron Nitride by the Combination of Spray-Pyrolysis of Borazine and Additive-free
Sintering of Derived Ultrafine Powders. Chem. Mater. 2009, 21, 2920–2929.
20. Bernard, S.; Salles, V.; Li, J.; Brioude, A.; Bechelany, M.; Demirci, U.B.; Miele, P. High-Yield
Synthesis of Hollow Boron Nitride Nano-Polyhedrons. J. Mater. Chem. 2011, 21, 8694–8699.
21. Salles, V.; Bernard, S.; Chiriac, R.; Miele, P. Structural and Thermal Properties of Boron Nitride
Nanoparticles. J. Eur. Ceram. Soc. 2012, 32, 1867–1871.
22. Moussa, G.; Demirci, U.B.; Malo, S.; Bernard, S.; Miele, P. Boron Nitride Nanopolyhedrons
with hollow core@Mesoporous Shell structure: From Design to solid-state hydrogen storage
application. J. Mater. Chem. A 2014, 2, 7717–7722.
23. Shore, S.G.; Parry, R.W. The crystalline compound ammonia boraone H3NBH3. J. Am. Chem. Soc.
1955, 77, 6084–6085.
24. Demirci, U.B.; Bernard, S.; Chiriac, R.; Toche, F.; Miele, P. Hydrogen release by thermolysis of
ammonia borane NH3BH3 and then hydrolysis of its by-product [BNHx]. J. Power Sources 2011,
196, 279–286.
25. Moussa, G.; Bernard, S.; Demirci, U.B.; Chiriac, R.; Miele, P. Room-temperature hydrogen release
from activated carbon-confined ammonia borane. Int. J. Hydr. Energy 2012, 37, 13437–13445.
26. Stephens, F.H.; Pons, V.; Baket, R.T. Ammonia-Borane: The Hydrogen Source par excellence.
Dalton Trans. 2007, 2613–2626.
27. Hamilton, C.W.; Baker, R.T.; Staubitz, A.; Manners, I. B–N compounds for chemical hydrogen
storage. Chem. Soc. Rev. 2009, 38, 279–293.
28. Sanyal, U.; Demirci, U.B.; Jagirdar, B.R.; Miele, P. Hydrolysis of ammonia borane as hydrogen
source: Fundamental issues and potential solutions towards implementation. ChemSusChem.
2011, 4, 1731–1739.
29. Staubitz, A.; Robertson, A.P.M.; Manners, I. Ammonia-borane and related compounds as
dihydrogen sources. Chem. Rev. 2010, 110, 1079–1124.
30. De Jong, P.E.; Adelhelm, P. Nanosizing and nanoconfinement: New strategies towards meeting
hydrogen storage goals. ChemSusChem. 2010, 3, 1332–1348.
31. Gutowska, A.; Li, L.; Shin, Y.; Wang, C.M.; Li, X.S.; Linehan, J.C.; Smith, R.S.; Kay, B.D.;
Schmid, B.; Shaw, W.; et al. Nanoscaffold Mediates Hydrogen Release and the Reactivity of
Ammonia Borane. Angew. Chem. Int. Ed. 2005, 44, 3578–3582.
32. Gadipelli, S.; Ford, J.; Zhou, W.; Wu, H.; Udovic, T.J.; Yildirim, T. Nanoconfinement and
catalytic dehydrogenation of ammonia borane by magnesium-metal-organic-framework-74.
Chem. Eur. J. 2011, 17, 6043–6047.

175
33. Srinivas, G.; Travis, W.; Ford, J.; Wu, H.; Guo, Z.-X.; Yildirim, T. Nanoconfined ammonia
borane in a flexible metal–organic framework Fe–MIL-53: Clean hydrogen release with fast
kinetics. J. Mater. Chem. A 2013, 1, 4167–4172.
34. Peng, Y.; Ben, T.; Jia, Y.; Yang, D.; Zhao, H.; Qiu, S.; Yao, X. Dehydrogenation of Ammonia
Borane Confined by Low-Density Porous Aromatic Framework. J. Phys. Chem. C 2012, 116,
25694.
35. Sutton, A.D.; Burrell, A.K.; Dixon, D.A.; Garner, E.B., III; Gordon, J.C.; Nakagawa, T.; Ott, K.C.;
Robinson, J.P.; Vasiliu, M. Regeneration of ammonia borane spent fuel by direct reaction with
hydrazine and liquid ammonia. Science 2011, 331, 1426–1429.

176

Design of Experiments: Optimizing the
Polycarboxylation/Functionalization of Tungsten
Disulfide Nanotubes
Daniel Raichman, David Strawser and Jean-Paul Lellouche
Abstract: Design of experiments (DOE) methodology was used to identify and optimize factors
that influence the degree of functionalization (polycarboxylation) of WS2 INTs via a modified acidic
Vilsmeier–Haack reagent. The six factors investigated were reaction time, temperature and the
concentrations of 2-bromoacetic acid, WS2 INTs, silver acetate and DMF. The significance of
each factor and the associated interactive effects were evaluated using a two-level factorial statistical
design in conjunction with statistical software (MiniTab® 16) based on quadratic programming.
Although statistical analysis indicated that no factors were statistically significant, time, temperature
and concentration of silver acetate were found to be the most important contributors to obtaining
maximum functionalization/carboxylation. By examining contour plots and interaction plots, it was
determined that optimal functionalization is obtained in a temperature range of 115–120 °C with a
reaction time of 54 h using a mixture of 6 mL DMF, 200 mg INTs, 800 mg 2-bromoacetic
acid and 60 mg silver acetate.
Reprinted from Inorganics. Cite as: Raichman, D.; Strawser, D.; Lellouche, J.-P. Design of
Experiments: Optimizing the Polycarboxylation/Functionalization of Tungsten Disulfide Nanotubes.
Inorganics 2014, 2, 455–467.
1. Introduction
Design of experiments (DOE) [1,2] is a widely used discipline applied in a variety of areas,
including engineering [3–5], social sciences [6] and natural sciences [7–9]. Design of experiments is
a powerful statistical methodology in its own right, with a number of software applications readily
available to aid researchers in both designing and globally optimizing multi-parametric experiments
to achieve the best results through analysis and interpretation. In this context, combining
software-based applications with a researcher’s experience and scientific intuition is a powerfully
growing trend that typically results in significant savings in time and materials. The design of
experiments methodology includes formal, planned experimentation with the goal of optimizing
a set of reaction parameters that may disclose synergism between reaction parameters. The
optimization typically comprises six main steps: (1) selection of variables and defining their
range of variation; (2) selection of responses; (3) experimental design selection; (4) performing the
designed experiments in random order; (5) determination of coefficients in a mathematical model;
and (6) predicting the response and evaluating the model relevance. In this context, we designed
a series of experiments to optimize globally the reaction conditions for maximizing the yield of
covalently functionalized tungsten disulfide inorganic fullerene-like nanotubes (WS2 INTs). This
specific functionalization reaction comprises a polycarboxylation technique [10,11], developed
recently in our laboratories, that uses a modified highly electrophilic Vilsmeier–Haack reagent [12].

177
It is well known that classical Vilsmeier–Haack reactions use DMF (secondary N-formyl amine)
and POCl3/SOCl2 to effect the formylation of a wide range of electrophilic substrates via intermediate
electrophilic iminium salts of Type A (Scheme 1a). Such Vilsmeier–Haack reactions have been
studied extensively and found to be quite versatile, leading to a number of oxygen and nitrogen
heterocycles [13–18], as well. For example, a modified Vilsmeier–Haack reagent that uses
ethyl chloroformate in place of POCl3 was found useful when reacted with active methylene
compounds [19]. Analogously, we have discovered that by using a mixture containing DMF (a 2nd
N-CHO amine) and O-alkylating 2-bromoacetic acid with catalysis by Ag(I)OAc, WS2 nanotubes
may be polycarboxylated readily and quite effectively according to the mechanism described in
Scheme 1c. Indeed, DMF removal or replacement with other polar, non-protic materials/solvents
(e.g., 1,4-dioxane, DME, etc.) leads to unsuccessful polycarboxylation. This strict requirement of
DMF (2nd N-CHO amine) as an essential component of the reaction mixture led us to propose
and detail a corresponding Vilsmeier–Haack-like reaction mechanism, displayed in Scheme 1c.
Interestingly, silver acetate (Ag(I)OAc) was also included as an essential reaction factor due to
its well-known ability to chemically trap halogens and, thereby, assist in halogen (Br) abstraction.
Depicted in Scheme 1b is the reaction of Ag(I)OAc with the Br atom of bromoacetic acid, leading
to the formation of the Vilsmeier–Haack complex of Type B1 with subsequent precipitation
of Ag(I)Br.
Because several reaction factors are involved in the polycarboxylation reaction mentioned above,
we selected a DOE methodology as the most economical means of optimizing the reaction factors
to produce the highest yields of polycarboxylation. In this context, the level of polycarboxylation
was determined indirectly by reacting the functionalized INTs with an excess of 1,3-diaminopropane,
resulting in a terminal primary amine that was quantified by the Kaiser test [20]. Due to a one-to-one
reaction between the diamine and carboxylic acid, the amount of terminal amine is equal to the amount
of carboxylic acid.
In addition to the Kaiser test; further confirmation of the successful functionalization of the
WS2 INTs was obtained by a combination of FT-IR; TGA and zeta potential analyses (see the
Supplementary Information).
Thus far, we have investigated this unique functionalization method only for the
polycarboxylation of WS2 nanotubes. However, we strongly believe it may prove applicable for
similar polycarboxylation of other layered dichalcogenide nanomaterials (MS(Se)2, M = Mo, Sn).

178
Scheme 1. (a) Generalized Vilsmeier–Haack generation of electrophilic iminium
salts; (b) Vilsmeier–Haack generation of DMF-based electrophilic iminium salt
complex with bromophilic Ag(I) assistance in halide abstraction; (c) sulfur-mediated
nucleophilic addition of the iminium salt complex to the outermost sulfur atoms of
WS2 INTs, producing the corresponding polycarboxylated INTs.

(a)

(b)

(c)
2. Results and Discussion
Because the level of surface functionalization of the WS2 INTs by the Vilsmeier–Haack-like
reaction is of critical importance in determining the coordination capability of the corresponding
optimally surface-engineered WS2 INTs, a statistically designed experiment was implemented using
the design of experiment (DOE) methodology. The goal of the design was to disclose an optimal set
of reaction conditions that would result in the maximized level of surface functionalization of the WS2
INTs. This DOE study enables varying more than one factor/reaction condition at a time for process
optimization, even when several influential factors are involved. In addition, this multi-factor
approach not only enables running fewer experiments, but enables the study of the interactions
between the reaction factors and how these interactions influence the final result. These advantages
are unavailable with the more commonly used one factor at a time (OFAT) optimization methods.

179
Based on our current process knowledge, four factors (reaction parameters) likely to affect
significantly the functionalization process were identified: DMF, silver acetate (Ag(I)OAc), WS2
INTs and 2-bromoacetic acid (BrCH2COOH). In addition, two other factors, time and temperature,
were included in the design. The experimental design and subsequent analysis of the significance
of each factor and associated interactive effects were performed using a two-level factorial statistical
design in conjunction with statistical software (MiniTab® 16) based on quadratic programming. Pareto
analysis (Figure 1) was used to disclose which reaction parameters are process active. Using the software
default value for Į of 0.05, all absolute magnitude effect values fall below the software error-calculated
reference line for statistical significance (vertical red line, value 0.3361), indicating that none of
the factors nor interactive effects are statistically significant. However, the software analysis
indicates that time, temperature and the amount of silver acetate are the three most important
contributors to obtaining maximum functionalization, and these warrant attention in the subsequent
analyses. From the prediction model that was obtained, an optimal functionalization yield was obtained
(1.21 mmol COOH groups/g) at an optimal temperature of 120 °C and a reaction time of 54 h.
Figure 1. Pareto chart of the effects of the factors on the amount of functionalization
obtained. Values to the left of the reference line (vertical red line) are not
statistically significant.

Figure 2 displays the effect of the two factors, temperature and time, on the amount of
functionalization obtained per gram of INTs. Clearly, a higher temperature (120 °C) for a longer
time (54 h) results in the highest level of functionalization.
To assess the effect of each of the other factors together with temperature on the amount
of functionalization obtained when the reaction is conducted for 54 h, a series of contour plots was
made. For several of the factors, two combinations of temperature and factor (WS2 INTs,
BrCH2COOH, DMF or AgOAc) concentration produce the highest amount of carboxylation. When
a higher temperature is used, the factor under evaluation does not affect the yield of carboxylation.
However, in a lower temperature range, the highest amount of carboxylation is achieved only
when the factor concentration is within a specific range.

180
Figure 2. Amount (mmol) of carboxylic acid functional groups found per gram of
WS2 INTs. Results are grouped by temperature and time. A higher temperature and
a longer time increase the level of functionalization.

Figure 3 displays the effect of the dual factors, temperature (°C) and quantity (mg) of INTs, on
the reaction mixture.
Figure 3. Contour plot of the amount (mmol) of carboxylation per gram of INTs
obtained as the dual factors, temperature and amount of INTs in the reaction mixture,
are varied at a constant reaction time of 54 h.

Two areas are revealed in which maximum carboxylation (0.4–0.6 mmol/g) can be obtained at
a 54-h reaction time; by using 200–300 mg of INTs in the mixture with the temperature maintained
in the range of 105–120 °C and using 200–600 mg INTs with a temperature range of 115–120 °C.
Interestingly, a slightly lower amount of carboxylation (0.2–0.4 mmol/g) can be obtained using an
INT level in the range of 200–600 mg with a decrease of temperature in a range of 80–100 °C. This
suggests that it may be possible to customize the yield of carboxylation by tuning the amount of
INTs in the reaction mixture together with the temperature.
Figure 4 displays the effect of the dual factors, temperature (°C) and quantity (mg, mmol)
of 2-bromoacetic acid in the reaction mixture. Two areas are revealed in which maximum carboxylation

181
(0.4–0.6 mmol/g) can be obtained at a 54-h reaction time; by using 800–1100 mg (5.76–7.92 mmol)
of 2-bromoacetic acid in the mixture with temperature maintained in a range of 105–120 °C and
using 800–2250 mg (5.76–16.2 mmol) of 2-bromoacetic acid with a temperature range of 115–120 °C.
As mentioned above, with the INT concentration, by tuning the amount of 2-bromoacetic acid
(800–2250 mg, 5.76–16.2 mmol) and the temperature (80–100 °C), it may be possible to produce
a controlled level of carboxylation within the range of 0.2–0.4 mmol/g.
Figure 4. Contour plot of the amount (mmol) of carboxylation per gram of INTs
obtained as the dual factors, temperature and amount of 2-bromoacetic acid in the
reaction mixture, are varied at a constant reaction time of 54 h.

Figure 5 displays the effect of the dual factors, temperature (°C) and quantity (mL) of DMF in
the reaction mixture. Two areas are revealed in which maximum carboxylation (0.4–0.6 mmol/g)
can be obtained at a 54-h reaction time; by using 2–2.5 mL of DMF in the mixture with
temperature maintained in a range of 110–120 °C, and using 2–6 mL DMF with a temperature
range of 115–120 °C. As mentioned above, for both INT and the 2-bromoacetic acid concentration,
by tuning the amount of DMF (2–6 mL) and temperature (80–100 °C), it may be possible to produce
a controlled level of carboxylation within the range of 0.2–0.4 mmol/g.
Figure 6 displays the effect of the dual factors, temperature (°C) and quantity (mg) of AgOAc in
the reaction mixture. Three zones of carboxylation yield are clearly identified, indicating that
several sets of AgOAc concentration and temperature can be used to achieve the selected level of
carboxylation. For example, maximum carboxylation (0.6–0.8 mmol/g) can be obtained at a 54-h
reaction time by using AgOAc in a range of 55–60 mg (0.33–0.36 mmol) in the mixture with
temperature maintained in a range of 115–120 °C. The following combinations of AgOAc/temperature
produce a lower carboxylation level of 0.4–0.6 mmol/g: 60 mg (0.36 mmol)/100–110 °C;
50–60 mg (0.30–0.36 mmol)/105–110 °C; 40–50 mg (0.24–0.30 mmol)/115–120 °C. In addition, a
carboxylation level of 0.2–0.4 can be achieved over the entire temperature range of 80–120 °C if
the AgOAc level is maintained within 20–35 mg (0.12–0.21 mmol). Similarly, this lower
carboxylation level is achievable over the entire range of 20–60 mg (0.12–0.36 mmol) AgOAc
by maintaining a lower temperature range of 80–100 °C.

182
Figure 5. Contour plot of the amount (mmol) of carboxylation per gram of INTs
obtained as the dual factors, temperature and amount of DMF in the reaction mixture,
are varied at a constant reaction time of 54 h.

Figure 6. Contour plot of the amount (mmol) of carboxylation per gram of INTs
obtained as the dual factors, temperature and amount of AgOAc in the reaction mixture,
are varied at a constant reaction time of 54 h.

Figure 7 displays the effect on carboxylation yield with the dual factors, temperature (°C) and
time (h). For this factor combination, four zones of carboxylation yield are identified. As noted with
the factors temperature and AgOAc, several sets of the two factors can be used to achieve a
selected level of carboxylation. As examples for achieving the highest carboxylation level
(0.6–0.8 mmol/g), a time range of 45–60 h at a temperature of 120 °C or a temperature in a range
of 110–120 °C for 54 h can be used. Similar analyses can be done to find ranges of temperature
and time to produce the lower carboxylation levels displayed in Figure 7.

183
Figure 7. Contour plot of the amount (mmol) of carboxylation per gram of INTs
obtained as the dual factors, temperature (°C) and time (h), are varied.

Analysis of the contour plots for a 54-h reaction time (Figures 3–7) indicates that only the
temperature range of 115–120 °C is common to all of the factors to produce the maximum
degree of carboxylation. Using this temperature range, the levels of each factor to achieve
maximum carboxylation are as follows. INTs 200–600 mg; 2-bromoacetic acid 800–2250 mg
(5.76–16.2 mmol); DMF 2–6 mL and AgOAc 55–60 mg (0.33–0.36 mmol).
Figure 8. Interaction plot of the effects of the factors on the degree of carboxylation
at a reaction time of 54 h.

The influence of the factors on the degree of carboxylation was examined by interaction
plots. Separate interaction plots were made for reaction times of 18 h and 54 h. Figure 8 displays
the interaction plot for a 54-h reaction time. The temperature column indicates that a reaction

184
temperature of 80 °C results in not only a much lower amount of carboxylation than a reaction
temperature of 120 °C, but, at the levels tested, changing the concentrations of the four factors, DMF,
INT, 2-bromoacetic acid and AgOAc, has little effect on the degree of carboxylation. In contrast,
at 120 °C, higher concentrations of DMF (6 mL) and AgOAc (60 mg, 0.36 mmol) with lower
concentrations of INTs (200 mg) and 2-bromoacetic acid (800 mg, 5.76 mmol) produce higher degrees
of carboxylation. Note that this analysis is in agreement with the findings of the contour plots
(Figures 3–7).
The AgOAc column shows similar effects, with the lower concentration of AgOAc producing
a lower amount of carboxylation than a higher concentration, and changing the concentration of
the other factors has little effect on the degree of carboxylation. When a higher concentration of
AgOAc is used, the degree of carboxylation is increased with a higher DMF concentration, but
with lower concentrations of INTs and 2-bromoacetic acid. The 2-bromoacetic acid column indicates
that a higher degree of carboxylation is obtained using a lower concentration of acid along with
a lower concentration of INTs and a higher concentration of DMF. Similarly, the WS2 INT column
indicates that using a lower concentration of INTs and a higher concentration of DMF produces a
higher degree of carboxylation.
Figure 9 displays the interaction plot for an 18-h reaction time.
Figure 9. Interaction plot of the effects of the factors on the degree of carboxylation
at a reaction time of 18 h.

The temperature column indicates that a reaction time of 18 h with optimized factors of 120 °C,
2 mL DMF, 60 mg (0.36 mmol) AgOAc, 200–600 mg INTs, and 800–2400 mg (5.76–17.3 mmol)
2-bromoacetic acid gives a maximum amount of carboxylation that is less than half the amount that
is obtained with optimized factors when the reaction time is 54 h.

185
3. Experimental Section
3.1. Materials
Tungsten disulfide nanotubes (WS2-INTs) were provided by NanoMaterials Ltd. (Yavne, Israel).
All reagents and solvents were purchased from commercial sources and used without further purification.
3.2. Methods
Carboxylated WS2 INTs
The carboxylated INTs were prepared using the conditions in Table 1. Details for Sample 1
are presented as a representative example.
Table 1. Sample parameters for the carboxylation reaction.
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

DMF
mL
2
6
2
6
2
6
2
6
2
6
2
6
2
6
2
6
4

WS2 INTs
mg
200
200
600
600
200
200
600
600
200
200
600
600
200
200
600
600
400

2-Bromoacetic acid
mg
800
800
800
800
2400
2400
2400
2400
800
800
800
800
2400
2400
2400
2400
1600

AgOAc
mg
20
20
20
20
20
20
20
20
60
60
60
60
60
60
60
60
40

Temp
°C
80
120
120
80
120
80
80
120
80
120
120
80
120
80
80
120
100

Time
h
18
18
54
54
54
54
18
18
54
54
18
18
18
18
54
54
36

Carboxylation
mmol/g
0.150
0.241
0.430
0.318
0.519
0.278
0.151
0.241
0.444
1.212
0.359
0.100
0.406
0.240
0.349
0.681
0.346

To a solution of 2-bromoacetic acid (0.80 g, 5.76 mmol) in DMF (2 mL) was added AgOAc
(20 mg, 0.12 mmol) and WS2 INTs (200 mg). The mixture was heated in an oil bath to 80 °C
and stirred for 18 h. After cooling to room temperature, the mixture was centrifuged and the
supernatant discarded. The solids were washed with ethanol followed by centrifugation (11,000 rpm
5 min) 5 times and dried under vacuum to obtain carboxylated WS2 INTs containing 0.150 mmol
carboxylic acid per gram of INT.
Quantification of the carboxylic acid was done indirectly by coupling 1,3-diaminopropane
to the carboxylic acid using EDC as a coupling agent, followed by the Kaiser test for terminal
amines [10,13].

186
4. Conclusions
By using the design of experiments (DOE) methodology, six reaction factors were identified
and optimized to maximize the degree of functionalization of inorganic WS2 INTs with a carboxylic
acid shell via a modified highly electrophilic Vilsmeier–Haack reaction. Initial studies produced
functionalized INTs with 0.5 mmol COOH groups per gram of INT as quantified by the Kaiser
test. Moreover, a statistically relevant DOE global optimization resulted in functionalization increasing
by a factor of 2.4 to 1.2 mmol of accessible COOH groups per gram of chemically-modified INTs. Time,
temperature and amount of silver acetate (Ag(I)OAc) were found to be the most important factors
that affect the functionalization yield.
In addition, contour plots of interacting reaction parameters suggest that it may be possible
to control the amount of polyCOOH functionalization by tuning the amounts of INTs, 2-bromoacetic
acid, DMF and Ag(I)OAc in the reaction mixture in conjunction with reaction temperature.
This attractive capability may prove valuable for optimizing the use of functionalized WS2 INTs in
important applications, such as functional nanoscale fillers for the mechanical reinforcement of
polymeric matrices with optimal interfacial phase interactions.
Supplementary Information
1. FT-IR
FT-IR spectra were recorded on an FT-IR Tensor 27 spectrometer (Bruker) using ATR.
Figure S1 displays the FT-IR spectrum of carboxylated WS2 INTs with peaks of interest marked.
The peaks can be assigned as follows (wavenumbers Q in cmí1). 3367, O-H stretch; 2881, C-H stretch;
1747, C=O stretch of carboxylic acid; 1467, C-H bend; 1360, C-H rock; 1039, C=S stretch (indicative
of covalent attachment to sulfur of the INTs); 946, O-H of carboxylic acid.
Figure S1. FT-IR spectrum of carboxylated WS2 INTs.

187
2. Thermogravimetric Analysis
Thermogravimetric analysis was performed on a TA Q600-0348, model SDT Q600
(Thermofinnigan) using a temperature profile of 25–800 °C at 10 °C/min under N2 flow (10 mL/min)
with sample masses ranging from 5-12 mg.
The TGA analysis of the inorganic nanotubes before and after carboxylation is displayed in Figure
S2. Weight losses were 1.39% and 17.12% for the untreated and carboxylated WS2 INTs respectively,
indicating that organic material is attached to the surface of the treated nanotubes.
Figure S2. TGA plots of WS2 INTs.

3. Zeta Potential
Nanoparticle surface charge (ȟ potential) was determined using a Zetasizer Nano-ZS (Malvern
Instruments Ltd. Worcestershire, UK) in water at 25 °C and 150 V. Dispersions were prepared with
an ElmaSonic S30 sonicator (Elma GmbH & Co., Singen, DE).5 mg of WS2 INTs or carboxylated
WS2 INTs were dispersed in 15 ml ultra-pure H2O (resitivity >18 Mohm•cm) by sonication for one
minute. The Zeta potential was measured immediately after sonication. The analysis showed -27.8
mV and -17.3 mV for the carboxylated and untreated WS2 INTs, respectively.
Acknowledgments
We thank NanoMaterials Ltd. for their generous gift of WS2 nanotubes and the Israel National
Nanotechnology Initiative Focal Technology Area proposal Inorganic Nanotubes: From
Nanomechanics to Improved Nanocomposites, (Reshef Tenne, program coordinator) for partial
funding of this research.
Author Contributions
Daniel Raichman performed most of the experimental work, prepared drafts of the experimental
section, prepared the initial images and researched the design of experiments. David Strawser assisted
with experimental procedures and design of experiments processing, finalized images and edited much

188
of the manuscript. Jean-Paul Lellouche initiated this WS2 INTs functionalization project, provided full
research assistance in the selection of the influential factors, checked the entire manuscript and
provided all of the necessary budget, laboratory facilities and intellectual support for conducting
the corresponding research work.
Conflicts of Interest
The authors declare no conflict of interest.
References
1.
2.
3.

4.

5.

6.
7.
8.
9.
10.
11.

12.

13.

Montgomery, D.C. Design and Analysis of Experiments, 6th ed.; Wiley: New York, NY, USA, 2005.
Engineering Statistics Handbook. Available online: http://www.itl.nist.gov/div898/
handbook/pri/section1/pri11.htm (accessed on 20 February 2014).
Castillo, P.; Magnin, J.P.; Velasquez, M.; Willisom, J. Modeling and Optimization of
Hydrogen Production by the Methodology of Design of Experiments (DOE): Interaction
between Lactate Concentration and Light Luminosity. Energy Procedia 2012, 29, 357–366.
Imandi, S.B.; Bandaru, V.V.R.; Somalanka, S.R.; Bandaru, S.R.; Garapati, H.R. Application
of Statistical Experimental Designs for the Optimization of Medium Constituents for the
Production of Citric Acid from Pineapple Waste. Bioresour. Technol. 2009, 99, 4445–4450.
Varrone, C.; Giussani, B.; Izzo, G.; Massini, G.; Marone, A.; Signorini, A.; Wang, A. Statistical
Optimization of Biohydrogen and Ethanol Production from Crude Glycerol by Microbial
Mixed Culture. Int. J. Hydrog. Energy 2013, 37, 16479–16488.
Schneider, S. Experimental Design in the Behavioral and Social Sciences; SAGE Publications:
Thousand Oaks, CA, USA, 2013.
Mandenius, C.K.; Brundin, A. Bioprocess Optimization Using Design-of-Experiments
Methodology. Biotechnol. Prog. 2008, 24, 1191–1203.
Yates, F. Imperial Bureau of Soil Science Technical Communication No. 35: The Design
and Analysis of Factorial Experiments; Imperial Bureau of Soil Science: London, UK, 1937.
Pilat, P.; Votruba, J.; Dobersky, P.; Prokop, A. Application of Mathematical Optimization
Methods in Microbiology. Folia Microbiol. 1976, 21, 391–405.
Raichman, D.; Strawser, D.A.; Lellouche, J.P. Covalent Functionalization/Polycarboxylation
of Tungsten Disulfide Inorganic Nanotubes. Nanoscale 2014, submitted for publication.
Lellouche, J.-P. Versatile chemical Modificaions of Hard Fullerene-Like Inorganic Tungsten
Disulfide (WS2) Nanotubes (WS2 INTs). U.S. Provisional Application for Patent No. 61/865,
15 1, 13 August 2013.
Vilsmeier, A.; Haack, A. Uber die Einwirkung von Halogen Phosphor auf Alkyl-formanilide.
Ein neue Methode zur Darstellung sekundarer und tertiarer p-Alkylamino-benzaldehyde.
Chem. Berichts 1927, 60, 119–122 (In German).
Sreenivasulu, M.; Rao, K.G.S. Vilsmeier reaction studies on some Į,ȕ-unsaturated alkenones.
Indian J. Chem. Sect. B 1989, 28, 594–595.

189
14. Lilienkampf, A.; Johansson, M.P.; Wahala, K. (Z)-1-Aryl-1-haloalkenes as Intermediates in
the Vilsmeier Haloformylation of Aryl Ketones. Org. Lett. 2003, 5, 3387–3390.
15. Ali, M.M.; Tasneem; Rajanna, K.C.; Saiprakash, P.K. An efficient and facile synthesis of
2-chloro-3-formyl quinolines frome acetanilides in micellar media by Vilsmeier–Haack
cyclisation. Synlett 2001, 2, 251–253.
16. Quiroga, J.; Trilleras, J.; Galvez, J.; Insuasty, B.; AbonÕa, R.; Nogueras, M.; Cobo, J.; Marchal, A.
C- and N-cyanoacetylation of 6-aminopyrimidines with cyanoacetic acid and acetic anhydride.
Tetrahedron Lett. 2008, 49, 5672–5675.
17. Giles, C.M.M.A.P.R. Synthesis Using Vilsmeier Reagents. New Directions in Organic and
Biological Chemistry, 1st ed.; CRC Press: Boca Raton, FL, USA, 1994.
18. Rajanna, K.C.; Ali, M.M.; Solomon, F.; Saiprakash, P.K. Vilsmeier-Haack Formylation of
Coumarin Derivatives. A Solvent Dependent Kinetic Study. Int. J. Chem. Kinet. 1996, 28,
865–872.
19. Ikawa, K.; Takami, F.; Fukui, Y.; Takuyama, K. A Novel Reagent for the Vilsmeier–Haack
Reaction. Tetrahedron Lett. 1969, 38, 3279–3281.
20. Kaiser, E.; Colescott, R.L.; Bossiriger, C.D.; Cook, P.I. Color Test for the Detection of
Free Terminal Amino Groups in the Solid-phase Synthesis of Peptides. Anal. Biochem. 1970, 34,
595–598.

190

Noble-Metal Chalcogenide Nanotubes
Nourdine Zibouche, Agnieszka Kuc, Pere Miró and Thomas Heine
Abstract: We explore the stability and the electronic properties of hypothetical noble-metal
chalcogenide nanotubes PtS2 , PtSe2 , PdS2 and PdSe2 by means of density functional theory
calculations. Our findings show that the strain energy decreases inverse quadratically with the
tube diameter, as is typical for other nanotubes. Moreover, the strain energy is independent of
the tube chirality and converges towards the same value for large diameters. The band-structure
calculations show that all noble-metal chalcogenide nanotubes are indirect band gap semiconductors.
The corresponding band gaps increase with the nanotube diameter rapidly approaching the respective
pristine 2D monolayer limit.
Reprinted from Inorganics. Cite as: Zibouche, N.; Kuc, A.; Miró, P.; Heine, T. Noble-Metal
Chalcogenide Nanotubes. Inorganics 2014, 2, 556–564.

1. Introduction
Inorganic nanotubes (INTs) are a class of materials that are very attractive for many applications
in nanotechnology due to their interesting physical and chemical properties, which arise from their
low dimensionality. In 1930, Pauling had already stated that mismatched layered materials may
form cylindrical shapes [1]. However, the first carbon-based tubular forms, namely carbon nanotubes
(CNTs) have been observed by Iijima in 1991 [2]. At the same time, the synthesis of WS2 NTs by
Tenne and co-workers [3], has demonstrated that these tubular systems are not limited to carbon, but
can also be obtained from any other layered compound. Later on, several INTs have been prepared
and produced such as transition-metal sulfides [3,4]. boron-nitrides [5,6], metal oxides [7,8], rare
earth oxide [9] and more recently misfit layered compounds [10].
WS2 and MoS2 NTs, being the first synthesised INTs, are semiconductors. They have
demonstrated excellent mechanical properties [11–16] and are known to be good solid lubricants [17].
They have also been suggested as scanning probe tips [18], catalysts [19], reinforcements for
composite materials [20], photo-transistors [21], gas storage and host materials [22,23], etc. Later,
other transition-metal chalcogenide (TMC) NTs have been reported such as TiS2 , NbS2 , ReS2 , TiSe2
and TaS2 [24–28]. Subsequently, many more TMC NTs can be expected due to the large manifold
of the layered TMC materials [29,30]. Nowadays, different techniques and strategies have been
employed and developed for the synthesis and growth of large amount of NTs such as chemical
transport technique [31], thermochemical decomposition [32] and in situ heating [33]. For example,
WS2 and MoS2 NTs were produced using gas-solids reactions at high temperatures by the reduction
of WO3 (MoO3 ) in the atmosphere of a mixture of H2 , N2 and H2 S gases [3,4,34].
In this work, we aim to extend the scope of inorganic nanotubular materials by investigating
noble-metal chalcogenide M X 2 single wall nanotubes, where M = Pt, Pd and X = S, Se. Tubular
forms based on these materials have not yet been experimentally observed, however, considering that

191
PtS2 , PtSe2 , PdS2 , PdSe2 belong to the family of layered TMCs, one can expect that such compounds
may also form nanotubes. Structure and electronic structure of these noble metal chalcogenides have
been subject to controversial debate in the 1950s and 1960s. Two possible phases of the Pd- and
Pt-based TMCs were suggested, namely orthorhombic (pyrite) and 1T . [35–38]. We have calculated
the relative stability of bulk and monolayered forms of these TMCs. While the orthorhombic phase
is preferable for the bulk PdS2 material (energy difference of 56 meV - all energies are given per
M X2 formula unit), the monolayers (MLs) favor the 1T arrangement (energy difference of 68
meV). For the heavier PdSe2 , we have found similar trends for the MLs, where the 1T form is
by 0.6 eV more stable than the orthorhombic one. Bulk PdSe2 is unstable in the orthorhombic form,
forbidding comparison with the 1T structure. Therefore, the choice of the 1T polytype to simulate
the single-wall NTs is justified. In view of the renaissance of layered materials and the advance of
experimental technology it is important to reexamine these phases and attempt exfoliation.
A recent study of bi- and monolayered noble-metal materials have shown interesting
quantum confinement effects and electromechanical properties, suggesting them for applications in
optoelectronics and flexible devices [39]. Therefore we have investigated, using density functional
theory (DFT), the stability of the M X 2 NTs and their structural and electronic properties. The
strain energy is found to be chirality independent and exhibits the characteristic dependence on the
tube diameter d (∼1/d2 ). The band structure analysis shows that noble-metal chalcogenide NTs are
all semiconducting in a similar way to their ML counterparts. Unlike MoS2 and WS2 NTs, where
the band gap is direct and indirect for zigzag and armchair, respectively, PtX2 and PdX2 NTs have
indirect band gaps which increase with the diameter.
2. Computational Details
All calculations were carried out using density functional theory with the PBE
(Perdew-Burke-Ernzerhof) exchange-correlation functional [40] as implemented in the Crystal09
code [41]. We have used an all electron triple-zeta valence basis set with one polarization function
for sulphur atoms [42], a scalar-relativistic pseudopotential with 18 valence electrons for platinum
atoms [43], Hay and Wadt effective core potentials with small core for palladium atoms [41], and
the relativistic multi-electron pseudopotential with six valence electrons for selenium atoms [44].
We have employed helical boundary conditions, as implemented in the Crystal09 code [41], for
the generation of the NT structures in order to reduce computational costs. Lattice vectors and
atomic positions of M X 2 MLs and NTs were fully optimized. The tube diameters considered here
are in the range of 10–50 Å, corresponding to chiral indices (10,0)–(32,0) and (6,6)–(24,24) for
zigzag and armchair NTs, respectively. The shrinking factors of 16 for MLs and 8 for NTs were
used, resulting in 30 and 5 k-points in the irreducible Brillouin zone, respectively, following the
Monkhorst-Pack sampling [45]. Band structures were calculated along the high symmetry k-points
using the M–Γ–K–M and X–Γ paths for MLs and NTs, respectively.

192
3. Results and Discussion
The monolayered noble-metal chalcogenides, considered here as large diameter NT limits, were
adopted in the 1T polymorph, with space group (P-3m1). Figure 1 shows the 1T geometry for a
monolayered structures in the left side, the top and the side views of zigzag and armchair NTs are
represented in the middle and the right parts, respectively. The optimized lattice parameters and
bond lengths with respect to the tube diameter are shown in Figure 2. Increasing the tube diameter
d, the lattice vectors of the tubes decrease for both zigzag and armchair chiralities. For (n,n) NTs,
the lattice vectors correspond to those of the MLs; however, due to the curvature they enlarge as the
diameters become smaller. Similar behaviour is observed for the bond lengths between metal and
chalcogen atoms (M –X), nevertheless, the convergence to the ML limit is much slower. Generally,
the bond lengths and lattice vectors of selenide NTs are larger than those of sulphide NTs, and the
same holds for the comparison of the platinum over the palladium forms.
The M –X bond lengths can be divided into two types, the inner and the outer wall bond lengths,
which are referred hereafter as M –X i and M –X o , respectively. The M –X o (M –X i ) are longer
(shorter) than the corresponding bond lengths found in the ML structures, and the deviations are
more pronounced for armchair NTs than for zigzag NTs, particularly for small d. For diameters
below 15 Å, the difference between outer and inner bond reaches 0.40 Å and 0.25 Å for armchair and
zigzag NTs, respectively. These numbers strongly reduce to about 0.10 Å difference for diameters
of at least 40 Å. Recently, we have reported that the bond lengths of MoS2 and WS2 NTs exhibit
opposite behaviour, where the zigzag NTs have longer (shorter) M –X o (M –X i ) bond lengths than
the armchair NTs [46].
Figure 1. 1T 2D (monolayer) and 1D (tubular) forms of noble metal chalcogenides. The
unit cell of 2D systems is shown.

193
Figure 2. Lattice parameters and bond lengths vs. diameter of M X 2 NTs.
(n,0)

2.70

----- ML Pt-X
----- ML Pd-X

6.6
PtS2
PtSe2
PdS2
PdSe2

2.55

M-X / Å

a / Å

6.4

6.2
3.9

(n,n)

MS2

2.40

(n,0) M-Xi
(n,0) M-Xo
(n,n) M-Xi
(n,n) M-Xo

2.85

3.8
PtS2 / ML
PtSe2 / ML
PdS2 / ML
PdSe2 / ML

3.7

MSe2

2.70

3.6
2.55
3.5
10

20

30

40

50

10

20

30

d /Å

40

50

d /Å

The stability of NTs can be expressed by the strain energy, EStrain , which is the difference of total
energy per atom of the tube and the respective ML. Generally, the strain energy of NTs is correlated
to the tube diameter through the archetypal relation EStrain ∼ 1/d2 . The calculated EStrain of M X 2
NTs with respect to their diameters (see Figure 3) follow the same dependence, where the strain
energies decrease quadratically with the diameter and converge to the same value for all systems.
The curves EStrain (d) were fitted to the equation EStrain = C/d2 with correlation coefficients greater
than 0.999 and the values of coefficients C are given in the Table 1 for each system. We note that
strain energies of the PdX 2 NTs are smaller than for PtX 2 NTs for small diameters, this means that
PdX 2 NTs are more stiff than PtX 2 NTs in that range. Furthermore, the strain energy of noble-metal
chalcogenide tubes is the same for both zigzag and armchair chiralities, whereas for MoS2 and WS2
counterparts, armchair NTs are more stable than zigzag NTs, especially for large diameters [46]. In
addition, the coefficients C and the strain energy values of the noble-metal chalcogenide NTs are
smaller than those of MoS2 and WS2 NTs. This means that noble-metal chalcogenide NTs are more
favorable and easier to form than MoS2 and WS2 NTs.
Figure 3. Strain energy vs. diameter of M X 2 nanotubes.

(n,0) PtS2
(n,n) PtS2
(n,0) PdS2
(n,n) PdS2

0.20
0.15
0.10
0.05
0.00

(n,0) PtSe2
(n,n) PtSe2
(n,0) PdSe2
(n,n) PdSe2

0.25

EStrain / eV

EStrain / eV

0.25

0.20
0.15
0.10
0.05

10

15

20

25

30

d/Å

35

40

45

50

0.00

10

15

20

25

30

d/Å

35

40

45

50

194
Table 1. Coefficients C ( in eV Å2 ) of the fitted curves EStrain = C/d2 .
System

Zigzag (n,0)

Armchair (n,n)

PtS2
PtSe2
PdS2
PdSe2
MoS2 [46]
WS2 [46]

36.60
42.47
29.71
32.50
57.50
58.14

35.71
41.22
28.46
31.38
50.90
59.68

We have also investigated the electronic structure of these noble-metal chalcogenide systems.
The M X 2 MLs are found to be semiconducting with indirect band gaps of 1.26 eV, 1.75 eV,
0.74 eV and 1.38 eV for PdS2 , PtS2 , PdSe2 , PtSe2 , respectively. This is in agreement with the
results of Miró et al. [39] for these MX2 MLs, where the obtained band gaps are 1.11 eV, 1.75 eV,
0.39 eV and 1.05 eV for PdS2 , PtS2 , PdSe2 and PtSe2 , respectively, using similar level of theory.
The band gaps Δ versus the tube diameter of all M X 2 NTs are plotted in Figure 4. Similar to
MoS2 and WS2 NTs, these band gaps of M X 2 NTs increase with the diameter and approach the
band gaps of their respective MLs. For small d, the band gaps of armchair NTs become larger than
zigazag NTs for all systems. Substituting Pd with Pt causes an increase in Δ, while replacing S by Se
decreases it. The band structures of noble-metal M X 2 MLs and NTs are depicted in Figure 5. Unlike
MoS2 and WS2 MLs, where the band gaps are direct with values of 1.9 and 2.1 eV, respectively [47],
the noble-metal MLs are all indirect bandgap semiconductors. This difference could be understood
in the electronic configuration of the metal elements, as well as, in the 2H and 1T symmetries
of MoS2 /WS2 and the noble-metal chalcogenides, respectively. The M X 2 chalcogenide NTs also
exhibit indirect band gaps, for both zigazags and armchairs and for all systems, whereas MoS2 and
WS2 NTs show direct and indirect band gaps for zigzag and armchair forms, respectively [46].

2.0

2.0

1.8

1.8

1.5

1.5

1.2

1.2

Δ / eV

Δ / eV

Figure 4. Band gap vs. diameter of M X 2 nanotubes.

1.0

(n,0) PtS2
(n,n) PtS2
(n,0) PdS2
(n,n) PdS2

0.8
0.5
0.2
10

15

20

25

30

d/Å

35

40

45

50

(n,0) PtSe2
(n,n) PtSe2
(n,0) PdSe2
(n,n) PdSe2

1.0
0.8
0.5
0.2
10

15

20

25

30

d/Å

35

40

45

50

195
Figure 5.
Band structures of monolayers, (32,0), (18,18) nanotubes of M X 2
systems, respectively.
(14,0)

ML

1.75 eV

PtS2

0.8

1.28 eV

1.18 eV

0.0

0.0

0.8
0.0

2.8
0.8

1.4

PdS2

1.26 eV

M

Γ

K

M

0.92 eV

0.0

X

(14,0)

ML
2.0
1.0

0.8
0.79 eV

0.0

0.0

Γ X

(8,8)

1.2
1.38 eV

PtSe2

0.6

Γ
1.2

0.88 eV

1.0 eV

0.6

0.0

0.0

0.0

-1.0
1.6

-0.6
0.8

-0.6
0.8

dd

0.8

PdSe2

0.74 eV

0.6 eV

0.53 eV

0.0

0.0
-0.8
M

0.4

Γ

K

M

-0.4
X

E-EF / eV

E-EF / eV

1.6

E-EF / eV

E-EF / eV

1.4

(8,8)

1.6

2.8

0.4
0.0

Γ X

-0.4
Γ

4. Conclusions
In analogy to the existing transition metal chalcogenide nanotubes, we have investigated
hypothetical noble-metal chalcogenide nanotubes (PdS2 , PdSe2 , PtS2 and PtSe2 NTs) through density
functional theory calculations. We have shown that formation of these nanotubes is possible, since
they have smaller strain energies than MoS2 or WS2 nanotubes. Furthermore, we have found that the
strain energy of the studied nanotubes is chirality independent and decreases inverse quadratically
with the tube diameter. Moreover, PdX 2 nanotubes are more stable than PtX 2 for nanotubes
with small diameters. We have also examined the electronic structure of noble-metal chalcogenide
monolayers and nanotubes, which are found to be all indirect band gap semiconductors in the ranges
of 0.6–1.1 eV (0.9–1.7 eV) and 0.3–0.8 eV (0.6–1.3 eV) for PdS2 (PtS2 ) and PdSe2 (PtSe2 ) NTs,
respectively. These NTs band gaps increase with the diameter rapidly approaching that of the
respective pristine 2D monolayer.
Acknowldgements
This work was supported by the German Research Council (Deutsche Forschungsgemeinschaft,
HE 3543/18-1), the European Commission (FP7-PEOPLE-2009-IAPP QUASINANO, GA 251149
and FP7-PEOPLE-2012-ITN MoWSeS, GA 317451).

196
Author Contributions
Nourdine Zibouche, Agnieszka Kuc, Pere Miró and Thomas Heine generated, analyzed and
discussed the results. Thomas Heine conceived this project. All authors contributed in writing
this paper.
Conflicts of Interest
The authors declare no conflict of interest.
References
1. Pauling, L. The structure of the chlorites. Proc. Natl. Acad. Sci. USA 1930, 16, 578–582.
2. Iijima, S. Helical Microtubules of Graphitic Carbon. Nature 1991, 354, 56–58.
3. Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Polyhedral and Cylindrical Structures of
Tungsten Disulfide. Nature 1992, 360, 444.
4. Margulis, L.; Salitra, G.; Tenne, R.; Talianker, M. Nested fullerene-like structures. Nature
1993, 365, 113–114.
5. Chopra, N.G.; Luyken, R.J.; Cherrey, K.; Crespi, V.H.; Cohen, M.L.; Louie, S.G.; Zettl, A.
Boron Nitride Nanotubes. Science 1995, 269, 966–967.
6. Yin, L.W.; Bando, Y.; Golberg, D.; Li, M.S. Growth of Single-Crystal Indium Nitride
Nanotubes and Nanowires by a Controlled-Carbonitridation Reaction Route. Adv. Mater. 2004,
16, 1833–1838.
7. Spahr, M.; Bitterli, P.; Nesper, R.; Mueller, M.; Krumeich, F.; Nissen, H.U. Redoxaktive
Nanoröhren aus Vanadiumoxid. Angew. Chem. 1998, 110, 1339–1342.
8. Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K. Titania Nanotubes Prepared by
Chemical Processing. Adv. Mater. 1999, 11, 1307–1311.
9. Yada, M.; Mihara, M.; Mouri, S.; Kuroki, M.; Kijima, T. Rare Earth (Er, Tm, Yb, Lu) Oxide
Nanotubes Templated by Dodecylsulfate Assemblies. Adv. Mater. 2002, 14, 309–313.
10. Radovsky, G.; Popovitz-Biro, R.; Stroppa, D.G.; Houben, L.; Tenne, R. Nanotubes from
Chalcogenide Misfit Compounds: Sn–S and Nb–Pb–S. Acc. Chem. Res. 2014, 47, 406–416.
11. Zhu, Y.Q.; Sekine, T.; Brigatti, K.S.; Firth, S.; Tenne, R.; Rosentsveig, R.; Kroto, H.W.;
Walton, D.R.M. Shock-Wave Resistance of WS2 Nanotubes. J. Am. Chem. Soc. 2003,
125, 1329–1333.
12. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Ivanovskaya, V.; Heine, T.;
Seifert, G.; Wagner, H.D.; Tenne, R. Mechanical properties of individual WS2 nanotubes.
Electron. Prop. Synth. Nanostruct. 2004, 723, 306–312.
13. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Seifert, G.; Tenne, R.
Mechanical behavior of individual WS2 nanotubes. J. Mater. Res. 2004, 19, 454–459.
14. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Rosentsveig, R.; Ivanovskaya, V.; Heine,
T.; Seifert, G.; Wagner, H.D.; Tenne, R. Direct tensile tests of individual WS2 nanotubes.
Mater. Sci. Forum 2005, 475–479, 4097–4102.

197
15. Kaplan-Ashiri, I.; Cohen, S.R.; Gartsman, K.; Ivanovskaya, V.V.; Heine, T.; Seifert, G.;
Wiesel, I.; Wagner, H.D.; Tenne, R. On the Mechanical Behavior of WS2 Nanotubes under
Axial Tension and Compression. Proc. Natl. Acad. Sci. USA 2006, 103, 523–528.
16. Kaplan-Ashiri, I.; Tenne, R. Mechanical Properties of WS2 Nanotubes. J. Clust. Sci. 2007, 18,
549–563.
17. Rapoport, L.; Bilik, Y.; Feldman, Y.; Homyonfer, M.; Cohen, S.R.; Tenne, R. Hollow
nanoparticles of WS2 as potential solid-state lubricants. Nature 1997, 387, 791–793.
18. Rothschild, A.; Cohen, S.R.; Tenne, R. WS2 nanotubes as tips in scanning probe microscopy.
Appl. Phys. Lett. 1999, 75, 4025, doi:10.1063/1.125526.
19. Komarneni, M.R.; Yu, Z.; Burghaus, U.; Tsverin, Y.; Zak, A.; Feldman, Y.;
Tenne, R. Characterization of Ni-Coated WS2 Nanotubes for Hydrodesulfurization Catalysis.
Isr. J. Chem. 2012, 52, 1053–1062.
20. Zhang, W.; Ge, S.; Wang, Y.; Rafailovich, M.H.; Dhez, O.; Winesett, D.A.; Ade, H.;
Shafi, K.V.; Ulman, A.; Popovitz-Biro, R. Use of functionalized WS2 nanotubes to produce
new polystyrene/polymethylmethacrylate nanocomposites. Polymer 2003, 44, 2109–2115.
21. Unalan, H.E.; Yang, Y.; Zhang, Y.; Hiralal, P.; Kuo, D.; Dalal, S.; Butler, T.; Cha, S.N.;
Jang, J.E.; Chremmou, K.; et al. ZnO Nanowire and WS2 Nanotube Electronics. IEEE Trans.
Electron Devices 2008, 55, 2988–3000.
22. Wang, G.X.; Bewlay, S.; Yao, J.; Liu, H.K.; Dou, S.X. Tungsten Disulfide Nanotubes for
Lithium Storage. Electrochem. Solid-State Lett. 2004, 7, A321–A323.
23. Hua, M.; Zhanliang, T.; Feng, G.; Jun, C.; Huatang, Y. Synthesis, characterization and hydrogen
storage capacity of MS2 (M = Mo, Ti) nanotubes. Front. Chem. China 2006, 1, 260–263.
24. Chen, J.; Li, S.L.; Tao, Z.L.; Gao, F. Low-temperature synthesis of titanium disulfide nanotubes.
Chem. Commun. 2003, 980–981.
25. Remskar, M.; Mrzel, A.; Jesih, A.; Leperevy, F. Metal-Alloyed NbS2 Nanotubes Synthesized
by the Self-Assembly of Nanoparticles. Adv. Mater. 2002, 14, 680–684.
26. Coleman, K.S.; Sloan, J.; Hanson, N.A.; Brown, G.; Clancy, G.P.; Terrones, M.; Terrones, H.;
Green, M.L.H. The Formation of ReS2 Inorganic Fullerene-like Structures Containing Re4
Parallelogram Units and Metal–Metal Bonds. J. Am. Chem. Soc. 2002, 124, 11580–11581.
27. Chen, J.; Tao, Z.L.; Li, S.L.; Fan, X.B.; Chou, S.L. Synthesis of TiSe2 Nanotubes/Nanowires.
Adv. Mater. 2003, 15, 1379–1382.
28. Li, P.; Stender, C.L.; Ringe, E.; Marks, L.D.; Odom, T.W. Synthesis of TaS2 Nanotubes from
Ta2 O5 Nanotube Templates. Small 2010, 6, 1096–1099.
29. Miro, P.; Audiffred, M.; Heine, T. An Atlas of Two-Dimensional Materials. Chem. Soc. Rev.
2014, 43, 6537–6554.
30. Nicolosi, V.; Chhowalla, M.; Kanatzidis, M.G.; Strano, M.S.; Coleman, J.N. Liquid Exfoliation
of Layered Materials. Science 2013, 340, doi:10.1126/science.1226419.
31. Remskar, M.; Skraba, Z.; Regula, M.; Ballif, C.; Sanjines, R.; Levy, F. New crystal structures
of WS2 : Microtubes, ribbons, and ropes. Adv. Mater. 1998, 10, 246–249.

198
32. Catherine, M.; Zelenski, P.K.D. Template Synthesis of Near-Monodisperse Microscale
Nanofibers and Nanotubules of MoS2 . J. Am. Chem. Soc. 1998, 120, 734–742.
33. Zhu, Y.Q.; Hsu, W.K.; Grobert, N.; Chang, B.H.; Terrones, M.; Terrones, H.; Kroto, H.W.;
Walton, D.R.M. Production of WS2 Nanotubes. Chem. Mater 2000, 12, 1190–1194.
34. Feldman, Y.; Wasserman, E.; Srolovitz, D.J.; Tenne, R. High-Rate, Gas-Phase Growth of MoS2
Nested Inorganic Fullerenes and Nanotubes. Science 1995, 267, 222–225.
35. Gronvold, F.; Rost, E. On the sulfides, selenides and tellurides of palladium. Acta Chem. Scand.
1956, 10, 1620–1634.
36. Kjekshus, A.; Gronvold, F. High temperature X-ray study of the thermal expansionn of PtS2 ,
PtSe2 , PtTe2 and PdTe2 . Acta Chem. Scand. 1959, 13, 1767–1774.
37. Gronvold, F.; Haraldsen, H.; Kjekshus, A. On the sulfides, selenides and tellurides of platinium.
Acta Chem. Scand. 1960, 14, 1879–1893.
38. Wilson, J.A.; Yoffe, A.D. The Transition Metal Dichalcogenides Discussion and Interpretation
of Observed Optical, Electrical and Structural Properties. Adv. Phys. 1969, 18, 193–335.
39. Miro, P.; Ghorbani-Asl, M.; Heine, T.
Two Dimensional Materials Beyond MoS2 :
Noble-Transition-Metal Dichalcogenides. Angew. Chem. Int. Ed. 2014, 53, 3015–3018.
40. Perdew, J.P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple.
Phys. Rev. Lett. 1996, 77, 3865, doi:10.1103/PhysRevLett.77.3865.
41. Dovesi, R.; Saunders, V.R.; Roetti, R.; Orlando, R.; Zicovich-Wilson, C.M.; Pascale, F.;
Civalleri, B.; Doll, K.; Harrison, N.M.; Bush, I.J.; et al. CRYSTAL09 User’s Manual; University
of Torino: Torino, Italy, 2009.
42. Peintinger, M.F.; Oliveira, D.V.; Bredow, T. Consistent Gaussian basis sets of triple-zeta valence
with polarization quality for solid-state calculations. J. Comput. Chem. 2013, 34, 451–459.
43. Doll, K. Adsorption on the Pt surface a comparison of a gradient corrected functional and a
hybrid functional. Surf. Sci. 2004, 573, 464 – 473.
44. Stoll, H.; Metz, B.; Dolg, M.
Relativistic energy-consistent pseudopotentials–recent
developments. J. Comput. Chem. 2002, 23, 767–778.
45. Monkhorst, H.J.; Pack, J.D. Special Points For Brillouin-Zone Integrations. Phys. Rev. B 1976,
13, 5188, doi:10.1103/PhysRevB.13.5188.
46. Zibouche, N.; Kuc, A.; Heine, T. From layers to nanotubes: Transition metal disulfides TMS2 .
Eur. Phys. J. B 2012, 85, 49, doi:10.1140/epjb/e2011-20442-1.
47. Kuc, A.; Zibouche, N.; Heine, T. Influence of quantum confinement on the electronic structure
of the transition metal sulfide TS2 . Phys. Rev. B 2011, 83, doi:10.1103/PhysRevB.83.245213.

MDPI AG
Klybeckstrasse 64
4057 Basel, Switzerland
Tel. +41 61 683 77 34
Fax +41 61 302 89 18
http://www.mdpi.com/
Inorganics Editorial Office
E-mail: inorganics@mdpi.com
http://www.mdpi.com/journal/inorganics

MDPI • Basel • Beijing • Wuhan
ISBN 978-3-03842-149-8
www.mdpi.com