A Thesis

Entitled

Non-hydrolytic Sol-gel (NHSG) Synthesis of Transition Metal Sulfides and Theoretical

Investigations

Xiuquan Zhou

Submitted to the Graduate Faculty as partial fulfillment of the requirements for

The Master of Science Degree in Chemistry

______Dr. Cora Lind-Kovacs, Committee Chair

______Dr. Terry Bigioni, Committee Member

______Dr. Eric Findsen, Committee Member

______Dr. Sanjay Khare, Committee Member

______Dr. Patricia Komuniecki Dean of College of Graduate Studies

The University of Toledo

May 2013

An Abstract of

Non-hydrolytic Sol-gel (NHSG) Synthesis of Transition Metal Sulfides and Theoretical Investigations

by

Xiuquan Zhou

Submitted to the Graduate Faculty as partial fulfillment of the requirements for The Master of Science Degree in Chemistry

The University of Toledo May 2013

Non-hydrolytic sol-gel (NHSG) synthesis provides an elegant approach to many solid-state materials, which was originally developed for the preparation of oxides.1 As they do not require high temperatures like conventional solid-state routes, access to thermodynamically metastable materials, which cannot be prepared through traditional solid-state routes, is possible. In this project, NHSG chemistry is explored for the synthesis of binary metal sulfides.

Sulfides, some of which are thermally unstable and highly oxygen sensitive, have applications in many areas, such as solar cells, catalysts, sensors, lubricants, semiconductors etc.2 Despite the widespread use of sulfides, they have been studied much less comprehensively than oxides. One of the reasons is the difficulty in synthesis because metastable sulfides are intolerant to high temperatures and oxygen. In NHSG routes, the reaction of a metal halide with a thioether is used to form a metal sulfide

iii

network at low temperatures. This can give access to materials that are not accessible by high temperature routes, and may lead to the discovery of new phases.3

Because synthesis of sulfides through NHSG is an unexplored area, the synthesis conditions must be optimized for each metal. In this thesis, the Cu-S and Ta-S systems were thoroughly explored using NHSG techniques. Amorphous tantalum sulfides were obtained in as-recovered samples, and heat treatments of such amorphous precursors resulted in crystallization of 1-TaS2 and 3R-TaS2 phases. For the synthesis of copper sulfides, precise phase selection of five different polymorphs, metastable hexagonal chalcocite, monoclinic chalcocite, djurleite, low digenite and covellite, was achieved by fine tuning synthetic parameters. In addition, a continuous phase evolution from copper rich phases towards copper deficient phases between chalcocites and djurleite was observed. All polymorphs could be obtained as nanoparticles.

Theoretical studies can complement experimental approaches, and may aid in understanding polymorph stabilities. In addition, theory can elucidate phase transition pathways between polymorphs, which may help design synthetic approaches to specific phases. As a starting project, pressure-induced phase transitions from the NaCl-type (B1) to the CsCl-type (B2) structure in BaS, BaSe and BaTe were studied using ab initio density functional theory computations in the local density approximation. The Buerger4 and WTM5 mechanisms were explored by mapping the enthalpy contours in two and four dimensional configuration space for the two mechanisms, respectively. Transition pressures for BaS, BaSe and BaTe were determined to be 5.5 GPa, 4.9 GPa and 3.4 GPa, respectively. From these configuration space landscapes, a low enthalpy barrier path was constructed for the transitions to proceed at three different pressures. We obtained

iv

barriers of 0.18, 0.16 and 0.15 eV/pair (17.4, 15.4 and 14.5 kJ/mol) for the Buerger mechanism, and 0.13, 0.13 and 0.12 eV/pair (12.5, 12.5 and 11.6 kJ/mol) for the WTM mechanism at the transition pressures for BaS, BaSe and BaTe, respectively, indicating that the WTM mechanism is slightly more favorable in these compounds.

v

Acknowledgements

A conclusion of my study at the University of Toledo (UT) will soon be reached, although the results will be quite different from what I expected when I first came here.

From both bitter and sweet experiences, I indeed have learned much more than what I expected as well. I would like to thank those who have helped and supported me in the past, as nothing was achieved alone.

First and most, I would like to thank my advisor Dr. Cora Lind-Kovacs for her guidance and inspiration. I would not have decided to continue to do research in the field of solid-state chemistry if I had not worked for her. She not only taught me valuable knowledge of sciences, but also taught me how to be a scientist. Admittedly, not all experiences were pleasant, but it was those unpleasant ones that revealed mistakes and helped me to change. I did not fully realize the importance of the role of an advisor until I reached the lowest point of my life, as many things I had taken for granted. A good advisor is a like a beacon that can guide lost ships to a safe port from rough seas in storms. Fortunately, I had a very good advisor, and for that matter, I will always be grateful.

I would like to thank Dr. Sanjay Khare, not only for the financial support of my last semester at UT, but also for leading me into the world of theoretical physics. He taught me a valuable lesson that a seemingly fine paper from a reputable group in a

vi

respectful journal could contain many vital mistakes by iterating a work related to our research. He also provided us laughter with many of his witty jokes, such as "always trying to roar like a tiger, even though end up with a meow".

I would like to thank my committee members, Dr. Terry Bigioni and Eric

Findsen. I would like to thank Pannee for all the patient help with the instruments, Steve for the help with glassware, and Youming and Tom for fixing many equipments. I would like to thank Jason Roehl, Dr. Khare's graduate student, for teaching me how to use

VASP and working together with me. Not for Jason, our publication on Journal of

Physics: Condensed Matter would not have been possible. I would also like to thank

Jason for showing me how to be a professional graduate student.

I would like to thank both current and former members of Lind-Kovacs group for eating, working and laughing together. I enjoyed and appreciated their friendships, especially for Xiaodong and Nate. In addition, I want to thank Nathalie Pedoussaut, Anne

Soldat, Christophe Heinrich, Martin Kluenker and Derek Mull as their work will be included in our papers.

I would like to thank the University of Toledo and all the faculty and stuff at the department of chemistry. I want to thank National Science Foundation (NSF) for funding our work on sulfides (DMR 1005911), providing computational resources (CNS 0855134 and CMMI 1234777) and providing SEM (CRIF 0840474). I would like to thank Ohio

Supercomputer Center (OSC) for providing additional computational resources and the

Advanced Photon Source (APS) for 16 mail-in samples.

vii

viii

Table of Contents

An Abstract of ...... iii

Acknowledgements ...... vi

Table of Contents ...... ix

List of Tables ...... xiii

List of Figures ...... xvi

1. Introduction ...... 1

1.1 Transition Metal Sulfides ...... 1

1.2 Literature Routes to Metal Sulfides ...... 7

1.2.1 Traditional High Temperature Routes ...... 8

1.2.2 Solution Based Routes ...... 9

1.2.3 Other Routes ...... 10

1.3 Non-hydrolytic Sol-gel (NHSG) Methods ...... 11

1.4 Theoretical Modeling in Solid State Chemistry ...... 14

1.5 Goals of Thesis Research ...... 16

2. Experimental Methods ...... 18

2.1. Synthetic Method ...... 18

2.1.1 Room Temperature Route ...... 19

2.1.2 Solvo-thermal Route ...... 20

2.1.3 Heat Treatment...... 21

ix

2.2. Characterization Methods ...... 21

2.2.1 CHNS Analysis ...... 21

2.2.2 Thermogravimetric and Differential Thermal Analysis ...... 25

2.2.3 Powder X-ray Diffraction ...... 26

2.2.4 Scanning Electron Microscopy and Energy Dispersive Spectroscopy ...... 28

3. Synthesis and Characterization of Copper Sulfides ...... 31

3.1 Introduction ...... 31

3.1.1 Applications of Copper Sulfides and Synthetic Challenges ...... 32

3.1.2 Copper Sulfide Polymorphs ...... 34

3.1.3 Literature Routes for the Preparation of Copper Sulfides ...... 39

3.2 Preparation of Copper Sulfides by Non-hydrolytic Sol-gel Synthesis ...... 42

3.3 Results and Discussion ...... 46

3.3.1 Elemental Analysis ...... 47

3.3.2 Neat Reactions ...... 49

3.3.3 CuCl and HMDST in Chloroform and Acetonitrile ...... 56

3.3.4 CuCl2 and HMDST in Solvents ...... 85

3.3.5 Synthesis with DTBS ...... 94

3.3.6 Phase and Morphology Evolution ...... 101

3.4 Effect of Synthetic Parameters ...... 104

3.4.1 Effect of Copper Sources ...... 104

3.4.2 Effect of Sulfur Sources ...... 104

3.4.3 Effect of Solvents ...... 105

3.4.4 Effect of Temperature ...... 105

x

3.5 Conclusions ...... 105

4. Synthesis and Characterization of Tantalum Sulfides ...... 107

4.1 Introduction ...... 107

4.1.1 Polymorphs and Applications of Tantalum Sulfides ...... 107

4.1.2 Literature Routes ...... 114

4.1.3 Preliminary Results ...... 116

4.2 Preparation of Tantalum Sulfides by Non-hydrolytic Sol-gel Synthesis ...... 117

4.3 Results and Discussion ...... 120

4.3.1 Phases and Elemental Analysis ...... 120

4.3.2 Influence of Synthetic Conditions ...... 130

4.4 Conclusions ...... 143

5. Study of B1 (NaCl-type) to B2 (CsCl-type) Pressure-induced Structural Phase

Transition in BaS, BaSe and BaTe Using ab initio Computations ...... 144

5.1 Introduction ...... 144

5.2 Transition mechanisms ...... 146

5.3 Computational method ...... 149

4. Modeling approach ...... 150

5.5 Results and discussion ...... 152

5.5.1 Transition pressure and structural parameters ...... 152

5.5.2 Potential energy surface and energy barriers ...... 157

5.5.3 Symmetry and coordination ...... 163

5.6 Conclusions ...... 165

6. Summary and Future Work ...... 166

xi

References ...... 169

Appendix A ...... 183

Appendix B ...... 189

xii

List of Tables

2.1: Analytical blanks during CHNS analysis and uncertainty estimates...... 24

3.1: Stability ranges and crystal systems of the 13 known copper sulfide polymorphs ... 34

3.2: Synthetic parameters for copper sulfide samples...... 43

3.3: PDF cards and structural parameters of observed copper sulfide polymorphs...... 47

3.4: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of copper chlorides and HMDST without solvent...... 50

3.5: CHNS analysis and EDS results of raw copper sulfide samples prepared by reactions of copper sources and HMDST without solvent ...... 50

3.6: In situ variable temperature X-ray diffraction results for the poorly crystallized phase of sample XZ1026...... 53

3.7: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in 15 mL (XZ1001 to XZ1007) and 10 mL CHCl3...... 58

3.8: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CHCl3...... 61

3.9: In situ variable temperature X-ray diffraction results for XZ1009 (hexagonal chalcocite), XZ1011 and XZ1020 (djurleite), and XZ1019 (digenite) raw samples...... 68

3.10: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CH3CN...... 71

xiii

3.11: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CH3CN...... 73

3.12: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CHCl3...... 87

3.13: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CH3CN...... 88

3.14: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CHCl3...... 89

3.15: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CH3CN...... 89

3.16: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl or CuCl2 and DTBS in CHCl3 or CH3CN...... 95

3.17: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl or CuCl2 with DTBS...... 96

3.18: X-ray diffraction results for as-recovered samples prepared with DTBS...... 97

4.1: Crystallographic parameters of tantalum sulfide polymorphs...... 108

4.2: Synthetic parameters for tantalum sulfide samples...... 119

4.3: X-ray diffraction results for raw and heated tantalum sulfide samples prepared with 3 mmol of halide in 15 mL of solvent...... 121

4.4: X-ray diffraction results for raw and heated tantalum sulfide samples prepared with 1 mmol of halide in 15 mL of solvent...... 122

4.5: Elemental analysis results for raw and heated tantalum sulfide samples...... 123

4.6: Elemental analysis results for selected CH and DM samples...... 124

xiv

4.7: XRD results for previous and current samples prepared with DTBS...... 131

4.8: XRD results for previous and current samples prepared by reactions between

HMDST and TaI5...... 132

4.9: XRD results for previous and current samples prepared by reactions between

HMDST and TaBr5...... 135

4.10: XRD results for previous and current samples prepared by reactions between

HMDST and TaCl5...... 139

5.1: Results for transition pressures and structural parameters from both present work and earlier work...... 156

5.2: Coefficients (Cm) of the cubic polynomial fits to the computed transition paths. ... 161

5.3: Energy barriers in eV/pair and kJ/mol (in parentheses) for the transition from B1 to

B2 phase for BaS, BaSe and BaTe...... 163

xv

List of Figures

1-1: Structures of a) zinc blende and b) wurtzite with anions and cations depicted by light and dark balls, respectively...... 3

1-2: Structures of CdI2 type compounds with a) a primitive cell and b) extended network.

Anions and cations are depicted by dark and light balls respectively...... 4

1-3: The unit cell of pyrite (FeS2) with cations and anions depicted as dark and light balls, respectively...... 5

1-4: Scheme illustrating the mechanism of non-hydrolytic sol-gel routes to metal oxides.

...... 12

2-1: Structures of a) tert-butylsulfide (DTBS) and b) hexamethyldisilathiane (HMDST).

...... 18

2-2: Components of a PE-2400 II CHNS/O analyzer...... 23

2-3: Air sensitive PXRD sample holder for Scintag XDS-2000 diffractrometer...... 27

2-4: a) SEM image of a copper sulfide sample and distributions of b) Cu, c) S, d) Cl and

(e) all three elements in the image...... 30

3-1: Phase diagram of the Cu-S system. (adapted from Ref. 27) ...... 35

3-2: Phase diagram of the Cu-S system in the Cu1.7S-Cu2S composition range...... 36

3-3: XRD patterns of sample XZ1013 prepared by reaction of CuCl and HMDST at RT without solvent for a) 3 h, b) 24 h and c) 96 h...... 51

xvi

3-4: XRD patterns of samples prepared by reactions of CuCl and HMDST without solvent at a) 70 °C for 7 d and b) 130 °C for 7.5 d...... 52

3-5: XRD patterns of samples prepared by reactions of CuCl2 and HMDST without solvent at a) RT for 3 h, b) RT for 96 h and c) 70 °C for 7 d...... 54

3-6: SEM images of a) and b) XZ1087raw, c) XZ1055raw, d) and e) XZ1061raw, and f)

XZ1084raw...... 56

3-7: XRD patterns of as-recovered samples prepared at RT for 24 h: a) ACS18 (0.4 mL

HMDST), b) ACS19 (0.8 mL HMDST) and c) ACS17 (2.4 mL HMDST)...... 63

3-8: XRD patterns of as-recovered samples prepared at RT for 7 d: a) XZ1102, b)

XZ1103, c) XZ1104 and d) an insert of the capillary data for XZ1104...... 64

3-9: XRD patterns of as-recovered samples of a) and b) XZ1015 collected on capillary and flat stages, respectively, c) XZ1011 and d) XZ1020...... 65

3-10: XRD patterns of identical samples a) XZ1025 and b) ACS21 prepared at 130 °C with 0.4 mL HMDST...... 66

3-11: XRD patterns of parallel reactions at 130 °C with 0.8 mL HMDST for a) 7 d

(XZ1088) and b) 15 d (XZ1089)...... 67

3-12: XRD patterns of samples prepared at RT with 0.4 mL HMDST for a) 5 min, b) 5 h, c) 24 h, d) 96 h and e) 168 h...... 74

3-13: XRD patterns of samples prepared at RT for 1.5 h with a) 0.8 mL HMDST and b)

2.4 mL HMDST, and c) for 7 d with 0.8 mL HMDST...... 75

3-14: XRD patterns of samples prepared at a) 70 °C with 0.8 mL HMDST, b) 70 °C with

2.4 mL HMDST and c) 130 °C with 2.4 mL HMDST...... 76

xvii

3-15: SEM images of RT samples prepared with 0.4 mL HMDST in CHCl3: a) and b)

ACS18, c) XZ1105raw and d) XZ1102raw...... 78

3-16: SEM images of RT samples prepared in CHCl3: a) ACS19, b) XZ1106raw, c)

XZ1103raw and d) XZ1104raw...... 79

3-17: SEM images of djurleite samples prepared at 70 °C in CHCl3: a) XZ1011raw (0.8 mL HMDST), and b) XZ1020raw (2.5 mL HMDST)...... 80

3-18: SEM images of low digenite samples prepared at 130 °C in CHCl3: a) and b)

XZ1022raw (2.5 mL HMDST), and c) and d) XZ1025raw (0.4 mL HMDST)...... 81

3-19: SEM images of samples prepared at RT in CH3CN a) XZ1111raw, b) XZ1097raw, c) XZ1112raw, and d) XZ1098raw...... 82

3-20: SEM images of samples prepared at RT with 0.4 mL HMDST in CH3CN: a)

XZ1113, b) XZ1110, c) XZ1114, d) XZ1053, e) XZ1086...... 83

3-21: SEM images of djurleite samples prepared with 0.8 mL HMDST at 130 °C in

CH3CN (XZ1031)...... 85

3-22: XRD patterns of chloroform samples prepared by reactions between CuCl2 and

HMDST...... 91

3-23: XRD patterns of samples prepared with 0.8 mL HMDST a) in CHCl3 at RT, and in

CH3CN at b) RT, c) 70 °C and d) 130 °C...... 92

3-24: SEM images of covellite samples prepared in CHCl3: a) XZ1032raw, b)

XZ1044raw, c) and d) XZ1027raw, and e) and f) XZ1038raw...... 93

3-25: SEM images of covellite samples prepared in CH3CN: a) and b) XZ1041, c)

XZ1045, and d) XZ1051...... 94

xviii

3-26: XRD patterns of CuCl2 samples prepared in 15 mL of CH3CN for 7 d with a) 2.0 mL DTBS (NPSu32), b) 4.0 mL DTBS (NPSu40) and c) 6.0 mL DTBS (NPSu78)...... 99

3-27: XRD patterns of a) XZ1081 (CuCl, 0.8 mL DTBS, CH3CN) and b) ACS9(CuCl,

0.8 mL DTBS, CH3CN)...... 100

3-28: SEM images of DTBS samples: a) XZ1075raw, b) and c) XZ1077raw, and d)

XZ1081raw...... 101

3-29: Conditions for targeted synthesis of several copper sulfide polymorphs for reactions of CuCl with HMDST...... 102

4-1: (a) Octahedral and (b) trigonal prismatic coordination sites for Ta atoms in TaS2, and

(c) a depiction of the (110) plane (shaded) of a primitive unit cell of TaS2...... 109

4-2: Cross sections of (110) planes of a) 1T-TaS2, b) 2H-TaS2, c) 3R-TaS2, and d) 6R-

TaS2...... 110

4-3: Schematic of the stacking of the trigonal prism columns along the b-axis in TaS3. 113

4-4: Projection of a pentagonal antiprism formed by Ta atoms and the bonded sulfur atoms...... 114

4-5: XRD patterns of as-recovered samples a) XZ2001, b) XZ2006 and c) XZ2012. ... 125

4-6: EDS spectrum of XZ2009 (TaI5+HMDST in CHCl3)...... 127

4-7: TGA (solid line) and DTA (dashed line) curves in air for pre-heated sample

XZ2006.T800. The sample was heated to 800 °C for 10 h before the TG/DTA run. .... 129

4-8: XRD patterns of a) as-recovered, b) 200 °C heat treated and c) 900 °C heat treated

CHTa16 (150 °C, 15 mL CH3CN, S/Ta = 3)...... 133

4-9: STEM images of CHTa16 (TaI5+HMDST in CH3CN at 150 °C for 7 d) at different magnifications...... 134

xix

4-10: XRD patterns after heat treatment to 800 °C for TaBr5/HMDST samples prepared for 7d with 15 mL of CH3CN: a) XZ2007, b) XZ2017 and c) XZ2018...... 135

4-11: XRD patterns after heat treatment to 700 °C for TaBr5/HMDST samples prepared for 7d with 15 mL of CH3CN: a) XZ2017 and b) XZ2018...... 136

4-12: SEM images of 800 °C heat treated HMDST samples prepared in 15 mL CH3CN: a) and b) XZ2007, c) XZ2017 and d) XZ2018...... 137

4-13: XRD patterns after heating to 800 °C for TaCl5/HMDST samples prepared at

100 °C for 7 d in chloroform...... 140

4-14: SEM images of 800 °C heat treated HMDST samples of a) and b) XZ2003, c) and d)

XZ2013 and e) and f) XZ2014...... 141

4-15: XRD patterns after heating to 800 °C for TaCl5/HMDST samples prepared with 15 mL of CH3CN for 7 d at a) 150 °C and b) at 100 °C...... 142

4-16: SEM images of 800 °C heat treated HMDST samples prepared in 15 mL CH3CN of a) XZ2004 (S/Ta=5.8, 150 °C) and b) XZ2022 (S/Ta=16.9, 100 °C)...... 143

5-1: The transition from the B1 to the B2 structure depicted in (a) rhombohedral unit cell

(used for Mechanism I) and (b) orthorhombic unit cell (used for Mechanism II)...... 147

5-2: Plots of enthalpy (eV/pair) of one pair of atoms as a function of pressure (GPa) of (a)

BaS, (b) BaSe, and (c) BaTe...... 154

5-3: Plots of cohesive energy (eV/pair) of one pair of atoms as a function of volume

(Å3/pair) of (a) BaS, (b) BaSe, and (c) BaTe for B1 and B2 structures...... 155

5-4: Contour plot of the computed enthalpy as a function of θ and a of (a) BaS at 5.5 GPa,

(b) BaSe at 4.8 GPa and (c) BaTe at 3.4 GPa, for Mechanism I...... 159

xx

5-5: Contour plot of the computed enthalpy as a function of x and c of (a) BaS at 5.5

GPa, (b) BaSe at 4.8 GPa and (c) BaTe at 3.4 GPa, for Mechanism II...... 160

5-6: The nearest and second nearest neighboring barium atoms of sulfur atoms in the transition state at 5.5 GPa of (a) Mechanism I and (b) Mechanism II...... 164

xxi

Chapter 1

1. Introduction

1.1 Transition Metal Sulfides

Various transition metal sulfide minerals have been discovered in nature, such as pyrite (FeS2), zinc blende (ZnS), cinnabar (HgS) and chalcocite (Cu2S). Their ores are very important sources of transition metals.1 The demand for transition metals has been growing over the past decade, in large part due to the rapid growth of new emerging economies, such as China and India. Thus, new mineral sources, such as seafloor massive sulfides (rich in Fe, Cu, Zn, Pb), become more attractive as the prices of the metals are increasing.2 In order to fully explore such resources, understanding their mechanism of formation is important. For example, seafloor massive sulfides are deposited by precipitation through undersea hydrothermal reactions, which are more common at volcanic sites.2 A detailed understanding of the formation of these undersea minerals may enable predicting what types of geological locations are likely to contain rich deposits.

Besides being important metal sources, transition metal sulfides are widely used in catalysis, solar cells, lubricants, electronics, sensors, etc.3-14 Because the electronegativity of sulfur is lower than oxygen, and the radius of the sulfide anion is larger than that of the oxide anion, the bonds between transition metals and sulfur ions

1

have more covalent character, and the sulfide ion is more polarizable.15-17 As a result of this more covalent bonding character, wider composition ranges are often observed for transition metal sulfides.18 In addition, because most of the transition metal sulfides can form a variety of nonstoichiometric polymorphs, their structures can be very diverse.16,17

Even the structures of simple binary compounds, such as CuS (covellite), can be very complex.17,19 Although the structural characterization of transition metal sulfides can be difficult, knowledge of atomic level structure is essential to understanding their properties.

Many binary transition metal sulfides can assume zinc blende, wurtzite, NiAs-

17 type, CdI2-type and pyrite structures or their derivatives. Of these, zinc blende, wurtzite and NiAs-type structures are commonly found in transition metal monosulfides.16,17 Many binary sulfides, such as ZnS and CdS, can adopt the zinc blende structure (F 3m), shown in Figure 1-1 a).16 In this type of structure, anions form a cubic close packed lattice, and cations occupy half of the tetrahedral holes, resulting in a 1:1 anion to cation ratio.

Sulfides that adopt the zinc blende structure can usually also adopt the wurtzite stucture

16,17 (P63mc) shown in Figure 1-1 b). In the wurtzite structure, anions form a hexagonal close packed lattice, and like in the zinc blende structure, cations occupy half of the tetrahedral holes. Besides these two structure types, the NiAs-type structure (P63/mmc) is also commonly found in transition metal sulfides, such as FeS, NiS, CoS and IrS.16,17 In the NiAs structure, anions form a hexagonal close packed lattice, and cations occupy all of the octahedral holes. Many transition metal sulfides can also form non-stoichiometric variants, such as monoclinic Cr1-xS, which result in distorted NiAs structures with different vacancy ordering patterns.16,17

2

Figure 1-1: Structures of a) zinc blende and b) wurtzite with anions and cations depicted by light and dark balls, respectively.

The CdI2-type structure (P m1), shown in Figure 1-2, is a common type of

17 structure for transition metal disulfides, such as TiS2 and ZrS2. This type of structure is similar to the NiAs-type structure, with hexagonal close packed anions. However, in the

CdI2-type structure, cations only occupy half of the octahedral holes, resulting in vacancies in every other metal layer. This results in sulfur-sulfur interactions of adjacent sulfur layers through van der Waals forces, analogous to graphite (Figure 1-2 b)). Other transition metal disulfides, such as MoS2 and WS2, adopt structures derived from the 3

17 CdI2-type structure. For example, the structure of MoS2 can be considered as a CdI2 structure with all the sulfur atoms displaced from their special positions, resulting in distorted octahedral coordination for Mo atoms.

Figure 1-2: Structures of CdI2 type compounds with a) a primitive cell and b) extended network. Anions and cations are depicted by dark and light balls respectively.

2- Many transition metals form compounds with the disulfide anion, S2 , which results in pyrite-type (FeS2) structures (Pa ), shown in Figure 1-3. Besides FeS2, the

transition metal disulfides CoS2, MnS2, NiS2 and RuS2 adopt the cubic pyrite structure.

16,17 2- 2- In the pyrite structure, S2 groups instead of S ions form a face-centered cubic

4

lattice, and cations occupy all the octahedral sites.17 In this arrangement, because sulfur atoms form S2 pairs, the overall ratio of metal to sulfur is 1:2. Besides the simple pyrite type structure, some transition metal sulfides are found to crystallize in pyrite

16 superstructures, such as Cu3FeS8.

Figure 1-3: The unit cell of pyrite (FeS2) with cations and anions depicted as dark and light balls, respectively.

Besides the structure types mentioned above, transition metals can form a variety of complex nonstoichiometric sulfides. The characterization of nonstoichiometric transition metal sulfides can be difficult, because in most cases solution based analytical methods and X-ray microanalysis cannot provide accurate compositional information for nonstoichiometric compounds. Therefore, the most reliable method to obtain accurate compositions of transition metal sulfides is the structure determination from perfect single crystals, which may not be available for every compound.17 In addition, perfect single crystals may not be representative of the corresponding powders or minerals.

5

However, structure determination of transition metal sulfides is crucial to understanding their properties, which dictate their applications. In the following paragraphs, some examples of applications of transition metal sulfides are described, and their properties are explained by the microscopic structures of these sulfides.

Transition metal sulfides are well known as catalysts for hydrodesulfurization, the removal of undesired sulfur during industrial processes, especially in petroleum processing.3,20 T. A. Pecoraro and R. R Chianelli found that the catalytic performance of sulfides of 4d and 5d metals was orders of magnitude better than the performance of 3d metals, and the reactivity increased towards the center of the d-block.3 It appeared that the catalytic activity of these metals was propotional to their d character.3 Therefore, they suggested that hydrodesulfurization performance of transition metal sulfides was related to the strength of their covalent bonding, which was confirmed by theoretical studies of their electronic structures and electron densities. The catalytic properties of transition metal sulfides have also been utilized in other reactions, such as hydrogen evolution4 and photo-oxidation of organics.21 In addition, some interesting studies showed that transition metal sulfides might have been catalysts for the formation of life due to their association with the metabolism of early earth life forms.22

Some transition metal sulfides with layered structures, such as MoS2 and WS2, can be used as dry lubricants.15 As shown in Figure 1-2, the neighboring sulfur layers in

MoS2 are held together by van der Waals forces. Because of the weak interaction between layers of sulfur atoms, they easily slide past each other, resulting in low friction coefficients. These properties of MoS2 make it an excellent dry lubricant, which can be used in conditions like high vacuum or space, where a wet lubricant cannot be applied.

6

Some transition metal sulfides, such as group 11 and 12 metal sulfides, are

9,13-15 excellent semiconductors due to their electronic properties. For example, Cu2-xS can form with very high copper deficiencies, resulting in cation vacancies, while CdS can be doped to create electron holes, which makes them good p-type and n-type

14,23 semiconductors, respectively. Because the band gaps of Cu2S and CdS are around 1.1 eV and 2.4 eV,23 respectively, they are suitable solar cell materials and were used in the

14,24 fabrication of early solar cells. Recently, a CuInS2/CdS based solar cell has been reported to achieve an efficiency of 10.2%.25 Sulfides with promising photovoltaic properties may be obtained by further exploration of the Cu-In sulfide system. Because of the semiconducting properties of transition metal sulfides, they are also suitable materials for capacitors and transistors.8,9 Moreover, recent studies of nanostructured transition metal sulfide materials showed intriguing properties that differ from those of bulk

26,27 materials. For example, a monolayer of MoS2 demonstrated promising properties for transistors and photoluminescence applications, showing a strong contrast to bulk MoS2, which is not a good semiconductor material.26,27 Thus, by adjusting the particle size and morphology of transition metal sulfides, their properties can be enhanced.

1.2 Literature Routes to Metal Sulfides

As described in the previous section, transition metal sulfides exhibit many interesting properties and have a variety of applications. However, their naturally occurring minerals usually contain large amounts of impurities, and cannot be directly used.17 In addition, some polymorphs of transition metal sulfides, such as air sensitive and metastable polymorphs, cannot be found in nature. Therefore, targeted synthetic methods are essential to obtain transition metal sulfides, especially if nanostructures are

7

desired. In general, the synthetic routes can be categorized into traditional routes, which usually involve high temperatures, and solution based routes. These will be discussed in sections 1.2.1 and 1.2.2, respectively. In addition, in order to make transition metal sulfide nanoparticles, many new methods have been developed as well. They are distinct from both the traditional high temperature routes and the widely applied soft chemistry routes. Some of these methods are discussed in section 1.2.3.

1.2.1 Traditional High Temperature Routes

The synthesis of transition metal sulfides by traditional routes is usually carried out at high temperatures in the solid state. The most common method to obtain transition metal sulfides is the direct reaction of metals with elemental sulfur. Except for Au, Pt and

Pd, most transition metals react directly with elemental sulfur at certain temperatures.28 In order to avoid oxidation, reactions are usually carried out by heating mixtures of elemental powders in sealed glass or quartz tubes.29,30 The composition of the products can be controlled by the initial ratio of the elements, but usually only thermodynamically stable compounds can be obtained. Direct reactions of the elements can also be carried out by ball milling or mechanical alloying through grinding fine powders of transition metals and sulfur at high temperatures.29-31

Vapor transport is another common method to obtain transition metal sulfides.

This type of reaction is usually carried out by reaction of a sulfur-containing vapor phase, such as H2S or CS2, with binary transition metal compounds, such as metal halides

28,32 or even oxides at high temperatures. For example, TiS2 can be obtained by heating

TiCl4 with H2S at 450 °C , and pure NbS2 can be prepared by heating Nb2O5 with CS2 or a

8

27,31 mixture of CS2 and H2S at above 900 °C . Sometimes, vapor transport is also used to purify or recrystallize metal sulfides. For example, the reaction

TaS2 + 2I2 ⇌ TaI4 + 2S (Equation 1-1), is used to obtain pure TaS2 single crystals.

Through traditional routes, which usually consist of simple single step reactions, pure and well crystallized transition metal sulfides can be obtained. However, metastable polymorphs are inaccessible through these high temperature routes. In addition, usually only well crystallized micron sized particles can be obtained, while nanoparticles with large surface areas cannot be synthesized through traditional routes.

1.2.2 Solution Based Routes

Unlike traditional routes, solution based routes are usually carried out in the liquid phase at lower temperatures. One common solution based method is to precipitate transition metal sulfides by reaction of soluble transition metal salts with sulfide salts or hydrogen disulfide in aqueous solutions, as shown in the following examples.

ZnSO4(aq) + (NH4)2S(aq) ZnS + (NH4)2SO4(aq) (Equation 1-2)

CdCl2(aq) + H2S CdS + 2HCl(aq) (Equation 1-3)

Because most transition metal sulfides are extremely insoluble in water, these aqueous precipitation reactions can occur rapidly at room temperature. However, both the starting materials and products involved in such reactions have to be stable in air and water.

Therefore, unstable or oxygen sensitive metal sulfide polymorphs cannot be synthesized by this method.

9

Another common solution based method is the solvo-thermal method, where soluble transition metal and sulfur sources are dissolved in solvents and then heated in autoclaves at low temperatures (usually less than 200 °C ) for a few hours to several days.

The metal precursors can be inorganic salts, such as halides, sulfates and nitrates, or organic compounds, such as acetates and acetylacetonates.33-35 The sulfur sources can be inorganic acids or salts, including hydrogen sulfide, polythionic acids, sodium polysulfides and thiosulfate, or organic compounds such as thiourea and thioacetamide.

33-36 The solvents used in such solvo-thermal reactions can be water or organic solvents, such as ethanol, aqueous ammonia and ethylene glycol.34,35,37,38 Many transition metal sulfides obtained by such solvo-thermal methods have been reported to be nanocrystalline. Thus, by carefully selecting the precursors and controlling the synthetic conditions, a number of transition metal sulfide polymorphs, including some metastable polymorphs and specific morphologies, were obtained through solvo-thermal routes.33-

35,37,38

1.2.3 Other Routes

In order to obtain homogeneous and stoichiometric products, syntheses by traditional solid state routes usually requires high temperatures, long reactions times and sometimes repeated mechanical grinding to overcome the diffusion barriers. In order to overcome the disadvantages of traditional solid state reactions, in the late 1980s and early

1990s, Kaner's group reported a series of rapid and low initiation temperature syntheses of transition metal sulfides from solid state reactions of transition metal halides with alkali metal sulfides.39-42 These solid state metathesis reactions are highly exothermic and self-propagating once the mixtures are ignited by a hot filaments or an intense flash of

10

light.39-42 Compared to traditional solid state methods, products synthesized by solid state metathesis reactions are usually highly crystalline and homogeneous, and the reaction time is very short (order of seconds).39-42

In order to obtain new transition metal sulfides and control their morphologies, new methods have been developed as well. For example, nanoparticle precursors can be used to control the morphology of the resulting sulfide nanoparticles.43 It was reported that MnS, FeS2, Co9S8 and Ni9S8 nanoparticles were synthesized by reactions of the corresponding oxide nanoparticles with sulfur at temperatures below or at 370 °C .43

Pyrolysis of transition metal complexes can be used to prepare transition metal sulfide

44,45 thin films. Cu2-xS and ZnS thin films were obtained by the pyrolysis of metal thiocarbamide chlorides.44,45 This method can be used to fabricate thin film solar cell devices as Cu2S and CuInS2, which are promising solar cell materials.

1.3 Non-hydrolytic Sol-gel (NHSG) Methods

Compared with traditional solid state routes, sol-gel processes provide low temperature synthetic routes to homogeneous inorganic materials such as metal oxides. In traditional sol-gel routes to metal oxides, the products are results of the hydrolysis of metal alkoxides.46 However, the reaction rates of sol-gel processes are dependent on the metals.46 This limits their usefulness for ternary or higher oxides for certain metal combinations, because the products can be inhomogeneous or simply mixtures of different metal oxides if hydrolysis rates are very different for the individual metals.

Hence, in order to improve sol-gel processes and obtain homogeneous metal oxides, alternative routes have been explored. The formation of silica from alkoxysilanes has

11

been known since the late 1940’s,47-49 and similar reactions were named non-hydrolytic sol-gel (NHSG) methods and more thoroughly developed and explored since the early

1990s.50-56 In NHSG routes, metal oxides are usually prepared by reactions of metal halides and secondary or tertiary ethers in organic solvents. Metal alkoxides can also be used as precursors, or they can be formed in situ. Corriu and coworkers51 studied the mechanisms of NHSG routes and proposed the mechanism depicted in Figure 1-4:

Scheme illustrating the mechanism of non-hydrolytic sol-gel routes to metal oxides..

Metal alkoxides are formed in situ, followed by breaking of the carbon-oxygen bond of the alkyl groups.51,52 Unlike traditional sol-gel methods, the limiting step in NHSG routes is the breaking of carbon-oxygen bonds, which is largely independent of the types of metals. Hence, NHSG routes are suitable for preparing metal oxides with very slow hydrolysis rate of the corresponding metal precursors, as well as oxides with multiple metal components that show different hydrolysis rates.

MXn R R X M X n-1 + O RO M Xn -1 M Ox + R X

Figure 1-4: Scheme illustrating the mechanism of non-hydrolytic sol-gel routes to metal oxides.

In the late 1980s and early 1990s, Schleich et al. reported the synthesis of transition metal sulfides by methods comparable to non-hydrolytic sol-gel routes to metal oxides.57-61 Similarly to the methods used for preparing metal oxides, metal sulfides can be obtained by reactions of metal halides and thioethers in organic solvents. The as-

12

recovered materials may be amorphous, in which case heat treatment may yield crystalline metal sulfide phases. Schleich's group successfully prepared a variety of transition metal sulfides, such as polymorphs of iron sulfides, vanadium sulfides, manganese sulfides, tungsten sulfides, molybdenum sulfides, titanium sulfides, chromium sulfides and niobium sulfides.57-61 It is also possible to obtain transition metal sulfide polymorphs with different stoichiometries by careful choice of oxidation states of the metal halide starting materials. For example, Schleich and Martin demonstrated the preparation of amorphous MoS2, MoS2.5 and MoS3 by reactions of hexamethyldisilthiane

57 (HMDST) with MoCl4, MoCl5 and MoF6, respectively. Interestingly, a novel molybdenum sulfide with an usual stoichiometry, MoS2.5, was prepared for the first time through this NHSG route. Similarly, an amorphous niobium sulfide with an usual stoichiometry of NbS2.5 were prepared by Bensalem and Schleich using NHSG methods.60 Although the as-recovered precipitates were generally amorphous, it is possible to obtain crystalline phases by heat treatment. Bensalem and Schleich obtained crystalline TiS2 and Ti3S4 by heating the amorphous TiS2 obtained by reactions of TiCl4 with HMDST and di-tert-butylsulfide (DTBS), respectively, to 650 °C.59 In addition, amorphous materials with different compositions, such as amorphous VS3 and VS4 could indicate that local coordination environment was different in these as-recovered amorphous sulfides, which required further analysis.61 These preliminary results by

Schleich's group indicate that it is possible to prepare various transition metal sulfides with different stoichiometries or even new polymorphs of sulfides through NHSG methods.

13

1.4 Theoretical Modeling in Solid State Chemistry

Theoretical modeling is widely used in various scientific fields. For solid state chemistry, computer modeling techniques have been applied to predict the structures and properties of materials. Moreover, computational methods are also applied to design of new materials for specific applications, and to guide experimental work for the creation of such new materials.62 For example, recently, Chookajorn et al.63 developed a theoretical method to predict the stability of nanocrystalline alloys, which are generally unstable, and then successfully synthesized a predicted W-Ti nanocrystalline alloy with promising high-temperature properties. There are many theoretical methods for modeling solid state materials, which can be categorized into three general types: Energy minimization,64-67 molecular dynamics68,69 and quantum Monte Carlo methods.70-72

Energy minimization is an approach used to determine the structural configuration with minimum energy for a given compound. In this approach, the total energy of each configuration is usually calculated from first-principle quantum mechanical methods. However, since the Schrödinger equation of a many electron system cannot be solved, an approximation method, such as density functional theory (DFT),73,74 is used to estimate the total energy of the system. Energy minimization is a fast and efficient approach to the optimized structure of a given material. Furthermore, once the structure is determined, materials properties, such as band gaps, band structures and elastic properties, can be obtained. However, the energy minimization approach can only calculate the energy at zero Kelvin because the kinetic energy of each atom simulated is omitted. In addition, depending on the starting simulation structure, a local energy

14

minimum instead of the global minimum may be reached. In this case, the obtained structure is not the most stable structure, which can result in prediction of an incorrect structure.75

Compared to energy minimization, usually a much larger group of atoms is sampled in molecular dynamics (MD) simulations. Moreover, a random kinetic energy is assigned to each atom, with the limiting condition that the kinetic energy of these atoms follows a Boltzmann distribution, and that the overall momentum of the system is zero.75

Hence, by applying Newton's law to these atoms, their motion can be simulated. Once the equilibrium state is reached, the structure of the system can be obtained. Because the equilibrium is established by the evolution of many finite time intervals, the lattice dynamics of the process are recorded, which is useful for characterizing phase transitions or reactions. In addition, unlike the energy minimization approach, temperature information is included in MD simulations. However, because the number of sampled atoms is much larger than for the energy minimization approach and time evolution usually requires at least 100 steps, MD simulations are highly time and resource consuming processes.75

Similar to MD, Monte Carlo (MC) techniques also simulate consecutive configurations of a collective of atoms, but a MC algorithm involves many random processes instead of time evolution. In MC simulations, millions of configurations can be generated, and in each configuration, all the atoms fulfill Boltzmann statistics. MC simulations are suitable for studying adsorption of organic liquids in porous solids, diffusion processes, and complex energy surfaces.75 MC techniques are very powerful

15

tools in molecular modeling, but similar to MD, they require extensive computational time.75

1.5 Goals of Thesis Research

The main goal of this thesis research is to explore the usefulness of the non- hydrolytic sol-gel (NHSG) method for the synthesis of transition metal sulfides. Although

NHSG methods are very promising for preparing metal sulfides, to our knowledge, they have not yet been thoroughly studied, and there are a number of synthetic parameters that need to be explored. In order to further understand the NHSG route, it is necessary to systematically study the influence of synthetic parameters, such as metal precursors, thioethers, solvents, temperature, concentration of starting materials and reaction time.

By exploring these synthetic variables, it may be possible to discover new metal sulfide polymorphs or metal sulfide nanoparticles with interesting morphologies as demonstrated by the thesis project of a former group member, Nathalie Pedoussaut.76,77

Theoretical modeling of possible phase transition paths of metal sulfides may help to improve control of phase selection, or to obtain new phases. Many transition metals can form several different polymorphs. Knowledge about pressure-induced or temperature-induced phase transition pathways may aid in designing specific synthesis conditions that result in these polymorphs. Moreover, it is even possible to predict new polymorphs by finding intermediate phases on the transition pathways. In addition, computational materials design is particular useful for guiding our future synthesis of metal sulfides with multiple metal components by predicting which compositions may exhibit promising properties. Therefore, in this thesis, we studied a relatively simple case,

16

pressure-induced phase transitions in barium chalcogenides, to demonstrate how theoretical work may help with our future experimental work.

17

Chapter 2

2. Experimental Methods

2.1. Synthetic Method

In this thesis project, metal sulfides were prepared through a non-hydrolytic sol- gel (NHSG) method. This process involves the reaction of metal halides with different sulfurizing agents, di-tert-butylsulfide (DTBS) or hexamethyldisilathiane (HMDST)

(Figure 2-1), in inert organic solvents like chloroform or acetonitrile. The reactions can be carried out either at room temperature (RT) or under solvo-thermal (ST) conditions.

a) b)

Figure 2-1: Structures of a) tert-butylsulfide (DTBS) and b) hexamethyldisilathiane (HMDST).

The two sulfur sources, DTBS and HMDST (Figure 2-1), have different

+ + reactivities. (CH3)3Si is more stable than (CH3)3C , and therefore (CH3)3Si is a better

18

leaving group. Hence, HMDST is more reactive than DTBS. Their different reactivities can be exploited to tune the reaction rate in order to obtain homogeneous metal sulfides.

Most of the chemicals used in this project are unstable with respect to oxygen and moisture, and as a result, all materials used were anhydrous and stored in a glove box under argon. The solvents, chloroform and acetonitrile, were purified by vacuum distillation with phosphorus pentoxide and calcium hydride as respective drying agents.

Synthetic variables explored in this project are different starting materials, solvents, metal to sulfur ratio, concentration of sulfur source, temperature and reaction time. Because of the pungent smells of the thioethers, exposed glassware and air were sprayed with bleach.

2.1.1 Room Temperature Route

Because the starting materials and NHSG routes are sensitive to moisture and oxygen, all reactions at room temperature took place in a glove box under argon. For a routine reaction, about 1 to 3 mmol of anhydrous metal halide were weighed into an

Erlenmeyer flask, and dissolved in 10 or 15 mL of solvent. For neat reactions, the step for the addition of solvents was skipped and no solvent was used. If a solvent was used, the solution was usually stirred for five to ten minutes. Then, between 1.5 and 32 mmol of sulfurizing agent were slowly added to the flask by syringe. After the addition of the sulfur source, the flask was capped to avoid evaporation of the solvent and sulfur source.

The solution was stirred for a few hours to several days. The product was recovered by filtration. The filtered precipitate was dried and then stored in a vial inside the glove box.

19

2.1.2 Solvo-thermal Route

Similarly to room temperature reactions, syntheses under solvo-thermal conditions were also carried out in closed systems to avoid contact with water and oxygen. For a typical reaction, inside a glove box under argon, about 1, 1.5 or 3 mmol of anhydrous metal halide were weighed into a glass ampoule and dissolved in 5 to 15 mL of solvent. For neat reactions, the step for the addition of solvents was skipped and no solvent was used. If a solvent was used, the solution was usually stirred for five to ten minutes. Then, between 2 and 19 mmol of sulfurizing agent were slowly added into the ampoule by syringe. The mixture was stirred for about 10 to 15 min, then the ampoule was capped with a septum and transferred outside the glove box, where it was immediately immersed in liquid nitrogen. After the contents of the ampoule were frozen, it was connected to a Schlenk line, and sealed under vacuum. While this procedure was followed for all reactions, it was realized towards the end of this project that it would be better to connect the ampoule to a Schlenk line before freezing in liquid nitrogen, as traces of atmospheric moisture could condense on the frozen ampoule contents while connecting the ampoule to the line. Also, after the ampoule is connected to the Schlenk line, a quick vacuum should be applied to remove trace oxygen that could have entered while connecting the ampoule to the Schlenk line. After warming to room temperature, the ampoule was transferred to an oven with an accuracy of ±1 °C and heated to a temperature ranging from 70 °C to 150 °C for 7 to 120 d. Samples prepared by solvo- thermal reactions were under autogenous pressure. It is not possible to independently vary temperature and pressure in a closed system. Ampoules were enclosed in a metal can so that glass would be contained in case of an explosion. The cooled ampoule was

20

transferred back into the glove box before opening, so that the sulfide product would not be oxidized. Powder-like precipitates were recovered by filtration. Sticky tar-like precipitates, which were usually stuck on the bottom of the ampoules and could not be filtered, were recovered as powders by decanting excess solvent and drying by residual solvent evaporation under vacuum. After the as-recovered precipitates were dried, they were stored in vials under argon.

2.1.3 Heat Treatment

As-recovered samples might contain organic residues or unreacted metal halides, and some of the raw samples were amorphous. Heat treatments were carried out to remove impurities and improve crystallinity. In order to avoid oxidation, powders were packed in a molybdenum boat inside a quartz tube inside a glove box. Titanium wire balls were used as upstream oxygen scavengers. Valves were connected to both ends of the quartz tube, so that the quartz tube could be closed off from air when it was taken out of the glove box. The assembly was placed into a tube furnace and connected to an argon tank. The outlet valve was connected to a bleach bubbler, which was used to destroy volatile sulfur-containing compounds. With this assembly, argon can flow through the tube and prevent oxidation of the powders. All parts of the assembly were pre-heated and kept either under vacuum or inside the glove box.

2.2. Characterization Methods 2.2.1 CHNS Analysis

CHNS analysis is a type of analytical method that gives the carbon, hydrogen, nitrogen and sulfur content in a sample. A Perkin Elmer PE-2400 II CHNS/O analyzer

21

was used for all CHNS analyses in this thesis. As shown in Figure 2-2, the instrument is based on frontal gas chromatography (GC) and consists of four main parts, combustion and reduction tubes, a mixing chamber, a GC column and a thermal conductivity detector.

Depending on the type of combustion and reduction tubes, the analyzer can be set to

CHN, CHNS or oxygen mode. For CHNS analysis, both combustion and reduction occur in the same tube. The upper part of the tube is filled with tungsten trioxide, which ensures that the combustion products are CO2, H2O, SO2 and NO2. The lower part of the tube is filled with copper wire, which can reduce NO2 to N2. The combustion and reduction products, CO2, H2O, SO2 and N2, are mixed in a mixing chamber and then transferred into a separation column under pressure. Then the separated gases are detected by a thermal conductivity detector.

In CHNS analysis, the absolute weight of carbon, hydrogen, nitrogen and sulfur can be directly measured and their contents are reported as weight percentage in each sample. In this project, the sulfur content and the residual carbon of a sample were measured by CHNS analysis. Samples used for CHNS analysis (ranging from about 1.0 mg to 6.0 mg) were weighed on an analytical balance and encapsulated in tin capsules. In order to avoid oxidation, tin capsules were weighed and transferred into the glove box.

The samples were then enclosed in folded capsules, which were taken out of the glove box and weighed again. For each new tube prepared, empty tin capsules (blanks) and cysteine (C6H12N2O4S2) standards were run before samples to calibrate the instrument.

Helium was used as carrier gas because of its high thermal conductivity, which is favorable for thermal conductivity detectors. All the chemicals used in CHNS analyses were purchased from Perkin Elmer.

22

Figure 2-2: Components of a PE-2400 II CHNS/O analyzer.

CHNS is an efficient and relatively accurate method for elemental analysis of compounds that contain carbon, hydrogen, nitrogen and sulfur. However, the accuracy of

CHNS analysis results depends strongly on the operating conditions. Similar to any analytical method, proper data evaluation has to be carried out to assess the reliability of the results. An example of data evaluation and uncertainty estimates is given in the following text.

Error! Reference source not found. shows the results of analytical blanks during a set of 73 runs of copper sulfide samples. Blank runs No.7 to No. 18 were carried out for initial calibration, and blanks 32 to 68 were inserted between sample runs to correct the baseline. The K2 factor used to calculate the absolute weight is also provided in the table. The uncertainty (σ) of each element is calculated using equation (2-1).

Uncertainty of element X (in mg) = STDX / (K2X × 100) (Equation 2-1)

23

Table 2.1: Analytical blanks during CHNS analysis and uncertainty estimates. No. C H N S BLANK 7 -17 426 706 47 BLANK 9 -5 407 705 38 BLANK 11 -10 380 701 44 BLANK 13 -9 346 699 33 BLANK 15 -9 334 699 32 BLANK 18 -8 364 699 44 BLANK 32 10 254 580 45 BLANK 45 -4 273 550 41 BLANK 57 -10 250 483 48 BLANK 65 -14 121 470 56 BLANK 68 -9 205 466 43 STD 6.9 93.4 105.4 6.8 K2 14.8 39.0 5.1 7.2 Uncertainty 0.0047 0.024 0.21 0.0095 (mg)

K2X is the K2 factor of element X and STDX is the standard deviation for element

X. The obtained absolute uncertainty (in mg) for each element is shown in Table 2.1.

Because the signals for hydrogen and nitrogen declined over time, their standard deviations are very large compared to carbon and sulfur. As most of the samples only contain negligible amounts of hydrogen and nitrogen, their errors were very large, making the hydrogen and nitrogen data useless. However, the carbon content can be used to investigate the completness of the reaction, and the sulfur content allows calculation of sample stoichiometry when combined with other methods.

As shown in Table 2.1, carbon and sulfur signals of blanks were very consistent and within the range suggested in the manual. Considering the actual amount of sulfur in the samples, the relative uncertainty (σ) for sulfur was usually less than 3%. However, because the carbon content in most of the samples was negligible, the relative uncertainty of carbon was very large (above 30%). For a large portion of the samples, the amount of

24

carbon detected was within the detection limit (3σ). Thus, measurements for carbon content do not accurately reflect the actual amount of carbon in the samples, but indicate how well reactions proceeded towards completion.

2.2.2 Thermogravimetric and Differential Thermal Analysis

Thermogravimetric and differential thermal analyses (TG/DTA) were carried out for elemental analysis on a TA Instruments SDT 2960 Simultaneous DTA-TGA. In a

TG/DTA instrument, a dual balance mechanism is applied through two balance cantilevers with thermocouples attached, allowing weight and temperature of a sample and a reference to be measured simultaneously. The cantilevers are enclosed by a tube furnace. Gas flow through the tube furnace can be used to create an inert or oxidizing atmosphere during a measurement. In addition, a constant gas flow facilitates removal of volatile contents. Most of the TG/DTA runs in this project were carried out by heating samples in platinum pans to 800 °C to 1000 °C with a temperature ramp of 10 °C/min in air with a flow rate above 100 mL/min.

During a measurement, the instrument records both the weight of the sample and the temperature difference between the sample and a reference pan as a function of temperature, which are plotted as thermogravimetric (TG) and differential thermal analysis (DTA) curves, respectively. Sudden weight gains or losses in a TG curve are usually due to oxidation or decomposition of the sample or organic residues. DTA curves show exothermic or endothermic events, which can provide information about phase transitions and the stability of samples. All TG plots presented in this thesis were baseline corrected.

25

The metal content of the samples was calculated from the weight before and after a TGA run in air. Both the initial and final weight were measured at room temperature. For each metal sulfide system, the residue of a TGA run was characterized by powder X-ray diffraction to ensure that the residue was a phase pure metal oxide. If the residue is composed of a metal oxide MOx, then the metal content in the initial sample can be calculated according to equation 2-2.

mM% = (mf × MM / Mox) / mi (Equation 2-2) where mM% is the weight percentage of the metal M, mi is the initial mass of the sample before the TGA run, mf is the final mass after the TGA run, MM is the molar mass of the metal and Mox is the molar mass of the metal oxide with the formula MOx.

The metal to sulfur (M/S) ratio in a metal sulfide sample can be calculated by combining CHNS and TGA data. TGA data alone can be used to calculate the M/S ratio if the starting sample is composed entirely of metal sulfide. Uncertainties were calculated from the results of several analytical blank runs, and consecutive runs for the same sample. These estimates reflect the error of the whole analysis process instead of only the balance.

2.2.3 Powder X-ray Diffraction

In this project, powder X-ray diffraction (PXRD) was used for phase identification of both as-recovered and heat treated samples. Most PXRD data were collected using Cu-Kα radiation on a Scintag XDS-2000 diffractrometer in Bragg-

Brentano theta-theta configuration with a Moxtek MXP-D1 solid-state detector. Two sample stages, a flat stage for room temperature measurements, and an Anton Parr heating stage with a platinum heater strip for variable temperature (VT) experiments,

26

were used to study the phases of samples. For several samples, due to small amounts of samples left or requirement for high signal to noise ratio, they were packed in 0.2 mm capillaries and scanned on a capillary stage with a PANalytical X’pert Pro diffractometer in Bragg-Brentano configuration with an X’Celerator detector. The collected data were analyzed in Jade 81 and compared to the ICDD PDF-2 database.2 For some samples,

Rietveld refinements were carried out using TOPAS-Academic V4.3

Samples measured at room temperature were packed inside a glove box in a special air sensitive sample holder (Figure 2-3). In this design, powders are protected from air by an O-ring seal and a Kapton dome, which is transparent to X-rays.

Figure 2-3: Air sensitive PXRD sample holder for Scintag XDS-2000 diffractrometer.

Samples characterized by variable temperature XRD were stored in vials filled with acetonitrile to minimize oxidation during transfer. To characterize a sample on the

VT stage, a slurry of the sample in acetonitrile was spread evenly on the platinum strip of

27

the VT stage. After the acetonitrile evaporated, a thin film of the sample was formed on the platinum strip. The temperature of the stage was controlled by a thermocouple attached to the strip. In order to prevent oxidation at high temperatures, a vacuum was applied to each sample measured on the VT stage.

Because the detector of the instrument occasionally generates voltage spikes, a program was written to filter these spikes and correct the data. Instructions for use of the software and the source code of the algorithm are provided in Appendix A.

2.2.4 Scanning Electron Microscopy and Energy Dispersive Spectroscopy

In this project, scanning electron microscopy (SEM) was used to study the morphology of both as-recovered and heat treated samples, and energy dispersive X-ray spectroscopy (EDS) was used to study the elemental composition and homogeneity of the samples. Samples were examined on a JEOL JSM-7500F field emission scanning electron microscope with a BRUKER XFlash 5010 series EDS detector. EDS data were collected from uncoated samples and analyzed using the QUANTAX Spirit 1.9 microanalysis software. Elemental analysis by EDS is subject to larger uncertainties than

CHNS and TGA, and was therefore mainly used when CHNS and TGA were not suitable, for example, for samples containing significant amounts of starting materials.

An example of an SEM image of a copper sulfide sample prepared with CuCl as the copper source is shown in Figure 2-4 a), and the corresponding elemental hyper maps of Cu, S, Cl and an overlay of all three elements are shown in Figure 2-4 b), c), d), and e), respectively. The boxes on the SEM images indicate the areas selected for microanalysis.

Besides the areas shown in Figure 2-4 a), areas selected from other parts of the sample

28

were also used to obtain the average elemental composition. A common error encountered during EDS analysis is due to microabsorption. Heavier elements may block low energy X-rays emitted by lighter elements. In this case, the X-ray energies used for microanalysis for sulfur and chlorine were 2.308 eV (S-Kα1) and 2.622 eV (Cl-Kα1), respectively, while the energy for copper was 8.048 eV (Cu- Kα1), which is significantly higher than those of sulfur and chlorine. Hence, it is possible for copper to block the low energy photons emitted by sulfur and chlorine, which leads to an overestimation of the sample’s copper content. For example, the dark area in Figure 2-4 c) and d) indicates that no S or Cl were present, but Figure 2-4 b) indicates the presence of Cu in this area. Hence, the results of this elemental hyper map seem to indicate the presence of large amounts of elemental copper, which is highly unlikely to be formed. However, the dark area in

Figure 2-4 c) and d) may also be a result of shadowing by the particles marked with an oval in Figure 2-4 a). As a result of the shadowing effects, the areas selected for analysis should not be surrounded by higher areas, which may block low energy X-rays from reaching the detector.

In order to calculate average compositions, it is assumed that all measured areas equally represent the sample so that the elemental compositions can be calculated by averaging the results of the measured areas.

29

Figure 2-4: a) SEM image of a copper sulfide sample and distributions of b) Cu, c) S, d) Cl and (e) all three elements in the image. The boxes in the SEM image indicate the areas selected for elemental analysis.

30

Chapter 3

3. Synthesis and Characterization of Copper Sulfides

3.1 Introduction

Copper sulfides have been known to mankind for thousands of years, and

1,2 naturally occurring CuSx (0.5 ≤ x ≤ 2) minerals are a major source of copper metal.

However, their properties are not quite understood yet because of the complicated structures of various non-stoichiometric polymorphs.1-8 Among the 13 known Cu-S binary phases, several have only been discovered or identified over the last few decades, such as djurleite in 19583 and geerite in 1980.6 After the discovery of a new polymorph, it usually took years or decades to ultimately obtain accurate structural information, which is essential to understand the properties of materials. Even though structure determination in the Cu-S system is no longer an obstacle, studying the properties of copper sulfides remains very challenging because of the high extent of disorder in the Cu-S system.2,5,9

Even the oxidation state of copper in the simplest Cu-S binary compound, covellite

(CuS), has been under discussion for the past decades.10

31

3.1.1 Applications of Copper Sulfides and Synthetic Challenges

Copper sulfides possess interesting semiconducting, electronic and optical properties, which can be applied to gas sensors, solar cells, capacitors and transistors. 9,11-

15 9 Naturally occurring Cu2S is often found to be highly copper deficient. This makes

Cu2-xS a good p-type semiconductor with low cost. It was used as a coating on CdS substrates during the 1980’s to fabricate early solar cells that achieved efficiencies up to

9%.9,16 However, because of the poor stability of copper sulfides in air and moisture, the performance of CuS/CdS based solar cells rapidly deteriorated under normal use conditions.17 Although this performance loss can be reversed by subsequent heating in

17 mixtures of hydrogen and air, the instability of Cu2S has limited its use in solar cells, as cost and downtime made the process unfeasible. Nevertheless, because of the recent advancements of nanotechnology and solar cell technology, copper sulfide nano-particles have attracted renewed attention, and have great potential in the fabrication of stable solar cell devices with high efficiencies.18-23 Copper sulfides appear in many novel nanostructured solar cell designs, such as 3D nanostructured and extremely thin absorber solar cells.24,25

In addition to the solar cell industry, copper sulfides also have applications as sensors, transistors and capacitors.11-15 Copper sulfides with hollow structure can be used

11,12 as ammonia gas sensors, and Cu2S nano-particles can be embedded into ordered mesoporous carbons to fabricate hydrogen peroxide sensors.13 A recent study showed that

CuxS may be a suitable electrode material in organic field-effect transistors, which have the potential to replace conventional silicon transistors.14 Chalcocite, a copper(I) sulfide polymorph, has potential applications as a supercapacitor.15 However, these potential

32

applications also bring substantial synthetic challenges with them.

Many modern applications demand not only specific stoichiometries and electrical properties, but also depend on particle size and morphology control. For instance, unlike early CuxS based solar cells, the new devices require copper sulfide nanoparticles with specific morphologies, such as quantum dots,21 thin films25 and nanowires.26 Traditional synthetic routes that rely on the reaction between elemental copper and sulfur at high temperature are not suitable for the production of nanoparticles as usually micron sized particles are obtained.27 It is also not feasible to use traditional coating methods like physical vapor deposition (PVD) and chemical vapor deposition

(CVD) when particle shapes other than thin films are desired. In addition, vapor deposition methods can be costly, as they require specialized equipment or expensive precursor materials. In comparison, chemical bath deposition (CBD) methods offer a relatively low cost approach, which allows in situ film growth and deposition on a substrate in solution.20 In order to apply CBD methods to coat copper sulfides, proper synthetic conditions for the targeted synthesis of specific copper sulfide phases with suitable morphologies must be established.

This chapter explores the non-hydrolytic sol-gel synthesis of copper sulfides.

NHSG routes are attractive for the in situ deposition of homogenous products under mild conditions. Thus, it is important to explore NHSG routes in the Cu-S binary system to establish control over phases and morphologies of the products. Moreover, NHSG routes may be extended to the synthesis of homogenous ternary or quaternary sulfides so that better photovoltaic materials may be discovered. For example, a CuInS2/CdS based solar cell has been reported to achieve an efficiency of 10.2%.28 By adding other metals to the

33

Cu-S system, increased resistance to oxidation may be achieved, which would benefit the long term performance.

3.1.2 Copper Sulfide Polymorphs

Copper sulfide forms 13 different polymorphs (arguably for roxbyite), most of which have very broad composition ranges.2,3,5 The sulfur atoms in hexagonal or monoclinic crystal systems usually form hexagonal close packed lattices (low chalcocite, high chalcocite, djurleite), while they usually form cubic close packed sublattices in cubic or distorted cubic crystal systems (digenite, tetragonal chalcocite). The stability ranges of these polymorphs are shown in Table 3.1. The phase diagram of the Cu-S system is shown in Figure 3-1, and a Cu-S phase diagram in the Cu2S-Cu1.75S composition range is shown in Figure 3-2.

Table 3.1: Stability ranges and crystal systems of the 13 known copper sulfide polymorphs (adapted from Ref. 2).

Crystal Phase (Mineral Name) Formula System Stability

Low Chalcocite Cu2-xS Monoclinic < 103 °C

High Chalcocite Cu2-xS Hexagonal 103 °C to 435 °C

Tetragonal Chalcocite Cu1.96S Tetragonal > 0.8 kbar * Low Digenite Cu1.77-1.83S Rhombohedral < 83 °C

High Digenite Cu1.8+xS Cubic > 76 °C

Djurleite Cu1.94-1.97S Monoclinic < 93 °C

Roxbyite Cu1.78S Monoclinic Not Available

Anilite Cu1.75S Orthorhombic < 76 °C

Geerite Cu1.6S Rhombohedral Not Available

Spionkopite Cu1.4S Hexagonal Not Available

Yallowite Cu9S8 Hexagonal < 157 °C Covellite CuS Hexagonal < 507 °C

Villamaninite CuS2 Cubic Not Available 1 * The stable composition at room temperature is from Cu1.77S to Cu1.79S.

34

Monoclinic chalcocite is the thermodynamically stable phase of Cu2S at room temperature.9 It transforms into hexagonal chalcocite at 103 °C. Above 435 °C, hexagonal chalcocite converts to a cubic phase, which is isostructural with high digenite, but shows a higher percent occupancy of the copper sites.2 Monoclinic chalcocite possesses a distorted structure closely related to hexagonal chalcocite, and its stoichiometry can vary from Cu1.98S to Cu2S, as Cu defects are favored in this structure.9,29 Unlike the other two chalcocite polymorphs, sulfur atoms in tetragonal chalcocite are cubic close packed. It is a high pressure phase ( > 0.8 kbar), but the material can be quenched to ambient conditions.30

Figure 3-1: Phase diagram of the Cu-S system. (adapted from Ref. 27)

35

Figure 3-2: Phase diagram of the Cu-S system in the Cu1.7S-Cu2S composition range. Cv, An, Dg, Dj, Lc and Hc indicate covellite, anilite, digenite, djurleite, low chalcocite and high chalcocite, respectively. (adapted from Ref. 31)

Low digenite, Cu1.77~1.83S, is thermodynamically stable up to about 76 °C to 83 °C, where the exact temperature depends on its composition,1 and transforms to high digenite at higher temperatures. High digenite adopts a face centered cubic structure in space group Fm m. In this cubic unit cell, sulfur atoms are found on the corners and face centers, while the copper atoms occupy 90% of the 8 tetrahedral site, but with slight displacements from the center, thus generating 24 possible equivalent positions with lower occupancies.2,3 Low digenite is related to high digenite by a rhombohedral distortion.2,3 High digenite forms solid solutions with different copper occupancies at

2,27 high temperatures, which allow its stoichiometry to vary from Cu1.73S to Cu2S. High digenite coexists with high chalcocite at high temperatures.

36

Djurleite was first discovered by Djurle in 1958 by sintering pressed pallets of elemental copper and sulfur powders at high temperatures.3 Of these samples, the one with a 1.96:1 copper to sulfur ratio resulted in the formation of a major djurleite phase with a small amount of tetragonal chalcocite after annealing at 75 °C for a month.

Roseboom also reported the phase found by Djurle from annealed samples synthesized by heating mixtures of elemental copper and sulfur with the composition close to Cu1.96S in sealed ampoules, but found pure djurleite was very difficult to obtain by this method.1

Because of the structural complexity of djurleite, its structure was not fully understood until the XRD analysis on a very rare untwinned single crystal was reported by Evans in

1979.5 The monoclinic unit cell of djurleite consists of two types of hexagonal close packed sulfur atoms (16 each). In each unit cell, 20 and 32 copper atoms form triangular coordination with sulfur atoms in and between sulfur layers, respectively. There are also nine copper atoms in distorted tetragonal coordination, and one in linear coordination with sulfur atoms.

Roxbyite was first claimed to be prepared by electrochemical oxidation of chalcocite,32,33 and was named after Roxby Downs, Australia, where the natural mineral of roxbyite was found.34 The compositions of electrochemically prepared roxbyite have

32 33 34 been reported to be Cu1.75S and Cu1.80-1.86S, and Mumme et al. reported compositions between Cu1.74S and Cu1.82S for several natural mineral samples, which contain small amount of iron. The composition of roxbyite in literature is usually referred to as Cu1.78S (mean of 1.74 and 1.82). Roxbyite was reported to be monoclinic in space group C2/m with a = 53.79 Å, b = 30.90 Å, c = 13.36 Å and β = 90.0°. However, the lattice parameters a and b of roxbyite are twice as those of djurleite, and the reflections of

37

roxbyite could also be found in djurleite. It is possible that the so-called roxbyite is a super structure of djurleite. Hence, unless there is atomic level structure determination for roxbyite, it cannot be confirmed as a new copper sulfide polymorph.34

Anilite, Cu1.75S or Cu7S4, exhibits an orthorhombic structure in space group

Pnma.35 It is stable below 76 °C, and decomposes to digenite and covellite above 76 °C .2

Geerite, yarrowite and spionkopite, which are not shown on the phase diagrams, were

6 found in minerals. Geerite, Cu1.60S, was discovered in 1980. It adopts a rhombohedral structure in space group R m.36 Its XRD pattern is very similar to cubic CuCl. Thus it

6,36 was reported as a cubic structure initially. Yarrowite, C9S8, and spionkopite, Cu39S28 were named after the locations where they were discovered, Yarrow Creek and Spionkop

Creek Valley, Alberta, Canada.37

Covellite has a hexagonal structure with 2/3 and 1/3 of copper atoms in tetrahedral and triangular coordination, respectively, and 2/3 of the sulfur atoms forming

38 S-S bonds. X-ray absorption spectra at the Cu LIII-edge and X-ray photoelectron spectroscopy indicated that only monovalent copper was present for both tetrahedrally and triangularly coordinated copper sites.10,39 Pattrick et. al.39 proposed that in each non- bridging sulfur layer, there were electron holes in the valance band of S2- ions, resulting

2- 2- in a formula of Cu(I)3S2 S ∙, where ∙ depicts electron holes. Villamaninite (CuS2) crystallizes in a pyrite type cubic structure,40 and similar to covellite, only monovalent copper was found in villamaninite.41

38

3.1.3 Literature Routes for the Preparation of Copper Sulfides

3.1.3.1 Solid State Routes

Copper sulfide syntheses by solid state methods usually involve the direct reaction between elemental copper and sulfur.1,3,27,30,42-44 To increase the reactivity, the starting materials are ground into fine powders. In order to ensure the purity of the starting materials, the copper powders are heated in a mixture of hydrogen (10%) and nitrogen (90%) to remove oxide impurities, and sulfur is sublimed in an evacuated glass

27 tube or recrystallized from CS2. High purity elemental copper and sulfur are then mixed at accurate stoichiometric ratios to obtain targeted binary phases. Reactions can be carried out by sintering a pellet of a pressed mixture of copper and sulfur in a sealed silica or quartz ampoule, or through mechanical alloying at temperatures between 500 °C and 700 °C.3,27,30,43 In order to prevent oxidation, the ampoule is usually pre-heated in vacuum, and the ball milling process may be carried out under argon. Alternatively, instead of mixing the starting elements at the beginning, they can also be placed separately at opposite ends of the same sealed ampoule.44 The sulfur end of the ampoule is heated at 425 °C, while the copper end is heated to above 450 °C.44 In this process, the sulfur is vaporized and reacts with the copper at the other end of the ampoule.44

Traditional solid state methods are well suited to produce a number of different copper sulfide polymorphs, such as monoclinic, hexagonal, and tetragonal chalcocite, djurleite, digenite, anilite and covellite, by controlling the stoichiometry, reaction temperature and annealing process. One disadvantage of these direct elemental reactions is that they require high temperatures or intensive ball-milling. The obtained products may be heterogeneous and require further heat treatments for very long times. In addition,

39

traditional solid state routes do not provide particle size or morphology control. The products are therefore rarely suitable for the fabrications of solar cells, sensors and electronic devices.

3.1.3.2 Solution Based Routes

Solution based methods usually allow the preparation of more homogeneous products whose phase purity and morphologies can be controlled. A typical solution based reaction is carried out by the reaction of a mixture of soluble copper and sulfur sources in a solvent at room temperature or mild temperatures between 60 °C and 180 °C for times ranging from several hours to several days. The copper sources can be inorganic copper salts, such as CuCl, CuCl2, CuSO4, and Cu(NO3)2, or copper compounds with organic anions, for example, copper(II) acetate and copper(II) acetylacetonate.22,45-47 The sulfurizing reagents can be inorganic acids or salts like hydrogen sulfide, polythionic acids, sodium polysulfides and thiosulfate, or organic compounds such as thiourea and thioacetamide.45-48 Common solvents used in solution based reactions include water, aqueous ammonia or organic solvents, such as ethanol, ethylene glycol and oleylamine.22,46,47 Solution based reactions using elemental copper and sulfur as starting materials have also been reported. For example, Yu et al. reported the synthesis of Cu2S nanowires by hydrothermal reaction of copper foil and sulfur powder in hydrazine and cetyltrimethylammonium bromide.46

Most of the solution based reactions produce copper sulfide nanostructures of various shapes. The crystal structure of the materials can generally be controlled through proper choice of starting materials, and morphology can be varied through changes in synthetic conditions such as stoichiometric ratios, concentration, and reaction time.

40

Syntheses of monoclinic chalcocite,22 hexagonal chalcocite,49,50 djurleite,51,52 digenite,52 roxbyite,50,51 and covellite53-55 nanoparticles have been reported. Various nanostructures such as nanoplates,18,55 nanowires,53 nanorods,52,56 nanospheres,49,57 nanotubes,57 and flower-like46,54 structures have been reported.

Electrochemical techniques,58,59 ultrasound,60 microwave,56,57 serial ionic exchange,61 ionic liquid,62 and photochemical methods63 can be used to assist the syntheses of copper sulfide nanoparticles. Electrochemical synthesis of copper sulfides is usually based on electrolysis of sulfur containing electrolytes with elemental copper electrodes as substrates. For example, Yang et. al reported the synthesis of covellite nanoparticles at room temperature by electrolysis of an aqueous solution of Na2S2O3 and

58 60 56,57 KNO3 with copper foils as electrodes. Ultrasound, microwave, and ultra-violet

(UV) radiation63 have been reported to promote solution based reactions that produce nano-sized covellite particles. Xiong et. al reported the synthesis of copper sulfide hollow spheres by serial ionic exchange.61

41

3.2 Preparation of Copper Sulfides by Non-hydrolytic Sol-gel Synthesis

To prepare copper sulfides, copper chlorides, CuCl or CuCl2, are reacted with different sulfur sources, di-tert-butylsulfide (DTBS) or hexamethyldisilthiane (HMDST), in nonreactive organic solvents like chloroform or acetonitrile. The reactions are carried out either at room temperature or by heating in a sealed ampoule. The exploration of various synthetic conditions includes temperature (from RT to 130 °C), sulfur to copper ratio (from 0.5 to 10), concentration of sulfur sources (from about 0.4 to 4.7 M), volume of solvent (0, 10 and 15 mL) and reaction time (from 5 min to 120 d). Part of the synthetic conditions with DTBS were explored by a former member of our group,

Nathalie Pedoussaut.64 Several samples were also prepared by an exchange student, Anne

Soldat. Previous samples prepared by Nathalie Pedoussaut and Anne Soldat are named with the prefixes "NPSu" and "ACS", respectively. For most of the samples, the amount of copper starting material was fixed at around 3 mmol to simplify comparison of synthetic conditions. About 1.5 mmol of CuCl2 were used for a few samples with high sulfur to copper ratios. All prepared samples in the current work with synthetic conditions are presented in Table 3.2.

42

Table 3.2: Synthetic parameters for copper sulfide samples. Metal Source Sulfur Source Solvent Conditions Rec. Ratio T Sample V V mass Name m (g) Name Name S:M (° Time (mL) (mL) (g) C)

XZ1001 CuCl 0.314 HMDST 2.3 CHCl3 15 3.4 RT 29 h 0.192

XZ1002 CuCl 0.318 HMDST 2.4 CHCl3 15 3.4 RT 3 h 0.202

XZ1003 CuCl 0.312 HMDST 2.3 CHCl3 15 3.5 RT 96 h 0.218

XZ1004 CuCl 0.316 HMDST 0.8 CHCl3 15 1.2 RT 3 h 0.221

XZ1005 CuCl 0.327 HMDST 6.4 CHCl3 15 9.2 RT 3 h 0.241

XZ1006 CuCl 0.315 HMDST 0.8 CHCl3 15 1.2 RT 24 h 0.240

XZ1007 CuCl 0.310 HMDST 6.8 CHCl3 15 10.3 RT 25 h 0.220

XZ1008 CuCl 0.307 HMDST 0.8 CHCl3 10 1.2 RT 3 h 0.233

XZ1009 CuCl 0.314 HMDST 2.4 CHCl3 10 3.6 RT 3 h 0.245

XZ1010 CuCl 0.300 HMDST 6.2 CHCl3 10 9.7 RT 3 h 0.224

XZ1011 CuCl 0.298 HMDST 0.8 CHCl3 10 1.3 70 7 d 0.186

XZ1012 CuCl 0.343 HMDST 0.4 CHCl3 10 0.5 RT 24 h 0.234 XZ1013a CuCl 0.520 HMDST 2.4 N/A N/A 2.2 RT 3 h ** XZ1013b CuCl 0.520 HMDST 2.4 N/A N/A 2.2 RT 24 h ** XZ1013c CuCl 0.520 HMDST 2.4 N/A N/A 2.2 RT 96 h **

XZ1014 CuCl 0.302 HMDST 0.8 CHCl3 10 1.2 130 7 d 0.238

XZ1015 CuCl 0.312 HMDST 0.4 CHCl3 10 0.6 70 7 d 0.228 XZ1016a CuCl 0.529 HMDST 2.4 N/A N/A 2.1 RT 2 h ** XZ1016b CuCl 0.529 HMDST 2.4 N/A N/A 2.1 RT 3 h **

XZ1017 CuCl 0.308 HMDST 0.4 CHCl3 10 0.6 130 7 d *

XZ1018 CuCl 0.317 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.287

XZ1019 CuCl 0.281 HMDST 0.8 CHCl3 10 1.3 100 7 d 0.217

XZ1020 CuCl 0.289 HMDST 2.5 CHCl3 10 4.1 70 7 d 0.231

XZ1021 CuCl 0.271 HMDST 0.3 CH3CN 10 0.6 RT 25 h 0.282

XZ1022 CuCl 0.274 HMDST 2.5 CHCl3 10 4.3 130 7 d 0.203

XZ1023 CuCl 0.276 HMDST 0.8 CH3CN 10 1.4 RT 96 h 0.196

XZ1024 CuCl 0.313 HMDST 2.4 CH3CN 10 3.6 RT 96 h 0.208

XZ1025 CuCl 0.313 HMDST 0.4 CHCl3 10 0.6 130 7 d 0.244 XZ1026 CuCl 0.432 HMDST 2.4 N/A N/A 2.6 RT 3 h 0.229

XZ1027 CuCl2 0.393 HMDST 0.8 CHCl3 10 1.3 RT 96 h 0.235

XZ1028 CuCl2 0.398 HMDST 2.4 CHCl3 10 3.8 RT 98 h 0.252

XZ1029 CuCl 0.303 HMDST 0.4 CH3CN 10 0.6 130 8 d 0.256

XZ1030 CuCl 0.301 HMDST 2.4 CH3CN 10 3.7 130 7 d 0.188

XZ1031 CuCl 0.317 HMDST 0.8 CH3CN 10 1.2 130 7 d 0.211

XZ1032 CuCl2 0.159 HMDST 2.4 CHCl3 10 9.6 RT 96 h 0.074

XZ1033 CuCl 0.308 HMDST 0.4 CH3CN 10 0.6 70 7 d 0.192

XZ1034 CuCl2 0.401 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.207

43

XZ1035 CuCl 0.305 HMDST 0.8 CH3CN 10 1.2 70 7 d 0.204

XZ1036 CuCl 0.289 HMDST 2.4 CH3CN 10 3.9 70 7 d 0.202

XZ1037 CuCl2 0.421 HMDST 0.8 CH3CN 10 1.2 RT 96 h 0.210

XZ1038 CuCl2 0.506 HMDST 2.4 CHCl3 10 3.0 130 7 d 0.323

XZ1039 CuCl2 0.443 HMDST 0.8 CHCl3 10 1.2 130 7 d 0.320

XZ1040 CuCl2 0.181 HMDST 2.4 CHCl3 10 8.5 130 7 d 0.118

XZ1041 CuCl2 0.411 HMDST 2.4 CH3CN 10 3.7 RT 96 h 0.234

XZ1042 CuCl2 0.409 HMDST 0.8 CHCl3 10 1.2 70 7 d 0.262

XZ1043 CuCl2 0.404 HMDST 2.4 CHCl3 10 3.8 70 7 d 0.266

XZ1044 CuCl2 0.178 HMDST 2.4 CHCl3 10 8.6 70 7 d 0.102

XZ1045 CuCl2 0.183 HMDST 2.4 CH3CN 10 8.4 RT 96 h 0.068

XZ1046 CuCl2 0.420 HMDST 0.8 CH3CN 10 1.2 70 7 d 0.236

XZ1047 CuCl2 0.414 HMDST 2.4 CH3CN 10 3.7 70 7 d 0.132

XZ1048 CuCl2 0.193 HMDST 2.4 CH3CN 10 7.9 70 7 d 0.058

XZ1049 CuCl2 0.447 HMDST 2.4 N/A N/A 3.4 RT 96 h 0.287

XZ1050 CuCl2 0.435 HMDST 0.8 CH3CN 10 1.2 130 7 d 0.293

XZ1051 CuCl2 0.398 HMDST 2.4 CH3CN 10 3.8 130 7 d 0.219

XZ1052 CuCl2 0.447 HMDST 0.8 CH3CN 15 1.1 130 7 d 0.312

XZ1053 CuCl 0.308 HMDST 0.4 CH3CN 10 0.6 RT 12 h 0.235

XZ1054 CuCl 0.311 HMDST 0.8 CHCl3 10 1.2 130 7.5 d * XZ1055 CuCl 0.302 HMDST 2.4 N/A N/A 3.7 130 7.5 d 0.217

XZ1056 CuCl 0.307 HMDST 0.8 CHCl3 10 1.2 130 7 d 0.264

XZ1057 CuCl2 0.452 HMDST 0.8 CHCl3 10 1.1 130 7 d 0.301

XZ1058 CuCl2 0.446 HMDST 0.4 CH3CN 10 0.6 130 7 d 0.245

XZ1059 CuCl 0.275 HMDST 0.8 CH3CN 10 1.4 RT 7 d 0.159

XZ1060 CuCl 0.307 HMDST 2.4 CH3CN 10 3.7 RT 7 d 0.130

XZ1061 CuCl2 0.468 HMDST 2.4 N/A N/A 3.3 70 7 d 0.267

XZ1062 CuCl2 0.404 HMDST 0.4 CH3CN 10 0.6 70 7 d 0.242

XZ1063 CuCl 0.330 HMDST 0.8 CH3CN 10 1.1 70 7 d 0.221

XZ1064 CuCl 0.378 DTBS 0.4 CHCl3 10 0.6 70 15 d 0.203†

XZ1065 CuCl 0.315 DTBS 2.4 CHCl3 10 4.2 70 120 d †

XZ1066 CuCl 0.376 DTBS 0.4 CH3CN 10 0.6 70 15 d 0.004†

XZ1067 CuCl 0.305 DTBS 2.4 CH3CN 10 4.3 70 120 d †

XZ1068 CuCl2 0.426 DTBS 0.4 CH3CN 10 0.7 70 120 d †

XZ1069 CuCl2 0.415 DTBS 2.4 CH3CN 10 4.3 70 120 d †

XZ1070 CuCl 0.340 DTBS 0.8 CHCl3 10 1.3 70 65 d 0.281†

XZ1071 CuCl 0.352 DTBS 0.8 CH3CN 10 1.3 70 65 d 0.029†

XZ1072 CuCl2 0.440 DTBS 0.8 CHCl3 10 1.4 70 65 d 0.24†

XZ1073 CuCl2 0.466 DTBS 0.8 CH3CN 10 1.3 70 65 d †

XZ1074 CuCl 0.304 DTBS 0.4 CH3CN 10 0.7 130 15 d 0.216

XZ1075 CuCl 0.301 DTBS 0.8 CH3CN 10 1.5 130 15 d 0.156

44

XZ1076 CuCl 0.323 DTBS 0.4 CHCl3 10 0.7 130 15 d 0.237

XZ1077 CuCl 0.339 DTBS 0.8 CHCl3 10 1.3 130 10 d 0.253

XZ1078 CuCl2 0.448 DTBS 0.8 CHCl3 10 1.3 130 10 d 0.209

XZ1079 CuCl2 0.425 DTBS 0.8 CH3CN 10 1.4 130 10 d 0.007

XZ1080 CuCl2 0.413 DTBS 2.0 CH3CN 10 3.6 100 12 d 0.067

XZ1081 CuCl 0.298 DTBS 0.8 CH3CN 10 1.5 100 12 d 0.173

XZ1082 CuCl 0.336 DTBS 0.8 CHCl3 10 1.3 100 12 d 0.171

XZ1083 CuCl 0.304 HMDST 0.8 CHCl3 10 1.2 130 7 d 0.178

XZ1084 CuCl2 0.689 HMDST 2.4 N/A N/A 2.2 70 ** 0.378

XZ1085 CuCl2 0.443 HMDST 0.8 CHCl3 10 1.2 130 15 d 0.296

XZ1086 CuCl 0.317 HMDST 0.4 CH3CN 10 0.6 RT ** 0.183 XZ1087 CuCl 0.305 HMDST 2.4 N/A N/A 3.7 70 7 d 0.155

XZ1088 CuCl 0.302 HMDST 0.8 CHCl3 10 1.2 130 7 d 0.187

XZ1089 CuCl 0.302 HMDST 0.8 CHCl3 10 1.2 130 15 d 0.178

XZ1090 CuCl 0.316 HMDST 0.4 CH3CN 10 0.6 RT 24 h 0.167

XZ1091 CuCl 0.317 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.197

XZ1092 CuCl 0.317 HMDST 0.4 CH3CN 10 0.6 RT 7 d 0.192

XZ1093 CuCl 0.313 HMDST 0.4 CH3CN 10 0.6 RT 90 h 0.053†

XZ1094 CuCl 0.314 HMDST 0.4 CH3CN 10 0.6 RT 117 h 0.036†

XZ1095 CuCl 0.278 HMDST 0.8 CH3CN 10 1.4 RT 24 h 0.192

XZ1096 CuCl 0.314 HMDST 2.4 CH3CN 10 3.6 RT 24 h 0.214

XZ1097 CuCl 0.288 HMDST 0.8 CH3CN 10 1.3 RT 96 h 0.198

XZ1098 CuCl 0.312 HMDST 2.4 CH3CN 10 3.6 RT 96 h 0.157

XZ1099 CuCl 0.320 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.195

XZ1100 CuCl 0.320 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.180

XZ1101 CuCl 0.318 HMDST 0.4 CH3CN 10 0.6 RT 96 h 0.209

XZ1102 CuCl 0.307 HMDST 0.4 CHCl3 10 0.6 RT 7 d 0.205

XZ1103 CuCl 0.310 HMDST 0.8 CHCl3 10 1.2 RT 7 d 0.129

XZ1104 CuCl 0.311 HMDST 2.4 CHCl3 10 3.6 RT 7 d 0.189

XZ1105 CuCl 0.311 HMDST 0.4 CHCl3 10 0.6 RT 4 d 0.128

XZ1106 CuCl 0.311 HMDST 0.8 CHCl3 10 1.2 RT 4 d 0.048

XZ1107 CuCl 0.308 HMDST 2.4 CHCl3 10 3.7 RT 4 d 0.116

XZ1108 CuCl 0.305 HMDST 0.4 CHCl3 10 0.6 RT 7 d 0.163

XZ1109 CuCl 0.310 HMDST 0.4 CHCl3 10 0.6 RT 7 d 0.199

XZ1110 CuCl 0.315 HMDST 0.4 CH3CN 10 0.6 RT 1.5 h 0.129

XZ1111 CuCl 0.286 HMDST 0.8 CH3CN 10 1.3 RT 1.5 h 0.211

XZ1112 CuCl 0.306 HMDST 2.4 CH3CN 10 3.7 RT 1.5 h 0.224

XZ1113 CuCl 0.302 HMDST 0.4 CH3CN 10 0.6 RT 300 s 0.187

XZ1114 CuCl 0.302 HMDST 0.4 CH3CN 10 0.6 RT 5 h 0.178 *Unsuccessful recovery because of sample handling **Tests for times required for complete reactions † Unsuccessful reaction (not copper sulfide) or impossible to recover 45

3.3 Results and Discussion

Seven copper sulfide polymorphs were observed by X-ray Diffraction (XRD) analysis in this thesis project: Low chalcocite, high chalcocite, tetragonal chalcocite, low digenite, high digenite, djurleite and covellite. The powder diffraction file (PDF) numbers and structural parameters of these polymorphs are shown in Table 3.3. Because of the similar hexagonal close packing of sulfur atoms in hexagonal chalcocite, monoclinic chalcocite and djurleite, positions of major XRD peaks of the three compounds are very close, which made definitive assignment of phase identity very difficult and not always possible with regular lab setups, especially when crystallinity was poor or other phases were present. Therefore, for certain samples, when clear assignments of phase identities between these three polymorphs were impossible, "chal/dj" was assigned to XRD patterns. Patterns that resembled monoclinic chalcocite or djurleite, but did not give strong enough peaks to assign phases with absolute certainty are labeled as "monoclinic chalcocite like", and "djurleite like". In addition, because the PDF cards selected for monoclinic chalcocite and djurleite only contained peaks up to 50° and 40° (2θ), respectively, positions and relative intensities at higher angles were calculated from the structures on which the view cards are based. These extended view cards are used for

XRD patterns presented in this chapter. Due to the large number of samples and significant amount of data presented in this chapter, the results are described grouped by reaction conditions in the following sub-sections.

46

Table 3.3: PDF cards and structural parameters of observed copper sulfide polymorphs.

Space Name PDF# Formula System Structural Parameters group a = 11.884 Å Low b = 13.494 Å 33-0490 Cu S Monoclinic P2 /c Chalcocite 2 c = 15.228 Å 1 β = 116.215° High a = 3.961 Å 26-1116 Cu S Hexagonal P6 /mmc Chalcocite 2 c = 6.722 Å 3 Tetragonal a = 3.996 Å 29-0578 Cu S Tetragonal P4 2 2 Chalcocite 1.96 c = 11.287 Å 3 1 a = 3.93 Å

Low Digenite 47-1748 Cu1.8S Rhombohedral c = 48.14 Å R m

High Digenite 71-4316 Cu1.8S Cubic a = 5.57 Å Fm m a = 26.897 Å b = 15.745 Å Djurleite 71-1383 Cu S Monoclinic P2 /c 1.94 c = 13.565 Å 1 β = 90.13° a = 3.794 Å Covellite 78-0877 CuS Hexagonal b = 16.341 Å P63/mmc

3.3.1 Elemental Analysis

Due to the complexity of copper sulfides, in order to obtain an accurate and reliable copper to sulfur ratio of a raw sample, a combination of several characterization methods, such as CHNS analysis, thermogravimetric analysis (TGA) and energy dispersive X-ray spectroscopy (EDS), is often necessary. CHNS analysis can provide the content of sulfur and residual carbon, and TGA can provide information about copper content. However, if a sample contains CuCl, TGA cannot provide accurate information about copper sulfide content, as CuCl may either evaporate or be oxidized. Therefore, elemental analysis was also carried out by EDS.

47

For samples without CuCl present in their XRD patterns, their chlorine contents obtained by EDS were usually less than 2 atom% and within the detection limit (3σ), indicating negligible amounts of CuCl in such samples. If CuCl was present in the XRD pattern of a sample, copper chloride particles could be found by SEM/EDS. For such samples, the Cu/S ratio was directly calculated from EDS data by using Equation 3-1 by area analysis:

ri = (atCu% - atCl%) / atS% (Equation 3-1) where ri is the Cu to S ratio, and atCu%, atCl% and ats% are the atomic percentages of Cu,

Cl and S, respectively, obtained by EDS. The average Cu/S ratio of a sample was then calculated by averaging the calculated ratios obtained from each measurement. Copper chloride particles were excluded from the analysis area as much as possible, as they did not represent the copper sulfide phases in these samples. However, small amounts of chlorine (2 to 7 at%) were still detected, as small CuCl particles may be covered by or fused with copper sulfide particles. The copper to sulfur ratios of such samples were calculated after correcting each measurement for the detected chlorine content.

For heat treated samples containing negligible amounts of CuCl, and samples for which EDS data were not available, the copper to sulfur ratios were calculated from

CHNS and TGA data using Equation 3-2.

r = (wt%Cu / MCu) / (wt%S / MS) (Equation 3-2) where wt%Cu and wt%S are the weight percentages of copper obtained by TGA and sulfur obtained by CHNS analysis, respectively, and MCu and MS are the molar mass for copper and sulfur, respectively.

48

For samples with only CHNS data available, the copper to sulfur ratios were calculated from the CHNS data using Equation 3-3 by assuming that only copper, sulfur and carbon are present in these samples. If other impurities were present, the true copper to sulfur ratios will be lower.

r = (100% - wt%C - wt%S) / wt%S × MS / MCu (Equation 3-3) where wt%C and wt%S are the weight percentages of carbon and sulfur obtained by

CHNS analysis, and MCu and MS are the molar mass for copper and sulfur.

Error propagation for all calculations was carried out by standard methods.65

3.3.2 Neat Reactions

As both sulfur sources are liquids, several samples were prepared by reactions of

CuCl or CuCl2 with HMDST in the absence of solvent. During the preparation of samples

XZ1013, XZ1016 and XZ1084, part of the precipitate was recovered at different reaction times to check the progress of the reaction. XRD and elemental analysis results of these samples are shown in Table 3.4 and Table 3.5, respectively.

Neat reaction between CuCl and HMDST at room temperature required several days to go to completion. XRD patterns of such neat reactions after 3 h, 24 h and 96 h are shown in Figure 3-3 a), b) and c), respectively. It was found that the core of the black powders recovered after three hours was white (the color of CuCl). It appears that

HMDST formed copper sulfide shells on CuCl particles, hence hindering further reaction.

However, when the reaction of XZ1026 was carried out with a finely ground powder in an ultrasound bath, it was almost complete after 3 h, resulting in a poorly crystalline phase that resembled XZ1013 after 96 h with only a trace amount of starting material

CuCl left.

49

Table 3.4: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of copper chlorides and HMDST without solvent. Copper Conditions Thermal Sample Phases Source T/°C t/h History 3 Raw CuCl XZ1013 CuCl RT 24 Raw CuCl and amorphous* 96 Raw Poorly Crystalline 2 Raw CuCl XZ1016 CuCl RT 3 Raw CuCl Poorly Crystalline and trace Raw CuCl XZ1026** CuCl RT 3 Monoclinic Chalcocite 400 °C, 3h Tetragonal Chalcocite XZ1087 CuCl 70 168 Raw Monoclinic Chalcocite Raw Monoclinic Chalcocite XZ1055 CuCl 130 180 400 °C, 3h Monoclinic Chalcocite 3 Raw Covellite XZ1084 CuCl RT 2 30 Raw Covellite Raw Covellite XZ1049 CuCl2 RT 96 Low Digenite Tetragonal 400 °C, 3h Chalcocite Raw Covellite XZ1061 CuCl 70 168 2 400 °C, 3h Low Digenite * Minor phase ** Ground to a fine powder and reacted in ultrasound bath

Table 3.5: CHNS analysis and EDS results of raw copper sulfide samples prepared by reactions of copper sources and HMDST without solvent. Cu/S ratios were calculated from CHNS and EDS data unless otherwise indicated. TGA CHNS EDS Cu/S Ratio Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% CHNS EDS XZ1026 0.7±0.3 19.4±0.6 64±6 35±4 1±1 2.1±0.1 1.8±0.3 XZ1049 0.6±1.8 31.9±3.7 55±2 45±2 0±1 1.1±0.1 1.2±0.1 XZ1049.T400 76±7 0.4±0.9 20.8±1.9 69±7 31±4 0±0 1.9±0.2† 2.2±0.3 XZ1055 0.4±0.3 19.8±0.6 70±5 29±5 0±1 2.0±0.1 2.4±0.6 XZ1061 0.5±0.7 31.3±1.5 56±8 43±7 1±1 1.1±0.1 1.3±0.3 XZ1061.T400 80±8 0.4±2.2 22.3±4.4 1.8±0.4† XZ1084* 72±5 0.7±0.7 29.4±1.2 XZ1087 0.7±0.7 19.1±1.1 * Sample recovered after 30 h † Ratio calculated from TGA and CHNS data

50

Figure 3-3: XRD patterns of sample XZ1013 prepared by reaction of CuCl and HMDST at RT without solvent for a) 3 h, b) 24 h and c) 96 h, indicating the disappearance of the CuCl phase (PDF card) over time.

As shown in Figure 3-4, neat reactions of CuCl with HMDST at both 70 °C and

130 °C resulted in monoclinic chalcocite, which is the thermodynamically stable phase at room temperature. The copper to sulfur ratios in the samples prepared at RT (XZ1026) and 130 °C (XZ1055) were 2.1±0.1 and 2.0±0.1, respectively (Table 3.5), which agree with typical copper to sulfur ratio of chalcocites (2:1). It is possible that the poorly crystalline phase obtained at room temperature was highly disordered chalcocite, which crystallizes into monoclinic chalcocite at higher reaction temperatures.

51

Figure 3-4: XRD patterns of samples prepared by reactions of CuCl and HMDST without solvent at a) 70 °C for 7 d and b) 130 °C for 7.5 d. Lines indicate the PDF card of monoclinic chalcocite.

Heating monoclinic chalcocite samples to 400 °C and slow cooling did not induce a phase change. In contrast, the same treatment for poorly crystalline samples prepared at

RT resulted in formation of monoclinic chalcocite. To follow the phase evolution, the poorly crystalline phase of as-recovered XZ1026 was also studied by in situ variable temperature XRD (Table 3.6). When the sample was heated above 150 °C, it converted to high digenite. This agrees with the Cu-S phase diagram, as the sample's composition from EDS analysis is Cu1.8S. At higher temperatures, high digenite can adopt a range of compositions from Cu1.75S to Cu2.0S. When the sample was cooled, the high digenite

52

phase transformed to tetragonal chalcocite and monoclinic chalcocite below 100 °C, suggesting loss of some sulfur at high temperature. Therefore, the monoclinic chalcocite phase obtained during ex situ studies was formed during slow cooling of digenite with a

2:1 Cu to S ratio. While the pattern of as-recovered XZ1026 showed poorly crystalline and amorphous phase, heating resulted in crystallization of the thermodynamically stable polymorph for its overall composition.

Table 3.6: In situ variable temperature X-ray diffraction results for the poorly crystallized phase of sample XZ1026.

Sample Thermal History Phases

Heating from RT to 100 °C Poorly Crystalline Heating from 150 °C to 600 °C Digenite Cooling from 500 °C to 400 °C Digenite XZ1026 Cooling to 200 °C Tetragonal Chalcocite and Digenite Monoclinic Chalcocite and Tetragonal Cooling from 100 °C to RT Chalcocite

During the use of the VT stage, it was noticed that there might be errors in the temperature calibration. Therefore, a CsCl standard was used to check the offset as it can transform from a simple cubic lattice to a face centered cubic lattice at 450 °C. However, after heating to a programmed temperature of 500 °C, its XRD pattern still showed the simple cubic structure. When heating to 700 °C directly form 500 °C, a transition to the face centered cubic structure was observed. These resulted indicated that there was an offset of more than 50 °C but less than 250 °C in the temperature calibration. In addition, the melting point for CsCl is 645 °C, and the fact that a solid phase was still observed at a programmed temperature of 700 °C further confirmed this offset. Because the temperature offset may not be linear and the thickness of the samples may also contribute to the exact deviation if the sample has low thermal conductivity, the true measurement

53

may be significantly lower than the programmed temperature. A more precise determination of the offset was not possible at the time of submission of this thesis as the instrument was broken.

All as-recovered samples prepared by neat reactions of CuCl2 and HMDST consisted of covellite. The CuCl2 starting material, which was a fine powder with smaller grain sizes than the CuCl starting material, seemed to reacted rapidly with HMDST and reactions were completed after 3h. In comparison, a sample prepared with finely ground

CuCl powder (XZ1026) still contained small amounts of CuCl after reacting for 3 h in an ultrasound bath. The copper to sulfur ratios of all CuCl2 samples were very close to 1

(Table 3.5), the typical ratio expected for covellite. However, the covellite samples prepared from neat reactions contained significant amounts of amorphous materials

(Figure 3-5), though the crystallinity could be increased by longer reaction times and higher temperatures.

Figure 3-5: XRD patterns of samples prepared by reactions of CuCl2 and HMDST without solvent at a) RT for 3 h (XZ1084), b) RT for 96 h (XZ1049) and c) 70 °C for 7 d (XZ1061). Lines indicate the PDF card of covellite.

54

Samples prepared by neat reactions consisted of nanoparticles (Figure 3-6). The monoclinic sample prepared by neat reaction of CuCl with HMDST at 70 °C (Figure 3-6 a) and b)) consisted of distorted spherical nanoparticles with sizes below 100 nm.

However, these nanoparticles were severely agglomerated and partially fused. For a similar sample prepared at 130 °C (Figure 3-6 c)), the agglomeration of nanoparticles was more obvious, and many of them fused into bigger structures. Covellite particles prepared by neat reactions consisted of highly agglomerated nanoplatelets (Figure 3-6 d) and e)).

The faces of these nanoplatelets tended to stick together in the agglomerates. The covellite nanoplatelets of the room temperature sample (XZ1084, 3 h) appeared to be smaller and less agglomerated (Figure 3-6 f)).

55

Figure 3-6: SEM images of a) and b) XZ1087raw (CuCl, 70 °C , monoclinic chalcocite), c) XZ1055raw (CuCl, 130°C , monoclinic chalcocite), d) and e) XZ1061raw (CuCl2, 70°C, covellite), and f) XZ1084raw (CuCl2, RT, 3 h, covellite). 3.3.3 CuCl and HMDST in Chloroform and Acetonitrile

Reactions between CuCl and HMDST gave interesting results, as five different copper sulfide polymorphs were obtained from as-recovered products by fine tuning reaction conditions, such as temperature, concentration and reaction time. However, the parameters necessary for phase control were initially unclear because of the large number of combinations of synthetic parameters. In addition, because of the similar hexagonal close packing of sulfur atoms in hexagonal chalcocite, monoclinic chalcocite and djurleite, positions of major XRD peaks of the three compounds are very close, which made definitive assignment of phase identity very difficult, especially when crystallinity was poor. However, accurate phase identification was essential to understand the phase evolution between these three polymorphs, and the confirmation of the formation of hexagonal chalcocite was especially important, as it is thermodynamically unstable at room temperature and should transform to monoclinic chalcocite. Therefore, XRD data with significantly improved signal to noise ratios were collected using specialized setups.

To obtain the best possible data, a selected subset of samples were packed inside Kapton

56

capillaries and sent for mail-in service at beamline 11BM at the Advanced Photon Source

(APS). Additional data were collected in 0.2 mm capillaries on a Panalytical X'pert Pro diffractrometer with an X’Celerator detector. These results are summarized in the following sub-sections.

3.3.3.1 XRD and Elemental Analyses for Samples Prepared in Chloroform

The results of XRD analysis and elemental analysis for raw and heated samples prepared by reactions of CuCl and HMDST in CHCl3 are shown in Table 3.7 and Table

3.8, respectively. Crystalline copper sulfide polymorphs were observed in all raw samples.

Temperature was the major factor that determined which polymorph was favored. CHNS analysis results showed that the carbon content of all samples was below 3%, and many of the values were within the detection limit, indicating a very small quantity of organic residue in each as-recovered sample. However, peaks corresponding to the starting material CuCl were seen in several samples, indicating incomplete reactions. Generally, the reactions of CuCl and HMDST in CHCl3 were successful.

57

Table 3.7: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in 15 mL (XZ1001 to XZ1007) and 10 mL CHCl3. Conditions Thermal Sample [HMDST] S:Cu Phases T/°C t/h History Raw Chal/Dj XZ1001 0.6 3.4 RT 29 200 °C, 3h Monoclinic Chalcocite Raw Hexagonal Chalcocite XZ1002 0.6 3.4 RT 3 200 °C, 3h Monoclinic Chalcocite 400 °C, 3h Monoclinic Chalcocite Raw Hexagonal Chalcocite XZ1003 0.6 3.5 RT 96 200 °C, 3h Monoclinic Chalcocite Raw Hexagonal Chalcocite 200 °C, 3h Monoclinic Chalcocite XZ1004 0.2 1.2 RT 3 400 °C, 3h Monoclinic Chalcocite 600 °C, 3h Monoclinic Chalcocite Hexagonal Chalcocite Raw CuCl Monoclinic Chalcocite 200 °C, 3h XZ1005 1.4 9.2 RT 3 CuCl Monoclinic Chalcocite 400 °C, 3h CuCl 600 °C, 3h Monoclinic Chalcocite Raw Monoclinic Chalcocite Like 200 °C, 3h Monoclinic Chalcocite XZ1006 0.2 1.2 RT 24 400 °C, 5h Monoclinic Chalcocite 600 °C, 3h Monoclinic Chalcocite Raw† Hexagonal Chalcocite 200 °C, 3h Monoclinic Chalcocite XZ1007 1.5 10.3 RT 25 400 °C, 3h Monoclinic Chalcocite 600 °C, 3h Monoclinic Chalcocite Hexagonal Chalcocite Raw CuCl Tetragonal Chalcocite 200 °C, 4h XZ1012 0.2 0.6 RT 24 Low Digenite 400 °C, 4h Tetragonal Chalcocite Monoclinic Chalcocite 600 °C, 4h Tetragonal Chalcocite* Hexagonal Chalcocite ACS18 0.2 0.6 RT 24 Raw† * CuCl 58

Monoclinic Chalcocite Like XZ1105 0.2 0.6 RT 96 Raw CuCl* XZ1102 0.2 0.6 RT 168 Raw Monoclinic Chalcocite XZ1108 0.2 0.6 RT 168 Raw Monoclinic Chalcocite XZ1109 0.2 0.6 RT 168 Raw Monoclinic Chalcocite Raw Hexagonal Chalcocite 200 °C, 3h Monoclinic Chalcocite XZ1008 0.4 1.2 RT 3 Monoclinic Chalcocite 400 °C, 3h Tetragonal Chalcocite* 600 °C, 3h Monoclinic Chalcocite ACS19 0.4 1.3 RT 24 Raw† Hexagonal Chalcocite XZ1106 0.4 1.2 RT 96 Raw Chal/Dj XZ1103 0.4 1.2 RT 168 Raw Chal/Dj Hexagonal Chalcocite Raw CuCl* Monoclinic Chalcocite 200 °C, 3h XZ1009 0.9 3.6 RT 3 CuCl* Monoclinic Chalcocite 400 °C, 3h CuCl* 600 °C, 3h Monoclinic Chalcocite ACS17 0.9 3.7 RT 24 Raw† Hexagonal Chalcocite XZ1107 0.9 3.7 RT 96 Raw Hexagonal Chalcocite XZ1104 0.9 3.6 RT 168 Raw† Hexagonal Chalcocite Hexagonal Chalcocite Raw CuCl Monoclinic Chalcocite 200 °C, 3h XZ1010 1.8 9.7 RT 3 CuCl Chal/Dj 400 °C, 3h CuCl 600 °C, 6h Monoclinic Chalcocite Raw† Djurleite Tetragonal Chalcocite 200 °C, 4h XZ1015 0.2 0.6 70 168 Low Digenite and CuCl* Low Digenite 400 °C, 4h Tetragonal Chalcocite

59

Djurleite Raw CuCl* Low Digenite and CuCl* 200 °C, 4h XZ1011 0.4 1.3 70 168 Tetragonal Chalcocite Monoclinic Chalcocite 400 °C, 4h Tetragonal Chalcocite Monoclinic Chalcocite 600 °C, 4h Tetragonal Chalcocite Djurleite Raw CuCl* Tetragonal Chalcocite XZ1020 1.0 4.1 70 168 200 °C, 3h CuCl Monoclinic Chalcocite 400 °C, 3h Tetragonal Chalcocite Low Digenite Raw CuCl* XZ1019 0.4 1.3 100 168 200 °C, 3h Low Digenite Low Digenite 400 °C, 3h Tetragonal Chalcocite Low Digenite Raw Covellite* CuCl* XZ1025 0.2 0.6 130 168 200 °C, 3h Mainly Low Digenite Low Digenite 400 °C, 3h Tetragonal Chalcocite* Low Digenite ACS21 0.2 0.6 130 168 Raw Covellite Low Digenite Raw Covellite XZ1014 0.4 1.3 130 168 Low Digenite 200 °C, 4h Covellite 400 °C, 4h Low Digenite Raw Low Digenite XZ1056 0.4 1.2 130 168 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Low Digenite XZ1083 0.4 1.2 130 168 Raw Covellite

60

XZ1088 0.4 1.2 130 160 Raw Low Digenite, Covellite XZ1089 0.4 1.2 130 350 Raw Low Digenite, Covellite ACS20 1.0 3.6 130 168 Raw Low Digenite Raw Low Digenite 200 °C, 3h Low Digenite XZ1022 1.0 4.3 130 168 Low Digenite 400 °C, 3h * Tetragonal Chalcocite * Minor phase or trace † Capillary data collected Table 3.8: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CHCl3. Wt% after HT TGA CHNS EDX Cu/S Ratio 200 °C 400 °C 600 °C Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% XZ1001 67±5 28±8 6±6 2.2±0.6* XZ1002 95% 0.4±0.8 18.4±1.5 67±5 28±5 5±6 2.3±0.5* XZ1002.T400 77±5 XZ1003 99% 97% 0.5±0.3 18.8±0.7 65±3 29±4 6±7 2.1±0.2* XZ1004 99% 95% 89% 0.6±0.5 19.3±1.0 67±7 30±3 3±1 2.2±0.3* XZ1004.T400 79±3 XZ1005 98% 89% 68% 0.4±1.2 13.0±2.3 65±4 28±7 4±6 2.1±0.4* XZ1005.T400 73±3 XZ1006 99% 95% 86% 0.4±0.3 19.3±0.5 67±4 29±4 4±2 2.2±0.4* XZ1006.T400 75±14 XZ1006.T600 70±6 30±5 1±2 XZ1007 100% 93% 83% 0.5±0.5 18.7±1.0 66±3 31±6 4±4 2.0±0.4* XZ1007.T400 74±2 2.8±0.8 17.9±1.6 2.1±0.2† XZ1008 99% 94% 68±6 28±6 4±4 2.3±0.7* XZ1008.T400 78±2 0.6±1.4 19.4±2.8 2.0±0.3† XZ1009 99% 87% 80% 0.3±0.4 17.9±0.9 67±5 28±6 4±2 2.2±0.6* XZ1009.T400 80±7 2.1±0.2‡ XZ1010 76% 65% 66% 0.5±0.2 13.1±0.4 67±2 30±1 4±1 2.1±0.3* XZ1010.T400 77±8 1.3±0.7 13.5±1.4 XZ1011 97% 86% 79% 0.4±0.2 19.6±0.5 68±6 30±5 2±2 2.2±0.6* XZ1012 98% 91% 84% 0.6±0.8 18.1±1.6 66±1 31±5 3±6 2.0±0.1* XZ1012.T400 78±2 0.7±0.4 19.8±0.7 2.0±0.1‡ XZ1014 99% 95% 0.4±0.3 23.0±0.6 59±17 40±19 1±3 1.5±0.9* XZ1014.T400 76±6 XZ1015 100% 99% 68±4 30±5 1±1 2.2±0.4* XZ1015.T400 77±2 0.3±0.3 20.3±0.6 67±9 33±9 1±1 1.9±0.1‡ XZ1019 96% 92% 0.5±0.2 21.0±0.4 66±4 31±1 3±3 2.0±0.3* XZ1019.T400 79±4 1.7±0.7 20.2±1.4 2.0±0.2† XZ1020 98% 94% 0.4±0.3 18.3±0.6 67±4 30±4 3±1 2.1±0.4* XZ1020.T400 79±3 0.3±0.3 19.3±0.6 2.0±0.1† XZ1022 90% 89% 0.5±0.4 21.8±0.9 64±1 34±3 2±3 1.9±0.1* XZ1022.T400 79±2 0.5±0.4 20.4±0.9 65±9 35±7 0±1 2.0±0.1† XZ1025 99% 96% 0.4±0.3 21.3±0.5 66±3 32±3 3±1 2.0±0.3* XZ1025.T400 80±7 0.3±0.4 20.7±0.7 67±3 33±3 1±1 2.0±0.2† XZ1056 98% 1.2±0.6 20.8±1.2 59±11 39±9 2±3 1.9±0.2** XZ1056.T400 82% 77±3 0.9±0.4 17.4±0.9 2.2±0.2† XZ1083 0.3±0.2 21.7±0.4 1.8±0.1** XZ1088 0.2±0.2 23.1±0.4 1.7±0.1** XZ1089 0.2±0.4 23.0±0.6 1.7±0.1** XZ1102 0.1±0.6 18.8±0.9 2.2±0.1** XZ1103 0.8±0.7 18.9±1.1 2.1±0.1** XZ1104 0.7±0.5 18.2±0.8 2.2±0.1** XZ1105 0.3±0.4 18.6±0.7 2.2±0.1** XZ1106 0.8±0.5 19.1±0.9 2.1±0.1** XZ1107 0.4±0.7 18.5±1.1 2.2±0.1** * EDS † CHNS and TGA ‡ TGA ** CHNS

61

Samples prepared with 15 mL of chloroform (XZ1001 to XZ1007), resulted in phases that resembled hexagonal chalcocite, except for XZ1006, the phase of which resembled monoclinic chalcocite. The rest of the samples were prepared with 10 mL chloroform because initially no obvious difference was observed for different solvent volumes. As shown in Table 3.7, despite different amounts of solvent and sulfur source, early samples (XZ1001 to XZ1012) prepared at RT resulted in similar phases (except for

XZ1006), but the completeness of reaction was influenced by the amounts of solvent and sulfur source as several samples contained CuCl starting material. It was found that samples with higher HMDST concentrations required longer reaction times. Similar to the neat reactions, because of the poor solubility of CuCl in chloroform, and if the undissolved particles form a copper sulfide shell, the overall contact area is reduced significantly. Further reaction is only possible after the precipitate is broken up and removed from the surface, thus increasing the reaction time. Reactions with lower

HMDST concentrations allowed copper sulfide to precipitate more slowly and from solution, and consequently more CuCl dissolved before being covered. Therefore, CuCl appeared more reactive with lower HMDST concentrations.

For RT reactions prepared for 24 h with different HMDST concentrations

(ACS17 and ACS18), patterns resembling hexagonal chalcocite were obtained (Figure

3-7). Even capillary data showed no distinct features of monoclinic chalcocite or djurleite, although the significant background suggests either disorder of the hexagonal chalcocite phases or presence of significant amounts of amorphous materials. For a set of three samples prepared for 7 d with varying HMDST concentrations (XZ1102 to XZ1104), the pattern of the sample prepared with the highest HMDST concentration (XZ1104) showed

62

a hexagonal chalcocite pattern with no distinct additional features, while the patterns of the other two samples showed emerging (XZ1103) or fully developed (XZ1102) peaks corresponding to monoclinic chalcocite with decreased HMDST concentration (Figure

3-8). These results clearly indicated a phase evolution from the initial hexagonal chalcocite phase towards the monoclinic chalcocite phase, and the conversion rate decreased with increasing HMDST concentration.

Figure 3-7: XRD patterns of as-recovered samples prepared at RT for 24 h: a) ACS18 (0.4 mL HMDST), b) ACS19 (0.8 mL HMDST) and c) ACS17 (2.4 mL HMDST), collected in 0.2 mm capillaries. Lines indicate PDF card of hexagonal chalcocite, and * indicates CuCl.

63

Figure 3-8: XRD patterns of as-recovered samples prepared at RT for 7 d: a) XZ1102 (0.4 mL HMDST), b) XZ1103 (0.8 mL HMDST), c) XZ1104 (2.4 mL HMDST) and d) an insert of the capillary data for XZ1104 between 25° and 40° (2θ). Lines indicate PDF card of monoclinic chalcocite, and * indicate peaks of hexagonal chalcocite.

For 70 °C reactions, djurleite, a more copper deficient phase than monoclinic chalcocite, was obtained. It was difficult to distinguish the poorly crystallized djurleite, which was difficult to distinguish from hexagonal chalcocite (Figure 3-9), as the phases only differ in the lower intensity peaks. Hence XRD data for some of these samples were also collected in a 0.2 mm capillary. The low intensity peaks of could be observed

(Figure 3-9 a) and b)) despite the broadness of the peaks. For higher temperature reactions at 100 °C and 130 °C, more copper deficient phases, like digenite and covellite could be obtained depending on reaction time and HMDST concentration.

64

Figure 3-9: XRD patterns of as-recovered samples of a) and b) XZ1015 (70 °C, 7 d, 0.4 mL HMDST) collected on capillary and flat stages, respectively, c) XZ1011 (70 °C, 7 d, 0.8 mL HMDST) and d) XZ1020 (70 °C, 7 d, 2.5 mL HMDST). Lines indicate PDF card of djurleite.

For the 130 °C samples prepared with the lowest HMDST concentration with a sulfur to copper ratio of about 0.6 (XZ1025 and ACS21), low digenite and covellite were observed. Sample ACS21 contained a more significant amount of covellite than XZ1025

(Figure 3-10). For samples with intermediate HMDST concentrations, a mixture of low digenite and covellite was first observed in XZ1014 (130 °C, 7 d, 0.8 mL HMDST).

Hence, two follow-up samples, XZ1056 and XZ1083, were synthesized under identical conditions as XZ1014 to reproduce the results. No covellite was found in as-recovered

XZ1056, but this phase was present in XZ1083. Because of this, two more parallel reactions with identical conditions except for the reaction times (7 d and 15 d for XZ1088 and XZ1089, respectively) were prepared.

65

Figure 3-10: XRD patterns of identical samples a) XZ1025 and b) ACS21 prepared at 130 °C with 0.4 mL HMDST. Lines indicate the PDF cards of covellite (bottom) and low digenite (top).

As shown in Figure 3-11, the amount of covellite phase increased significantly for longer reaction times. Similarly to XZ1088, covellite was observed as a minor phase in other 7 d reactions, such as XZ1014 and XZ1083. Therefore, the covellite phase found in these samples was likely formed by conversion of the existing digenite phase. In contrast to these samples, no covellite was found in samples XZ1022 and ACS20 prepared with the highest HMDST concentration (1.0 M, 2.5 mL). These results indicated that at higher temperatures, phase evolutions towards more copper deficient phases, low digenite and covellite, are possible. In addition, it appeared that conversion was favored at lower HMDST concentrations. Since the copper to sulfur ratio for some samples was less than 1, complete conversion to covellite was not possible.

66

Figure 3-11: XRD patterns of parallel reactions at 130 °C with 0.8 mL HMDST for a) 7 d (XZ1088) and b) 15 d (XZ1089). Lines indicate PDF cards of low digenite (bottom) and covellite (top).

Heat treatments of hexagonal chalcocite and poorly crystallized djurleite samples tended to result in formation of monoclinic chalcocite, although some gave a mixture of low digenite and tetragonal chalcocite or monoclinic chalcocite and tetragonal chalcocite.

Digenite samples were more stable, as the phase did not change after heating to 200 °C.

However, further heat treatment to 400 °C caused partial conversion from the digenite phase to tetragonal chalcocite, and resulted in a mixture of digenite and tetragonal chalcocite.

In order to understand the phase transformations of these samples during heat treatment, in situ variable temperature XRD analysis was carried out with selected samples. XZ1009, XZ1011 and XZ1020, and XZ1019 were analyzed to study the 67

temperature-induced phase transitions of hexagonal chalcocite, djurleite and digenite, respectively. These results are summarized in Table 3.9.

Table 3.9: In situ variable temperature X-ray diffraction results for XZ1009 (hexagonal chalcocite), XZ1011 and XZ1020 (djurleite), and XZ1019 (digenite) raw samples.

Sample Thermal History Phases

Heating from RT to 600 °C Hexagonal Chalcocite XZ1009 Cooling from 500 °C to 200 °C Hexagonal Chalcocite Cooling from 100 °C to RT Monoclinic Chalcocite Heating from RT to 100 °C Djurleite Heating from 150 °C to 600 °C Hexagonal Chalcocite and Digenite XZ1011 Cooling from 500 °C to 200 °C Hexagonal Chalcocite and Digenite Cooling from 100 °C to RT Monoclinic Chalcocite Heating from RT to 100 °C Djurleite

Heating from 150 °C to 600 °C Hexagonal Chalcocite* and Digenite XZ1020 Cooling from 500 °C to 300 °C Hexagonal Chalcocite and Digenite**

Monoclinic Chalcocite with trace Cooling from 200 °C to RT Tetragonal Chalcocite Heating from RT to 600 °C Digenite XZ1019 Cooling from 500 °C to 300 °C Digenite Cooling from 200 °C to RT Tetragonal Chalcocite and Digenite* * Trace at 300 °C ** Trace at 200 °C

Upon heating, the hexagonal chalcocite sample remained hexagonal chalcocite up to 600 °C , and converted to the thermodynamically stable monoclinic chalcocite when cooled below 100 °C . For the djurleite samples, mixtures of hexagonal chalcocite and high digenite were observed at high temperature. When cooled below 100 °C , formation of monoclinic chalcocite was observed.

The digenite sample (XZ1019) showed no significant phase changes up to 600 °C.

However, during cooling from 600 °C to 200 °C, a tetragonal chalcocite phase appeared, and only a trace amount of the digenite phase was observed at room temperature.

68

Because low digenite transforms to high digenite, which can adopt a wide range of stoichiometries in a solid solution, the Cu/S ratio can increase with temperature if the sample loses sulfur during heat treatment. As the heat treatment was not carried out in a sealed system, any sulfur lost would not be recovered during cooling. Therefore, this experiment resulted in tetragonal chalcocite (Cu1.96S) as the major phase.

All as-recovered djurleite samples contained a fair amount of amorphous material, thus their overall copper to sulfur ratios could deviate from the allowed region for djurleite (Cu1.93S to Cu1.97S). In addition, djurleite is only stable up to 93 °C , and upon heating, the djurleite samples formed a mixture of hexagonal chalcocite and digenite.

During cooling, hexagonal chalcocite transformed to monoclinic chalcocite (Cu1.98S to

Cu2S). However, if a sample was copper deficient, digenite solid solutions may remain during cooling and transform to tetragonal chalcocite (Cu1.96S). This explains why trace tetragonal chalcocite was observed in several samples.

The sulfur loss of digenite samples after heat treatments explains why some low digenite samples, such as XZ1019, XZ1022 and XZ1025, resulted in the formation of tetragonal chalcocite after heat treatment at 400 °C. The copper to sulfur ratios in raw

XZ1019, XZ1022 and XZ1025 were close to 1.9. However, after heat treatment at 400 °C, their ratios increased to an average of about 2.0, which is close to chalcocite compositions. While accurate changes of the copper to sulfur ratios cannot be determined due to the large errors, the trend that the ratio increases after heat treatment is still obvious. Because the heat treatments were carried out in flowing argon, some sulfur was lost at high temperatures. Therefore, the transformations between low digenite and chalcocites were not reversible.

69

3.3.3.2 XRD and Elemental Analyses for Samples Prepared in Acetonitrile

Reactions between CuCl and HMDST in acetonitrile produced crystalline phases in the as-recovered powders. Most of these samples were heated to 400 °C , and some of them were heated to 200 °C or 600 °C . The results of XRD and CHNS analysis and EDS microanalysis on as-recovered and heat treated samples, and TGA results for previously heated samples are summarized in Table 3.10 and Table 3.11. When comparing the XRD and EDS results in this section with the previous section, the chlorine content of the samples prepared in acetonitrile is generally lower than for those prepared in chloroform, which can be attributed to the better solubility of CuCl in acetonitrile. CHNS analysis gave carbon contents below 1.5% for all samples, and in most cases below the detection limit. Therefore, the organic residues in these samples were also negligible. Thus, the impurities in the as-recovered samples prepared in acetonitrile were generally low.

Several samples (XZ1018, XZ1021, XZ1023, XZ1024 and XZ1101) were prepared at room temperature in a flask covered with a Petri dish. It was found that covellite could be formed with evaporation of significant amounts of the solvent. Because the exact conditions were very difficult to control, the results were not completely reproducible. Therefore, these samples are not further discussed in this chapter.

70

Table 3.10: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CH3CN. [HMDST] Conditions Thermal Sample S:Cu Phases (M) T/°C t/h History XZ1113 0.2 0.6 RT 5 min Raw Monoclinic Chalcocite XZ1110 0.2 0.6 RT 1.5 Raw Monoclinic Chalcocite XZ1114 0.2 0.6 RT 5 Raw Monoclinic Chalcocite Like Raw Djurleite Like XZ1053 0.2 0.6 RT 12 Low Digenite 400 °C, 3h Tetragonal Chalcocite XZ1090 0.2 0.6 RT 24 Raw‡ Djurleite Like 96 Raw Djurleite XZ1086 0.2 0.6 RT 180 Raw Djurleite XZ1091 0.2 0.6 RT 96 Raw‡ Djurleite XZ1099 0.2 0.6 RT 96 Raw Djurleite XZ1100 0.2 0.6 RT 96 Raw Djurleite XZ1092 0.2 0.6 RT 168 Raw‡ Djurleite XZ1111 0.4 1.4 RT 1.5 Raw‡ Hexagonal Chalcocite XZ1095 0.4 1.4 RT 24 Raw Hexagonal Chalcocite XZ1097 0.4 1.3 RT 96 Raw Hexagonal Chalcocite Raw Hexagonal Chalcocite Like XZ1059† 0.4 1.4 RT 168 400 °C, 3h Monoclinic Chalcocite XZ1112 0.9 3.7 RT 1.5 Raw‡ Hexagonal Chalcocite XZ1096 0.9 3.6 RT 24 Raw Hexagonal Chalcocite XZ1098 0.9 3.6 RT 96 Raw Hexagonal Chalcocite Raw Hexagonal Chalcocite Like XZ1060† 0.9 3.7 RT 168 Monoclinic Chalcocite 400 °C, 3h Tetragonal Chalcocite Raw Djurleite XZ1033 0.2 0.6 70 168 Tetragonal Chalcocite 200 °C, 3h Djurleite Djurleite Raw CuCl XZ1035 0.4 1.2 70 168 Chal/Dj 400 °C, 3h CuCl Raw Djurleite XZ1063 0.4 1.1 70 168 Tetragonal Chalcocite 400 °C, 3h Low Digenite

71

Djurleite Raw CuCl* XZ1036 0.9 3.9 70 168 Djurleite 400 °C, 3h Tetragonal Chalcocite* Raw Djurleite Tetragonal Chalcocite 200 °C, 3h Djurleite XZ1029 0.2 0.6 130 190 Low Digenite Tetragonal Chalcocite 400 °C, 3h Djurleite Raw Djurleite XZ1031 0.4 1.2 130 168 Low Digenite 200 °C, 3h Tetragonal Chalcocite Raw Djurleite Low Digenite 200 °C, 3h XZ1030 0.9 3.7 130 168 Tetragonal Chalcocite Low Digenite 400 °C, 3h Tetragonal Chalcocite * Minor phase or trace † Room temperature reactions in sealed ampoules stirred outside the glovebox ‡ Capillary data collected

72

Table 3.11: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl and HMDST in CH3CN. Wt% after HT TGA CHNS EDS Cu/S Ratio 200 °C 400 °C Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% CHNS/TGA EDS XZ1029 99% 97% 0.5±0.9 20.0±1.7 65±14 35±14 0±1 2.0±0.2† 1.9±1.1 XZ1029.T400 79±3 0.2±0.9 19.1±1.8 68±7 31±4 0±0 2.1±0.2‡ 2.2±0.3 XZ1030 99% 98% 0.3±0.5 20.8±0.9 69±4 31±4 0±0 1.9±0.1† 2.2±0.4 XZ1030.T400 79±2 68±9 32±7 0±0 2.0±0.3‡ 2.1±0.6 XZ1031 99% Spilled 0.7±0.5 19.7±1.0 68±4 31±4 1±1 2.0±0.1† 2.2±0.4 XZ1033 98% Spilled 0.6±1.4 18.8±2.9 67±6 32±5 1±2 2.2±0.3† 2.1±0.5 XZ1035 90% 0.4±0.6 12.2±1.3 68±9 28±7 4±1 2.3±0.7 XZ1035.T400 0.2±0.6 13.9±1.2 67±2 30±1 4±1 2.1±0.2 XZ1036 97% 0.7±0.9 18.8±1.8 69±3 30±3 2±2 2.2±0.2† 2.3±0.3 XZ1036.T400 79±3 0.6±0.6 19.8±1.1 69±7 30±7 1±1 2.0±0.1‡ 2.3±0.8 XZ1053 97% 1.4±1.5 20.1±3.0 66±6 31±2 3±4 2.0±0.3† 2.1±0.5 XZ1053.T400 80±6 0.3±1.7 20.0±3.4 61±17 39±16 0±0 2.0±0.4‡ 1.6±0.8 XZ1059 Spilled 0.8±1.5 18.6±3.0 64±22 35±23 1±2 2.2±0.4† XZ1059.T400 79±3 0.1±0.5 19.5±1.0 2.1±0.1‡ XZ1060 Spilled 1.2±0.6 20.3±1.2 68±10 30±12 2±7 2.0±0.1† 2.1±0.1 XZ1060.T400 80±4 0.6±1.6 19.8±3.2 2.0±0.3‡ XZ1063 97% 0.7±1.6 20.2±3.1 58±8 42±7 0±0 2.0±0.3† 1.4±0.3 XZ1063.T400 79±4 1.9±0.1* XZ1090 0.1±0.8 19.9±1.3 2.0±0.1† XZ1091 0.7±0.9 20.5±1.5 1.9±0.1† XZ1092 0.7±1.1 19.0±1.8 2.1±0.2† XZ1095 0.7±1.1 18.9±1.8 2.1±0.2† XZ1096 0.8±0.7 18.8±1.2 2.2±0.1† XZ1097 0.8±1.2 18.5±1.9 2.2±0.2† XZ1098 0.8±1.0 18.9±1.6 2.1±0.2† * TGA

† CHNS ‡ CHNS and TGA ** EDS

Interesting phase evolutions were observed for room temperature reactions in acetonitrile. For samples prepared at room temperature with the lowest HMDST concentration (0.4 mL, 0.2 M) in Teflon capped flasks or sealed ampoules, a phase evolution from monoclinic chalcocite to djurleite was observed (Figure 3-12). It was very surprising that a well defined monoclinic chalcocite was recovered after the reaction was carried out for only five minutes. However, with increased reaction time, the crystallinity of the as-recovered samples decreased, and peaks that did not belong to monoclinic chalcocite appeared, although the whole pattern still resembled monoclinic chalcocite

(Figure 3-12 b)). For reactions carried out for more than 24 h, XRD patterns of as-

73

recovered samples resembled djurleite. These results clearly indicated a continuous phase evolution from monoclinic chalcocite to djurleite.

Figure 3-12: XRD patterns of samples prepared at RT with 0.4 mL HMDST for a) 5 min (XZ1113), b) 5 h (XZ1114), c) 24 h (XZ1090), d) 96 h (XZ1091) and e) 168 h (XZ1092). Lines indicate PDF cards of monoclinic chalcocite and djurleite, respectively.

For similar room temperature reactions with higher HMDST concentrations, hexagonal chalcocite, which is a metastable phase at room temperature, was recovered.

For short time reactions (1.5 h), the phases were clearly hexagonal chalcocite (Figure

74

3-13). For similar reaction prepared for seven days (XZ1059, Figure 3-13 c) and

XZ1060), the XRD patterns still resembled hexagonal chalcocite, but they contained additional weak features. These results indicated that the hexagonal chalcocite phase was more stable at higher HMDST concentrations, and disorder was induced much slower.

Hence, relatively phase pure hexagonal chalcocite can be prepared with high HMDST concentrations for short reaction times.

Figure 3-13: XRD patterns of samples prepared at RT for 1.5 h with a) 0.8 mL HMDST (XZ1111) and b) 2.4 mL HMDST (XZ1112), and c) for 7 d with 0.8 mL HMDST (XZ1059). (a) was collected over two different scans. Lines indicate PDF card of hexagonal chalcocite.

For high temperature reactions (70 °C and 130 °C), only djurleite was recovered.

The crystallinity of the djurleite phases increased with preparation temperature (Figure

3-14), and samples prepared at higher temperature contained less amorphous materials.

75

Figure 3-14: XRD patterns of samples prepared at a) 70 °C with 0.8 mL HMDST (XZ1063), b) 70 °C with 2.4 mL HMDST (XZ1036) and c) 130 °C with 2.4 mL HMDST (XZ1030). Lines indicates PDF card of djurleite.

5 Heat treatments of djurleite samples (Cu1.94S to Cu1.97S) usually resulted in mixtures of low digenite (typically Cu1.8S) and tetragonal chalcocite (typically Cu1.96S).

As shown in Table 3.11, the weight loss during heat treatment (up to 400 °C) was usually less than 4%. In comparison, less than 1% of weight loss is necessary to change the most sulfur rich djurleite composition to the chalcocite composition (from Cu1.94S to Cu2S).

The higher weight loss during heat treatments may also be due to residual carbon or residual CuCl. The inability to separate the weight loss contributions of S, C and CuCl makes it impossible to calculate the composition change from the weight loss data. In addition, the copper to sulfur ratios of the as-recovered djurleite samples were very close to 2, but the errors were larger than 0.1, and thus larger than the change expected during conversion to chalcocite. Because the composition regions of these copper sulfide

76

polymorphs are very close, the accuracy of available analytical techniques was not suitable to distinguish them. In addition, because it takes only about 1.6% extra sulfur to completely convert the most sulfur rich djurleite phase (CuS0.515, 80.48 g/mol) to a typical digenite phase (CuS0.556, 81.79 g/mol), it is possible that these samples formed high digenite solid solutions at high temperatures because of extra sulfur rich amorphous materials. Upon cooling, depending on the Cu to S ratio, high digenite solid solutions could transform to low digenite and tetragonal chalcocite. Because the copper to sulfur ratios before and after heat treatment could not be precisely measured, it not clear what caused such phase transitions.

3.3.3.3 Morphology

The morphologies of the samples were characterized by scanning electron microscopy (SEM). Most as-recovered samples consisted of nanoparticles, although agglomeration was significant in many cases. SEM studies also helped to characterize phase evolutions.

Samples prepared in chloroform at room temperature mainly consisted of agglomerated nanospheres and distorted nanospheres (Figure 3-15 and Figure 3-16).

Particle size changed with reaction time and HMDST concentration, and was found to be intricately related to the phase observed in the XRD patterns. For samples prepared with the lowest HMDST concentration, short reaction times (ACS18, 24 h, Figure 3-15 a) and b)) gave hexagonal chaclocite, which consisted of particles with sizes between 10 and 50 nm. For a sample with longer reaction time (XZ1105, 96 h), the phase resembled monoclinic chalcocite, and the sample consisted of particles with sizes between 10 to 100 nm (Figure 3-15 c)). For 7 d of reaction (XZ1102), the sample consisted of well defined

77

monoclinic chalcocite, and particle sizes ranged from 50 to 150 nm (Figure 3-15d)).

These results indicated that the hexagonal chalcocite phase is stabilized for the smallest particle sizes, and as the particle size increases, hexagonal chalcocite slowly converts to monoclinic chalcocite. Samples with a wide size distribution like XZ1105 may consist of a mixture of hexagonal and monoclinic particles, which agrees with the difficult to interpret X-ray patterns.

Figure 3-15: SEM images of RT samples prepared with 0.4 mL HMDST in CHCl3: a) and b) ACS18 (24 h, hexagonal chalcocite), c) XZ1105raw (96 h, monoclinic chalcocite like) and d) XZ1102raw (7 d, monoclinic chalcocite).

Chloroform samples prepared at room temperature with higher concentrations show similar phase/particle size relations (Figure 3-16). Hexagonal chalcocite samples

(ACS19 and XZ1104) consisted of severely agglomerated nanoparticles with sizes below

78

50 nm, and the two samples that resembled monoclinic chalcocite consisted of both small particles and particles with diameters of about 100 nm. This suggested that the phase evolution was related to the stability of nanoparticles.

Figure 3-16: SEM images of RT samples prepared in CHCl3: a) ACS19 (0.8 mL HMDST, 24 h), b) XZ1106raw (0.8 mL HMDST, 96 h), c) XZ1103raw (0.8 mL HMDST, 7 d) and d) XZ1104raw (2.4 mL HMDST, 7d). a) and d) consist of hexagonal chalcocite, and b) and c) consist phases starting to resembled monoclinic chalcocite.

Djurleite samples prepared at 70 °C in chloroform consisted of highly distorted nanospheres, and the particles were severely agglomerated and fused (Figure 3-17). The nanocrystallinity observed in the SEM explained the broad peaks observed in their XRD patterns.

79

Figure 3-17: SEM images of djurleite samples prepared at 70 °C in CHCl3: a)

XZ1011raw (0.8 mL HMDST), and b) XZ1020raw (2.5 mL HMDST).

For samples prepared at 130 °C in chloroform, the sample prepared with the highest HMDST concentration (XZ1022, Figure 3-18 a) and b)) consisted of highly agglomerated low digenite nanospheres similar to the morphology of those prepared at lower temperatures. The sample prepared with the lowest HMDST concentration

(XZ1025, Figure 3-18 c) and d)) consisted of agglomerated networks spanning a few micrometers. It was also found that these agglomerates consisted of nanoparticles, and the surfaces of larger particles were covered by much smaller nanoparticles with diameters of a few nanometers. Given that this sample was composed of low digenite with trace amounts of covellite, it is possible that during the formation of the covellite phase, copper ions diffused out from the bulk to the surface and immediately reacted with extra HMDST in the solvent, thus resulting in smaller particles on the surfaces.

80

Figure 3-18: SEM images of low digenite samples prepared at 130 °C in CHCl3: a) and b) XZ1022raw (2.5 mL HMDST), and c) and d) XZ1025raw (0.4 mL HMDST). XZ1025 also contained trace amounts of covellite.

Similarly to samples prepared in chloroform, samples prepared in acetonitrile at room temperature also consisted of nanospheres and distorted nanospheres (Figure 3-19 and Figure 3-20), and all hexagonal chalcocite samples were composed of very small particles. For samples reacted for 1.5 h, the samples prepared at higher HMDST concentrations (XZ1111, Figure 3-19 a) and XZ1112 Figure 3-19 c)) mainly consisted of severely agglomerated nanospheres, many of which were only a few nanometers in diameter. The small particle sizes in these two samples are consistent with the broad peaks observed for their XRD patterns. The sample prepared for 1.5 h with the lowest

HMDST concentration (XZ1110, Figure 3-20 b), monoclinic chalcocite) also contained a

81

few nanoplatelets with diameters between 50 and 100 nm. Hexagonal chalcocite samples prepared with higher HMDST concentrations for longer reaction time (4 d) also consisted of small nanoparticles (Figure 3-19 b) and d)). Similar to RT samples prepared in chloroform, these resulted further corroborated that hexagonal chalcocite was more stable for very small particle sizes.

Figure 3-19: SEM images of samples prepared at RT in CH3CN a) XZ1111raw (0.8 mL HMDST, 1.5 h), b) XZ1097raw (0.8 mL HMDST, 96 h), c) XZ1112raw (2.4 mL HMDST, 1.5 h), and d) XZ1098raw (2.4 mL HMDST, 96 h). All samples consist of hexagonal chalcocite.

82

Figure 3-20: SEM images of samples prepared at RT with 0.4 mL HMDST in CH3CN: a) XZ1113raw (5 min), b) XZ1110raw (1.5 h), c) XZ1114raw (5 h) d) XZ1053raw (12 h), e) XZ1086 for 96 h and f) XZ1086 for 7.5 d. a) and b) consist of monoclinic chalcocite, c) and d) consist of phases resemble monoclinic chalcocite and djurleite, respectively, and e) and f) consist of djurleite.

83

For samples prepared with the lowest HMDST concentration, particle growth was fast enough that phase pure monoclinic chalcocite was recovered after as little as 5 min of reaction (XZ1113, Figure 3-20 a)). For samples with reaction times up to 5 h (Figure

3-20 a-c), particle morphologies were similar and consistent with monoclinic chalcocite.

For reactions carried out for 12 h or longer, XRD patterns of as-recovered samples started to resemble djurleite. For the 12 h reaction, the sample consisted of larger particles with diameters between 100 nm and 200 nm. In addition, nanospheres with diameters of a few nanometers were observed on the surfaces of the larger particles. This could be caused by diffusion of copper atoms to the surface and subsequent reaction with solution sulfur sources. These results suggested that as particles grow larger, copper deficiencies are favored, resulting in a phase evolution from monoclinic chalcocite toward djurleite.

Phase pure djurleite was recovered after 4 and 7.5 d, and samples contained more platelets as the reaction time increased (Figure 3-20 e) and f)).

It is possible that this observed morphology evolution was related to the phase evolution. Hexagonal chalcocite could be thermodynamically stable for small nanoparticles, while monoclinic chalcocite is the most stable for larger particles. Hence, samples prepared with higher HDMST concentrations resulted in hexagonal chalcocite, and the hexagonal chalcocite phase remained stable for longer reaction times due to low growth rates in higher HMDST concentrations. As a smaller number of nuclei is formed, crystals grow much faster in low HMDST concentrations, thus resulting in a swift phase shift from hexagonal chalcocite to monoclinic chalcocite. In addition, copper deficient phases appear to be thermodynamically favored for larger particles. The diffusion of Cu atoms results in a slow and continuous composition shift towards djurleite.

84

For djurleite samples prepared at higher temperatures, similar morphologies to those prepared at room temperature with much larger agglomerates were observed

(Figure 3-21 a)). For larger agglomerates in a sample prepared at 130 °C (XZ1031), the surfaces were covered with much smaller particles with sizes below 100 nm. As the

200 °C heat treatment of this sample resulted in formation of low digenite and tetragonal chalcocite, it is possible that this sample contained sulfur rich amorphous materials.

Figure 3-21: SEM images of djurleite samples prepared with 0.8 mL HMDST at 130 °C in CH3CN (XZ1031).

3.3.4 CuCl2 and HMDST in Solvents

With CuCl2 as the copper source, all samples prepared by reactions with HMDST in chloroform and acetonitrile resulted in covellite. Most of these samples were heated to

400 °C , and some were heated to 200 °C. The XRD results for chloroform and acetonitrile samples are shown in Table 3.12 and Table 3.13, respectively. Elemental analysis results are shown in Table 3.14 and Table 3.15, respectively. XRD analysis showed that most samples contained non-negligible amounts of amorphous material, especially for samples prepared at room temperature. For reaction at 130 °C, a large amount of CuCl, which might be a side product of the reaction of CuCl2 and HMDST in

85

chloroform, was found in as-recovered sample XZ1039 (0.4 M HMDST, 130 °C for 7 d).

A similar sample (XZ1057) prepared under identical conditions, except that the CuCl2 starting material was finely ground, showed a reduced amount of CuCl. Reacting an identical sample for 15 d reduced the amount of CuCl to a trace. It appeared that the reaction rate under these conditions was very slow. The carbon content of these samples was generally low, often below the detection limit, except for samples prepared in chloroform at 130 °C (Table 3.14). Heat treatments resulted in conversion to low digenite and tetragonal chalcocite.

86

Table 3.12: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CHCl3. Conditions Thermal Sample [HMDST] S:Cu Phases T/°C t/h History Raw Covellite 200 °C, 3h Covellite XZ1027 0.4 1.3 RT 96 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Raw Covellite 200 °C, 3h Covellite XZ1028 0.9 3.8 RT 98 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Raw Covellite XZ1032** 0.9 9.6 RT 96 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Raw Covellite XZ1042 0.4 1.3 70 168 400 °C, 3h Low Digenite Raw Covellite XZ1043 0.9 3.8 70 168 Low Digenite 400 °C, 3h Covellite Raw Covellite XZ1044** 0.9 8.6 70 168 Low Digenite 400 °C, 3h Covellite* Covellite Raw CuCl XZ1039 0.4 1.2 130 168 Low Digenite 400 °C, 3h Tetragonal Chalcocite CuCl Covellite Raw CuCl XZ1057 0.4 1.1 130 168 Low Digenite 400 °C, 3h Tetragonal Chalcocite CuCl Covellite XZ1085 0.4 1.2 130 15 d Raw CuCl* Raw Covellite XZ1038 0.9 3.0 130 168 400 °C, 3h Low Digenite Raw Covellite XZ1040** 0.9 8.5 130 168 Low Digenite 400 °C, 3h Tetragonal Chalcocite * Minor phase or trace ** Samples prepared with about 1.5 mmol CuCl2

87

Table 3.13: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CH3CN. Conditions Thermal Sample [HMDST] S:Cu Phases T/°C t/h History Raw Covellite XZ1034 0.2 0.6 RT 96 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Raw Covellite XZ1037 0.4 1.2 RT 96 400 °C, 3h Low Digenite Raw Covellite XZ1041 0.9 3.7 RT 96 Low Digenite 400 °C, 3h Tetragonal Chalcocite* Raw Covellite XZ1045** 0.9 8.4 RT 96 400 °C, 3h Low Digenite Raw Covellite XZ1062 0.2 0.6 70 168 400 °C, 3h Low Digenite Raw Covellite XZ1046 0.4 1.2 70 168 400 °C, 3h Low Digenite Raw Covellite XZ1047 0.9 3.7 70 168 400 °C, 3h Low Digenite Raw Covellite XZ1048** 0.9 8.0 70 168 Low Digenite 400 °C, 3h Tetragonal Chalcocite Raw Covellite XZ1058 0.2 0.6 130 168 400 °C, 3h Low Digenite Raw Covellite XZ1050 0.4 1.2 130 168 Low Digenite 400 °C, 3h Covellite Raw Covellite XZ1051 0.9 3.8 130 168 Low Digenite 400 °C, 3h Covellite Raw Covellite XZ1052*** 0.2 1.1 130 168 400 °C, 3h Low Digenite * Minor phase or trace ** Samples prepared with about 1.5 mmol CuCl2 *** Samples prepared with 15 mL solvent

88

Table 3.14: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CHCl3. Wt% after HT TGA CHNS EDS Cu/S Ratio 200 °C 400 °C Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% CHNS/TGA EDS XZ1027 97% 84% 0.5±0.8 29.2±1.5 60±13 38±15 2±5 1.2±0.1* 1.5±0.7 XZ1027.T400 78±4 0.6±1.2 20.9±2.3 66±7 34±4 0±0 1.9±0.2† 2.0±0.3 XZ1028 95% 82% 0.6±0.7 29.4±1.3 58±12 41±12 1±1 1.2±0.1* 1.4±0.7 XZ1028.T400 79±4 0.7±1.1 21.3±2.2 67±7 33±4 0±0 1.9±0.2† 2.1±0.3 XZ1032 85% 0.8±0.3 26.7±0.6 57±4 42±5 1±2 1.4±0.1* 1.3±0.2 XZ1032.T400 76±6 70±6 29±7 1±1 2.4±0.7 XZ1038 85% 3.0±1.5 31.3±2.9 52±8 46±7 2±1 1.1±0.1* 1.1±0.3 XZ1038.T400 74±3 2.8±1.2 22.5±2.4 66±7 34±4 0±0 1.7±0.2† 1.9±0.3 XZ1039 70% 3.5±1.1 22.6±2.2 49±3 44±7 7±4 1.0±0.2 XZ1039.T400 70±4 3.7±1.0 18.6±2.1 67±5 29±2 3±4 1.9±0.2† 2.2±0.3 XZ1040 85% 3.7±1.0 31.5±2.0 53±4 46±4 1±1 1.0±0.1* 1.1±0.2 XZ1040.T400 74±4 3.6±1.3 22.5±2.7 65±7 35±4 0±0 1.7±0.2† 1.8±0.3 XZ1042 87% 0.3±1.4 31.1±2.8 56±2 43±2 1±1 1.1±0.1* 1.3±0.1 XZ1042.T400 76±4 0.4±1.6 22.3±3.2 62±4 38±4 0±0 1.7±0.3† 1.7±0.3 XZ1043 88% 0.7±0.7 31.3±1.4 55±7 44±9 1±2 1.1±0.1* 1.3±0.4 XZ1043.T400 76±3 0.7±1.1 22.8±2.3 59±6 41±5 0±0 1.7±0.2† 1.4±0.2 XZ1044 0.7±1.2 31.3±2.4 54±10 45±9 1±1 1.1±0.1* 1.2±0.5 XZ1044.T400 78±8 0.7±1.0 23.2±2.0 63±15 37±15 0±0 1.7±0.2† 1.7±1.1 XZ1057 82% 2.3±1.0 29.3±2.0 57±15 41±13 2±2 1.2±0.1* 1.4±0.8 XZ1057.T400 79±17 2.1±1.6 21.0±3.2 * Calculated from CHNS Data † Calculated from CHNS and TGA data

Table 3.15: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl2 and HMDST in CH3CN. Wt% after HT TGA CHNS EDS Cu/S Ratio 200 °C 400 °C 600 °C Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% CHNS/TGA EDS XZ1034 76% 0.8±1.1 31.8±2.2 54±9 45±8 1±1 1.1±0.1* 1.2±0.4 XZ1034.T400 78±4 0.6±1.6 21.2±3.2 68±7 32±4 0±0 1.9±0.3† 2.1±0.3 XZ1037 84% 0.6±1.4 30.9±2.8 55±4 44±5 2±3 1.1±0.1* 1.2±0.2 XZ1037.T400 76±7 0.4±1.9 21.7±3.7 65±7 35±4 0±0 1.8±0.4† 1.9±0.3 XZ1041 81% 0.8±1.1 30.5±2.1 54±4 45±6 1±1 1.1±0.1* 1.2±0.2 XZ1041.T400 79±7 0.4±0.9 21.2±1.8 68±3 32±3 0±0 1.9±0.2† 2.2±0.3 XZ1045 84% 1.0±2.5 31.5±5.1 55±4 44±3 1±1 1.1±0.2* 1.2±0.2 XZ1045.T400 77±8 1.0±1.1 23.1±2.2 63±10 37±9 0±0 1.7±0.2† 1.7±0.5 XZ1046 84% 0.6±1.7 31.0±3.3 60±16 39±16 0±0 1.1±0.1* 1.6±1.1 XZ1046.T400 78±7 0.4±1.5 20.7±3.0 67±2 33±2 0±0 1.9±0.3† 2.0±0.2 XZ1047 84% 0.9±1.9 30.6±3.8 57±1 43±1 0±0 1.1±0.1* 1.3±0.1 XZ1047.T400 76±6 0.7±1.6 22.4±3.2 65±9 35±9 0±0 1.7±0.3† 1.8±0.7 XZ1048 83% 1.1±1.3 30.9±2.5 56±3 44±3 0±0 1.1±0.1* 1.3±0.2 XZ1048.T400 76±4 2.1±1.7 21.3±3.4 68±7 32±4 0±0 1.8±0.3† 2.1±0.4 XZ1050 91% 1.2±0.7 31.5±1.5 58±7 41±8 0±0 1.1±0.1* 1.4±0.4 XZ1050.T400 68±10 0.9±1.3 26.2±2.6 61±10 39±9 0±0 1.3±0.2† 1.6±0.5 XZ1051 90% 1.2±1.8 31.2±3.7 56±6 44±5 0±0 1.1±0.1* 1.3±0.2 XZ1051.T400 73±7 1.5±2.6 25.4±5.3 56±6 44±5 0±0 1.4±0.3† 1.3±0.2 XZ1052 84% 0.8±1.1 32.0±2.3 56±6 44±5 0±0 1.1±0.1* 1.3±0.2 XZ1052.T400 78±5 0.9±1.9 21.6±3.7 1.8±0.3† XZ1058 80% 2.7±1.7 29.7±3.5 49±12 47±12 4±1 1.2±0.1* 1.0±0.5 XZ1058.T400 74±15 2.4±1.0 21.4±2.0 XZ1062 83% 0.9±1.3 28.6±2.5 60±9 36±5 4±1 1.2±0.1* 1.5±0.3 XZ1062.T400 76±5 0.4±0.8 22.3±1.5 1.7±0.2† * Calculated from CHNS Data † Calculated from CHNS and TGA data

89

For 200 °C heat treatment of covellite samples XZ1027 and XZ1028, the covellite phase was retained after heating. This is consistent with the phase diagram, which shows stoichiometric covellite (CuS) to be stable up to 507 °C . For 400 °C heat treatments, the cooled powders had transformed into low digenite. In some of these heat treated samples, tetragonal chalcocite was present as a minor phase as well. Because these heat treatments were carried out under flowing argon in an open system, instead of under equilibrium conditions in a closed system (lack of saturated sulfur vapor), it is possible that covellite samples lost sulfur and decomposed below 500 °C. This was confirmed by CHNS analysis, as sulfur content in CuCl-free samples decreased from an average of approximately 31% to an average of about 21% after heating to 400 °C (Table 3.14 and

Table 3.15). This would result in an average change of copper to sulfur ratio from approximately 1.1 to 1.9, and an average weight loss of about 13% if samples only contained sulfur and copper. This approximate composition change agreed with the phase change from covellite to digenite, and the sulfur loss was close to the average weight loss after heat treatments (about 16%). Therefore, covellite samples could lose sulfur and convert to copper rich phases during heat treatments.

The XRD patterns of as-recovered and heat treated XZ1028 (RT, 2.4 mL HMDST,

CHCl3) are shown in Figure 3-22 a) and d), respectively. Heat treatment at 200 °C did not induce any phase transition, but significantly increased the crystallinity of the sample, although both still contained a large amount of amorphous material. It was found that the crystallinity of the as-recovered samples also increased with preparation temperature

(Figure 3-22). The amorphous contribution in samples prepared at 130 °C (XZ1038,

Figure 3-22c)) was negligible. Hence, crystallinity could be increased both by higher

90

reaction temperature and subsequent heat treatment, while reaction completeness depended on reaction temperature.

Figure 3-22: XRD patterns of chloroform samples prepared at a) RT with 2.4 mL HMDST (XZ1028), b) 70 °C with 2.4 mL HMDST (XZ1043), c) 130 °C with 2.4 mL HMDST (XZ1038) and d) 200 °C heat treatment of XZ1028. Lines indicate the PDF card of covellite.

Samples prepared in acetonitrile showed similar trends as those prepared in chloroform (Figure 3-23). For identical room temperature reactions, the sample prepared in acetonitrile (Figure 3-23 b)) showed better crystallinity than the one prepared in chloroform (Figure 3-23 a)). For acetonitrile samples, the crystallinity of the as-recovered samples did not increase much at 70 °C, but increased significantly at 130 °C (Figure

3-23). Similar to chloroform samples, samples prepared in acetonitrile at both room temperature and 70 °C contained significant amounts of amorphous material, while the amorphous component was negligible in samples prepared at 130 °C.

91

Figure 3-23: XRD patterns of samples prepared with 0.8 mL HMDST a) in CHCl3 at RT (XZ1027), and in CH3CN at b) RT (XZ1037), c) 70 °C (XZ1046) and d) 130 °C (XZ1050). Lines indicate the PDF card of covellite.

All covellite samples prepared in chloroform consisted of severely agglomerated and partially fused nanoparticles (Figure 3-24). In many of these samples, covellite nanoparticles appeared to be agglomerated into stick-like structures with widths around

100 nm (Figure 3-24 a) and c)). High magnification images (Figure 3-24 d)) showed that the elongated structures it can be seen that the bundles were composed of much smaller nanorods and fused nanospheres. It is interesting that these samples tended to agglomerate along one direction. For solvo-thermo reactions, (XZ1044, 2.4 mL and 0.9

M HMDST, 70 °C , Figure 3-24 b)), micron sized structures formed by agglomerated nanosheets were observed. Samples prepared at 130 °C also consisted of agglomerated nanorods and fused nanospheres tended to align in one direction.

92

Figure 3-24: SEM images of covellite samples prepared in CHCl3: a) XZ1032raw (0.9 M HMDST, RT, 96 h), b) XZ1044raw (0.9 M HMDST, 70 °C, 7 d), c) and d) XZ1027raw, (0.4 M HMDST, RT, 96 h), and e) and f) XZ1038raw (0.9 M HMDST, 130 °C, 7 d).

Covellite samples prepared in acetonitrile consisted of thin nanoplatelets (Figure

3-25). The thickness of the nanoplatelets could range from a few nanometers to 20 nm in the same sample (Figure 3-25 a) and b)). Nanoplatelets also formed irregular (Figure 3-25

93

c)) and flower-like (Figure 3-25 d)) agglomerates. The particle morphologies of samples prepared at higher temperatures were similar to those prepared at room temperature except that agglomeration was more severe and particle sizes were larger.

Figure 3-25: SEM images of covellite samples prepared in CH3CN: a) and b) XZ1041 (0.9 M HMDST, RT, 96 d) c) XZ1045, (1.4 mmol CuCl2, 0.9 M HMDST, RT, 96 h), and d) XZ1051 (0.9 M HMDST, 130 °C, 7 d).

3.3.5 Synthesis with DTBS

Copper sulfide samples prepared by reactions of CuCl or CuCl2 with DTBS in chloroform or acetonitrile were mainly explored by a former group member, Nathalie

94

Pedoussaut. Several samples were also reproduced by an exchange student, Anne Soldat.

A summary of the results of both previous and current work are presented in this section.

The ten samples (XZ1064 to XZ1073) prepared at 70 °C were generally unsuccessful, as none or very little precipitate was recovered even when reaction times were extended to 120 d. Unknown phases were recovered from a few samples. This is consistent with previous work, and no further analysis was attempted. Crystalline copper sulfide precipitates were recovered from most samples prepared at 100 °C and 130 °C, and their elemental and XRD analysis results are shown in Table 3.16 and Table 3.17, respectively. In order to comprehensively understand the influence of synthetic conditions, a summary of the phases in the as-recovered samples of both previous and current work is shown in Table 3.18.

Table 3.16: Elemental analysis results of raw and heated copper sulfide samples prepared by reactions of CuCl or CuCl2 and DTBS in CHCl3 or CH3CN. Wt left TGA CHNS EDS Cu/S Ratio 400 °C Cu Wt% C Wt% S Wt% Cu Atom% S Atom% Cl Atom% XZ1074 45±4 12.4±1.6 14.4±1.3 68±8 29±4 3±1 2.2±0.4* XZ1074.T400 65% 6.9±1.4 19.8±1.2 67±3 32±3 1±1 2.1±0.3* XZ1075 3.4±2.6 19.5±2.2 68±5 29±5 3±8 2.2±0.4* XZ1075.T400 84% 79±7 1.9±1.6 20.5±1.3 67±4 32±3 1±4 2.0±0.2† XZ1076 75±2 0.3±1.3 18.8±1.0 65±11 32±11 3±1 2.0±0.1† XZ1076.T400 93% 0.2±1.4 19.8±1.2 68±8 31±7 1±1 2.0±0.1‡ XZ1077 77±2 0.3±1.8 21.4±1.5 67±7 33±7 1±1 1.8±0.1† XZ1077.T400 98% 0.1±0.9 21.3±0.8 66±14 34±14 0±0 1.9±0.1‡ XZ1078 75±5 0.5±1.8 22.4±1.5 63±15 35±13 3±2 1.7±0.2† XZ1078.T400 94% 0.5±4.3 21.3±3.5 70±13 30±12 0±0 1.9±0.3‡ XZ1080 80±7 1.6±2.5 20.2±2.1 66±3 32±2 2±1 2.0±0.3† XZ1080.T400 91% 68±11 31±11 1±1 2.2±1.2* XZ1081 81±3 1.1±1.8 19.5±1.4 67±7 30±7 3±1 2.1±0.2† XZ1081.T400 97% 0.4±2.0 20.0±1.6 2.0±0.2‡ XZ1082 0.3±1.8 9.0±1.5 * Calculated from EDS Data † Calculated from CHNS and TGA data ‡ Calculated from CHNS data

95

Table 3.17: X-ray diffraction results for raw and heated copper sulfide samples prepared by reactions of CuCl or CuCl2 with DTBS. [DTBS] Conditions Thermal Sample Solvent S:Cu Phases (M) T/°C t/h History Low Digenite Raw Unknown Crystalline Phase

XZ1074 0.2 CH3CN 0.7 130 15 d Low Digenite 400 °C, 3h Tetragonal Chalcocite Unknown Crystalline Phase Raw Low Digenite XZ1075 0.4 CH CN 1.5 130 15 d Low Digenite 3 400 °C, 3h Tetragonal Chalcocite Low Digenite Raw CuCl Unknown Phase* XZ1076 0.2 CHCl 0.7 130 15 d 3 Low Digenite 400 °C, 3h Tetragonal Chalcocite CuCl* Raw Low Digenite XZ1077 0.4 CHCl 1.3 130 10 d Low Digenite 3 400 °C, 3h Tetragonal Chalcocite* Raw Djurleite XZ1081 0.4 CH CN 1.5 100 12 d Monoclinic Chalcocite 3 400 °C, 3h Tetragonal Chalcocite Major CuCl XZ1082 0.4 CHCl 1.3 100 12 d Raw 3 Chal/Dj Low Digenite Raw ** Covellite XZ1078 0.4 CHCl3 1.3 130 10 d Low Digenite 400 °C, 3h Tetragonal Chalcocite

** Raw Djurleite XZ1080 0.9 CH3CN 3.6 100 12 d 400 °C, 3h Tetragonal Chalcocite * Minor phase or trace ** CuCl2 as the copper source

96

Table 3.18: X-ray diffraction results for as-recovered samples prepared with DTBS. NPSu and ACS samples were prepared by Nathalie Pedoussaut and Anne Soldat, respectively. NPSu samples and ACS8 were prepared with 15 mL solvents, and the rest of the samples were prepared with 10 mL. "(r)" indicates an additional identical sample was reproduced. Cu [DTBS] Conditions Sample Solvent S:Cu Phases Source (M) T/°C t/d

XZ1081 CuCl 0.4 CH3CN 1.5 100 12 Djurleite Low Digenite ACS9 CuCl 0.4 CH CN 1.5 100 12 3 Djurleite Major CuCl XZ1082 CuCl 0.4 CHCl 1.3 100 12 3 Chal/Dj

ACS10 CuCl 0.4 CHCl3 1.3 100 12 Low Digenite, CuCl XZ1080 CuCl2 0.9 CH3CN 3.6 100 12 Djurleite

ACS7 CuCl2 0.9 CH3CN 3.6 100 12 Low Digenite

NPSu45 CuCl2 0.6 CHCl3 3.7 100 11 Low Digenite

ACS8 CuCl2 0.6 CHCl3 3.7 100 11 Low Digenite Low Digenite XZ1074 CuCl 0.2 CH CN 0.7 130 15 3 Unknown Phase

XZ1075 CuCl 0.4 CH3CN 1.5 130 15 Low Digenite Low Digenite NPSu33 CuCl 0.6 CH CN 3.7 130 7 3 Unknown Phase

XZ1076 CuCl 0.2 CHCl3 0.7 130 15 Low Digenite, CuCl Poorly Crystalline Low NPSu68* CuCl 0.6 CHCl 1.1 130 7 3 Digenite, CuCl

XZ1077 CuCl 0.4 CHCl3 1.3 130 10 Low Digenite

NPSu31(r) CuCl 0.6 CHCl3 3.7 130 7 Low Digenite

NPSu39(r) CuCl 1.2 CHCl3 7.3 130 7 Low Digenite

NPSu77 CuCl 1.6 CHCl3 11.0 130 7 Low Digenite

XZ1079 CuCl2 0.4 CH3CN 1.4 130 10 6.5 mg recovered Low Digenite NPSu34(r) CuCl 0.6 CH CN 3.7 130 7 2 3 Unknown Phase

NPSu69* CuCl2 0.6 CHCl3 1.1 130 7 CuCl, Unknown phase

XZ1078 CuCl2 0.4 CHCl3 1.3 130 10 Low Digenite, Covellite

NPSu32(r) CuCl2 0.6 CHCl3 3.7 130 7 Low Digenite, Covellite

NPSu40(r) CuCl2 1.2 CHCl3 7.3 130 7 Low Digenite, Covellite

NPSu78 CuCl2 1.6 CHCl3 11.0 130 7 Low Digenite * 10 mmol copper source

97

Reactions of CuCl with DTBS at 130 °C in both solvents resulted in the formation of low digenite. For samples with low DTBS concentration (XZ1074) or shorter reaction time (NPSu33) in acetonitrile, a second unknown phase was also present. Elemental analysis of these samples indicated that XZ1074 and NPSu33 contained about 12% and 9% of carbon, respectively. Although the unknown crystalline phase was not observed in

XZ1075, it also contained about 3% of carbon. This indicates that there were significant amounts of organic residues in all samples prepared in acetonitrile at 130 °C. In addition,

CHNS analysis confirmed non-zero amounts of nitrogen. It has been reported that acetonitrile can form complexes with Cu+, and it is possible that higher DTBS concentrations or longer reaction times are needed to overcome complexing effects and form digenite.66

For CuCl2 samples reacted in acetonitrile at 130 °C, little or no precipitate was recovered for most samples (Table 3.2). When some powder could be recovered (higher

DTBS concentration (0.65 M, NPSu34 and NPSu34r), a mixture of low digenite and an unknown phase was observed. Formation of soluble complexes with acetonitrile may be responsible for the poor yields. For CuCl2 samples reacted in chloroform at 130 °C

(Figure 3-26), a mixture of low digenite and covellite was recovered. The amount of covellite decreased with increasing concentration of DTBS. It is possible that low digenite converted to covellite, a more copper deficient phase, and the rate decreased with increasing the DTBS concentration, similar to results found for reactions between

CuCl and HMDST at 130 °C.

98

Figure 3-26: XRD patterns of CuCl2 samples prepared in 15 mL of CH3CN for 7 d with a) 2.0 mL DTBS (NPSu32), b) 4.0 mL DTBS (NPSu40) and c) 6.0 mL DTBS (NPSu78). Lines indicate PDF cards of covellite (bottom) and low digenite (top).

Reactions at 100 °C gave digenite, djurleite or a mixture of djurleite and digenite.

Two samples prepared under identical conditions (CuCl, 0.4 M DTBS) except for particle size of the copper chloride starting material (XZ1081, coarse powder, ACS9, fine powder) resulted in phase pure djurleite and a mixture of djurleite and digenite, respectively

(Figure 3-27 . It is possible that the kinetics for XZ1081 were slower due to the larger particle size, and that longer reaction times would also have resulted in partial conversion to digenite. This would be similar to the phase evolution observed for reactions of CuCl with HMDST in CHCl3. It may be possible to develop conditions for the targeted synthesis of djurleite or digenite phases for reactions at 100 °C, but the slow reaction rate, as evident from the presence of significant amounts of unreacted CuCl after 12 d of reaction, makes these reaction conditions inefficient.

99

Figure 3-27: XRD patterns of a) XZ1081 (CuCl, 0.8 mL DTBS, CH3CN) and b) ACS9(CuCl, 0.8 mL DTBS, CH3CN). Lines indicate PDF cards of low digenite (bottom) and djurleite (top), respectively.

The morphologies of the DTBS samples were also studied. Similarly to HMDST samples prepared in acetonitrile, nanoplatelets were also formed by reactions of CuCl with DTBS in acetonitrile (Figure 3-28 a)), although the diameters of the nanoplatelets were much larger than those prepared with HMDST. Reactions in chloroform resulted in micron sized agglomerates. However, nanoparticles were observed at higher magnifications. For the djurleite sample XZ1081, it was also found that the surfaces of larger particles was covered by smaller nanospheres with diameters of a few nanometers.

It is possible that these nano sized attachments to the surface were formed by reactions of diffused copper ions with DTBS. This further confirmed the phase evolution from djurleite phase toward more copper deficient low digenite.

100

Figure 3-28: SEM images of DTBS samples: a) XZ1075raw (CuCl, CH3CN, 0.8 mL DTBS at 130 °C) b) and c) XZ1077raw (CuCl, CHCl3, 0.8 mL DTBS at 130 °C), and d) XZ1081 (CuCl, CH3CN, 0.8 mL DTBS at 100 °C).

3.3.6 Phase and Morphology Evolution

Throughout this project, various copper sulfide polymorphs were obtained, and phase evolutions from less copper deficient phases towards more copper deficient phases were observed. The phase control was closely related to phase evolutions of copper sulfide nanoparticles. For reactions with CuCl2 and HDMST, only covellite was obtained.

For synthesis with DTBS, djurleite, low digenite and covellite were recovered, but precise phase control was difficult. Most interesting results were obtained for reactions between CuCl and HMDST, and precise phase control by fine tuning synthetic

101

parameters was possible. Reaction conditions resulting in different polymorphs are summarized in Figure 3-29.

Figure 3-29: Conditions for targeted synthesis of several copper sulfide polymorphs for reactions of CuCl with HMDST.

In conclusion, it was found that higher temperatures, longer reaction times and lower HMDST concentrations could facilitate phase evolutions towards more copper deficient phases. Continuous phase evolutions from hexagonal chalcocite to monoclinic chalcocite and from monoclinic chalcocite to djurleite were observed. XRD analysis revealed that intermediate phases between hexagonal chalcocite, monoclinic chalcocite and djurleite could exist. It is possible that these intermediate structures were variations of the base structures to accommodate higher copper deficiencies. Although chalcocites

102

and djurleite were reported to have composition regions, it is the first time that continuous changes and phase evolutions between these structures were observed.

The phase evolutions were also supported by morphology studies. Hexagonal chalcocite formed by room temperature reactions in acetonitrile consisted of small nanospheres or distorted nanospheres, while monoclinic chalcocite and djurleite nanoparticles were larger and consisted of platelets. According to the phase diagram of the Cu-S system, monoclinic chalcocite is the thermodynamically stable phase at room temperature, while the direct formation of the metastable hexagonal chalcocite phase was observed for many cases. Given that the sizes of the obtained hexagonal chalcocite nanoparticles were much smaller than those of monoclinic chalcocite, it is possible that hexagonal chalcocite is energetically favored for small particles. Hence, the hexagonal chalcocite phase could be stabilized by small nanoparticles. However, as the particles grew larger during reactions, the more stable monoclinic polymorph was formed. Copper deficiency appears to be energetically favored in the chosen solution media, resulting in diffusion of copper atoms to the surface of particles and a continuous structural change towards the more copper deficient djurleite phase.

In addition to the size effect on the stability of different polymorphs, the solvent conditions also played an important role. At room temperature, higher HDMST concentration always resulted in slower phase evolution and smaller particle sizes. It is possible that higher HMDST concentrations could result in more terminal sulfur atoms on the surfaces of copper sulfide particles, which could increase the energy barrier for copper atoms to diffuse into solution. In addition, higher HMDST concentrations could result in larger numbers of nucleation centers for initial reaction, as the possibility for

103

collisions between reactants is higher, resulting in smaller particle sizes due to parallel growth of the nucleation centers. This is supported by the sizes of the particles observed for different HMDST concentrations.

In addition to phase evolutions observed during synthesis, it was also found that chalcocite and djurleite samples could convert to digenite during grinding. Because of the difficulty to clearly distinguish hexagonal chalcocite, monoclinic chalcocite and djurleite,

16 samples were packed in capillaries and measured at the Advanced Photon Source

(APS). To facilitate packing, each of these samples was ground for about half an hour in air. Based on the data obtained from the APS, grinding resulted in digenite in all cases.

These results provided additional evidence that phase evolutions towards more copper deficient phases were favored and could be caused by defects.

3.4 Effect of Synthetic Parameters

3.4.1 Effect of Copper Sources

Various copper sulfide phases were obtained for samples prepared with CuCl.

Samples prepared with CuCl2 tended to result in covellite except for reactions with DTBS in acetonitrile. Therefore, CuCl2 is a suitable precursor for covellite, while CuCl can be used to target different phases.

3.4.2 Effect of Sulfur Sources

+ + HMDST is more reactive than DTBS, as (CH3)3Si is more stable than (CH3)3C .

Reactions with HMDST can go to completion in a few hours while they take days or weeks with DTBS. Moreover, reactions with DTBS at room temperature or even up to

104

70 °C were unsuccessful. More copper sulfide polymorphs were obtained with HMDST than with DTBS.

3.4.3 Effect of Solvents

The solubility of CuCl and CuCl2 in chloroform is poor, but both halides dissolve well in acetonitrile. Because of this, less unreacted CuCl was found for samples prepared with acetonitrile. Also, it is possible to obtain products with better homogeneity in acetonitrile. Phase selection can be achieved by using different solvents as the phase evolution differs in different solvent environments.

3.4.4 Effect of Temperature

Temperature can be used to achieve phase selection along with solvent choice.

Higher temperatures usually facilitate phase evolution towards more copper deficient phases. Increasing temperature usually results in better crystallinity, but also more agglomeration of nanoparticles.

3.5 Conclusions

Monoclinic chalcocite, hexagonal chalcocite, djurleite, low digenite and covellite nanoparticles were successfully prepared by reactions of copper halides with HMDST or

DTBS without solvent or in acetonitrile or chloroform. Hexagonal chalcocite nanospheres, monoclinic chalcocite nanospheres, djurleite nanospheres and nanoplatelets, low digenite nanospheres and flower-like covellite nanostrcutures were obtained, although they were not uniform and tended to agglomerate. Hence, precise phase control

105

of copper sulfide polymorphs was achieved by fine tuning the synthetic parameters, and nanoparticles of these polymorphs could be obtained.

High digenite was observed during in situ variable temperature XRD analysis.

Low temperature heat treatments (<200 °C) can increase the crystallinity of covellite, digenite and monoclinic chalcocite samples without inducing phase transitions. In addition, tetragonal chalcocite was obtained from heat treatments, but most tetragonal chalcocite samples contained other impurity phases. Phase pure tetragonal chalcocite was recovered from several reactions, indicating that targeted synthesis may be possible.

Precise stoichiometry control is necessary, which would require careful and reproducible adjustments to amount of material heated, argon flow rate, and length of heat treatment.

106

Chapter 4

4. Synthesis and Characterization of Tantalum Sulfides

4.1 Introduction

4.1.1 Polymorphs and Applications of Tantalum Sulfides

Tantalum can form a number of sulfide polymorphs, which can be generally

1-5 categorized into three different types: TaS2, TaS3 and Ta6Sn subsulfides (n=1, 3, 4). The crystallographic parameters of these polymorphs are shown in Table 4.1. The structures and properties of these tantalum sulfide polymorphs are briefly described in the following paragraphs. As the stabilities of these polymorphs are not fully understood, there is no phase diagram reported for the Ta-S binary system.

TaS2 is a layered sulfide that adopts structures related to the CdI2-type structure

(Figure 1-2). 1T-TaS2, 2H-TaS2, 3R-TaS2, 4H-TaS2, and 6R-TaS2 polymorphs are known, where the numbers indicate how many layers of TaS2 are in the unit cell, and capital letters indicate the crystal system.2,6

107

Table 4.1: Crystallographic parameters of tantalum sulfide polymorphs. Space Phase System Structural Parameters Ref. group

1T-TaS2 Trigonal a = 3.36 Å, c = 5.90 Å P m 2

2H-TaS2 Hexagonal a = 3.32 Å, c = 12.10 Å P63/mmc 2

3R-TaS2 Rhombohedral a = 3.32 Å, c = 17.9 Å R3m 2

4H-TaS2 Hexagonal a = 3.33 Å, c = 23.62 Å P63/mmc 6

6R-TaS2 Rhombohedral a = 3.34 Å, c = 35.85 Å R3m 2 a = 36.804 Å, b = 15.173 Å, o-TaS Orthorhombic C222 7 3 c = 3.340 Å 1 a = 9.515 Å, b = 3.341 Å, m-TaS Monoclinic P2 /m 8 3 c = 14.912 Å, β = 109.99° 1 a = 7.478, b = 17.222 Å, Ta S Orthorhombic Abm2 5 3 2 c = 5.605 Å a = 7.381 Å, b = 5.574 Å, Ta S Orthorhombic Pbcm 4 2 c = 15.195 Å a = 14.158 Å, b = 5.284 Å, Ta S Monoclinic C2/c 4 6 c = 14.789 Å, β = 118.01°

Jellinek investigated the TaS2 system, and first provided correct structural

2 information for the 1T, 2H, 3R and 6R types of TaS2 in 1962. 1T-TaS2 adopts the CdI2 structure, where anions form hexagonal close packed arrays, and cations occupy half of the octahedral sites. The octahedral coordination of Ta atoms, which are located at the center of two parallel layers of sulfur atoms, is shown in Figure 4-1 a) and Figure 4-2 a).

For the superstructures of TaS2, due to different positioning of sulfur atoms, sulfur atoms can form a trigonal prismatic coordination environment of the Ta atoms, where sulfur atoms stack on top of each other (Figure 4-1 b)). The stacking of Ta and S atoms in 1T-

108

TaS2 and TaS2 superstructures can be depicted by the cross sections of the (110) planes

(Figure 4-1 c)) of their unit cells (Figure 4-2). Unlike the octahedral coordination in 1T-

TaS2, in 2H- and 3R-TaS2, sulfur atoms form a trigonal prismatic coordination environment of the tantalum atoms, while in 6R-TaS2, there are both octahedral and trigonal prismatic coordination environments. Waszczak et. al. first reported 4H-TaS2, and concluded that its structure was analogous to 4H-TaSe2, but no structural refinement was attempted.6

Figure 4-1: (a) Octahedral and (b) trigonal prismatic coordination sites for Ta atoms (darker) in TaS2 with the (110) planes depicted by dashed lines, and (c) a depiction of the (110) plane (shaded) of a primitive unit cell of TaS2.

109

Figure 4-2: Cross sections of (110) planes of a) 1T-TaS2, b) 2H-TaS2, c) 3R-TaS2, and d) 6R-TaS2. Anions and cations are depicted by large and small spheres, respectively. (according to Ref. 2)

Because the octahedral sites between every other S-S layer are empty, it is possible for tantalum atoms to partially fill these sites to form nonstoichiometric compounds of composition Ta1+xS2. Except for 1T-TaS2, nonstoichiometric compounds

1 have been found for all polymorphs of TaS2. Exact compositions can be difficult to

110

determine due to the difficulty in measuring small amounts of Ta atoms in these nonstoichiometric polymorphs. Jellinek estimated x to be around 0.2-0.35, 0.15 and 0.2 in

2 2H-, 3R- and 6R-Ta1+xS2, respectively. For any given TaS2 polymorph, increasing Ta occupancy of the interstitial sites results in expansion of the c-axis (longer stacking distance). For example, the values of the unit cell parameter c (stacking direction) for 3R-

TaS2, 3R-Ta1.08S2, and 3R-Ta1.15S2 are 17.9 Å, 18.01 Å (determined from single crystal data) and 18.2 Å, respectively.2,9

Similar to Ta1+xS2, other ions or molecules can insert between TaS2 layers to form intercalation complexes.10-12 For example, stearamide, a quite large molecule, can from intercalation compounds with 2H-TaS2, resulting in an increase of the TaS2 layer distance to 51 Å.10 Moreover, the intercalated compounds are superconductors if the host itself is a superconductor (such as 2H-TaS2). The critical temperatures of 2H-TaS2 intercalation compounds are between 2-5 K, a slight increase from the 0.8 K value of the host.10 Due to the low critical temperature, there is no actual application for TaS2 intercalation compounds as superconductors, but alkali metal intercalated TaS2 can be used as an

11,12 electrode material for alkali metal batteries. 2H-TaS2 is metallic with good

1 conductivity, and alkali metal intercalated TaS2 can exhibit both good conductivity and high alkali ion mobility, which makes it suitable for electrodes in alkali metal batteries.

Besides applications of intercalation compounds, layered TaS2 is also a suitable material for dry lubricants, because of the weak van der Waals interactions between sulfur layers.1

Nanostructured TaS2 may exhibit additional desirable properties. For example,

TaS2 nanowires exhibit slightly higher critical temperatures than the corresponding bulk

111

13 TaS2. Because of the good field emission current density of TaS2 nanobelts, they have potential applications as field emission devices, where 1D nanostructures are desirable.14

Two polymorphs of TaS3 in the orthorhombic and monoclinic crystal systems are

7,8 known, but there is no atomic level structure available for orthorhombic TaS3. In monoclinic TaS3, sulfur atoms form columns of stacked trigonal prisms along the b-axis, and tantalum atoms are located near the center of each trigonal prism and coordinated by six sulfur atoms (Figure 4-3). Adjacent columns interact through weak van der Waals forces. Hence, similar to TaS2, monoclinic TaS3 also exhibits a layered structure, and tantalum atoms are also 6-coordinated. Although the crystal structure of the orthorhombic

TaS3 phase remains unknown, by comparing the crystallographic data of the two TaS3 polymorphs, Meerschaut et. al.8 suggested that the orthorhombic structure could consist of similar trigonal prisms as the monoclinic structure. This prediction was based on the observation that the lattice parameter b of the monoclinic structure (3.341 Å) is nearly equal to the lattice parameter c of the orthorhombic structure (3.340 Å). As this is the distance between two adjacent tantalum atoms in the same column or the height of a trigonal prism, it is possible that similar trigonal prisms are formed along the c-axis in the orthorhombic structure.

112

Figure 4-3: Schematic of the stacking of the trigonal prism columns along the b-axis in TaS3. Sulfur atoms are at the corner and tantalum atoms are at the center of each trigonal prism.

Compared to TaS2 and TaS3, the structures of the subsulfides Ta6Sn (n=1, 3, 4) are more complicated. Ta3S2, Ta2S and Ta6S crystallize in orthorhombic, orthorhombic and monoclinic systems, respectively.4,5 In order to understand the structural connections between these subsulfide polymorphs, Kim et. al. proposed a hypothetic structure of

5 Ta6S5. In Ta6S5, Ta5 pentagons form columns of stacked pentagonal antiprisms (Figure

4-4). For each pentagonal antiprism, there is an extra tantalum atom at the center of the antiprism, and five sulfur atoms are positioned outside every other triangular face. These sulfur atoms connect with neighboring arrays. The structures of Ta3S2, Ta2S and Ta6S can be considered as missing one, two and four sulfur atoms outside each pentagonal antiprism. Kim et. al. calculated the band structures of Ta6Sn, and predicted increasingly

113

5 metallic properties from Ta6S5 to Ta3S2 and Ta2S. Although the structures of these subsulfides have been solved, their properties were not adequately studied.

Figure 4-4: Projection of a pentagonal antiprism formed by Ta atoms and the bonded sulfur atoms. Tantalum and sulfur atoms are indicated by filled and open circles, respectively.

4.1.2 Literature Routes

The synthesis of tantalum sulfides is usually carried out by direct reaction of the elements at high temperature, or by chemical vapor transport. All stoichiometric tantalum sulfide polymorphs were initially prepared through solid state routes.

TaS2 powders can be prepared by heating elemental tantalum and sulfur in a sealed quartz tube at 877 °C for a week and subsequently quenching in cold water.15

Then, single crystal 1T-TaS2 can be obtained by applying a temperature gradient from

847 °C to 947 °C to a sealed quartz tube containing a mixture of TaS2 powder and I2 for

15 about a week. 2H-TaS2 single crystals can be formed through iodine vapor transport by heating a sealed quartz tube containing a mixture of TaS2 powder and I2 in a two-zone

114

tube furnace with the low and high temperatures at 800 °C and 1000 °C , respectively, for

16 a week. 3R-Ta1+xS2 single crystals can be prepared through NH4Cl vapor transport in a sealed silica ampoule with low and high temperatures at about 700 °C to 800 °C and

9 1100 °C , respectively. Figueroa et. al. reported the synthesis of 6R-TaS2 by heating mixtures of tantalum powder and sulfur in a sealed quartz tube at 947 °C for 8 d, and then annealing at 647 °C for 4 d.17 It was also found that when tantalum foil was used as

17 starting material, 2H-TaS2 was obtained under the same synthetic conditions.

Similarly to TaS2, TaS3 could also be prepared by heating stoichiometric amounts of elemental Ta and S at high temperatures, or by recrystallization through vapor transport, but the products usually contained impurities such as TaS2. Kikkawa et. al. reported the synthesis of phase pure o-TaS3 by reactions between tantalum and sulfur powders at 700 °C or 900 °C under 1 to 3 GPa pressure in a cubic anvil press for 30

18 min. When the obtained o-TaS3 was heated at 700 °C under 2 GPa pressure for 30 min,

18 it converted to the monoclinic phase with a small amount of TaS2 impurity.

Ta2S and Ta6S were prepared by stoichiometric reactions of the elements in a

4 tungsten container at 1600 °C . Single crystals of Ta3S2 were a side product of the attempted synthesis of Ta9S6Ru2 by heating TaS2, Ta and Fe powders in a 3:6:2 ratio at

5 1000 °C with trace TaBr5 as transport reagent.

Besides high temperature solid state routes, tantalum sulfides have also been prepared by solution based low temperature routes, for which the as-recovered products were amorphous. Chianelli and Dines reported the synthesis of amorphous TaS2 through

19 reaction of TaCl5 and Li2S in ethyl acetate by stirring overnight. Carmalt et. al. reported the synthesis of amorphous tantalum sulfides by reactions of TaCl5 and

115

20 hexamethyldisilthiane in diethylether or toluene. Heat treatments at 800 °C under H2S atmosphere for the amorphous product obtained from reactions in diethylether resulted in crystallization of 1T-TaS2. However, microanalysis indicated that the sample contained significant amounts of silicon.

4.1.3 Preliminary Results

The Ta-S system was previously investigated by two group members, Christophe

Heinrich and Derek Mull. Their work is briefly summarized in the following paragraphs, and lists of previously prepared samples (CH and DM samples) are shown in Appendix B.

The synthesis with DTBS was mainly explored with TaCl5 as the starting material, especially under solvo-thermal conditions. During TG/DTA under N2, all the as- recovered DTBS samples showed a sharp weight loss up to 80% around 300 °C , indicating the presence of significant amounts of organic residues or tantalum halides. In addition, XRD patterns of several as-recovered DTBS samples, such as CHTa3 and

CHTa6, showed well crystallized unknown phases, which disappeared after heat treatments to 300 °C . This result indicated that the unknown crystalline phase was most likely organic instead of a tantalum sulfide. Although further elemental analysis methods, such as CHNS and EDS, were required to confirm the carbon and halide contents in these samples, it was clear that these samples contained large amounts of impurities, and that the completeness of reactions was low. Furthermore, no well crystallized tantalum sulfides were obtained from heat treatments on DTBS samples that were prepared with

TaCl5. The heat treatment for the sample prepared with TaI5 (CHTa18) at RT resulted in

TaO2I, and the reactions of DTBS with TaBr5 under solvo-thermal conditions were unexplored.

116

In contrast to the unsuccessful heat treatments of DTBS samples, crystalline 1T-,

3R- and 6R-TaS2 phases were obtained by heat treatments of as-recovered HMDST samples. However, many of these samples also contained tantalum oxide impurities, which indicated a leak in the experimental setup. The reactivity of the starting tantalum halides was very different. TaF3 was always obtained when TaF5 was used as a starting material. In addition, TaF3 could not be removed by heat treatments up to 600 °C . Hence,

TaF5 was not a good starting material. Generally, for syntheses with HMDST, TaBr5 and

TaCl5 appeared to be better starting materials than TaF5 and TaI5. However, further elemental analysis was required to assess the reaction completeness for these synthetic conditions.

In summary, room temperature syntheses and syntheses with DTBS under solvo-

21 thermal conditions for TaCl5 had been well explored. Crystalline TaS2 polymorphs had been obtained through heat treatments of these as-recovered samples. However, a few synthetic conditions, such as synthesis with HMDST under solvo-thermal conditions, were unexplored, and further elemental analysis was required to assess the reaction completeness. Hence, the goal of the present work was to explore the synthetic conditions not studied previously, and to gain a comprehensive understanding of this system.

4.2 Preparation of Tantalum Sulfides by Non- hydrolytic Sol-gel Synthesis

To prepare tantalum sulfides, 1 or 3 mmol of the tantalum pentahalides, TaCl5,

TaBr5 and TaI5, were reacted with two different sulfur sources, hexamethyldisilthiane

(HMDST) or di-tert-butylsulfide (DTBS), in nonreactive organic solvents like chloroform or acetonitrile. The reactions were carried out either at room temperature for 117

three days or by heating in a sealed ampoule for seven days. The exploration of various synthetic conditions included temperature (RT, 100 °C and 150 °C), sulfur to tantalum ratio (3 to 19), concentration of sulfur sources (0.2 to 2.1 M), and volume of solvent (5 or

15 mL). The initial 12 samples of the current work focused on the previously unexplored conditions, especially for synthesis with HMDST under solvo-thermo conditions. In addition, a few conditions for reactions with DTBS under solvo-thermal conditions and

HMDST at room temperature were also studied. The last ten samples of the current work focused on optimizing the synthetic parameters so that higher reaction completeness could be achieved. All samples prepared and synthetic conditions used are presented in

Table 4.2.

Tar-like precipitates were formed for most of the reactions of the initial 12 samples. For samples from XZ2001 to XZ2012 (except for XZ2006 and XZ2008), the sticky tar-like precipitates, which were stuck on the bottom of the ampoules and could not be filtered, were recovered as powders by decanting excess solvent and drying by evaporation under vacuum. For samples XZ2006 and XZ2008, large chunks of glass-like precipitates were recovered by filtration and powders were recovered by grinding these chunks. For sample XZ2010, an extremely sticky tar was formed at the bottom of the ampoule. After the ampoule was cut open inside the glovebox, the solution was discarded, and the tar-like precipitate was subjected to vacuum overnight. However, very little powder (88 mg) was recovered, and the major part of the sample was still very sticky tar.

This sample was considered unsuccessful, as it was impossible to recover enough powder for analysis.

118

Table 4.2: Synthetic parameters for tantalum sulfide samples.

Metal Source Sulfur Source Solvent Conditions Ratio Recovered Sample S:M mass (g) V T Time Name mmol Name mmol Name (mL) (°C) (d)

XZ2001 TaCl5 3.0 HMDST 9.5 CHCl3 15 3.2 100 7 1.062

XZ2002 TaCl5 3.0 HMDST 9.5 CH3CN 15 3.2 100 7 0.995

XZ2003 TaCl5 3.0 HMDST 19.0 CHCl3 15 6.2 100 7 0.683

XZ2004 TaCl5 3.3 HMDST 19.0 CH3CN 15 5.8 150 7 0.461

XZ2005 TaCl5 3.1 HMDST 9.5 CHCl3 15 3.0 150 7 0.731

XZ2006* TaCl5 3.0 HMDST 19.0 CHCl3 15 6.4 150 7 1.075

XZ2007 TaBr5 3.1 HMDST 9.5 CH3CN 15 3.0 150 7 0.779

XZ2008* TaBr5 3.1 HMDST 9.5 CHCl3 15 3.1 150 7 0.791

XZ2009 TaI5 3.0 HMDST 9.5 CHCl3 15 3.1 150 7 0.987

XZ2010 TaCl5 2.9 DTBS 8.9 CHCl3 15 3.0 150 7 0.088

XZ2011 TaBr5 2.9 DTBS 8.9 CHCl3 15 3.0 150 7 1.087

XZ2012 TaBr5 3.0 DTBS 8.9 CH3CN 15 3.0 150 7 1.431

XZ2013† TaCl5 1.0 HMDST 19.0 CHCl3 5 18.8 100 7 0.256

XZ2014 TaCl5 1.0 HMDST 19.0 CHCl3 15 18.8 100 7 0.224

XZ2015 TaCl5 1.0 HMDST 19.0 CH3CN 15 18.8 100 7 **

XZ2016 TaCl5 1.0 HMDST 3.1 CH3CN 15 3.0 RT 3 0.229

XZ2017 TaBr5 1.0 HMDST 19.0 CH3CN 15 18.9 100 7 0.278

XZ2018 TaBr5 1.0 HMDST 19.0 CH3CN 15 19.3 150 7 0.323

XZ2019 TaI5 1.0 HMDST 19.0 CH3CN 15 19.0 100 7 0.282

XZ2020 TaI5 1.0 HMDST 19.0 CH3CN 15 18.6 150 7 0.320

XZ2021† TaI5 1.0 HMDST 19.0 CH3CN 15 18.5 RT 3 0.367 XZ2022† TaCl5 1.1 HMDST 19.0 CH3CN 15 16.9 100 7 0.367 * large chunks of glass-like precipitate recovered by filtration ** failed to recover by filtration (filtrate discarded) † recovered by evaporation

For samples XZ2014, and XZ2016 to XZ2020, gel-like sediments were found at the bottom of the ampoules or flasks, and they could be suspended in solvents by agitation. The suspensions could be filtered and wet powders and very viscous gels were recovered on filter paper. Powders were obtained by further drying under vacuum. For samples XZ2013, XZ2021 and XZ2022, precipitates appeared to be powders with very small particle sizes, and therefore, they were recovered by evaporation of the solvents.

119

XZ2022 was a repeat of XZ2015, as XZ2015 could not be recovered by filtration, possibly because of the small particle sizes of its precipitate. Although the precipitate of

XZ2015 appeared to be recovered on filter paper, the actual amount was too low for analysis.

4.3 Results and Discussion 4.3.1 Phases and Elemental Analysis

No crystalline tantalum sulfide phase was obtained in any of the as-recovered samples. Heat treatments to 400 °C , 700 °C or 800 °C were carried out to induce crystallization (Table 4.3 and Table 4.4). The corresponding elemental analysis results of these samples are shown in Table 4.5. Elemental analysis for selected samples that were prepared previously are shown in Table 4.6. The EDS analysis of previous samples was carried out by Martin Kluenker. As-recovered samples from current work in this chapter are referred to as XZ# (i.e. XZ2001), and heat treated XZ samples are referred to as

XZ#.T*, where * indicates the temperature in centigrade to which the samples were heated. For example, the heat treatment of XZ2001 to 800 °C is referred as XZ2001.T800.

For some of these samples, the suffix, "_MK", indicates that the heat treatment was carried out by Martin Kluenker. Previous samples prepared by Christophe Heinrich and

Derek Mull are named with the prefixes "CHTa" and "DM", respectively.

120

Table 4.3: X-ray diffraction results for raw and heated tantalum sulfide samples prepared with 3 mmol of halide in 15 mL of solvent. Conditions Thermal Sample Halide Solvent S:Ta Phases T/°C t/d History Amorphous Raw Unknown phase XZ2001 TaCl CHCl 3 100 7 5 3 Poorly crystalline 3R-TaS 800 °C, 3 h 2 Ta2O5 Raw Amorphous Unknown phase XZ2002 TaCl5 CH3CN 3 100 7 Poorly crystalline 3R-TaS 800 °C, 15 h 2 Ta2O5 Raw Amorphous XZ2003 TaCl5 CHCl3 6 100 7 Unknown phase 800 °C, 12 h 3R-TaS2 Raw Amorphous XZ2004 TaCl5 CH3CN 6 150 7 Unknown phase 800 °C, 12 h Poorly crystalline 3R-TaS2 Raw Amorphous

XZ2005 TaCl5 CHCl3 3 150 7 Poorly crystalline 3R-TaS2 800 °C, 10 h Ta2O5 Amorphous Raw Unknown phase XZ2006 TaCl5 CHCl3 6 150 7 Poorly crystalline 3R-TaS2 800 °C, 10 h Ta2O5 Raw Amorphous XZ2007 TaBr5 CH3CN 3 150 7 800 °C, 10 h 3R-TaS2 Raw Amorphous

XZ2008 TaBr5 CHCl3 3 150 7 Poorly crystalline 3R-TaS2 800 °C, 10 h Ta2O5 Amorphous Raw Unknown phase XZ2009 TaI5 CHCl3 3 150 7 Poorly crystalline 3R-TaS2 800 °C, 10 h Ta2O5 Raw Amorphous Poorly crystalline phase XZ2011* TaBr5 CHCl3 3 150 7 800 °C, 10 h resembled 6R-TaS2 and

Ta2O5 Raw Unknown crystalline phase

XZ2012* TaBr5 CH3CN 3 150 7 Poorly crystalline phase 800 °C, 10 h resembled 6R-TaS 2 * DTBS sample

121

Table 4. 4: X-ray diffraction results for raw and heated tantalum sulfide samples prepared with 1 mmol of halide in 15 mL of solvent. Conditions Thermal Sample Halide Solvent S:Ta Phases T/°C t/d History Raw Amorphous 400 °C, 7 h Amorphous XZ2013* TaCl5 CHCl3 19 100 7 700 °C, 7 h Amorphous

800 °C, 8 h 3R-TaS2 Raw Amorphous

XZ2014 TaCl5 CHCl3 19 100 7 700 °C, 7 h Amorphous

800 °C, 8 h 3R-TaS2 Raw Amorphous

XZ2016 TaCl5 CH3CN 3 RT 3 Poorly crystalline 3R-TaS 800 °C, 8 h 2 Ta2O5 Raw Amorphous

XZ2017 TaBr5 CH3CN 19 100 7 700 °C, 5 h 1T-TaS2

800 °C, 8 h 3R-TaS2 Raw Amorphous

XZ2018 TaBr5 CH3CN 19 150 7 700 °C, 5 h 1T-TaS2

800 °C, 8 h 3R-TaS2 Raw Amorphous

XZ2019 TaI5 CH3CN 19 100 7 Poorly crystalline 3R-TaS2 800 °C, 8 h Ta2O5 Raw Amorphous

700 °C, 7 h Amorphous and Ta2O5 XZ2020 TaI5 CH3CN 19 150 7 Poorly crystalline 3R-TaS 800 °C, 8 h 2 Ta2O5 Raw Amorphous

XZ2021 TaI5 CH3CN 19 RT 3 Poorly crystalline 3R-TaS 800 °C, 8 h 2 Ta2O5 Raw Amorphous XZ2022 TaCl5 CH3CN 17 100 7 800 °C, 8 h Poorly crystalline 3R-TaS2 * 5 mL solvent

122

Table 4.5: Elemental analysis results for raw and heated tantalum sulfide samples. Elemental ratios of heated samples calculated from both a combination of TGA and CHNS data and EDS data are both provided.

Wt left TGA CHNS EDS TaSx∙yTaX5 S/Ta 800 °C Ta Wt% C Wt% S Wt% Ta At% S At% Cl At% Br At% I At% x y TGA+CHNS EDS XZ2001 36±3 7.1±1.4 12.2±1.2 17±5 27±3 56±7 4.6 1.9 XZ2001.T800* 21% 49±3 25.1±1.3 17.5±1.1 33±7 67±7 2.0±0.2 2.0±0.5 XZ2002 63±2 6.8±2.8 15.3±2.5 38±7 58±8 4±2 1.5 0.02 XZ2002.T800* 78% 79±3 1.9±2.0 11.9±1.7 55±9 45±9 0.9±0.1 0.8±0.2 XZ2003 49±3 7.2±0.9 29.2±0.8 24±8 72±9 4±8 3.1 0.04 XZ2003.T800 65% 71±5 5.3±0.9 20.4±0.8 35±5 65±5 1.6±0.1 1.9±0.3 XZ2004 55±7 10.4±1.5 25.2±1.3 35±6 63±6 2±1 1.8 0.01 XZ2004.T800 66% 70±4 5.9±1.6 20.4±1.4 37±2 63±2 1.6±0.2 1.7±0.1 XZ2005 38±6 11.0±1.0 12.4±0.9 18±6 34±2 49±5 4.2 1.2 XZ2005.T800* 17% 42.6±2.5 11.4±2.2 XZ2005.T800_MK* 42% 48±9 62±11 38±11 0.6±0.2 XZ2006 35±4 10.1±1.0 11.5±0.8 21±5 37±7 42±8 3.0 0.68 XZ2006.T800* 14% 26±5 47.8±2.3 12.2±2.0 2.7±0.7 XZ2007 54±3 6.9±1.5 24.1±1.3 26±9 44±8 31±16 2.2 0.32 XZ2007.T800 81% 77±4 4.6±1.1 22.6±0.9 37±2 63±2 1.6±0.1 1.7±0.1 XZ2008 37±3 10.2±1.9 11.7±1.7 26±5 44±5 28±4 2±2 2.2 0.30 XZ2008.T800* 17% 36.8±2.0 13.2±1.7 XZ2008.T800_MK* 51% 51±9 59±10 41±10 0.7±0.2 XZ2009 29±3 12.2±0.6 10.8±0.5 22±9 44±7 22±3 12±6 2.4 0.44 XZ2009.T800* 14% 49.4±7.4 12.1±6.5 XZ2009.T800_MK* 24% 29±7 XZ2011 28.6±0.9 6.8±0.8 33±7 35±4 30±5 2±2 1.3 0.25 XZ2011.T800* 51% 41±6 40.5±0.9 9.2±0.8 44±7 55±9 1.2±0.2 1.3±0.3 XZ2012 19.8±0.8 6.9±0.7 23±7 58±15 19±9 3.0 0.20 XZ2012.T800 25% 56±4 22.5±1.4 14.8±1.2 42±9 58±9 1.5±0.2 1.4±0.4 XZ2013 7.4±0.7 37.0±1.3 19±3 77±5 4±4 4.2 0.04 XZ2013.T800 71% 69±7 4.6±1.4 20.8±2.8 33±3 67±3 1.7±0.3 2.0±0.2 XZ2014 8.0±0.8 37.4±1.7 21±4 77±4 2±1 3.7 0.02 XZ2014.T800_II 68% 68±6 6.3±1.7 21.5±3.3 37±6 63±6 1.8±0.3 1.7±0.3 XZ2016 5.3±0.2 19.9±0.4 52±8 46±8 3±2 0.9 0.10 XZ2016.T800* 80% 74±1 0.8±0.3 14.4±0.7 52±5 48±5 1.1±0.1 0.9±0.1 XZ2017 9.8±1.0 24.4±1.9 37±3 58±3 5±1 1.6 0.03 XZ2017.T800 80% 68±6 2.4±0.8 21.3±1.6 37±3 63±3 1.8±0.2 1.7±0.2 XZ2018 9.3±1.2 21.8±2.4 41±3 55±1 4±2 1.4 0.02 XZ2018.T800 79% 69±3 5.8±1.4 21.2±2.8 35±7 65±7 1.7±0.2 1.9±0.4 XZ2019 5.5±0.6 13.1±1.1 43±5 50±3 7±2 1.2 0.03 XZ2019.T800* 76% 74±4 3.1±1.4 16.4±2.9 39±4 61±4 1.3±0.2 1.5±0.2 XZ2020 7.8±0.8 18.9±1.5 35±6 59±6 6±3 1.7 0.03 XZ2020.T800* 78% 71±4 5.8±0.8 20.1±1.5 37±7 63±7 1.6±0.2 1.7±0.3 XZ2021 4.2±0.7 14.1±1.5 32±5 53±5 15±2 1.8 0.10 XZ2021.T800* 44% 0.5±0.8 17.0±1.6 XZ2021.T800_MK* 55% 70±9 30±9 0.4±0.1 XZ2022 7.6±1.3 14.9±2.5 50±3 34±3 16±4 0.7 0.07 XZ2022.T800 80% 73±5 1.5±0.6 20.1±1.1 41±3 59±3 1.5±0.1 1.5±0.2 * Oxidized during heat treatment

123

Table 4.6: Elemental analysis results for selected CH and DM samples. Elemental ratios of heated samples were calculated from EDS data.

CHNS EDS TaSx∙yTaX5 S/Ta C Wt% S Wt% Ta At% S At% Cl At% Br At% I At% x y (EDS) CHTa1 12±8 19±2 68±7 N/A N/A CHTa4 31±6 40±6 29±2 1.6 0.23 CHTa5 60±4 6±2 34±3 0.1 0.13 CHTa7 8.1±0.9 21.2±1.7 28±7 66±5 5±2 2.5 0.04 CHTa7_I_T800 4.1±0.3 8.5±0.6 CHTa7_II_T700 40±4 60±4 1.5±0.2 CHTa7_II_T900 3.5±0.6 18.9±1.1 CHTa8 3.5±0.5 17.7±1.0 34±3 58±3 8±1 1.8 0.05 CHTa8_I_T700 0.3±0.4 16.4±0.7 CHTa8_III_T900 0.4±0.7 19.2±0.7 42±6 58±6 2±1 1.4±0.2 CHTa10 4.9±0.4 14.8±0.9 34±4 56±6 9±2 1.7 0.06 CHTa10_I_T800 1.0±0.7 20.5±1.4 CHTa10_II_T700 0.7±1.0 23.5±0.9 35±2 65±2 1.9±0.1 CHTa10_III_T700 1.0±2.5 21.4±2.2 37±3 63±5 1.7±0.2 CHTa12 2.8±0.8 5.5±1.6 50±12 29±16 22±9 0.6 0.10 CHTa12_I_T600 1.1±0.3 4.6±0.5 CHTa12_II_T800 1.0±0.5 14.0±1.1 49±2 51±2 1.1±0.1 CHTa14 6.6±0.8 8.6±1.7 55±5 29±6 17±2 0.6 0.06 CHTa14_I_T800 4.1±0.3 14.2±0.7 CHTa14_II_T800 4.3±0.8 16.6±1.6 43±3 57±3 1.3±0.1 CHTa15 4.0±0.5 17.7±0.9 34±9 54±9 12±2 1.7 0.08 CHTa15_T700 1.3±0.3 21.4±0.5 43±5 57±5 1.3±0.2 CHTa16_I_T700 5.6±0.5 19.9±1.1 42±6 58±6 1.4±0.2 CHTa16_II_T900 5.0±0.4 19.4±0.8 41±5 59±5 1.4±0.2 CHTa21 6.5±1.2 15.3±2.3 CHTa21_I_T900 0.4±0.6 14.6±1.1 CHTa21_II_T900 0.4±0.8 19.8±1.6 42±2 58±2 1.4±0.1 DM4 3.3±0.6 18.2±1.3 31±4 61±3 8±2 2.1 0.06 DM4.T400 1.2±0.7 16.8±1.4 39±9 59±10 1±2 1.5 0.01 DM13 2.1±0.8 19.0±1.6 DM13.T400 0.5±1.8 20.4±3.6

In Table 4.5, Equation 4-1 was used to calculate the sulfur to tantalum ratios from a combination of CHNS and TGA data:

r = (wt%S / MS) / (wt%Ta / MTa) (Equation 4-1) where wt%Ta and wt%S are the weight percentages of tantalum obtained by TGA

(Equation 2-2) and sulfur obtained by CHNS analysis, respectively, and MTa and MS are the molar masses for copper and sulfur, respectively. In Table 4.5 and Table 4.6,

Equation 4-2 was used to calculate the sulfur to tantalum ratios from EDS data:

r = at%S / at%Ta (Equation 4-2)

124

where at%S and at%Ta are the atomic percentages of sulfur and tantalum, respectively, obtained by EDS. The calculated elemental ratios only give the correct tantalum sulfide

formula for unoxidized samples. A sample with Ta2O5 may have a smaller S/Ta ratio.

Most of the as-recovered samples were amorphous except for a few samples,

XZ2001, XZ2006, XZ2009 and XZ2012, which contained unknown phases (Figure 4-5).

The crystalline features of these samples could not be matched to any existing tantalum sulfide polymorph. Except for XZ2012, these samples contained fairly large amounts of amorphous materials. These unknown phases were not retained in the heated samples, and the samples lost about 80% of their weight after heat treatments. It is possible that these unknown phases were mostly composed of organic residues. This was also supported by elemental analysis (Table 4.5), as these samples contained about 10% to 20% of carbon, but only about 10% of sulfur.

Figure 4-5: XRD patterns of as-recovered samples a) XZ2001, b) XZ2006 and c) XZ2012.

125

In order to induce crystallization, heat treatments were carried out for as- recovered samples. After heating to 800 °C , crystalline or poorly crystalline 3R-TaS2 was obtained for most samples, but Ta2O5 were also found in many of these samples. It was not surprising to observe oxidation after heat treatments, as similar results were obtained during previous studies. It was obvious that there was a leak in the experimental setup, and the samples contained materials that were more reactive to oxygen than the titanium scavengers at some temperature. However, using the same experimental setup, many samples were not oxidized, and the oxidation appeared to be random. This inconsistency of oxidation for the same experimental setup could indicate that the samples oxidized are substantially different from samples not oxidized. It is possible that Ta2O5 was formed by reactions between oxygen or moisture from small leak and materials that are more reactive towards these impurities than tantalum sulfides, such as tantalum halides.

By comparing the X-ray diffraction results with the results of elemental analysis, it was found that most of the oxidized samples lost significant amounts of weight (50% to

85 %) after heat treatments, except for XZ2002, XZ2016, XZ2019 and XZ2020 (20 to 25% of weight loss). According to the EDS results (an example is shown in Figure 4-6), such as-recovered samples with large weight losses after heat treatment also contained significant amounts of halides (about 30 to 50 atom%), in contrast to less than 10 atom% of halides in non-oxidized samples. It is possible that the Ta2O5 found in heat treated samples was formed by oxidation of tantalum halides. Although most as-recovered samples contained halides, no halide peaks were found in EDS spectra of samples heated to 800 °C . This indicates that the halides evaporated or were oxidized at high temperature.

Evaporation is possible, considering the low boiling point of tantalum halides (233 °C ,

126

345 °C , and 543 °C for TaCl5, TaBr5 and TaI5, respectively). Small amounts of tantalum oxide were found when TaCl5 and TaI5 were heated above their boiling temperatures under the same setup as a typical heat treatment. However, for a similar experiment for

TaBr5, no tantalum oxide was found in the residue, which was about 1% of the starting weight. These results indicated that TaBr5 is more stable towards oxygen than TaCl5 and

TaI5 below their respective boiling points. This explains why XZ2007 was not oxidized after heat treatment despite a Br content of about 30%. In addition, because the as- recovered samples could contain significant amounts of halides, which could evaporate or be oxidized during TGA runs in air, the tantalum content of raw samples calculated from

TGA could be very inaccurate.

x 0.001 cps/eV

250

200

150 Cl S Ta I Ta S Cl I Ta I Ta

100

50

0 1 2 3 4 5 6 7 8 9 keV

Figure 4-6: EDS spectrum of XZ2009 (TaI5+HMDST in CHCl3), indicating the presence of Ta, S, Cl and I.

127

Although no crystalline halide starting materials were observed by X-ray diffraction, the halides could crystallize and evaporate upon heating. It was also noticed that TaBr5 and TaI5 could react with chloroform. As a result, large amounts of chlorine were found in the samples prepared in chloroform with TaBr5 (XZ2008 and XZ2011) and

TaI5 (XZ2009, shown in Figure 4-6) as starting materials.

Besides the weight loss of halides and organic residues, sulfur loss was also observed. This is most obvious for samples in which the sulfur content decreased after heat treatment. If a weight loss is not accompanied by a sulfur loss, the relative sulfur content in this sample will increase after heat treatment. This was observed for some samples, however, most of them still lost sulfur, as their increased sulfur contents were not proportional to the overall weight loss. For example, the sulfur content in XZ2001 was about 12%, and it increased to about 18% after heat treatment with about 80% of overall weight loss. Thus, about 70% (∆S% = 18%×(100%-80%)/12%) of the sulfur originally present was lost after heat treatment.

All as-recovered samples contained about 5% to 30% of carbon. For samples with high weight losses, the carbon content could increase drastically as halides accounted for most of the weight loss. For example, the carbon content in XZ2006 was about 10%, and it increased to about 50% after heat treatment. Given that there was an 85% overall weight loss, only about 75% (remaining C% = 50%×(100%-85%)/10%) of the carbon still remained in the sample. Hence, despite the drastic increase in the carbon content, the absolute amount of carbon was actually reduced after heat treatment. This result was also confirmed by TGA, as XZ2006.T800 lost about 70% of its weight during a TGA run to

800 °C in air. This high weight loss was caused by oxidation of the carbon in this sample.

128

Given that the carbon residue was stable up to 800 °C in an inert atmosphere, the carbon in the heat treated samples was likely to be amorphous carbon or carbides instead of organic residues bonded to the metals. Upon heating, the organic residues in as-recovered samples may have decomposed to amorphous carbon, which remained in the sample as an impurity even after heat treatments.

Figure 4-7: TGA (solid line) and DTA (dashed line) curves in air for pre-heated sample XZ2006.T800. The sample was heated to 800 °C for 10 h before the TG/DTA run.

In general, the as-recovered samples were amorphous and could contain significant amounts of halides and organic residues. Crystalline tantalum sulfides could be obtained after heat treatments, although the heat treated samples may contain impurities. After heat treatments, the halide impurities could be removed, while the carbon from organic residues could still remain in the sample as amorphous carbon or

129

carbides. Depending on the as-recovered products, Ta2O5 could be formed after heat treatments. Hence, as-recovered samples with higher reaction completeness could result in less oxide impurities after heat treatments. In order to find desirable synthetic conditions that lead to high reaction completeness, the current and previous results are summarized and discussed in the next section.

4.3.2 Influence of Synthetic Conditions

4.3.2.1 Synthesis with DTBS

The results of both previous and current samples prepared with DTBS are summarized in Table 4.7. In general, syntheses with DTBS were not successful, as no crystalline tantalum sulfides were prepared for as-recovered and heat treated samples.

The reaction completeness was low, and samples contained about 20% to 70% of halides

(Table 4.5 and Table 4.6). Consequently, the heat treatments of these samples resulted in large amounts of tantalum oxide, except for XZ2012, the halide impurity of which was bromide. Heat treatments of DTBS samples prepared with TaBr5 (XZ2011 and XZ2012) and TaCl5 (CHTa6) at 150 °C resulted in similar poorly crystalline phases that resembled

6R-TaS2. However, it was impossible to confirm the phases in these samples were tantalum sulfide due to the poor crystallinity and large amounts of impurities (41% and

23% of carbon for XZ2011.T800 and XZ2012.T800, respectively). Hence, it was unlikely that crystalline tantalum sulfides could be obtained by using DTBS, and optimization for synthetic parameters was not attempted with DTBS.

130

Table 4.7: XRD results for previous and current samples prepared with DTBS. Solvent Conditions Thermal Sample Halides S:Ta Results Name V (mL) T/°C t/d History

CHTa5 TaCl5 CHCl3 50 6 RT 0.5 h 500 °C Ta2O5

CHTa13 TaCl5 CHCl3 75 3 RT 0.5 h 500 °C Major Ta2O5

CHTa20 TaCl5 CHCl3 75 3 RT 0.5 h 600 °C Major Ta2O5

CHTa1 TaCl5 CHCl3 15 3 100 7 300 °C amorphous

XZ2010 TaCl5 CHCl3 15 3 150 7 Raw 0.088 g recovered

CHTa17 TaCl5 CH3CN 45 3 RT 1 Raw Failed

CHTa3 TaCl5 CH3CN 15 3 100 7 Raw 0.103 g recovered

CHTa11 TaCl5 CH3CN 15 3 100 7 Raw 0.016 g recovered Poorly crystalline phase CHTa6 TaCl5 CH3CN 15 3 150 7 900 °C resembled 6R-TaS2 Poorly crystalline phase

XZ2011 TaBr5 CHCl3 15 3 150 7 800 °C, 10 h resembled 6R-TaS2 and

Ta2O5 Poorly crystalline phase XZ2012 TaBr5 CH3CN 15 3 150 7 800 °C, 10 h resembled 6R-TaS2 CHTa18 TaI5 CHCl3 75 3 RT 0.5 h 400 °C TaO2I

4.3.2.2 Reactions between HMDST and TaI5

The results of the samples prepared by reactions between HMDST and TaI5 are summarized in Table 4.8. Previously, no reaction was attempted in chloroform, as the solubility of TaI5 in chloroform was very low. In the current work, one sample (XZ2009) was prepared in chloroform under solvo-thermal conditions (150 °C ). The as-recovered sample contained large amounts of impurities (12 wt% carbon, 22 at% Cl and 12 at% I), indicating very low reaction completeness. In addition, because TaI5 reacts with chloroform, no further attempt was made for reactions in chloroform.

131

Table 4.8: XRD results for previous and current samples prepared by reactions between HMDST and TaI5. Solvent Conditions Thermal Sample S:Ta Phases Name V (mL) T/°C t/d History

Poorly crystalline 3R-TaS2 XZ2009 CHCl3 15 3 150 7 800 °C, 10 h Ta2O5

DM5 CH3CN 15 1 RT 3 Raw TaI5

CHTa19 CH3CN 75 3 RT 0.5 h 460 °C Ta2O5

400 °C 1T-TaS2

CHTa12* CH3CN 30 3 RT 1 700 °C 1T-TaS2 and Ta2O5

800 °C 3R-TaS2 and Ta2O5

Poorly crystalline 3R-TaS2 XZ2021† CH3CN 15 19 RT 3 800 °C, 8 h Ta2O5

Poorly crystalline 3R-TaS2 XZ2019† CH3CN 15 19 100 7 800 °C, 10 h Ta2O5

DM9 CH3CN 15 1 150 7 600 °C Amorphous Poorly crystalline phase Raw resembled 1T-TaS2 CHTa16 CH3CN 15 3 150 7 Poorly crystalline phase 900 °C, 7 h resembled 3R-TaS2

700 °C, 7 h Amorphous and Ta2O5

XZ2020† CH3CN 15 19 150 7 Poorly crystalline 3R-TaS2 800 °C, 8 h Ta2O5 * 2 mmol halide † 1 mmol halide

For samples prepared in acetonitrile, heat treatment to or above 800 °C resulted in partial oxidation, except for CHTa16. For the two samples (XZ2018 and XZ2019) prepared under solvo-thermal conditions in acetonitrile with high S/Ta ratios (about 19), only small amounts of halide impurities (less than 10 atom%) were measured by EDS.

The weight losses of these samples after heat treatments were only about 20%, which might indicate less impurity. It was unexpected that they were oxidized after heat treatments.

132

CHTa16 was the only sample without oxidation after heating to above 800 °C .

Interestingly, as-recovered CHTa16 consisted of a poorly crystalline phase that resembled 1T-TaS2 (Figure 4-8 a)). The phase started to resemble the 3R-TaS2 phase after heat treatment at 200 °C or higher (Figure 4-8 b) and c)). Because of the broad peaks in this sample, it was also studied by STEM. The results of the STEM analysis indicated that tantalum sulfide samples with broad peaks could be nanocrystalline with coherence lengths of about 5 nm (Figure 4-9).

Figure 4-8: XRD patterns of a) as-recovered, b) 200 °C heat treated and c) 900 °C heat treated CHTa16 (150 °C , 15 mL CH3CN, S/Ta = 3). Line positions for 1T-TaS2 (bottom) and 3R-TaS2 (top) are shown.

133

Figure 4-9: STEM images of CHTa16 (TaI5+HMDST in CH3CN at 150 °C for 7 d) at different magnifications.

4.3.2.3 Reactions between HMDST and TaBr5

The results of samples prepared with HMDST and TaBr5 are shown in Table 4.9.

Because TaBr5 is less reactive with oxygen than TaCl5 and TaI5, no oxidation was found in heated treated TaBr5 samples prepared in acetonitrile. Similar to TaI5, TaBr5 can also react with chloroform, resulting in chlorine impurities in as-recovered samples.

Subsequently, heat treatment of the sample prepared in chloroform (XZ2008) resulted in oxidation. Therefore, relatively pure tantalum sulfides could be obtained by using TaBr5 and acetonitrile after heat treatments, and acetonitrile is a better solvent than chloroform in this case.

The 800 °C heat treatments of as recovered XZ2007, XZ2017 and XZ2018 all resulted in crystallization of 3R-TaS2 (Figure 4-10). The sample prepared at 100 °C

(XZ2017) exhibited better crystallinity than those prepared at 150 °C . The heat treatments of two samples prepared at 150 °C with different sulfur to tantalum ratios (3.0 for XZ2007, and 19.3 for XZ2018) and different amounts of halide (3 mmol and 1 mmol for XZ2007 and XZ2018, respectively) resulted in similar crystallinity.

134

Table 4.9: XRD results for previous and current samples prepared by reactions between HMDST and TaBr5. Solvent Conditions Thermal Sample Halide S:Ta Phases Name V (mL) T/°C t/d History

Poorly crystalline 3R-TaS2 XZ2008 TaBr5 CHCl3 15 3 150 7 800 °C, 10 h Ta2O5

700 °C, 15 h 1T- and 6R-TaS2 CHTa10 TaBr5 CH3CN 45 3 RT 2 700 °C, 5 h 1T-TaS2

DM4 TaBr5 CH3CN 15 1 RT 3 400 °C Amorphous and 1T-TaS2

700 °C, 5 h 1T-TaS2 XZ2017* 1 mmol CH3CN 15 19 100 7 800 °C, 8 h 3R-TaS2

DM7 TaBr5 CH3CN 15 1 150 7 600 °C Amorphous

XZ2007 TaBr5 CH3CN 15 3 150 7 800 °C, 10 h 3R-TaS2

700 °C, 5 h 1T-TaS2 XZ2018* 1 mmol CH3CN 15 19 150 7 800 °C, 8 h 3R-TaS2 *1 mmol halide

Figure 4-10: XRD patterns after heat treatment to 800 °C for TaBr5/HMDST samples prepared for 7d with 15 mL of CH3CN: a) XZ2007 (S/Ta=3.0, 150 °C), b) XZ2017 (S/Ta=18.9, 100 °C) and c) XZ2018 (S/Ta=19.3, 150 °C). Lines indicate peak positions of 3R-TaS2.

135

In contrast to the 3R-TaS2 phase obtained at 800 °C , the 1T-TaS2 phase was obtained for heat treatments at 700 °C . The 1T-TaS2 phase was obtained for 700 °C heat treatments of two solvo-thermally prepared samples, XZ2017 and XZ2018, although they still contained significant amounts of amorphous materials (Figure 4-11). Heat treatments for a RT sample, CHTa10 (TaBr5+HMDST in CH3CN at RT), to 700 °C for 5 h also resulted in the 1T-TaS2 phase, but it started to transform into the 6R-TaS2 phase when heated longer, as the 15 h heat treatment resulted in crystallization of a mixture of 1T- and 6R-TaS2. This result demonstrated that the crystallization of different TaS2 polymorphs could be controlled. However, a similarly solvo-thermally prepared TaI5 sample (XZ2020) did not result in the 1T-TaS2 phase after heating to 700 °C . The different crystallization behaviors could be caused by different particle sizes and bonding environment in the amorphous materials.

Figure 4-11: XRD patterns after heat treatment to 700 °C for TaBr5/HMDST samples prepared for 7d with 15 mL of CH3CN: a) XZ2017 (S/Ta=18.9, 100 °C) and b) XZ2018 (S/Ta=19.3, 150 °C). Lines indicate peak positions of 1T-TaS2.

136

The morphologies of the 3R-TaS2 phases prepared by reactions between TaBr5 and HMDST in acetonitrile were also studied (Figure 4-12). The 800 °C heat treatment of

XZ2007 resulted in fused and severely agglomerated particles. As shown in Figure 4-12 b), the micron sized fused structures consisted of numerous nanoplatelets. This indicated that the larger agglomerates in this sample were formed by fusion of nanoplatelets from the as-recovered sample. XZ2017.T800 and XZ2018.T800 consisted of agglomerated nanoparticles in shapes of spheres and round disks with diameters around 300 nm. The agglomerates in these samples appeared to be much smaller than the fused structures in heat treated XZ2004. It appeared that samples prepared with higher S/Ta ratios resulted in smaller particle sizes.

Figure 4-12: SEM images of 800 °C heat treated HMDST samples prepared in 15 mL CH3CN: a) and b) XZ2007 (S/Ta=3.0, 150 °C), c) XZ2017 (S/Ta=18.9, 100 °C) and d) XZ2018 (S/Ta=19.3, 150 °C).

137

4.3.2.1 Reaction between HMDST and TaCl5

The results of samples prepared with HMDST and TaCl5 are summarized in Table

4.10. For samples prepared with TaCl5, it was found that by tuning the starting S/Ta ratio, temperature and solvent, high reaction completeness could be achieved. No crystalline

TaS2 without the presence of Ta2O5 was obtained by heating samples prepared at RT.

Under solvo-thermo conditions, samples prepared in chloroform at 150 °C resulted in large amounts of halides and organic residues (XZ2005 and XZ2006), and subsequently large amounts of amorphous carbon and Ta2O5 impurities for their heat treated samples.

Two similar samples, XZ2001 (S/Ta=3) and XZ2003 (S/Ta=6), prepared at 100 °C resulted in oxidation and well crystallized 3R-TaS2 phase (Figure 4-13), respectively.

Samples prepared with larger S/Ta ratios at 100 °C also resulted in higher reaction completeness. Hence, for reactions in chloroform, a temperature of 100 °C is preferred, and higher S/Ta ratios improve reaction completeness. For samples prepared in acetonitrile, heat treatments of those prepared at RT and 100 °C with S/Ta 3 resulted in oxidation, while heat treatments of a sample with S/Ta=3 (CHTa7) prepared at 150 °C resulted in relatively phase pure TaS2. In addition, EDS analysis showed that the samples prepared at 150 °C , XZ2004 and CHTa7, contained about 2 at% and 5 at% of chlorine, respectively, indicating higher reaction completeness. When the S/Ta ratio was increased to 17 (XZ2022), a relatively pure 3R-TaS2 phase was obtained after heating to 800 °C , although the as-recovered sample contained about 16 at% of chlorine. Hence, reactions in acetonitrile at 150 °C could improve reaction completeness.

138

Table 4.10: XRD results for previous and current samples prepared by reactions between HMDST and TaCl5. Solvent Conditions Thermal Sample S:Ta Phases Name V (mL) T/°C t/d History

500 °C 1T-TaS2

700 °C 1T-TaS2 and Ta2O5 CHTa14 CHCl3 75 3 RT 0.5 h Poorly crystalline 3R-TaS2 800 °C Ta2O5

Poorly crystalline 3R-TaS2 XZ2001 CHCl3 15 3 100 7 800 °C, 3 h Ta2O5

XZ2003 CHCl3 15 6 100 7 800 °C, 12 h 3R-TaS2 700 °C, 7 h Amorphous XZ2014* CHCl3 15 19 100 7 800 °C, 8 h 3R-TaS2 400 °C, 7 h Amorphous

XZ2013* CHCl3 5 19 100 7 700 °C, 7 h Amorphous

800 °C, 8 h 3R-TaS2

Poorly crystalline 3R-TaS2 XZ2005 CHCl3 15 3 150 7 800 °C, 10 h Ta2O5

Poorly crystalline 3R-TaS2 XZ2006 CHCl3 15 6 150 7 800 °C, 10 h Ta2O5

Poorly crystalline 3R-TaS2 600-800 °C CHTa8 CH3CN 15 3 RT 0.3 h Ta2O5

900 °C 3R-TaS2 and Ta2O5

CHTa15 CH3CN 45 2 RT 1 500-700 °C 1T-TaS2 and Ta2O5 700 °C Mostly amorphous CHTa21 CH3CN 45 3 RT 1 900 °C 3R-TaS2 and Ta2O5

Poorly crystalline 3R-TaS2 XZ2016* CH3CN 15 3 RT 3 800 °C, 8 h Ta2O5

Poorly crystalline 3R-TaS2 XZ2002 CH3CN 15 3 100 7 800 °C, 15 h Ta2O5

XZ2022* CH3CN 15 17 100 7 800 °C, 8 h Poorly crystalline 3R-TaS2

800 °C Poorly crystalline 3R-TaS2 CHTa7 CH3CN 15 3 150 7 900 °C 3R-TaS2 XZ2004 CH3CN 15 6 150 7 800 °C, 12 h Poorly crystalline 3R-TaS2 * 5 mL solvent ** previously heated at 700 °C for 5 h and 800 °C for 3 h. 139

Figure 4-13: XRD patterns after heating to 800 °C for TaCl5/HMDST samples prepared at 100 °C for 7 d with a) 15 mL of CHCl3, S/Ta=6.2 (XZ2003), b) 5 mL CHCl3, S/Ta=18.8 (XZ2013) and c) 15 mL CHCl3, S/Ta=18.8 (XZ2014). Line positions for 3R-TaS2 are shown, and × indicates voltage spikes.

For successful samples, crystallinity was influenced by reaction conditions. The crystallinity of XZ2003.T800 was clearly better than the crystallinity of XZ2013.T800 and XZ2014.T800. It is possible that XZ2003 was better crystallized due to longer heat treatment (12 h instead of 8 h). Alternatively, the broader peaks in XZ2013.T800 and

XZ2014.T800 may also be caused by nanocrystallinity (Figure 4-9). The peak broadening could be caused by shorter coherence lengths. The more crystalline sample XZ2003.T800 contained both fused micron sized hexagonal plate-like particles and fused spherical nanoparticles. The particles in XZ2014.T800 were similar to XZ2003.T800, but with smaller sizes. For XZ2013.T800, the particles were smaller and much less agglomerated

140

and fused than for XZ2003.T800 and XZ2014.T800. XZ2013.T800 consisted of distorted spherical nanoparticles with diameters around 100 to 300 nm and some nanoplatelets.

Among all three samples, particles in heat treated XZ2013 were the smallest, and the

XRD peaks for this sample were the broadest. This suggests that the apparent lower crystallinity was caused by small crystallite sizes.

Figure 4-14: SEM images of 800 °C heat treated HMDST samples of a) and b) XZ2003 (15 mL CHCl3, S/Ta=6.2), c) and d) XZ2013 (5 mL CHCl3, S/Ta=18.8) and e) and f) XZ2014 (15 mL CHCl3, S/Ta=18.8).

141

Heat treatments of HMDST samples prepared in acetonitrile (XZ2022 and

XZ2004) to 800 °C resulted in poorly crystallized phases that resembled 3R-TaS2 with very broad peaks (Figure 4-15). CHTa7 from previous studies also resulted in poorly crystallized 3R-TaS2 after heating to 700 °C and 800 °C , but a well defined 3R-TaS2 phase was obtained after it was heated to 900 °C . This suggests that samples prepared in acetonitrile contained much smaller crystallites than the samples prepared in chloroform, which was confirmed by SEM (Figure 4-16). It can be clearly seen that the agglomerates in XZ2022.T800 (Figure 4-16 b)) consisted of agglomerated spherical nanoparticles with diameters below 50 nm.

Figure 4-15: XRD patterns after heating to 800 °C for TaCl5/HMDST samples prepared with 15 mL of CH3CN for 7 d at a) 150 °C (S/Ta=5.8, XZ2004) and b) at 100 °C (S/Ta=16.9, XZ2022). Line positions for 3R-TaS2 are shown.

142

Figure 4-16: SEM images of 800 °C heat treated HMDST samples prepared in 15 mL CH3CN of a) XZ2004 (S/Ta=5.8, 150 °C ) and b) XZ2022 (S/Ta=16.9, 100 °C ). 4.4 Conclusions

No crystalline TaS2 phases were obtained by heat treatments of samples prepared with DTBS. The reaction completeness of such samples was very low and they usually contained large amounts of impurities. For the synthesis with HMDST, reaction completeness could be tuned by synthetic parameters. Although no crystalline TaS2 was obtained from as-recovered samples, 1T-TaS2 and 3R-TaS2 could be obtained by 700 °C and 800 °C heat treatment of amorphous tantalum sulfides, respectively. In order to obtain relatively pure TaS2, as-recovered amorphous samples with higher reaction completeness are desired. In general, high S/Ta ratios improve the reaction completeness.

For samples prepared with HMDST and TaCl5, reactions at 100 °C in chloroform and reactions at 150 °C in acetonitrile showed higher completeness. Because TaBr5 and TaI5 can react with chloroform, acetonitrile was a better solvent for these halides. Although as-recovered samples prepared with HMDST and TaBr5 in acetonitrile could contain large amounts of unreacted TaBr5, relatively pure TaS2 still could be obtained for their heat treatments because TaBr5 is less sensitive to oxygen.

143

Chapter 5

5. Study of B1 (NaCl-type) to B2 (CsCl- type) Pressure-induced Structural Phase Transition in BaS, BaSe and BaTe Using ab initio Computations

5.1 Introduction

Possible mechanisms of phase transition of binary compounds from the 6- coordinated B1 structure (NaCl-type, Fm m) to the 8-coordinated B2 structure (CsCl- type, Pm m) have been studied both experimentally and theoretically over the past decades.1-8 Two transition mechanisms are generally discussed in this type of phase transition. Buerger proposed the first mechanism, in which the angle of a rhombohedral primitive cell with the B1 structure increases from 60° to 90° and results in the B2 structure.1 Watanabe, Tokonami and Morimoto2 discussed a second possibility, referred to as the WTM mechanism, based on their X-ray studies and optical observations on the phase transition of CsCl. In this mechanism, the phase transition from the B1 to the B2 structure involves a translational displacement between two adjacent layers of atoms.

Stoke and Hatch3 studied all the common subgroups of the B1 and B2 structures for more possibilities. They suggested that a third mechanism, the P21/m mechanism, should be

144

considered when an intermediate phase is observed on the transition pathway, though the

Buerger and WTM mechanism are more energetically favorable. Toledano et al.9 suggested another mechanism, where the B1/intermediate and intermediate/B2 transitions are described in space groups Pnma and Pbcm, respectively. Both the P21/m and Toledano pathways provide good explanations for the B1 to B2 phase transitions with intermediate states.9,10 While several theoretical investigations on the transition mechanisms have been published, only a few papers have characterized the transition mechanisms by modeling their energy hypersurfaces.5,6

The barium chalcogenides BaX (X = S, Se and Te), henceforth referred to as

BaX when all 3 compounds are involved, are potential materials for light-emitting diodes and laser diodes.11 They exhibit reversible pressure-induced phase transitions from the B1 to the B2 phase at moderate pressures (4.8 GPa to 6.5 GPa).12-14 No intermediate phase has been observed during the transition. A phase space region of BaS where the two phases coexisted above or below the transition pressure during the pressure increase or release, respectively, was reported by Yamaoka et al.12 With further application of pressure, the B2 phases of BaX are stable until their metallization pressures of about 80,

52 and 20 GPa, for X = S, Se and Te, respectively.14-16

In this project, we present a density functional theory (DFT) based ab initio computational study on the Buerger and WTM mechanisms for BaX. There has been no experimental observation of any intermediate phase for the B1 to B2 phase transition of

BaX. Hence, the P21/m and Toledano pathways are not studied in our work. Many theoretical investigations on the pressure-induced phase transitions of barium chalcogenides have focused on obtaining the transition pressure, band structure, density

145

of states and elastic properties.11, 17-21 The study of the energy hypersurfaces and resulting pathways for the transition with the two competing mechanisms has been left unexplored.

In this work, we report such a study of the energy hypersurfaces of BaX at pressures above, at, and below the transition pressure. Transition paths and energy barriers of the two different mechanisms are compared. Structural changes along the transition path are described. Our results are compared with earlier experimental and theoretical work. A. M.

Pendas et al.5 mentioned that it could be difficult to study the transition dynamics and kinetics experimentally due to impurities, defects, thermal and mechanical history and experimental parameters. In addition, nucleation processes can have profound influence on the pressure-induced B1 to B2 phase transition, and the transition may not be purely homogeneous.22, 23 However, by using idealized theoretical models for both the Buerger and WTM mechanisms, the microscopic lattice dynamics of the two different transition mechanisms can be predicted and compared. Therefore, our work can further elucidate the transition mechanisms from the B1 to the B2 structure. Section 5.2 describes the

Buerger and WTM mechanisms for the B1 to B2 structural transition; section 5.3 describes our computational method; section 5.4 describes the modeling approach; section 5.5 describes and discusses our results and section 5.6 gives the conclusions.

5.2 Transition mechanisms

The Buerger mechanism1 (Mechanism I) and WTM mechanism 2 (Mechanism

II), shown in Figure 5-1, can be characterized in one of the two common subgroups of the

B1 and B2 structures, R m and Pmmn, respectively.6

146

Figure 5-1: The transition from the B1 to the B2 structure depicted in (a) rhombohedral unit cell (used for Mechanism I) and (b) orthorhombic unit cell (used for Mechanism II).

147

As shown in Figure 5-1 a), Mechanism I is the Buerger mechanism, in which the rhombohedral cell with a space group of R m is compressed along its diagonal and results in a transformation to the B2 structure. The variables in the transition for

Mechanism I are the angle (θ) between two cell edges and lattice constant a of the unit cell. Mechanism II is the WTM mechanism (Figure 5-1 b)), in which atoms in the (002) plane of the orthorhombic unit cell, with a space group of Pmmn translate by half a lattice vector ( 1/2) causing a transformation to the B2 structure. The variables in this transition are the lattice constants a, b and c, which are the magnitude of lattice vectors 1, 2, and

3, respectively and the internal parameter x, which is the ratio between the distance that the atoms in (002) plane have moved and the magnitude of the lattice vector 1. The limiting values of x are 0 and 0.5 for B1 and B2 structures while those of θ are 60° and

90°, respectively.

For BaX in Mechanism I, the B1 and B2 structures can be represented by a two- atom rhombohedral primitive cell, with the Ba and X atoms on the corners and centers of the unit cell, respectively. The cell is defined by the magnitude 'a' of each lattice vector and the angle θ between any two of them. Thus a and θ are the only two variables in

Mechanism I. When θ = 60°, the two-atom rhombohedral cell when extended through all space becomes isomorphic with an eight-atom face-centered cubic cell, which is the B1 structure. When θ = 90°, the rhombohedral cell becomes a primitive cubic cell, which is the B2 structure. During the pressure-induced phase transition from the B1 to the B2 structure, the rhombohedral unit cell is compressed along its body diagonal, which results in a decrease of the unit cell lattice constant a, accompanied with an opening of the angle

θ from 60° to 90°.

148

In Mechanism II, the B1 and B2 structures can be represented by a four-atom orthorhombic unit cell with three orthogonal lattice vectors ( 1, 2, and 3), and an internal parameter 'x'. For the B1 structure, the positions for the two Ba atoms are the corners and the body center of the unit cell. Two X atoms are located on two opposing face centers, which are perpendicular to the vector 3. Four additional X atoms lie on centers of the four edges parallel to the vector 3. During the pressure-induced phase transition from the B1 to the B2 structure, atoms on the (002) plane of BaX slide along the [100] direction. The resulting relative displacement is called the internal parameter x.

It is defined as the ratio of the magnitudes of the displacement of the moving atoms and that of the lattice vector 1. When the transition is complete, the Ba atoms that occupies the body centers of the B1 structure are located on the face centers of the (100) planes of the B2 structure, while the X atoms are moved from their positions in the B1 structure to the face centered sites of the (010) planes. Therefore, x increases from 0 to 0.5 as the transition goes from the B1 to the B2 phase. The phase transition of Mechanism II is characterized by the lattice constants a, b and c, which are the magnitudes of the lattice vectors 1, 2, and 3, respectively, and the internal parameter x. Because of geometric constraints, the relations between a, b and c are a = b = c and a = b = c for the

B1 and B2 structures, respectively.

5.3 Computational method

All the ab initio calculations in this work are performed by using the Vienna Ab- initio Simulation Package (VASP)24-27 codes within the local density approximation

(LDA) to density functional theory (DFT).28,29 The electron-ion interactions are treated

149

by ultrasoft-Vanderbilt type pseudo potentials30 as supplied by Kresse et al.31 Three cutoff energies, 210 eV, 170 eV and 140 eV, are applied to the valance electronic wave functions expanded in a plane-wave basis set for BaS, BaSe and BaTe, respectively. A

Monkhorst-Pack32 generated 6×6×6 k-point grid was used for the Brillouin-zone integrations in all calculations with both Mechanism I and Mechanism II. Tests using cutoff energies and k-points mentioned above reached a convergence better than 1 meV.

The minimum energy for each configuration with the B1 or the B2 structure at a certain pressure was obtained by fully relaxing all atoms and lattice constants until a force convergence to less than 0.01 eV/Å was achieved. In order to fully explore the lattice space of the four dimensional orthorhombic unit cell in the WTM mechanism, more than

200,000 structure configurations were constructed.

4. Modeling approach

The potential energy surface (PES)5,33 is approached by obtaining enthalpy as a function of structural parameters, which can be described as HI=HI(θ,a) and

HII=HII(x,a,b,c) for Mechanism I and Mechanism II, respectively. Because H=U+PV, where U is internal energy, P is pressure and V is volume, each value of the enthalpy in these energy hypersurfaces was obtained by combining the computed value of the total energy (U) of each structure with the product of the pressure and volume of the corresponding unit cell.

Since the energy hypersurfaces of Mechanism I are two dimensional (2D), the phase transition can be completely characterized by the two variables θ and a, which determine the size and shape of the rhombohedral primitive cell. For Mechanism I, each

PES of a fixed pressure was computed for 13 uniformly distributed values of θ between

150

57° and 93°. For each value of θ, 12 to 14 values for the variable a were chosen from about 0.1 Å below the lattice constant a of the B2 structure to about 0.1 Å above the lattice constant a of the B1 structure. The three lattice vectors of Mechanism I are defined by:

i j = (m+ ) dm (i, j= 1, 2 and 3)

where j are unit vectors along the Cartesian axes, is the Kronecker delta

2 -1/2 function, m ( – 1)/3, dm (3m +2m+1) , and the corresponding position vectors for each atom are defined by:

= (0, 0, 0),

= (0.5, 0.5, 0.5),

in relative coordinates. The energy hypersurfaces in Mechanism II are four dimensional (4D), which would involve building a four dimensional grid. This would require prohibitive computational time. Therefore, instead of 4D energy hypersurfaces, we constructed two dimensional PES with x and c as independent variables and a and b as the dependent ones. For each fixed x and c, the enthalpy is minimized by finding the minimum value of a computed 2D enthalpy contour of 11×11 mesh grids with a and b as variables. Each PES of a fixed pressure was computed for 11 uniformly distributed values of x between 0 and 0.5. For each value of x, 12 to 14 values for the variable c were chosen from about 0.1 Å below the lattice constant c of the B2 structure to about 0.1 Å above the lattice constant a of the B1 structure. The three lattice vectors of Mechanism II are 1 = (a, 0, 0), 2 = (0, b, 0), and 3 = (0, 0, c), and the corresponding position vectors for each atom in relative coordinates are:

151

= (0, 0, 0),

= (0.5-x, 0.5, 0.5),

= (0.5, 0.5, 0),

= (1-x, 0, 0.5).

The transition path from the B1 structure to the B2 structure is defined by connecting the points on the energy hyper-surface that minimize the enthalpy. Starting at the initial B1 structure, the transition path connects points that result in the minimum increase of enthalpy up to the point that enthalpy starts to decrease. At this point, the transition path connects points that result in the maximum decrease of enthalpy to the final B2 structure. The transition energy barriers are obtained by finding the maximum enthalpy elevation on the transition path from the B1 to the B2 structure. In order to obtain accurate values of transition energy barriers, hypersurfaces of the surroundings of the saddle points with denser mesh grids are obtained. A convergence of 1 meV is achieved for each consecutive step on the transition path in the high resolution region.

5.5 Results and discussion 5.5.1 Transition pressure and structural parameters

At the transition pressure (Pt) of BaX (X = S, Se and Te), the enthalpy of the

B1 structure is equal to that of the B2 structure. Therefore, the transition pressure can be calculated by solving the roots of the fitted linear equations of enthalpy as a function of pressure of the B1 and B2 structures. Figure 5-2 shows the plots of enthalpy (eV/pair) vs. pressure (GPa) for all BaX, and the corresponding energy (eV/pair) vs. volume plots are shown in Figure 5-3 for comparison. For each material, linear regression fits to the data

152

points of the respective enthalpy-pressure lines of the B1 and B2 phases give the coefficient of determination R2 values of 0.999. For BaS, BaSe and BaTe the computed transition pressures are 5.47 GPa, 4.87 GPa and 3.42 GPa, respectively.

153

Figure 5-2: Plots of enthalpy (eV/pair) of one pair of atoms as a function of pressure (GPa) of (a) BaS, (b) BaSe, and (c) BaTe. Circles and triangles indicate computed data points of enthalpy-pressure line of B1 and B2 structures respectively.

154

Figure 5-3: Plots of cohesive energy (eV/pair) of one pair of atoms as a function of volume (Å3/pair) of (a) BaS, (b) BaSe, and (c) BaTe for B1 and B2 structures.

The values of the calculated Pt, as well as other structural parameters, are compared with experimental and theoretical data in Table 5.1. Lattice constants, a (Å), at zero and transition pressures for the B1 and B2 structures of BaS, BaSe and BaTe, the transition pressure (Pt) and percentage change in volume (△V/VB1) as the phase changes from B1 to B2 and the zero pressure bulk modulus B0 for the B1 and B2 structures are

155

given in Table 5.1. Results from earlier work are given for comparison where available, with experimental data in parentheses and theoretical data in square brackets. The agreement between our results and other experimental and theoretical results is reasonable considering differences in different computational methods as well as experimental conditions used.

Table 5.1: Results for transition pressures and structural parameters from both present work and earlier work with experimental data in parentheses and theoretical data in square brackets, respectively.

We found that the B1 phase of all BaX compounds exhibits lower cohesive energy (U) than the corresponding B2 phase at zero pressure and therefore results in lower enthalpy, which explains why the B1 structure is the thermodynamically stable phase in nature at low temperature and pressure.12-14 With increasing pressure, the cohesive energy of the B1 phase remains lower than that of the B2 phase. However, the

difference in enthalpy ( - ), between the enthalpies of the B1 and B2 phases becomes smaller. This is because the B1 structure exhibits a much larger volume than the

B2 structure at the same pressure, causing the PV term to become important. As an

156

example, the specific volumes of the B1 and B2 phases at the respective transition pressures are 57.8 and 49.4 Å3/pair for BaS, 63.3 and 54.6 Å3/pair for BaSe, and 76.4 and

3 66.4 Å /pair for BaTe. When the pressure increases so that P Pt, the B1 phase is no longer energetically favorable, and the phase transition to the B2 phase occurs, which results in a significant reduction in volume and increase in coordination number. Our calculations are consistent with experimental observations.12-14

5.5.2 Potential energy surface and energy barriers

The potential energy surfaces (PES) of BaS at 0, 5.5 and 8.0 GPa, BaSe at 0, 4.9 and 8.0 GPa, and BaTe at 0, 3.4 and 5.5 GPa for both mechanisms were computed. The

PES at different pressures for a given material with a fixed mechanism is very similar.

Hence only the PES plots at the transition pressure are shown in Figure 5-4 and Figure

5-5, for Mechanism I and Mechanism II, respectively. The scale bar indicates the enthalpy difference between a specific structure and the B1 structure of the material at the corresponding pressure. Two minima corresponding to the B1 and B2 structures, whose enthalpies are equal at the transition pressure, are found on each PES plot. The dotted line on each PES plot indicates the computed transition pathway, whose cubic order polynomial fit is given in Table 5.2, from the B1 to the B2 structure. The equations used

for fitting are: a (Å) = θ for Mechanism I, while a (Å) = , b (Å) =

and c (Å) = for Mechanism II. The corresponding coefficient of determination (R2) fits giving the quality of the fit is shown in the last column. Values close to 1 in the last column signify the goodness of the fit. The transition paths corresponding to these best fits are shown by the dashed lines in Figure 5-4 and Figure

5-5.

157

158

Figure 5-4: Contour plot of the computed enthalpy as a function of θ and a of (a) BaS at 5.5 GPa, (b) BaSe at 4.9 GPa and (c) BaTe at 3.4 GPa, for Mechanism I. The scale bars indicate the enthalpy in eV.

159

Figure 5-5: Contour plot of the computed enthalpy as a function of x and c of (a) BaS at 5.5 GPa, (b) BaSe at 4.9 GPa and (c) BaTe at 3.4 GPa, for Mechanism II. The scale bars indicate the enthalpy in eV.

160

Table 5.2: Coefficients (Cm) of the cubic polynomial fits to the computed transition paths for the lattice constant a as a function of angle θ in Mechanism I, and lattice constants a, b and c as a function of the internal parameter x in Mechanism II of BaS, BaSe and BaTe.

2 Material Mechanism Parameter C3 C2 C1 C0 R -1.2262 3.1849 -2.9061 1.2939 0.994 I a ×10-5 ×10-3 ×10-1 ×101 BaS a 11.089 -7.2777 -0.50539 4.3506 0.994 II b -5.6161 3.7126 1.2326 4.2948 0.982 c -11.228 6.7915 -2.4688 6.1496 0.991 4.2023 -6.9674 9.8855 5.4641 0.995 I a ×10-6 ×10-4 ×10-3 ×100 BaSe a 13.961 -9.7152 -0.018261 4.4900 0.997 II b -4.1537 3.8111 0.88618 4.4414 0.986 c -26.608 17.761 -4.0371 6.3401 0.957 -2.5658 8.7250 -1.1069 8.7929 0.994 I a ×10-6 ×10-4 ×10-1 ×102 BaTe a 17.022 -11.731 0.18954 4.7705 0.997 II b -6.5722 5.1451 0.99374 4.7318 0.989

c -21.201 13.439 -3.3871 6.7444 0.987

For Mechanism I, there are only two variables, the angle θ and the lattice constant a, represented by the x and y axes on the contours, respectively. From Figure 5-4, only one transition pathway, whose shape is very close to the diagonal on the PES, is found possible for each material. These transition pathways suggest that both the angle and lattice constant change proportionately and simultaneously during the phase transition in mechanism I.

For mechanism II, the orthorhombic unit cell is determined by four variables, the three lattice constants (a, b and c) and the internal parameter x. The change of the three lattice constants (a, b and c) is not completely independent. It is found that an increase of the internal parameter x is always accompanied with decreases of the lattice constants a

161

and c, and an increase of the lattice constant b. The two structural parameters x and c, which are used to visualize the PES, are represented by the abscissa and ordinate, respectively. On the transition pathway from the B1 to the B2 structure, the change of the internal parameter x is generally faster than the change of the lattice constant c before the saddle point, while the change of the lattice constant c is generally faster after the saddle point. Compared with Mechanism I, the PES plots of Mechanism II are smoother even though there is a steep and high energy elevation on the left bottom corner of each PES.

No other metastable phase, i.e. minimum other than the B1 and B2 phases, is found on any of the PES plots. This indicates that there is no intermediate phase in this phase transition for both types of mechanisms, which is consistent with experimental observations for BaS.12-14

The transition energy barriers of BaS at 0, 5.5 and 8.0 GPa, BaSe at 0, 4.9 and 8.0

GPa, and BaTe at 0, 3.4 and 5.5 GPa for both mechanisms are shown in Table 5.3. From the corresponding PES plots, it is found that the transition state, whose structure exhibits the maximum enthalpy on the transition path, occurs early on the transition path for both

Mechanism I and Mechanism II. Moreover, the energy barrier decreases for both mechanisms as the pressure increases, which suggests that the kinetics of the transition will proceed faster at higher pressure, consistent with experimental observations for

BaS.10 The energy barriers for a given compound at the same pressure for Mechanism I are higher, by about 0.02 to 0.05 eV/pair (2 to 5 kJ/mol) than those for Mechanism II.

These small differences in energy per pair suggest that Mechanism II is marginally favored over Mechanism I during the phase transition from B1 to B2 for all BaX.

162

Table 5.3: Energy barriers in eV/pair and kJ/mol (in parentheses) for the transition from B1 to B2 phase for BaS, BaSe and BaTe at three different pressures with Mechanism I and Mechanism II. Pressure Energy barrier (GPa) [eV/pair (kJ/mol)] Mechanism Mechanism I II Material 0 0.35 (33.8) 0.32 (30.9) BaS 5.5 0.18 (17.4) 0.13 (12.5) 8 0.12 (11.6) 0.08 (7.7) 0 0.32 (30.9) 0.30 (28.9) BaSe 4.9 0.16 (15.4) 0.13 (12.5) 8 0.09 (8.7) 0.07 (6.8) 0 0.28 (27.0) 0.25 (24.1) BaTe 3.4 0.15 (14.5) 0.12 (11.6) 5.5 0.10 (9.6) 0.08 (7.7)

5.5.3 Symmetry and coordination

The phase transition leads to an 8-coordianted B2 structure from a 6- coordinated B1 structure. Figure 5-6 gives a comparison between the coordination environments of sulfur atoms of BaS in the transition state at the transition pressure for both Mechanism I and Mechanism II. The four nearest neighbors on the (110) plane for

Mechanism I and the (002) plane for Mechanism II form a rectangle, which is retained during the entire transition path. The sulfur atoms are located in the center of these rectangles. In Mechanism I, the nearest and second nearest neighbors on the (1 0) plane enclose a parallelogram with the sulfur atom on its inversion center and the nearest neighbors on its shorter diagonal. In Mechanism II, the nearest and second nearest neighbors on the (020) plane enclose a rectangle with the (020) plane as its mirror plane.

163

For both mechanisms, the symmetry and coordination environments of Se and Te atoms in BaSe and BaTe, respectively, are similar to those of the S atom.

Figure 5-6: The nearest (darker or green online) and second nearest (lighter, or orange online) neighboring barium atoms of sulfur atoms in the transition state at 5.5 GPa of (a) Mechanism I and (b) Mechanism II.

Due to the symmetry of the unit cell, the six nearest neighbors of the X atom in

Mechanism I are always equidistant, and this distance increases from the B1 to B2 structure. In contrast, as the transition proceeds, the two second nearest neighbors move closer to X until it finally has 8 nearest neighbors in total in the B2 phase. In Mechanism

II, symmetry requirements are not as stringent on the nearest neighbor distances of the X atom. The distances to the X atom from nearest neighbors in the (002) and (020) planes, defined as D1 and D2 respectively, are not mandated by symmetry to be equal. However, our analysis reveals that they are equal within the error bars associated with our computations along the transition path. As shown in Figure 5-6, our calculations (with an error of ±0.02 Å) reveal a D1 and D2 of 3.08 and 3.06 Å respectively for the transition state of BaS at 5.5 GPa, 3.18 and 3.16 Å for BaSe at 4.9 GPa and 3.40 and 3.39 Å for

BaTe at 3.4 GPa. This was not analyzed in earlier experimental work2 and may be tested 164

upon further experimental investigation. Therefore, the major difference between

Mechanism I and Mechanism II is the movement of the two planes adjacent to the (110) plane and the (002) plane respectively. The two planes in Mechanism I move in opposite directions while the two planes in Mechanism II move in the same direction.

5.6 Conclusions

In this project, we have studied the pressure-induced B1 to B2 phase transition of BaX (X = S, Se and Te) for both the Buerger (I) and WTM (II) mechanisms at three different pressures by using first-principles calculations. By constructing energy hypersurfaces, we have proposed modeled transition paths and obtained the energy barriers of the phase transitions, which indicate that the WTM mechanism is marginally favored for the pressure-induced phase transition of all three BaX compounds. No intermediate state was found during the pressure-induced phase transition from the B1 to

B2 phase for BaX. We discovered that the coordination number of the X atoms remains 6 throughout the transition as mandated by symmetry in mechanism I and despite any such mandate in mechanism II until the final B2 structure is reached.

165

Chapter 6

6. Summary and Future Work

In this work, Ta-S and Cu-S systems were thoroughly explored using NHSG methods, and pressure-induced phase transitions from the NaCl-type (B1) to the CsCl- type (B2) structure in BaS, BaSe and BaTe were studied using ab initio density functional theory computations in the local density approximation. The completeness of research goals and possible future work are briefly summarized in this chapter.

For the tantalum sulfide project, previously unexplored conditions were studied in the current work, and experiments aimed at optimizing synthetic conditions to achieve high reaction completeness were carried out. The results obtained in this project fulfill the main research goal, and no further work is necessary under the current research trajectory.

For the copper sulfide project, previously unexplored synthetic parameters for reactions between copper chlorides and DTBS or HMDST were studied. Moreover, synthetic conditions for reactions between CuCl and HMDST in acetonitrile and chloroform were finely tuned for the targeted synthesis of several copper sulfide polymorphs. By controlling temperature, HMDST concentration, reaction time and solvent type (chloroform or acetonitrile), five different copper sulfide polymorphs could

166

be prepared. The as-recovered phases mainly consisted of nanoparticles. It was found that particle size affected phase stability, resulting in the recovery of hexagonal chalcocite nanoparticles, a thermodynamically stable high temperature phase, at room temperature.

In addition, kinetically controlled phase evolutions between some polymorphs were observed. Such interesting results were beyond initial expectations, and fulfilled the goal of this project. However, new questions surfaced along with these results, and to answer these questions will improve the understanding of this system and provide useful information that may be applicable for phase and morphology control with other systems.

The main question is how phase selection occurs, and how phase evolution and particle sizes are related. As continuous phase evolutions were possibly caused by copper deficiencies based on the typical compositions of chalcocites and djurleite, it is important to further confirm these results by precisely measuring the compositions of the as- recovered phases. Although CHNS analysis could provide sulfur contents with errors around 0.01 mg, the current TGA instrument was unable to provide equivalent accuracy, which made distinguishing copper compositions between Cu1.94S and Cu2S impossible.

Although it is possible to obtain more meaningful data with a new TGA instrument or by measuring copper contents using ICP-AES, it can still be difficult to confirm the hypotheses even with more accurate measurements, as trace amounts of amorphous materials or unreacted CuCl can have a significant influence on this narrow composition region. In addition, formation of hexagonal chalcocite nanoparticles on the surface of particles by Cu diffusion results in measurements of phase mixtures. Hence, besides experimental means, insightful information can be obtained by theoretical modeling.

167

For the first step of theoretical modeling, hexagonal chalcocite, monoclinic chalcocite and djurleite nanoparticles of different sizes can be modeled by an energy minimization approach to obtain the stable phases as a function of particle sizes.

Although modeling for solids in a vacuum will be different from the actual solvent environment, a good estimate can still be obtained. In order to study the influence of copper deficiency on the stability of nanoparticles, copper atom(s) close to the surface can be removed from nanoparticles, and then minimum energies for copper deficient systems can be obtained by allowing the structures to relax. Thus, a phase diagram as a function of particle size and composition can be obtained, which will be very useful to compare with the experimental results. The results obtained from theoretical modeling can be further used to predict crystal growth in the copper sulfide system, which is helpful for further controlling morphologies and compositions during experiments. As diffusion (Cu atoms) has been modeled in Dr. Khare’s group, this modeling approach is feasible. However, further studies may require a dedicated Ph. D student.

For the theoretical modeling of the B1 to B2 pressure-induced phase transitions in barium chalcogenides, the goal of demonstrating the capability of modeling a simple sulfide system was fulfilled. To further advance in this direction, more complicated structure changes and phase transitions with intermediate phases can be modeled. As demonstrated in the copper sulfide project, it is possible to observe gradual phase evolutions. Hence, further modeling of phase transition pathways will not only be helpful for understanding phase transitions in transition metal sulfides, but may also reveal intermediate metastable phases.

168

References

Abstract

(1) Mutin, P. H.; Vioux, A. Chem. Mat. 2009, 21, 582-596.

(2) Rao, C. N. R.; Pisharody, K. P. R. Prog. Solid State Chem. 1976, 10, Pt. 4, 207-

70.

(3) Schleich, D. M.; Martin, M. J. J. Solid State Chem. 1986, 64, 359-364.

(4) Buerger, M. Phase Transformations in Solids; John Wiley & Sons, 1951.

(5) Watanabe, M.; Tokonami, M.; Morimoto, N. Acta Crystallogr., Sect. A 1977,

A33, 294-8.

Chapter 1

(1) Vaughan, D. J. Rev. Mineral. Geochem. 2006, 61, 1-5.

(2) Hoagland, P.; Beaulieu, S.; Tivey, M. A.; Eggert, R. G.; German, C.; Glowka, L.;

Lin, J. Mar. Pol. 2010, 34, 728-732.

(3) Pecoraro, T. A.; Chianelli, R. R. J. Catal. 1981, 67, 430-445.

(4) Bonde, J.; Moses, P. G.; Jaramillo, T. F.; Norskov, J. K.; Chorkendorff, I.

Faraday Discuss. 2008, 140, 219-231.

169

(5) Sagade, A. A.; Sharma, R.; Sulaniya, I. J. Appl. Phys 2009, 105, 043701-043701.

(6) Yu, X. L.; Wang, Y.; Chan, H. L. W.; Cao, C. B. Micropor. Mesopor. Mat. 2009,

118, 423-426.

(7) Bo, X.; Bai, J.; Wang, L.; Guo, L. Talanta 2010, 81, 339-345.

(8) Gu, W.; Jin, W.; Wei, B.; Zhang, J.; Wang, J. Appl. Phys. Lett. 2010, 97, 243303-

243303.

(9) Stevi, Z.; Mirjana, R. J. Power Sources 2006, 160, 1511-1517.

(10) Di Benedetto, F.; Borgheresi, M.; Caneschi, A.; Chastanet, G.; Cipriani, C.;

Gatteschi, D.; Pratesi, G.; Romanelli, M.; Sessoli, R. Eur. J. Mineral. 2006, 18,

283-287.

(11) Rapoport, L.; Fleischer, N.; Tenne, R. J. Mater. Chem. 2005, 15, 1782-1788.

(12) Alonso, G.; Chianelli, R. R. J. Catal. 2004, 221, 657-661.

(13) Rohrbach, A.; Hafner, J.; Kresse, G. J. Phys.-Condes. Matter 2003, 15, 979-996.

(14) Choi, J. Y.; Kim, K. J.; Yoo, J. B.; Kim, D. H. Sol. Energy 1998, 64, 41-47.

(15) Chemistry of the Elements (2nd Edition); Greenwood, N. N.; Earnshaw, A., Eds.;

Elsevier, 1998.

(16) Makovicky, E. Rev. Mineral. Geochem. 2006, 61, 7-125.

(17) Mineralogical Society of America Short Course Notes: Sulfide Mineralogy;

Ribbe, P. H., Ed.: Blacksburg, VA, USA, 1974; Vol. 1.

(18) van der Laan, G.; Pattrick, R. A. D.; Henderson, C. M. B.; Vaughan, D. J. J. Phys.

Chem. Solids 1992, 53, 1185-1190.

(19) Goh, S. W.; Buckley, A. N.; Lamb, R. N. Miner. Eng. 2006, 19, 204-208.

170

(20) Oviedo-Roa, R.; Martinez-Magadan, J. M.; Illas, F. J. Phys. Chem. B 2006, 110,

7951-7966.

(21) Wilcoxon, J. P. J. Phys. Chem. B 2000, 104, 7334-7343.

(22) Cody, G. D. Annu. Rev. Earth Planet. Sci. 2004, 32, 569-599.

(23) Lukashev, P.; Lambrecht, W. R. L.; Kotani, T.; van Schilfgaarde, M. Phys. Rev. B

2007, 76, 195202.

(24) Hench, T. L.; Bragagnolo, J. A.; Delgado, J. E.; Leyman, P. F.; Motta, P. F. J.

Vac. Sci. Technol. A 1983, 1, 360-363.

(25) Scheer, R.; Walter, T.; Schock, H. W.; Fearheiley, M. L.; Lewerenz, H. J. Appl.

Phys. Lett. 1993, 63, 3294-3296.

(26) Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, V.; Kis, A. Nat.

Nanotechnol. 2011, 6, 147-150.

(27) Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.; Kim, J.; Chim, C.; Galli, G.; Wang, F.

Nano Lett. 2010, 10, 1271-1275.

(28) Inorganic Chemistry Series; Zhang, Q., Ed.; Science Press: Beijing, China, 1984;

Vol. 5.

(29) Blachnik, R.; Müller, A. Thermochim. Acta 2001, 366, 47-59.

(30) Blachnik, R.; Müller, A. Thermochim. Acta 2000, 361, 31-52.

(31) Ohtani, T.; Koh, K.; Motoki, M.; Ohshima, K. Mater. Res. Bull. 1995, 30, 1495-

1504.

(32) Wold, A.; Dwight, K. J. Solid State Chem. 1992, 96, 53-58.

(33) Jiang, X.; Xie, Y.; Lu, J.; He, W.; Zhu, L.; Qian, Y. J. Mater. Chem. 2000, 10,

2193-2196.

171

(34) Kumar, P.; Gusain, M.; Nagarajan, R. Inorg. Chem. 2011, 50, 3065-3070.

(35) Yu, X.; An, X. Mater. Lett. 2010, 64, 252-254.

(36) Ancutienė, I.; Janickis, V. Cent. Eur. J. Chem. 2010, 8, 1281-1287.

(37) Lotfipour, M.; Machani, T.; Rossi, D. P.; Plass, K. E. Chem. Mat. 2011, 23, 3032-

3038.

(38) S. Liu; Q. Shi; Cheng, F. J. At. Mol. Phy. 2010, 27, 372-376.

(39) Bonneau, P. R.; Shibao, R. K.; Kaner, R. B. Inorg. Chem. 1990, 29, 2511-2514.

(40) Bonneau, P. R.; Jarvis, R. F.; Kaner, R. B. Nature 1991, 349, 510-512.

(41) Bonneau, P. R.; Jarvis, R. F., Jr.; Kaner, R. B. Inorg. Chem. 1992, 31, 2127-32.

(42) Treece, R. E.; Gillan, E. G.; Kaner, R. B. Comments Inorg. Chem. 1995, 16, 313-

337.

(43) Muthuswamy, E.; Brock, S. L. J. Am. Chem. Soc. 2010, 132, 15849-15851.

(44) Krunks, M.; Leskelä, T.; Mannonen, R.; Niinistö, L. J. Therm. Anal. Calorim.

1998, 53, 355-364.

(45) Krunks, M.; Madarász, J.; Leskelä, T.; Mere, A.; Niinistö, L.; Pokol, G. J. Therm.

Anal. Calorim. 2003, 72, 497-506.

(46) Brinker, C. J.; Scherer, W. Sol-Gel Science: The Physics and Chemistry of Sol-

Gel Processing; Academic Press: San Diego, 1990.

(47) Gerrard, W.; Kilburn, K. D. J. Chem. Soc. 1956, 1536-1539.

(48) Gerrard, W.; Woodhead, A. H. J. Chem. Soc. 1951, 519-522.

(49) Ridge, D.; Todd, M. J. Chem. Soc. 1949, 2637-2640.

(50) Corriu, R. J. P.; Leclercq, D.; Lefevre, P.; Mutin, P. H.; Vioux, A. J. Mater.

Chem. 1992, 2, 673-674.

172

(51) Acosta, S.; Corriu, R. J. P.; Leclercq, D.; Lefevre, P.; Mutin, P. H.; Vioux, A. J.

Non-cryst. Solids 1994, 170, 234-242.

(52) Corriu, R. J. P.; Leclercq, D. Angew. Chem. Int. Edit. 1996, 35, 1420-1436.

(53) Andrianainarivelo, M.; Corriu, R.; Leclercq, D.; Mutin, P. H.; Vioux, A. J. Mater.

Chem. 1996, 6, 1665-1671.

(54) Vioux, A.; eacute; Mahandrimanana; Mutin, P. H.; Leclercq, D.; Corriu J. Mater.

Chem. 1997, 7, 279-284.

(55) Vioux, A. Chem. Mat. 1997, 9, 2292-2299.

(56) Mutin, P. H.; Vioux, A. Chem. Mat. 2009, 21, 582-596.

(57) Schleich, D. M.; Martin, M. J. J. Solid State Chem. 1986, 64, 359-364.

(58) Martin, M. J.; Qiang, G. H.; Schleich, D. M. Inorg. Chem. 1988, 27, 2804-2808.

(59) Bensalem, A.; Schleich, D. M. Mater. Res. Bull. 1988, 23, 857-868.

(60) Bensalem, A.; Schleich, D. M. Mater. Res. Bull. 1990, 25, 349-356.

(61) Bensalem, A.; Schleich, D. M. Inorg. Chem. 1991, 30, 2052-2055.

(62) Computational Materials Design; Saito, T., Ed.; Springer: Berlin, Germany, 1999.

(63) Chookajorn, T.; Murdoch, H. A.; Schuh, C. A. Science 2012, 337, 951-954.

(64) Parker, S. C. Solid State Ionics 1983, 8, 179-186.

(65) Filippini, G.; Gramaccioli, C. M.; Simonetta, M.; Suffritti, G. B. Acta

Crystallogr., Sect. A 1976, 32, 259-264.

(66) Patel, A.; Price, G. D.; Mendelssohn, M. J. Phys. Chem. MInerals 1991, 17, 690-

699.

(67) Barron, T. H. K.; Collins, J. G.; White, G. K. Adv. Phys. 1980, 29, 609-730.

(68) Alder, B. J.; Wainwright, T. E. J. Chem. Phys. 1959, 31, 459-466.

173

(69) Rahman, A. Phys. Rev. A 1964, 136, A405-A411.

(70) Metropolis, N.; Ulam, S. J. Am. Stat. Assoc 1949, 44, 335-341.

(71) Metropolis, N.; Rosenbluth, A. W.; Rosenbluth, M. N.; Teller, A. H.; Teller, E. J.

Chem. Phys. 1953, 21, 1087-1092.

(72) Yashonath, S.; Thomas, J. M.; Nowak, A. K.; Cheetham, A. K. Nature 1988, 331,

601-604.

(73) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 137, 1697-705.

(74) Hohenberg, P.; Kohn, W. Phys. Rev. B 1964, 136, B864.

(75) Computer Modelling in In organic Ctrystallography; Catlow, C. R. A., Ed.;

Academic Press: London, UK, 1997.

(76) Pedoussaut, N. M.; Lind, C. Inorg. Chem. 2008, 47, 392-394.

(77) Pedoussaut, N. M.S. Thesis, University of Toledo, 2008.

Chapter 2

(1) In MDI Jade; 8 ed.; Materials Data Inc.: Livermore, CA.

(2) In ICDD PDF-2 database; ICDD, 2006.

(3) In TOPAS-Academic; 4 ed.; Coelho Software: Brisbane, Australia, 2006.

Chapter 3

(1) Roseboom, E. H. Econ. Geol. 1966, 61, 641-672.

(2) Fleet, M. E. Rev. Mineral. Geochem. 2006, 61, 365-419.

(3) Djurle, S. Acta Chem. Scand. 1958, 12, 1415-1426.

174

(4) Evans, H. T. Zeitschrift für Kristallographie 1979, 150, 299-320.

(5) EVANS, H. T. Science 1979, 203, 356-358.

(6) Goble, R. J.; Robinson, G. Can. Mineral. 1980, 18, 519-523.

(7) Evans, H. T. Am. Mineral. 1981, 66, 807-818.

(8) Posfai, M.; Buseck, P. R. Am. Mineral. 1994, 79, 308-315.

(9) Lukashev, P.; Lambrecht, W. R. L.; Kotani, T.; van Schilfgaarde, M. Phys. Rev. B

2007, 76, 195202.

(10) Goh, S. W.; Buckley, A. N.; Lamb, R. N. Miner. Eng. 2006, 19, 204-208.

(11) Sagade, A. A.; Sharma, R.; Sulaniya, I. J. Appl. Phys 2009, 105, 043701-043701.

(12) Yu, X. L.; Wang, Y.; Chan, H. L. W.; Cao, C. B. Micropor. Mesopor. Mat. 2009,

118, 423-426.

(13) Bo, X.; Bai, J.; Wang, L.; Guo, L. Talanta 2010, 81, 339-345.

(14) Gu, W.; Jin, W.; Wei, B.; Zhang, J.; Wang, J. Appl. Phys. Lett. 2010, 97, 243303-

243303.

(15) Stevi, Z.; Mirjana, R. J. Power Sources 2006, 160, 1511-1517.

(16) Hench, T. L.; Bragagnolo, J. A.; Delgado, J. E.; Leyman, P. F.; Motta, P. F. J.

Vac. Sci. Technol. A 1983, 1, 360-363.

(17) Partain, L. D.; McLeod, P. S.; Duisman, J. A.; Peterson, T. M.; Sawyer, D. E.;

Dean, C. S. J. Appl. Phys 1983, 54, 6708-6720.

(18) Li, F.; Kong, T.; Bi, W.; Li, D.; Li, Z.; Huang, X. Appl. Surf. Sci. 2009, 255,

6285-6289.

(19) Lu, Y.; Yi, G.; Jia, J.; Liang, Y. Appl. Surf. Sci. 2010, 256, 7316-7322.

175

(20) Sam, M.; Bayati, M. R.; Mojtahedi, M.; Janghorban, K. Appl. Surf. Sci. 2010,

257, 1449-1453.

(21) Lin, M. C.; Lee, M. W. Electrochem. Commun. 2011, 13, 1376-1378.

(22) Lotfipour, M.; Machani, T.; Rossi, D. P.; Plass, K. E. Chem. Mat. 2011, 23, 3032-

3038.

(23) Zou, J.; Jiang, J.; Huang, L.; Jiang, H.; Huang, K. Solid State Sci. 2011, 13, 1261-

1267.

(24) Page; lt; fnoteref; gt; Miles; Niitsoo; Olivia; Itzhaik, Y.; Cahen, D.; Hodes, G.

Energy Environm. Sci. 2009, 2, 220-223.

(25) Isac, L.; Popovici, I.; Enesca, A.; Duta, A. Energy Procedia 2010, 2, 71-78.

(26) Ghahremaninezhad, A.; Asselin, E.; Dixon, D. G. Electrochem. Commun. 2011,

13, 12-15.

(27) Blachnik, R.; Müller, A. Thermochim. Acta 2000, 361, 31-52.

(28) Scheer, R.; Walter, T.; Schock, H. W.; Fearheiley, M. L.; Lewerenz, H. J. Appl.

Phys. Lett. 1993, 63, 3294-3296.

(29) Rao, C. N. R.; Pisharody, K. P. R. Prog. Solid State Chem. 1976, 10, Part 4, 207-

270.

(30) Ohtani, T.; Koh, K.; Motoki, M.; Ohshima, K. Mater. Res. Bull. 1995, 30, 1495-

1504.

(31) Potter, R. W. Econ. Geol. 1977, 72, 1524-1542.

(32) Cavallotti, P.; Salvago, G. Electrochim. Metal. 1969, 4, 181-210.

(33) Koch, D. F. A.; McIntyre, R. J. j. Electroanal. Chem. Interficial Elechochem.

1976, 71, 285-296.

176

(34) Mumme, W. G.; Sparrow, G. J.; Walker, G. S. Mineral Mag. 1988, 52, 323-330.

(35) K. Koto, N. M. Acta Crystallogr., Sect. B 1969, B26, 915.

(36) Goble, R. J. Can. Min. Metall. Bull. 1988, 81, 110-114.

(37) Goble, R. J. Can. Mineral. 1980, 18, 511-518.

(38) Evans, H. T.; Konnert, J. A. Am. Mineral. 1976, 61, 996-1000.

(39) Pattrick, R. A. D.; Mosselmans, J. F. W.; Charnock, J. M.; England, K. E. R.;

Helz, G. R.; Garner, C. D.; Vaughan, D. J. Geochim. Cosmochim. Acta 1997, 61,

2023-2036.

(40) King, H. E.; Prewitt, C. T. Am. Mineral. 1979, 64, 1265-1271.

(41) van der Laan, G.; Pattrick, R. A. D.; Henderson, C. M. B.; Vaughan, D. J. J. Phys.

Chem. Solids 1992, 53, 1185-1190.

(42) Chemistry of the Elements (2nd Edition); Greenwood, N. N.; Earnshaw, A., Eds.;

Elsevier, 1998.

(43) Tezuka, K.; Sheets, W. C.; Kurihara, R.; Shan, Y. J.; Imoto, H.; Marks, T. J.;

Poeppelmeier, K. R. Solid State Sci. 2007, 9, 95-99.

(44) Dutrizac, J. E.; MacDonald, R. J. C. Mater. Res. Bull. 1973, 8, 961-971.

(45) Jiang, X.; Xie, Y.; Lu, J.; He, W.; Zhu, L.; Qian, Y. J. Mater. Chem. 2000, 10,

2193-2196.

(46) Yu, X.; An, X. Mater. Lett. 2010, 64, 252-254.

(47) Kumar, P.; Gusain, M.; Nagarajan, R. Inorg. Chem. 2011, 50, 3065-3070.

(48) Ancutienė, I.; Janickis, V. Cent. Eur. J. Chem. 2010, 8, 1281-1287.

(49) Li, S.; Wang, H.; Xu, W.; Si, H.; Tao, X.; Lou, S.; Du, Z.; Li, L. S. J. Colloid

Interf. Sci. 2009, 330, 483-487.

177

(50) Lim, W. P.; Wong, C. T.; Ang, S. L.; Low, H. Y.; Chin, W. S. Chem. Mat. 2006,

18, 6170-6177.

(51) Li, W.; Shavel, A.; Guzman, R.; Rubio-Garcia, J.; Flox, C.; Fan, J.; Cadavid, D.;

Ibanez, M.; Arbiol, J.; Morante, J. R.; Cabot, A. Chem. Commun. 2011, 47,

10332-10334.

(52) Kruszynska, M.; Borchert, H.; Bachmatiuk, A.; Rümmeli, M. H.; Büchner, B.;

Parisi, J.; Kolny-Olesiak, J. ACS Nano 2012, 6, 5889-5896.

(53) Yan, H.; Wang, W.; Xu, H. J. Cryst. Growth 2008, 310, 2640-2643.

(54) Chen, L.; Yu, W.; Li, Y. Powder Technol. 2009, 191, 52-54.

(55) Chu, L.; Zhou, B.; Mu, H.; Sun, Y.; Xu, P. J. Cryst. Growth 2008, 310, 5437-

5440.

(56) Liao, X.; Chen, N.; Xu, S.; Yang, S.; Zhu, J. J. Cryst. Growth 2003, 252, 593-598.

(57) Thongtem, T.; Phuruangrat, A.; Thongtem, S. Mater. Lett. 2010, 64, 136-139.

(58) Yang, Y. J.; Hu, S. J. Solid State Electr. 2008, 12, 1405-1410.

(59) Fotouhi, L.; Rezaei, M. Microchim. Acta 2009, 167, 247-251.

(60) Behboudnia, M.; Khanbabaee, B. J. Cryst. Growth 2007, 304, 158-162.

(61) Xiong, S.; Zeng, H. C. Angew. Chem. Int. Edit. 2012, 124, 973-976.

(62) Xu, C.; Wang, L.; Zou, D.; Ying, T. Mater. Lett. 2008, 62, 3181-3184.

(63) Mathew, S. K.; Rajesh, N. P.; Ichimura, M.; Udayalakshmi. Mater. Lett. 2008, 62,

591-593.

(64) Pedoussaut, N. M.S. Thesis, University of Toledo, 2008.

(65) Quantitative Chemical Analysis; 6th ed.; Harris, D. C., Ed.; W. H. Freeman and

Company: New York, US, 2003.

178

(66) Kolthoff, I. M.; Coetzee, J. F. J. Am. Chem. Soc. 1957, 79, 1852-1858.

Chapter 4

(1) Rao, C. N. R.; Pisharody, K. P. R. Prog. Solid State Chem. 1976, 10, Pt. 4, 207-

70.

(2) Jellinek, F. J. Less Common Met. 1962, 4, 9-15.

(3) Bjerkelund, E.; H., F. J.; Kjekshus, A. Acta Chem. Scand. 1966, 20, 1836-1842.

(4) Franzen, H. F.; Smeggil, J. G. J. Am. Chem. Soc. 1969, 91, 2814-2815.

(5) Kim, S. J.; Nanjundaswamy, K. S.; Hughbanks, T. Inorg. Chem. 1991, 30, 159-

164.

(6) Di Salvo, F. J.; Bagley, B. G.; Voorhoeve, J. M.; Waszczak, J. V. J. Phys. Chem.

Solids 1973, 34, 1357-1362.

(7) Bjerkelund, E.; Kjekshus, A. Z. Anorg. Allg. Chem. 1964, 328, 235-242.

(8) Meerschaut, A.; Guemas, L.; Rouxel, J. J. Solid State Chem. 1981, 36, 118-123.

(9) Gotoh, Y.; Akimoto, J.; Oosawa, Y. J. Alloy. Compd. 1998, 270, 115-118.

(10) Gamble, F. R.; Osiecki, J. H.; Cais, M.; Pisharody, R.; DiSalvo, F. J.; Geballe, T.

H. Science 1971, 174, 493-497.

(11) Johnson, W. B.; Worrell, W. L. Synth. Met. 1982, 4, 225-248.

(12) Hernan, L.; Morales, J.; Sanchez, L.; Tirado, J. L. Electrochim. Acta 1994, 39,

2665-2671.

(13) Dunnill, C. W.; Edwards, H. K.; Brown, P. D.; Gregory, D. H. Angew. Chem. Int.

Edit. 2006, 45, 7060-7063.

(14) Wu, X.-C.; Tao, Y.-R.; Gao, Q.-X. Chem. Commun. 2009, 6008-6010.

179

(15) Thompson, A. H.; Gamble, R. F.; Revelli, J. F. Solid State Commun. 1971, 9, 981-

985.

(16) Conroy, L. E.; Pisharody, K. R. J. Solid State Chem. 1972, 4, 345-350.

(17) Figueroa, E.; Kuo, Y. K.; Olinger, A.; Lloyd, M. A.; Bastin, L. D.; Petrotsatos, S.;

Chen, Q.; Dobbs, B.; Dev, S.; Selegue, J. P.; DeLong, L. E.; Brock, C. P.; Brill, J.

W. J. Solid State Chem. 1995, 114, 486-490.

(18) Kikkawa, S.; Koizumi, M.; Yamanaka, S.; Onuki, Y.; Inada, R.; Tanuma, S. J.

Solid State Chem. 1981, 40, 28-33.

(19) Chianelli, R. R.; Dines, M. B. Inorg. Chem. 1978, 17, 2758-2762.

(20) Carmalt, C. J.; Dinnage, C. W.; Parkin, I. P.; Peters, E. S.; Molloy, K.; Colucci,

M. A. Polyhedron 2003, 22, 1255-1262.

(21) Mull, D. Honor's Thesis, Advisor: Lind, C. Unversity of Toledo, 2010.

Chapter 5

(1) Buerger, M. Phase Transformations in Solids; Smoluchowski, R.; Mayer, J. E.;

Weyl, W. A., Eds.; John Wiley & Sons, 1951.

(2) Watanabe, M.; Tokonami, M.; Morimoto, N. Acta Crystallogr., Sect. A 1977,

A33, 294-8.

(3) Stokes, H. T.; Hatch, D. M. Phys. Rev. B 2002, 65, 144114.

(4) Onodera, A.; Kawano, S.; Nakai, Y.; Achiwa, N. Physica B 1992, 180-181, 279-

80.

(5) Pendas, A. M.; Luana, V.; Recio, J. M.; Florez, M.; Francisco, E.; Blanco, M. A.;

Kantorovich, L. N. Phys. Rev. B 1994, 49, 3066-3074.

180

(6) Sims, C. E.; Barrera, G. D.; Allan, N. L.; Mackrodt, W. C. Phys. Rev. B 1998, 57,

11164-11172.

(7) Catti, M. J. Phys.-Condes. Matter 2004, 16, 3909-3921.

(8) Mota, R. C.; Branicio, P. S.; Rino, J. P. Europhys. Lett. 2006, 76, 836-841.

(9) Toledano, P.; Knorr, K.; Ehm, L.; Depmeier W. Phys. Rev. B 2003, 67, 144106.

(10) Stokes, H. T.; Hatch, D. M.; Dong, J.; Lewis, J. P. Phys. Rev. B 2004, 69, 174111.

(11) Bouhemadou, A.; Khenata, R.; Zegrar, F.; Sahnoun, M.; Baltache, H.; Reshak, A.

H. Comput. Mater. Sci. 2006, 38, 263-270.

(12) Yamaoka, S.; Shimomura, O.; Nakazawa, H.; Fukunaga, O. Solid State Commun.

1980, 33, 87-89.

(13) Grzybowski, T. A.; Ruoff, A. L. Phys. Rev. B 1983, 27, 6502-6503.

(14) Grzybowski, T. A.; Ruoff, A. L. Phys. Rev. Lett. 1984, 53, 489.

(15) Ruoff, A.; Grzybowski, T. A. Solid State Physics Under Pressure; Minomura, S.,

Ed.; Terra Scientific: Tokyo, 1985.

(16) Weir, S. T.; Vohra, Y. K.; Ruoff, A. L. Phys. Rev. B 1987, 35, 874-876.

(17) Jha, P. K.; Sakalle, U. K.; Sanyal, S. P. J. Phys. Chem. Solids 1998, 59, 1633-

1637.

(18) Durandurdu, M. Chem. Phys. 2010, 367, 80-82.

(19) Potzel, O.; Taubmann, G. J. Solid State Chem. 2011, 184, 1079-1084.

(20) Kalpana, G.; Palanivel, B.; Rajagopalan, M. Phys. Rev. B 1994, 50, 12318-12325.

(21) Lin, G. Q.; Gong, H.; Wu, P. Phys. Rev. B 2005, 71, 085203-085207.

(22) Zahn, D.; Leoni, S. Phys. Rev. Lett. 2004, 92, 250201.

(23) Zahn, D.; Hochrein, O.; Leoni, S. Phys. Rev. B 2005, 72, 094106.

181

(24) Kresse, G. Thesis, Technische Universiät Wien, 1993.

(25) Kresse, G.; Hafner, J. Phys. Rev. B 1993, 47, 558.

(26) Kresse, G.; Furthmüller, J. Comput. Mater. Sci. 1996, 6, 15-50.

(27) Kresse, G.; Furthmüller, J. Phys. Rev. B 1996, 54, 11169.

(28) Hohenberg, P.; Kohn, W. Phys. Rev. B 1964, 136, 864.

(29) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 137, 1697-705.

(30) Vanderbilt, D. Phys. Rev. B 1990, 41, 7892.

(31) Kresse, G.; Hafner, J. J. Phys.-Condes. Matter 1994, 6, 8245-8257.

(32) Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13, 5188.

(33) Cai, J.; Chen, N. X. J. Phys.-Condes. Matter 2007, 19.

(34) Weir, S. T.; Vohra, Y. K.; Ruoff, A. L. Phys. Rev. B 1986, 33, 4221.

182

Appendix A

Instructions for the voltage spike filter and source code of the algorithm

Because the detector of the Scintag XDS-2000 diffractrometer (Dinosaur) at the

University of Toledo occasionally generates voltage spikes, a program called

"Triceratops" was written to filter these spikes and correct the data. In order to load data in Triceratops, PXRD data files collected using the Scintag XDS-2000 must be exported as plain text files. To load such text files, a data loading window can be called up by clicking "Open" in the "File" menu (Figure A-1), and exported *.txt files can be chosen.

Commands in the "Plot" menu provide operations for data visualization in the plot area, and by clicking "Reset" the loaded data will be shown in the data plot area, as the example with voltage spikes shows in Figure A-1.

After the data are loaded, a value between 0 and 1 must be entered in the text box next to the label "Sigma range", otherwise the default value 0.19 will be applied by the program. This value determines the portion of the data with the lowest intensities that will be used to estimate the baseline or noise level. For example, a value of 0.19 means

19% of the lowest intensity values (counts) will be treated as noise. This value gives users an option to handle different situations. Once the sigma value is entered, voltage

183

spikes can be filtered by simply clicking the button "Voltage Spike Filter". An example data file after voltage spike filtering is shown in Figure A-2, and the peak positions that were filtered are listed in the "Results" column. If users are satisfied with the results, corrected data can be saved by clicking "Save" in the "File" menu, and the names of the saved files will be shown. For more advanced users that are interested in understanding or modifying the algorithm, the source code for the algorithm written in Visual Basic 6.0 is provided at the end of this section.

Figure A-1: User interface of the voltage spike filtering program "Triceratops".

184

Figure A-2: Plot after the data shown in Figure A-1 was filtered.

185

Source Code:

Private Sub VotSpik_Click() Dim sum, ssum, n, z, inter, Cps(5000), IfOut, XRpk, Rsigma n = 1 sum = 0: ssum = 0 List1.Clear sum = 0: ssum = 0: List1.Clear Rsigma = Val(Text2.Text) If Rsigma > 1 Or Rsigma < 0 Then Rsigma = 0.19 Text2.Text = 0.19 End If 'Find the Sigma for 19% (Rsigma) of the lowest intensities For i = 1 To nd Cps(i) = Ydata(i) Next i inter = Cps(1) For i = 1 To nd For j = i To nd If Cps(i) > Cps(j) Then inter = Cps(i) Cps(i) = Cps(j) Cps(j) = inter End If Next j Next i i = 0 Do Until (i > Rsigma * nd) sum = sum + Cps(i) i = i + 1 Loop i = i - 1 For k = 1 To i ssum = ((sum / i) - Cps(i)) ^ 2 + ssum Next k Sigma = (ssum / i) ^ 0.5

'Find the intensity increase larger than 3 sigma

i = 2 Do Until (i >= nd - 10) inter = 0 'Intensity back to baseline in 5 pts If Ydata(i + 1) - Ydata(i) > 3 * Sigma Then For z = 5 To 9

186

If Abs(Ydata(i + z) - Ydata(i)) > 3 * Sigma Then inter = 1 'Intensity NOT back to baseline after 5 points Next z If inter = 0 Then For z = 1 To 4 Ydata(i + 5 - z) = Ydata(i) + (Ydata(i + 5) - Ydata(i)) * z / 5 List1.AddItem "intensity replaced at:" + Format$(Xdata(i + 5 - z), "00.000") Next z i = i + 9 Else IfOut = 0: XRpk = 1 Do Until (IfOut <> 0) Or (i + XRpk + 10 >= nd) If Abs(Ydata(i + XRpk) - Ydata(i)) > 3 * Sigma Then IfOut = 1 For z = 1 To 10 If Abs(Ydata(i + XRpk + z) - Ydata(i + XRpk)) > 3 * Sigma Then IfOut = 0 Next z Else IfOut = 2 For z = 1 To 10 If Abs(Ydata(i + XRpk + z) - Ydata(i)) > 3 * Sigma Then IfOut = 0 Next z End If XRpk = XRpk + 1 Loop XRpk = XRpk - 1 If IfOut = 1 Then 'Ramp shaped i = i + XRpk End If If IfOut = 2 Then 'Peak Dim Yin(3000), Xin(3000), Rin, Xmax, Rmax Rin = 1 'Determine intensity increase of the peak For z = 1 To XRpk If Ydata(i + z) - Ydata(i + z - 1) > 0 Then Xmax = z Next z For z = 1 To Xmax If Ydata(i + z) - Ydata(i + z - 1) > 0 Then Yin(Rin) = Ydata(i + z): Xin(Rin) = Xdata(i + z) Rin = Rin + 1 End If Next z Rmax = Rin - 1 For z = Xmax To XRpk If Ydata(i + z) - Ydata(i + z - 1) < 0 Then

187

Yin(Rin) = Ydata(i + z): Xin(Rin) = Xdata(i + z) Rin = Rin + 1 End If Next z Rin = Rin - 1 If (Yin(Rmax) - Yin(Rmax - 1) > Yin(Rmax - 1) - Yin(Rmax - 2)) And (Yin(Rmax) - Yin(Rmax + 1) > Yin(Rmax + 1) - Yin(Rmax + 2)) And (Yin(Rmax) - Yin(Rmax - 1) > 3 * Sigma) And (Yin(Rmax) - Yin(Rmax + 1) > 3 * Sigma) Then Ydata(i + Xmax) = (Ydata(i + Xmax - 1) + Ydata(i + Xmax + 1)) / 2 List1.AddItem "intensity replaced at:" + Format$(Xdata(i + Xmax), "00.000") Else For z = 1 To Rmax If ((Xin(z + 1) - Xin(z)) / (Xdata(2) - Xdata(1)) > 5) And (Yin(z) - Yin(z - 1) > 3 * Sigma) Then Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1))) = 0.5 * (Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1)) - 1) + Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1)) + 1)) List1.AddItem "intensity replaced at:" + Format$(Xin(z), "00.000") End If Next z For z = Rmax To Rin If ((Xin(z + 1) - Xin(z)) / (Xdata(2) - Xdata(1)) > 5) And (Yin(z) - Yin(z + 1) > 3 * Sigma) Then Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1))) = 0.5 * (Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1)) - 1) + Ydata((Xin(z) - Xdata(1)) / (Xdata(2) - Xdata(1)) + 1)) List1.AddItem "intensity replaced at:" + Format$(Xin(z), "00.000") End If Next z End If i = i + XRpk End If i = i + 1 End If Else i = i + 1 End If Loop For z = 1 To nd Yo(z) = Ydata(z) Next z End Sub

188

Appendix B

Lists of CH and DM tantalum sulfide samples

Table B-1: Conditions and XRD results of tantalum sulfide samples prepared by Christophe Heinrich (CHTa samples).

Metal Source Sulfur Source Solvent Conditions Ratio Recovered Sample XRD Results S:M mass (g) V T Name mmol Name mmol Name Time (mL) (°C)

CHTa1 TaCl5 3 DTBS 9 CHCl3 15 3 100 7 d 1.241 N/A

CHTa2 TaCl5 3 DTBS 9 CH2Cl2 7.5 3 100 7 d 0.250 N/A

CHTa3 TaCl5 3 DTBS 9 CH3CN 15 3 100 7 d 0.103 Unknown crystalline phase

CHTa4 TaCl5 3 DTBS 18 N/A N/A 6 100 7 d 1.527 Amorphous Amorphous and unknown CHTa5 TaCl 3 DTBS 18 CHCl 50 6 RT 0.5 h 0.471 5 3 crystalline phase

CHTa6 TaCl5 3 DTBS 9 CH3CN 15 3 150 7 d 1.434 Unknown crystalline phase

CHTa7 TaCl5 3 HMDST 9 CH3CN 15 3 150 7 d 0.875 Amorphous

CHTa8 TaCl5 3 HMDST 9 CH3CN 15 3 RT 0.3 h 0.866 Amorphous

CHTa9 TaF5 3 HMDST 9 CH3CN 45 3 RT 2 d 0.828 Amorphous and TaF3

CHTa10 TaBr5 3 HMDST 9 CH3CN 45 3 RT 2 d 1.045 Amorphous

CHTa11 TaCl5 3 DTBS 9 CH3CN 15 3 100 7 d 0.016 N/A

CHTa12 TaI5 2 HMDST 6 CH3CN 30 3 RT 1 d 0.995 Amorphous Amorphous and unknown CHTa13 TaCl 3 DTBS 9 CHCl 75 3 RT 0.5 h 0.668 5 3 phase

CHTa14 TaCl5 3 HMDST 9 CHCl3 75 3 RT 0.5 h 0.782 Amorphous

CHTa15 TaCl5 3 HMDST 5.7 CH3CN 45 2 RT 1 d 0.730 Amorphous Amorphous/nanocrystalline CHTa16 TaI5 3 HMDST 9 CH3CN 15 3 150 7 d 0.950 TaS2

CHTa17 TaCl5 3 DTBS 9 CH3CN 45 3 RT 1 d failed N/A

CHTa18 TaI5 3 DTBS 9 CHCl3 75 3 RT 0.5 h 2.204 TaI5

CHTa19 TaI5 3 HMDST 9 CHCl3 75 3 RT 0.5 h 2.186 TaI5

CHTa20 TaCl5 3 DTBS 9 CHCl3 75 3 RT 0.5 h 0.597 Unknown crystalline phase

CHTa21 TaCl5 3 HMDST 9 CH3CN 45 3 RT 1 d 0.690 Amorphous CHTa22 TaI5 2 DTBS 6 CH3CN 30 3 RT 1 d 0.479 N/A

189

Table B-2: Conditions and XRD results of tantalum sulfide samples prepared by Derek Mull (DM samples).

Metal Source HMDST CH CN Conditions 3 Ratio Recovered Sample XRD Results S:M mass (g) T Name mmol mmol V (mL) Time (°C)

DM1 TaCl5 3 3 15 1 150 7 d 1.13 Amorphous

DM2 TaCl5 3 3 15 1 RT 0.5 h 0.81 Amorphous

DM3 TaF5 3 3 15 1 RT 1 d 0.82 TaF3

DM4 TaBr5 3 3 15 1 RT 3 d 0.81 Amorphous

DM5 TaI5 3 3 15 1 RT 3 d 1.46 TaI5

DM6 TaF5 3 3 15 1 150 7 d 0.85 TaF3

DM7 TaBr5 3 3 15 1 150 7 d 0.94 Amorphous

DM8 TaF5 3 3 15 1 RT 3 d 0.28 TaF3

DM9 TaI5 3 3 15 1 150 7 d 0.98 Amorphous

DM10 TaCl5 3 3 15 1 RT 0.5 h 0.80 Amorphous

DM11 TaCl5 3 3 0 1 RT 7 d 0.77 Unknown

DM12 TaCl5 3 18 0 6 RT 7 d 0.09 N/A DM13 TaCl5 3 3 7.5 1 RT 0.8 h 0.27 Amorphous

190