Alexandre Henrique Pinto

Green Chemistry Approach for The Synthesis of

Transition Metal Sulfides based on Cu2ZnSnS4

(CZTS) and Etching of their Impurities

A DISSERTATION

SUBMITTED TO THE FACULTY OF

UNIVERSITY OF MINNESOTA

BY

Alexandre Henrique Pinto

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Professor R. Lee Penn

JULY

2017

© Alexandre Henrique Pinto 2017

i

ACKNOWLEDGMENTS

During these five years in Graduate School, many people contributed to the

accomplishment of this Doctoral work. Among them, I firstly recognize all the

support from my parents and brother, who stood these five years without my

presence. Second, I would like to acknowledge all the support and love from my

wife Caroline Barbosa, who postponed many of her career goals to follow me in

this journey to live and establish a family in a foreign country.

This Doctoral research would not be possible without the support from my advisor

Professor R. Lee Penn, and collaborators Professor Eray Aydil, and Dr. Seung

Wook Shin. When anyone read this dissertation, make sure that the content

present here is not only outcome from my own and lonely thoughts or efforts.

Instead, it is a result of many extensive discussions, which some of them lasted

for months or years. To perform this research, I had to learn to operate many

pieces of equipment to characterize my samples, and the University of Minnesota

Characterization Facility staff were essential in providing me all the knowledge and

resources necessary. I would like to nominally acknowledge three people from

their staff: Dr. Jason Myers, Dr. Nick Seaton, and Dr. Bing Luo for all their

availability and patience throughout these five years.

I consider the opportunity to mentor undergraduate students in research one of the

highlights in Graduate School. I was very lucky to find many good ones. All of them

were very open to learn new things, and committed to the research. Fortunately, I

can say that all of them contributed substantially to the development of my

ii

research. So, Mr. Arafat Akintomiwa Akinlabi (Summer 2014), Ms. Aastha Sharma

(Summer 2015), Ms. Elianna Isaac (Summer 2016), and Ms. Emily Foster (Fall

2016, Spring 2017), I would like to say thank you so much for all your time and

efforts.

The beginning in Graduate School was very hard. Besides all the adjustments

necessary to adapt to the life in a new educational system, the complexity of the

course work made the things overwhelming. On that point, some Professors were

very helpful talking to me and helping me to understand these differences. Maybe

for them it was just few minutes of their day, but for me those talks and advices

strengthened my self-confidence and helped me to stay in Graduate School. In

that sense, I would like to acknowledge Professors Kate Martin (Center of

Education Innovation), Professor Chris Macosko (Chemical Engineering and

Materials Science), and Dr. Chris Thurber (former Teaching Assistant and

Graduate Student at Chemical Engineering and Materials Science). I am also

thankful for all Professors who I had to take some course from, they are Professors:

Andreas Stein, Kent Mann, Connie Lu, Aaron Massari, K. Andre Mkhoyan, and

Wayne Gladfelter.

During these years, my interest about teaching grew further, and some people

were very important teaching me about those topics. Professor John Dwyer (St.

Catherine University), Professor Deb Wingert (Center of Education Innovation),

and Professor Bill Rozaitis (Center of Education Innovation), I would like to say

iii

thanks for opening a new door called Teaching in Higher Education, and revealing

many details from this area.

Finally, I say thank you to all Lee Penn Research Group members that I have the

chance to share office and learn a lot from them on these five years. Nathan

Burrows, Jennifer Soltis, Jennifer Strehlau, Amanda Vindedahl, Ryan Knutson,

Kairat Sabirov, Tom Webber, Jeannette Voelz, Suyue Chen, Clare Johnston, Clay

Easterday, Anna Abfalterer, and Louis Corcoran.

iv

DEDICATION

This dissertation is specially dedicated to my soon to be born son Anthony, to my

wife Caroline da Silva Barbosa, to my parents Olavio Ferreira Pinto and Evanilde

Jacques Pinto, to my brother Daniel Fernando Pinto, and to my grandmother Maria

de Souza.

My dear Grandma, I still remember when I was about five, and you told that you

would not live much to see me become a Doctor. I am so glad and thankful to God

that you were wrong, and we can enjoy this moment with you very healthy and

well.

v

ABSTRACT

Green Chemistry comprises a set of good practices leading to more sustainable

and environmentally friendly chemical processes. In the first chapter, this

dissertation introduces the Green Chemistry Principles and shows how these

Principles can be applied towards the synthesis of transition metal chalcogenides.

The next chapter presents results focused on the synthesis of the multinary sulfide

Cu2ZnSnS4 (CZTS) using microwave as heating source. The control over the

crystalline phase of CZTS was studied as a function of variations in synthetic

conditions, such as source of sulfur excess, initial oxidation states of Cu and Sn

sources, temperature, and time. A model explaining two different behaviors

according to the sulfur excess source is proposed. The third chapter uses the

concepts learned from the previous one to develop the synthesis of solid solutions

between Cu2ZnSnS4 and Cu2CoSnS4, generating compounds with the formula

Cu2(Zn1-xCox)SnS4. Thin films were prepared from aqueous dispersions of these

Cu2(Zn1-xCox)SnS4 compounds, and the stability of the films upon annealing in

sulfur atmosphere was analyzed. The fourth chapter describes the development

of a milder etching solution based on a mixture of ethylenediamine and 2-

mercaptoethanol to eliminate undesired copper sulfide (Cu2-XS) or copper selenide

(Cu2-XS) phases from CZTS thin films. The development of this etching solution

represents a viable alternative to the widely used etching methods based on

potassium cyanide (KCN) use. The fifth chapter extends the application of the

etching solution to etch other common undesired phases such as ZnS, SnS2, and

CuxZnySnz, and a possible mechanism is proposed for the etching process based

in a Lewis acid-base reaction between ethylenediamine and 2-mercaptoethanol.

vi

TABLE OF CONTENTS

ACKNOWLEDGMENTS…………………………………………………………………………………………………. i

DEDICATION……………………………………………………………………………………………………………….. iv

ABSTRACT…………………………………………………………………………………………………………………… v

TABLE OF CONTENTS…………………………………………………………………………………………………... vi

LIST OF TABLES……………………………………………………………………………………………………......... ix

LIST OF FIGURES……………………………………………………………………………………………………..…… xi

CHAPTER 1……………………………………………………………………………………………………….…………….. 1

1. INTRODUCTION………………………………………………………………………………………………….. 1

1.1 Green Chemistry Definition and Historical Perspective………………………………… 1

1.2 Green Chemistry and Nanotechnology………………………………………………………… 4

1.3 Strategies to make a nanoparticle synthetic process greener……………………..... 6

1.4 Successful Examples of the Application of Green Chemistry Principles in the

Synthesis of Transition Metal Chalcogenides…………………………………………………......

10

1.4.1 Wet-Chemical Synthesis of Multinary Transition Metal Chalcogenides… 10

CHAPTER 2………………………………………………………………………………………………………………….….. 20

2. Cu2ZnSnS4 (CZTS) phase control by changing the initial oxidation states of the

cations and sulfur excess source in microwave solvothermal synthesis…………......

20

2.1 Introduction…………………………………………………………………………………………………. 20

2.2 Experimental Procedure………………………………………………………………………………. 24

2.2.1 Materials…………………………………………………………………………………………….. 24

2.2.2 CZTS Nanoparticles Synthesis………………………………………………………........ 25

2.2.3 Experiments varying Copper and Tin Oxidation States………………………… 26

2.2.4 Experiments varying the Copper and Tin Oxidation States…….……………. 26

2.2.5 Product labelling convention………………………………………………………………. 27

2.2.6 Characterization…………………………………………………………………………………. 27

2.3 Results and Discussion…………………………………………………………………………………. 30

2.3.1 Effect of S:M ratio on the phase of the nanocrystals synthesized using

Cu(II) and Sn(II) reagents………………………………………………………………………………

30

2.3.2 Effect of Cu and Sn initial oxidation states……………………………………........ 34

2.3.3 Effect of excess sulfur source …….………………………………………………………. 35

2.3.4 On the mechanism and precursors to CZTS formation ……………………..... 37

2.4 Conclusions and Conclusions………….……………………………………………………………. 46

CHAPTER 3…………………………………………………………………………………….……………………………….. 50

3. Green Synthesis of Cu2(Zn1-x,Cox)SnS4 Nanocrystals and Formation of

Polycrystalline Thin Films from Their Aqueous Dispersions………………………………...

50

3.1 Introduction……………………………………………………………………………………..…………. 50

3.2 Experimental Procedure………………………………………………………………………………. 53

vii

3.2.1 Materials………………………………………………………………………………………….... 53

3.2.2 Synthesis of Cu2(Zn1-xCox)SnS4 nanocrystals……………………………………….. 53

3.2.3 Formation of Polycrystalline films………………………………………………………. 54

3.2.4 Characterization…………………………………………………………………………………. 55

3.3 Results and Discussion…………………………………………………………………………………. 58

3.3.1 Cu2(Zn1-xCox)SnS4 nanocrystals……………………………………………………………. 58

3.3.2 Polycrystalline Cu2(Zn1-xCox)SnS4 thin films………………………………………….. 66

3.4 Conclusions………………………………………………………………………………………............ 73

CHAPTER 4……………………………………………………………………………………………………………………… 74

4. Selective removal of Cu2−x(S,Se) phases from Cu2ZnSn(S,Se)4 thin films………….... 74

4.1 Introduction…………………………………………………………………………………………......... 74

4.2 Experimental Procedure………………………………………………………………………………. 76

4.2.1 Materials…………………………………………………………………………………………… 76

4.2.2 Synthesis of Cu2−xS, wurtzite CZTS, and kesterite CZTS nanocrystals……. 77

4.2.3 Precursor coating……………………………………………………………………………..... 78

4.2.4 Cu2−x(S,Se) and kesterite CZTSSe thin films………………………………………… 79

4.2.5 Etching………………………………………………………………………………………………. 79

4.2.6 Characterization…………………………………………………………………………………. 80

4.3 Results and Discussion………………………………………………………………………………… 81

4.4 Conclusions…………………………………………………………………………………………………. 95

CHAPTER 5…………………………………………………………………………………………………………………… 97

5. Etching Mechanism of Cu-, Zn-, and Sn-Containing Sulfides in Ethylenediamine

and 2-Mercaptoethanol Mixture…………………………………………………………………………

97

5.1 Introduction…………………………………………………………………………………………………. 97

5.2 Experimental Procedures…………………………………………………………………………..... 99

5.2.1 Materials…………………………………………………………………………………………….. 99

5.2.2 Synthesis of Cu2-xS, ZnS, SnS2, CuxSnySz, Cu2ZnSnS4 nanocrystals………… 100

5.2.3 Preparation of heterogeneous Cu2ZnSnS4 thin films from nanocrystal

coatings ……………………………………………………………….…………………………………..….

1 101

5.2.4 Preparation of etching solutions and selective etching procedure………. 1 102

5.2.5 Characterization………………………………………………………………………………… 103

5.3 Results and Discussion……………………………………………………………………………….... 105

5.3.1 Etching Cu-, Zn-, and Sn-binary sulfides nanocrystals………………………… 105

5.3.2 Etching CTS and CZTS nanocrystals…………………………………………………….. 111

5.3.3 Explaining the inability to etch SnS2 nanocrystals ……………..………………. 112

5.3.4 Selective etching SnS2, ZnS, CTS from Cu2ZnSnS4 thin films after

annealing in sulfur…………………………………………………………………………………........

114

5.4 Conclusions……………………………………………………………………………………………….... 1 116

BIBLIOGRAPHY……………………………………………………………………………………………………………….. 118

References Chapter 1……………………………………………………………………………………………….. 118

References Chapter 2……………………………………………………………………………………………….. 124

viii

References Chapter 3……………………………………………………………………………………………….. 128

References Chapter 4……………………………………………………………………………………………….. 134

References Chapter 5………………………………………………………………………………………………. 138

APPENDIX A……………………………………………………………………………………………………………………. 142

A1 Rietveld Refinement Procedure….…………………………………………………………….…………. 142

A2 Fraction of wurtzite and kesterite phases in the product synthesized at 160 °C

using various S:M ratios …………………………………………………………..………………………………

144

A3 Crystallite size estimate using the Scherrer equation for

Cu(II)_Sn(II)_Tu_S:M_160 °C………………………………………………………………………………….....

145

A4 STEM-HAADF Elemental Mapping of the Cu(II)_Sn(II)_Tu_1.9_160 °C and

Cu(II)_Sn(II)_Tu_6.2_160 °C Samples………………………………………………………………………..

145

A5 Elemental Composition of Cu(N)_Sn(L)_Tu_S:M_160 °C……………………………………... 146

A6 Microwave Synthesis of CZTS using other excess sulfur sources………………………… 147

A7 Infrared Spectroscopy of the Samples Cu(II)_Sn(II)_Tu_S:M_160 °C……………………. 147

A8 Effect of S:M ratio and Sn oxidation state, and excess sulfur source on phase

composition………………………………………………………………………………………………………………

149

A9 Synthesis of Zn-Sn intermediates at room temperature………………………………………. 1 151

A10 Experiments varying Sn oxidation state at room temperature…………………………… 155

A11 Effect of temperature on the morphology of products Cu(II)_Sn(II)_Tu_1.9_T

and Cu(II)_Sn(II)_Tu_6.2_T………………………………………………………………………………………..

156

A12 Elemental Analysis of the Product Cu(II)_Sn(IV)_TGacid_1.9_100 °C…………………. 160

A13 Influence of the water on the CZTS phase composition……………………………………… 160

APPENDIX B..……………………………………………………………………………………………………………….... 163

B1 Measurement and Control of Co fraction, x in Nanocrystals………………………………… 163

B2 Rietveld Refinement and XRD Simulation Details………………………………………………… 164

B3 XRD patterns from Cu2(Zn1-xCox)SnS4 Nanocrystals as a function of x on an

expanded scale………………………………………………………………………………………………………….

165

APPENDIX C..………………………………………………………………………………………………………………….. 166

ix

LIST OF TABLES

Table 1.1 Comparison between molecular chemistry and materials chemistry context for different concepts.

5

Table 1.2 Band Gap energy of different binary, ternary, and quaternary transition metal chalcogenides.

11

Table 2.1. Summary of the dominant phases formed using different combinations of tin and excess sulfur sources and with low and high S:M ratios. TGAcid is thioglycolic acid and MPAcid is 3-mercaptopropionic acid.

37

Table 3.1 Lattice parameters of the Cu2(Zn1-xCox)SnS4 phases.

61

Table A1. Wyckoff position of the Cu, Zn, Sn, and S used to simulate wurtzite CZTS, space group P63mc (#186).

142

Table A2. Fractions (in %) of wurtzite and kesterite phases and Goodness of Fit (GoF) for triplicate trials of the product synthesized at 160 °C using various S:M ratios Cu(II)_Sn(II)_Tu_S:M_160 °C.

144

Table A3. Crystallite size estimates for Cu(II)_Sn(II)_Tu_S:M_160 °C. Sizes were calculated using the Scherrer equation, a shape factor of 0.9 and the peak at 2θ≈32°, (112) of the wurtzite and the (002) of the kesterite phases. Deviation from a spherical shape and differences in the shapes of kesterite and wurtzite nanocrystals in multiphase mixtures may change these estimates slightly.

145

Table A4. Elemental composition determined from SEM-EDS for nanocrystal products synthesized using Cu and Sn precursors with different oxidation states and S:M ratios, i.e., Cu(N)_Sn(L)_Tu_S:M, for N= I or II, L= II or IV, and S:M = 1.9 or 6.2. All nanocrystals were synthesized at 160 °C. Compositions are given with respect to S, where the sulfur EDS intensity was normalized to 4.

146

x

Table A5. Fraction (in %) of the phases estimated from Rietveld refinement for the samples Cu(II)_Sn(L)_XCS_S:M_160 °C, where L = II or IV, XCS = Tu, Cyst, TGacid, or MCPacid, and S:M = 1.9 or 3.6.

150

Table A6. Composition of the product synthesized with Cu(II) and Sn(IV) reagents, thiourea, and S:M=1.9 at 25 °C (e.g., Cu(II)_Sn(IV)_Tu_1.9_25 °C). Elemental values are in atom %.

156

Table A7. TEM-EDS analysis of the product Cu(II)_Sn(IV)_TGacid_1.9_100 °C. Elemental composition is in atom %. Composition from three locations are shown.

160

Table A8. Correlation between hydration water and phase composition of different CZTS products

161

Table B1. Wyckoff positions and occupational parameters (sof) used in the Rietveld refinement and XRD simulations.

164

Table C1. Elemental ratio for (a) CuxSnySz and (b) CZTS nanocrystals obtained by TEM EDS and XPS characterization before etching. Elemental ratio Data from TEM EDS were collected from different 5 points.

172

xi

LIST OF FIGURES

Figure 1.1 Number of publications per year for the term Green Chemistry, from 1990 to 2016.

3

Figure 1.2 Unit cell of wurtzite and kesterite CZTS.

13

Figure 2.1 (a) XRD patterns from nanocrystals synthesized using Cu2+, and Sn2+ at 160 °C using different S:M ratios. Standard XRD patterns for wurtzite (simulated) and kesterite (PDF# 00-0026-0575) CZTS are shown the bottom. Peaks labeled as Ω are consistent with Cu3SnS4 (PDF 00-036-0218), α with Cu4SnS4 (PDF 00-029-0584), and х with elemental sulfur (PDF 00-008-0247) (b) Phase composition estimated from Rietveld refinement analysis of the XRD patterns shown in (a). Error bars were determined from triplicate analyses.

31

Figure. 2.2 Raman spectra of the nanocrystals synthesized from Cu2+, and Sn2+ at 160 °C, using different S:M ratios.

32

Figure.2.3 (a) TEM, (b) SAED (c) HR-TEM images of the Cu(II)_Sn(II)_Tu_1.9_160 °C and d) TEM, (e) SAED (f) HR-TEM of the Cu(II)_Sn(II)_Tu_6.2_160 °C nanocrystals.

34

Figure 2.4 XRD patterns for the nanocrystals (a) Cu(I)_Sn(II)_Tu_S:M_160 °C (bottom), and Cu(I)_Sn(IV)_Tu_S:M_160 °C (top) (b) Cu(II)_Sn(II)_Tu_S:M_160 °C (bottom), and Cu(II)_Sn(IV)_Tu_S:M_160 °C (top).

35

Figure 2.5 XRD patterns of the CZTS samples using different sources of sulfur excess, Sn oxidation states, and S:M=1.9 and 3.6. Sulfur excess sources: a) thiourea, b) L-cysteine, c) thioglycolic acid, d) 3-mercaptopropionic acid.

36

Figure 2.6 XRD pattern from the nanocrystals synthesized at different temperatures, T from Sn(II) source: (a) Cu(II)_Sn(II)_Tu_1.9_T, (b) Cu(II)_Sn(II)_Tu_6.2_T; • denotes the diffraction from the Zn-Sn intermediate.

39

xii

Figure 2.7 SEM images of the nanocrystals (a) Cu(II)_Sn(II)_Tu_1.9_25 °C, (b) Cu(II)_Sn(II)_Tu_1.9_160 °C, (c) Cu(II)_Sn(II)_Tu_6.2_25 °C, and (d) Cu(II)_Sn(II)_Tu_1.9_160 °C. (a) and (c) are the Zn-Sn glycolate precursor nanocrystals whereas (b) and (d) are the CZTS nanocrystals formed from these precursors. (See Appendix A Figure A10 for additional SEM images.)

43

Figure 2.8. XRD pattern of the samples Cu(II)_Sn(II)_TGacid_1.9_T and Cu(II)_Sn(II)_TGacid_1.9_T, with T between 25 and 160 °C. The sample Cu(II)_Sn(II)_TGacid_1.9_25 °C did not produce any solid. Peaks labelled as Δ are consistent with Cu2-xS (See also Table A7 in the Appendix A for the elemental analysis results).

46

Figure 2.9. a) Reaction scheme for S excess source containing NH2 group, according to the S:M variation, leading to kesterite or wurtzite CZTS; b) Reaction scheme for S excess source without NH2 group, leading mostly to wurtzite CZTS.

49

Figure 3.1 (a) XRD patterns, (b) lattice parameters, and (c) sizes of Cu2(Zn1-xCox)SnS4 nanocrystals as a function of Co fraction, x. Simulated XRD patterns of wurtzite Cu2ZnSnS4 and Cu2CoSnS4 are shown as stick patterns. Lines in (b) are linear extrapolation (Vegard’s law) between the lattice parameters of Cu2ZnSnS4 and Cu2CoSnS4. Nanocrystal sizes were obtained from the Scherrer equation using the measured XRD patterns. Co Kα emission was used for XRD.

60

Figure 3.2 (a) Raman spectra and (b) A1 mode peak positions of Cu2(Zn1- xCox)SnS4 nanocrystals as a function of x.

62

Figure 3.3 Representative TEM images of Cu2(Zn1-xCox)SnS4 nanocrystals with; (a) x = 0, (b) x = 0.25, (c) x = 0.4, (d) x = 0.6, (e) x = 0.75, and (f) x = 1, respectively. TEM samples were made by drop casting nanocrystal dispersion in methanol onto Ni carbon mesh grids.

64

Figure 3.4 Optical absorption spectra of Cu2(Zn1-xCox)SnS4 nanocrystals as a function of Co fraction, x. The absorption spectra were obtained from films drop cast from nanocrystal dispersions in methanol on soda lime glass substrates after compression. Air was used as a baseline.

66

xiii

Figure 3.5 XRD patterns from Cu2(Zn1-xCox)SnS4 polycrystalline thin films formed on Mo-coated soda lime glass substrates via thermal annealing in sulfur of Cu2(Zn1-xCox)SnS4 nanocrystal coatings. Annealing was conducted in 500 Torr of sulfur at 600 °C for 1 hour. 3.0 x 10-5 mol of Na was added to the annealing ampule as described in the experimental procedure. XRD patterns of kesterite Cu2ZnSnS4 (00-026-0575), kesterite Cu2CoSnS4 (00-026-0513), tetragonal Cu1.96S (00-012-0224), cubic Co0.24Zn0.76S (00- 047-1656), and cubic CoS2 (98-001-3473) are also shown for comparison. Co Kα emission was used for XRD.

68

Figure 3.6 (a-d) Raman spectra and (e) A1 mode peak positions of Cu2(Zn1-xCox)SnS4 thin films formed on Mo coated soda lime glass substrates by thermal annealing of Cu2(Zn1-xCox)SnS4 nanocrystal coatings. A1 mode Raman peaks were collected from large grains in the Cu2(Zn1-xCox)SnS4 films as shown in (e). Annealing conditions were the same as for Figure 5.5. The Raman spectra for Cu2(Zn1-xCox)SnS4 thin films at x ≥ 0.6 (b,c,d) were collected from different grains, which were selected to show the different phases present.

70

Figure 3.7 Plan view and cross sectional FE-SEM images of the Cu2(Zn1- xCox)SnS4 thin films formed on Mo coated soda lime glass substrates by thermal annealing Cu2(Zn1-xCox)SnS4 nanocrystal coating; ((a) and (b)) x = 0, ((c) and (d)) x = 0.25, ((e) and (f)) x = 0.4, ((g) and (h)) x = 0.6, ((i) and (j)) x = 0.75, and ((k) and (l)) x = 1. Annealing conditions were the same as for Figure 3.5.

72

Figure 4.1 X-ray diffraction patterns for Cu2-xS (a), kesterite-CZTS and wurtzite-CZTS NCs (b) prepared by microwave assisted solvothermal process. X-ray diffraction patterns were collected from nanocrystals drop cast and dried on a glass substrate from aqueous dispersions at room temperature.

82

Figure 4.2 X-ray diffraction patterns for Cu2-xS (a), kesterite-CZTS and wurtzite-CZTS nanocrystals (b) prepared by microwave assisted solvothermal process. X-ray diffraction patterns were collected from nanocrystals drop cast and dried on a glass substrate from aqueous dispersions at room temperature.

84

xiv

Figure 4.3 Plan view FE-SEM images of films (a) before and (b) after etching in the etchant solution for 20 minutes. Higher magnification FE-SEM images are shown in the insets.

86

Figure 4.4 X-ray diffraction patterns from annealed films before and after etching in the etchant solution for 20 minutes. CZTS and copper sulfide diffraction patterns are shown for comparison.

87

Figure 4.5 Raman scattering maps from annealed CZTS films (a & b) before (c & d) after etching in e solution. Raman scattering images from (a & c) Cu2−xS and (b & d) CZTS were produced by selectively filtering scattering at 475 cm−1 and 338 cm−1, respectively. Representative Raman spectra of CZTSe films (e) before and (f) after etching. In (e) and (f), spectra from two different regions are shown. The Raman spectra labelled Regions (i) and (ii) in (e) are representative of spectra collected from dark (CZTS) and bright orange (Cu2−xS) regions in (a), respectively. The Raman spectra labelled regions (i) and (ii) in (f) are representative of spectra collected from any location on the surface after etching.

90

Figure 4.6 XRD patterns for CZTSe thin film containing a small fraction of Cu2−xSe before and after etching in the etchant solution for 20 minutes.

91

Figure 4.7 Raman scattering maps from annealed CZTSe films (a & b) before (c & d) after etching in etchant solution. Raman scattering images from (a & c) Cu2−xSe and (b & d) CZTSe were produced by selectively filtering scattering at 260 cm−1 and 196 cm−1, respectively. Representative Raman spectra of CZTSe films (e) before and (f) after etching. In (e) and (f), spectra from two different regions are shown. The Raman spectra labelled Regions (i) and (ii) in (e) are representative of spectra collected from dark (CZTSe) and bright orange (Cu2−xSe) regions in (a), respectively. The Raman spectra labelled Regions (i) and (ii) in (f) are representative of spectra collected from any location on the surface after etching. CuSe, Cu2Se and CZTSe diffraction patterns are also shown for comparison.

92

Figure 4.8 Plane view FE-SEM images of films (a) before and (b) after etching in etchant solution for 20 minutes. Higher magnification FE-SEM images are shown in the insets. Hexagonal crystals protruding from the surface are copper selenide grains while the remaining large crystals are CZTSSe grains. These copper selenide grains are removed selectively when the film is immersed the etchant solution for 20 minutes.

95

xv

Figure 5.1 Photographs of 6 mg of Cu2-xS, ZnS, SnS2, CTS, and CZTS nanocrystals before (top row) and at specific time intervals after introduction of 1 mL of etching solution (one part by volume 2-mercaptoethanol with four parts per volume ethylenediamine).

105

Figure 5.2 FTIR spectra of (a) ethylenedimine, (b) 2-mercaptoethanol, (c) Etching Solution (20% 2 mercaptoethanol, 80% ethylenediamine, by volume), (d) CZTS kept for 2 h in neat 2-mercaptoethanol, (e) CZTS kept for 2 h in neat ethylenediamine, and (f) CZTS kept for 2 h in etching solution. All the spectra were collected in transmittance mode.

107

Figure 5.3. Conductivity versus the ethylenediamine volume percent (a), and mole ratio between HS/NH2 groups (b). The point having ethylenediamine volume percent equal to 100% has an undefined NH2/HS mol ratio.

108

Figure 5.4. H1-NMR spectra of the neat ethylenediamine, 2-mercaptoethanol and mixtures in different proportions of these two solvents (top). Possible species produced during Lewis acid-base reaction between ethylenediamine and 2-mercaptoethanol (bottom).

110

Figure 5.5. X-ray diffraction patterns of as-synthesized and post-annealed SnS2 nanocrystals before and after etching process (left). Right photo images for visual appearance of annealed SnS2 nanocrystals as function of etching times. Annealing conditions are at 600 °C for 1 hour under 500 Torr of sulfur partial pressure. SnS2-ethyelediamine intercalated compound diffraction pattern was collected from Ref. 40 CoKα used as a X-ray sources.

114

Figure 5.6. Raman scattering maps from annealed thin films from mixture dispersions of (i) CZTS and SnS2, (ii) CZTS and ZnS, (iii) CZTS and CTS before and after etching (rows A and B) and the representative Raman spectra of annealed films collected from different regions shown in maps (rows C and D). Annealing conditions are at 600 °C for 1 hour under 500 Torr of sulfur partial pressure.

116

Figure A1. (a) TEM image (same as Figure 2.3a) of Cu(II)_Sn(II)_Tu_1.9_160 °C nanocrystals with boxes around the oblate

145

xvi

nanocrystals. (b) Higher magnification TEM images from another region of the same sample, where the oblate nanocrystals can be seen more clearly. Figure A2. STEM-HAADF elemental map for a) Cu(II)_Sn(II)_Tu_1.9_160 °C and b) Cu(II)_Sn(II)_Tu_6.2_160 °C

146

Figure A3. XRD patterns from products synthesized at 160 °C without thiourea, varying the tin oxidation state, and using either L-cysteine or thioglycolic acid as the only sulfur source.

147

Figure A4. FTIR transmission spectra of the Cu(II)_Sn(II)_Tu_1.9_160 °C (red), and Cu(II)_Sn(II)_Tu_6.2_160 °C (blue) nanocrystals. Reference spectra for ethylene glycol (Fisher Scientific), and thiourea are also shown for comparison.

149

Figure A5. a) XRD pattern from Cu(II)_Sn(II)_Tu_1.9_25 °C product and XRD patterns for coordination compounds of Zn and Sn that matched the best. b) Molecular Structure of the compounds possibly present in the Zn-Sn intermediate.

151

Figure A6. Elemental composition determined using SEM-EDS for a) Cu(II)_Sn(II)_Tu_1.9_T and b) Cu(II)_Sn(II)_Tu_6.2_T as a function of temperature, T, from 25 to 160 °C.

152

Figure A7. XRD patterns from products Cu(II)_Sn(L)_XCS_1.9_25 °C, where L = II or IV, and XCS = thiourea (Tu), L-cysteine (cyst), 3-mercaptopropionic acid (MCPacid), thioglycolic acid (TGacid).

154

Figure A8. SEM image of the Zn-Sn intermediate formed by the synthesis where L-cysteine was used as the excess sulfur source (e.g., Cu(II)_Sn(II)_cyst_1.9_25 °C).

155

Figure A9. XRD pattern from products synthesized with Cu(II) and Sn(IV) reagents at 25 °C with S:M=1.9 and S:M=6.2 (i.e., Cu(II)_Sn(IV)_Tu_1.9_25 °C and Cu(II)_Sn(IV)_Tu_6.2_25 °C). This data shows that only elemental sulfur forms at room temperature when Sn(IV) is used in the synthesis.

156

xvii

Figure A10. SEM images from samples: (a) Cu(I)_Sn(II)_Tu_1.9_160 °C (b) Cu(I)_Sn(II)_Tu_6.2_160 °C (c) Cu(II)_Sn(II)_Tu_1.9_160 °C (d) Cu(II)_Sn(II)_Tu_6.2_160 °C (e) Cu(I)_Sn(IV)_Tu_1.9_160 °C (f) Cu(I)_Sn(IV)_Tu_6.2_160 °C (g) Cu(II)_Sn(IV)_1.9_Tu_160 °C (h) Cu(II)_Sn(IV)_Tu_6.2_160 °C.

158

Figure A11. SEM images from products a) Cu(II)_Sn(II)_Tu_1.9_75 °C, b) Cu(II)_Sn(II)_Tu_1.9_100 °C, c) Cu(II)_Sn(II)_Tu_1.9_130 °C, d) Cu(II)_Sn(II)_Tu_6.2_75 °C, e) Cu(II)_Sn(II)_Tu_6.2_100 °C, f) .Cu(II)_Sn(II)_Tu_1.9_130 °C.

159

Figure A12. XRD pattern of the products Cu(II)_Sn(II)_Tu_1.9_160C, with (top) and without addition of 5 mmol of water (bottom).

162

Figure B1. Comparison of the nominal Co fraction, x in Cu2(Zn1-x,Cox)SnS4, with the Co fraction, X, measured using ICP-MS (blue triangles) and SEM-EDS (red squares). The measured cobalt fraction, X is defined as X=CCo/(CZn+CCo), where Ci is the concentration of species i (Co or Zn). The nominal Co fraction, x, is the Co fraction in the precursor solutions calculated from the ratio of the moles of CoAc2 to the sum of the moles of CoAc2 and ZnAc2 (i.e., [CoAc2]/[CoAc2 + ZnAc2]). The elemental composition determined by ICP-MS and SEM-EDS reveal the CCo/(CZn+CCo) ratio is very close to expected x value in Cu2(Zn1-

xCox)SnS4, as shown in the Figure B1.

163

Figure B2. XRD patterns from Cu2(Zn1-xCox)SnS4 nanocrystals as a function of the nominal Co fraction, x on an expanded scale. Simulated XRD patterns of wurtzite Cu2ZnSnS4 and Cu2CoSnS4 are shown as stick patterns.

165

Figure C1. X-ray diffraction patterns for Cu2-xS, ZnS, SnS2, CuxSnySz, Cu2ZnSnS4 nanocrystals. X-ray diffraction patterns were collected from nanocrystals drop cast and dried from aqueous dispersion at room temperature. The bottom panels show the expected diffraction patterns for kesteite Cu2ZnSnS4 (ICDD : 00-026-0575), Cu2SnS3 (ICDD : 00-027-0198), Cu4SnS4 (ICDD : 00-027-0296), SnS2 (ICDD : 00-023-0677), ZnS (ICCD : 00-005-0566), and CuS (ICDD : 00-006-0464), respectively.

167

Figure C2. Qualitative etching rate of different proportions by volume of the ethylenediamine-2-mercaptoethanol mixture. In each experiment were added 7 mg of kesterite CZTS and 1 mL .of etching mixture.

168

xviii

Figure C3. Photographs for reaction between metal salts and neat ethylenediamine or 2-mercaptoethanol. For each experiment 17 mg of Cu(Ac)2.H2O, 15 mg of Zn(Ac)2.2H2O, 25 mg of SnCl4.5H2O were added to 4 mL ethylenediamine or 1 mL of 2-mercaptoethanol.

169

Figure C4. X-ray diffraction patterns for the samples shown tin the top part of the figure C3. The powder samples formed from the reaction between Cu(I) Acetate and ethylenediamine and Sn(IV) Chloride and ethylenediamine are amorphous. The powder formed from the reaction between Zn(II) Acetate and ethylenediamine is crystalline, and present peaks very similar to what is observed for [Zn(en)3](NO3)2 and [Zn(en)3]F2. These similarities lead us to conclude that the powder observed from the reaction between Zn(II) Acetate and ethylenediamine is [Zn(en)3]Acetate2. The standard pattern for [Zn(en)3](NO3)2 and [Zn(en)3]F2 were obtained from Cambridge Structural Database (CSD), with the respective reference codes: RAVJAZ for [Zn(en)3](NO3)2 and VAPKAY for [Zn(en)3]F2. On the best of our knowledge, there is no reference pattern for [Zn(en)3]Acetate2 in CSD database.

170

Figure C5. Scheme of (a) Lewis acid-base reaction when ethylenediamine and 2-mercaptoethanol are mixed and (b) etching mechanism of metal-sulfide nanocrystals.

171

Figure C6. Elemental ratio for (a) CZTS and (b) CTS nanocrystals obtained by TEM EDS as function of etching times.

172

Figure C7. Scheme for etching mechanism regarding CuxSnySz and Cu2ZnSnS4 nanocrystals.

173

Figure C8. XRD pattern of the CZTS thin film containing SnS2 as impurity (shown on Figure 6) before (bottom pattern) and after (top pattern) etching. The absence of the SnS2 peaks around 24, 27, and 39 ° on the pattern after etching confirms the removal of SnS2, as its shown by Raman spectroscopy on Figure 5.6. Although Figures 5.1, and 5.5 indicate the inability of the etching solution to etch SnS2, in the thin film case, this removal may be explained by delamination of SnS2 from the underlying material as a result if intercalation with ethylenediamine.

174

1

CHAPTER 1

1. INTRODUCTION

1.1 - Green Chemistry Definition and Historical Perspective

The term Green Chemistry refers to the strategies for the production and

use of safer chemical products as replacements for hazardous substances. In this

sense, hazardous substances can be defined in a broad way as any substance

representing any physical, such as injury as result of short or long term exposure;

environmental, such as water or air pollution; or toxicological risks, such as

mutations or cancer.1

Although the search for safer reagents and solvents has been an ongoing

process in modern chemistry, the term Green Chemistry was coined in the

beginning of the 1990’s decade, soon after the establishment of the Pollution

Prevention Act of 1990.2 Among different proposals, this Act included source

reduction as desirable in relation to waste management and pollution control and

more cost-effective production and operation procedures to reduce or prevent

pollution generated by industries.

In 1998, Anastas and Warner published the book Green Chemistry: Theory

and Practice. This book presents for the first time the 12 Principles of Green

Chemistry,3 which serve as guidelines for good practices regarding minimization

of chemical waste production, mitigation of harmful or hard to treat byproducts,

atom economy, development of materials with reasonable degradation period after

the end of their lifecycles, and search for safer chemical sources, renewable

feedstocks, and energy efficient processes. The 12 Principles of Green Chemistry

2

function like an instruction manual for those professionals willing to develop

products and processes more aligned with the Green Chemistry concept. These

Principles also have contributed to the popularization of Green Chemistry, since

they work as a concise and accessible consulting resource. As the 12 Principles

of Green Chemistry will be continuously recalled throughout this chapter, they are

presented here to provide a quick reference to the reader.3

Principle 1 – Prevent the Waste

Principle 2 – Atom Economy

Principle 3 – Less Hazardous Chemical Synthesis

Principle 4 – Designing Safer Chemicals

Principle 5 – Safer Solvents and Auxiliaries

Principle 6 – Design for Energy Efficiency

Principle 7 – Use of Renewable Feedstocks

Principle 8 – Reduce Derivatives

Principle 9 – Catalyst reagents are preferred over stoichiometric ones

Principle 10 – Design for Degradation

Principle 11 – Real-time Analysis for Pollution Prevention

Principle 12 – Inherently Safer Chemicals for Accident Prevention

Since the proposal of the Pollution Prevention Act, the Green Chemistry

field has grown substantially in the scientific literature. A search on the database

Web of Science revealed that the number of scientific articles published having the

term Green Chemistry increased from 1 in 1990 to 1,305 in 2016, with an

3

accumulated total of 10,964 papers having the term Green Chemistry, either in its

content or title, published in this 26 year period.4

Another sign of the increasing interest in Green Chemistry is the steady

increase in the number of journals dedicated to communicate the Green Chemistry

scientific advances. For instance, Green Chemistry, published by Royal Society of

Chemistry (RSC), was created in 1999;5 Environmental Chemistry Letters,

published by Springer, was created in 2003;6 ACS Sustainable Chemistry &

Engineering, published by American Chemical Society (ACS), was created in

2013;7 and Journal of Environmental Chemical Engineering, published by Elsevier,

was also created in 2013.8

1990 1995 2000 2005 2010 20150

200

400

600

800

1000

1200

1400

Years

Nu

mb

er

of

Pu

bli

ca

tio

ns

Figure 1.1 Number of publications per year for the term Green Chemistry, from 1990 to 2016.

4

Green Chemistry is not a field that has remained confined within the limits

of the academic realm. Instead, many initiatives have been created over the last

27 years having Green Chemistry as underlying principle. In the U.S., the US

Presidential Green Chemistry Challenge Award, established in 1995 by the

Environmental and Protection Agency (EPA), has been promoting the

advancement of Green Chemistry by awarding innovative solutions from academe,

industry, and small businesses.9 A similar award called the European Sustainable

Chemistry Award has been granted in Europe by the European Chemical Sciences

since 2010.10

Green Chemistry has been expanding to address challenges arising from

new developments in both educational and technological fields. Some examples

of Green Chemistry broad scope can be noticed, for instance, in waste valorization

and use in further applications, green development of catalysts for renewable

energy technologies based on splitting of water molecule, the advent of the

nanotechnology and analysis of the impact of materials in the nanoscale while

interacting with soil, atmosphere, water, and biological systems.11–13

1.2 - Green Chemistry and Nanotechnology

Nanotechnology is defined as the development of materials with nanometer

dimensions (10-9 nm), and such materials often have surprising properties (i.e.

optical, mechanical, thermal, electrical) that are not observed for these materials

when in the macro or bulk scale.

5

Many of the terms and parameters related to Green Chemistry were defined

considering a molecular structure as a model to assess safety and toxicological

properties. While dealing with nanomaterials, besides molecular structure, other

factors such as being crystalline or amorphous, crystalline structure, surface area,

particle size, porosity, and so on can play substantial roles regarding how a

nanomaterial should be evaluated regarding its production, life cycle, toxicity, and

disposal. Recently, Hutchison14 published a paper presenting a comparison

between the materials context and the molecular context, which are presented with

some adaptation in table 1.1

Concept Molecular Context Materials Context

Composition Defined by molecular formula

Core and surface composition difficult to

define; may vary according to sample

shape and size Size/shape Defined molecular

structure and shape Often a mixture of sizes and shapes, dependent

on synthetic method

Dispersity Single and continuous composition and structure

Characterized by distributions of

composition and structural features

Purity Purification procedure is intimately related to

molecular structure (i.e. chromatography)

Small molecule impurities coming from

surface coating or unreacted precursors significantly influence

properties

Table 1.1 Comparison between molecular chemistry and materials chemistry context for

different concepts.

6

Toxicity Easier to assess based on molecular structure

May be inherent to the composition, but also

can be related to particle size, shape, or surface

coating In the next section, some strategies used to make a synthetic procedure

greener are presented.

1.3 - Strategies to make a nanoparticle synthetic process greener

Any change in the synthetic process that eliminates or replaces a hazardous

reagent or solvent,15 or is consistent with one of the 12 Principles of Green

Chemistry will likely result in an overall process that is more environmentally

friendly and less hazardous. Also, it is important to point out that the changes

necessary to make the process greener must not compromise the quality of the

final product. Green Chemistry, when successfully implemented, results in the

green production of high performance products. If performance is compromised,

then, the process does not yield a useful product.

There are many opportunities to make the synthesis of transition metal

chalcogenides greener. Hutchison summarized some of these opportunities for the

synthesis of gold nanoparticles stabilized with thiol groups.14 Due to the similarity

between the steps involved in the wet-chemical synthesis of gold and transition

metal chalcogenide nanoparticles, the opportunities summarized by Hutchison are

completely transferable to the synthesis of transition metal chalcogenides. Hence,

based on that, we outline the opportunities to green up the synthesis of transition

7

metal chalcogenides alongside the Green Chemistry Principles addressed by

making the change.

Opportunity 1 – Safer reagents and solvents

The selection of safer reagents and solvents that are unsuitable for

producing high quality materials represents a waste of time and resources. Thus,

the best course of action is to first examine results from related work in order to

reasonably predict whether a reaction or procedure will be successful using the

greener precursors and solvents. Also, working at small scales in the initial stages

can represent an economy of time and resources. If the procedure did not work

well in small scale, then one would not proceed to a larger scale procedure. Finally,

careful examination of the safety materials associated with each chemical is crucial

for preventing problems arising from the combination of incompatible materials or

the production of toxic byproducts. These strategies specifically address Green

Chemistry Principles 1-5 and 12.

Opportunity 2 – Use more efficient energy input sources

The wet-chemical synthesis of transition metal chalcogenides requires

some source of energy input, which is often provided by heating the solution

containing the starting materials to temperatures above 200 °C. Often, this heating

procedure is carried out using a reflux apparatus, which requires the consumption

of many liters of water to cool the reflux column. Alternatives to refluxing include

reactions assisted by microwaves or ultrasound.16 Furthermore, procedures that

enable the synthesis at lower temperature or even at room temperature represent

8

a greener process. The use of more efficient energy input sources complies with

Green Chemistry Principles number 1 and 6.

Opportunity 3 – Eliminate or minimize byproducts

The reduction or elimination of byproducts can mean little to no post-

synthesis purification is required. Indeed, the separation of the desired product

from the reaction medium as well as from the undesired byproducts often

represents the most waste generating step.

Opportunity 4 – Avoid using unnecessary additives and steps

In the synthesis of transition metal chalcogenides, it is common to use

capping agents, which are often surfactants, to obtain a certain size and

anisotropic shape for the nanoparticles. In many cases surfactants, are necessary

to obtain a particular anisotropic shape. However, in some cases, the growth can

be controlled by the solvent, by varying the amount of a certain starting material,

or adjusting other parameters like temperature, pH, or ionic strength. Avoiding

unnecessary additives means less post-synthesis purification is required and

fewer reagents are required overall.

Nanoparticle synthesis methods, commonly produce nanoparticles in some

non-polar solvent. To use these nanoparticles for some application often requires

dispersion in a polar solvent. When this happens, it is necessary to replace the

capping agent that makes the particle dispersible in the non-polar solvent with

another capping agent that makes particle dispersible in a polar solvent. This

9

ligand exchange procedure consumes time and additional solvent and reagents.

In many cases, ligand exchange can be avoided by simply choosing a synthetic

route that yields the nanoparticles with surface chemistry that is suitable for the

final application. Avoiding unneeded additives and unnecessary ligand exchange

steps directly address Principles number 2, 5, and 6.

Opportunity 5 – Greener purification procedures

Commonly employed purification procedures include washing nanoparticles

with a solvent that can solubilize only the byproducts; size-exclusion

chromatography, and dialysis, where products can be separated from the

byproducts due to difference in size; and evaporation, when the byproducts are

volatile and the nanoparticles are not. All these procedures require the use of

additional reagents, particularly solvents, which makes the purification procedure

one of the most difficult steps to green up. An ideal synthetic procedure will

produce the desired product in both high yield and high purity. Indeed, even trace

impurities can drastically compromise the performance of devices. In order to

reduce the total solvent required, in dialysis, for example, sequential dialysis

against smaller volumes of pure solvent as compared to one dialysis step against

a large volume of solvent will not only generate less waste but also generate a

product of higher purity. Judicious selection of solvent may also mean that the post

dialysis solvent could be recycled by passing through a purification column, for

example. Alternative purification procedures should be investigated in order to

10

select the method that will be the greenest possible but still yield the final product

with the required purity.

1.4 - Successful Examples of the Application of Green Chemistry Principles

in the Synthesis of Transition Metal Chalcogenides

Transition metal chalcogenides (TMC) constitute an important class of

materials and include different types of oxides, sulfides, selenides, and tellurides.

Increasing interest has been devoted to TMC due to their technological

applications, in different fields such as photocatalysis, sensors, solar cells,

supercapacitors, electrocatalysis, heterogeneous catalysis, and many other

applications.17 In view of the growing need and interest for this class of materials,

it is necessary to find ways to produce them through solution chemistry synthetic

routes that minimize the environmental impacts and health related risks. In order

to demonstrate that it is possible to produce TMC by greener routes, throughout

this section we will show some successful examples where changes in the

synthetic process yielded substantial improvements by decreasing environmental

impacts and biological risks.

1.4.1 - Wet-Chemical Synthesis of Multinary Transition Metal

Chalcogenides

Binary chalcogenides often contain cadmium as the metal cation, which is

a toxic and carcinogenic element for humans18,19 and an environmental concern

11

regarding its segregation and disposal.20 Additionally, binary chalcogenides have

band-gaps higher than 2.5 eV, which is suitable for the absorption of the UV

radiation of the electromagnetic spectra, whereas ternary and quaternary

chalcogenides typically have band-gaps ranging from 0.5 to 2.0 eV, which is

suitable for the absorption of the visible light. A more detailed description of the

band-gap values can be observed on table 1.2.21–29 Thus, the ternary and binary

chalcogenides are more suitable for applications relying on the absorption of solar

energy, such as absorption layers of thin film solar cells.30

Material Band Gap (eV)

CdS 2.60 CdSe 1.74 CdTe 1.48 ZnS 3.78

ZnSe 2.82 ZnTe 2.39

CuInS2 1.53 CuInSe2 1.05

Cu(In0.5Ga0.5)S2 1.60 Cu(In0.5Ga0.5)Se2 1.30

CuGaSe2 1.68 CuFeS2 0.60 AgInS2 1.87

Cu2SnS3 1.15 Cu2SnSe3 0.70 Cu3SnS4 1.22

Cu2ZnSnS4 1.50 Cu2ZnSnSe4 1.02 Cu2FeSnS4 1.28

Table 1.2 Band Gap energy of different binary, ternary, and quaternary transition metal

chalcogenides.

12

Cu2FeSnSe4 1.10 Cu2CdSnS4 1.40 Cu2CdSnSe4 0.96

This dissertation focuses primarily on the quaternary chalcogenide,

Cu2ZnSnS4, which will be hereafter referred to as CZTS. CZTS is comprised of

non-toxic and earth-abundant elements, features that align with Green Chemistry

Principles. In addition, its direct band-gap around 1.5 eV and high-absorption

coefficient (> 104 cm-1) make CZTS a suitable material for many applications,

including use as the absorber layer in thin film solar cells,30 counter-electrode

material in dye sensitized solar cells,31 visible-light photocatalyst for water pollutant

degradation,32 as well as use in heterostructures with Pt or Au for H2 evolution in

water splitting systems.33

CZTS can occur in five different crystalline structures: kesterite, stannite,

wurtzite-stannite, wurtzite-kesterite, orthorrombic CZTS, and primitive-mixed Cu-

Au (PMCA). At room temperature, the thermodynamically stable crystalline phase

is kesterite, which is only 2.8 meV/atom more stable than stannite.34 Kesterite and

stannite are both tetragonal structures, whereas, wurtzite-stannite and wurtzite-

kesterite are hexagonal structures. There are only few reliable reports about the

synthesis of the orthorrombic CZTS,35 whereas the successful synthesis of the

PMCA phase has not yet been reported and, to date, has been subject only of

theoretical studies, to the best of our knowledge.36–38

13

The first synthesis of CZTS dates from 1967, by Nitsche and co-workers,

where they prepared needle-like single-crystalline CZTS by heating to 1100 °C in

a sealed quartz ampoule containing stoichiometric amount of elemental Cu, Zn,

Sn, and S in the presence of gaseous iodide.39 Later, Ito and Nakazawa

determined the optical band gap and absorption coefficient for the material, in

1988.40 The interest in CZTS was low in the following decades but increased

starting in 2001, when Katagiri et al. reported the preparation of CZTS thin films

having dense microstructure and assembly of a solar cell using those films as

absorber layer, which resulted in a solar cell with efficiency of 2.49%.41

The wet-chemistry synthesis of CZTS was reported in 2010, when Mitzi

and co-workers used a hybrid solution-particle approach dissolving pre-formed

particles of Cu2S-S, SnSe-Se in hydrazine combined with the in situ formation of

Wurtzite Kesterite

Figure 1.2 Unit cell of wurtzite and kesterite CZTS.

14

ZnSe(N2H4) or ZnS(N2H4) particles, due to addition of Zn powder in the reaction

mixture. The resulting dispersion was successive deposited onto Mo coated soda-

lime glass and annealed at 540 °C, rendering dense CZTSSe thin films. By using

those films as absorber layer, they were able to assemble a solar cell with an

efficiency of 9.7%.42 Mitzi’s group continued working on the development and

improvement of the hydrazine based synthesis of CZTS and CZTSSe and obtained

the highest efficiency reported for a CZTS based thin film solar cell in 2014, which

was 12.6%.43 Although hydrazine is a versatile solvent enabling the dissolution of

many different cations at room temperature,44 its high toxicity, flammability,

pyrophoricity, and carcinogenicity make it a solvent hard to handle safely and

therefore unsuitable for large scale applications.45

Oleylamine (OAm) is a less hazardous and more environmentally friendly

alternative to hydrazine for the synthesis of CZTS-based materials. The first wet-

chemical synthesis of CZTS using oleylamine as a solvent was reported by Guo

et al. They used a hot-injection method in which a solution of copper

acetylacetonate, zinc acetylacetonate, tin (IV) bis(acetylacetonate) dibromide, and

elemental sulfur were injected in hot OAm (225 °C).46 This method produced CZTS

nanoparticles with diameter around 10-20 nm, and a thin film of these

nanoparticles was annealed in selenium vapor and then used as absorber layer in

a solar cell device producing an efficiency equal to 0.80%. This same procedure

has been successfully used to produce CZTS and CZTGS (Cu2Zn(Sn,Ge)S4) with

different particle size distributions and elemental ratios according to the reaction

time.47 Khare et al.48 reported the successful synthesis of CZTS quantum dots with

15

controlled size. Their results demonstrate that the crystallite size can be controlled

by the reaction temperature and reagents concentration, which was controlled by

adding different volumes of OAm and octadecene. The smallest particles (2 nm)

produced at 150 °C and volume of OAm equal to 3 mL, and the largest particles

(7 nm) produced at 175 °C, with 1.5 mL of OAm and 1.5 mL of octadecene. The

absorption spectra of the resulting materials demonstrate that CZTS particles

exhibit a strong quantum size effect in this size regime, with the smallest particles

exhibiting a band gap of 1.8 eV, and 1.5 eV for the largest ones.

For many applications, stable dispersions of particles in water are desirable.

CZTS produced using hot-injection in OAm method produces CZTS nanoparticles

that are capped with OAm and oleic acid molecules,48 making them readily

dispersible in organic solvents. Tosun et al.49 developed a ligand exchange

method that replaces the OAm and oleic acid molecules with S2- ions, rendering

readily dispersible in polar solvents, including water. Despite the success of the

ligand exchange procedure, it introduces the need for post-synthetic steps and use

of additional reagents. It would be preferable to prepare water dispersible CZTS

nanoparticles directly from the synthesis, without the need of additional post-

synthetic steps.

Hydrothermal synthesis can be a Green and viable option for CZTS

synthesis; however, it requires careful control of pH or the addition of complexing

agents, since Zn+2, Sn+2/+4 tend to easily form oxides when in contact with

water.50,51 Wang and co-workers52 prepared spherical CZTS nanoparticles

hydrothermally reaction at 200 °C for 24 h, by using Cu+2, Zn+2, and Sn+2 chlorides

16

and thioacetamide as the sulfur source. Their results demonstrate that the addition

of ammonia played a crucial role in the synthesis by controlling pH, the

thioacetamide decomposition rate, and coordinating with dissolved Cu+2 and Zn+2

cations. The absence of ammonia or its replacement with other bases like NaOH

produces a lot of undesired byproducts alongside CZTS.52 Additionally, the CZTS

nanoparticles produced in that work were used to deposit CZTS thin films on

electrodes used for photocatalytic evolution of H2, showing the applicability of

these nanoparticles produced via hydrothermal synthesis. Another successful

example of the hydrothermal synthesis of CZTS used citric acid as a complexing

agent, with equimolar amounts of citric acid and total metal cations added and

thiourea as the sulfur source in that paper, and the best reaction condition was 190

°C for 24 hours.53 Alternatively, ethylenediaminetetraacetic acid (EDTA) was used

as the complexing agent and L-cysteine as the sulfur source, slight excess of Zn+2,

and a metallic cation to sulfur mole ratio of 2.5, phase pure kesterite CZTS was

produced at 180 °C for at least 12 hours.54

Generally speaking, the successful hydrothermal synthesis of CZTS

requires relatively long reaction times (ca. 12-24 hours). To speed up the reaction

and reduce the overall energy required, microwave heating represents a promising

alternative to conventional heating. An important consideration is the capacity of a

given solvent to absorb microwaves, which is measured by a parameter called

dielectric loss factor (tan δ). In general, tan δ can be classified as high (tan δ >0.5),

medium (0.1 < tan δ < 0.5), and low (tan δ <0.1).55 Water has a tan δ of 0.123,

which is on the very low side of the medium range. Ethylene glycol, which does

17

not have any toxicity if not ingested,56 has a high tan δ of 1.350,55 which has led to

its use for the solvothermal synthesis of CZTS using microwave heating.

The initial report of microwave assisted synthesis of CZTS using ethylene

glycol as solvent was made by Flynn and co-workers in 2012,57 they prepared

CZTS nanoparticles by reacting the chlorides salts of Cu+, Zn+2, and Sn+4 and

thioacetamide in ethylene glycol, the at 190 °C for just 30 minutes. The product

nanoparticles were spherical and agglomerated in larger clusters. Those

nanoparticles dispersed in ethylene glycol were used to prepare absorber layer in

a solar cell, and the resulting cell had an efficiency of 0.25%.57 CZTS nanoflowers

were prepared from Cu+2 and Zn+2 acetates, SnCl4, and thiourea at 230 °C for 90

minutes.58 The addition of PVP was shown to not make any difference on the

synthesis of spherical agglomerated kesterite CZTS nanoparticles in a reaction

system constituted by CuCl2, ZnCl2, SnCl2, and Na2S, with temperatures equal to

160 or 180 °C, and reaction duration ranging from 15 to 60 minutes.59 Ghedya and

co-workers60 used a domestic microwave oven at 190 °C for 3 minutes to produce

ethylene glycol dispersions of CZTS from Cu+2, and Zn+2 acetates, SnCl2 and

thiourea, they did not present the XRD pattern of the product right after the

microwave heating, instead, they used those dispersions to prepare thin films

annealed in different atmospheres.

The formation of dense thin films generally requires deposition and

annealing steps. The deposition is generally carried out through procedures such

as spin coating,61 dip coating,62 drop cast,31,63 doctor blade,60,64 or aerosol jet

printing.65 Whereas the annealing requires temperatures as high as 200 to 600 °C,

18

and generally some type of non-oxidizing atmosphere, such as vacuum, argon or

nitrogen gas.60,66–71 These additional steps represent an increase on the cost and

energy input. Aiming to reduce the cost and energy associated with the deposition

and annealing steps, Knutson et al.72 developed a procedure where the CZTS

nanoparticles are synthesized, directly deposited onto conductive substrates (like

molybdenum coated glass), and annealed all inside the same microwave in

sequential steps. The authors dissolved in ethylene glycol Cu2+, Zn2+, Sn2+ salts,

thioglycolic acid and thiourea. In the same vial, there was submerged a substrate

coated in one side with molybdenum. The combination of the high dielectric loss

factor of the solvent with the high conductivity of the Mo contained in one side of

the substrate lead the CZTS particles to be selectively deposited directly onto the

substrate. The whole deposition process took only 15 minutes, and reached a

maximum temperature of 80 °C. This deposition could be carried out more than

once, in order to increase the film thickness. It is important to point out that all the

nanoparticles produced are deposited onto the conductive side of the substrate,

which could be confirmed by the optical transparency of the solvent when the

deposition step is completed. The film microstructure can be densified by carrying

out an annealing procedure, by replacing the ethylene glycol by octadecene, and

heating up at 160 °C for 30 minutes in the same microwave. This sequential

synthesis-deposition-annealing all in the same reaction apparatus demonstrates

how Green Chemistry Principles can be used as guideline to look for more energy

efficient and safer synthetic process implementation.

19

The microwave synthesis of CZTS using ethylene glycol as a solvent is a

promising method for producing high purity CZTS in a sustainable fashion. In the

next chapters of this dissertation, results from experiments using different sources

of sulfur excess, different oxidation states for copper and tin reagents,

temperature, and reaction duration will be presented, with particular focus on how

these experimental variables impact the phases produced. Then, the preparation

using microwave assisted solvothermal route of solid solutions of Cu2ZnSnS4 and

Cu2CoSnS4 (CZTS-CCTS) and annealing of their thin films will be discussed.

Finally, a new and milder process for etching impurity phases from CZTS thin films

will be presented. The results of dissolution experiments combined with

characterization of the etching solution and the solid materials before and after

exposure to the etching solution enable elucidation of why some phases etch away

rapidly while CZTS etches slowly.

20

CHAPTER 2

2. Cu2ZnSnS4 (CZTS) phase control by changing the initial oxidation states

of the cations and sulfur excess source in microwave solvothermal

synthesis

2.1 - Introduction

Thin film solar cells based on direct band gap semiconductors such

as Cu(In,Ga)Se2 (CIGS) and CdTe require a hundred times less absorber

material than silicon solar cells and are therefore expected to be less

expensive.1,2 Solar cells based on CIGS have already reached efficiencies

as high as 21.7%, comparable to silicon solar cells.3 However, CIGS-based

solar cells are difficult to implement in large scale because indium is scarce

and high in demand as an element in transparent conductive oxides used

widely in the display industry.4 A potential substitute for CIGS in thin film

solar cells is Cu2ZnSnS4 (CZTS). CZTS is a p-type semiconductor with a

band gap around 1.5 eV and high absorption coefficient (> 104 cm-1) so that

a few micrometer thick film absorbs nearly all photons with energies greater

than the band gap. Moreover, CZTS is comprised of abundant and non-toxic

elements,5–7 and CZTS-based solar cells have already reached efficiencies

as high as 12.7%.8

CZTS can crystallize in kesterite, stannite, primitive mixed Cu-Au

(PMCA), and wurtzite crystal structures.9 Besides these, other structures

derived from wurtzite do also form, such as, wurtzite-derived monoclinic, and

21

wurtzite-derived orthorhombic, which are also known as wurtzite-kesterite

and wurtzite-stannite phases, respectively.10,11 The thermodynamically

stable crystalline phase of CZTS is tetragonal kesterite (space group �4�),

whereas wurtzite (P63mc) is considered to be metastable.10 Because the

kesterite and stannite phases differ only in the ordering of the Cu+ and Zn2+

cations, routinely used characterization methods such as X-ray diffraction

and Raman spectroscopy cannot distinguish between these two phases

regardless of the underlying lattice (tetragonal or hexagonal, i.e., wurtzite).

Perhaps the only characterization technique that can possibly distinguish

between kesterite and stannite phases is neutron diffraction. Indeed, using

neutron diffraction Schorr et al.12 have shown that tetragonal kesterite forms

when the synthesis is carried out in a way to favor the equilibrium phase.

However, Schorr et al. also showed that there could still be Cu+ and Zn2+

cation disorder in the product. The analogous study to distinguish between

wurtzite-kesterite and wurtzite-stannite was recommended by Regulacio et

al.13 but to our knowledge has not been carried out yet. In this manuscript,

we refer to the tetragonal kesterite phase as kesterite and presume that this

is the ground state structure. We refer to the wurtzite phase without any

qualification on whether it is kesterite-wurtzite or stannite-wurtzite because

we cannot tell the difference without conducting neutron diffraction on nearly

phase pure samples. Of course, cation disorder can also be present but is

also difficult to quantify.

22

In 2011, Lu et al.14 synthesized the metastable wurtzite phase of CZTS

by injecting metal chloride solutions in dodecanethiol into a hot mixture of

dodecanethiol, oleylamine, and oleic acid. They hypothesized that

dodecanethiol played an important role in obtaining the metastable wurtzite

phase. Indeed, most methods for synthesizing wurtzite CZTS employ a

surfactant such as hexadecanethiol, dodecanethiol, oleylamine, and

trioctylamine during synthesis.15,16 Regulacio et al.13 showed that, in one

mechanism, the role of the surfactant is to stabilize intermediate products

that help the formation of wurtzite CZTS while restricting the formation of the

tetragonal phase. Regulacio et al.13 also found that wurtzite CZTS formation

was favored when long chain alkanethiols helped nucleation and growth of

hexagonal Cu1.94S nanocrystals, which in turn allowed wurtzite CZTS to

template on these Cu1.94S nanocrystals. The heterostructured wurtzite-

CZTS-Cu1.94S nanocrystals eventually converted to wurtzite via cation

diffusion.13 Thus, the surfactants stabilize an intermediate that locks in the

crystal structure.

Another approach to synthesize wurtzite CZTS phase is via cation

substitution into a wurtzite intermediate. For instance, Wang and co-workers

prepared wurtzite CZTS by diffusing Zn into wurtzite Cu2SnS3 and replacing

half the copper cations with zinc.17 This idea is related to that used by

Regulacio because it allows the system to choose the final crystal structure

through an intermediate nanocrystal with hexagonal symmetry. Li et al.18

varied the phase composition of their product from wurtzite to kesterite CZTS

23

by varying the sulfur sources mixed into oleylamine. Using elemental sulfur

dissolved in oleylamine produced kesterite CZTS via a rapid reaction,

whereas moderating the reaction rate by using dodecanethiol allowed the

formation of a copper sulfide intermediate (Cu7S4) and produced wurtzite

CZTS. Lin et al.19 used a microwave solvothermal method and varied the

volume fraction of ethylenediamine and water as solvents. When only

ethylenediamine was used they obtained wurtzite CZTS whereas a 50%-

50% mixture of ethylenediamine and water lead to kesterite CZTS. Again,

wurtzite formation was preceded by the formation of a sulfide, this time Cu2S,

as the intermediate.

Thus, the work to date shows that wurtzite is obtained when the

reaction is slowed and copper sulfide intermediates are allowed to form

though not all the copper sulfide intermediates identified by various groups

were the same. The reaction rate and the formation of the intermediate could

be controlled by varying the capping ligands, sulfur source, or the solvent.

One approach to making thin polycrystalline kesterite CZTS films for

solar cells is to anneal coatings cast from nanocrystal dispersions in sulfur

or selenium vapour.20–23 During annealing the microstructure of the coatings

cast from wurtzite or kesterite nanocrystals evolve differently.24 For example,

grain growth is faster, begins at lower temperature, and is more facile when

metastable wurtzite nanocrystals are transformed into kesterite grains. This

motivates the search for controlled synthetic routes that are able to

24

selectively give one crystalline phase (e.g., wurtzite) over the other (e.g.,

kesterite).24,25

Herein, we studied several factors that play a key role in the control of

the CZTS phase from wurtzite to kesterite. These factors include the sulfur

source, molar ratio of sulfur to total metal ions, and the initial oxidation state

of the tin reagent. Surfactant or capping agents were not used in any of the

syntheses. There are multiple paths to the metastable wurtzite phase.

Instead of the copper sulfides that appeared to be the key intermediate in

previous work, we show that the intermediate that leads to the wurtzite phase

in our work is a Zn-Sn glycolate complex, whose formation is controlled by

the oxidation state of Sn and the presence of an amino group in the sulfur

source. The absence of surfactants is a also a significant advantage of our

method, since long chain organic ligands are difficult to remove from particle

surfaces after synthesis, and can hinder charge transport in films prepared

from these particles.26 Additionally, the microwave based solvothermal

approach described here is reproducible, fast, and scalable. In a typical

synthesis significant amounts (0.4 g) of CZTS particles can be synthesized

in just 20 minutes.

2.2 Experimental

2.2.1 Materials

Copper (II) acetate monohydrate (CuAc2.H2O ACS reagent, >98%

Sigma-Aldrich), copper (I) acetate (CuAc 97% Sigma Aldrich), zinc acetate

25

dihydrate (ZnAc2.2H2O ACS reagent Acros Organic), tin (II) chloride (SnCl2

98% Sigma Aldrich), tin (IV) chloride pentahydrate (SnCl4.5H2O Sigma

Aldrich), thiourea (CH4N2S ≥ 99.0% Sigma Aldrich), L-cysteine (C3H7NO2S

97% Sigma Aldrich), thioglycolic acid (C2H4O2S 98% Sigma Aldrich), 3-

mercaptopropionic acid (C3H6O2S ≥ 99.0% Sigma Aldrich) ethylene glycol

(Fisher Scientific), methanol (Sigma Aldrich), and ethanol (Decon – 200

Proof) were used as received and without additional purification.

2.2.2 CZTS Nanoparticles Synthesis

In a typical synthesis, 1.7 × 10-3 mol of CuAc2·H2O, 1.0 × 10-3 mol of

ZnAc2·2H2O, and 1.0 × 10-3 mol of SnCl2 were added to 30 mL ethylene

glycol with stirring. Variable amounts of thiourea were added to achieve a

sulfur to total (Cu, Zn and Sn) metal ion molar ratio (S:M) between 1 and 6.2,

i.e., S:M = nS/(nCu+nZn+nSn), where ni is the number of moles of species i.

After sonicating for 30 minutes, this reaction mixture was sealed in a Teflon

vial, which was placed inside a SiC sleeve and loaded into an Anton Parr

Multiwave Pro microwave. The reaction mixture was heated from room

temperature to 160 °C in 15 minutes, using 300 W of power. Following this

temperature ramp, the reaction mixture was maintained at 160 °C for 5

minutes. After 5 minutes, the power was turned off, and the reaction mixture

was cooled to 55 °C in 20 minutes via convective flow induced by the fans

in the microwave. The microwave is equipped with a turntable, which spins

the vial at 2 rpm during the entire procedure. Magnetic stirring was not used

26

during the synthesis. The reaction mixture temperature was measured using

an infrared sensor.

After cooling, the contents of the Teflon vial were transferred to a

centrifuge tube and centrifuged at 8500 rpm (8450 rcf) for 15 minutes. The

supernatant was discarded, ethanol was added, and the resulting dispersion

in ethanol was centrifuged at 8450 rcf for 5 minutes. This centrifugation-

redispersion cycle was repeated four times. Finally, the precipitate was

dispersed in methanol. The effect of the excess sulfur sources other than

thiourea was studied by adding appropriate moles (to achieve a particular

S:M ratio) of L-cysteine, thioglycolic acid or 3-mercaptopropionic acid to the

mixture of CuAc2·H2O (1.7 × 10-3 mol), ZnAc2·2H2O (1.0 × 10-3 mol), SnCl2

(1.0 × 10-3 mol), and thiourea (4.0 × 10-3 mol) in 30 mL ethylene glycol.

2.2.3 Experiments varying the synthesis temperature

To study the effect of temperature, the microwave heating was

stopped during the ramp at different temperatures. For instance, a reported

synthesis temperature of 75 °C means that the microwave heating was

turned off when the temperature reached 75 °C; after the microwave heating

was turned off, the reaction mixture immediately started to cool.

2.2.4 Experiments varying the Copper and Tin Oxidation States

27

In addition to the synthesis with CuAc2·H2O and SnCl2, the following

combinations of copper and tin reagents were also used: CuAc and SnCl2,

CuAc and SnCl4·5H2O; and CuAc2·H2O and SnCl4.5H2O. In all these

experiments, 1.7 × 10-3 mol of the copper reagent, 1.0 × 10-3 mol of

ZnAc2·2H2O, and 1.0 × 10-3 mol of the tin reagent were used. Thiourea was

added in the appropriate amounts to vary the S:M ratios between 1.9 and

6.2. The ultrasonication and microwave heating procedures were the same

as used in the synthesis with CuAc2·H2O and SnCl2.

2.2.5 Product labelling convention

Because of the large number of synthesis variables, we use a

convenient product-labelling scheme. The synthesis products are labeled as

Cu(N)_Sn(L)_XCS_S:M_T, where N and L represent copper and tin

oxidation states, respectively; XCS represents the source of sulfur excess,

where Tu, Cyst, TGacid and MCPacid stand for thiourea, L-cysteine,

thioglycolic acid, and 3-mercaptopropionic acid, respectively. S:M is the

sulfur to total metal ions ratio; and T is the maximum temperature reached

during the microwave heating.

2.2.6 Characterization

Products were characterized using X-Ray Diffraction (XRD), Raman

scattering, optical and infrared spectroscopy, and various electron

microscopies. Specifically, XRD patterns from the products were collected

28

using a Pananalytical X’Pert Pro X-ray diffractometer (Co Kα radiation with

a wavelength of 1.7890 Å) equipped with a X’Celerator detector. The XRD

patterns were collected from 16 to 85° (2θ), with an effective step size of

0.0167° and 50 s dwell time per step. The crystalline phase percentage was

estimated using Rietveld refinement, with the details provided in the

Appendix A.

Raman spectra were collected from dried powders using a Witec

Confocal micro-Raman spectrometer with a green 532 nm laser as the

excitation source. The laser power was fixed at 1 mW, and each spectrum

was integrated over 150 seconds. Optical absorption spectra were collected

from diluted methanol dispersions of the nanocrystals, using an Agilent 8453

spectrophotometer. For Infrared spectroscopy analyses, nanocrystals were

drop cast from methanol dispersions onto NaCl windows, which were then

dried in air. Infrared spectra were collected using a Nicolet Series II Magna-

IR System 750 FTIR in the transmittance mode.

For SEM analyses, the nanocrystals dispersed in methanol were

sonicated for 15 minutes, drop cast onto silicon substrates, and dried in air.

The dried nanocrystals were then examined in a JEOL 6500 SEM, at an

acceleration voltage of 5 kV. Elemental composition of the nanocrystals was

determined using Energy Dispersive X-Ray Spectroscopy (EDS) using a

Thermo-Noran Vantage system equipped with an EDS detector coupled to

the JEOL 6500 SEM. The acceleration voltage was adjusted to 15 kV for all

EDS measurements, the EDS spectra were collected in the standardless

29

mode, with the ratio of the elements calculated based on the reference

spectra provided by the software System SIX. For each sample, the

elemental compositions were determined at ten different locations and these

values were averaged.

Samples for TEM characterization were prepared by drop casting

nanocrystals onto Cu TEM grids (SPI 200 mesh holey carbon coated) from

methanol dispersions after sonicating for at least 30 minutes. The

nanocrystals were imaged using a FEI T12 microscope, with acceleration

voltage of 120 kV. High-resolution images were collected using a FEI Tecnai

G2 F-30, with acceleration voltage of 300 kV. Selected area electron

diffraction patterns were collected using this same microscope. Scanning

transmission electron microscopy (STEM) was performed using an FEI

Tecnai G2 field-emission S/TEM operating at an accelerating voltage of 80

kV. High-angle annular dark field (HAADF) images were collected using an

E. A. Fischione annular detector. Energy-dispersive X-ray spectroscopy

(EDS) spectra were collected using the ChemiSTEM EDX spectrometer.

EDS maps were collected while rastering the beam over the sample, which

facilitated minimization of beam damage. A probe current of ∼0.1 nA was

used, and maps were collected over a minimum of five minutes. Data were

analyzed using ESPRIT software (version 1.9.4).

30

2.3 Results and Discussion

2.3.1 Effect of S:M ratio on the phase of the nanocrystals synthesized

using Cu(II) and Sn(II) reagents

When Cu(II) and Sn(II) precursors were used and the sole sulfur

source was thiourea, the kesterite-to-wurtzite ratio in the synthesis product

could be changed by varying the S:M ratio. Figure 2.1a shows the XRD

patterns from nanocrystals synthesized using Cu(II) and Sn(II) reagents and

thiourea as the sole sulfur source while varying the S:M ratio. The

stoichiometric reaction mixture (S:M = 1) results in the formation of impurity

phases such as Cu3SnS4, Cu4SnS4, and elemental sulfur. The kesterite-to-

wurtzite ratio increases with increasing S:M ratio (Figure 2.1b). The

nanocrystal product is comprised of approximately 50% wurtzite and 50%

31

kesterite for S:M between 2 and 4. When S:M is greater than 4, the kesterite

fraction rises to 80%. (See Appendix A Table A2 for details).

Figure 2.1 (a) XRD patterns from nanocrystals synthesized using Cu2+, and Sn2+ at 160 °C

using different S:M ratios. The sulfur source was thiourea. (i.e., Cu(II)_Sn(II)_Tu_S:M_160 oC where S:M was varied between 1.0 and 6.2. XRD patterns for wurtzite (simulated) and

kesterite (PDF# 00-0026-0575) CZTS are shown at the bottom. Peaks labeled as Ω are

consistent with Cu3SnS4 (PDF 00-036-0218), α with Cu4SnS4 (PDF 00-029-0584), and х

with elemental sulfur (PDF 00-008-0247) (b) Phase composition estimated from Rietveld

refinement analysis of the XRD patterns shown in (a). Error bars were determined from

triplicate analyses, as described in the Appendix A.

32

The nanocrystal size estimated using Scherrer equation, for the peak

located at 2θ around 32°, corresponding to the (112) planes of the wurtzite

phase, and (002) planes of the kesterite phase, varied between 6.4 and 3.7

nm and decreased with increasing S:M. The product containing mostly

kesterite had the smallest average crystallite size. (Appendix A Table A3).

Figure 2.2 Raman spectra of the nanocrystals synthesized from Cu2+, and Sn2+ at 160 °C,

while varying the S:M ratio. The sulfur source was thiourea: (i.e.,

Cu(II)_Sn(II)_Tu_S:M_160 oC where S:M was varied between 1.0 and 6.2.

Raman spectra (Figure 2.2) are consistent with the XRD results and

exhibit a high intensity peak at 328 cm-1 for all products prepared with a S:M

of 1.9 or greater. The most intense Raman peak from CZTS is expected

around 336 – 339 cm-1.27,28 However, it is common to observe this peak

shifted to lower wavenumbers (e.g., 328-331 cm-1). This peak shift is thought

to be due to inhomogeneity within the disordered cation sublattice.29,30 Peaks

related to secondary phases such as Cu2S, SnS, SnS2, ZnS, Cu2SnS3, and

Cu3SnS4 were not detected, though some of these compounds (ZnS,

33

Cu2SnS3, and Cu3SnS4) have partially overlapping peaks with CZTS and

presence of amounts undetectable via XRD and Raman can not be ruled

out.28

Bright field TEM images of the Cu(II)_Sn(II)_Tu_1.9_160 °C

nanocrystals (Figure 2.3a) revealed the presence of two types of nanocrystal

shapes. Some nanocrystals were anisotropic and oblate while others were

more spheroidal (See also Figure A1 in Appendix A). Anisotropic oblate

nanocrystal shapes have been observed previously when wurtzite CZTS

was present.31 The observation of two types of morphology is consistent with

the XRD results, which demonstrate that the product is a mixture of kesterite

and wurtzite CZTS: we associate the oblate morphology with wurtzite CZTS

and the spheroid morphology with kesterite CZTS. The SAED patterns

(Figures 2.3b and 2.3e) exhibit rings with d-spacings that are consistent with

both the wurtzite and the kesterite phases. The SAED pattern (Figure 2.3b)

is consistent with the XRD results, showing more intense reflections related

to the wurtzite phase than for kesterite phase. High-resolution images from

an oblate nanocrystal show lattice fringes with spacing consistent with the

(002) of wurtzite CZTS (Figure 2.3c). This would be consistent with our

association of the oblate anisotropic nanocrystals with the wurtzite phase. In

contrast, in the TEM image of Cu(II)_Sn(II)_Tu_6.2_160 °C (Figure 2.3f), the

nanocrystals appear spheroidal in shape, and the reflections consistent with

the kesterite phase are substantially more intense, which is consistent with

the XRD results. The spheroid nanocrystals showed lattice fringes with

34

spacings that match the (112) d-spacing of kesterite (Figure 2.3f). The

Cu(II)_Sn(II)_Tu_1.9_160 °C and Cu(II)_Sn(II)_Tu_6.2_160 °C samples

were analyzed using HAADF-STEM. The STEM-EDS maps revealed that

Cu, Zn, Sn and S are homogeneously distributed in both samples (Figure A2

in Appendix A).

Figure 2.3. (a) TEM, (b) SAED (c) HR-TEM images of the Cu(II)_Sn(II)_Tu_1.9_160 °C

and d) TEM, (e) SAED (f) HR-TEM of the Cu(II)_Sn(II)_Tu_6.2_160 °C nanocrystals.

2.3.2 Effect of Cu and Sn initial oxidation states

Varying the oxidation states of the Sn and Cu precursors led to the

conclusion that the initial tin oxidation state is an important parameter and is

one of the factors in the synthesis that determines the crystalline phase of

the product. To support this conclusion Figure 2.4 shows the XRD patterns

35

from nanocrystals synthesized using different pairings of Cu(I) and Cu(II)

reagents with Sn(II) and Sn(IV) reagents while holding S:M constant at 1.9

or 6.2: the excess sulfur was provided using thiourea. The nanocrystal

compositions are in Appendix A Table A4. When the S:M is 6.2, the majority

phase is always kesterite, regardless of the oxidation states of Sn and Cu.

However, when the S:M is 1.9, the dominant product is wurtzite when Sn(II)

is used and kesterite when Sn(IV) is used, regardless of the copper oxidation

state.

Figure 2.4. XRD patterns for (a) Cu(I)_Sn(II)_Tu_S:M_160 °C (bottom), and

Cu(I)_Sn(IV)_Tu_S:M_160 °C (top) (b) Cu(II)_Sn(II)_Tu_S:M_160 °C (bottom), and

Cu(II)_Sn(IV)_Tu_S:M_160 °C (top) .

2.3.3 Effect of excess sulfur source

We also studied the effect of changing the excess sulfur source by

substituting L-cysteine, thioglycolic acid or 2-mercaptoethanol for excess

thiourea. We conducted experiments where we varied the excess sulfur

source, Sn oxidation state (Sn(II) or Sn(IV)), and S:M ratio, S:M = 1.9 or 3.6.

Figure 2.5 shows the XRD patterns from the products of these experiments.

There are two noticeable trends with the source of sulfur excess. First, the

36

molecules containing an amino group (NH2), e.g., thiourea and L-cysteine,

produce mostly kesterite CZTS when Sn(IV) is used as tin source but tend

to produce wurtzite when Sn(II) is used. On the other hand, molecules

without NH2 groups, like thioglycolic acid and 3-mercaptopropionic acid,

when used as source of excess sulfur always produce wurtzite CZTS,

regardless of the initial tin oxidation state.

Figure 2.5. XRD patterns of the CZTS nanocrystals synthesized using different excess sulfur

sources, Sn oxidation states, and S:M=1.9 or 3.6. Sulfur excess sources were a) thiourea,

b) L-cysteine, c) thioglycolic acid, d) 3- mercaptopropionic acid. (See also Table A5 for

wurtzite and kesterite fractions calculated via Rietveld refinement).

37

2.3.4 On the mechanism and precursors to CZTS formation

Table 2.1 summarizes the dominant phases obtained with different

excess sulfur sources, tin sources and S:M ratios. Thiourea was always

necessary in at least stoichiometric amounts to form CZTS. Use of other

sulfur sources without thiourea did not produce CZTS (see Appendix A

Figure A3). While there are combinations of synthesis variables that lead to

predominantly wurtzite or kesterite phases, there does not appear to be neat

and obvious trends, perhaps with the exception of the trend with amino

groups: it appears that presence of the amino group in the excess sulfur

source differentiates the Sn(II) and Sn(IV) sources. Infrared spectra of the

surfaces of these nanocrystals were consistent with the presence of ethylene

glycol (Appendix A Figure A4). The explanation of Table 2.1 requires

experiments aimed at revealing the formation mechanisms. Towards this

end we conducted experiments at lower temperatures in attempts to slow

the formation of the CZTS nanocrystals and observe any reaction

intermediates.

Table 2.1. Summary of the dominant phases formed using different combinations of tin and

excess sulfur sources and with low and high S:M ratios. TGAcid is thioglycolic acid and

MPAcid is 3-mercaptopropionic acid.

Excess S Low S:M High S:M

Source Sn(II) Sn(IV) Sn(II) Sn(IV)

Thiourea W K K K

L-cysteine W K W K

TGacid W W W W

MPacid W W W W

38

Considering the possibility that the preferential synthesis of one phase

over another may be related to the differences in their formation mechanisms

we explored the synthesis at temperatures lower than 160 °C. Figure 2.6a

shows the XRD patterns from products synthesized at various temperatures

between 25 °C and 160 °C for S:M=1.9 with Sn(II) source and thiourea,

conditions that favor the formation of the wurtzite phase. At and below 100

°C, we find multiple diffraction peaks The reaction of the Zn and Sn(II) in

ethylene glycol yields a product with 1:1 Zn:Sn ratio, with XRD peaks

identified with black diamonds in Figure 2.6 (See also Figures A5 and A6 in

Appendix A), corresponding to Zn-Sn glycolates, as discussed by Wang,32

39

Ng,33 and Das.34 The XRD peaks for the wurtzite CZTS begin to appear at

130 °C. Zn-Sn glycolate intermediates are the only products below 100 °C.

Figure 2.6 XRD pattern from the nanocrystals synthesized at different temperatures, T from

Sn(II) source: (a) Cu(II)_Sn(II)_Tu_1.9_T, (b) Cu(II)_Sn(II)_Tu_6.2_T; ♦ denotes the

diffraction from the Zn-Sn intermediate.

It is well known that Zinc (II) acetate can react with thiourea to form

the bisthiourea zinc acetate complex (Zn((NH2CSNH2)2(CH3COO)2).35,36 It is

also known that SnCl2 can form complexes with thiourea. For instance, when

40

Sn2+ and thiourea are mixed in 1:1 mole ratio, they form thioureatin(II)

chloride (Sn(NH2CSNH2)Cl2), and when the Sn:thiourea ratio is between 1:2

and 1:6, the pentathioureadi-[tin(II)chloride]dihydrate

(Sn2(NH2CSNH2)5Cl4.2H2O) complex.37 Thus, the formation of a Zn-Sn

complex is not surprising. Unfortunately, we have been unable to grow large

single crystals of this Zn-Sn intermediate to conduct X-ray diffraction, which

leaves us unable to assign a definite formula and structure. However, after

a thorough search on Cambridge Structural Database (CSD), we found that

the XRD pattern of this Zn-Sn intermediate resembles a pattern formed by a

mixture of thioureatin(II) chloride (CSD code: CAPWEV)38 and a zinc

succinate thiourea complex called catena-((m2-Succinato-O,O')-

bis(thiourea-S)-zinc) (CSD code: FELXEA).39 Based on the similarity

between the XRD pattern of this zinc succinate thiourea complex and the

Zn-Sn intermediate, one possibility is that the intermediate is a mixture of the

thioureatin(II) chloride complex and another complex that contains glycolate

and thiourea ligands.39 Further information about the XRD patterns of the

thioureatin(II) chloride and catena-((m2-Succinato-O,O')-bis(thiourea-S)-

zinc) can be found in the Appendix A (Figure A5, the additional data in

Figures A6-A9 and the accompanying discussion). Another possibility is that

the Zn-Sn intermediate may be a bi-nuclear complex containing Zn and Sn.

A Zn-Sn intermediate is also formed when L-cysteine is used as the excess

sulfur source but not with other sulfur sources such as TGAcid and MPAcid,

implicating the amino group in the formation of this complex intermediate.

41

Another possibility is that the Zn-Sn intermediate could be a compound with

the formula similar to {[Zn(amine containing ligand)x]2[Sn2S6]}n. Similar

compounds containing Ni instead of Zn with this structure have already been

synthesized at room temperature and reported in the literature.40

In contrast, when the synthesis is conducted using higher sulfur

excess (S:M = 6.2), Zn-Sn intermediates are observed only up to 75 °C, with

CZTS peaks just beginning to appear at lower temperature. However, these

CZTS peaks are those expected from the kesterite phase and not the

wurtzite phase. For S:M=6.2 at and above 100 °C only the XRD peaks from

kesterite CZTS are observed, with no evidence of wurtzite formation. The

XRD data from the products synthesized with a high S:M ratio (S:M=6.2)

show that the kesterite phase is formed directly, without the formation of the

wurtzite phase. Similarly, the XRD data from the products synthesized with

a low S:M ratio (S:M=1.9) show that the wurtzite phase is formed directly

without the formation of the kesterite phase. In other words, there is no phase

transition from wurtzite to kesterite or vice versa during synthesis. This

indicates that the S:M ratio directly influences the nucleation and growth of

CZTS, an issue that we will revisit shortly.

The same low temperature experiments conducted with the Sn(IV)

precursor did not yield the Zn-Sn glycolate complex. Instead, the XRD and

SEM-EDS data show that the room temperature synthesis starting with

Cu(II) and Sn(IV) precursors resulted in the formation only of elemental

sulfur, with no evidence for the presence of any Zn-Sn intermediates,

42

regardless of the S:M ratio (Appendix A Figure A9 and Table A6). The fact

that thiourea does bind Sn(IV) can be explained based on Pearson’s Hard

and Soft Acid Base Theory (HSAB)41, since thiourea is a soft base and thus

more likely to bind to Sn(II) and Zn(II), which are both borderline acids, than

to Sn(IV), which is a hard acid. Consequently, thiourea remains in the

solution and decomposes, producing elemental sulfur. At higher

temperatures Cu(II) and Sn(IV) precursors always lead to the formation of

the kesterite phase. This observation implicates Zn-Sn glycolate

intermediates as the precursors in the formation of the wurtzite phase when

thiourea is used as sulfur excess source. That the Zn-Sn intermediates also

form with other amine containing sulfur sources such as L-cysteine and

leads to wurtzite formation supports this conclusion.

Figure 2.7 compares SEM images of the Zn-Sn intermediate products,

specifically, Cu(II)_Sn(II)_Tu_1.9_25 °C (Figure2.7a) and

Cu(II)_Sn(II)_Tu_6.2_25 °C (Figure 2.7c) with SEM images of the CZTS

nanocrystals produced with the same tin and sulfur sources at 160oC,

specifically, Cu(II)_Sn(II)_Tu_1.9_160 °C (Figure 2.7b) and

Cu(II)_Sn(II)_Tu_6.2_160 °C (Figure 2.7d). The morphologies of the Zn-Sn

glycolate intermediates are sensitive to S:M ratio. When the S:M ratio

employed is 1.9, the Zn-Sn intermediate particles are micron-sized

hexagonal prisms (Figure 2.7a), and the product formed after heating to 160

oC is wurtzite particles agglomerated into submicron spherules. In contrast

an S:M ratio of 6.2 leads to the formation of hexagonal plates that have holes

43

and this donut morphology is retained in the kesterite product that forms

upon heating to 160 oC (Figure 2.7c). The similarity between this

morphology and those of CZTS nanocrystals (e.g., Figure 2.7d) is striking

and immediately reinforces the conclusion that these Zn-Sn glycolate

intermediates are precursors to CZTS formation. It thus seems that the

hexagonal donuts transform to kesterite CZTS directly.

Figure 2.7 SEM images of the nanocrystals (a) Cu(II)_Sn(II)_Tu_1.9_25 °C, (b)

Cu(II)_Sn(II)_Tu_1.9_160 °C, (c) Cu(II)_Sn(II)_Tu_6.2_25 °C , and (d)

Cu(II)_Sn(II)_Tu_1.9_160 °C. (a) and (c) are the Zn-Sn glycolate precursor nanocrystals

whereas (b) and (d) are the CZTS nanocrystals formed from these precursors. (See

Appendix A Figure A10 for additional SEM images.)

Even though the morphologies of the particles synthesized with

S:M=1.9 and S:M=6.2 (e.g., at 25 °C) are very different, their XRD patterns

are nearly indistinguishable. These different precursor morphologies lead to

the formation of wurtzite CZTS when S:M=1.9 and kesterite CZTS when

44

S:M=6.2. We believe that the morphologies of the Zn-Sn glycolate precursor

mediates the rate of transformation to CZTS. We hypothesize that the

formation of wurtzite CZTS is favored by slow conversion of the Zn-Sn

intermediate precursor to CZTS via Cu diffusion. The hollow morphology

presents a larger surface area facilitating faster diffusion of Cu into the

intermediate and faster transformation, which favors the formation of the

kesterite phase. In contrast, the transformation is slower with larger

hexagonal prisms with less surface area, which slows the transport rate of

Cu into the precursor particles favoring the slow formation of the wurtzite

phase.

Table 2.1 also shows that wurtzite is always the major CZTS phase

when thioglycolic acid or 3-mercaptopropionic acid are used as the excess

sulfur source in the synthesis, regardless of the S:M ratio or the Sn oxidation

state. This is in direct contrast to synthesis where amine containing excess

sulfur sources are used where both the S:M ratio and the Sn oxidation state

affects whether one obtains kesterite or wurtzite phases. We now show that

the mechanism of the wurtzite CZTS formation when thioglycolic acid or 3-

mercaptopropionic acid are used as the excess sulfur source is different than

the mechanism discussed above where the precursor is a Zn-Sn glycolate.

We further show that when thioglycolic acid or 3-mercaptopropionic acid are

used the formation mechanism is similar to that previously revealed by

Regulacio and others, and is via a copper sulfide intermediate as discussed

in the introduction.

45

Again, to reveal the intermediates we conducted the CZTS synthesis

at temperatures lower than 160 °C, varying the Sn oxidation state. We

changed the excess sulfur source to a molecule that does not have an amino

group and cannot form the Zn-Sn glycolate precursor.

Figure 2.8 shows the XRD patterns from nanocrystals synthesized

using thioglycolic acid as the excess sulfur source. Indeed, the Zn-Sn

glycolate intermediate is not observed at any temperature. Instead, an

amorphous powder is formed at temperatures lower than 75 °C. This

amorphous powder begins to convert to crystalline Cu2-xS, above 100 oC.

We see clear diffraction peaks from Cu2-xS (e.g., sample

Cu(II)_Sn(IV)_TGacid_1.9_100 °C in Figure 2.8b) when the tin source is

Sn(IV) and weaker and broader diffraction peaks when the tin source is

Sn(II) (e.g., sample Cu(II)_Sn(II)_TGacid_1.9_100 °C in Figure 2.8a) . We

thus conclude that when the excess sulfur source does not have an amino

group, the synthesis of the wurtzite CZTS phase proceeds through an Cu2-

xS intermediate. The presence of Cu2-xS as intermediate leading to the

formation of wurtzite CZTS and CuInS2 was previously observed and

reported in the literature.42,43

46

Figure 2.8. XRD pattern of the samples Cu(II)_Sn(II)_TGacid_1.9_T and

Cu(II)_Sn(II)_TGacid_1.9_T, with T between 25 and 160 °C. The sample

Cu(II)_Sn(II)_TGacid_1.9_25 °C did not produce any solid. Peaks labelled as Δ are

consistent with Cu2-xS (See also Table A7 in the Appendix A for the elemental analysis

results).

2.4 Summary and Conclusions

Figure 2.9 is a schematic summary of our findings. Our main

conclusion is the existence of two distinct pathways to wurtzite CZTS

nanocrystals. One of these (Figure 2.9a) is via copper sulfide intermediates

as previously reported and discussed in the literature. 13,42 In this pathway,

47

the oxidation state of the tin and S:M ratio does not play a role in determining

the phase of the CZTS: it seems that wurtzite CZTS is formed as long as

copper sulfides are formed as intermediates. The second pathway (Figure

2.9b) is new and involves an intermediate Zn-Sn glycolate that appears to

form only when the sulfur precursor contains –NH2 groups. In this second

pathway, the oxidation state of the tin and S:M ratio does play a significant

role. The former determines whether the glycolate intermediate can form and

the latter determines the morphology of the precursor and thus the

transformation rate of the precursor to CZTS. The Zn-Sn glycolate forms only

when the Sn source is in the +2 oxidation state and does not form when the

tin is in +4 oxidation state.

In the new mechanism (Figure 2.9b) the first step is the formation of

Zn-Sn intermediate precursor particles upon mixing of the reagent solutions.

This reaction proceeds even at room temperature. The morphology of these

Zn-Sn intermediate particles depends on the S:M ratio. At low thiourea

concentration (i.e., S:M=1.9) large hexagonal prisms form. On the other

hand, at higher thiourea concentrations (i.e., S:M=6.2), dissolution-

reconstitution rates will increase and faster localized Ostwald ripening will

produce hollow hexagonal plate-like structures.44–47 The second step is the

transformation of the Zn-Sn intermediate precursor to CZTS, in either the

wurtzite or the kesterite phase. This transformation requires the diffusion into

and reaction of copper with the intermediate. After Cu(II) adsorbs on to the

Zn-Sn intermediates, the high temperature in the microwave promotes the

48

mixing, interdiffusion, and reaction of the species present in the Zn-Sn

intermediate particles and Cu(II) ions. The rate of this transformation

depends on the morphology of the intermediate. Small, hollow structures

promote faster conversion while large hexagonal prims slow down the

transformation. A major consequence of the morphological difference

between the products synthesized with S:M=1.9 and S:M=6.2 is that the

interdiffusion rate of the elements within the hexagonal prism shaped Zn-Sn

intermediate particles (obtained with S:M=1.9) is slower than the

interdiffusion rate in the hollow hexagonal plate-like particles (obtained with

S:M=6.2). (It is easier to diffuse into a small hollow structure with high

surface to volume ratio that it is to diffuse into a larger one with low surface

to volume ratio.) Consequently, the Zn-Sn intermediates prepared with

S:M=1.9 require a higher temperature in order to convert to CZTS. To test

this hypothesis, we examined the SEMs of the products whose XRD patterns

are shown in Figure 6 (See Figure A11 in Appendix A). These SEM images

show that, indeed, the large hexagonal prisms (S:M=1.9) still persist at 100

°C with small spherical CZTS particles growing around them while the hollow

hexagonal plates (S:M=6.2) have been completely converted. The slower

diffusion rate of Cu into the hexagonal prism shaped large Zn-Sn-

intermediate particles (obtained with S:M=1.9) slows the transformation of

the intermediate to CZTS. This slow transformation leads to the formation of

the metastable wurtzite phase.

49

Figure 2.9. a) Reaction scheme for S excess source containing NH2 group, according to the

S:M variation, leading to kesterite or wurtzite CZTS; b) Reaction scheme for S excess

source without NH2 group, leading mostly to wurtzite CZTS.

50

CHAPTER 3

3. Green Synthesis of Cu2(Zn1-x,Cox)SnS4 Nanocrystals and

Formation of Polycrystalline Thin Films from Their Aqueous Dispersions

3.1 Introduction

Kesterite Cu2ZnSnS4 (CZTS) and the related compound

Cu2ZnSn(Sx,Se1-x)4 (CZTSSe) are being considered as absorber layers in

thin-film solar cells1,2 because they have high absorption coefficients in the

visible range of the electromagnetic spectrum3 (>10-4 cm-1), direct and easily

tunable band gaps between 1.0 eV to 1.5 eV by adjusting the S-to-Se ratio,4

and a high theoretical efficiency (~32%).5 Moreover, CZTSSe is comprised

of abundant metals with low toxicity relative to other solar cell materials such

as CdTe. Although CZTSSe-based thin-film solar cells have already

achieved a record efficiency of 12.7%,6 this performance still falls

significantly short of the 21.7% efficiency reached in Cu(In,Ga)Se2 (CIGS)-

based solar cells7 and of theoretical efficiencies (32.4% for CZTS and 31%

for CZTSe).8

Alarmingly, the initial rapid rise achieved in record efficiencies

between 2010 and 2014 has stalled. Among the reasons responsible for this

underperformance is the relatively large open circuit voltage, Voc, deficit (600

mV) with respect to the theoretical maximum (i.e., the band gap, Eg, divided

by the electron charge).9,10 This deficit is much larger for CZTSSe than for

CIGS-based solar cells (440 mV).6 One possible reason proposed for this

51

relatively high Voc-deficit is band gap and band edge fluctuations due to

cation disorder and antisite defects such as CuZn, ZnCu, ZnSn, and

SnZn.11,12 Moreover, Cu and Sn have multiple oxidation states and favor the

formation of deep-level defects in CZTSSe.13 These defects may increase

the non-radiative carrier recombination rate and lower Voc.11 Higher Voc are

achieved by substituting some of the cations in CZTSSe. For example,

replacing Sn with Ge improves Voc by reducing defect concentrations.14 The

main problem, however, is the similar sizes of Cu+ (0.91 Å), Cu2+ (0.87 Å),

and Zn2+ (0.88 Å) cations, which allows them to easily substitute for each

other resulting in substantial cation disorder in wurtzite and kesterite phases

of CZTS.10,11 One solution to this problem is to substitute a smaller cation

such as Co2+ (0.68 Å) for Zn2+ (0.88 Å) to form Cu2(Zn1-xCox)SnS4 alloys with

band gaps similar to CZTS.15,16 The smaller size of the Co2+ is expected to

result in reduced cation disorder and antisite defects in Cu2CoSnS4 as

compared to Cu2ZnSnS4. Thus, Cu2(Zn1-xCox)SnS4 alloys may lead to solar

cells with higher Voc.

In one approach to making polycrystalline thin films for solar cells,

CZTS coatings are cast on substrates from nanocrystal dispersions, and the

resulting coatings are annealed in sulfur or selenium vapor.17–20 CZTS

nanocrystals have been synthesized using a variety of routes, and methods

for controlling their size and composition have been explored.21–24 Making

Cu2(Zn1-xCox)SnS4 films using this method will require the synthesis of

nanocrystals with the desired cation composition, x. A few articles have

52

reported the synthesis of Cu2CoSnS4 nanocrystals.25–30 To our knowledge,

there is only one recent report of stannite Cu2(Zn1-xCox)SnS4 (0 ≤ x ≤ 1)

nanocrystals across the entire composition range: Huang et al. synthesized

stannite Cu2(Zn1-xCox)SnS4 nanocrystals by reacting metal salts with

thiourea in oleylamine for more than 24 hours at temperatures greater than

260 °C.31 More recently, Thompson et al. synthesized wurtzite Cu2(Zn1-

xCox)SnS4 nanocrystals from metal salts in octadecene using

trioctylphosphine oxide but only for x between 0 and 0.11. To our knowledge,

wurtzite Cu2(Zn1-xCox)SnS4 nanocrystals and polycrystalline thin films have

not been synthesized previously across the entire composition range (i.e., 0

≤ x ≤ 1). Moreover, previous synthesis of Cu2CoSnS4 nanocrystals used

chemicals such as trioctylphosphine oxide and oleylamine,25,30,31 which are

known to be cytotoxic.32,33

Herein, we report a synthesis of Cu2(Zn1-xCox)SnS4 nanocrystals (0 ≤

x ≤ 1) and thin polycrystalline films formed by annealing coatings cast from

dispersions of these nanocrystals. Specifically, we made Cu2(Zn1-

xCox)SnS4 nanocrystals using a microwave-assisted solvothermal synthesis

in ethylene glycol and investigated the formation of polycrystalline Cu2(Zn1-

xCox)SnS4 films on Mo-coated soda lime glass substrates when coatings

cast from aqueous dispersions of these nanocrystals are annealed in sulfur

vapor. Advantages of microwave synthesis have been reviewed by

Baghbanzadeh et al.: they include significant reduction in synthesis times

and efficient use of energy by coupling it selectively with the solvent.34

53

3.2 Experimental Procedure

3.2.1 Materials

Copper (II) acetate monohydrate (Cu(II)Ac2·H2O ACS reagent, > 98%,

Sigma-Aldrich), zinc acetate dihydrate (ZnAc2·2H2O ACS reagent Acros

Organic), cobalt (II) acetate tetrahydrate (Co(II)Ac2·4H2O, > 97%, Acros

Organic), tin (II) chloride (Sn(II)Cl2 98% Sigma Aldrich), thiourea (CH4N2S, ≥

99.0%, Sigma Aldrich), sodium thioglycolate (HSCH2COONa ≥ 96.5%,

Sigma Aldrich), ethylene glycol (Fisher Scientific), methanol (Sigma Aldrich),

and ethanol (Decon-200 Proof) were used as received.

3.2.2 Synthesis of Cu2(Zn1-xCox)SnS4 nanocrystals

In a typical Cu2ZnSnS4 synthesis, 1.7 × 10-3 mol of Cu(II)Ac2·H2O, 1.0

× 10-3 mol of ZnAc2·2H2O, and 1.0 × 10-3 mol of Sn(II)Cl2 were added to 30

mL ethylene glycol while stirring. To synthesize Cu2CoSnS4, the 1.0 × 10-3

mol of ZnAc2·2H2O was replaced by 1.0 × 10-3 mol of CoAc2·4H2O. To

synthesize Cu2(Zn1-xCox)SnS4 solid solutions, the amounts of ZnAc2·2H2O

and CoAc2·4H2O were adjusted according to the desired solid solution

composition, x, where x is the nominal Co fraction in the precursor solutions

as calculated from the ratio of the moles of CoAc2 to the sum of the moles of

CoAc2 and ZnAc2 (i.e., [CoAc2]/[Co(II)Ac2 + ZnAc2]). Following, 4.0 × 10-3

mol of thiourea and 3.4 × 10-3 mol of HSCH2COONa were added to the metal

54

acetate solution. After sonicating for 30 minutes, this mixture was sealed in

a Teflon vial, placed inside a SiC sleeve, and loaded into an Anton Parr

Multiwave Pro microwave. The solution temperature was measured using

an infrared sensor. The solution was heated from room temperature to 160

°C in 15 minutes and maintained at 160 °C for 5 minutes. Following, the

power was turned off, and the solution was cooled to 55 °C in approximately

20 minutes with the aid of the fans inside the microwave. The microwave is

equipped with a turntable, which spins at 2 revolutions per minute (RPM)

during the entire synthesis, but the solution is not stirred.

3.2.3 Formation of Polycrystalline films

After synthesis, the wurtzite Cu2(Zn1-xCox)SnS4 nanocrystals were

centrifuged at 8450 RCF for 15 minutes, ethylene glycol supernatant

discarded, and the remaining nanocrystals dispersed in water to make

36±10 mg/ml dispersions. These dispersions do agglomerate in

approximately one day and must be redispersed by sonication before use.

The nanocrystal coatings were cast from dispersions that were sonicated for

at least 1 hour prior to casting. Soda lime glass substrates covered with 700

nm of Mo were coated with the wurtzite Cu2(Zn1-xCox)SnS4 nanocrystals by

drop casting 400 μL of the aqueous dispersions onto 6.25 cm2 area defined

by a metal frame. After drop casting, the coatings were dried at room

temperature for 2 days as described previously.35,36 The substrates coated

with nanocrystals were annealed in air at 200 °C for 10 minutes to remove

55

any residual water trapped in between the nanocrystals. These nanocrystal

coatings were porous and rough because the nanocrystals agglomerate as

the dispersion cast on the substrate dries. The coatings were compacted

using a hydraulic press at 6000 kPa for 15 seconds (Caver Autopellet 3887)

as described previously.37,38 A thin sheet of Kapton film was placed on top

of the nanocrystal coating to prevent the nanocrystals from adhering to the

press.

The compacted wurtzite Cu2(Zn1-xCox)SnS4 nanocrystal coatings

were placed in pre-cleaned quartz ampoules (1 cm inner diameter and 10

cm length) with 14 mg of solid sulfur and 3×10-5 moles of NaOH, which was

introduced as described by Johnson et al.39 Following, the quartz ampoule

was evacuated to 10-6 Torr and flamed-sealed. The sealed quartz ampoule

at room temperature was loaded into a preheated furnace at 600 °C. At this

temperature, 14 mg of sulfur corresponds to approximately 500 Torr of sulfur

vapor. Details of the annealing process can be found in our previous

papers.19,37,39 After maintaining the quartz ampoule at 600 °C for 1 hour, the

furnace was turned off and the ampoule allowed to cool naturally to 150 °C

before removing it from the furnace. The furnace took approximately 4 hours

to cool to 150 °C.

3.2.4 Characterization

The nanocrystal coatings were characterized, before and after

annealing in sulfur, using X-Ray Diffraction (XRD), Raman scattering, optical

56

absorption, and field emission scanning electron microscopy (FE-SEM). The

as synthesized nanocrystals were also characterized using transmission

electron microscopy (TEM). Specifically, XRD patterns from the products

were collected using a Pananalytical X’Pert Pro X-ray diffractometer (Co Kα

radiation with a wavelength of 1.7890 Å) equipped with an X’Celerator

detector. The XRD patterns were collected from 20 to 70 ° (2θ) with an

effective step size of 0.0167° and 50 s dwell time per step. Lattice

parameters were calculated by Rietveld refinement using the X’pert High

Score Plus software. Raman spectra were collected from dried powders and

annealed thin films using a Witec Confocal micro-Raman spectrometer, with

a green (532 nm) laser as the excitation source. The laser power was fixed

at 1 mW, and each spectrum was integrated over 150 seconds. For FE-SEM

examination, the nanocrystals dispersed in methanol were sonicated for 1

hour, drop cast onto soda lime glass substrates, and dried in air. The dried

nanocrystal coatings and annealed thin films were then examined in a JEOL

6500 FE-SEM, at an acceleration voltage of 5 kV. Elemental composition of

the nanocrystals was determined using energy dispersive X-Ray

spectroscopy (EDS) using a Thermo-Noran Vantage system equipped with

an EDS detector coupled to the JEOL 6500 FE-SEM. The acceleration

voltage was adjusted to 15 kV for all EDS measurements. The ratios of the

elemental concentrations were calculated based on the reference spectra

provided by the software system SIX and converted to atomic %. For each

sample, the elemental compositions were determined at ten different

57

locations and the values averaged. Grain sizes were estimated from SEM

images of thin films using the determined using the intercept method: lines

were drawn randomly on several plan view SEMs and the average grain

sizes were determined from the number of grain boundaries the line

intersected per unit length. Samples for TEM characterization were prepared

by drop casting nanocrystals onto Ni TEM grids (SPI 200 mesh holey carbon

coated) from methanol dispersions after sonicating for at least 30 minutes.

The nanocrystals were imaged using a FEI T12 TEM, operating at an

accelerating voltage of 120 kV. High-resolution images were collected using

a FEI Tecnai G2 F30, with an accelerating voltage of 300 kV. The elemental

composition of the nanocrystals was also measured using inductively

coupled plasma mass spectroscopy (ICP-MS) using a Thermo Scientific

XSERIES 2 ICP-MS with ESI PC3 Peltier cooled spray chamber, SC-FAST

injection loop, and SC-4 auto sampler. Samples were diluted as appropriate

and 40 ppb of Indium internal standard was added. The powders were

digested in a mixture of 4 mL of concentrated trace metal grade HNO3, 1 mL

of deionized water, 1.5 mL of concentrated trace metal grade HCl, and 1.5

mL of concentrated trace metal grade HF. The powders were digested by

heating this dispersion in a CEM Corp Discover SP-D microwave to 150 °C

for 4 minutes. We use a capital X to denote the measured (by ICP-MS and

SEM-EDS) Co fraction, which is defined as CCo/(CZn+CCo) in the

nanocrystals, where Ci is the concentration of species i (Co or Zn). We use

small x in Cu2(Zn1-xCox)SnS4 nanocrystals to denote the nominal Co fraction

58

calculated from the ratio of the moles of Co (II) acetate monohydrate to the

sum of Co (II) acetate monohydrate and Zn (II) acetate monohydrate in the

precursor solutions. These values (X and x) agree to within measurement

error and are virtually the same as shown in the Appendix B Figure B1. Even

though Na is present in the synthesis, it was not detected in the product. If

present its concentration is below the detection limit of EDS (~0.3%).

3.3 Results and Discussion

3.3.1 Cu2(Zn1-xCox)SnS4 nanocrystals

Figure 3.1 shows (a) XRD patterns, (b) lattice parameters, and (c)

sizes of Cu2(Zn1-xCox)SnS4 nanocrystals as a function of x. All XRD patterns

show three strong peaks located near 30.5°, 32°, 35°, which correspond to

diffraction from the (100), (002), and (101) planes in wurtzite Cu2(Zn1-

xCox)SnS4 (P63mc). Additionally, the XRD patterns exhibit several weak

diffraction peaks near 46°, 55°, 60°, 66°, corresponding the (102), (110),

(103), and (112) planes, also in wurtzite Cu2(Zn1-xCox)SnS4.40 No secondary

phases such as Zn-, Sn-, Cu- and Co-based sulfide compounds are detected

regardless of x. Generally, the relative intensities of the XRD peaks are

consistent with those expected for a powder diffraction pattern of the wurtzite

phase, indicating that the kesterite and/or stannite phases are present in

very low concentrations or completely absent: were a significant fraction of

59

these phases present, the (002) peak intensity would have increased

substantially as compared to the (100) and (101) peaks (See Appendix B

Figure B2). In fact, careful examination of the XRD patterns for samples with

x=0 and x=0.05 show that the (002) peak is slightly higher than expected for

wurtzite, indicating the possible presence of small amounts of the kesterite

phase. The a and c lattice parameters in the wurtzite Cu2(Zn1-xCox)SnS4

nanocrystals decrease with increasing Co fraction, x, because the ionic

radius of Co2+ (0.68 Å) is smaller than the ionic radius of Zn2+ (0.88 Å). For

x=0 and x=1, wurtzite Cu2ZnSnS4 and wurtzite Cu2CoSnS4 the calculated

lattice parameters are in excellent agreement with those reported previously

(Table 3.1).4,41,42

60

Figure 3.1 (a) XRD patterns, (b) lattice parameters, and (c) sizes of Cu2(Zn1-xCox)SnS4

nanocrystals as a function of Co fraction, x. Simulated XRD patterns of wurtzite Cu2ZnSnS4

and Cu2CoSnS4 are shown as stick patterns. Lines in (b) are linear extrapolation (Vegard’s

law) between the lattice parameters of Cu2ZnSnS4 and Cu2CoSnS4. Nanocrystal sizes were

obtained from the Scherrer equation using the measured XRD patterns. Co Kα emission was

used for XRD. Also see Appendix B Figure B2 for XRD patterns on expanded scale between

2θ =30o and 2θ=37o, and for simulation details.

The sizes of the Cu2(Zn1-xCox)SnS4 nanocrystals, calculated from the

full width at half maximum of the (002) diffraction peak using the Scherrer

equation remains approximately constant up to x=0.5 and decreases

thereafter. The decrease in the nanocrystal size with increasing doping or

alloying (i.e. increasing x) has been observed previously in other systems43

and may be expected when two different types of ions compete with the

same lattice position because this competition can reduce both the

nucleation and the growth of the nanocrystals.44,45 For example, lattice strain

61

induced upon incorporation of ions with different radii can increase critical

nucleation size.

Table 3.1 Lattice parameters of the Cu2(Zn1-xCox)SnS4 phases

Figure 3.2 shows the Raman spectra of the Cu2(Zn1-xCox)SnS4

nanocrystals as a function of nominal Co fraction, x. The A1 mode peak

position shifts from 328 cm-1 for x=0 (Cu2ZnSnS4) to 319 cm-1 for x=1

(Cu2CoSnS4). The A1 mode peak in wurtzite CZTS has been reported to

range from 325 cm-1 to 335 cm-1.50–54 There is evidence that A1 mode peak

position may depend on the shape and sizes of the nanocrystals. For

example, Li et al. reported the A1 mode peak position for wurtzite CZTS

nanoplates at 335 cm-1 as compared to 325 cm-1 for oblate rice shaped

nanocrystals.19–23 Our nanocrystals were a mixture of oblate and spherical

nanocrystals and gave rise to a broad Raman scattering that peaked at 328

cm-1 for wurtzite CZTS (x=0). To our knowledge, there is no Raman

Crystal Structure

Space Group a (Å) c (Å)

Cu2ZnSnS4 (x=0) Ref 46

Wurtzite P63mc 3.8387 6.3388

Cu2CoSnS4 (x=1) Ref 47

Wurtzite P63mc 3.806 6.295

Cu2ZnSnS4 (x=0) Ref 48

Kesterite I4�2m 5.427 10.848

Cu2CoSnS4 (x=1) Ref 49

Kesterite I4�2m 5.396 10.789

Cu2ZnSnS4 (x=0) This work

wurtzite P63mc 3.837 6.339

Cu2CoSnS4 (x=1) This work

wurtzite P63mc 3.809 6.281

62

scattering reported for wurtzite Cu2CoSnS4. Gillorin et al. synthesized

kesterite Cu2CoSnS4 and reported a broad Raman scattering peaking at 315

cm-1 for 3 nm diameter nanocrystals.25 For larger Cu2CoSnS4 kesterite

crystals (e.g., 100s of nm), Krishnaiah observed the A1 peak at 325 cm-1.29

Huang et al. reported a shift from 335 cm-1 to 325 cm-1 for stannite or

kesterite Cu2(Zn1-xCox)SnS4 as x increased from 0 to 1.31 Thus, substituting

Co for Zn in kesterite CZTS appears to shift the A1 mode from 337 cm 1 to

325 cm-1. Such Raman shift to lower wavenumbers has also been observed

when Zn was substituted with Fe to form Cu2(Zn1-xFex)SnS4 and Co to form

Cu2(Zn1-xCox)SnS4.55

Figure 3.2 (a) Raman spectra and (b) A1 mode peak positions of Cu2(Zn1- xCox)SnS4

nanocrystals as a function of x.

Both alloying and small nanocrystal size broadens the Raman peaks

and reduces scattering intensity, making it difficult to ascertain the phase

purity by Raman alone. However, there is no obvious presence of large

amounts of secondary phases such as Cu-, Zn-, Co-, and Sn-sulfides via

63

XRD. The impurity phases that are most difficult to distinguish from Cu2(Zn1-

xCox)SnS4 are Cu3SnS4 and ZnS. ZnS Raman scattering at ~350 cm-1 is

weak and can easily be masked by the main broad A1 mode.56–58 Similarly,

the most intense Raman scattering from Cu3SnS4 is at 318 cm-1.56 To

complicate matters, these phases have overlapping XRD peaks with

Cu2(Zn1-xCox)SnS4. However, the presence of both of these phases in

significant amounts would alter the intensity ratios of the (100), (002), and

(101) XRD peaks in Figure 3.1. This is not the case and the diffraction

intensity ratios closely match those expected from a powder of Cu2(Zn1-

xCox)SnS4.

Figure 3.3 shows representative high resolution (HR)-TEM images of

Cu2(Zn1-xCox)SnS4 nanocrystals. The Cu2(Zn1-xCox)SnS4 nanocrystals are

either spherical or oblate. Oblate nanocrystals have been observed

previously when wurtzite Cu2ZnSnS4 and Cu2CoSnS4 nanocrystals were

synthesized in other solvents.26,41 On average the nanocrystals observed in

the TEM are smaller for larger x, which is consistent with sizes obtained from

XRD patterns (Figure 3.1c). The HR-TEM images show lattice fringes with

64

spacings consistent with the (101), (102), (100) planes of wurtzite CZTS and

wurtzite Cu2CoSnS4.

Figure 3.3 Representative TEM images of Cu2(Zn1-xCox)SnS4 nanocrystals with; (a) x = 0,

(b) x = 0.25, (c) x = 0.4, (d) x = 0.6, (e) x = 0.75, and (f) x = 1, respectively. TEM samples

were made by drop casting nanocrystal dispersion in methanol onto Ni carbon mesh grids.

The Cu2(Zn1-xCox)SnS4 nanocrystals show strong absorption in the

visible region of the electromagnetic spectrum, with the absorption edge

(Figure 3.4) blue shifting from ~1120 nm (1.1 eV) to ~ 920 nm (1.35 eV) as

x is increased. The band gap of wurtzite Cu2ZnSnS4 has been reported

between 1.4 and 1.5 eV, but these analyses rely on plotting (αhν)2 vs. (hν),

arbitrarily deciding the region where this plot appears linear, and

extrapolating this line to (αhν)=0 to find where it intercepts the (hν) axis: this

intercept is an estimate of the band gap energy, Eg.59 However, this analysis

is fraught with difficulties and potential pitfalls, particularly when there is

significant reflection and scattering or when there is absorption near the

65

band edge due to plasmons or tail states. Indeed, all the reported

measurements for wurtzite CZTS show absorption below the reported band

gap values.41,46,59 (See also the discussion in Appendix B) Using density

functional theory within hybrid functional PBE0, Zhao et al. calculated

wurtzite CZTS band gap to be 1.372 eV.60 Like the previous experimental

absorption spectra, absorption from our wurtzite CZTS films also begins to

rise at ~ 1200 nm, indicating either a lower band gap value than the reported

1.43 eV or a significant absorption tail below the conduction band edge,

possibly due to defects. There is only one report of optical absorption in

Cu2(Zn1-xCox)SnS4, but these are in kesterite or stannite nanocrystals.31

Based on the changes in the Raman peak widths, the authors claimed that

the crystal structure changed from kesterite to stannite, around x=0.3-0.6,

as x was increased. The XRD, however, shows that these crystals were

either kesterite or stannite and not wurtzite. These authors reported that the

band gap for Cu2(Zn1-xCox)SnS4 (kesterite or stannite) red shifted from 1.45

eV at x=0 to 1.21 eV at x=1. These values are again determined from

extrapolation of the (αhν)2 vs. (hν) plot, which showed a significant tail and

absorption below the extrapolated values. In fact, one cannot discern a band

edge from the unprocessed absorption spectra in this article. Our absorption

edge gradually blue shifts to ~1.35 eV as more Co is substituted for Zn (i.e.,

x increases), the opposite trend reported by Huang et al.31 (See also the

discussion in Appendix B.) One possibility is that the nanocrystals become

increasingly more quantum confined as their sizes decrease with increasing

66

Co concentration (see Figure 3.1), but Khare et al. showed that the CZTS

nanocrystal sizes must be approximately 3 nm or less to observe this

effect.61 Thus, quantum confinement is an unlikely explanation for the shift

observed in the spectra of Figure 3.4.

Figure 3.4 Optical absorption spectra of Cu2(Zn1-xCox)SnS4 nanocrystals as a function of Co

fraction, x. The absorption spectra were obtained from films drop cast from nanocrystal

dispersions in methanol on soda lime glass substrates after compression. Air was used as a

baseline.

3.3.2 Polycrystalline Cu2(Zn1-xCox)SnS4 thin films

Encouraged by the formation of wurtzite Cu2(Zn1-xCox)SnS4 solid

solutions without any detectable secondary phases, we proceeded to study

the formation of polycrystalline Cu2(Zn1-xCox)SnS4 thin films from coatings

comprising the nanocrystals. We made polycrystalline Cu2(Zn1-xCox)SnS4

thin films by annealing, in sulfur vapor, coatings drop cast onto Mo-coated

67

soda lime glass substrates from aqueous dispersions of the nanocrystals.

Figure 3.5 shows the XRD patterns and lattice parameters of the

polycrystalline films formed on Mo-coated soda lime glass substrates by

annealing coatings of Cu2(Zn1-xCox)SnS4 nanocrystals. The XRD patterns

collected from the annealed films with x=0 are consistent with kesterite CZTS

phase, Mo, and MoS2. When annealed, the wurtzite nanocrystals transform

into larger kesterite grains. When x is less than 0.4, no secondary phases

were detected by XRD, and the films appear to consist primarily of

polycrystalline Cu2(Zn1-xCox)SnS4. On the other hand, when x ≥ 0.6,

secondary phases such as Cu1.96S and Co0.24Zn0.76S were also detected.

We assign the peak appearing at 33.4o to cubic Co0.24Zn0.76S. This peak

appears at around x=0.75 and increases in intensity as the Co concentration

is increased and x reaches 1. It may seem odd that this particular

composition of this alloy forms, but Becker and Lutz et al. showed that there

is a limit to solubility of Co in zinc blende ZnS. It so happens that 24% is

indeed the solubility limit at 600 oC, our annealing temperature.62 The

diffraction peaks at 22o and 45.6o are from tetragonal Cu1.96S. The Mo peak

at 38° overlaps with diffraction from some of the secondary phases, most

notably the strong (103) diffraction from Cu1.96S also at 38°. One can still see

the Cu1.96S diffraction emerging from the top of the broad Mo diffraction for

films with x=0.6 and x=0.75. While wurtzite Cu2(Zn1-xCox)SnS4

forms across the entire composition range, kesterite Cu2(Zn1-

xCox)SnS4 via this route could only be formed up to x=0.4. It is interesting to

68

note that Cu2(Zn1-xFex)SnS4 transforms from kesterite to stannite with

changes in the cation ordering when x ≈ 0.4-0.5.49,63,64 Because Co2+ and

Fe2+ have similar sizes, like Cu2(Zn1-xFex)SnS4, kesterite Cu2(Zn1-xCox)SnS4

also becomes destabilized near this value of x.31 For this reason, we surmise

that the formation of kesterite from the wurtzite structure may be-come

inaccessible around x ≈ 0.4-0.5, leading instead to phase separation into

Cu1.96S and Co0.24Zn0.76S.

Figure 3.5 XRD patterns from Cu2(Zn1-xCox)SnS4 polycrystalline thin films formed on Mo-

coated soda lime glass substrates via thermal annealing in sulfur of Cu2(Zn1-xCox)SnS4

nanocrystal coatings. Annealing was conducted in 500 Torr of sulfur at 600 °C for 1 hour.

3.0 x 10-5 mol of Na was added to the annealing ampule as described in the experimental

procedure. XRD patterns of kesterite Cu2ZnSnS4 (00-026-0575), kesterite Cu2CoSnS4 (00-

026-0513), tetragonal Cu1.96S (00-012-0224), cubic Co0.24Zn0.76S (00- 047-1656), and cubic

CoS2 (98-001-3473) are also shown for comparison. Co Kα emission was used for XRD.

69

The lattice parameters of polycrystalline Cu2(Zn1-xCox)SnS4 thin films

extracted from the XRD decrease from a=0.5429 nm and c= 1.0832 nm at

x=0 to a=0.5425 nm and c= 1.0828 nm at x=0.4, and this trend is consistent

with the results obtained with the Cu2(Zn1-x,Cox)SnS4 nanocrystals (Figure

3.1(b)). For x ≥ 0.4, we can not determine the true x in the kesterite phase

since there are significant secondary phases in the films.

Figure 3.6 shows the Raman spectra and A1 mode peak positions for

polycrystalline Cu2(Zn1-xCox)SnS4 thin films formed on Mo-coated soda lime

glass substrates. The film made from nanocrystals without Co(x=0) exhibit a

strong peak located at 337±2 cm-1, which corresponds to the A1 mode

Raman peak of the kesterite CZTS structure. As x increases to x=0.4 (Figure

3.6(a)), this peak shifts towards lower wavenumbers (324 cm-1), broadens,

and decreases in intensity. This behavior would be expected if Co was

substituting for Zn and forming a solid solution. If the Co and Zn were to

phase segregate into Cu2ZnSnS4 and Cu2CoSnS4 domains, we would

expect two sharper peaks, with one at 337 cm-1 for Cu2ZnSnS4 and another

around 324 cm-1 for Cu2CoSnS4.29,31 For x ≤ 0.4, we do not observe any

other Raman scattering that can be assigned to secondary phases.

However, we begin seeing Raman scattering from secondary phases in films

synthesized with x ≥ 0.6, in addition to that from kesterite Cu2(Zn1-xCox)SnS4,

which appears between 337 cm-1 (x=0) and 324 cm-1 (x=1). For example,

Figure 3.6(b) shows Raman spectra collected from three different locations

on the same film (x=0.6). Region (i) shows intense Raman scattering from

70

copper-sulfur vibrations characteristic of copper sulfides (475 cm-1). An

additional peak at 400 cm-1 can be identified both in this region (i) and

elsewhere on the film, e.g., region (ii), and is assigned to characteristic Co-

S vibrations in Co0.24Zn0.76S, which are expected around 390-400 cm-1.65,66

Region (ii) in the Figure 3.6(b) shows Raman scattering from Cu2(Zn1-

xCox)SnS4 and Co0.24Zn0.76S but not from Cu1.96S. The vast majority of the

film, however, is Cu2(Zn1-xCox)SnS4 and exhibits Raman scattering like that

shown for region (iii) in Figure 3.6(b). In some regions like region (i) in Figure

3.6(d) a particularly strong Raman peak at 476 cm-1 from copper sulfide is

observed when x=1.

Figure 3.6 (a-d) Raman spectra and (e) A1 mode peak positions of Cu2(Zn1- xCox)SnS4 thin

films formed on Mo coated soda lime glass substrates by thermal annealing of Cu2(Zn1-

xCox)SnS4 nanocrystal coatings. A1 mode Raman peaks were collected from large grains in

the Cu2(Zn1-xCox)SnS4 films as shown in (e). Annealing conditions were the same as for

Figure 5.5. The Raman spectra for Cu2(Zn1-xCox)SnS4 thin films at x ≥ 0.6 (b,c,d) were

collected from different grains, which were selected to show the different phases present.

Figure 3.7 shows plan view and cross sectional FE-SEM images of

the polycrystalline thin films formed on Mo-coated soda lime glass substrates

via thermal annealing of Cu2(Zn1-xCox)SnS4 nanocrystal coatings in sulfur.

71

FE-SEM images for the polycrystalline Cu2(Zn1-xCox)SnS4 thin film for x=0

show a 2.0 µm thick film consisting of submicron sized grains ranging from

0.2 to 2.4 µm. The grains appear densely packed though some voids are

visible. As Co fraction, x, increases, the microstructure and grain size

undergo dramatic changes. When x = 0.25 and 0.4, we observe large

abnormal grains that have grown to over 2 µm on top of a layer comprised

of smaller 200 to 800 nm grains. The average grain size, increases from 0.6

± 0.3 to 1.1 ± 0.5 µm as x increases from 0.6 to 1. While the annealed

Cu2(Zn1-xCox)SnS4 thin films show micron sized grains, there are also many

voids. Further optimization of the annealing parameters; such as S partial

72

pressure, temperature, annealing time as well as Na concentration; will be

needed.

Figure 3.7 Plan view and cross sectional FE-SEM images of the Cu2(Zn1- xCox)SnS4 thin

films formed on Mo coated soda lime glass substrates by thermal annealing Cu2(Zn1-

xCox)SnS4 nanocrystal coating; ((a) and (b)) x = 0, ((c) and (d)) x = 0.25, ((e) and (f)) x =

0.4, ((g) and (h)) x = 0.6, ((i) and (j)) x = 0.75, and ((k) and (l)) x = 1. Annealing conditions

were the same as for Figure 3.5.

73

3.4 Conclusions

We report the synthesis of wurtzite Cu2(Zn1-xCox)SnS4 nanocrystals

across the entire composition range using a microwave assisted

solvothermal method. The Cu2(Zn1-xCox)SnS4 nanocrystals have an

average size that decreases from 8 to 4 nm as x increases from 0 to 1. The

absorption band edge blue shifted from 1.1 eV for Cu2ZnSnS4 (x=0) to 1.35

eV for Cu2CoxSnS4 (x=1). These values are lower than those predicted by

density functional theory calculations and previous attempts at determining

the optical band gap for Cu2ZnSnS4 (x=0) and Cu2CoxSnS4 (x=1). Either the

band gap of wurtzite Cu2ZnSnS4 (x=0) and Cu2CoxSnS4 (x=1) are lower, or

these materials have significant band tails due to defects. The lattice

parameters for Cu2(Zn1-x,Cox)SnS4 nanocrystals also decrease with

increasing x. Finally, we made polycrystalline Cu2(Zn1-xCox)SnS4 thin films

by annealing, in sulfur vapor, coatings comprised of Cu2(Zn1-xCox)SnS4

nanocrystals. Upon annealing, the coatings transformed to larger grained

(100s of nm to several microns) kesterite films if x ≤ 0.4. At higher x (x ≥ 0.6)

annealing wurtzite nanocrystals produced secondary phases such as

Co0.24Zn0.76S, Cu1.96S and Co2S, in addition to kesterite Cu2(Zn1-xCox)SnS4.

74

CHAPTER 4

4 Selective removal of Cu2−x(S,Se) phases from Cu2ZnSn(S,Se)4 thin

films

4.1 Introduction

Kesterite, Cu2ZnSnS4 (CZTS), and the related alloys Cu2ZnSn(S,Se)4

(CZTSSe) are prospective candidates as environmentally sustainable light

absorbing materials for low cost thin-film solar cells because they are composed

of earth-abundant elements with low toxicity.1–3 CZTS and CZTSSe have optical

properties well-suited for use in solar cells, including a direct and easily tunable

band gap energy, from 1 eV to 1.5 eV, by adjusting the S-to-Se ratio, and a high

absorption coefficient (>104 cm−1) in the ultraviolet and visible regions of the

electromagnetic spectrum.4–6 The theoretical maximum efficiency for single

junction solar cells based on CZTS is near 30% and is comparable to the

theoretical maximum efficiencies for silicon and other commercialized thin film

solar cells based on copper indium gallium selenide (CIGS) and CdTe.4–8 CZTSSe

based thin-film solar cells have been developing rapidly during the last decade and

have already achieved a record efficiency of 12.7%.9 This rapid progress in the

laboratory now motivates the development of scalable techniques compatible with

industrial mass production.

Despite promising properties and rapid development, the highest power

conversion efficiency of CZTSSe based solar cells is still far below that of

commercial thin-film solar cells based on CIGS (21.7% record efficiency)10 or CdTe

This chapter is published at: Pinto, A. H.; Shin, S. W.; Aydil, E. S.; Penn, R. L. Selective Removal of Cu2−x(S,Se) Phases from Cu2ZnSn(S,Se)4 Thin Films. Green Chem. 2016, 18, 5814–5821.

75

(19.6% record efficiency).3 Among the factors responsible for this

underperformance is the formation of impurity phases during the synthesis of

CZTSSe thin films.11–13 Impurity phases can be present on the surface, near the

back contact (e.g., Mo), or within the absorber layer and usually form during

annealing in S and Se vapor that is common for the synthesis of CZTSSe films.

The impurity phases located near the back contact have been traced to the

reaction of the CZTSSe film with Mo.12,13 To avoid this parasitic decomposition

reaction and to improve back contact stability, diffusion barrier layers such as i-

ZnO, TiN, and TiB2 have been introduced between the Mo and the CZTSSe

film.12,14–17 The impurity phases located on the surface of the CZTSSe film can be

removed by chemical etching. While several etchants have been developed to

selectively remove impurity phases such as Zn(S,Se), Cu2Sn(S,Se)3, and

Sn(S,Se)x,11,18–21 most have focused on removing the conductive Cu2−x(S,Se)

phase, a common cause of electrical shunts across the solar cell electrodes.22

Specifically, potassium cyanide (KCN) has emerged as a widely used etchant to

remove Cu2−x(S,Se) phases from CIGS thin films.22 The efficiencies of solar cells

made from CIGS thin films etched in KCN increased dramatically as compared to

those made from films that were not etched in KCN.22 Similar improvements have

been reported for CZTSe thin film solar cells.11,23 However, KCN is toxic and, in

some circumstances, reactions of KCN can produce the highly toxic gas HCN.24,25

Its use, particularly in mass production, is both a safety and cost concern. A less

toxic chemical that etches Cu2−x(S,Se) selectively over CZTSSe is desirable.

Recently, Webber et al. reported the remarkable ability of 1,2-ethylendithiol and

76

1,2-ethylenediamine mixtures for dissolving V2VI3 chalcogenides (V: As, Sb, and

Bi; VI: S, Se, and Te) at room temperature and atmospheric pressure.26,27 Inspired

by this report, we hypothesized that this mixture might selectively remove the

Cu2−x(S,Se) phases in CZTSSe thin films at room temperature and atmospheric

pressure.

Herein, we report the development of a selective chemical etchant for

Cu2−x(S,Se) over CZTS or CZTSSe that is much less toxic than KCN. The etchant

is based on a mixture of mercapto-alcohol and ethylenediamine. To demonstrate

the efficacy of this selective etchant, we compare the dissolution rates of kesterite,

wurtzite CZTS, and Cu2−xS in a mixture of 2-mercaptoethanol and

ethylenediamine, hereafter referred as the etchant solution. Surprisingly, the

Cu2−xS nanocrystals dissolve within a few seconds in the solution, while wurtzite

CZTS persists for several hours. Exploiting this significant difference in etching

rates, we show that this etchant mixture also selectively removes Cu2−x(S,Se) from

thin films comprising CZTS (or CZTSSe) and Cu2−x(S,Se).

4.2 Experimental Procedure

4.2.1 Materials

Copper(II) acetate monohydrate (Cu(CO2CH3)2·H2O) ACS reagent, >98%

Sigma Aldrich), Zinc(II) acetate dihydrate (Zn(CO2CH3)2·2H2O ACS reagent Acros

Organic), Tin(II) Chloride (SnCl2 98% Sigma Aldrich), thiourea (CH4N2S > 99%

Sigma Aldrich), sodium thioglycolate (C2H3O2SNa > 96.5% Sigma Aldrich),

ethylene glycol (C2H6O2 Fisher Scientific), copper(I) chloride (CuCl 95% Acros

77

Organic), ethylenediamine (C2H8N2 > 99.0% Fluka), and 2-mercaptoethanol

(C2H6OS > 99% Sigma Aldrich) were used without as received, without any

additional process.

4.2.2 Synthesis of Cu2−xS, wurtzite CZTS, and kesterite CZTS nanocrystals

To synthesize wurtzite CZTS nanocrystals, 1.7 × 10−3 mol (0.3394 g) of

(Cu(CO2CH3)2·H2O), 1.0 × 10−3 mol (0.2195 g) of Zn(CO2CH3)2·2H2O, and 1.0 ×

10−3 mol (0.1935 g) of SnCl2, 4.0 × 10−3 mol (0.3045 g) of thiourea, and 3.4 × 10−3

mol (0.4088 g) of sodium thiglycolate were added to 30 mL of ethylene glycol under

stirring. This solution was sonicated for 30 minutes before loading it in a Teflon vial

and then placed inside a SiC sleeve of a commercial Anton Parr Multiwave Pro

microwave reactor. The microwave was programmed to heat the samples from

room temperature to 160 °C in 15 minutes. The microwave power during this

temperature ramp was 300 mW. Following the ramp procedure, the temperature

was maintained at 160 °C in 5 minutes, and after that, the power was turned off.

The products in the vial were allowed to cool naturally to 55 °C before being

removing the vial from the microwave reactor. The microwave is equipped with a

turn table, which spins at 2 revolutions per minute (RPM) during the entire

procedure. The temperature was measured using an infrared sensor, and the

reaction media enclosed in the Teflon vials were not stirred.

To synthesize kesterite CZTS nanocrystals, the same amounts of

(Cu(CO2CH3)2·H2O), Zn(CO2CH3)2·2H2O, and SnCl2, as in the synthesis of

wurtzite CZTS nanocrystals, were used. However, the total thiourea added to the

78

reactant solution was 2.46 × 10−2 mol (1.8726 g), and sodium thioglycolate was

omitted from the synthesis. All other procedures were the same as in wurtzite

CZTS nanocrystal synthesis.

To synthesize Cu2−xS nanocrystals, 2.0 × 10−3 mol (0.2084 g) of CuCl, 1.0

× 10−3 mol (0.0761 g) of CH4N2S, and 1.0 × 10−3 mol (0.1114 g) of C2H3O2SNa

were added to 30 mL C2H6O2 under stirring. This mixture was then sonicated and

loaded into the microwave reactor as described above for CZTS nanocrystal

synthesis. All other procedures were the same as in the wurtzite and kesterite

CZTS nanocrystals synthesis.

4.2.3 Precursor coating

Wurtzite CZTS and Cu2−xS nanocrystal dispersions were prepared by

mixing 800 μL of 44 ± 10 mg ml−1 wurtzite CZTS nanocrystals dispersed in ethylene

glycol, with 200 μL of 36 ± 10 mg ml−1 Cu2−xS nanocrystals, also dispersed in

ethyleneglycol. The mixed dispersion was centrifuged at 8500 RPM for 15 minutes.

Following, the supernatant was discarded and 5 mL of deionized water was added

to form stable aqueous nanocrystals dispersion. The Mo-coated substrates (700

nm thick Mo layer) were coated with nanocrystals by drop casting 400 μL of the

dispersion onto a 6.25 cm2 area defined by a metal frame followed by drying at

room temperature for 2 days.28 Specifically, we mixed Cu2−xS and wurtzite CZTS

nanocrystals and casted coatings on Mo-coated soda lime glass substrates. The

wurtzite CZTS was chosen due to its slowest etching rate in comparison to

kesterite CZTS, as it was shown on results and discussion session. The coatings

79

were heated for 10 minutes at 200 °C in air to remove any remaining solvent. The

drop cast coating was very porous and its surface was rough. This porous and

rough coating was compacted using a hydraulic press at 6000 kPa for 15 seconds

(Caver AutoPellet 3887) as described by Williams et al.28 A thin sheet of Kapton

film was placed on the coatings to prevent the coating from adhering to the press.

4.2.4 Cu2−x(S,Se) and kesterite CZTSSe thin films

The compact coating was placed in pre-cleaned quartz ampoule (1 cm inner

diameter and 9 cm length) with solid sulfur (10 mg) or selenium (11 mg). Following,

the ampoule was evacuated to 10−6 Torr and flamed-sealed. The sealed quartz

ampoule at room temperature was loaded into a furnace pre-heated to 600 °C.

After keeping the ampoules and its contents at 600 °C for 1 hour, the furnace was

turned off and the ampoule was allowed to cool naturally to 150 °C before removing

it from the furnace.29,30 It took approximately 4 hours for the furnace to cool to

150 °C.

4.2.5 Etching

The etching solution was prepared by adding 1 mL of 2-mercaptoethanol to

4 mL of ethylenediamine in a 20 mL vial at room temperature, without stirring.

About 20 mg of dried Cu2−xS, kesterite CZTS, or wurtzite CZTS nanocrystals were

weighed, placed in this solution and observed visually and using optical absorption

to determine if the nanocrystals had dissolved. Thin films cast from nanocrystal

dispersions and then annealed in S or Se were also placed in this solution for a

80

period of 20 minutes. At the end of these 20 minutes, the film was taken out the

solution and dried on a heating plate at 100 °C for 10 minutes. Rinsing was very

effective in removing any residual etchant as confirmed by attenuated total

reflection Fourier transform infrared (ATR-FTIR) spectroscopy. All etching

experiments were conducted at room temperature and atmospheric pressure.

After the heat treatment at 600 °C, wurtzite CZTS was converted to kesterite

CZTS.

4.2.6 Characterization

The optical absorbance of the etchant solutions including the nanocrystals

was measured using Ultraviolet-Visible (UV-Vis) spectroscopy (Agilent 8453) at

room temperature. The nanocrystals and the thin films were characterized using

X-ray diffraction (XRD), before and after etching. XRD was recorded with a Bruker

D8 Discover system that employed a Cu Kα X-ray source, 0.8 mm beam collimator,

and Hi-Star 2D area detector. The spot size of the X-ray beam on the sample was

800 μm. Surface morphology of the thin film was examined using a field emission

scanning electron microscope (FE-SEM JEOL 6500, at 10 keV) before and after

etching. The Raman scattering from the films was imaged, before and after

etching, using confocal Raman microscopy. Specifically, a Witec alpha300R

confocal Raman microscope equipped with a UHTS300 spectrometer and a

DV401 CCD detector operating at −60 °C was used. The excitation source was an

argon ion laser at 532 nm (average power 0.38 mW) and the excitation spot is

approximately 300 nm. Raman scattering was collected in the backscattering

81

geometry. The scattered light was collected, filtered to remove the contributions at

the excitation wavelength and dispersed using a 600 lines per mm grating. Raman

scattering images were acquired from selected 100 μm × 100 μm regions.

4.3 Results and discussion

X-ray diffraction patterns collected from the nanocrystalline products used

as powders or in the preparation of thin films for the etching experiments described

in this chapter. Figure 4.1 reveals that the products obtained matched the expected

powder diffraction from wurtzite CZTS, kesterite CZTS and Cu2−xS (including CuS

and Cu2S). Impurity phases such as ZnS, SnS2, and Cu2SnS3 were not detected

within the detection limitations of XRD. The average sizes estimated from the

Scherrer equation applied to the full width at half maximum of the most intense

peak in each pattern were 9 nm, 3 nm, and 15 nm, for wurtzite CZTS, kesterite

CZTS, and Cu2−xS nanocrystals, respectively.

82

Figure 4.2 shows the optical absorption spectra from (a) Cu2−xS, (b)

kesterite CZTS, and (c) wurtzite CZTS nanocrystals dispersed in the etchant

solution after 1 minute, 2 hours, and 6 hours. The photographs in each inset show

the visual appearance of the respective dispersion after 1 minute, 2 hours, and 6

hours. Before the nanocrystals are dispersed, the etchant solution appears clear

and light yellow (similar to the photograph labelled 6 hours in Figure 4.2(a)).

Figure 4.1 X-ray diffraction patterns for Cu2-xS (a), kesterite-CZTS and wurtzite-CZTS NCs

(b) prepared by microwave assisted solvothermal process. X-ray diffraction patterns were

collected from nanocrystals drop cast and dried on a glass substrate from aqueous dispersions

at room temperature.

83

Remarkably, the etchant solution does not change its color when Cu2−xS

nanocrystals are added. Nearly all of the Cu2−xS nanocrystals dissolve within

seconds, with only a few specks of agglomerates remaining after 1 minute (see

the bottom of the vial in Figure 4.2(a) after 1 minute). The initial dispersions of

wurtzite or kesterite are opaque and black. After six hours, the kesterite CZTS

nanoparticles completely dissolve, resulting in a light yellow solution, whereas, the

wurtzite CZTS dispersion remains black and opaque, even after six hours.

84

The optical absorption from kesterite CZTS nanocrystals in the etchant

solution after 1 minute and 2 hours is higher and noisier than the absorption from

Figure 4.2 X-ray diffraction patterns for Cu2-xS (a), kesterite-CZTS and wurtzite-CZTS

nanocrystals (b) prepared by microwave assisted solvothermal process. X-ray diffraction patterns

were collected from nanocrystals drop cast and dried on a glass substrate from aqueous dispersions

at room temperature.

85

the etchant without the nanocrystals. This is because nearly all the light is

absorbed. Once all the kesterite CZTS nanocrystals have dissolved, the absorption

deceases and the spectrum of the etchant solution is recovered. Wurtzite CZTS

nanocrystals remain dispersed in the solution and absorb nearly all the light in this

spectral range even after 6 hours. We conclude that the etching (dissolution) rate

of Cu2−xS is the highest followed by kesterite CZTS, and the etching rate of wurtzite

CZTS is the slowest.

Encouraged by the dramatic differences in these dissolution rates, we

designed an experiment to determine whether Cu2−xS phases could be removed

selectively from films containing a mixture of Cu2−xS and kesterite CZTS. We made

the films by annealing, in sulfur, coatings cast from nanocrystal dispersions as

described in the Experimental section. To form films that contained both CZTS and

Cu2−xS, coatings were drop cast from nanocrystal dispersions containing both

wurtzite CZTS and Cu2−xS nanocrystals.

Figure 4.3 shows plan-view FE-SEM images of these annealed films before

(Figure 4.3(a)) and after (Figure 4.3(b)) etching in the etchant solution for 20

minutes. Hexagonal crystals with sizes ranging from 500 nm to 2.5 μm are clearly

visible on the surface of the film before it was etched. This hexagonal crystal habit

is typical of CuS and the hexagonal form (so called high chalcocite phase) of Cu2S

(Figure 4.3(a)).31 The areal density of hexagonal crystals on the annealed thin film

is approximately 10−2 hexagonal crystals per μm2. The annealed thin film is

composed of 0.2–1 μm wide grains with some cracks and voids. The hexagon-

shaped crystals are removed after the film is immersed in the etchant solution for

86

20 minutes (Figure 4.3(b)). With the exception of the absence of hexagonal

crystals, the morphology and microstructure after etching is similar to that

observed before etching. Selective removal of copper sulfide crystals from the film

over kesterite CZTS is consistent with the nanocrystal dissolution experiments

shown in Figure 4.2.

Figure 4.3 Plan view FE-SEM images of films (a) before and (b) after etching in the etchant

solution for 20 minutes. Higher magnification FE-SEM images are shown in the insets.

87

The etchant solution etches copper sulfides faster than kesterite CZTS.

Figure 4.4 shows the XRD patterns from the annealed films before and after

etching in the etchant solution for 20 minutes. The XRD pattern recorded before

etching shows strong diffraction peaks located at 28.5°, 47.3°, and 56.1°, which

correspond to the kesterite CZTS structure (ICDD-ref.: 00-034-1246), as well as

weak diffraction peaks at 29.2°, 31.8°, 47.8°, and 52.8°, which correspond to the

hexagonal copper sulfide structure (ICDD-ref.: 98-000-0176). Other diffraction

peaks can all be assigned to kesterite CZTS, Mo, or MoS2.32 The copper sulfide

diffraction peaks are weak because of the small amounts in the CZTS film.

Nevertheless, consistent with the SEM images, the peaks from copper

sulfide disappear from the diffraction pattern collected after etching the film in the

Figure 4.4 X-ray diffraction patterns from annealed films before and after etching in the etchant

solution for 20 minutes. CZTS and copper sulfide diffraction patterns are shown for comparison.

88

etchant solution for 20 minutes. Finally, we examined the films using confocal

Raman microscopy both before and after etching. Specifically, we mapped the

spatial distribution of Raman scattering from copper sulfide and kesterite CZTS

phases in the annealed thin films before and after etching in the etchant solution

for 20 minutes. Raman scattering maps were constructed by filtering all scattering

except in a narrow band (±10 cm−1) around the most intense characteristic Raman

scattering peaks of Cu2−xS and kesterite CZTS at 475 cm−1 and 338 cm−1,

respectively. Raman spectra from CuS and Cu2−xS are indistinguishable. While we

refer only to Cu2−xS in the following, both CuS and Cu2−xS contribute to the

scattering intensity at 475 cm−1.33 Figure 4.5 shows Raman scattering maps for

films annealed in S before ((a) Cu2−xS and (b) CZTS)) and after ((c) Cu2−xS and (d)

CZTS)) etching. The Raman scattering image for Cu2−xS (475 cm−1) in Figure

4.5(a) shows many randomly distributed regions where scattering at 475 cm−1 is

very strong. The distribution of these regions on the surface is similar to the

distribution of the hexagonal copper sulfide crystals observed with SEM (e.g.,

Figure 4.3(a)) and strongly suggests that these regions have one or more copper

sulfide crystals. The Raman scattering map for CZTS (338 cm−1) shows a uniform

distribution of CZTS (Figure 4.5(b)) and this map does not change significantly

after etching (Figure 4.5(d)). On the other hand, the Raman scattering from copper

sulfide at 475 cm−1 (Figure 4.5(c)) disappears after etching, suggesting removal of

the copper sulfide phases by the etchant solution.

Figure 4.5(e) and (f) show representative spectra from two different regions

on the film, before and after etching, respectively. The Raman spectra labelled

89

Region (i) in Figure 4.5(e) is representative of spectra collected from regions where

there are no copper sulfide grains (i.e., dark regions in Figure 4.5(a)), and the

spectrum shows a strong peak at 338 cm−1 and weak peak at 278 cm−1, both from

CZTS. The Raman spectra labelled Region (ii) in Figure 4.5(e) is representative of

spectra collected from regions where there are copper sulfide grains (i.e., bright

orange regions in Figure 4.5(a)). Thus, the spectrum shows a very strong peak at

475 cm−1 (copper sulfides) as well as weak peaks at 264 cm−1 (copper sulfides)

and 419 cm−1 (MoS2).1,22 Regardless of the location on the surface, we do not find

any peaks at 475 cm−1 after etching (Figure 4.5(f)), which is consistent with

complete removal of copper sulfides within the detection limits of Raman

scattering.

90

We conducted the same study described above on selenized thin films.

Figure 4.6 shows the XRD pattern from a coating, initially comprised of CZTS and

Cu2−xS nanocrystals, after selenization (labelled as ‘before etching’). The

Figure 4.5 Raman scattering maps from annealed CZTS films (a & b) before (c & d) after etching

in e solution. Raman scattering images from (a & c) Cu2−xS and (b & d) CZTS were produced

by selectively filtering scattering at 475 cm−1 and 338 cm−1, respectively. Representative Raman

spectra of CZTSe films (e) before and (f) after etching. In (e) and (f), spectra from two different

regions are shown. The Raman spectra labelled Regions (i) and (ii) in (e) are representative of

spectra collected from dark (CZTS) and bright orange (Cu2−xS) regions in (a), respectively. The

Raman spectra labelled regions (i) and (ii) in (f) are representative of spectra collected from any

location on the surface after etching.

91

diffraction peaks are shifted to smaller angles with respect to those expected from

CZTS and is observed nearly at the diffraction angles expected for CZTSe.29

Specifically, the XRD pattern shows the three strong diffraction peaks

corresponding to the (112), (204), and (312) planes in kesterite CZTSe (ICDD-

ref.:00-052-0868). This indicates that Se has replaced nearly all the S in the

nanocrystals during selenization.29,34,35 In addition, the XRD pattern shows weak

diffractions from copper selenide (ICDD-ref.: 03-065-3562) as well as from Mo and

MoSe2.

Figure 4.6 also shows XRD patterns from CZTSe films before and after

etching in the etchant solution for 20 minutes. The two XRD patterns are identical

Figure 4.6 XRD patterns for CZTSe thin film containing a small fraction of Cu2−xSe before

and after etching in the etchant solution for 20 minutes.

92

except for diffraction peaks from copper selenide, which are absent in the

diffraction pattern collected after etching. Thus, the XRD indicates that copper

selenides are removed selectively over CZTSe from the films by the etchant

solution.

Figure 4.7 Raman scattering maps from annealed CZTSe films (a & b) before (c & d) after

etching in etchant solution. Raman scattering images from (a & c) Cu2−xSe and (b & d) CZTSe

were produced by selectively filtering scattering at 260 cm−1 and 196 cm−1, respectively.

Representative Raman spectra of CZTSe films (e) before and (f) after etching. In (e) and (f),

spectra from two different regions are shown. The Raman spectra labelled Regions (i) and

(ii) in (e) are representative of spectra collected from dark (CZTSe) and bright orange

(Cu2−xSe) regions in (a), respectively. The Raman spectra labelled Regions (i) and (ii) in (f)

are representative of spectra collected from any location on the surface after etching. CuSe,

Cu2Se and CZTSe diffraction patterns are also shown for comparison.

93

Raman spectroscopy supports the conclusion that copper scattering maps

were constructed by filtering all scattering except in a narrow band (±10 cm−1)

around the most intense Raman scattering peaks characteristics of Cu–Se

vibrations in CuSe and Cu2−xSe (260 cm−1) and of kesterite CZTSe (196 cm−1).

Figure 4.7 compares Raman scattering maps for selenized thin films before

and after etching. Figure 4.7(a) and (c) show the Raman scattering maps for

copper selenide before and after etching, respectively, while Figure 4.7(b) and (d)

show the same for CZTSe. Figure 4.7(a) shows that the selenized film contains

copper selenide grains scattered randomly on the surface of the film. Comparison

of Figure 4.7(a) with Figure 4.7(c) shows that Raman scattering from copper

selenides at 260 cm−1 disappears after etching, while Raman scattering from

CZTSe at 196 cm−1 (Figure 4.7(b) and (d)) does not change significantly. Figure

4.7(e) and (f) show representative Raman spectra before and after etching in the

etchant solution, respectively. Again, spectra from two different regions are shown.

The Raman spectra labelled Region (i) in Figure 4.7(e) is representative of spectra

collected from regions where there are no copper selenide grains (i.e., dark regions

in Figure 4.7(a)). Thus, the spectrum only shows a strong peak at 196 cm−1 and

weaker peaks at 173 cm−1 and 230 cm−1, which all correspond to the expected

peaks from the kesterite CZTSe. The Raman spectra labelled Region (ii) in Figure

4.7(e), on the other hand, is representative of spectra collected from the bright

orange regions in Figure 4.7(a) and show the copper selenide peak at 260

cm−1.34,35 The CZTSe Raman peaks are observed both before and after etching

regardless of where the spectra is collected on the surface indicating that copper

94

selenide phases are removed selectively from the CZTSe film by etching in the

etchant solution.

Figure 4.8 shows the plane view FE-SEM images of the CZTSe thin films

containing copper selenide impurity phases before (a) and after (b) etching in the

etchant solution. There are many hexagonal copper selenide crystals with sizes

ranging from 1 μm to 5 μm protruding from the film.36 We counted approximately

73 hexagonal crystals in a 100 μm × 100 μm, corresponding to an areal density

equal to 0.7 × 10−2 hexagonal crystals per μm2, which is similar to the density of

copper sulfide hexagonal crystals on the annealed films shown in Figure 4.3.

Clearly, the hexagonal crystals are removed from the film after etching (Figure

4.8(b)).

95

4.4 Conclusions

In summary, we report a new safer etchant, a mixture of ethylenediamine

and 2-mercaptoethanol, for facile and selective removal of copper sulfide and

copper selenide phases from kesterite CZTS(Se) thin films at room temperature

and atmospheric pressure. Copper sulfides and copper selenides are completely

Figure 4.8 Plane view FE-SEM images of films (a) before and (b) after etching in etchant

solution for 20 minutes. Higher magnification FE-SEM images are shown in the insets.

Hexagonal crystals protruding from the surface are copper selenide grains while the

remaining large crystals are CZTSSe grains. These copper selenide grains are removed

selectively when the film is immersed the etchant solution for 20 minutes.

96

and selectively removed by immersing the films in the etchant for 20 minutes. This

etchant may prove an effective substitute for the widely used toxic and hazardous

KCN to remove copper sulfide and copper selenide impurity phases commonly

encountered while forming thin films of CIGS, CZTS, and CZTSe for solar cells.

We showed, using various analytical methods, that the new etchant, like its toxic

alternative, KCN, removes the Cu2−x(S,Se) phases from CZTS films. It is well

known that removal of these impurity phases with KCN improves solar cell

efficiencies and is sometimes the difference between working and non-functional

devices.22 The new approach is greener, simpler, safer and at least as effective in

removing Cu2−x(S,Se) phases as the current KCN based etching approach. Finally,

these results suggest that this etchant could be effective in removing Cu2−xSe from

CIGS. It may also be effective in etching metal–tellurium compounds. While we do

not yet understand the origin of this selectivity at the atomistic level, one possibility

is that Cu complexation with etchant molecules is much faster and facile than

complexation with the other cations (i.e., Zn and Sn). In CZTS and CZTSe, the

surface of the nanocrystals (or grains) may get depleted of Cu quickly, but once

the surface becomes rich in Zn and Sn, etching essentially stops. Testing this

mechanistic hypothesis will be the subject of a future study.

97

CHAPTER 5

5. Etching Mechanism of Cu-, Zn-, and Sn-Containing Sulfides in

Ethylenediamine and 2-Mercaptoethanol Mixture

5.1 Introduction

Metal chalcogenide based semiconductors have been widely used in

transparent conductive electrodes, low emission windows, phase change memory,

thin film transistors, quantum dot light-emitting diodes, and thin-film solar cells.1–6

Among metal chalcogenides, Cu-based chalcogenides such as Cu2ZnSn(S,Se)4

(CZTSSe) and Cu(In,Ga)Se2 (CIGS) are potential absorber materials for thin-film

solar cells due to their direct band gap, high optical absorption coefficients, easily

tunable band gap, and high therotical performance.1,7–16 Various undesired phases

such as Cu2-x(S,Se)x, Zn(S,Se), Snx(S,Se)y, and CuxSny(S,Se)z can be formed

during annealing of CZTSSe and CIGS, and these impurities compromise the

electrical properties and can drastically reduce device performance.9 The

formation of Cu2-x(S,Se) phases in CIGS can be prevented by adjusting process

parameters due to wide composition window for CIGS-based absorber layers with

high performance.9,17 In contrast, impurity phases must be removed from CZTSSe

because even small changes in the CZTSSe composition compromises

performance. The stoichiometry that results in the production of phase pure CZTS

is an extremely narrow one, which means that tight control during synthesis is

essential to achieve a phase pure material with suitable stoichiometry.

98

Furthermore, a somewhat Cu-poor and Zn-rich composition is ideal for producing

high performance absorber layers in solar cells. Subsequent annealing in sulfur

vapor, which is a necessary step to produce thin films composed of grains of

suitable crystallinity and size, commonly results in the formation of Cu2-xSx.9,17

The ideal etching procedure will remove the undesired impurities in one step

and leave CZTSSe intact. Several etchants have been developed to selectively

remove undesired phases such as Cu2-x(S,Se)x, Zn(S,Se), Snx(S,Se)y,

CuxSny(S,Se)z, and several studies have reported enhanced performace after

selective etching to remove such impurities.18–23 Bromine(Br)-based solutions

have been demonstrated to selectively etch Zn(S,Se) and CuxSny(S,Se)z, and

potassium cynide (KCN) has been demonstrated to selectively etch Cu2-

x(S,Se)x.18–20 However, both are toxic. In the case of KCN, highly toxic HCN gas

can be produced.11,24 Thus, less toxic or non-toxic chemicals that etch selectively

undesired metal-chalcogenides on CZTSSe surface are desirable.14,15,25

Recent work using mixtures of 1,2-ethylenedithoil and 1,2-ethylenediamine

to dissolve a wide range of selenide and telluride compounds26,27 lead us to posit

that a similar mixture might serve as an effective etchant for impurities formed

during CZTSSe thin film preparation and annealing. Indeed, thin films of mixed

metal-chalcogenides have been prepared by solution deposition from solutions

prepared by dissolving metal-chalcogenides in thiol-amine mixtures followed by

evaporation and then annealing.3–6,28–30 Inspired by those results, Pinto et al. used

the approach to effectively remove Cu2-x(S,Se) from CZTSSe thin films under

99

ambient conditions.11 The approach represents a greener etching solution as

compared to the commonly employed KCN approach as described previously.

Herein, we report results that enable elucidation of the mechanism by which

Cu, Zn, and Sn-containing sulfides are dissolved in ethylenediamine and 2-

mercaptoethanol under ambient conditions. We quantitatively compare dissolution

rates of Cu2-xS, ZnS, SnS2, CuxSnySz (CTS), kesterite Cu2ZnSnS4 (CZTS) in

mixtures of ethylenediamine and 2-mercaptothanol and characterize materials

after exposure. Results from the characterization of materials before, during, and

after etching enable elucidation of the mechanism by which the materials dissolve

and whether the CZTS is altered in the time required to remove impurity phases.

Results demonstrate that the relatively slow dissolution of CZTS means that a high

quality thin film remains after treatment with the ethylenediamine and 2-

mercaptoethanol solution.

5.2 Experimental Procedures

5.2.1 Materials

Copper (II) acetate monohydrate (CuAc2∙H2O ACS reagent, >98% Sigma-

Aldrich), zinc acetate dihydrate (ZnAc2∙2H2O ACS reagent Acros Organic), tin (II)

chloride (SnCl2 98% Sigma Aldrich), tin (IV) chloride pentahydrate (SnCl4∙5H2O

Sigma Aldrich), thiourea (CH4N2S ≥ 99.0% Sigma Aldrich), ethylene glycol (Fisher

Scientific), methanol (Sigma Aldrich), and ethanol (Decon– 200 Proof) were used

as received, without further purification.

100

5.2.2 Synthesis of Cu2-xS, ZnS, SnS2, CuxSnySz, Cu2ZnSnS4 nanocrystals

Reagents were added to 30 mL of ethylene glycol, and resulting mixtures

were sonicated for 45 minutes, until all the powders become completely dispersed.

In the case of binary sulfides, such as Cu2-xS, ZnS, and SnS2, 1.0 x 10-3 mol of the

metal cation source and 24.6 x 10-3 mol of thiourea were used. The metal cation

sources were CuAc2∙H2O, ZnAc2∙2H2O, and SnCl4∙5H2O. In case of the ternary

sulfide, CTS, 2 x 10-3 mol of CuAc2∙H2O, and 1.0 x 10-3 mol of SnCl2 were used.

For the quaternary sulfide, CZTS, 1.7 x 10-3 mol of CuAc2.H2O, 1.0 x 10-3 mol of

ZnAc2∙2H2O, and 1.0 x 10-3 mol of SnCl2 were used. Then, the reaction mixtures

were transferred to a Teflon vessel (Anton Parr), enclosed in SiC sleeves, which

were enclosed in a Anton Parr Multiwave Pro microwave. The microwave was set

to increase the temperature from room temperature up to 160 °C in 15 minutes

and then held at 160 °C for 5 minutes. After the end of the microwave heating, the

reaction vessel was cooled down to 55 °C using forced room temperature air. Once

the reaction vessels had been cooled, they were dissembled and the contents

centrifuged at 8450 rcf for 15 minutes. The supernatant was discarded, and the

resulting solid material was washed by resuspending it in ethanol, centrifuging,

discarding the supernatant, and repearing these steps five times. Lastly, materials

were resuspended in methanol.

X-ray diffraction patterns collected from the products of each respective

synthesis (Appendix C Figure C1) matched the expected diffraction from CuS,

ZnS, SnS2, CuxSnySz (CTS) including Cu2SnS3 and Cu4SnS4, and kesterite CZTS

structures. The average nanocrystal sizes estimated from the Scherrer equation

101

applied to the full with at half maximum of the most intense peak in each pattern

were, 7 nm, 2 nm, 2 nm, 3 nm, and 3 nm for Cu2-xS, ZnS, SnS2, CTS, and kesterite

CZTS nanocrystals, respectively.

5.2.3 Preparation of heterogeneous Cu2ZnSnS4 thin films from nanocrystal

coatings

Mixed dispersions of known mass loading (ranging from 44 - 1 mg/mL) were

prepared by mixing appropriate volumes of dispersions of each of the above

phases, with the goal of producing a dispersion that contained ca. 98-99% CZTS

and 1-2% impurity phase. The resulting mixed dispersions were then centrifuged

(8450 rcf) and the ethylene glycol supernatant discarded. Then, the solid material

was resuspended in 5 mL purified water. The resulting dispersion was sonicated

for 1 hour and then drop casted onto molybedenum coated (700 nm thick) soda-

lime glass substrates over an area of 6.25 cm2 as defined by a tightly fitted metal

frame. These were then allowed to dry under ambient conditions, as described by

previous works.11–13,16,31

The resulting films were then heated in air to 200 °C and held at that

temperature for 10 minutes in order to remove residual solvent. The resulting

coatings were compacted using a hydraulic press at 6000 kPa for 15 seconds

(Caver Autopellet 3887) as described previously.16,31 A thin sheet of Kapton film

was placed between the nanocrystal coating and the press surface to prevent loss

of nanocrystals by adhesion to the press surface.

102

Resulting samples were thn annealed in sulfur vapor following our

previously used protocol.11–13,16,31 Each sample was placed in a pre-cleaned quartz

ampoule (1 cm inner diameter and 10 cm length) with 14 mg of solid sulfur. Quartz

ampoules were then evacuated to 10-6 Torr and flame-sealed. The resulting

ampoules were then loaded into a furnace that had been preheated to 600 °C. At

this temperature, 14 mg of sulfur corresponds to approximately 500 Torr of sulfur

vapor pressure. After one hour, the furnace was turned off and allowed to cool

naturally to 150 °C (ca. four hours), at which point the ampoule was removed and

allowed to fully cool to ambient temperature.

5.2.4 Preparation of etching solutions and selective etching procedure

Solutions of varying volume fractions of 2-mercaptoethanol and

ethylenediamine were prepared by mixing appropriate volumes of each liquid. The

conductivity of solutions containing varying volume fractions of each component

was measured using an Accumet AP65 Series Handheld Conductivity meter.

Etching solutions were prepared by mixing one part by volume 2-

mercaptoethanol with four parts per volume ethylenediamine at room temperature.

This proportion was chosen based on the study of the CZTS etching efficiency of

different proportions of the mixture as shown on Figure C2 in Appendix C. To semi-

quantitatively determine the relative rates of etching of each of the prepared sulfide

materials, 1 mL of etching solution and 6 mg of sulfide powder was added to

individual wells of a colorless and transparent well plate. The resulting dispersions

were photographed at regular time intervals. To determine whether undesired

103

phases could be effectively etched from the annealed films. The annealed films on

the molybedenum coated soda-lime glass substrates, as described previously,

were placed into the etching solution for 20 minutes. After removal, the resulting

films were rinsed with purified water and dried by heating to 100 °C for 10 minutes

on a hot plate.

5.2.5 Characterization

Products were characterized using X-Ray Diffraction (XRD) using a

Pananalytical X’Pert Pro X-ray diffractometer (Co Kα radiation with a wavelength

of 1.7890 Å) equipped with a X’Celerator detector. The XRD patterns were

collected from 16 to 85 ° (2θ), with an effective step size of 0.0167 ° and 50 s

dwell time per step. The Raman scattering from the films was imaged, before and

after etching, using confocal Raman microscopy. Specifically, a Witec alpha300R

confocal Raman microscope equipped with a UHTS300 spectrometer and a

DV401 CCD detector operating at −60 °C was used. The excitation source was an

argon ion laser at 532 nm (average power 0.38 mW), and the excitation spot was

approximately 300 nm. Raman scattering was collected using backscattering

geometry. The scattered light was collected, filtered to remove the contributions at

the excitation wavelength, and dispersed using a 600 lines per mm grating. Raman

scattering images were acquired from selected 50 μm×50 μm regions. For infrared

spectroscopy (IR), either solvents or dispersions were dropped onto a NaCl plate,

in case of dispersions, they were dried by blowing hot air before the beginning of

spectra collection. The IR spectra were collected in the transmittance mode using

104

a Nicolet Series II Magna-IR System 750 Fourier transform (FT) IR, with resolution

of 4 cm-1. For TEM characterization, the methanol dispersions of the samples were

sonicated for 30 minutes and drop cast onto Au TEM grids. TEM images were

collected using a FEI T12 microscope, with acceleration voltage of 120 kV and

equipped with an Energy Dispersive X-Ray Spectroscopy (EDS) detector. EDS

data were collected TEM from 5 different regions. X-Ray Photoelectron

Spectroscopy (XPS) spectra was performed using a Phi VersaProbe III X-

ray/ultraviolet photoelectron spectrometer. H1-Nuclear Magnetic Resonance (H1-

NMR) spectra were collected at room temperature from 200 µL of the etching

solution mixed with 800 µL of deuterated dimethyl sulfoxide (DMSO), using a

Bruker AV-500 MHz NMR spectrometer.

105

5.3 Results and Discussion

5.3.1 Etching Cu-, Zn-, and Sn-containing sulfides nanocrystals

Figure 5.1 Photographs of 6 mg of Cu2-xS, ZnS, SnS2, CTS, and CZTS nanocrystals before (top

row) and at specific time intervals after introduction of 1 mL of etching solution (one part by

volume 2-mercaptoethanol with four parts per volume ethylenediamine).

The rate of dissolution of each of the transition metal sulfides varies

drastically with composition. Figure 5.1 shows a time series of photographs of each

transition metal sulfide (labeled across the top) taken at regular time intervals

(labeled along the left-hand side) after addition of the 20:80 (by volume) 2-

mercaptoethanol-ethylenediamine etching solution. The dissolution rates can be

106

ranked from fastest to slowest: ZnS, Cu2-xS, CTS, and CZTS. The well containing

SnS2 contains solid material at all time points, and this will be discussed later.

Notably, nearly all the ZnS, Cu2-xS has dissolved after just one-half hour while CTS

dissolution was accomplished around 6 hours and little to no CZTS has dissolved.

To elucidate the mechanism by which these phases are dissolved, CZTS

nanocrystals were dispersed in neat ethylenediamine or 2-mercaptoethanol,

washed, and then characterized using transmittance FTIR spectroscopy. First, it is

important to note that CZTS did not dissolve in the neat liquids. Figure 5.2 presents

FTIR spectra of ethylenediamine (5.2-a), 2-mercaptoethanol (5.2-b), and the

materials after exposure to either 2-mercaptoethanol (5.2-d) or ethylenediamine

(5.2-e) followed by washing. No bands consistent with the presence of 2-

mercaptoethanol are observed for materials soaked in 2-mercaptoethanol, while

many bands consistent with the presence of ethylenediamine (e.g., 1595, 1459

and 1318 cm-1) are observed for materials soaked in ethylenediamine. Thus, we

can conclude that 2-mercaptoethanol does not bind to the transition metal sulfides

whereras ethylenediamine remains after washing. Interestingly, the

ethylenediamine band located at 1595 cm-1, which is related to the asymmetric

deformation of N-H group, is shifted to lower wavenumber (1581 cm-1) in spectrum

5.2-e, which could indicate binding of ethylenediamine to CZTS via N-H group.

107

The FTIR spectrum of the etching solution (20:80 2-

mercaptoethanol:ethylenediamine; Figure 5.2-c) does not exhibit the 2-

mercaptoethanol band located at 2562 cm-1, which is related to the H-S bond. The

absence of the H-S band in the spectrum of the etching solution may indicate a

Lewis acid-base reaction between the two molecules. Interestingly, CZTS exposed

to the etching solution exhibits bands consistent with the presence of both

components, with a band at 1055 cm-1 that is related to stretching of the C-C-O of

2-mercaptoethanol and bands at 1595, 1459, and 1318 cm-1 that are consistent

with the presence of ethylenediamine.32–34 However, no band at 2562 cm-1, is

observed. Thus, we hypothesize that the products of Lewis acid-base reaction sorb

to the metal sulfide surface.

Figure 5.2 FTIR spectra of (a) ethylenediamine, (b) 2-mercaptoethanol, (c) Etching Solution (20%

2 mercaptoethanol, 80% ethylenediamine, by volume), (d) CZTS kept for 2 h in neat 2-

mercaptoethanol, (e) CZTS kept for 2 h in neat ethylenediamine, and (f) CZTS kept for 2 h in

etching solution. All the spectra were collected in transmittance mode.

108

Based on the FTIR spectra, we hypothesized that the etching solution might

be conductive due to the presence of ionic species. Figure 5.3. shows the

conductivity versus the ethylenediamine volume percentage on the etching

solution (bottom X-axis). Neither 2-mercaptoethanol nor ethylenediamine are

conductive as neat liquids. The mixtures have substantial conductivity, and the

maximum (5.9 mS) is observed at the 20:80 2-mercaptoethanol-ethylenediamine

volume ratio. Figure 5.3 also presents the conductivity versus the mole ratio

between NH2 and HS groups (top X-axis). This maximum conductivity (5.9 mS)

coincides with NH2/HS mole ratio of 9.37. As the NH2/HS mole ratio increases

further, the conductivity decreases substantially. The decrease in conductivity

when ethylenediamine volume percent is equal to 90% can be explained by a

maximum possible concentration of deprotonated 2-mercaptoethanol is quite low,

since NH2/HS is around 21.

Figure 5.3. Conductivity versus the ethylenediamine volume percent (a), and mole ratio between

HS/NH2 groups (b). The point having ethylenediamine volume percent equal to 100% has an

undefined NH2/HS mol ratio.

109

H1-NMR spectroscopy results (Figure 5.4) indicate that deprotonation of H-

S group happens for every mixture between 2-mercaptoethanol and

ethylenediamine. However, there are two meaningful differences between the

spectra of the mixtures in different 2-mercaptoethanol-ethylenediamine volume

ratios, they are: (i) The peak related to the protons residing on N (labelled as A on

Figure 5.4) steadily shift as the concentration of 2-mercaptoethanol increases,

which is consistent with rapidly exchanging protons spending more and more time

on S as compared to N. (ii) The peak related to CH2 group of 2-mercaptoethanol

increases its integrated intensity as the amount of 2-mercaptoethanol increases

(see peak labelled as E on Figure 5.4). The conductivity in conjunction with the H1-

NMR results lead us to hypothesize that the maximum conductivity is reached at

ca. 9.37 NH2/HS due to the equilibrium, leading to a condition that maximizes the

presence of charged species.

To further elucidate the roles of 2-mercaptoethanol and ethylenediamine

play in dissolving solids, the salts containing Cu(I) (CuAc), Zn(II) (ZnAc2), and

Sn(IV) (SnCl4) ions were placed into neat ethylenediamine or 2-mercaptoethanol.

When the Cu(I) and Sn(IV) salts were added to ethylenediamine, black (Cu(I)) or

white Sn(IV) amorphous material precipitated (Figure C3, in Appendix C). When

the Zn(II) salt was added to ethylenediamine, a white crystalline solid formed

(Figure C3). We surmise that [Zn(en)3](acetate)2 was formed, and although we

could not obtain a reference pattern or crystal structure for direct comparison, salts

of [Zn(en)3]2+ are known, and XRD patterns for two related compounds are shown

on Figure C4 in the Appendix C. When Cu(I) acetate was added to 2-

110

mercaptoethanol, elemental sulfur precipitated, which is consistent with the

oxidation of the thiol S to elemental sulfur. This result can be explained by the fact

that Cu(I) ions can catalyze the cleavage of C-S bonds.35 Interestingly, when ZnAc2

or SnCl4 are added into neat 2-mercaptoethanol, both salts are readily solubilized.

In both cases, the observed result can be explained by the formation of soluble

complexes of Zn(II)36,37 and Sn(IV) and 2-mercaptoethanol.38

Figure 5.4. H1-NMR spectra of the neat ethylenediamine, 2-mercaptoethanol and mixtures in

different proportions of these two solvents (top). Possible species produced during Lewis acid-

base reaction between ethylenediamine and 2-mercaptoethanol (bottom).

111

We hypothesize a sequence of reaction steps to explain the etching of

binary sulfides, Cu2-xS and ZnS. First, a Lewis acid-base reaction between

ethylenediamine and 2-mercaptoethanol occurs, resulting in an equilibrium mixture

of neutral ethylenediamine, singly protonated ethylenediamine (+H3NCH2CH2NH2),

doubly protonated ethylenediamine (+H3NCH2CH2NH3+), neutral 2-

mercaptoethanol, and thiol-deprotonated 2-mercaptoethanol (-

SCH2CH2OH). Species that adsorb to form surface complexes on the

chalcogenide surface are most likely the neutral ethylenediamine and the

deprotonated 2-meraptoethanol. Then, the surface bound metal cation detach,

resulting in the formation of metal cation-organic complexes in solution. Based on

this results, in combination with the lack of chalcogenide solubility in neat

ethylenediamine and neat 2-mercaptoethanol, we can surmise that the

deprotonated 2-mercaptoethanol plays a particularly important role in this ligand-

assisted dissolution. Finally, sulfides could bind the ethylenediamine species,

forming N,N’diamine-polysulfide species.39 These steps are summarized on Figure

C5 in Appendix C.

5.3.2 Etching CTS and CZTS nanocrystals

The etching of CTS and CZTS is much slower than the etching of Cu2-xS

and ZnS. In general, the rate of dissolution of nanocrystals will depend on size,

microstructure, and composition. The average sizes of the CTS and CZTS

nanoparticles (ca. 3nm) are smaller than the average size of the fastest dissolving

112

Cu2-xS (ca. 7 nm), which means that size can not explain the slower dissolution of

CTS and CZTS.

We then hypothesized that the distribution of the elements within CTS and

CZTS nanocrystals might influence their rates of dissolution in the etching solution.

Comparing elemental analysis obtained via EDS, which yields a bulk result, and

XPS, which yields a composition describing the elements residing at and near the

surface, reveals that the as prepared CTS and CZTS are deficient in copper at and

near the surface (Table C1 in Appendix C). Upon exposure to the etching solution,

the bulk composition of the CZTS nanocrystals becomes more and more copper

rich, which indicates that the Zn(II) and Sn(IV) are preferentially dissolved away

(Figure C6 a). A similar trend is observed with CTS dissolution, although the

preferential etching away of tin is less pronounced in CTS than the preferential

etching away of zinc and tin in CZTS (Figure C6 b). These steps are summarized

on Figure C7 in Appendix C.

5.3.3 Explaining the inability to etch SnS2 nanocrystals

SnS2 does not dissolve in the etching solution (Figure 5.1). XRD results on

Figure 5.5 demonstrate that the initially reddish orange powder is composed of the

expected hexagonal SnS2 (PDF 00-023-0677). Based on the position of the extra

XRD peak (at 2 theta equal to 9 °) in the initial powder, we hypothesize that it can

be related to formation of an intercalation compound composed by SnS2 layers

and ethylene glycol molecules.

113

After 14 hours in the etching solution, the XRD pattern exhibits no peaks for

SnS2. Instead, two peaks at 10 ° and 21 ° two theta are observed, and these are

consistent with an intercalation compound composed by SnS2 layers separated by

ethyenediamine.40 These results suggest that ethylenediamine can intercalate

phase pure SnS2 and replace the intercalated ethylene glycol. To test whether

ethylenediamine can indeed intercalate phase pure SnS2, the initial powder was

annealed to remove the ethylene glycol, and the XRD pattern of this annealed

material is consistent with the phase pure SnS2. Upon addition to the etching

solution, the same color change from reddish orange to black, and XRD pattern of

the black material is consistent with the SnS2-ethylenediamine intercalation

compound. Thus, we conclude that SnS2 does not dissolve in the etching solution,

but rather it is transformed to a stable intercalation compound.

114

Figure 5.5. X-ray diffraction patterns of as-synthesized and post-annealed SnS2 nanocrystals before

and after etching process (left). Right photo images for visual appearance of annealed SnS2

nanocrystals as function of etching times. Annealing conditions are at 600 °C for 1 hour under 500

Torr of sulfur partial pressure. SnS2-ethyelediamine intercalated compound diffraction pattern was

collected from Ref. 40 CoKα used as a X-ray sources.

5.3.4 Selective etching SnS2, ZnS, CTS from Cu2ZnSnS4 thin films after

annealing in sulfur

We recently reported the successful removal of Cu2-x(S,Se) phases from

annealed CZTSSe thin films.11 Here, we prepared thin films using mixed

dispersions of (i) CZTS and SnS2, (ii) CZTS and ZnS, and (iii) CZTS and CTS and

annealed them in sulfur vapor (500 Torr) at 600 °C for one hour. Before exposure

to the etching solution, Raman scattering maps (Figure 5.6 row A) and spectra

(Figure 5.6 row C) that the annealed thin films contain CZTS and the intentionally

added impurity phase. All Raman spectra exhibit a weak peak at 405 cm-1, and this

peak arises from the under layer of MoS2 that forms upon annealing in sulfur

115

vapor.11 After immersion in the etching solution for 20 minutes at room

temperature, the Raman scattering maps (Figure 5.6 row B) and spectra (Figure

5.6 row D) demonstrate the loss of the impurity phases. These results are

consistent with expectation based on semiquantitative dissolution experiments

with the phase pure powders, with the exception of the apparent removal of SnS2,

which was not expected. No evidence for SnS2 was observed in the post-etching

maps nor spectra, which may suggest that the SnS2 may have delaminated from

the underlying material as a result if intercalation with ethylenediamine. The XRD

data of the CZTS containing SnS2 (Figure C8, in Appendix C) before and after

etching also confirms the elimination of SnS2 in the thin film after the etching.

116

5.4 Conclusions

In summary, our experimental results demonstrate that a 80:20 mixture of

ethylenediamine and 2-mercaptoethanol is an effective etchant for the removal of

SnS2, ZnS, and CTS from annealed thin films of primarily CZTS. Our previous work

demonstrated the successful removal of Cu2-xS from annealed films.

Cu2-xS, ZnS, and CTS are removed via ligand-assisted dissolution. SnS2 is

intercalated with ethylenediamine upon immersion in the etching solution; thus, we

surmise that it is removed from the annealed thin films by delamination upon

intercalation rather than dissolution.

Figure 5.6. Raman scattering maps from annealed thin films from mixture dispersions of (i)

CZTS and SnS2, (ii) CZTS and ZnS, (iii) CZTS and CTS before and after etching (rows A and

B) and the representative Raman spectra of annealed films collected from different regions

shown in maps (rows C and D). Annealing conditions are at 600 °C for 1 hour under 500 Torr

of sulfur partial pressure.

117

The successful application of this etching solution to remove copper

selenide suggests this approach can be generalized to remove other undesired

selenides and tellurides.

118

BIBLIOGRAPHY

References Chapter 1

(1) Anastas, P. T.; Kirchhoff, M. M. Origins, Current Status, and Future Challenges of Green Chemistry. Acc. Chem. Res. 2002, 35, 686–694.

(2) 42 U. S. C. Pollution Prevention Act of 1990; 1990.

(3) Anastas, P. T.; Warner, J. H. Green Chemistry: Theory and Practice; Oxford University Press: New York, 1998.

(4) Web of Science search for the term Green Chemistry from 1990 to 2016 https://www.webofknowledge.com, accessed March 2017.

(5) Clark, J. H. Green Chemistry : Challenges and Opportunities. Green Chem. 1999, 1, 1–8.

(6) Lichtfouse, E. Introduction to Environmental Chemistry Letters. Environ.

Chem. Lett. 2003, 1, 1–1.

(7) Allen, D. Welcome to ACS Sustainable Chemistry & Engineering. ACS

Sustain. Chem. Eng. 2013, 1, 1–1.

(8) Fatta-Kassinos, D.; Lee, Y.; Lim, T. T.; Lima, E. C. Editorial. J. Environ.

Chem. Eng. 2013, 1, 1–1.

(9) United States Environmental Protection Agency (EPA). Presidential Green Chemistry Challenge Winners https://www.epa.gov/greenchemistry/presidential-green-chemistry-challenge-winners (accessed Apr 1, 2017).

(10) European Chemical Sciences (EuCheMS). European Sustainable Chemistry Award http://www.euchems.eu/awards/european-sustainable-chemistry-award/ (accessed Apr 1, 2017).

(11) Arancon, R. A. D.; Lin, C. S. K.; Chan, K. M.; Kwan, T. H.; Luque, R. Advances on Waste Valorization: New Horizons for a More Sustainable Society. Energy Sci. Eng. 2013, 1, 53–71.

(12) Kudo, A.; Miseki, Y. Heterogeneous Photocatalyst Materials for Water Splitting. Chem. Soc. Rev. 2009, 38, 253–278.

(13) Tratnyek, P. G.; Johnson, R. L. Nanotechnologies for Environmental Cleanup. Nano Today 2006, 1, 44–48.

(14) Hutchison, J. E. The Road to Sustainable Nanotechnology: Challenges, Progress and Opportunities. ACS Sustain. Chem. Eng. 2016, 4, 5907–5914.

(15) National Fire Protection Agency (NFPA). NFPA 704 Standard System for the Identification of the Hazards of Materials and Emergency Response

119

http://www.nfpa.org/codes-and-standards/all-codes-and-standards/list-of-codes-and-standards?mode=code&code=704 (accessed May 1, 2017).

(16) Baig, R. B. N.; Varma, R. S. Alternative Energy Input: Mechanochemical, Microwave and Ultrasound-Assisted Organic Synthesis. Chem. Soc. Rev. 2012, 41, 1559–1584.

(17) Gao, M.-R.; Xu, Y.-F.; Jiang, J.; Yu, S.-H. Nanostructured Metal Chalcogenides: Synthesis, Modification, and Applications in Energy Conversion and Storage Devices. Chem. Soc. Rev. 2013, 42, 2986–3017.

(18) Fowler, B. A. Monitoring of Human Populations for Early Markers of Cadmium Toxicity: A Review. Toxicol. Appl. Pharmacol. 2009, 238, 294–300.

(19) Järup, L.; Akesson, A. Current Status of Cadmium as an Environmental Health Problem. Toxicol. Appl. Pharmacol. 2009, 238, 201–208.

(20) Fthenakis, V. M. Life Cycle Impact Analysis of Cadmium in CdTe PV Production. Renew. Sustain. Energy Rev. 2004, 8, 303–334.

(21) Chen, S.; Gong, X. G.; Walsh, A.; Wei, S.-H. Electronic Structure and Stability of Quaternary Chalcogenide Semiconductors Derived from Cation Cross-Substitution of II-VI and I-III-VI2 Compounds. Phys. Rev. B 2009, 79, 165211–165220.

(22) Aldakov, D.; Lefrançois, A.; Reiss, P. Ternary and Quaternary Metal Chalcogenide Nanocrystals: Synthesis, Properties and Applications. J. Mater.

Chem. C 2013, 1, 3756–3776.

(23) Bouaziz, M.; Amlouk, M.; Belgacem, S. Structural and Optical Properties of Cu2SnS3 Sprayed Thin Films. Thin Solid Films 2009, 517, 2527–2530.

(24) Bouaziz, M.; Ouerfelli, J.; Amlouk, M.; Belgacem, S. Structural and Optical Properties of Cu3SnS4 Sprayed Thin Films. Phys. status solidi 2007, 204, 3354–3360.

(25) Marcano, G.; Rincón, C.; de Chalbaud, L. M.; Bracho, D. B.; Pérez, G. S. Crystal Growth and Structure, Electrical, and Optical Characterization of the Semiconductor Cu2SnSe3. J. Appl. Phys. 2001, 90, 1847–1853.

(26) Yan, C.; Huang, C.; Yang, J.; Liu, F.; Liu, J.; Lai, Y.; Li, J.; Liu, Y. Synthesis and Characterizations of Quaternary Cu2FeSnS4 Nanocrystals. Chem. Commun. 2012, 48, 2603–2605.

(27) Meng, X.; Deng, H.; He, J.; Zhu, L.; Sun, L.; Yang, P.; Chu, J. Synthesis of Cu2FeSnSe4 Thin Film by Selenization of RF Magnetron Sputtered Precursor. Mater. Lett. 2014, 117, 1–3.

120

(28) Cui, Y.; Wang, G.; Pan, D. Synthesis and Photoresponse of Novel Cu2CdSnS4 Semiconductor Nanorods. J. Mater. Chem. 2012, 22, 12471–12473.

(29) Liu, F. S.; Zheng, J. X.; Huang, M. J.; He, L. P.; Ao, W. Q.; Pan, F.; Li, J. Q. Enhanced Thermoelectric Performance of Cu2CdSnSe4 by Mn Doping: Experimental and First Principles Studies. Sci. Rep. 2015, 4, 5774 (7pp).

(30) Henry, C. H. Limiting Efficiencies of Ideal Single and Multiple Energy Gap Terrestrial Solar Cells. J. Appl. Phys. 1980, 51, 4494–4500.

(31) Du, Y.-F.; Fan, J.-Q.; Zhou, W.-H.; Zhou, Z.-J.; Jiao, J.; Wu, S.-X. One-Step Synthesis of Stoichiometric Cu2ZnSnSe4 as Counter Electrode for Dye-Sensitized Solar Cells. ACS Appl. Mater. Interfaces 2012, 4, 1796–1802.

(32) Kush, P.; Deka, S. Anisotropic Kesterite Cu2ZnSnSe4 Colloidal Nanoparticles: Photoelectrical and Photocatalytic Properties. Mater. Chem. Phys. 2015, 162, 608–616.

(33) Yu, X.; Shavel, A.; An, X.; Luo, Z.; Ibáñez, M.; Cabot, A. Cu2ZnSnS4-Pt and Cu2ZnSnS4-Au Heterostructured Nanoparticles for Photocatalytic Water Splitting and Pollutant Degradation. J. Am. Chem. Soc. 2014, 136, 9236–9239.

(34) Chen, S.; Walsh, A.; Luo, Y.; Yang, J. H.; Gong, X. G.; Wei, S. H. Wurtzite-Derived Polytypes of Kesterite and Stannite Quaternary Chalcogenide Semiconductors. Phys. Rev. B 2010, 82, 1–8.

(35) Jiang, H.; Dai, P.; Feng, Z.; Fan, W.; Zhan, J. Phase Selective Synthesis of Metastable Orthorhombic Cu2ZnSnS4. J. Mater. Chem. 2012, 22, 7502–7506.

(36) Khare, A.; Himmetoglu, B.; Johnson, M.; Norris, D. J.; Cococcioni, M.; Aydil, E. S. Calculation of the Lattice Dynamics and Raman Spectra of Copper Zinc Tin Chalcogenides and Comparison to Experiments. J. Appl. Phys. 2012, 111, 083707–083715.

(37) Chen, S.; Gong, X. G.; Walsh, A.; Wei, S.; Chen, S.; Gong, X. G.; Walsh, A.; Wei, S. Crystal and Electronic Band Structure of Cu2ZnSnS4 (X=S and Se) Photovoltaic Absorbers: First-Principles Insights. Appl. Phys Lett 2013, 94, 041903–041905.

(38) Ramkumar, S. P.; Gillet, Y.; Miglio, A.; Setten, M. J. Van; Gonze, X.; Rignanese, G. First-Principles Investigation of the Structural , Dynamical , and Dielectric Properties of Kesterite , Stannite , and PMCA Phases of Cu2ZnSnS4. Phys. Rev. B 2016, 94, 224302–224312.

(39) Nitsche, R.; Sargent, D. F.; Wild, P. Crystal Growth of Quaternary 122464 Chalcogenides by Iodine Vapor Transport. J. Cryst. Growth 1967, 1, 52–53.

(40) Ito, K.; Nakazawa, T. Electrical and Optical Properties of Stannite-Type

121

Quaternary Semiconductor Thin Films. Jpn. J. Appl. Phys. 1988, 27, 2094–2097.

(41) Katagiri, H.; Ishigaki, N.; Ishida, T.; Saito, K. Characterization of Cu2ZnSnS4 Thin Films Prepared by Vapor Phase Sulfurization. Jpn. J. Appl. Phys. 2001, 40, 500–504.

(42) Todorov, T. K.; Reuter, K. B.; Mitzi, D. B. High-Efficiency Solar Cell with Earth-Abundant Liquid-Processed Absorber. Adv. Mater. 2010, 22, E156–E159.

(43) Wang, W.; Winkler, M. T.; Gunawan, O.; Gokmen, T.; Todorov, T. K.; Zhu, Y.; Mitzi, D. B. Device Characteristics of CZTSSe Thin-Film Solar Cells with 12.6% Efficiency. Adv. Energy Mater. 2014, 4, 1–5.

(44) Yuan, M.; Mitzi, D. B. Solvent Properties of Hydrazine in the Preparation of Metal Chalcogenide Bulk Materials and Films. Dalt. Trans. 2009, 6078–6088.

(45) Hydrazine Safety Data Sheet (SDS), Sigma Aldrich http://www.sigmaaldrich.com/catalog/product/sial/215155?lang=en&region=US, accessed June 2017.

(46) Guo, Q.; Hillhouse, H. W.; Agrawal, R. Synthesis of Cu2ZnSnS4 Nanocrystal Ink and Its Use for Solar Cells. J. Am. Chem. Soc. 2009, 131, 11672–11673.

(47) Collord, A. D.; Hillhouse, H. W. Composition Control and Formation Pathway of CZTS and CZTGS Nanocrystal Inks for Kesterite Solar Cells. Chem.

Mater. 2015, 27, 1855–1862.

(48) Khare, A.; Wills, A. W.; Ammerman, L. M.; Norris, D. J.; Aydil, E. S. Size Control and Quantum Confinement in Cu2ZnSnS4 Nanocrystals. Chem. Commun. 2011, 47, 11721–11723.

(49) Tosun, B. S.; Chernomordik, B. D.; Gunawan, A. A.; Williams, B.; Mkhoyan, K. A.; Francis, L. F.; Aydil, E. S. Cu2ZnSnS4 Nanocrystal Dispersions in Polar Liquids. Chem. Commun. 2013, 49, 3549–3551.

(50) Becheri, A.; Dürr, M.; Lo Nostro, P.; Baglioni, P. Synthesis and Characterization of Zinc Oxide Nanoparticles: Application to Textiles as UV-Absorbers. J. Nanoparticle Res. 2008, 10, 679–689.

(51) Das, S.; Chaudhuri, S.; Maji, S. Ethanol−Water Mediated Solvothermal Synthesis of Cube and Pyramid Shaped Nanostructured Tin Oxide. J. Phys. Chem.

C 2008, 112, 6213–6219.

(52) Wang, J.; Zhang, P.; Song, X.; Gao, L. Surfactant-Free Hydrothermal Synthesis of Cu2ZnSnS4 (CZTS) Nanocrystals with Photocatalytic Properties. RSC

Adv. 2014, 4, 27805–27810.

(53) Sarkar, S.; Bhattacharjee, K.; Das, G. C.; Chattopadhyay, K. K. Self-

122

Sacrificial Template Directed Hydrothermal Route to Kesterite-Cu2ZnSnS4 Microspheres and Study of Their Photo Response Properties. CrystEngComm 2014, 16, 2634–2644.

(54) Xia, Y.; Chen, Z.; Zhang, Z.; Fang, X.; Liang, G. A Nontoxic and Low-Cost Hydrothermal Route for Synthesis of Hierarchical Cu2ZnSnS4 Particles. Nanoscale

Res. Lett. 2014, 9, 208 (7 pp).

(55) Kappe, C. O. Controlled Microwave Heating in Modern Organic. Angew.

Chemie-International Ed. 2004, 43, 6250–6284.

(56) Ethylene glycol Safety Data Sheet (SDS), Sigma Aldrich http://www.sigmaaldrich.com/catalog/product/sial/324558?lang=en&region=US, accessed June 2017.

(57) Flynn, B.; Wang, W.; Chang, C.; Herman, G. S. Microwave Assisted Synthesis of Cu2ZnSnS4 Colloidal Nanoparticle Inks. Phys. Status Solidi 2012, 209, 2186–2194.

(58) Long, F.; Mo, S.; Zeng, Y.; Chi, S.; Zou, Z. Synthesis of Flower-Like Cu2ZnSnS4 Nanoflakes via a Microwave-Assisted Solvothermal Route. Int. J.

Photoenergy 2014, 2014, 618789 (4 pp).

(59) Yan, X.; Michael, E.; Komarneni, S.; Brownson, J. R.; Yan, Z. F. Microwave-Hydrothermal/solvothermal Synthesis of Kesterite, an Emerging Photovoltaic Material. Ceram. Int. 2014, 40, 1985–1992.

(60) Ghediya, P. R.; Chaudhuri, T. K. Doctor-Blade Printing of Cu2ZnSnS4 Films from Microwave-Processed Ink. J. Mater. Sci. Mater. Electron. 2014, 26, 1908–1912.

(61) Tian, Q.; Wang, G.; Zhao, W.; Chen, Y.; Yang, Y.; Huang, L.; Pan, D. Versatile and Low-Toxic Solution Approach to Binary, Ternary, and Quaternary Metal Sulfide Thin Films and Its Application in Cu2ZnSn(S,Se)4 Solar Cells. Chem.

Mater. 2014, 26, 3098–3103.

(62) Suryawanshi, M.; Shin, S. W.; Ghorpade, U.; Song, D.; Hong, C. W.; Han, S.-S.; Heo, J.; Kang, S. H.; Kim, J. H. A Facile and Green Synthesis of Colloidal Cu2ZnSnS4 Nanocrystals and Their Application in Highly Efficient Solar Water Splitting. J. Mater. Chem. A 2017, 5, 4695–4709.

(63) Chernomordik, B. D.; Béland, A. E.; Trejo, N. D.; Gunawan, A. A.; Deng, D. D.; Mkhoyan, K. A.; Aydil, E. S. Rapid Facile Synthesis of Cu2ZnSnS4 Nanocrystals. J. Mater. Chem. A 2014, 2, 10389–10395.

(64) Park, B.-I.; Hwang, Y.; Lee, S. Y.; Lee, J.-S.; Park, J.-K.; Jeong, J.; Kim, J. Y.; Kim, B.; Cho, S.-H.; Lee, D.-K. Solvent-Free Synthesis of Cu2ZnSnS4

123

Nanocrystals: A Facile, Green, up-Scalable Route for Low Cost Photovoltaic Cells. Nanoscale 2014, 6, 11703–11711.

(65) Williams, B. A.; Mahajan, A.; Smeaton, M. A.; Holgate, C. S.; Aydil, E. S.; Francis, L. F. Formation of Copper Zinc Tin Sulfide Thin Films from Colloidal Nanocrystal Dispersions via Aerosol-Jet Printing and Compaction. ACS Appl.

Mater. Interfaces 2015, 7, 11526–11535.

(66) Ren, Y.; Ross, N.; Larsen, J. K.; Rudisch, K.; Scragg, J. J. S.; Platzer-Björkman, C. Evolution of Cu2ZnSnS4 during Non-Equilibrium Annealing with Quasi-in Situ Monitoring of Sulfur Partial Pressure. Chem. Mater. 2017, 29, 3713–3722.

(67) Chernomordik, B. D.; Béland, A. E.; Deng, D. D.; Francis, L. F.; Aydil, E. S. Microstructure Evolution and Crystal Growth in Cu2ZnSnS4 Thin Films Formed by Annealing Colloidal Nanocrystal Coatings. Chem. Mater. 2014, 26, 3191–3201.

(68) Chernomordik, B. D.; Ketkar, P. M.; Hunter, A. K.; Béland, A. E.; Deng, D. D.; Aydil, E. S. Microstructure Evolution during Selenization of Cu2ZnSnS4 Colloidal Nanocrystal Coatings. Chem. Mater. 2016, 28, 1266–1276.

(69) Williams, B. A.; Smeaton, M. A.; Trejo, N. D.; Francis, L. F.; Aydil, E. S. Effect of Nanocrystal Size and Carbon on Grain Growth during Annealing of Copper Zinc Tin Sulfide Nanocrystal Coatings. Chem. Mater. 2017, 29, 1676–1683.

(70) Johnson, M. C.; Wrasman, C.; Zhang, X.; Manno, M.; Leighton, C.; Aydil, E. S. Self-Regulation of Cu/Sn Ratio in the Synthesis of Cu2ZnSnS4 Films. Chem.

Mater. 2015, 27, 2507–2514.

(71) Pinto, A. H.; Shin, S. W.; Aydil, E. S.; Penn, R. L. Selective Removal of Cu2−x(S,Se) Phases from Cu2ZnSn(S,Se) 4 Thin Films. Green Chem. 2016, 18, 5814–5821.

(72) Knutson, T. R.; Hanson, P. J.; Aydil, E. S.; Penn, R. L. Synthesis of Cu2ZnSnS4 Thin Films Directly onto Conductive Substrates via Selective Thermolysis Using Microwave Energy. Chem. Commun. 2014, 50, 5902–5904.

124

References Chapter 2

(1) Green, M. A. Thin-Film Solar Cells: Review of Materials, Technologies and Commercial Status. J. Mater. Sci. Mater. Electron. 2007, 18, 15–19.

(2) Wolden, C. A.; Kurtin, J.; Baxter, J. B.; Repins, I.; Shaheen, S. E.; Torvik, J. T.; Rockett, A. A.; Fthenakis, V. M.; Aydil, E. S. Photovoltaic Manufacturing: Present Status, Future Prospects, and Research Needs. J. Vac. Sci. Technol. 2011, 29, 030801–030815.

(3) Green, M. A.; Emery, K.; Hishikawa, Y.; Warta, W.; Dunlop, E. D. Solar Cell Efficiency Tables (Version 45). Prog. Photovoltaics 2015, 23, 1–9.

(4) Liu, H.; Avrutin, V.; Izyumskaya, N.; Özgür, Ü.; Morkoç, H. Transparent Conducting Oxides for Electrode Applications in Light Emitting and Absorbing Devices. Superlattices Microstruct. 2010, 48, 458–484.

(5) Steinhagen, C.; Panthani, M. G.; Akhavan, V.; Goodfellow, B.; Koo, B.; Korgel, B. A. Synthesis of Cu2ZnSnS4 Nanocrystals for Use in Low-Cost Photovoltaics. J. Am. Chem. Soc. 2009, 131, 12554–12555.

(6) Huang, T. J.; Yin, X.; Qi, G.; Gong, H. CZTS-Based Materials and Interfaces and Their Effects on the Performance of Thin Film Solar Cells. Phys. Status Solidi

- Rapid Res. Lett. 2014, 8, 735–762.

(7) Collord, A. D.; Hillhouse, H. W. Composition Control and Formation Pathway of CZTS and CZTGS Nanocrystal Inks for Kesterite Solar Cells. Chem.

Mater. 2015, 27, 1855–1862.

(8) Kim, J.; Hiroi, H.; Todorov, T. K.; Gunawan, O.; Kuwahara, M.; Gokmen, T.; Nair, D.; Hopstaken, M.; Shin, B.; Lee, Y. S.; et al. High Efficiency Cu2ZnSn(S,Se)4 Solar Cells by Applying a Double In2S3/CdS Emitter. Adv. Mater. 2014, 26, 7427–7431.

(9) Khare, A.; Himmetoglu, B.; Johnson, M.; Norris, D. J.; Cococcioni, M.; Aydil, E. S. Calculation of the Lattice Dynamics and Raman Spectra of Copper Zinc Tin Chalcogenides and Comparison to Experiments. J. Appl. Phys. 2012, 111, 083707–083715.

(10) Chen, S.; Walsh, A.; Luo, Y.; Yang, J. H.; Gong, X. G.; Wei, S. H. Wurtzite-Derived Polytypes of Kesterite and Stannite Quaternary Chalcogenide Semiconductors. Phys. Rev. B 2010, 82, 1–8.

(11) Jiang, H.; Dai, P.; Feng, Z.; Fan, W.; Zhan, J. Phase Selective Synthesis of Metastable Orthorhombic Cu2ZnSnS4. J. Mater. Chem. 2012, 22, 7502–7506.

(12) Schorr, S.; Hoebler, H.-J.; Tovar, M. A Neutron Diffraction Study of the Stannite-Kesterite Solid Solution Series. Eur. J. Mineral. 2007, 19, 65–73.

125

(13) Regulacio, M. D.; Ye, C.; Lim, S. H.; Bosman, M.; Ye, E.; Chen, S.; Xu, Q.-H.; Han, M.-Y. Colloidal Nanocrystals of Wurtzite-Type Cu2ZnSnS4: Facile Noninjection Synthesis and Formation Mechanism. Chem. A Eur. J. 2012, 18, 3127–3131.

(14) Lu, X.; Zhuang, Z.; Peng, Q.; Li, Y. Wurtzite Cu2ZnSnS4 Nanocrystals: A Novel Quaternary Semiconductor. Chem. Commun. 2011, 47, 3141–3143.

(15) Singh, A.; Geaney, H.; Laffir, F.; Ryan, K. M. Colloidal Synthesis of Wurtzite Cu2ZnSnS4 Nanorods and Their Perpendicular Assembly. J. Am. Chem. Soc. 2012, 134, 2910–2913.

(16) Lin, X.; Kavalakkatt, J.; Kornhuber, K.; Abou-Ras, D.; Schorr, S.; Lux-Steiner, M. C.; Ennaoui, A. Synthesis of Cu2ZnxSnySe1+x+2y Nanocrystals with Wurtzite-Derived Structure. RSC Adv. 2012, 2, 9894–9898.

(17) Wang, Y.; Wei, M.; Fan, F.; Zhuang, T.; Wu, L.; Yu, S.; Zhu, C. Phase-Selective Synthesis of Cu2ZnSnS4 Nanocrystals through Cation Exchange for Photovoltaic Devices. Chem. Mater. 2014, 26, 5492–5498.

(18) Li, Z.; Lui, A. L. K.; Lam, K. H.; Xi, L.; Lam, Y. M. Phase-Selective Synthesis of Cu2ZnSnS4 Nanocrystals Using Different Sulfur Precursors. Inorg. Chem. 2014, 53, 10874–10880.

(19) Lin, Y.-H.; Das, S.; Yang, C.-Y.; Sung, J.-C.; Lu, C.-H. Phase-Controlled Synthesis of Cu2ZnSnS4 Powders via the Microwave-Assisted Solvothermal Route. J. Alloys Compd. 2015, 632, 354–360.

(20) Hillhouse, H. W.; Beard, M. C. Solar Cells from Colloidal Nanocrystals: Fundamentals, Materials, Devices, and Economics. Curr. Opin. Colloid Interface

Sci. 2009, 14, 245–259.

(21) Guo, Q.; Hillhouse, H. W.; Agrawal, R. Synthesis of Cu2ZnSnS4 Nanocrystal Ink and Its Use for Solar Cells. J. Am. Chem. Soc. 2009, 131, 11672–11673.

(22) Chernomordik, B. D.; Béland, A. E.; Deng, D. D.; Francis, L. F.; Aydil, E. S. Microstructure Evolution and Crystal Growth in Cu2ZnSnS4 Thin Films Formed by Annealing Colloidal Nanocrystal Coatings. Chem. Mater. 2014, 26, 3191–3201.

(23) Chernomordik, B. D.; Ketkar, P. M.; Hunter, A. K.; Béland, A. E.; Deng, D. D.; Aydil, E. S. Microstructure Evolution during Selenization of Cu2ZnSnS4 Colloidal Nanocrystal Coatings. Chem. Mater. 2016, 28, 1266–1276.

(24) Mainz, R.; Singh, a; Levcenko, S.; Klaus, M.; Genzel, C.; Ryan, K. M.; Unold, T. Phase-Transition-Driven Growth of Compound Semiconductor Crystals from Ordered Metastable Nanorods. Nat. Commun. 2014, 5, 3133.

(25) Yang, W.-C.; Miskin, C. K.; Hages, C. J.; Hanley, E. C.; Handwerker, C.;

126

Stach, E. A.; Agrawal, R. Kesterite Cu2ZnSn(S,Se)4 Absorbers Converted from Metastable, Wurtzite-Derived Cu2ZnSnS4 Nanoparticles. Chem. Mater. 2014, 26, 3530–3534.

(26) Lin, X.; Kavalakkatt, J.; Ennaoui, A.; Lux-Steiner, M. C. Cu2ZnSn(S, Se)4 Thin Film Absorbers Based on ZnS, SnS and Cu3SnS4 Nanoparticle Inks: Enhanced Solar Cells Performance by Using a Two-Step Annealing Process. Sol.

Energy Mater. Sol. Cells 2015, 132, 221–229.

(27) Himmrich, M.; Haeuseler, H. Far Infrared Studies on Stannite and Wurtzstannite Type Compounds. Spectrochim. Acta Part A Mol. Spectrosc. 1991, 47, 933–942.

(28) Cheng, A.-J.; Manno, M.; Khare, A.; Leighton, C.; Campbell, S. A.; Aydil, E. S. Imaging and Phase Identification of Cu2ZnSnS4 Thin Films Using Confocal Raman Spectroscopy. J. Vac. Sci. Technol. 2011, 29, 051203 (11pp).

(29) Valakh, M. Y.; Kolomys, O. F.; Ponomaryov, S. S.; Yukhymchuk, V. O.; Babichuk, I. S.; Izquierdo-Roca, V.; Saucedo, E.; Perez-Rodriguez, A.; Morante, J. R.; Schorr, S.; Bodnar, I. V. Raman Scattering and Disorder Effect in Cu2ZnSnS4. Phys. Status Solidi - Rapid Res. Lett. 2013, 7, 258–261.

(30) Ahmad, R.; Brandl, M.; Distaso, M.; Herre, P.; Spiecker, E.; Hock, R.; Peukert, W. A Comprehensive Study on the Mechanism behind Formation and Depletion of Cu2ZnSnS4 (CZTS) Phases. CrystEngComm 2015, 17, 6972–6984.

(31) Arora, L.; Gupta, P.; Chhikara, N.; Singh, O. P.; Muhunthan, N.; Singh, V. N.; Singh, B. P.; Jain, K.; Chand, S. Green Synthesis of Wurtzite Copper Zinc Tin Sulfide Nanocones for Improved Solar Photovoltaic Utilization. Appl. Nanosci. 2015, 5, 163–167.

(32) Wang, Y.; Jiang, X.; Xia, Y. A Solution-Phase, Precursor Route to Polycrystalline SnO2 Nanowires That Can Be Used for Gas Sensing under Ambient Conditions. J. Am. Chem. Soc. 2003, 125, 16176–16177.

(33) Ng, S. H.; Chew, S. Y.; Dos Santos, D. I.; Chen, J.; Wang, J. Z.; Dou, S. X.; Liu, H. K. Hexagonal-Shaped Tin Glycolate Particles: A Preliminary Study of Their Suitability as Li-Ion Insertion Electrodes. Chem. - An Asian J. 2008, 3, 854–861.

(34) Das, J.; Evans, I. R.; Khushalani, D. Zinc Glycolate: A Precursor to ZnO. Inorg. Chem. 2009, 48, 3508–3510.

(35) Jayalakshmi, D.; Kumar, J. Growth and Characterization of Bis Thiourea Zinc Acetate (BTZA). Cryst. Res. Technol. 2006, 41, 37–40.

(36) Lydia Caroline, M.; Vasudevan, S. Growth and Characterization of Pure and Doped Bis Thiourea Zinc Acetate: Semiorganic Nonlinear Optical Single Crystals.

127

Curr. Appl. Phys. 2009, 9, 1054–1061.

(37) Cassidy, J. E.; Moser, W.; Donaldson, J. D.; Jelen, A.; Nicholson, D. G. Thiourea Complexes of tin(II) Compounds. Journal of the Chemical Society A:

Inorganic, Physical, Theoretical, 1970, 173–175.

(38) Harrison, P. G.; Haylett, B. J.; King, T. J. The Crystal and Molecular Structure of dichloro(thiourea)tin(II). Inorganica Chim. Acta 1983, 75, 259–264.

(39) Burrows, A. D.; Donovan, A. S.; Harrington, R. W.; Mahon, M. F. Backbone Flexibility and Counterion Effects on the Structure and Thermal Properties of Di(thiourea)zinc Dicarboxylate Coordination Polymers. Eur. J. Inorg. Chem. 2004, 2, 4686–4695.

(40) Hilbert, J.; Näther, C.; Weihrich, R.; Bensch, W. Room-Temperature Synthesis of Thiostannates from {[Ni(tren)]2[Sn2S6]}n . Inorg. Chem. 2016, 55, 7859–7865.

(41) Pearson, R. G. Hard and Soft Acids and Bases, HSAB, Part 1: Fundamental Principles. J. Chem. Educ. 1968, 45, 581–587.

(42) Li, M.; Zhou, W.-H.; Guo, J.; Zhou, Y.-L.; Hou, Z.; Jiao, J.; Zhou, Z.; Du, Z.; Wu, S. Synthesis of Pure Metastable Wurtzite CZTS Nanocrystals by Facile One-Pot Method. J. Phys. Chem. C 2012, 116, 26507–26516.

(43) Li, J.; Bloemen, M.; Parisi, J.; Kolny-Olesiak, J. Role of Copper Sulfide Seeds in the Growth Process of CuInS2 Nanorods and Networks. ACS Appl. Mater.

Interfaces 2014, 6, 20535–20543.

(44) Morin, S. A.; Forticaux, A.; Bierman, M. J.; Jin, S. Screw Dislocation-Driven Growth of Two-Dimensional Nanoplates. Nano Lett. 2011, 11, 4449–4455.

(45) Dziedzic, R. M.; Gillian-Daniel, A. L.; Petersen, G. M.; Martinez-Hernandez, K. J. Microwave Synthesis of Zinc Hydroxy Sulfate Nanoplates and Zinc Oxide Nanorods in the Classroom. J. Chem. Educ. 2014, 91, 1710–1714.

(46) Yu, H.; Yu, J.; Liu, S.; Mann, S. Template-Free Hydrothermal Synthesis of CuO / Cu2O Composite Hollow Microspheres. Chem. Mater. 2007, 19, 4327–4334.

(47) Lou, X. W.; Archer, L. A.; Yang, Z. Hollow Micro-/nanostructures: Synthesis and Applications. Adv. Mater. 2008, 20, 3987–4019.

128

References Chapter 3

(1) Katagiri, H.; Sasaguchi, N.; Hando, S.; Hoshino, S.; Ohashi, J.; Yokota, T. Preparation and Evaluation of Cu2ZnSnS4 Thin Films by Sulfurization of E-B Evaporated Precursors. Sol. Energy Mater. Sol. Cells 1997, 49, 407–414.

(2) Mitzi, D. B.; Gunawan, O.; Todorov, T. K.; Wang, K.; Guha, S. The Path towards a High-Performance Solution-Processed Kesterite Solar Cell. Sol. Energy

Mater. Sol. Cells 2011, 95, 1421–1436.

(3) Ito, K.; Nakazawa, T. Electrical and Optical Properties of Stannite-Type Quaternary Semiconductor Thin Films. Jpn. J. Appl. Phys. 1988, 27, 2094–2097.

(4) Todorov, T. K.; Reuter, K. B.; Mitzi, D. B. High-Efficiency Solar Cell with Earth-Abundant Liquid-Processed Absorber. Adv. Mater. 2010, 22, E156–E159.

(5) Shockley, W.; Queisser, H. J. Detailed Balance Limit of Efficiency of p-n Junction Solar Cells. J. Appl. Phys. 1961, 32, 510–519.

(6) Kim, J.; Hiroi, H.; Todorov, T. K.; Gunawan, O.; Kuwahara, M.; Gokmen, T.; Nair, D.; Hopstaken, M.; Shin, B.; Lee, Y. S.; et al. High Efficiency Cu2ZnSn(S,Se)4 Solar Cells by Applying a Double In2S3/CdS Emitter. Adv. Mater. 2014, 26, 7427–7431.

(7) Jackson, P.; Hariskos, D.; Wuerz, R.; Kiowski, O.; Bauer, A.; Friedlmeier, T. M.; Powalla, M. Properties of Cu(In,Ga)Se2 Solar Cells with New Record Efficiencies up to 21.7%. Phys. Status Solidi – Rapid Res. Lett. 2015, 9, 28–31.

(8) Ki, W.; Hillhouse, H. W. Earth-Abundant Element Photovoltaics Directly from Soluble Precursors with High Yield Using a Non-Toxic Solvent. Adv. Energy

Mater. 2011, 1, 732–735.

(9) Altamura, G.; Vidal, J. Impact of Minor Phases on the Performances of CZTSSe Thin-Film Solar Cells. Chem. Mater. 2016, 28, 3540–3563.

(10) Yan, C.; Liu, F.; Sun, K.; Song, N.; Stride, J. A.; Zhou, F.; Hao, X.; Green, M. Boosting the Efficiency of Pure Sulfide CZTS Solar Cells Using the In/Cd-Based Hybrid Buffers. Sol. Energy Mater. Sol. Cells 2016, 144, 700–706.

(11) Halliday, D. P.; Claridge, R.; Goodman, M. C. J.; Mendis, B. G.; Durose, K.; Major, J. D. Luminescence of Cu2ZnSnS4 Polycrystals Described by the Fluctuating Potential Model. J. Appl. Phys. 2013, 113, 223503–223512.

(12) Van Puyvelde, L.; Lauwaert, J.; Smet, P. F.; Khelifi, S.; Ericson, T.; Scragg, J. J.; Poelman, D.; Van Deun, R.; Platzer-Björkman, C.; Vrielinck, H. Photoluminescence Investigation of Cu2ZnSnS4 Thin Film Solar Cells. Thin Solid

Films 2015, 582, 146–150.

129

(13) Washio, T.; Nozaki, H.; Fukano, T.; Motohiro, T.; Jimbo, K.; Katagiri, H. Analysis of Lattice Site Occupancy in Kesterite Structure of Cu2ZnSnS4 Films Using Synchrotron Radiation X-Ray Diffraction. J. Appl. Phys. 2011, 110, 74511–74514.

(14) Collord, A. D.; Hillhouse, H. W. Germanium Alloyed Kesterite Solar Cells with Low Voltage Deficits. Chem. Mater. 2016, 28, 2067–2073.

(15) Cui, Y.; Deng, R.; Wang, G.; Pan, D. A General Strategy for Synthesis of Quaternary Semiconductor Cu2MSnS4 (M = Co2+, Fe2+, Ni2+, Mn2+) Nanocrystals. J. Mater. Chem. 2012, 22, 23136-23140.

(16) Thompson, M. J.; Blakeney, K. J.; Cady, S. D.; Reichert, M. D.; Pilar-Albaladejo, J. Del; White, S. T.; Vela, J. Cu2ZnSnS4 Nanorods Doped with Tetrahedral, High Spin Transition Metal Ions: Mn2+ , Co2+ , and Ni2+. Chem. Mater. 2016, 28, 1668–1677.

(17) Guo, Q.; Hillhouse, H. W.; Agrawal, R. Synthesis of Cu2ZnSnS4 Nanocrystal Ink and Its Use for Solar Cells. J. Am. Chem. Soc. 2009, 131, 11672–11673.

(18) Hillhouse, H. W.; Beard, M. C. Solar Cells from Colloidal Nanocrystals: Fundamentals, Materials, Devices, and Economics. Curr. Opin. Colloid Interface

Sci. 2009, 14, 245–259.

(19) Chernomordik, B. D.; Béland, A. E.; Deng, D. D.; Francis, L. F.; Aydil, E. S. Microstructure Evolution and Crystal Growth in Cu2ZnSnS4 Thin Films Formed by Annealing Colloidal Nanocrystal Coatings. Chem. Mater. 2014, 26, 3191–3201.

(20) Akhavan, V. a.; Goodfellow, B. W.; Panthani, M. G.; Steinhagen, C.; Harvey, T. B.; Stolle, C. J.; Korgel, B. a. Colloidal CIGS and CZTS Nanocrystals: A Precursor Route to Printed Photovoltaics. J. Solid State Chem. 2012, 189, 2–12.

(21) Steinhagen, C.; Panthani, M. G.; Akhavan, V.; Goodfellow, B.; Koo, B.; Korgel, B. A. Synthesis of Cu2ZnSnS4 Nanocrystals for Use in Low-Cost Photovoltaics. J. Am. Chem. Soc. 2009, 131, 12554–12555.

(22) Zhou, H.; Hsu, W.-C.; Duan, H.-S.; Bob, B.; Yang, W.; Song, T.-B.; Hsu, C.-J.; Yang, Y. CZTS Nanocrystals: A Promising Approach for next Generation Thin Film Photovoltaics. Energy Environ. Sci. 2013, 6, 2822.

(23) Yang, H.; Jauregui, L. A.; Zhang, G.; Chen, Y. P.; Wu, Y. Nontoxic and Abundant Copper Zinc Tin Sulfide Nanocrystals for Potential High-Temperature Thermoelectric Energy Harvesting. Nano Lett. 2012, 12, 540–545.

(24) Collord, A. D.; Hillhouse, H. W. Composition Control and Formation Pathway of CZTS and CZTGS Nanocrystal Inks for Kesterite Solar Cells. Chem.

Mater. 2015, 27, 1855–1862.

130

(25) Gillorin, A.; Balocchi, A.; Marie, X.; Dufour, P.; Chane-Ching, J. Y. Synthesis and Optical Properties of Cu2CoSnS4 Colloidal Quantum Dots. J. Mater. Chem. 2011, 21, 5615–5619.

(26) Zhang, X.; Bao, N.; Lin, B.; Gupta, A. Colloidal Synthesis of Wurtzite Cu2CoSnS4 Nanocrystals and the Photoresponse of Spray-Deposited Thin Films. Nanotechnology 2013, 24, 105706.

(27) Murali, B.; Krupanidhi, S. B. Facile Synthesis of Cu2CoSnS4 Nanoparticles Exhibiting Red-Edge-Effect: Application in Hybrid Photonic Devices. J. Appl. Phys. 2013, 114, 144312.

(28) Murali, B.; Madhuri, M.; Krupanidhi, S. B. Solution Processed Cu2CoSnS4 Thin Films for Photovoltaic Applications. Cryst. Growth Des. 2014, 14, 3685–3691.

(29) Krishnaiah, M.; Bhargava, P.; Mallick, S. Low-Temperature Synthesis of Cu2CoSnS4 Nanoparticles by Thermal Decomposition of Metal Precursors and Study of Its Structural, Optical and Electrical Properties for Photovoltaic Application. RSC Adv. 2015, 5, 96928–96933.

(30) Zhong, J.; Wang, Q.; Cai, W. Rapid Synthesis of Flower-like Cu2CoSnS4 Microspheres with Nanoplates Using a Biomolecule-Assisted Method. Mater. Lett. 2015, 150, 69–72.

(31) Huang, K.-L.; Huang, C.-H.; Lin, W.-T.; Fu, Y.-S.; Guo, T.-F. Solvothermal Synthesis and Tunable Bandgap of Cu2(Zn1−xCox)SnS4 and Cu2(Fe1−xCox)SnS4 Nanocrystals. J. Alloys Compd. 2015, 646, 1015–1022.

(32) Hoshino, A.; Fujioka, K.; Oku, T.; Suga, M.; Sasaki, Y. F.; Ohta, T.; Yasuhara, M.; Suzuki, K.; Yamamoto, K. Physicochemical Properties and Cellular Toxicity of Nanocrystal Quantum Dots Depend on Their Surface Modification. Nano Lett. 2004, 4, 2163–2169.

(33) Andelman, T.; Gordonov, S.; Busto, G.; Moghe, P. V.; Riman, R. E. Synthesis and Cytotoxicity of Y2O3 Nanoparticles of Various Morphologies. Nanoscale Res. Lett. 2010, 5, 263–273.

(34) Baghbanzadeh, M.; Carbone, L.; Cozzoli, P. D.; Kappe, C. O. Microwave-Assisted Synthesis of Colloidal Inorganic Nanocrystals. Angew. Chemie Int. Ed. 2011, 50, 11312–11359.

(35) Chernomordik, B. D.; Béland, A. E.; Trejo, N. D.; Gunawan, A. A.; Deng, D. D.; Mkhoyan, K. A.; Aydil, E. S. Rapid Facile Synthesis of Cu2ZnSnS4 Nanocrystals. J. Mater. Chem. A 2014, 2, 10389–10395.

(36) Chernomordik, B. D.; Ketkar, P. M.; Hunter, A. K.; Béland, A. E.; Deng, D. D.; Aydil, E. S. Microstructure Evolution During Selenization of Cu2ZnSnS4

131

Colloidal Nanocrystal Coatings. Chem. Mater. 2016, 28, 1266–1276.

(37) Williams, B. A.; Mahajan, A.; Smeaton, M. A.; Holgate, C. S.; Aydil, E. S.; Francis, L. F. Formation of Copper Zinc Tin Sulfide Thin Films from Colloidal Nanocrystal Dispersions via Aerosol-Jet Printing and Compaction. ACS Appl.

Mater. Interfaces 2015, 7, 11526–11535.

(38) Williams, B. A.; Smeaton, M. A.; Holgate, C. S.; Trejo, N. D.; Francis, L. F.; Aydil, E. S. Intense Pulsed Light Annealing of Copper Zinc Tin Sulfide Nanocrystal Coatings. J. Vac. Sci. Technol. A Vacuum, Surfaces, Film. 2016, 34, 51204 (13 pp).

(39) Johnson, M.; Baryshev, S. V.; Thimsen, E.; Manno, M.; Zhang, X.; Veryovkin, I. V.; Leighton, C.; Aydil, E. S. Alkali-Metal-Enhanced Grain Growth in Cu2ZnSnS4 Thin Films. Energy Environ. Sci. 2014, 7, 1931–1938.

(40) Pinto, A. H.; Shin, S. W.; Aydil, E. S.; Penn, R. L. Selective Removal of Cu 2−x (S,Se) Phases from Cu2ZnSn(S,Se)4 Thin Films. Green Chem. 2016, 18, 5814–5821.

(41) Lu, X.; Zhuang, Z.; Peng, Q.; Li, Y. Wurtzite Cu2ZnSnS4 Nanocrystals: A Novel Quaternary Semiconductor. Chem. Commun. 2011, 47, 3141–3143.

(42) Khare, A.; Himmetoglu, B.; Johnson, M.; Norris, D. J.; Cococcioni, M.; Aydil, E. S. Calculation of the Lattice Dynamics and Raman Spectra of Copper Zinc Tin Chalcogenides and Comparison to Experiments. J. Appl. Phys. 2012, 111, 083707–083715.

(43) Ghosh, S.; Saha, M.; De, S. K. Tunable Surface Plasmon Resonance and Enhanced Electrical Conductivity of In Doped ZnO Colloidal Nanocrystals. Nanoscale 2014, 6, 7039-7051.

(44) Bryan, J. D.; Schwartz, D. A.; Gamelin, D. R. The Influence of Dopants on the Nucleation of Semiconductor Nanocrystals from Homogeneous Solution. J.

Nanosci. Nanotechnol. 2005, 5, 1472–1479.

(45) Bose, R.; Manna, G.; Pradhan, N. Surface Doping for Hindrance of Crystal Growth and Structural Transformation in Semiconductor Nanocrystals. J. Phys.

Chem. C 2013, 117, 20991–20997.

(46) Zhang, W.; Zhai, L.; He, N.; Zou, C.; Geng, X.; Cheng, L.; Dong, Y.; Huang, S. Solution-Based Synthesis of Wurtzite Cu2ZnSnS4 Nanoleaves Introduced by α-Cu2S Nanocrystals as a Catalyst. Nanoscale 2013, 5, 8114–8121.

(47) Shi, L.; Li, Y.; Zhu, H.; Li, Q. Well-Aligned Quaternary Cu2CoSnS4 Single-Crystalline Nanowires as a Potential Low-Cost Solar Cell Material. Chempluschem 2014, 79, 1638–1642.

132

(48) Das, S.; Krishna, R. M.; Ma, S.; Mandal, K. C. Single Phase Polycrystalline Cu2ZnSnS4 Grown by Vertical Gradient Freeze Technique. J. Cryst. Growth 2013, 381, 148–152.

(49) Gulay, L. D.; Nazarchuk, O. P.; Olekseyuk, I. D. Crystal Structures of the Compounds Cu2CoSi(Ge,Sn)S4 and Cu2CoGe(Sn)Se4 J. Alloys Compd. 2004, 377, 306–311.

(50) Zhao, Y.; Qiao, Q.; Zhou, W.-H.; Cheng, X.-Y.; Kou, D.-X.; Zhou, Z.-J.; Wu, S.-X. Wurtzite Cu2ZnSnS4 Nanospindles with Enhanced Optical and Electrical Properties. Chem. Phys. Lett. 2014, 592, 144–148.

(51) Li, M.; Zhou, W.-H.; Guo, J.; Zhou, Y.-L.; Hou, Z.-L.; Jiao, J.; Zhou, Z.-J.; Du, Z.-L.; Wu, S.-X. Synthesis of Pure Metastable Wurtzite CZTS Nanocrystals by Facile One-Pot Method. J. Phys. Chem. C 2012, 116, 26507–26516.

(52) Singh, A.; Geaney, H.; Laffir, F.; Ryan, K. M. Colloidal Synthesis of Wurtzite Cu2ZnSnS4 Nanorods and Their Perpendicular Assembly. J. Am. Chem. Soc. 2012, 134, 2910–2913.

(53) Tan, J. M. R.; Lee, Y. H.; Pedireddy, S.; Baikie, T.; Ling, X. Y.; Wong, L. H. Understanding the Synthetic Pathway of a Single-Phase Quarternary Semiconductor Using Surface-Enhanced Raman Scattering: A Case of Wurtzite Cu2ZnSnS4 Nanoparticles. J. Am. Chem. Soc. 2014, 136, 6684–6692.

(54) Guan, H.; Hou, H.; Yu, F.; Li, L. Synthesis of Wurtzite Cu2ZnSnS4 Thin Films Directly on Glass Substrates by the Solvothermal Method. Mater. Lett. 2015, 159, 200–203.

(55) Khadka, D. B.; Kim, J. Structural Transition and Band Gap Tuning of Cu2(Zn,Fe)SnS4 Chalcogenide for Photovoltaic Application. J. Phys. Chem. C 2014, 118, 14227-14237.

(56) Cheng, A.-J.; Manno, M.; Khare, A.; Leighton, C.; Campbell, S. A.; Aydil, E. S. Imaging and Phase Identification of Cu2ZnSnS4 Thin Films Using Confocal Raman Spectroscopy. J. Vac. Sci. Technol. 2011, 29, 051203 (11pp).

(57) Johnson, M.; Manno, M.; Zhang, X.; Leighton, C.; Aydil, E. S. Substrate and Temperature Dependence of the Formation of the Earth Abundant Solar Absorber Cu2ZnSnS4 by Ex Situ Sulfidation of Cosputtered Cu-Zn-Sn Films. J. Vac. Sci.

Technol. A Vacuum, Surfaces, Film. 2014, 32, 61203 (12pp).

(58) Johnson, M. C.; Wrasman, C.; Zhang, X.; Manno, M.; Leighton, C.; Aydil, E. S. Self-Regulation of Cu/Sn Ratio in the Synthesis of Cu2ZnSnS4 Films. Chem.

Mater. 2015, 27, 2507–2514.

(59) Kang, C.-C.; Chen, H.-F.; Yu, T.-C.; Lin, T.-C. Aqueous Synthesis of

133

Wurtzite Cu2ZnSnS4 Nanocrystals. Mater. Lett. 2013, 96, 24–26.

(60) Zhao, Z.; Ma, C.; Cao, Y.; Yi, J.; He, X.; Qiu, J. Electronic Structure and Optical Properties of Wurtzite-Kesterite Cu2ZnSnS4 Phys. Lett. A 2013, 377, 417–422.

(61) Khare, A.; Wills, A. W.; Ammerman, L. M.; Norris, D. J.; Aydil, E. S. Size Control and Quantum Confinement in Cu2ZnSnS4 Nanocrystals. Chem. Commun. 2011, 47, 11721–11723.

(62) Becker, W.; Lutz, H. D. Phase Studies in the Systems CoS-MnS, CoS-ZnS, and CoS-CdS. Mater. Res. Bull. 1978, 13, 907–911.

(63) Kevin, P.; Malik, M. A.; Mcadams, S.; O’Brien, P. Synthesis of Nanoparticulate Alloys of the Composition Cu2Zn1– XFeXSnS4 : Structural, Optical, and Magnetic Properties. J. Am. Chem. Soc. 2015, 137, 15086–15089.

(64) Bonazzi, P.; Bindi, L.; Bernardini, G. P.; Menchetti, S. A Model For the Mechanism of Incorporation of Cu, Fe and Zn in the Stannite - Kesterite Series, Cu2FeSnS4 - Cu2ZnSnS4. Can. Mineral. 2003, 41, 639–647.

(65) Chen, C.-J.; Chen, P.-T.; Basu, M.; Yang, K.-C.; Lu, Y.-R.; Dong, C.-L.; Ma, C.-G.; Shen, C.-C.; Hu, S.-F.; Liu, R.-S. An Integrated Cobalt Disulfide (CoS2) Co-Catalyst Passivation Layer on Silicon Microwires for Photoelectrochemical Hydrogen Evolution. J. Mater. Chem. A 2015, 3, 23466–23476.

(66) Lyapin, S. G.; Utyuzh, A. N.; Petrova, A. E.; Novikov, A. P.; Lograsso, T. A.; Stishov, S. M. Raman Studies of Nearly Half-Metallic Ferromagnetic CoS2. J.

Phys. Condens. Matter 2014, 26, 396001 (7pp).

134

References Chapter 4

(1) Johnson, M.; Baryshev, S. V.; Thimsen, E.; Manno, M.; Zhang, X.; Veryovkin, I. V.; Leighton, C.; Aydil, E. S. Alkali-Metal-Enhanced Grain Growth in Cu2ZnSnS4 Thin Films. Energy Environ. Sci. 2014, 7, 1931–1938.

(2) Polizzotti, A.; Repins, I. L.; Noufi, R.; Wei, S.-H.; Mitzi, D. B. The State and Future Prospects of Kesterite Photovoltaics. Energy Environ. Sci. 2013, 6, 3171-3182.

(3) Winkler, M. T.; Wang, W.; Gunawan, O.; Hovel, H. J.; Todorov, T. K.; Mitzi, D. B. Optical Designs That Improve the Efficiency of Cu2ZnSn(S,Se)4 Solar Cells. Energy Environ. Sci. 2014, 7, 1029-1036.

(4) Ghorpade, U.; Suryawanshi, M.; Shin, S. W.; Gurav, K.; Patil, P.; Pawar, S.; Hong, C. W.; Kim, J. H.; Kolekar, S. Towards Environmentally Benign Approaches for the Synthesis of CZTSSe Nanocrystals by a Hot Injection Method: A Status Review. Chem. Commun. 2014, 50, 20–23.

(5) Ghorpade, U. V; Suryawanshi, M. P.; Shin, W. Wurtzite CZTS Nanocrystals and Phase Evolution to Kesterite Thin Film for Solar Energy Harvesting. Phys.

Chem. Chem. Phys. 2015, 17, 19777–19788.

(6) Katagiri, H.; Jimbo, K.; Maw, W. S.; Oishi, K.; Yamazaki, M.; Araki, H.; Takeuchi, A. Development of CZTS-Based Thin Film Solar Cells. Thin Solid Films 2009, 517, 2455–2460.

(7) Mitzi, D. B.; Gunawan, O.; Todorov, T. K.; Wang, K.; Guha, S. The Path towards a High-Performance Solution-Processed Kesterite Solar Cell. Sol. Energy

Mater. Sol. Cells 2011, 95, 1421–1436.

(8) Shin, S. W.; Pawar, S. M.; Park, C. Y.; Yun, J. H.; Moon, J. H.; Kim, J. H.; Lee, J. Y. Studies on Cu2ZnSnS4 (CZTS) Absorber Layer Using Different Stacking Orders in Precursor Thin Films. Sol. Energy Mater. Sol. Cells 2011, 95, 3202–3206.

(9) Kim, J.; Hiroi, H.; Todorov, T. K.; Gunawan, O.; Kuwahara, M.; Gokmen, T.; Nair, D.; Hopstaken, M.; Shin, B.; Lee, Y. S.; et al. High Efficiency Cu2ZnSn(S,Se)4 Solar Cells by Applying a Double In2S3/CdS Emitter. Adv. Mater. 2014, 26, 7427–7431.

(10) Jackson, P.; Hariskos, D.; Wuerz, R.; Kiowski, O.; Bauer, A.; Friedlmeier, T. M.; Powalla, M. Properties of Cu(In,Ga)Se2 Solar Cells with New Record Efficiencies up to 21.7%. Phys. Status Solidi – Rapid Res. Lett. 2015, 9, 28–31.

(11) Buffière, M.; Brammertz, G.; Sahayaraj, S.; Batuk, M.; Khelifi, S.; Mangin, D.; El Mel, A. A.; Arzel, L.; Hadermann, J.; Meuris, M.; et al. KCN Chemical Etch

135

for Interface Engineering in Cu2ZnSnSe4 Solar Cells. ACS Appl. Mater. Interfaces 2015, 7, 14690–14698.

(12) Scragg, J. J.; Ericson, T.; Kubart, T.; Edoff, M.; Platzer-Björkman, C. Chemical Insights into the Instability of Cu2ZnSnS4Films during Annealing. Chem.

Mater. 2011, 23, 4625–4633.

(13) Scragg, J. J.; Watjen, J. T.; Edoff, M.; Ericson, T.; Kubart, T.; Platzer-Bjorkman, C. A Detrimental Reaction at the Molybdenum Back Contact in Cu2ZnSn(S,Se)4 Thin-Film Solar Cells. J. Am. Chem. Soc. 2012, 134, 19330–19333.

(14) Cui, H.; Liu, X.; Liu, F.; Hao, X.; Song, N.; Yan, C. Boosting Cu2ZnSnS4 Solar Cells Efficiency by a Thin Ag Intermediate Layer between Absorber and Back Contact. Appl. Phys. Lett. 2014, 104, 041115–041118.

(15) Liu, F.; Sun, K.; Li, W.; Yan, C.; Cui, H.; Jiang, L.; Hao, X.; Green, M. A. Enhancing the Cu2ZnSnS4 Solar Cell Efficiency by Back Contact Modification: Inserting a Thin TiB2 Intermediate Layer at Cu2ZnSnS4/Mo Interface. Appl. Phys.

Lett. 2014, 104, 051105–051109.

(16) Shin, B.; Zhu, Y.; Bojarczuk, N. A.; Jay Chey, S.; Guha, S. Control of an Interfacial MoSe2 Layer in Cu2ZnSnSe4 Thin Film Solar Cells: 8.9% Power Conversion Efficiency with a TiN Diffusion Barrier. Appl. Phys. Lett. 2012, 101, 053903–053906.

(17) Lopez-Marino, S.; Placidi, M.; Perez-Tomas, A.; Llobet, J.; Izquierdo-Roca, V.; Fontane, X.; Fairbrother, A.; Espindola-Rodriguez, M.; Sylla, D.; Perez-Rodriguez, A.; et al. Inhibiting the absorber/Mo-Back Contact Decomposition Reaction in Cu2ZnSnSe4 Solar Cells: The Role of a ZnO Intermediate Nanolayer. J. Mater. Chem. A 2013, 1, 8338–8343.

(18) Xie, H.; Sánchez, Y.; López-Marino, S.; Espíndola-Rodríguez, M.; Neuschitzer, M.; Sylla, D.; Fairbrother, A.; Izquierdo-Roca, V.; Pérez-Rodríguez, A.; Saucedo, E. Impact of Sn(S, Se) Secondary Phases in Cu2ZnSn(S, Se)4 Solar Cells: A Chemical Route for Their Selective Removal and Absorber Surface Passivation. ACS Appl. Mater. Interfaces 2014, 6, 12744–12751.

(19) Fairbrother, A.; García-Hemme, E.; Izquierdo-Roca, V.; Fontané, X.; Pulgarín-Agudelo, F. A.; Vigil-Galán, O.; Pérez-Rodríguez, A.; Saucedo, E. Development of a Selective Chemical Etch To Improve the Conversion Efficiency of Zn-Rich Cu2ZnSnS4 Solar Cells. J. Am. Chem. Soc. 2012, 134, 8018–8021.

(20) Furuta, K.; Sakai, N.; Kato, T.; Sugimoto, H.; Kurokawa, Y.; Yamada, A. Improvement of Cu2ZnSn(S,Se)4 Solar Cell Efficiency by Surface Treatment. Phys.

Status Solidi Curr. Top. Solid State Phys. 2015, 4, 1–4.

136

(21) Miyazaki, H.; Aono, M.; Kishimura, H.; Katagiri, H. Surface Etching of CZTS Absorber Layer by Br-Related Solution. Phys. Status Solidi 2015, 12, 741–744.

(22) Wang, Y.-C.; Cheng, H.-Y.; Yen, Y.-T.; Wu, T.-T.; Hsu, C.-H.; Tsai, H.-W.; Shen, C.-H.; Shieh, J.-M.; Chueh, Y.-L. Large-Scale Micro- and Nanopatterns of Cu(In,Ga)Se2 Thin Film Solar Cells by Mold-Assisted Chemical-Etching Process. ACS Nano 2015, 9, 3907–3916.

(23) Buffière, M.; Mel, A. A. El; Lenaers, N.; Brammertz, G.; Zaghi, A. E.; Meuris, M.; Poortmans, J. Surface Cleaning and Passivation Using (NH4)2S Treatment for Cu(In,Ga)Se2 Solar Cells: A Safe Alternative to KCN. Adv. Energy Mater. 2015, 5, 1–7.

(24) John Emsley. The Elements of Murder: A History of Poison; Oxford University Press; New Ed edition, 2006.

(25) International Cyanide Management Institute. International Cyanide Management Code - Environmental & Health Effects http://www.cyanidecode.org/cyanide-facts/environmental-health-effects.

(26) Webber, D. H.; Buckley, J. J.; Antunez, P. D.; Brutchey, R. L. Facile Dissolution of Selenium and Tellurium in a Thiol–amine Solvent Mixture under Ambient Conditions. Chem. Sci. 2014, 5, 2498-2502.

(27) Webber, D. H.; Brutchey, R. L. Alkahest for V2VI3 Chalcogenides: Dissolution of Nine Bulk Semiconductors in a Diamine-Dithiol Solvent Mixture. J.

Am. Chem. Soc. 2013, 135, 15722–15725.

(28) Williams, B. A.; Mahajan, A.; Smeaton, M. A.; Holgate, C. S.; Aydil, E. S.; Francis, L. F. Formation of Copper Zinc Tin Sulfide Thin Films from Colloidal Nanocrystal Dispersions via Aerosol-Jet Printing and Compaction. ACS Appl.

Mater. Interfaces 2015, 7, 11526–11535.

(29) Chernomordik, B. D.; Ketkar, P. M.; Hunter, A. K.; Béland, A. E.; Deng, D. D.; Aydil, E. S. Microstructure Evolution during Selenization of Cu2ZnSnS4 Colloidal Nanocrystal Coatings. Chem. Mater. 2016, 28, 1266–1276.

(30) Chernomordik, B. D.; Béland, A. E.; Deng, D. D.; Francis, L. F.; Aydil, E. S. Microstructure Evolution and Crystal Growth in Cu2ZnSnS4 Thin Films Formed by Annealing Colloidal Nanocrystal Coatings. Chem. Mater. 2014, 26, 3191–3201.

(31) Kar, P.; Farsinezhad, S.; Zhang, X.; Shankar, K. Anodic Cu2S and CuS Nanorod and Nanowall Arrays: Preparation, Properties and Application in CO2 Photoreduction. Nanoscale 2014, 6, 14305–14318.

137

(32) Cheng, A.-J.; Manno, M.; Khare, A.; Leighton, C.; Campbell, S. A.; Aydil, E. S. Imaging and Phase Identification of Cu2ZnSnS4 Thin Films Using Confocal Raman Spectroscopy. J. Vac. Sci. Technol. 2011, 29, 51203-51213.

(33) Minceva-Sukarova, B.; Najdoski, M.; Grozdanov, I.; Chunnilall, C. J. Raman Spectra of Thin Solid Films of Some Metal Sulfides. J. Mol. Struct. 1997, 410–411, 267–270.

(34) Shin, S. W.; Kim, I. Y.; Gurav, K. V.; Jeong, C. H.; Yun, J. H.; Patil, P. S.; Lee, J. Y.; Kim, J. H. Band Gap Tunable and Improved Microstructure Characteristics of Cu2ZnSn(S1−x,Sex)4 Thin Films by Annealing under Atmosphere Containing S and Se. Curr. Appl. Phys. 2013, 13, 1837–1843.

(35) Shin, S. W.; Han, J. H.; Park, Y. C.; Agawane, G. L.; Jeong, C. H.; Yun, J. H.; Moholkar, A. V.; Lee, J. Y.; Kim, J. H. A Facile and Low-Cost Synthesis of Promising Absorber Materials on Cu2ZnSn(Sx,Se1−x)4 Nanocrystals Consisting of Earth Abundant Elements with Tunable Band Gap Characteristics. J. Mater. Chem. 2012, 22, 21727-21732.

(36) Gu, Y.; Su, Y.; Chen, D.; Geng, H.; Li, Z.; Zhang, L.; Zhang, Y. Hydrothermal Synthesis of Hexagonal CuSe Nanoflakes with Excellent Sunlight-Driven Photocatalytic Activity. CrystEngComm 2014, 16, 9185–9190.

138

References Chapter 5

(1) Zhao, X.; Lu, M.; Koeper, M. J.; Agrawal, R. Solution-Processed Sulfur Depleted Cu(In,Ga)Se2 Solar Cells Synthesized from a Monoamine–dithiol Solvent Mixture. J. Mater. Chem. A 2016, 4, 7390–7397.

(2) Suryawanshi, M.; Shin, S. W.; Ghorpade, U.; Song, D.; Hong, C. W.; Han, S.-S.; Heo, J.; Kang, S. H.; Kim, J. H. A Facile and Green Synthesis of Colloidal Cu2ZnSnS4 Nanocrystals and Their Application in Highly Efficient Solar Water Splitting. J. Mater. Chem. A 2017, 5, 4695–4709.

(3) Zhang, R.; Szczepaniak, S. M.; Carter, N. J.; Handwerker, C. a.; Agrawal, R. A Versatile Solution Route to Efficient Cu2ZnSn(S,Se)4 Thin-Film Solar Cells. Chem. Mater. 2015, 27, 2114–2120.

(4) Tian, Q.; Wang, G.; Zhao, W.; Chen, Y.; Yang, Y.; Huang, L.; Pan, D. Versatile and Low-Toxic Solution Approach to Binary, Ternary, and Quaternary Metal Sulfide Thin Films and Its Application in Cu2ZnSn(S,Se)4 Solar Cells. Chem.

Mater. 2014, 26, 3098–3103.

(5) Zhang, R.; Cho, S.; Lim, D. G.; Hu, X.; Stach, E. A.; Handwerker, C. A.; Agrawal, R. Metal–metal Chalcogenide Molecular Precursors to Binary, Ternary, and Quaternary Metal Chalcogenide Thin Films for Electronic Devices. Chem.

Commun. 2016, 846, 31–39.

(6) Arnou, P.; Van Hest, M. F. A. M.; Cooper, C. S.; Malkov, A. V.; Walls, J. M.; Bowers, J. W. Hydrazine-Free Solution-Deposited CuIn(S,Se)2 Solar Cells by Spray Deposition of Metal Chalcogenides. ACS Appl. Mater. Interfaces 2016, 8, 11893–11897.

(7) Ghorpade, U. V; Suryawanshi, M. P.; Shin, S. W.; Hong, C. W.; Kim, I.; Moon, J. H.; Yun, J. H.; Kim, J. H.; Kolekar, S. S. Wurtzite CZTS Nanocrystals and Phase Evolution to Kesterite Thin Film for Solar Energy Harvesting. Phys. Chem.

Chem. Phys. 2015, 17, 19777–19788.

(8) Ghorpade, U.; Suryawanshi, M.; Shin, S. W.; Gurav, K.; Patil, P.; Pawar, S.; Hong, C. W.; Kim, J. H.; Kolekar, S. Towards Environmentally Benign Approaches for the Synthesis of CZTSSe Nanocrystals by a Hot Injection Method: A Status Review. Chem. Commun. 2014, 50, 20–23.

(9) Polizzotti, A.; Repins, I. L.; Noufi, R.; Wei, S.-H.; Mitzi, D. B. The State and Future Prospects of Kesterite Photovoltaics. Energy Environ. Sci. 2013, 6, 3171-3182.

(10) Gang, M. G.; Shin, S. W.; Hong, C. W.; Gurav, K. V; Gwak, J.; Yun, J.; Lee, J. Y.; Kim, J. H. Sputtering Processed Highly Efficient Cu2ZnSn(S,Se)4 Solar Cells by a Low-Cost, Simple, Green, and up-Scalable Strategy. Green Chem. 2015, 18,

139

700–711.

(11) Pinto, A. H.; Shin, S. W.; Aydil, E. S.; Penn, R. L. Selective Removal of Cu2−x(S,Se) Phases from Cu2ZnSn(S,Se)4 Thin Films. Green Chem. 2016, 18, 5814–5821.

(12) Chernomordik, B. D.; Ketkar, P. M.; Hunter, A. K.; Béland, A. E.; Deng, D. D.; Aydil, E. S. Microstructure Evolution During Selenization of Cu2ZnSnS4 Colloidal Nanocrystal Coatings. Chem. Mater. 2016, 28, 1266–1276.

(13) Chernomordik, B. D.; Béland, A. E.; Deng, D. D.; Francis, L. F.; Aydil, E. S. Microstructure Evolution and Crystal Growth in Cu2ZnSnS4 Thin Films Formed by Annealing Colloidal Nanocrystal Coatings. Chem. Mater. 2014, 26, 3191–3201.

(14) Bag, S.; Gunawan, O.; Gokmen, T.; Zhu, Y.; Todorov, T. K.; Mitzi, D. B. Low Band Gap Liquid-Processed CZTSe Solar Cell with 10.1 % Efficiency. Energy

Environ. Sci. View 2012, 5, 7060–7065.

(15) Bag, S.; Gunawan, O.; Gokmen, T.; Zhu, Y.; Mitzi, D. B. Hydrazine-Processed Ge-Substituted CZTSe Solar Cells. Chem. Mater. 2012, 24, 4588–4593.

(16) Williams, B. A.; Smeaton, M. A.; Trejo, N. D.; Francis, L. F.; Aydil, E. S. Effect of Nanocrystal Size and Carbon on Grain Growth during Annealing of Copper Zinc Tin Sulfide Nanocrystal Coatings. Chem. Mater. 2017, 29, 1676–1683.

(17) Mitzi, D. B.; Gunawan, O.; Todorov, T. K.; Wang, K.; Guha, S. The Path towards a High-Performance Solution-Processed Kesterite Solar Cell. Sol. Energy

Mater. Sol. Cells 2011, 95, 1421–1436.

(18) López-Marino, S.; Sánchez, Y.; Placidi, M.; Fairbrother, A.; Espindola-Rodríguez, M.; Fontané, X.; Izquierdo-Roca, V.; Lõpez-García, J.; Calvo-Barrio, L.; Pérez-Rodríguez, A.; Saucedo, E. ZnSe Etching of Zn-Rich Cu2ZnSnSe4: An Oxidation Route for Improved Solar-Cell Efficiency. Chem. - A Eur. J. 2013, 19, 14814–14822.

(19) Miyazaki, H.; Aono, M.; Kishimura, H.; Katagiri, H. Surface Etching of CZTS Absorber Layer by Br-Related Solution. Phys. Status Solidi 2015, 12, 741–744.

(20) Buffière, M.; Brammertz, G.; Sahayaraj, S.; Batuk, M.; Khelifi, S.; Mangin, D.; El Mel, A. A.; Arzel, L.; Hadermann, J.; Meuris, M.; et al. KCN Chemical Etch for Interface Engineering in Cu2ZnSnSe4 Solar Cells. ACS Appl. Mater. Interfaces 2015, 7, 14690–14698.

(21) Erkan, M. E.; Chawla, V.; Repins, I.; Scarpulla, M. A. Interplay between Surface Preparation and Device Performance in CZTSSe Solar Cells: Effects of

140

KCN and NH4OH Etching. Sol. Energy Mater. Sol. Cells 2015, 136, 78–85.

(22) Xie, H.; Sánchez, Y.; López-Marino, S.; Espíndola-Rodríguez, M.; Neuschitzer, M.; Sylla, D.; Fairbrother, A.; Izquierdo-Roca, V.; Pérez-Rodríguez, A.; Saucedo, E. Impact of Sn(S, Se) Secondary Phases in Cu2ZnSn(S,Se)4 Solar Cells: A Chemical Route for Their Selective Removal and Absorber Surface Passivation. ACS Appl. Mater. Interfaces 2014, 6, 12744–12751.

(23) Mousel, M.; Redinger, A.; Djemour, R.; Arasimowicz, M.; Valle, N.; Dale, P.; Siebentritt, S. HCl and Br2-MeOH Etching of Cu2ZnSnSe4 Polycrystalline Absorbers. Thin Solid Films 2013, 535, 83–87.

(24) Cummings, T. F. The Treatment of Cyanide Poisoning. Occup. Med. (Chic.

Ill). 2004, 54, 82–85.

(25) Liu, W.; Mitzi, D. B.; Yuan, M.; Kellock, A. J.; Jay Chey, S.; Gunawan, O. 12% Efficiency CuIn(Se,S)2 Photovoltaic Device Prepared Using a Hydrazine Solution Process. Chem. Mater. 2010, 22, 1010–1014.

(26) Webber, D. H.; Brutchey, R. L. Alkahest for V2VI3 Chalcogenides: Dissolution of Nine Bulk Semiconductors in a Diamine-Dithiol Solvent Mixture. J.

Am. Chem. Soc. 2013, 135, 15722–15725.

(27) Webber, D. H.; Buckley, J. J.; Antunez, P. D.; Brutchey, R. L. Facile Dissolution of Selenium and Tellurium in a Thiol–amine Solvent Mixture under Ambient Conditions. Chem. Sci. 2014, 5, 2498-2502.

(28) Buckley, J. J.; McCarthy, C. L.; Del Pilar-Albaladejo, J.; Rasul, G.; Brutchey, R. L. Dissolution of Sn, SnO, and SnS in a Thiol-Amine Solvent Mixture: Insights into the Identity of the Molecular Solutes for Solution-Processed SnS. Inorg. Chem. 2016, 55, 3175–3180.

(29) Liu, Y.; Yao, D.; Shen, L.; Zhang, H.; Zhang, X.; Yang, B. Alkylthiol-Enabled Se Powder Dissolution in Oleylamine at Room Temperature for the Phosphine-Free Synthesis of Copper-Based Quaternary Selenide Nanocrystals. J. Am.

Chem. Soc. 2012, 134, 7207–7210.

(30) Yang, W.; Duan, H. S.; Cha, K. C.; Hsu, C. J.; Hsu, W. C.; Zhou, H.; Bob, B.; Yang, Y. Molecular Solution Approach to Synthesize Electronic Quality Cu2ZnSnS4 Thin Films. J. Am. Chem. Soc. 2013, 135, 6915–6920.

(31) Williams, B. A.; Mahajan, A.; Smeaton, M. A.; Holgate, C. S.; Aydil, E. S.; Francis, L. F. Formation of Copper Zinc Tin Sulfide Thin Films from Colloidal Nanocrystal Dispersions via Aerosol-Jet Printing and Compaction. ACS Appl.

Mater. Interfaces 2015, 7, 11526–11535.

(32) Socrates, G. Infrared and Raman Characteristic Group Frequencies; 3rd

141

ed.; John Wiley & Sons, Inc.: West Sussex - England, 2001.

(33) Barick, K. C.; Aslam, M.; Prasad, P. V.; Dravid, V. P.; Bahadur, D. Nanoscale Assembly of Amine-Functionalized Colloidal Iron Oxide. J. Magn.

Magn. Mater. 2009, 321, 1529–1532.

(34) Zhang, X.; Wang, S. Voltametric Behavior of Noradrenaline at 2-Mercaptoethanol Self-Assembled Monolayer Modified Gold Electrode and Its Analytical Application. Sensors 2003, 3, 61–68.

(35) Huang, C.; Gou, S.; Zhu, H.; Huang, W. Cleavage of C−S Bonds with the Formation of a Tetranuclear Cu(I) Cluster. Inorg. Chem. 2007, 46, 5537–5543.

(36) Cornell, N. W.; Crivaro, K. E. Stability Constant for the Zinc-Dithiothreitol Complex. Anal. Biochem. 1972, 47, 203–208.

(37) Vasák, M.; Kägi, J. H. R.; O. Hill, H. A. Zinc(II), cadmium(II), and mercury(II) Thiolate Transitions in Metallothionein. Biochemistry 1981, 20, 2852–2856.

(38) Holmes, R. R.; Shafieezad, S.; Chandrasekhar, V.; Holmes, J. M.; Day, R. O. Pentacoordinated Molecules. 71. Hydrolysis Reactions Leading to Ring-Containing Hexacoordinated Distannoxanes. Tin-Sulfur vs Tin-Oxygen Bonding. J. Am. Chem. Soc. 1988, 110, 1174–1180.

(39) Li, Y.; Ding, Y.; Liao, H.; Qian, Y. Room-Temperature Conversion Route to Nanocrystalline Mercury Chalcogenides HgE (E=S,Se,Te). J. Phys. Chem. Solids 1999, 60, 965–968.

(40) Toh, M. L.; Tan, K. J.; Wei, F. X.; Zhang, K. K.; Jiang, H.; Kloc, C. Intercalation of Organic Molecules into SnS2 Single Crystals. J. Solid State Chem. 2013, 198, 224–230.

142

APPENDIX A:

SUPPORTING INFORMATION FOR

CHAPTER 2

2. Cu2ZnSnS4 (CZTS) phase control by changing the initial oxidation states

of the cations and sulfur excess source in microwave solvothermal

synthesis

.

A1 Rietveld Refinement Procedure

All the Rietveld refinement was performed using the software XPert

HighScore Plus. The wurtzite structure (space group P63mc, #186) was simulated

using a = 0.38353 nm, b = 0.38353 nm, c = 0.63008 nm, α = 90°, β = 90°, γ = 120°.

Wyckoff positions of the Cu, Zn, Sn, and S used in the simulations are given in

Table A1.

The kesterite structure was simulated according to the cif file Cu2ZnSnS4

ICSD 262388 from the Inorganic Crystal Structure Database (ICSD). Starting with

Table A1. Wyckoff position of the Cu, Zn, Sn, and S used to simulate wurtzite CZTS, space

group P63mc (#186).

Element Wyckoff

Position

X y z sof Biso

Cu 2b 1/3 2/3 0 0.333 0.500

Zn 2b 1/3 2/3 0 0.333 0.500

Sn 2b 1/3 2/3 0 0.333 0.500

S 2b 1/3 2/3 0.375 1 0.500

143

this information the kesterite and wurtzite crystalline structures were refined using,

the following steps in this order:

1) Scale Factors for kesterite and wurtzite structures were refined at the same

time.

2) The 4 first background parameters were refined.

3) Zero Shift was refined.

4) The a and b lattice parameters for one phase were refined, followed by the

refinement of a and b lattice parameters for the other phase.

5) The c lattice parameter for one phase was refined, followed by the c lattice

parameter for the other phase.

6) After refining all lattice parameters (a, b, c) for both phases, the lattice

parameters (a, b, c) for both phases were kept constant, while profile

parameter U was refined first for one of the phases followed by the other

phase.

7) The parameter peak shape 1 for one phase was refined followed by the

other phase. The profile parameter U and peak shape 1 are the parameters

related to the peak width.

The fraction (percentage) of wurtzite, and kesterite phases and the Goodness

of Fit (GoF) for each refinement are shown in Table A2. The values presented in

Table A2 were used to produce Figure 2.1b of Chapter 2. Three results for each

nanocrystal product were calculated and averaged and their standard deviation

calculated, producing the error bars shown in Figure 2.1 b.

144

A2 Fraction of wurtzite and kesterite phases in the product synthesized at 160 °C using various S:M ratios

Table A2. Fractions (in %) of wurtzite and kesterite phases and Goodness of Fit (GoF) for

triplicate trials of the product synthesized at 160 °C using various S:M ratios

Cu(II)_Sn(II)_Tu_S:M_160 °C.

S:M Wurtzite (%) Kesterite (%) GoF

64.4 35.6 2.82

1.9 44.4 55.6 2.27

74.2 25.8 2.09

52.8 47.2 4.29

2.7 34.9 65.1 2.60

63.8 36.2 2.81

49.5 50.5 4.67

3.2 62.7 37.3 2.87

33.7 66.3 2.48

47.0 53.0 4.77

3.6 62.3 37.7 3.02

35.5 64.5 2.02

11.4 88.6 2.49

4.5 35.5 64.5 3.60

25.2 74.8 1.86

13.8 86.2 4.23

5.3 13.4 86.6 4.21

39.5 60.5 6.93

24.5 75.5 1.74

6.2 13.5 86.5 3.41

28.7 71.3 2.51

145

A3 Crystallite size estimate using the Scherrer equation for

Cu(II)_Sn(II)_Tu_S:M_160 °C

A4 STEM-HAADF Elemental Mapping of the Cu(II)_Sn(II)_Tu_1.9_160 °C and

Cu(II)_Sn(II)_Tu_6.2_160 °C Samples

Table A3. Crystallite size estimates for Cu(II)_Sn(II)_Tu_S:M_160 °C. Sizes were calculated

using the Scherrer equation, a shape factor of 0.9 and the peak at 2θ≈32°, (112) of the wurtzite

and the (002) of the kesterite phases. Deviation from a spherical shape and differences in the

shapes of kesterite and wurtzite nanocrystals in multiphase mixtures may change these

estimates slightly.

S:M Size (nm)

1.9 6.4

2.7 5.0

3.2 4.1

3.6 4.4

4.5 4.3

5.3 3.7

6.2 3.7

Figure A1. (a) TEM image (same as Figure 2.3a) of Cu(II)_Sn(II)_Tu_1.9_160 °C

nanocrystals with boxes around the oblate nanocrystals. (b) Higher magnification TEM images

from another region of the same sample, where the oblate nanocrystals can be seen more

clearly.

146

A5 Elemental Composition of Cu(N)_Sn(L)_Tu_S:M_160 °C

Figure A2. STEM-HAADF elemental map for a) Cu(II)_Sn(II)_Tu_1.9_160 °C and b)

Cu(II)_Sn(II)_Tu_6.2_160 °C

Table A4. Elemental composition determined from SEM-EDS for nanocrystal products

synthesized using Cu and Sn precursors with different oxidation states and S:M ratios, i.e.,

Cu(N)_Sn(L)_Tu_S:M, for N= I or II, L= II or IV, and S:M = 1.9 or 6.2. All nanocrystals were

synthesized at 160 °C. Compositions are given with respect to S, where the sulfur EDS intensity

was normalized to 4.

Sample Cu Zn Sn S

Cu(I)_Sn(II)_1.9_Tu_160 °C 2.7 1.4 1.0 4.0

Cu(I)_Sn(II)_6.2_Tu_160 °C 1.6 0.8 0.9 4.0

Cu(II)_Sn(II)_1.9_Tu_160 °C 2.6 1.2 1.0 4.0

Cu(II)_Sn(II)_6.2_Tu_160 °C 2.3 1.1 1.1 4.0

Cu(I)_Sn(IV)_1.9_Tu_160 °C 3.2 0.4 1.0 4.0

Cu(I)_Sn(IV)_6.2_Tu_160 °C 2.0 1.1 1.1 4.0

Cu(II)_Sn(IV)_1.9_Tu_160 °C 2.8 1.1 0.9 4.0

Cu(II)_Sn(IV)_6.2_Tu_160 °C 2.4 1.2 1.1 4.0

147

A6 Microwave Synthesis of CZTS using other excess sulfur sources

To verify if the presence of thiourea is necessary to form phase pure CZTS,

we performed experiments without using thiourea. Instead, we used either L-

cysteine or thioglycolic acid as the only source of sulfur, varying the tin oxidation

state and fixing the S:M ratio equal to 1.9 with the final temperature 160 °C. The

XRD patterns of these samples are presented on Figure A4. Figure A4 shows that

CZTS could not be detected in any of the products, regardless of the tin oxidation

state or the sulfur source. These results reveal that thiourea is necessary.

A7 Infrared Spectroscopy of the Samples Cu(II)_Sn(II)_Tu_S:M_160 °C

Since the wurtzite to kesterite product ratio depends on S:M ratio, we

considered the possibility that the S:M ratio might affect the surface chemistry of

the product and hence also the product phases. The adsorption of surfactant

molecules could serve to alter the relative surface energy of different facets, which

could result in the formation of a metastable phase. Accordingly, we collected and

Figure A3. XRD patterns from products synthesized at 160 °C without thiourea, varying the

tin oxidation state, and using either L-cysteine or thioglycolic acid as the only sulfur source.

148

examined infrared (IR) spectra from nanocrystals synthesized using different S:M

ratios to determine if these nanocrystals exhibited differences in ligands that cap

their surfaces. These IR spectra showed that the surface of the nanocrystals were

capped with ethylene glycol regardless of the S:M ratio used in the synthesis. For

example, Figure A3 shows the IR spectra of nanocrystals synthesized using two

different S:M ratios (Cu(II)_Sn(II)_Tu_S:M_160 °C, where S:M=1.9 or 6.2) before

they were washed with ethanol and dried. Infrared spectra of ethylene glycol and

thiourea are also shown for comparison. The IR spectra of nanocrystals

Cu(II)_Sn(II)_Tu_S:M_160 °C with S:M equal to 1.9 and 6.2 show bands at 882,

1033, 1086, and 3329 cm-1. The first band is assigned to stretching of C-C-O

groups, the next two to stretching of C-O groups, and the last one to the stretching

of O-H groups in ethylene glycol, respectively. The sharp and intense band at 1601

cm-1 present in thiourea spectra, which is characteristic of in-plane deformation of

N-H group, is absent in both Cu(II)_Sn(II)_Tu_S:M_160 °C spectra. These

149

observations are consistent with the presence only of ethylene glycol on the

nanoparticle surface.

Since there is no obvious difference in the surface ligands when S:M=1.9,

and S:M=6.2 we also rule out the possibility that the wurtzite stabilization is due to

changes in the surface energies of different crystal facets under these two

synthesis conditions. This leaves the possibility that the preferential synthesis of

one phase over another may be related to the differences in their formation

mechanism at different S:M ratios.

A8 Effect of S:M ratio and Sn oxidation state, and excess sulfur source on

phase composition

Figure 2.5 of the main text shows the XRD patterns of the CZTS samples

prepared from Cu(II), different S:M ratios, different tin oxidation states (Sn(II) or

Sn(IV)), and different excess sulfur excess sources, such as thiourea, L-cysteine,

thioglycolic acid and 3-mercaptopropionic acid. In order to complement the data

presented in Figure 2.5, Table A5 presents the quantitative phase percentage

Figure A4. FTIR transmission spectra of the Cu(II)_Sn(II)_Tu_1.9_160 °C (red), and

Cu(II)_Sn(II)_Tu_6.2_160 °C (blue) nanocrystals. Reference spectra for ethylene glycol

(Fisher Scientific), and thiourea are also shown for comparison.

150

estimated by Rietveld Refinement. Table A5 confirms the trend that when the sulfur

excess is provided by a molecule without an NH2 group wurtzite is formed as the

major phase, regardless of the S:M ratio and tin oxidation state. On the other hand,

when the sulfur excess source contains an NH2 group, mostly wurtzite is formed

when tin has the oxidation state +2 and mostly kesterite formed when tin has the

oxidation state +4, regardless of the S:M ratio.

Table A5. Fraction (in %) of the phases estimated from Rietveld refinement for the samples

Cu(II)_Sn(L)_XCS_S:M_160 °C, where L = II or IV, XCS = Tu, Cyst, TGacid, or MCPacid,

and S:M = 1.9 or 3.6.

Sample Wurtzite Kesterite

Cu(II)_Sn(II)_Tu_1.9_160 °C 61.0 ± 12.4 39.0 ± 12.4

Cu(II)_Sn(II)_Tu_3.6_160 °C 48.3 ± 11.0 51.7 ± 11.0

Cu(II)_Sn(IV)_Tu_1.9_160 °C 33.7 ± 5.6 66.3 ± 5.6

Cu(II)_Sn(IV)_Tu_3.6_160 °C 33.7 ± 7.3 66.3 ± 7.3

Cu(II)_Sn(II)_Cyst_1.9_160 °C 78.7 ± 7.3 21.3 ± 7.3

Cu(II)_Sn(II)_Cyst_3.6_160 °C 61.8 ± 4.9 38.2 ± 4.9

Cu(II)_Sn(IV)_Cyst_1.9_160 °C 31.0 ± 5.0 69.0 ± 5.0

Cu(II)_Sn(IV)_Cyst_3.6_160 °C 52.6 ± 4.5 47.4 ± 4.5

Cu(II)_Sn(II)_TGacid_1.9_160 °C 70.8 ± 6.3 29.2 ± 6.3

Cu(II)_Sn(II)_TGacid_3.6_160 °C 81.4 ± 3.8 18.6 ± 3.8

Cu(II)_Sn(IV)_TGacid_1.9_160 °C 62.3 ± 4.4 37.7 ± 4.4

Cu(II)_Sn(IV)_TGacid_3.6_160 °C 72.6 ± 3.7 27.4 ± 3.7

Cu(II)_Sn(II)_MCPacid_1.9_160 °C 64.6 ± 9.0 35.4 ± 9.0

Cu(II)_Sn(II)_MCPacid_3.6_160 °C 75.8 ± 8.6 24.2 ± 8.6

Cu(II)_Sn(IV)_MCPacid_1.9_160 °C 81.6 ± 4.9 18.4 ± 4.9

Cu(II)_Sn(IV)_MCPacid_3.6_160 °C 66.9 ± 3.2 33.1 ± 3.2

151

A9 Synthesis of Zn-Sn intermediates at room temperature

A white powder formed when the synthesis was conducted at 25 °C with

Cu(II) and Sn(II) precursors and thiourea (e.g. products Cu(II)_Sn(II)_Tu_S:M_25

°C). Formation of this powder persists at low temperature but is converted to CZTS

at higher temperatures (see main text). This white powder contains Zn and Sn,

and we refer to it as the Zn-Sn intermediate or Zn-Sn glycolate intermediate in the

text. The XRD pattern from this product is shown in the main text. An extensive

search on Cambridge Structural Database (CSD) was carried out, looking for Zn

or Sn coordination compounds containing thiourea and glycolate groups. Although

the search did not reveal any compound that matches completely with the XRD

pattern obtained for from our white powder, it resembled XRD patterns that could

originate from a mixture of the compounds thioureatin(II) chloride

Figure A5. a) XRD pattern from Cu(II)_Sn(II)_Tu_1.9_25 °C product and XRD patterns for

coordination compounds of Zn and Sn that matched the best. b) Molecular Structure of the

compounds possibly present in the Zn-Sn intermediate.

152

(Sn(NH2CSNH2)Cl2) (CSD code: CAPWEV) and catena-((m2-Succinato-O,O')-

bis(thiourea-S)-zinc) (CSD code: FELXEA). Figure A5 shows the XRD pattern from

these compounds and the white powder product of Cu(II)_Sn(II)_Tu_1.9_25 °C.

The similarity of the XRD pattern of this zinc succinate thiourea complex and the

Zn-Sn intermediate prompted us to hypothesize that the intermediate is a mixture

of the complex thioureatin(II) chloride and complex containing glycolate and

thiourea ligands, where the glycolate ligands would be able help form polymeric

units. Moreover, the composition of this white powder product (i.e.

Cu(II)_Sn(II)_S:M_25 °C) always had 1:1 Zn:Sn ratio and did not contain any

detectable copper (Figure A6, see the data point at 25 °C).

In order to see if the other sulfur excess sources, besides thiourea, could

produce the same Zn-Sn intermediate, the room temperature synthesis was

carried out either with Sn(II) or Sn(IV) and with 3.4 x 10-3 mol of cysteine,

Figure A6. Elemental composition determined using SEM-EDS for a) Cu(II)_Sn(II)_Tu_1.9_T

and b) Cu(II)_Sn(II)_Tu_6.2_T as a function of temperature, T, from 25 to 160 °C.

153

thioglycolic acid, or mercaptopropionic acid added as the excess sulfur source, in

addition to the stoichiometric amount of thiourea (4.0 x 10-3 mol), to reach S:M=1.9.

Figure A7 shows the XRD pattern from products synthesized using different sulfur

excess sources and Sn initial oxidation state at 25 oC. When the sulfur excess is

provided by L-cysteine and Sn initial oxidation state is +2 (e.g.,

Cu(II)_Sn(II)_cyst_1.9_25 °C), the same Zn-Sn intermediate forms. The XRD

pattern is the same as that obtained from the product formed using thiourea (e.g.,

Cu(II)_Sn(II)_Tu_1.9_25 °C). See Figure 2.6 of the main text for the latter XRD

pattern, also reproduced in Figure A7 for easy comparison. In contrast, when the

sulfur source does not contain NH2 groups, (e.g., 3-mercaptopropionic acid and

thioglycolic acid), this Zn-Sn intermediate does not form.

154

Figure A8 shows an SEM image of the Zn-Sn intermediate formed by the

synthesis where L-cysteine was used as the excess sulfur source (e.g.,

Cu(II)_Sn(II)_cyst_1.9_25 °C). The crystals of this product have the same

anisotropic hexagonal rod-like morphology as the Zn-Sn intermediate formed by

the synthesis where thiourea was the excess sulfur source (e.g.,

Cu(II)_Sn(II)_Tu_1.9_25 °C, figure 2.7a in the main text).

Figure A7. XRD patterns from products Cu(II)_Sn(L)_XCS_1.9_25 °C, where L = II or IV,

and XCS = thiourea (Tu), L-cysteine (cyst), 3-mercaptopropionic acid (MCPacid), thioglycolic

acid (TGacid).

155

A10 Experiments varying Sn oxidation state at room temperature

To explore the identity of the Zn-Sn precursor that leads to wurtzite CZTS,

the Cu oxidation state was fixed at (II) while the Sn oxidation state was varied

between (II) and (IV) for S:M=1.9 or 6.2. Microwave heating was not used in these

experiments. After sonication at room temperature, the solid products were

collected by centrifugation and characterized. When using the Sn(IV) reagent, a

yellow powder was obtained both for S:M=1.9 and S:M=6.2. The XRD pattern for

both samples matched the XRD pattern for elemental sulfur (Figure A9). When the

solid product of the synthesis with Cu(II) and Sn(IV) reagents with S:M=1.9 at 25 °C

(i.e., Cu(II)_Sn(IV)_Tu_1.9_25 °C) was analyzed, it was found that it consisted only

of sulfur (Table A6). The production of elemental sulfur can be explained by the

decomposition of thiourea in acidic media, according to the following reaction:

SC(NH2)2(l) + H+ → NH3(g) + H+ + HCN(g) + 1/8 S8(s)

Figure A8. SEM image of the Zn-Sn intermediate formed by the synthesis where L-cysteine

was used as the excess sulfur source (e.g., Cu(II)_Sn(II)_cyst_1.9_25 °C).

156

A11 Effect of temperature on the morphology of products Cu(II)_Sn(II)_Tu_1.9_T and Cu(II)_Sn(II)_Tu_6.2_T

Figure A10 shows the SEM images of the products whose XRD patterns

are shown in Figure 2.6 of the main text. Figure A10a shows that synthesis with

S:M=1.9 at 75 °C produces only hexagonal prisms. At higher temperatures, 100

and 130 °C, these hexagonal prisms are still present but are now mixed with

smaller spherical particles. Synthesis with S:M=6.2 at 75 °C produces small

Table A6. Composition of the product synthesized with Cu(II) and Sn(IV) reagents, thiourea,

and S:M=1.9 at 25 °C (e.g., Cu(II)_Sn(IV)_Tu_1.9_25 °C). Elemental values are in atom %.

Element Spectra

1

Spectra

2

Spectra

3

Spectra

4

Spectra

5

Spectra

6

Spectra

7

Spectra

8

Spectra

9

Spectra

10

Cu 0.44 0 0 0.83 0.36 0 0 0.99 0 0.85

Zn 0 0 0 0 0 0 0 1.31 0 0

Sn 0 0.67 0 0 0.59 0 0 1.61 0 0

S 99.56 99.33 100 99.17 99.05 100 100 96.08 100 99.15

Figure A9. XRD pattern from products synthesized with Cu(II) and Sn(IV) reagents at 25 °C

with S:M=1.9 and S:M=6.2 (i.e., Cu(II)_Sn(IV)_Tu_1.9_25 °C and Cu(II)_Sn(IV)_Tu_6.2_25

°C). This data shows that only elemental sulfur forms at room temperature when Sn(IV) is used

in the synthesis.

157

spherical particles mixed with the hexagonal plates. Synthesis with S:M=6.2 at 100

and 130 °C produces primarily spherical particles. These results are in agreement

with the XRD results presented in Figure 2.6 of the main text, which shows that the

Zn-Sn intermediates persist until 130 °C for S:M=1.9, whereas, for S:M=6.2, the

Zn-Sn glycolates (associated with the hexagonal morphology) are present only

until 75 °C, and the phase pure CZTS kesterite is obtained starting from 100 °C. It

appears that higher sulfur concentration transforms the glycolates into CZTS at

lower temperature and hexagonal glycolates persist to higher temperatures only

when S:M is low.

The elemental analysis by SEM-EDS of the white powder, the Zn-Sn

precursor synthesized with S:M=1.9 and S:M=6.2 at room temperature, revealed

that it does not contain detectable Cu, and the Zn to Sn ratio is 1 (see Figure A6).

This confirms that the white precipitate at room temperature is the Zn-Sn

intermediate. The powder synthesized with S:M=6.2 had a larger amount of sulfur

than the powder synthesized with S:M=1.9 (see Fig. S5). By 75 °C, fractions of Zn

and Sn decrease in both products synthesized with S:M=1.9 and S:M=6.2. This

decrease is accompanied by the appearance of Cu, approximately 15% in the

product synthesized at 75 °C. At and above 100 °C, the product synthesized with

S:M=6.2 has a composition very close to that expected for CZTS. In contrast, with

S:M=1.9 the amount of sulfur increases more slowly, getting close to the expected

CZTS composition only at 160 °C.

158

Figure A10. SEM images from samples: (a) Cu(I)_Sn(II)_Tu_1.9_160 °C (b)

Cu(I)_Sn(II)_Tu_6.2_160 °C (c) Cu(II)_Sn(II)_Tu_1.9_160 °C (d) Cu(II)_Sn(II)_Tu_6.2_160

°C (e) Cu(I)_Sn(IV)_Tu_1.9_160 °C (f) Cu(I)_Sn(IV)_Tu_6.2_160 °C (g)

Cu(II)_Sn(IV)_1.9_Tu_160 °C (h) Cu(II)_Sn(IV)_Tu_6.2_160 °C.

159

These results corroborate the hypothesis that, at room temperature, only

Zn-Sn intermediates are formed, while copper remains in solution without

incorporation into the solid. As the temperature increases, copper incorporates into

the glycolates and the glycolates transform into CZTS. Also, as described in the

main text, the intermediates synthesized with S:M=6.2 have a higher amount of

sulfur than the intermediates with S:M=1.9, which accelerates the transformation

to CZTS, lowering the temperature CZTS is formed to 100 °C for S:M=6.2.

Figure A11. SEM images from products a) Cu(II)_Sn(II)_Tu_1.9_75 °C, b)

Cu(II)_Sn(II)_Tu_1.9_100 °C, c) Cu(II)_Sn(II)_Tu_1.9_130 °C, d) Cu(II)_Sn(II)_Tu_6.2_75

°C, e) Cu(II)_Sn(II)_Tu_6.2_100 °C, f) .Cu(II)_Sn(II)_Tu_1.9_130 °C.

160

A12 Elemental Analysis of the Product Cu(II)_Sn(IV)_TGacid_1.9_100 °C

As shown in Figure 2.8b in the main text, the sample

Cu(II)_Sn(IV)_TGacid_1.9_100 °C forms a crystalline compound identified as

copper sulfide by XRD. This product was placed onto Au TEM grids and elemental

analysis by TEM-EDS was performed. The results are presented in Table A7.

A13 Influence of the water on the CZTS phase composition

Considering that the syntheses were carried out using the following

reagents: Copper (II) acetate monohydrate (CuAc2·H2O), copper (I) acetate

(CuAc), zinc acetate dehydrate (ZnAc2·2H2O), tin (II) chloride (SnCl2), tin (IV)

chloride pentahydrate (SnCl4·5H2O), and thiourea (CH4N2S), it is necessary to ask

whether the water present in the starting reagents influences the CZTS crystalline

phase produced in the synthesis. Table A8 summarizes the results presented in

Figure 2.4 of the main text, and lists the total amount of water introduced in the

synthesis via the reagents.

Table A7. TEM-EDS analysis of the product Cu(II)_Sn(IV)_TGacid_1.9_100 °C. Elemental

composition is in atom %. Composition from three locations are shown.

Element

Spectrum

1

Spectrum

2

Spectrum

3 Average Std Dev

Cu 89.1 57.6 89.7 78.8 15.0

Zn 0 0 0 0 0.0

Sn 0 0 0 0 0.0

S 10.9 42.4 10.3 21.2 15.0

161

To check if this difference in the hydration water amount has some influence

on the crystalline phases produced in the synthesis, we conducted a synthesis

(specifically, Cu(II)_Sn(II)_1.9_160 C) while including an additional 5 mmols of

deionized water in ethylene glycol, to bring the total water level to 9 mmol of H2O.

As shown in Figure A12, the product was comprised mostly of wurtzite, identical

to the XRD pattern obtained when no additional water was included in the

synthesis (e.g., product Cu(II)_Sn(II)_Tu_1.9_160 C in Figure 2.4. We conclude

that water introduced through starting reagents has no influence on the CZTS

crystalline phases produced.

Table A8. Correlation between hydration water and phase composition of different CZTS

products

Sample Number of mols of

hydration H2O (mmols)

Phase Composition

Cu(I)_Sn(II)_S:M_160 C 2 Wurtzite (S:M = 1.9),

Kesterite (S:M = 6.2)

Cu(I)_Sn(IV)_S:M_160 C 7 Mostly Kesterite, regardless

S:M ratio

Cu(II)_Sn(II)_S:M_160 C 4 Wurtzite (S:M = 1.9),

Kesterite (S:M = 6.2)

Cu(II)_Sn(IV)_S:M_160 C 9 Mostly Kesterite, regardless

S:M ratio

162

Figure A12. XRD pattern of the products Cu(II)_Sn(II)_Tu_1.9_160C, with (top) and without

addition of 5 mmol of water (bottom).

163

APPENDIX B:

SUPPORTING INFORMATION FOR

CHAPTER 3

3. Green Synthesis of Cu2(Zn1-x,Cox)SnS4 Nanocrystals and Formation of

Polycrystalline Thin Films from Their Aqueous Dispersions

B1 Measurement and Control of Co fraction, x in Nanocrystals

Figure B1. Comparison of the nominal Co fraction, x in Cu2(Zn1-x,Cox)SnS4, with the Co fraction,

X, measured using ICP-MS (blue triangles) and SEM-EDS (red squares). The measured cobalt

fraction, X is defined as X=CCo/(CZn+CCo), where Ci is the concentration of species i (Co or Zn).

The nominal Co fraction, x, is the Co fraction in the precursor solutions calculated from the ratio

of the moles of CoAc2 to the sum of the moles of CoAc2 and ZnAc2 (i.e., [CoAc2]/[CoAc2 +

ZnAc2]). The elemental composition determined by ICP-MS and SEM-EDS reveal the

CCo/(CZn+CCo) ratio is very close to expected x value in Cu2(Zn1-xCox)SnS4, as shown in the Figure

B1.

164

B2 Rietveld Refinement and XRD Simulation Details

Rietveld refinement was performed using the software XPert HighScore

Plus. The wurtzite structure (space group P63mc, #186) was simulated using an

initial guess of a = 3.8353 Å, b = 3.8353 Å, c = 6.3276 Å while keeping α = 90°, β

= 90°, γ = 120° and. For Cu2ZnSnS4 (x=0) and Cu2CoSnS4 (x=1). The Wyckoff

positions of the atoms are given in Table B1. For the solid solutions Cu2(Zn1-

xCox)SnS4, the occupational parameter (sof) was adjusted according to the relative

amounts of Zn and Co present in each sample (i.e., x). The Rietveld refinement

was carried out in the following order: (1) Scale factors were refined; (2) the 4 first

background parameters were refined; (3) zero shift was refined; (4) the a and b

lattice parameters were refined; (5) the c lattice parameter refined; (6) the profile

parameter, U, was refined; (7) the peak shape parameter, peak shape 1, was

refined; (8) the profile parameter, V, was refined; (9) the peak shape parameter

peak shape 2 was refined.

Table B1. Wyckoff positions and occupational parameters (sof) used in the Rietveld refinement

and XRD simulations.

Element Wyckoff

Position

x y z sof Biso

Cu 2b 1/3 2/3 0 0.333 0.500

Zn 2b 1/3 2/3 0 0.333 (1-x) 0.500

Co 2b 1/3 2/3 0 0.333 x 0.500

Sn 2b 1/3 2/3 0 0.333 0.500

S 2b 1/3 2/3 0.375 1 0.500

165

B3 XRD patterns from Cu2(Zn1-xCox)SnS4 Nanocrystals as a function of x on

an expanded scale

Figure B2. XRD patterns from Cu2(Zn1-xCox)SnS4 nanocrystals as a function of the nominal Co

fraction, x on an expanded scale. Simulated XRD patterns of wurtzite Cu2ZnSnS4 and Cu2CoSnS4

are shown as stick patterns.

166

APPENDIX C:

SUPPORTING INFORMATION FOR

CHAPTER 5

5. Etching Mechanism of Cu-, Zn-, and Sn-Containing Sulfides in

Ethylenedimaine and 2-Mercaptoethanol Mixture

167

Figure C1. X-ray diffraction patterns for Cu2-xS, ZnS, SnS2, CuxSnySz, Cu2ZnSnS4 nanocrystals. X-

ray diffraction patterns were collected from nanocrystals drop cast and dried from aqueous

dispersion at room temperature. The bottom panels show the expected diffraction patterns for

kesteite Cu2ZnSnS4 (ICDD : 00-026-0575), Cu2SnS3 (ICDD : 00-027-0198), Cu4SnS4 (ICDD : 00-

027-0296), SnS2 (ICDD : 00-023-0677), ZnS (ICCD : 00-005-0566), and CuS (ICDD : 00-006-

0464), respectively.

168

Figure C2. Qualitative etching rate of different proportions by volume of the ethylenediamine-2-

mercaptoethanol mixture. In each experiment were added 7 mg of kesterite CZTS and 1 mL .of

etching mixture.

Figure C2 shows the qualitative etching rate of different proportions of the

mixture ethylenediamine and 2-mercaptoethanol for etching kesterite CZTS.

These results show that the mixtures constituted by 20%, 50%, or 80%

ethylenediamine etched higher amount of CZTS. These results agree with the

conductivity shown on Figure 5.3, since these three mixture proportions present

the higher conductivity values among the ones analyzed.

169

Figure C3. Photographs for reaction between metal salts and neat ethylenediamine or 2-

mercaptoethanol. For each experiment 17 mg of Cu(Ac)2.H2O, 15 mg of Zn(Ac)2.2H2O, 25 mg of

SnCl4.5H2O were added to 4 mL ethylenediamine or 1 mL of 2-mercaptoethanol.

170

Figure C4. X-ray diffraction patterns for the samples shown tin the top part of the figure C3. The

powder samples formed from the reaction between Cu(I) Acetate and ethylenediamine and Sn(IV)

Chloride and ethylenediamine are amorphous. The powder formed from the reaction between

Zn(II) Acetate and ethylenediamine is crystalline, and present peaks very similar to what is

observed for [Zn(en)3](NO3)2 and [Zn(en)3]F2. These similarities lead us to conclude that the

powder observed from the reaction between Zn(II) Acetate and ethylenediamine is

[Zn(en)3]Acetate2. The standard pattern for [Zn(en)3](NO3)2 and [Zn(en)3]F2 were obtained from

Cambridge Structural Database (CSD), with the respective reference codes: RAVJAZ for

[Zn(en)3](NO3)2 and VAPKAY for [Zn(en)3]F2. On the best of our knowledge, there is no reference

pattern for [Zn(en)3]Acetate2 in CSD database.

10 20 30 40 50 60 70

Sn(IV) Chloride + Ethylenediamine

Zn(II) Acetate + Ethylenediamine

Inte

nsi

ty (

arb

. u

nit

s)

2θθθθ (CoKαααα, °)

Cu(I) Acetate + Ethylenediamine

[Zn(en)3]F

2 - CSD Code: VAPKAY

[Zn(en)3](NO

3)2 - CSD Code: RAVJAZ

171

Figure C5. Scheme of (a) Lewis acid-base reaction when ethylenediamine and 2-mercaptoethanol

are mixed and (b) etching mechanism of metal-sulfide nanocrystals.

172

Figure C6. Elemental ratio for (a) CZTS and (b) CTS nanocrystals obtained by TEM EDS as

function of etching times.

Table C1. Elemental ratio for (a) CuxSnySz and (b) CZTS nanocrystals obtained by TEM EDS and

XPS characterization before etching. Elemental ratio Data from TEM EDS were collected from

different 5 points.

Materials Characterizations elements

Cu Zn Sn S

CuxSnySz TEM-EDS 2.2 ± 0.6 1.0 ± 0.1 3.0 ± 0.0

XPS 1.5 1.1 3.0

CZTS TEM-EDS 2.8 ± 1.7 1.1 ± 0.6 1.2 ± 0.2 4.0 ± 0.0

XPS 1.5 1.7 2.7 4.0

173

Figure C7. Scheme for etching mechanism regarding CuxSnySz and Cu2ZnSnS4 nanocrystals.

The elemental composition in the bulk (as provided by TEM-EDS), and

the surface (as provided by XPS) either on CTS and CZTS nanocrystals before

etching, as shown in table C1, reveals that the surface of the CTS and CZTS

nanocrystals is very Cu-poor, whereas the bulk is very Cu-rich. Then, the

deprotonated forms of ethylenediamine and 2-mercaptoethanol bind to the CTS

or CZTS nanoparticle surface, dissolving first the ions of those elements more

abundant in the nanoparticle surface, i.e., Sn(IV) in CTS case, and Sn(IV), and

Zn(II) in CZTS case. After the dissolution of the ions of the elements present in

the surface, the 2-mercaptoethanol species would be able to access and react to

the Cu ions present in the nanoparticle bulk, causing the dissolution of these

ions. Finally, sulfides could bind the ethylenediamine species, forming

N,N’diamine-polysulfide species.

174

Figure C8. XRD pattern of the CZTS thin film containing SnS2 as impurity (shown on Figure 6)

before (bottom pattern) and after (top pattern) etching. The absence of the SnS2 peaks around 24,

27, and 39 ° on the pattern after etching confirms the removal of SnS2, as its shown by Raman

spectroscopy on Figure 5.6. Although Figures 5.1, and 5.5 indicate the inability of the etching

solution to etch SnS2, in the thin film case, this removal may be explained by delamination of SnS2

from the underlying material as a result if intercalation with ethylenediamine.

20 30 40 50 60 70

•••• ••••

••••

•••• ••••

•••• ••••

••••

••••

♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦

♦♦♦♦

After Etching

Before EtchingIn

ten

sity

(a

rb.u

nit

)

2θθθθ (°°°°)

•••• CZTS

♦♦♦♦SnS2