Thermoelectric Behavior of Low Thermal Conductivity Cu-based and IV-V Chalcogenides by Alan Anthony Olvera A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Materials Science and Engineering) in the University of Michigan 2017 Doctoral Committee: Professor Pierre F. P. Poudeu, Chair Professor Bart Bartlett Professor John Heron Professor Emmanouil Kioupakis Alan Anthony Olvera [email protected] ORCID iD: 0000-0002-1511-5543 © Alan Anthony Olvera 2017 DEDICATION To my friends and family. ii ACKNOWLEDGMENTS Within the past five years, my advisor and mentor Professor Pierre Ferdinand P. Poudeu has greatly influenced my approach to solving challenging and unrelenting problems, to which I am incredibly grateful for. The scientific freedom and insight that Professor Poudeu provided during my doctoral studies is something that I can never forget and that I hope to someday pass on to an aspiring scholar. He not only provided this to myself but to every other lab member who I have witnessed pass through the Poudeu lab, whether a high school student, undergraduate student, or a close member of the lab. I am truly thankful to have been a part of the Poudeu lab. I would also like to sincerely thank Professor Manos Kioupakis and Professor John Heron, both of which have in part supported my previous and current research projects. They have graciously offered their knowledge and lab space to myself, where many great ideas have been able to come to fruition. Further, I want to thank Dr. Guangsha Shi, Logan Williams, Jihang Lee, and Kelsey Mengle in the Kioupakis group and Peter Meisenheimer and Steve Novakov in the Heron group who have provided comprehensive work beyond my own capabilities. I additionally thank Professor Bart Bartlett who was very welcoming in letting me use his lab for optical measurements when I was a young graduate student and whose charismatic personality reminded me that graduate school is more than just “by the book” studies, but also a creative outlet for truly inspiring and novel research. I don’t think I could say enough about my colleagues and friends who I have worked with in the Poudeu group. Professor Poudeu had recently moved to the University of Michigan when I joined the group and, despite experiencing a huge move from New Orleans to Michigan that was then followed by a lab renovation, the members of my group were more than welcoming. Dr. Pranati Sahoo, Dr. Yuanfeng Liu, Dr. Honore Djieutedjeu, and Dr. Erica Chen were all major influences on my graduate studies. I am also extremely thankful for my most recent lab mates whose dedication to research leave me impressed and whose friendships I wish to continue. Thanks again for the great time Juan Lopez, Dr. Nicholas Moroz, Ruiming Liu, Joseph Casamento, Grey Garett, Brandon Buchanan, Lueda Shemitraku, Stephanie Tarczynski, Yiqiao Huang, and Sophia Kwon. Next, I want to thank Professor Ctirad Uher and Dr. Alex Page and Trevor Bailey whose invaluable discussion and lab equipment have made a huge impact on the quality of my work. iii Lastly, I would like to thank my mom Maria and my dad Antonio who worked their entire life in Mexico and the United States just to be able to provide my sister and brothers Maricruz, Chris, Vico, Marcos, and Alex an unconditional opportunity to succeed in the paths we chose for ourselves. iv TABLE OF CONTENTS Dedication ii Acknowledgments iii List of Figures viii List of Equations xix List of Appendices xxi Abstract xxii Introduction 1 1.1 Motivation and outline of thesis 1 1.2 A brief introduction to thermoelectrics 5 1.3 Challenges in material optimization 7 1.3.1 Power factor optimization 10 1.3.2 Reducing contributions to thermal conductivity 15 1.4 Current materials 22 1.4.1 Cu-based chalcogenides 22 1.4.1.1 Binary Cu M (M = S, Se, Te) superionic materials 23 2 1.4.1.2 Ternary Cu-based superionic chalcogenides 24 1.4.1.3 Diamond-like Cu-chalcogenides 26 1.4.1.4 IV-V chalcogenides 27 1.5 Designing sustainable methods for Cu-based chalcogenides 27 1.6 Performance issues with current materials 29 Guided Solid-State Reaction: Redox-Induced Direct Structural Transformation from CuSe to CuInSe 31 2 2 2.1 Introduction 31 2.2 Synthesis 32 2.2.1 CuSe precursor 32 2 2.2.2 (1-x)CuSe / (x)CuInSe composites 32 2 2 v 2.3 Characterization 33 2.3.1 Powder X-ray diffraction (XRD) 33 2.3.2 Transmission electron microscopy (TEM) 33 2.3.3 Differential scanning calorimetry (DSC) 33 2.3.4 Fourier transform infrared spectroscopy (FTIR) 34 2.3.5 X-ray photoelectron spectroscopy (XPS) 34 2.4 Results and discussion 34 2.5 Concluding Remarks 45 Partial Indium Solubility Induces Chemical Stability and Colossal Thermoelectric Figure of Merit in Cu Se 46 2 3.1 Broader context 46 3.2 Introduction 46 3.3 Cu Se with nanoinclusions 49 2 3.4 Thermoelectric properties 52 3.5 Chemical stability 59 3.6 Conclusions 60 3.7 Material and methods 60 Stoichiometric and Structural Dependence on the Thermoelectric Properties of the Superionic System Cu Ag Se 63 4-x x 2 4.1 Introduction 63 4.2 Experimental details 65 4.2.1 Synthesis 65 4.2.2 X-ray diffraction 65 4.2.3 Thermal characterization 65 4.2.4 Electrical property characterization 65 4.2.5 Hall effect 66 4.2.6 X-ray photoelectron spectroscopy (XPS) 66 4.2.7 Phase morphology 66 4.3 Structural and morphological characterization 67 4.4 Electrical and thermal transport 70 4.5 Conclusion 74 vi Pb Bi Se : A Lillianite (4,5L) Homologue with Promising Thermoelectric 7 4 13 Properties 76 5.1 Introduction 76 5.2 Experimental Section 78 5.2.1 Synthesis 78 5.2.2 Characterization 78 5.2.2.1 X-ray powder diffraction (XRD) 78 5.2.2.2 Differential scanning calorimetry (DSC) 78 5.2.2.3 Diffuse reflectance Fourier transform infrared (FTIR) spectroscopy 79 5.2.2.4 Transport properties measurement 79 5.2.3 Crystal structure determination 79 5.3 First-principles calculations 83 5.4 Results and Discussion 83 5.4.1 Synthesis and characterization 83 5.4.2 Crystal structure 84 5.4.3 Electronic structure 87 5.4.4 Charge transport properties 89 5.5 Conclusion 93 Future Work 95 6.1 CuSe structural templating for synthesis of new materials 95 2 6.2 Exploring the phase diversity of Sn-Bi-Se compounds for TE applications 100 6.3 High entropy chalcogenides (HEC) 101 Appendix A Supplementary Information for Chapter 3 106 Appendix B Supplementary Information for Chapter 4 120 References 124 vii LIST OF FIGURES Figure 1-1 Graphical representation comparing the most commonly studied thermoelectric materials.† ................................................................................................................................................... 2 Figure 1-2 Thermoelectric figure of merit, ZT, as a function of year demonstrating the significant advancements that have been accomplished within the last decade. Chalcogenide-based systems dominate in the number of state-of-the-art thermoelectric materials due to their unique thermal and electronic properties.14 ...................................................................................................................... 3 Figure 1-3 Number of publications on Cu-based chalcogenide thermoelectrics over the span of five decades, 1965-2017. From 2010 to 2017, the number of related studies surged largely due to a publication on Cu Se by Liu et al.1 .......................................................................................................... 4 2 Figure 1-4 Schematic representation of a thermoelectric couple showing the direction of carrier flow when under a temperature differential. ........................................................................................ 6 Figure 1-5 Relationship between Seebeck coefficient, electrical conductivity, and power factor as a function of carrier concentration. For thermoelectric materials, there is a maximum achievable power factor with a corresponding carrier concentration. Tuning carrier concentration via doping is one successful method to balance the S and σ. ................................................................... 9 Figure 1-6 (Left) Power factor of enhancement consisting of (Si Ge ) (Si B ) (black squares) 80 20 70 100 5 30 nanocomposites as compared to uniformly doped Si Ge B (red circles) and SiGe alloy used 86 14 1.5 in radioisotope generators. (Right) Schematic representation of nanocomposite used in modulation doping.16 .............................................................................................................................. 11 Figure 1-7 Schematic representation of (Left) band gap divergence/convergence and (Right) the effects of ∆E on the onset of bipolar conduction in the Seebeck coefficient. Enlarging the g band gap will shift the onset of bipolar conduction to higher temperatures, while narrowing the band gap will do the opposite. .............................................................................................................. 11 Figure 1-8 Sb content dependence of bipolar conduction onset (T ) and Hall carrier concentration max (n ) vs temperature. HP and HD designate hot pressed and hot deformation. HP is standard H powder sintering, while HD is a secondary hot press to introduce higher defect concentration viii in the bulk material. Because T depends on band structure, defects introduced to the system max have no effect on shifting T .22 .......................................................................................................... 12 max Figure 1-9 A) Schematic representation of the effect that resonant doping has on the local density of states at the top valence band of PbTe (Dashed line indicates pure PbTe). The ZT is increased when E is somewhere within the energy range E . B) ZT values of Tl-doped PbTe F R as compared to Na-doped PbTe. C) Pisarenko plot at 300 K demonstrating the enhancement in Seebeck as compared to typical trend of Seebeck vs hole concentration in other published work. Adapted from Heremans et al.9,10 ............................................................................................... 13 Figure 1-10 A) Schematic representation of the band structure in PbTe with changes in the relative energy of the valance bands as a function of M in the solid solution Pb M Te. Using a solid 1-x x solution in PbTe forces both valence bands to move away from the conduction band but move closer to each other. B) Pisarenko plot comparing the enhanced Seebeck coefficient of solid solution Pb M Te as compared to Tl-doped PbTe and Na-doped PbTe.14 ................................ 14 1-x x Figure 1-11 Schematic representation of the A) zinc blende structure and B) tetragonal chalcopyrite structure with their associated band structures. Decreasing the splitting parameter, ∆ , towards CF zero creates a pseudo-cubic band structure for the chalcopyrite materials. C) Calculating the splitting parameter as function of η shows that D) ZT is maximized when η ≈ 0. Adapted from Zhang et al.20 ............................................................................................................................................. 15 Figure 1-12 Schematic representation of various dimensional phonon scattering sources. Each dimensional discontinuity serves as scattering centers for various ranges of phonon wavelength. ................................................................................................................................................................... 16 Figure 1-13 A) Lattice thermal conductivity as a function of copper concentration with an alloy model of interstitial Cu and substitutional Cu. B) Comparison of different SnTe solid solutions showing Cu Te as the most effective phonon scattering source. Adapted from Pei et al.18 ........ 17 2 Figure 1-14 Bright-field transmission electron micrographs showing dislocation density in A) Pb Sb Se and B) Pb Sb Se solid solutions. Temperature-dependent C) lattice thermal 0.95 0.033 0.97 0.02 conductivity and D) ZT for all compositions of Pb Sb Se. k min represents the calculated 1-x 2x/3 L minimal lattice thermal conductivity for PbSe.13 ................................................................................ 18 Figure 1-15 The effects of grain size reduction on thermal conductivity as demonstrated by grain size reduced A) strontium titanate and B) SiGe. Adapted from Koumoto et al. and Rowe et al. 12 ................................................................................................................................................................... 19 ix