UCLA UCLA Electronic Theses and Dissertations

Title Tetrides and Pnictides for Fast-Ion Conductors, Phosphor-Hosts, Structural Materials and Improved Thermoelectrics

Permalink https://escholarship.org/uc/item/9068g8b2

Author Hick, Sandra Marie

Publication Date 2013

Supplemental Material https://escholarship.org/uc/item/9068g8b2#supplemental

Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital Library University of California UNIVERSITY OF CALIFORNIA Los Angeles

Tetrides and Pnictides for Fast-Ion Conductors,

Phosphor-Hosts, Structural Materials and Improved Thermoelectrics

A dissertation submitted in partial satisfaction of the

requirements for the degree Doctor of Philosophy

in Chemistry

Sandra Marie Hick

2013

ABSTRACT OF THE DISSERTATION

Tetrides and Pnictides for Fast-Ion Conductors,

Phosphor-Hosts, Structural Materials and Improved Thermoelectrics

Sandra Marie Hick

Doctor of Philosophy in Chemistry

University of California, Los Angeles, 2013

Professor Richard B. Kaner, Chair

New routes to solid states materials are needed for discovery and the realization of improved reactions. By utilizing reactive approaches such as solid-state metathesis, pyrolysis, and nitride fluxes new routes to hard materials, phosphor hosts,

and fast ion conductors were developed.

The fast ion conductor lithium silicon nitride, Li2SiN2, was produced in a metathesis reaction between silicon chloride, SiCl4 and lithium nitride, Li3N, initiated in a conventional microwave oven. The Li2SiN2 produced had a conductivity of 2.70

x 10-3 S/cm at 500 °C.

Colorless millimeter-sized crystals of Ca16Si17N34 were synthesized at high temperatures from a flux generated in situ from reaction intermediates. The

ii compound was found to crystallize in the cubic space group F-43m (a = 14.8882 Å).

The powder X-ray diffraction pattern of Ca16Si17N34 matches that of a phase identified

as cubic CaSiN2 in the 1960s but never structurally characterized. In contrast to the

2+ orthorhombic phase of CaSiN2, in which Ca sits in octahedral sites, this cubic phase has Ca2+ in cubic sites that makes it an interesting host for new phosphors and gives rise to unique crystal field splitting.

A carbon-free silazane was formed by reacting metallic silicon in a melt of

sodium amide; pyrolysis of this material offers a convenient route to Si3N4. The use

of oxides for solid-state metathesis reactions was investigated for the formation of

ternary silicon nitrides for phosphor hosts. Red, green, and blue phosphors were

rapidly produced using rare earth oxides dopants. Solid-state metathesis reactions

were investigated for the production of nano-scale B4C and ZrC. The addition of a refractory, hard material significantly improves the mechanical properties of existing thermoelectric materials.

iii The dissertation of Sandra Marie Hick is approved.

M. Frederick Hawthorne

Bruce S. Dunn

Richard B. Kaner, Committee Chair

University of California, Los Angeles

2013

iv Dedication

This work is dedicated to the memories of Glenn and Cindie Blair, Grandma Delores,

Grandpa Orville, and my dear friend Susannah.

For my husband, Richard. Thank you for believing in me.

v Table of Contents

ABSTRACT OF THE DISSERTATION ...... ii

Dedication ...... v

List of Tables ...... xviii

Acknowledgements ...... xix

Vita ...... xxii

Publications and Presentations ...... xxiv

Chapter 1: Introduction to Alternative Solid-State Synthesis

...... 1

References ...... 3

Chapter 2: Microwave Initiated Solid-State Metathesis

Routes to Li2SiN2 ...... 7

Summary ...... 7

Introduction ...... 7

Experimental ...... 9

Reagents ...... 9

Synthesis ...... 9

vi Characterization ...... 11

Results and Discussion ...... 12

Conclusions ...... 19

Acknowledgments ...... 19

References ...... 20

Chapter 3: Synthesis and Crystal Structure of Cubic

Ca16Si17N34 ...... 24

Summary ...... 24

Introduction ...... 24

Reagents ...... 26

Single Crystal Synthesis ...... 26

Powder Synthesis ...... 27

Refractive Index ...... 28

Characterization ...... 28

Results ...... 29

Discussion ...... 33

Conclusions ...... 37

Acknowledgements ...... 38

Supporting Information ...... 38

References ...... 38

vii Chapter 4: Synthesis of Si3N4 via a Polysilazane

Intermediate ...... 42

Introduction ...... 42

Experimental ...... 47

Chemicals ...... 47

Method ...... 47

Characterization ...... 51

Results ...... 51

Discussion ...... 54

Conclusions ...... 58

References ...... 58

Chapter 5: Luminescence in Ternary Silicon Nitrides ...... 63

Introduction ...... 63

Experimental ...... 65

Chemicals ...... 65

Synthesis ...... 66

LiSi2N3 Phosphor Synthesis ...... 66

MgSiN2 Phosphor Synthesis ...... 67

XRD Analysis ...... 67

Luminescence ...... 68

Results and Discussion ...... 68

viii Fluorescence ...... 70

Conclusions ...... 73

References ...... 74

Chapter 6: Solid-State Metathesis Routes to Refractory

Carbides ...... 78

Introduction ...... 78

Experimental ...... 81

Chemicals ...... 81

Synthetic Preparation ...... 82

Torch Synthesis ...... 83

Microwave Synthesis ...... 83

XRD Analysis ...... 84

Results ...... 84

Quantitative Analysis of Calcium Carbide (CaC2) for Carbon and Calcium

Content ...... 90

Thermoelectric Composite ...... 90

Discussion ...... 91

Conclusions ...... 92

References ...... 93

ix Chapter 7: Future Work ...... 98

Introduction ...... 98

Calcium Silicon Nitride ...... 98

Refractive Carbides ...... 98

Luminescence in Ternary Nitrides ...... 99

Terbium Silicide ...... 101

Experimental ...... 102

Results ...... 103

Future Work ...... 107

Silicon Nitride ...... 108

References ...... 108

x List of Figures

Figure 1–1. High-speed photo frames of the solid-state metathesis (SSM) reaction

between MoCl5 and Na2S producing crystalline MoS2. The time from

the reaction’s self-initiation (time index = 0.04 s) to completion is 2

seconds. The reaction generates so much heat that after completion the

products continue to glow red hot...... 1

Figure 2–1. A photograph taken through the door of a conventional microwave oven

(2.45 GHz) showing the reaction: 3 SiCl4 + 4 Li3N + 1.5 LiNH2 + 6

NH4Cl in a capped vial after 1 minute at 100% power. The reaction

image has been enhanced to emphasize the reaction zone and the outline

of the glass vial has been added in white for clarity. Note that the input

of microwave energy induces multiple reaction sites as indicated by the

arrows...... 13

Figure 2–2. Powder X-ray diffraction patterns of the washed Li2SiN2 product formed

from the reactions (A) 3 SiI4 + 4 Li3N + 1.5 LiNH2 + 2 NH4Cl; and (B)

3 SiCl4 + 4 Li3N + 1.5 LiNH2 + 6 NH4Cl. A simulated pattern (C) for

Li2SiN2 is provided as a reference (JC-PDS 23-0365). The large, broad

peak near 15 degrees 2Θ is due to the specimen holder...... 14

Figure 2–3. Scanning electron micrograph at a magnification of 250X of washed

Li2SiN2 powder formed from the reaction of 3 SiCl4 + 4 Li3N + 1.5

LiNH2 + 6 NH4Cl...... 15

xi Figure 2–4. A Nyquist plot of the impedance for a Li2SiN2 pellet. The measurement

was performed at 500 ˚C in a frequency range from 0.1 Hz to 100 kHz.

The minimum impedance of 131 ohms corresponds to a conductivity of

2.7 x 10-3 S/cm...... 18

Figure 3–1. Optical microscope images of the transparent Ca16Si17N34 crystals

produced in this study. The left image is under 50x magnification and

the right is under 500x...... 30

Figure 3–2. Scanning electron microscope images of Ca16Si17N34 separated from the

reaction mass using a 170-mesh screen. This nitride exhibits cubic

crystals often in the form of cuboctahedra (right)...... 30

Figure 3–3. The unit cell for cubic Ca16Si17N34 (left) consists of clusters of SiN4 units

(gray tetrahedra), while that of orthorhombic CaSiN2 (right) consists of

SiN4 units in rings. The calcium atoms have been removed to highlight

the connectivity of the SiN4 units...... 32

Figure 3–4. The Ca1 and Ca3 positions are 6-coordinate with the Ca1 position in a

severely distorted octahedron. The Ca2 position is 8-coordinate in an

elongated cubic site...... 34

Figure 3–5. The powder X-ray diffraction pattern simulated from the structural data

(top) matches the measured pattern (bottom) and the deleted JC-PDS

file 40-1151 (middle)...... 34

Figure 3–6. The powder X-ray diffractogram of the product produced from the

reaction 16CaCN2 + 17Si + N2→ Ca16Si17N34 + 16C shows that the

xii sample contains silicon (*) and graphite (o) impurities with the

remaining peaks matching the pattern expected for Ca16Si17N34...... 37

Figure 4–1. The furnace used to melt the NaNH2 and subsequently react it with

silicon. The assembly is in a glove box under an inert atmosphere of

helium...... 49

Figure 4–2. The assembly used to wash the product from the reaction of silicon with

NaNH2. The blue color in the liquid ammonia is due to solvated

electrons...... 50

Figure 4–3. The furnace assembly for the annealing of the intermediate product from

the reaction of silicon with NaNH2 under a stream of ultrapure helium.

Any traces of water are removed from the helium stream by flowing the

gas through the Cr2+ scavenging column, a cryo-trap filled with 4 Å

molecular sieves, and finally a Mn2+ scavenging column. The sample is

held in a graphite boat loaded into an alumina tube while under helium.

This tube is transferred to the furnace via a glove bag placed around the

outlet...... 50

Figure 4–4. The bottom XRD pattern is the result of a reaction between Si and

NaNH2 and subsequent washing in liquid NH3 followed by annealing

under dry He atmosphere at 1500 ˚C. The top XRD pattern is a

simulated pattern for alpha Si3N4 (JCPDS 41-0360)...... 52

Figure 4–5. The FTIR spectrum for the intermediate formed by the reaction of

silicon with NaNH2. The Si-H stretching band along with the Si-N and

xiii N-H stretches provide evidence for a possible polysilazane polymer

intermediate...... 53

Figure 4–6. Solid-state MAS-NMR spectra of the intermediate formed by the

29 reaction between Si and KNH2. The coupled spectrum for Si is dashed

and the decoupled spectrum is solid. In the uncoupled spectrum, the

asymmetric peak at ~ -43 ppm is evidence of more than one type of Si

environment. The peak at ~ -79 ppm is due to elemental Si that was an

impurity in the product. The coupled spectrum, shows interaction

between 29Si and 1H. The predominant asymmetric peak indicates the

presence of more than one type of Si environment. The elemental Si

peak is not present since there would be no 1H interaction...... 54

Figure 5–1. Powder X-ray diffraction pattern of LiSi2N3 as synthesized (bottom

trace), 0.20 mol Eu doped into the LiSi2N3 lattice (middle), and 0.20

mol Tb doped into the LiSi2N3 lattice (top trace)...... 69

Figure 5–2. Powder X-ray diffraction pattern of MgSiN2 as synthesized (bottom

trace), 0.10 mol Eu doped into the MgSiN2 lattice (middle), and

simulated pattern of MgSiN2 (JC-PDS 25-0530) (top trace)...... 69

Figure 5–3. Emission spectra of Eu doped LiSi2N3 (emission at 475 nm, excitation

wavelength of 360 nm), MgSiN2 doped with Eu (emission at 618 nm,

excitation wavelength of 380 nm) and europium oxide (excitation

wavelength of 380 nm)...... 71

xiv Figure 5–4. Emission spectra of Tb doped LiSi2N3 (broad emission centered around

546 nm, excitation wavelength of 340 nm), LiSi2N3 and terbium oxide.

...... 72

Figure 5–5. Emission spectra of Tm doped LiSi2N3 (broad emission centered around

410 nm, excitation wavelength of 300 nm). Unexpectedly, LiSi2N3 (429

nm) and thullium oxide (374 nm) also appear to emit...... 73

Figure 6–1. Vicker’s hardness of common materials...... 79

Figure 6–2. Crack pinning in a nanoparticle reinforced material. Crack propagation

stops after a crack encounters a nanoparticle...... 80

Figure 6–3. Powder X-ray diffraction pattern of the washed B4C product formed

from the reaction MgB2 + C2Cl6 + S that was initiated by a resistively

heated nichrome wire in a bomb-type reactor. A simulated pattern for

B4C is provided as a reference (JC-PDS 35-0798). The remaining peaks

can be attributed to graphite (JC-PDS 08-0415) and an unknown

impurity...... 85

Figure 6–4. Powder X-ray diffraction pattern of the washed ZrC product formed

from the microwave initiated reaction of ZrCl4 + 2 CaC2 overlaid with a

simulated pattern of ZrC (JC-PDS 35-0784). The remaining peaks can

be attributed to two different phases of ZrO2 (•, monoclinic, JC-PDS 37-

1484 and *, tetragonal, JC-PDS 17-0923) and graphite ( JC-PDS 26-

1079)...... 87

xv Figure 6–5. Powder X-ray diffraction pattern of the washed ZrC formed from the

reaction of ZrCl4 + 2 CaC2 and initiated by an oxy-methane torch. The

simulated pattern of ZrC (JC-PDS 35-0784) is overlaid. Two different

phases of ZrO2 (•, monoclinic, JC-PDS 37-1484 and *, tetragonal, JC-

PDS 17-0923) and graphite ( JC-PDS 26-1079) account for the

remaining peaks...... 88

Figure 6–6. Comparison of newly machined CoSb3 (top) and B4C/CoSb3 composite

(bottom). The CoSb3 sample shows tooling marks and broken edges

from the machining process, while the B4C/CoSb3 sample shows no

such damage...... 91

Figure 7–1. Powder XRD pattern of the product of SiO2 + Mg3N2 + Li3N + Li2O +

Ca3N2 + Tb4O7 initiated by an oxy-methane torch and heated for

approximately 25 seconds (overlaid with the JCPDS diffraction pattern

number 01-072-2000 for hexagonal TbSi2). The reaction SiO2 + Mg3N2

+ Li3N (w/ Li2O impurities) + Ca3N2 + Tb4O7 produced a similar pattern

when heated between a range of 40-90 seconds. The remaining peaks

match with the JCPDS diffraction pattern number 00-017-0901 for

cubic silicon...... 104

Figure 7–2. Powder XRD pattern of the product of CaSi2 + Tb4O7 reacted by

mechanochemical initiation in a ball mill (overlaid with the JCPDS

diffraction pattern number 00-013-0383 for orthorhombic TbSi2). The

xvi remaining peaks match with the JCPDS diffraction pattern number 01-

078-2500 for cubic silicon...... 105

Figure 7–3. Powder XRD pattern of the product of CaSi + Tb4O7 initiated by an oxy-

methane torch (overlaid with the JCPDS diffraction pattern number 01-

072-2000 for TbSi2). The reaction was unsuccessful in producing TbSi2.

The product corresponds to calcium silicate (Ca2SiO4) and unreacted

CaSi...... 106

Figure 7–4. Powder XRD pattern of the product of CaSi2 + MgSi2 + Tb4O7 initiated

by an oxy-methane torch and heated for 40 seconds (overlaid with the

JCPDS diffraction pattern number 01-072-2000 for hexagonal TbSi2).

The remaining peaks match with the JCPDS diffraction pattern numbers

00-017-0901 and 03-065-0387 for cubic silicon and silicon dioxide,

respectively...... 107

xvii List of Tables

Table 3–1. Ca16Si17N34, forming reactions. The reaction marked with a (*) produced

the most crystalline product...... 31

Table 3–2. Crystal data and structural refinement parameters for cubic Ca16Si17N34. 31

Table 3–3. Atomic coordinates and equivalent isotropic displacement parameters

3 (x10 ) for cubic Ca16Si17N34. U(eq) is defined as one third of the trace

of the orthogonalized Uij tensor...... 32

Table 6–1. Results of attempted B4C forming reactions broken down by initiation

method...... 86

Table 6–2. Results of ZrC Forming Reactions Broken Down by Initiation Method 89

xviii Acknowledgements

At UCLA, I’d first like to thank Ric Kaner for supporting me in the pursuit of the science contained in this dissertation. In the Kaner group, this entire dissertation would not have been possible without the support, experiments, sharing of knowledge, and friendship of Richard G. Blair. The same can be said of fellow group

members Art Anderson and Sabah Bux.

My stay at UCLA would have been more miserable if not for some great

undergraduate researchers. Dominick Daurio and Kim Nguyen were members of the

group upon my arrival to UCLA; I am glad that I got to share the lab with them. The

chapter on phosphors could not have been possible without the research of Julie Chen

and Cindy N. Tran; they inspired me to keep trying new reactions and keep going

where Richard Blair’s research with the red phosphor left off. Cindy was a great

partner in the research that lead us into the violet phosphors (Chapter 5). Marc A.

Rodriguez and Velveth Klee stayed with me despite the rough times and drudgery of

the quest for carbides (Chapter 6). Marc’s hard work and determination pushed the

serendipitous discovery of a metathesis route to terbium silicide beyond the initial

reaction and reagent hiccups (Chapter 7).

The phosphor work could not have moved forward without help from the Zink

Research Group (Payam Minoofar, Johnny Skelton, Paul Sierocki) and the Schwartz

Research Group (Ian Craig, Alexander Ayzner).

xix I owe many thanks to other members of the Kaner Group and my friends in the greater LA area for providing me with support and encouragement. Special thanks go to: RoseMary Anderson, Robert Cumberland, Andrea Danek, Eduardo

Falcão, Jiaxing Huang, Becky Janes, Erika Karnell, Suzette and Beau Lavine, Mark

Lee, Julia and Dan Mack, and Shabnam Virji without whom with working in the lab for such long hours would have been unbearable. Without the friendship of Chris,

Denise and Wesley, Derrick, Grant, Ian, John, Lisa, Maeline, Noel, Pete, Sam, Steve and Claire, Tad, Tony, Ty, Zu and Barb, Los Angeles would not have been so

inviting.

Chapter 2 is reproduced with permission from Anderson, Arthur J.; Blair,

Richard G.; Hick, Sandra M.; Kaner, Richard B., “Microwave Initiated Solid-State

Metathesis Routes to Li2SiN2,” Journal of Materials Chemistry, 2006, 16, 1318-1322

Chapter 3 is reproduced with permission from Hick, Sandra M.; Miller,

Mattheu I.; Kaner, Richard B.; Blair, Richard G., “Synthesis and Crystal Structure of

Chemical Society.

This work is supported by the National Science Foundation Division of

Materials Research (DMR-0453121) and equipment grants (DMR-0315828 and

DMR-0114002).

I would be remiss if I did not thank the people who taught me well before I arrived at

UCLA. Virginia Bryan, Eric Voss, and Doug Eder did a great job of mentoring and

xx inspiring me to higher standards. Without the love, support and patience of my family, this entire pursuit would not have been possible. To my grandparents Orville,

Evelyn, Dolores, Joe, and Berta; my parents, Tom and Karen; my brother Brian; my aunts and uncles; and to my in-laws, Glenn and Cindie, Andy, Sarah, and the aunts, uncles and cousins – Thank You.

xxi Vita

1991 Phi Eta Sigma Member Southern Illinois University, Edwardsville Edwardsville, IL

1991-1995 Presidential Scholarship Southern Illinois University, Edwardsville Edwardsville, IL

1993-1994 Laboratory Assistant Division of Drug Analysis United States Food and Drug Administration St. Louis, MO

1994 Undergraduate Teaching Assistant Southern Illinois University Edwardsville Edwardsville, IL

1995 Laboratory Assistant Division of Drug Analysis United States Food and Drug Administration St. Louis, MO

1995 Bachelor of Science, Chemistry - ACS Southern Illinois University Edwardsville Edwardsville, IL

1995-1997 Teaching Assistant Southern Illinois University Edwardsville Edwardsville, IL

xxii 1996 Recipient Research Grants for Graduate Students Southern Illinois University, Edwardsville Edwardsville, IL

1998 Master of Science, Chemistry Southern Illinois University Edwardsville Edwardsville, IL

1998-1999 Lecturer Department of Chemistry Southern Illinois University Edwardsville Edwardsville, IL

1998-2001 Assistant to the Director Undergraduate Assessment & Program Review Southern Illinois University Edwardsville Edwardsville, IL

2001-2002 Teaching Associate University of California, Los Angeles Los Angeles, CA

2002 Fellowship Materials Creation and Training Program The National Science Foundation University of California, Los Angeles Los Angeles, CA

2002-2004 National Science Foundation – IGERT – Materials Creation Training Program Fellow University of California, Los Angeles Los Angeles, CA

xxiii 2002-2006 Research Associate University of California, Los Angeles Los Angeles, CA

2003 Scientific Consultant Rockwell Scientific Company Thousand Oaks, CA

2004 Student Travel Award Solid-State Gordon Conference New London, NH

2007-2009 Visiting Scientist University of Central Florida Orlando, FL

2010-2012 Chemical Inventory and Safety Assistant University of Central Florida Orlando, FL

2012-2013 Chemical Safety and Security Coordinator University of Central Florida Orlando, FL

Publications and Presentations

(1) Restrepo, D., S.M. Hick, C. Griebel, J. Alarcon, K. Giesler, Y. Chen, N. Orlovskaya, and R.G. Blair, Size Controlled Mechanochemical Synthesis of

ZrSi2. Chem. Commun. , 2013. 49(7): p. 707-709.

(2) Nash, D., K. Rosario, S.M. Hick, V. Turkowski, T. Rahman, and R.G. Blair. Transition Metal Cluster Compounds for the Fluorescent Identification and Trace Detection of Substances of Abuse. in 88th ACS Annual Florida Meeting and Exposition, 19 May. 2012. Tampa, FL.

(3) Hick, S.M., M.I. Miller, R.B. Kaner, and R.G. Blair, Synthesis and Crystal

Structure of Cubic Ca16Si17N34. Inorg. Chem., 2012. 51(23): p. 12626-12629.

xxiv (4) Blair, R.G., S.M. Hick, and J.H. Truitt, Soluble sugars produced according to a process of non-aqueous solid acid catalyzed hydrolysis of cellulosic materials,Patent US CIP 61/721,316, 2012

(5) Griebel, C., S. Hick, and R. Blair. Mechanochemical Synthesis of Group 1 Carbides. in 87th ACS Annual Florida Meeting and Exposition, 13 May. 2011. Tampa, FL.

(6) Hick, S.M., C. Griebel, D.T. Restrepo, J.H. Truitt, E.J. Buker, C. Bylda, and R.G. Blair, Mechanocatalysis for Biomass-Derived Chemicals and Fuels. Green Chem., 2010. 12(3): p. 468-474.

(7) Hick, S.M. and R.G. Blair. Mechanochemical Synthesis of Nanocrystalline Conductive Silicides. in 239th ACS National Meeting, 24 March. 2010. San Francisco, CA.

(8) Hick, S., C. Griebel, and R. Blair. Mechanochemical Synthesis of Nanocrystalline Conductive Silicides. in 86th ACS Annual Florida Meeting and Exposition, 13 May. 2010. Tampa, FL.

(9) Rosario, K., S. Hick, and R. Blair. Novel Precipitation Agents for the Rapid Identification of Pharmaceuticals. in 85th ACS Annual Florida Meeting and Exposition, 14 May. 2009. Orlando, FL.

(10) Hick, S.M., C. Griebel, and R.G. Blair, Mechanochemical Synthesis of Alkaline Earth Carbides and Intercalation Compounds. Inorg. Chem., 2009. 48(5): p. 2333-2338.

(11) Blair, R.G., S.M. Hick, and J.H. Truitt, Solid Acid Catalyzed Hydrolysis of Cellulosic Materials,Patent US20090118494A1, 2009

(12) Blair, R.G., S.M. Hick, and J.H. Truitt, Solid Acid Catalyzed Hydrolysis of Cellulosic Materials,Patent WO2009061750A1, 2009

(13) Griebel, C., S.M. Hick, and R.G. Blair. Mechanochemical Synthesis of Alkaline Earth Carbides and Intercalation Compounds. in Florida Inorganic & Materials Symposium, September. 2008. Gainesville, FL.

(14) Tran, C.N., S.M. Hick, and R.B. Kaner. Ternary Silicon Nitride Phosphors through Solid-State Metathesis: the Search for Blue. in 40th Western Regional Meeting of the American Chemical Society, 22-25 January. 2006. Anaheim, CA, United States.

xxv (15) Rodriguez, M.A., V. Klee, S.M. Hick, and R.B. Kaner. Refractory Carbides via Rapid Solid-State Synthesis. in Abstracts, 40th Western Regional Meeting of the American Chemical Society, 22-25 January. 2006. Anaheim, CA, United States.

(16) Anderson, A.J., R.G. Blair, S.M. Hick, and R.B. Kaner, Microwave Initiated

Solid-State Metathesis Routes to Li2SiN2. J. Mater. Chem., 2006. 16(14): p. 1318-1322.

(17) Hick, S.M., C.N. Tran, R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Lanthanide Doped Nitride Phosphors. in 229th ACS National Meeting, 15 March. 2005. San Diego, CA.

(18) Anderson, A.J., R.G. Blair, S.M. Hick, and R.B. Kaner. Metathesis Routes to Binary and Ternary Silicon Nitrides. in 229th ACS National Meeting, 15 March. 2005. San Diego, CA.

(19) Viculis, L.M., J.J. Mack, S.-C. Huang, J. Huang, S.M. Hick, Z. Ding, E. Falcao, F. Wudl, and R.B. Kaner. Intercalation And Exfoliation Routes To Nanoscrolled Materials. in 227th ACS National Meeting, 28 March. 2004. Anaheim, CA.

(20) Hick, S.M., C.N. Tran, R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Lanthanide Doped Nitride Phosphors. in Gordon Research Conference, Solid-State Chemistry I, 25-30 July. 2004. New London, New Hampshire.

(21) Hick, S.M., C.N. Tran, R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Lanthanide Doped Nitride Phosphors. in UC Inorganic Materials Conference, 27 August. 2004. University of California, Santa Barbara, California.

(22) Hick, S.M., C.N. Tran, R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Lanthanide Doped Nitride Phosphors. in Materials Creation and Training Program Symposium, 5 November. 2004. University of California, Los Angeles, California.

(23) Hick, S.M., R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Europium Doped Nitride Phosphors. in 227th ACS National Meeting, 30 March. 2004. Anaheim, CA.

xxvi (24) Hick, S.M., R.G. Blair, and R.B. Kaner. Rapid Solid State Synthesis of Europium Doped Nitride Phosphors. in Materials Creation and Training Program Symposium, 14 November. 2003. University of California, Los Angeles, California.

(25) Hick, S.M., R.B. Kaner, M.E. Barr, A. Lupinetti, and P. Berbon. Novel Actinide Storage Materials and Metal Matrix Composites.” Materials Creation and Training Program Symposium. in Materials Creation and Training Program Symposium, 14 November. 2002. University of California, Los Angeles, California.

(26) Bernaix, L.W., D.J. Eder, S.M. Hick, and C.A. Schmidt. Principles and Good Practices for Authentic Assessment with an Emphasis on Nursing and the Health Sciences. in The 2000 IUPUI Assessment Institute, 7 November. 2000. Indiana University-Purdue University Indianapolis, Indianapolis, Indiana.

(27) Hick, S.M. Writing Across the Curriculum. in Engaged Learning Conference, 23 November. 1998. outhern Illinois University Edwardsville, Edwardsville, Illinois.

(28) Eder, D.J. and S.M. Hick. Writing Across the Curriculum. in Presentation to Science Faculty, 17 April. 1998. Alton Senior High School, Alton, Illinois.

(29) Hick, S.M. and E.J. Voss. The Synthesis and Characterization of Lanthanide Diolates. in 21st Annual William J. Probst Memorial Lecture Series Student Research Poster Session, 23 April. 1997. Southern Illinois University Edwardsville, Edwardsville, Illinois.

(30) Hick, S.M. and E.J. Voss. The Synthesis and Characterization of Lanthanide Diolates. in 213th ACS National Meeting, 15 April. 1997. San Francisco, CA.

(31) Hick, S.M. The Synthesis and Characterization of Lanthanide Diolates. in Graduate Research Seminar, Department of Chemistry, 9 April. 1997. Southern Illinois University Edwardsville, Edwardsville, Illinois.

(32) Hick, S.M. and E.J. Voss. The Synthesis and Characterization of Lanthanide Diolates. in 89th Annual Meeting, Illinois State Academy of Science. 18 October. 1996. Illinois Wesleyan University, Bloomington, Illinois.

(33) Hick, S.M. Solution-Liquid-Solid Growth of Crystalline III-V Semiconductors: An Analogy to Vapor-Liquid-Solid Growth. in Graduate Literature Seminar, Department of Chemistry, 13 March. 1996. Southern Illinois University Edwardsville, Edwardsville, Illinois.

xxvii (34) Friend, R.J., S.J. Wall, S.M. Hick, L.A. Miller, and E.J. Voss. Chemistry club on campus and in the community: Activities of the 1994-95 SIUE student affiliates chapter. in 210th ACS National Meeting, August 20-24. 1995. Chicago, IL, United States.

xxviii

Chapter 1: Introduction to Alternative

Solid-State Synthesis

“From a little spark may burst a mighty flame.” — Dante Alighieri, Paradiso, Canto 1, line 34

Dante could have easily been describing a solid-state reaction; in particular,

self-propagating high-temperature synthesis (SHS) and solid-state metathesis (SSM) reactions. SHS and SSM reactions are initiated by a small amount of energy (an electrical spark, heat from a torch, or even grinding of the reactants) and produce tremendous amounts of heat; sometimes, these reactions even produce flames in an oxygen-rich atmosphere.

Figure 1–1. High-speed photo frames of the solid-state metathesis (SSM) reaction between MoCl5 and Na2S producing crystalline MoS2. The time from the reaction’s self-initiation (time index = 0.04 s) to completion is 2 seconds. The reaction generates so much heat that after completion the products continue to glow red hot.

SHS and SSM are two methods for carrying out a solvent-free reaction

between solids. The methods of producing solid-state materials include the

traditional “shake and bake,” the novel self-propagating high-temperature synthesis

(SHS), and more exotic syntheses (such as with an arc-melter). Other common synthetic techniques used to synthesize materials in the solid-state include ball- milling, carbothermal reduction, flux growth, hydrothermal, intercalation, melt

processing and vapor phase transport to name a few. 1 These methods can be used

over a wide range of temperatures. Many methods of solid-state synthesis require

large amounts of energy and several days before the final product is in hand.2

Solid-state (and solid-solid) reactions have been used to produce a variety of materials including: sulfides,3, 4 nitrides,5-11 carbides,12-14 borides,11, 15, 16 conducting polymers,17 binary and ternary compounds,9, 18, 19 alloys and solid solutions.20

This dissertation presents the results of efforts to find new routes to hard

materials, create superior reinforcement materials, a faster route to more robust solid-

state lighting, and a few serendipitous discoveries. Much of the work in the following

pages is a result of the search for a new route to silicon nitride and the realization that

just because the desired product is not formed, does not mean that what was formed is

useless and without merit.

Rather than be driven by a single synthetic method (such as solid-state

metathesis), the production of materials drives this dissertation. The second chapter

of this work explores a novel SSM initiation method (microwave) to react an

industrial by-product, SiCl4, and lithium nitride to synthesize a fast ion conductor,

Li2SiN2. In Chapter 3, the synthesis and characterization of a long debated structure

of cubic-calcium silicon nitride is presented utilizing a more traditional approach, the

prolonged heating of solids. A new route to α-Si3N4 utilizing an amide melt

combined with high-temperature annealing of the first reported inorganically derived

polysilazane intermediate is detailed in Chapter 4. In Chapter 5, solid-state

metathesis reactions are used to explore the synthesis of light-emitting materials.

Chapter 6 combines many of the initiation methods from the previous chapters to explore refractory carbides. Along the way, many different reaction pathways were investigated to achieve the results above and some of them yielded unexpected results, these reactions and other future work are discussed in Chapter 7.

References

(1) Stein, A., S.W. Keller, and T.E. Mallouk, Turning down the heat: design and

mechanism in solid-state synthesis. Science, 1993. 259(5101): p. 1558-64.

(2) Rao, R.R. and T.S. Kannan, Nitride-bonded silicon nitride from slip-cast Si +

Si3N4 compacts. Journal of Materials Research, 2002. 17(2): p. 386-395.

(3) Parkin, I.P. and A.T. Rowley, Metathesis routes to tin and lead chalcogenides.

Polyhedron, 1993. 12(24): p. 2961-4.

(4) Bonneau, P.R., R.F. Jarvis, Jr., and R.B. Kaner, Solid-state metathesis as a

quick route to transition-metal mixed dichalcogenides. Inorganic Chemistry,

1992. 31(11): p. 2127-32.

(5) Janes, R.A., M.A. Low, and R.B. Kaner, Rapid Solid-State Metathesis Routes

to Aluminum Nitride. Inorganic Chemistry, 2003. 42(8): p. 2714-2719.

(6) Cumberland, R.W., R.G. Blair, C.H. Wallace, T.K. Reynolds, and R.B. Kaner,

Thermal Control of Metathesis Reactions Producing GaN and InN. Journal of

Physical Chemistry B, 2001. 105(47): p. 11922-11927.

(7) O'Loughlin, J.L., C.H. Wallace, M.S. Knox, and R.B. Kaner, Rapid Solid-

State Synthesis of Tantalum, Chromium, and Molybdenum Nitrides. Inorganic

Chemistry, 2001. 40(10): p. 2240-2245.

(8) Wallace, C.H., S.-H. Kim, G.A. Rose, L. Rao, J.R. Heath, M. Nicol, and R.B.

Kaner, Solid-state metathesis reactions under pressure: A rapid route to

crystalline gallium nitride. Applied Physics Letters, 1998. 72(5): p. 596-598.

(9) Blair, R.G., A. Anderson, and R.B. Kaner, A Solid-State Metathesis Route to

MgSiN2. Chemistry of Materials, 2005. 17(8): p. 2155-2161.

(10) Shemkunas, M.P., G.H. Wolf, K. Leinenweber, and W.T. Petuskey, Rapid

synthesis of crystalline spinel tin nitride by a solid-state metathesis reaction.

Journal of the American Ceramic Society, 2002. 85(1): p. 101-104.

(11) Gibson, K., M. Stroebele, B. Blaschkowski, J. Glaser, M. Weisser, R.

Srinivasan, H.-J. Kolb, and H.-J. Meyer, Solid state metathesis reactions in

various applications. Zeitschrift fuer Anorganische und Allgemeine Chemie,

2003. 629(10): p. 1863-1870.

(12) Nartowski, A.M., I.P. Parkin, M. Mackenzie, and A.J. Craven, Solid state

metathesis: synthesis of metal carbides from metal oxides. Journal of

Materials Chemistry, 2001. 11(12): p. 3116-3119.

(13) Nartowski, A.M., I.P. Parkin, M. MacKenzie, A.J. Craven, and I. MacLeod,

Solid state metathesis routes to transition metal carbides. Journal of Materials

Chemistry, 1999. 9(6): p. 1275-1281.

(14) Hick, S.M., C. Griebel, and R.G. Blair, Mechanochemical Synthesis of

Alkaline Earth Carbides and Intercalation Compounds. Inorg. Chem., 2009.

48(5): p. 2333-2338.

(15) Rao, L., E.G. Gillan, and R.B. Kaner, Rapid synthesis of transition-metal

borides by solid-state metathesis. Journal of Materials Research, 1995. 10(2):

p. 353-61.

(16) Lupinetti, A.J., J.L. Fife, E. Garcia, P.K. Dorhout, and K.D. Abney, Low-

Temperature Synthesis of Uranium Tetraboride by Solid-State Metathesis

Reactions. Inorganic Chemistry, 2002. 41(9): p. 2316-2318.

(17) Huang, J., J.A. Moore, J.H. Acquaye, and R.B. Kaner, Mechanochemical

Route to the Conducting Polymer Polyaniline. Macromolecules, 2005. 38: p.

317-321.

(18) Anderson, A.J., R.G. Blair, S.M. Hick, and B. Kaner Richard, Microwave

initiated solid-state metathesis routes to Li2SiN2. Journal of Materials

Chemistry, 2006. 16: p. 1318-1322.

(19) Blair, R.G., E.G. Gillan, N.B.K. Nguyen, D. Daurio, and R.B. Kaner, Rapid

Solid-State Synthesis of Titanium Aluminides. Chemistry of Materials, 2003.

15(17): p. 3286-3293.

(20) Kaner, R.B., S.K. Bux, J.-P. Fleurial, and M. Rodriguez, Rapid solid-state

metathesis routes to nanostructured silicon-germanium,Patent

US20110318250A1, 2011

Chapter 2: Microwave Initiated Solid-State

Metathesis Routes to Li2SiN2

Summary

The fast ion conductor lithium silicon nitride, Li2SiN2, is produced in a metathesis reaction between silicon chloride, SiCl4 and lithium nitride, Li3N, initiated

in a conventional microwave oven. Lithium amide, LiNH2 and ammonium chloride,

NH4Cl, serve to control the temperature and enhance the yield of Li2SiN2. A reaction mechanism is proposed based on the exothermicity of forming the by-product salt

lithium chloride with silicon nitride, Si3N4, as a likely intermediate. By varying the

stoichiometry of the reactants the phase LiSi2N3 can also be produced. The Li2SiN2

product is characterized using powder X-ray diffraction, scanning electron

microscopy and complex impedance spectroscopy. A pressed pellet of Li2SiN2 has a

conductivity of 2.70 x 10-3 S/cm at 500 °C.

Introduction

The synthesis and properties of lithium silicon nitrides have been studied for a number of years due to potential applications as fast ion conductors and sensors.1-5

Anderson, A.J., R.G. Blair, S.M. Hick, and R.B. Kaner, Microwave Initiated Solid-

State Metathesis Routes to Li2SiN2. J. Mater. Chem., 2006. 16(14): p. 1318-1322. - Reproduced by permission of The Royal Society of Chemistry

Since lithium ions can move rapidly through an essentially static silicon nitride

framework,2 these materials are potentially useful as solid-state electrolytes in lithium batteries.6 Additionally, lithium silicon nitrides are of interest as the active components in nitrogen sensors.7

Most methods for preparing lithium silicon nitrides involve reacting Si3N4 with Li3N at elevated temperatures.8 This requires a significant input of energy and is limited to producing the most thermodynamically stable nitride under the reaction conditions employed. Additionally, silicon nitride is a relatively expensive starting material. A less costly route to lithium silicon nitrides could lead to their wider use.

Double displacement reactions, also known as solid-state metathesis (SSM) reactions, have shown promise in the rapid synthesis of a wide range of crystalline refractory ceramics9-11 and other inorganic solids.12-15 These reactions are driven by the formation of stable salts that can be subsequently washed away leaving behind the desired product. By choosing appropriate precursors, it is possible to control crystallite size16 and form high quality solid-solutions and ceramic composites from

these displacement reactions.17

Although some metathesis reactions initiate spontaneously,16 most are ignited by heating e.g. by passing a current through a nichrome wire. The use of microwave energy has proven to be another viable method for initiating solid-state reactions.18, 19

Microwave initiated synthesis has even produced phases different (e.g. TiAl) from

that formed by traditional SSM reactions initiated with a resistively heated wire.19

Here we examine the synthesis of Li2SiN2 via SSM. By utilizing SiCl4 as a silicon source and Li3N as a nitrogen source, lithium silicon nitrides can be produced rapidly. The use of SiCl4 as a precursor to Li2SiN2 is particularly attractive since

SiCl4 is an industrial by-product created during the isolation of various metals from

20 silicate minerals. The addition of LiNH2 and NH4Cl allows the formation of phase- pure Li2SiN2. While reactions initiated by a heated filament in a steel vessel under an

inert atmosphere have shown some positive results, the use of microwave heating is

demonstrated to be a novel method to drive these reactions to completion. Product

characterization includes powder X-ray diffraction, scanning electron microscopy and

complex impedance measurements.

Experimental

Reagents

The reagents SiCl4 (99.999% - Aldrich); Li3N (99.5% - Cerac); LiNH2 (95% - Alpha

Aesar); NH4Cl (99.999% - Cerac); and SiI4 (99.9% - Cerac) were used as received.

Lump Si (99.9999% - Cerac) was ground in an agate mortar and pestle and passed through a 170 mesh screen.

Synthesis

Li2SiN2 was prepared from SiCl4 by combining SiCl4, Li3N, LiNH2 and NH4Cl in a

helium-filled glove box. Approximately 0.6 ml of SiCl4 was used for each reaction.

In an agate mortar and pestle Li3N, LiNH2 and NH4Cl were ground in the desired ratio and transferred to a 20 ml borosilicate glass vial. The SiCl4 was then added.

In a comparable reaction, SiI4 was used as a silicon source by combining approximately 0.6 g SiI4 with enough Li3N, LiNH2 and NH4Cl to achieve the desired molar ratio. The reagents were ground in a helium-filled glove box and transferred to a 20 ml vial.

Reactions were initiated using microwave energy in a 500 watt Panasonic household microwave oven (2.45 GHz) with an alumina disc as a reaction platform in

place of the standard glass turntable. Caution! These reactions reach high

temperatures rapidly. Care should be taken in the choice of a reaction vessel and

when increasing the reaction scale. The reaction vessel can become extremely hot

and may soften. Microwave energy was applied to the reaction mixture at the 100%

power setting. Reactions initiated after approximately 1-3 minutes of applied energy.

Energy input was stopped upon observation of reaction initiation. Caution!

Continued input of microwave energy can lead to severe discharges in the microwave

cavity. Further heating may also lead to an explosion due to excessive gas pressure

produced by the reaction.

Isolation of Li2SiN2 was achieved by washing the reaction product with water.

Initially, 10 ml of distilled water was added to the reaction vial and sonication was applied for 10 minutes (50 watt, Fisher Scientific Sonicator). The solid product was removed from the liquid by centrifugation. Isolation was completed by rinsing twice

more with water, each rinse was followed by centrifugation. The isolated product

was dried under vacuum.

Characterization

The powder products of the SiX4 reactions (X = Cl, I) were analyzed using a

powder X-ray diffractometer (Scintag XDS-2000) with CuKα radiation (λ = 1.5418

Å). Product samples were scanned from 10 to 100° 2θ at 0.1 degree intervals with a 2

second count time per interval. The patterns obtained were then analyzed and

interpreted using Scintag based software and the Joint Committee on Powder

Diffraction Standards (JC-PDS) database.

Product particle sizes were observed using scanning electron microscopy

(Cambridge Stereoscan 250 SEM).

Crystallite size and lattice strain of Li2SiN2 produced from SiI4 and SiCl4

precursors were determined using the method of integral line breadths.21 In accordance with this method, a plot of the area under each diffraction peak versus its d-spacing gives a linear relationship. The slope of this line contains information on the crystallite size while the y-intercept indicates the lattice strain. Size broadening was assumed to follow a Gaussian profile and strain broadening was modeled as a

Cauchy profile.

Electrical conductivity was determined using complex impedance

spectroscopy. The product of the SiCl4 reaction was pelletized under 5000 pounds of force. These were taken at 500 °C in a frequency range of 0.1 Hz to 100 KHz. The

pellet was sputter coated with gold on both faces and silver epoxy was used to attach

platinum wires for electrical contacts. The epoxy was cured for approximately 8

hours at 150 °C. The entire assembly was then placed in an oven and connected to a

complex impedance spectroscopy system. The system consisted of a power booster

interface (EG&G 263A), a frequency response detector (EG&G Model 1025), a

potentiostat/galvanostat (EG&G Model 273A), and a bipolar operational power

supply/amplifier (Kepco).

Results and Discussion

Several viable pathways to Li2SiN2 were found using microwave initiated reactions. Figure 2-1 shows a typical SSM reaction progressing in the microwave chamber. The image has been enhanced to aid in discerning the reaction zone.

Arrows indicate the multiple reaction sites induced by the microwave energy.

(Caution! The reactions reach significant temperatures and prolonged microwave heating can result in melting of the borosilicate glass vial or catastrophic vessel failure.)

cap

Reaction initiation

glass vial Figure 2–1. A photograph taken through the door of a conventional microwave oven

(2.45 GHz) showing the reaction: 3 SiCl4 + 4 Li3N + 1.5 LiNH2 + 6 NH4Cl in a capped vial after 1 minute at 100% power. The reaction image has been enhanced to emphasize the reaction zone and the outline of the glass vial has been added in white for clarity. Note that the input of microwave energy induces multiple reaction sites as indicated by the arrows.

When either SiI4 or SiCl4 are reacted with Li3N and/or LiNH2 and/or NH4Cl,

the same product, Li2SiN2, is produced (Figure 2-2). However, the molar ratios of

reagents required for each reaction are different. Both SiCl4 or SiI4 reactions utilize the addition of NH4Cl and LiNH2 to assist in the formation of crystalline Li2SiN2.

Ammonium chloride serves to lower the reaction temperature by absorbing energy

through sublimation at 613 K and creates an ammonia atmosphere which helps with

13 nitridation. However, reactions with less than 1.5 moles of LiNH2 result in the formation of the unwanted phase LiSi2N3. The narrow diffraction peaks, in Figure 2-

2, for the Li2SiN2 product suggest good crystallinity. The most crystalline Li2SiN2,

produced from SiCl4, utilizes a SiCl4, Li3N, LiNH2 and NH4Cl molar ratio of 3:4:1.5:6 as indicated by the reaction given in Equation 2-1. With SiI4,

Li3N, LiNH2 and NH4Cl, a molar ratio of 3:4:1.5:2 gives the most crystalline product

(Equation 2-2). Reactions involving SiCl4 produce agglomerates in the size regime of

5 µm (Figure 2-3).

Figure 2–2. Powder X-ray diffraction patterns of the washed Li2SiN2 product formed from the reactions (A) 3 SiI4 + 4 Li3N + 1.5 LiNH2 + 2 NH4Cl; and (B) 3 SiCl4 + 4 Li3N + 1.5 LiNH2 + 6 NH4Cl. A simulated pattern (C) for Li2SiN2 is provided as a reference (JC-PDS 23-0365). The large, broad peak near 15 degrees 2Θ is due to the specimen holder.

1400

1200

1000 A

800

600

Intensity (counts) B

400

200 C

0 10 20 30 40 50 60 70 Angle (2e) Figure 2–3. Scanning electron micrograph at a magnification of 250X of washed

Li2SiN2 powder formed from the reaction of 3 SiCl4 + 4 Li3N + 1.5 LiNH2 + 6 NH4Cl.

6 SiI 4 + 8 Li3N + 3 LiNH2 + 4 NH4Cl ⎯→

6 Li2SiN2 + 13 LiI + 2 LiCl + 3 NH3 + 11 HI + 2 HCl (2-1)

6 SiCl 4 + 8 Li3N + 3 LiNH2 + 12 NH4Cl ⎯→

6 Li2SiN2 + 15 LiCl + 11 NH3 + 21 HCl (2-2)

21 Using the method of integral line breadths, it was found that Li2SiN2

produced from SiI4 had a crystallite size of 680 Å; while the same material produced

from SiCl4 had a crystallite size of 890 Å. The upper limit of the lattice strain (e) was determined to be 0.29% for Li2SiN2 produced from either halide.

The driving force for solid-state metathesis reactions is the formation of a

9 stable salt by-product. For the Li2SiN2 producing reactions, the formation of either

22 LiCl (∆Hf = -408.8 kJ/mol) or LiI (∆Hf = –271.1 kJ/mol) provides the driving force.

The fact that less NH4Cl is required for a reaction producing Li2SiN2 from SiI4 is consistent with their heat of formation since producing LiI gives off 34% less energy

than producing the same amount of LiCl. The formation of Li2SiN2 is known to be

favored at moderately high temperatures.8 However, if the reaction is allowed to get too hot, the product quality suffers. While the temperature can be controlled by

adding NH4Cl, when too much is added then the reaction does not reach the

temperatures required for Li2SiN2 synthesis and little product is produced.

The initiation of metathesis reactions has been shown to occur when a phase

change takes place in one of the reactants.19 This concept is consistent with the

longer microwave initiation times observed for SiI4 reactions (about 3 minutes)

compared to the SiCl4 reactions (about 1.3 minutes) considering the boiling points for

22 SiCl4 (b.p. = 57.6 °C) versus SiI4 (b.p. = 288 °C).

In order to probe the nature of these Li2SiN2 reactions, compressed pellets of elemental silicon and Li3N were heated by microwave energy. These reactions produced both Li2SiN2 and LiSi2N3. When a reaction was run with an Si:Li3N ratio of

3:5, phase pure Li2SiN2 was produced. A Si:Li3N ratio greater than 1:2 produced material that was mostly LiSi2N3. This implies that the amount of Li3N present

strongly affects which lithium silicon nitride phase is synthesized. Although it is

interesting that lithium silicon nitrides can be made from pure silicon, the use of SiCl4

as a precursor may be more attractive because it is an industrial by-product.20

Additionally, since SiCl4 is a liquid at room temperature, it is much easier to purify

than elemental silicon.

When SiCl4 was heated by itself in a microwave, it decomposed to elemental

8 silicon. The traditional synthesis of Li2SiN2 involves the reaction of Si3N4 with Li3N.

This implies that Li2SiN2 is thermodynamically more stable than Si3N4 under the

synthetic conditions used. In fact, high-temperature nitridation of silicon23 is used to

produce Si3N4. This suggests that SiX4 reactions may involve an elemental silicon

intermediate and any Si3N4 produced in the reaction zone is converted to Li2SiN2 in

the presence of Li3N. The following reaction sequence is thus likely to occur under

the application of microwave energy: SiCl4 dissociates to Si and Cl2 (Equation 2-3).

3 SiCl4 ⎯⎯→ 3 Si + 6 Cl2 (2-3)

Li3N reacts with the free Cl2 formed to produce LiCl and N2 (Equation 2-4).

4 Li3N + 6 Cl2 ⎯⎯→ 12 LiCl + 2 N2 (2-4)

The N2 product reacts with Si to form Si3N4 (Equation 2-5).

2 N2 + 3 Si ⎯⎯→ Si3N4 (2-5)

The Si3N4 then very rapidly reacts with excess Li3N to form Li2SiN2 (Equation 2-6).

Si3N4 + 2 Li3N ⎯⎯→ 3 Li2SiN2 (2-6)

Because Li2SiN2 is of interest as a lithium fast-ion conductor, it is important to

measure its conductivity. A sample of Li2SiN2, produced from SiCl4, was compacted

3 to give a pellet with a density of 1.510 g/cm . Since the density of bulk Li2SiN2 is

2.525 g/cm3, the test pellet was 59% of dense.24 In order to measure the ac conductivity of the material, complex impedance spectroscopy was utilized. At 500

˚C, a Nyquist plot of the impedance was generated (Figure 2-4). By using this method it is possible to successfully separate the capacitance and resistance due to intergrain boundaries from that of the bulk material.25 A resistance of 131 ohms was

found at minimum Zim. The conductivity, σ, was calculated using the relationship

given in Equation 2-7:

σ = L/(R*A) (2-7)

where σ is the conductivity in S/cm, L is the thickness of the pellet in cm and R is the

120

100 ) 1 80

40 Imaginary Impedance (

0 50 100 150 200 250 300 350 Real Impedance (1)

Figure 2–4. A Nyquist plot of the impedance for a Li2SiN2 pellet. The measurement was performed at 500 ˚C in a frequency range from 0.1 Hz to 100 kHz. The minimum impedance of 131 ohms corresponds to a conductivity of 2.7 x 10-3 S/cm.

resistance in ohms. This gives a maximum conductivity of 2.70 x 10-3 S/cm at 500

˚C. This conductivity is comparable to a previously reported value of 4.1 x 10-3 S/cm at the same temperature on a 72% compacted sample.7 The slightly lower value

observed here is likely due to the lower degree of compaction.

Conclusions

Solid-state metathesis reactions can provide an attractive alternative to the

traditional synthesis of lithium silicon nitride fast-ion conductors since they utilize an

industrial by-product (SiCl4) and can be initiated easily in a conventional microwave

oven. A metathesis reaction between SiCl4, SiI4, or silicon, with Li3N and LiNH2

acting as nitrogen and lithium sources respectively, produces Li2SiN2 in seconds. The

product phase (Li2SiN2 versus LiSi2N3) can be controlled by the amount of LiNH2 or

Li3N present in the reaction mixture; while the product yield and crystallinity can be

controlled by the addition of NH4Cl. The amount of lithium in the mixed nitride can

be tuned by varying the amount of LiNH2 in a given reaction mixture.

Acknowledgments

The authors thank Kim Nguyen and Matt Gave (NSF Summer Program in

Solid State Chemistry) for help with metathesis reactions, Dr. Jie Guan of GE Power

Systems for assistance with complex impedance measurements, and Marty Kory of

Honeywell Air Frame Systems for assistance with SEM images. This work is supported by the National Science Foundation Division of Materials Research.

References

(1) David, J., Y. Laurent, J.P. Charlot, and J. Lang, Crystallographic study of a I-

IV2-V3 nitride. Wurtzite-type tetrahedral structure of lithium silicon nitride

(LiSi2N3). Bulletin de la Societe Francaise de Mineralogie et de

Cristallographie, 1973. 96(1): p. 21-4.

(2) Guiheen, J.V., Masters Thesis. Materials Science and Engineering. 1995:

University of California, Los Angeles.

(3) Juza, R., H.H. Weber, and E. Meyer-Simon, The ternary nitrides and

oxonitrides of elements of the fourth group. Z. Anorg. Allg. Chem., 1953. 273:

p. 48-64.

(4) Whitney, E.D. and R.F. Giese, Jr., Preparation of a new ternary lithium

silicon nitride, LiSi2N3, and the high-pressure synthesis of magnesium silicon

nitride, MgSiN2. Inorg. Chem., 1971. 10(5): p. 1090-2.

(5) Yamane, H., S. Kikkawa, and M. Koizumi, Preparation and electrochemical

properties of double-metal nitrides containing lithium. J. of Power Sources,

1987. 20(3-4): p. 311-15.

(6) MacFarlane, D.R., J. Huang, and M. Forsyth, Lithium-doped plastic crystal

electrolytes exhibiting fast ion conduction for secondary batteries. Nature

(London), 1999. 402(6763): p. 792-794.

(7) Bhamara, M.S. and D.J. Fray, The electrochemical properties of Li3AlN2 and

Li2SiN2. J. of Mat. Sci., 1995. 30(21): p. 5381-8.

(8) Lang, J. and J.P. Charlot, Li3N-Si3N4 system. Revue de Chimie Minerale,

1970. 7(1): p. 121-31.

(9) Gillan, E.G. and R.B. Kaner, Synthesis of Refractory Ceramics via Rapid

Metathesis Reactions between Solid-State Precursors. Chem. Mater., 1996.

8(2): p. 333-43.

(10) Gillan, E.G. and R.B. Kaner, Rapid, energetic metathesis routes to crystalline

metastable phases of zirconium and hafnium dioxide. J. Mater. Chem., 2001.

11(7): p. 1951-1956.

(11) Parkin, I.P. and A.T. Nartowski, Fast Metathesis Routes to Tungsten and

Molybdenum Carbides. J. Mater. Sci. Lett., 1999. 18: p. 267-268.

(12) Bonneau, P.R., J.B. Wiley, and R.B. Kaner, Metathetical Precursor Route to

Molybdenum Disulfide. Inorganic Synthesis, 1995. 30: p. 33-37.

(13) Cumberland, R.W., R.G. Blair, C.H. Wallace, T.K. Reynolds, and R.B. Kaner,

Thermal Control of Metathesis Reactions Producing GaN and InN. J. of Phys.

Chem. B, 2001. 105(47): p. 11922-11927.

(14) Jarvis, R.F., Jr., R.M. Jacubinas, and R.B. Kaner, Self Propagating Metathesis

Routes to Metastable Group 4 Phosphides. Inorg. Chem., 2000. 39: p. 3243-

3246.

(15) O'Loughlin, J.L., C.H. Wallace, M.S. Knox, and R.B. Kaner, Rapid Solid-

State Synthesis of Tantalum, Chromium, and Molybdenum Nitrides. Inorg.

Chem., 2001. 40(10): p. 2240-2245.

(16) Bonneau, P.R., R.F. Jarvis, Jr., and R.B. Kaner, Rapid Solid-State Synthesis of

Materials from Molybdenum Disulfide to Refractories. Nature (London),

1991. 349: p. 510-512.

(17) Rao, L., E.G. Gillan, and R.B. Kaner, Rapid Synthesis of Transition Metal

Borides by Solid-State Metathesis. J. Mater. Res., 1995. 10: p. 353-361.

(18) Houmes, J.D. and H.-C. zur Loye, Microwave synthesis of ternary nitride

materials. J. Solid State Chem., 1997. 130(2): p. 266-271.

(19) Blair, R.G., E.G. Gillan, N.B.K. Nguyen, D. Daurio, and R.B. Kaner, Rapid

Solid-State Synthesis of Titanium Aluminides. Chem. Mater., 2003. 15(17): p.

3286-3293.

(20) Morgan, P.E.D. and E.A. Pugar, Synthesis of silicon nitrdies (Si3N4) with

emphasis on silicon-sulfur-nitrogen chemistry. J. Am. Ceram. Soc., 1985. 68:

p. 699-703.

(21) Klug, H.P. and L.E. Alexander, X-Ray Diffraction Procedures for

Polycrystalline and Amorphous Materials. 2nd ed. 1974, New York: John

Wiley & Sons.

(22) Standard Thermodynamic Properties of Chemical Substances, in CRC

Handbook of Chemistry and Physics, Internet Version 2005, D.R. Lide,

Editor. 2005, CRC Press: Boca Raton, FL.

(23) Greenwood, N.N. and A. Earnshaw, Chemistry of the Elements. 1984, New

York: Pergamon Press.

(24) Bhamara, M.S., Ph.D. Dissertation. Department of Materials Science and

Metallurgy. 1994: University of Cambridge.

(25) Irvine, J.T.S., D.C. Sinclair, and A.R. West, Electroceramics:

Characterization by Impedance Spectroscopy. Adv. Mater., 1990. 2(3): p.

132-138.

Chapter 3: Synthesis and Crystal Structure

of Cubic Ca16Si17N34

Summary

Since the late 1960s, the exact structure of cubic calcium silicon nitride has

been a source of debate. This chapter offers evidence that the cubic phase CaSiN2

described in the literature is actually Ca16Si17N34. Presented here is a method for synthesizing single crystals of cubic-calcium silicon nitride from calcium nitride and elemental silicon under flowing nitrogen at 1500 °C. The colorless millimeter-sized

crystals of Ca16Si17N34 with a refractive index (n25) = 1.620 were found to be cubic (a

= 14.8882 Å) and belong to the space group F-43m. The synthesis of bulk, powdered

cubic-Ca16Si17N34 from calcium cyanamide and silicon is also discussed. Ca16Si17N34

is a relatively air-stable refractory ceramic. In contrast to the orthorhombic phase of

2+ 2+ CaSiN2, in which Ca sits in octahedral sites, this cubic phase has Ca in cubic sites

that makes it an interesting host for new phosphors and gives rise to unique crystal

field splitting.

Introduction

Nitrides and oxynitrides have attracted great interest due to their potential as

host lattices for fluorescent species.1-3 The structure and crystal field splitting

Reproduced with permission from Hick, S.M., M.I. Miller, R.B. Kaner, and R.G.

accessible in ternary silicon nitrides offers the promise of new colors from emitting

species.4-8 Unfortunately, nitrides are notoriously difficult to crystallize. A variety of

methods 9 have been attempted to obtain single crystals of nitride materials. These include flux growth 10, 11, vapor transport 12, 13, growth under pressure 14, and melt growth 15. However, ternary silicon nitrides do not readily lend themselves to any of

these methods since they are not soluble in most of the fluxes used for crystal growth,

do not crystallize by vapor transport, and decompose before melting.

Two different phases of calcium silicon nitride (CaSiN2) have been reported in

the literature. Early work indicated a cubic unit cell with an a parameter of 14.8642

Å 16-18. However, this structure has been disputed and it has been suggested that the

cubic phase is, in fact, an oxynitride. Work by Ottinger et al. produced an orange,

orthorhombic phase (Pbca) with a = 5.129(3) Å, b = 10.224(1) Å, and c = 14.821(4)

19 Å by heating Ca, CaSi2 and Ca3N2 in a sealed niobium tube at 1400 ˚C . The same

structure was later obtained by heating Ca, NaN3 and Si in a sealed tantalum tube at

1000 ˚C 20.

Until now, single crystals of the initially reported cubic phase have not yet

been synthesized. Previously, the cubic phase was synthesized by the reaction

(Equation 3-1):

Si3N4 + Ca3N2 → 3CaSiN2 (3-1)

We have been able to obtain near-millimeter sized crystals of a calcium deficient

nitride (Ca16Si17N34) using a reaction (Equation 3-2) analogous to that used to prepare

21 a germanium anti-perovskite lattice with the formula Ca3GeN .

51Si + 16Ca3N2 + N2 → 3 Ca16Si17N34 (3-2)

The structure of this compound exhibits similar characteristics to other nitridosilicates

22.

Experimental

Reagents

The following reagents were used as received: Ca3N2 (Cerac; 99%, 200 mesh) and CaCN2 (Alfa Aesar, technical grade). Silicon powder was prepared from virgin polyfine Si (99.999% Alfa Aesar), by grinding under argon for 30 minutes in a SPEX mixer mill using a tungsten carbide vial and 3 0.5” tungsten carbide balls and then sieving to -270 mesh in an argon-filled glovebox. The reagents were stored in an argon-filled glovebox.

Single Crystal Synthesis

Single crystals of Ca16Si17N34 were synthesized by preparing a reaction mixture consisting of Si and Ca3N2 in a molar ratio of 1:1. The reagents were ground

together in a synthetic sapphire mortar and pestle; the powder was transferred to a 12

mm diameter die and pressed into a pellet under a load of 10,000 lbs while in the

glovebox. The pellet was placed in a hexagonal boron nitride (binderless grade AX05

Saint-Gobain Ceramic Materials) boat or powdered graphite-lined graphite boat. The boat was loaded into a 2” diameter alumina tube in a high temperature furnace utilizing silicon carbide heating elements. Initial syntheses were performed by removing the boat from the glove box in a plastic bag and quickly transferring to the

furnace tube in air. Later syntheses were loaded through a specially constructed

glove box antechamber directly connected to the furnace tube. Using this method the

reactant mass was never in contact with air. The system was sealed and a flow of

high purity nitrogen (ultra-high purity nitrogen further purified using a Cr2+ column)

gas was maintained. The system was heated to 1500 ˚C at a heating rate of 100 ˚C/h

and held at temperature for two hours then cooled to room temperature at a rate of

100 ˚C/h. Upon removal from the furnace, the product was washed with distilled

water and characterized.

Powder Synthesis

A slight excess of calcium cyanamide, CaCN2 (5 grams, 62 mmol), was mixed with silicon powder (1.578 g, 56 mmol). The reagents were ground together in an artificial sapphire mortar and pestle; the powder was transferred to a 12 mm diameter die and pressed into a pellet under a load of 10,000 lbs while under inert atmosphere.

The pellet was placed in a powdered graphite-lined graphite boat. Initial syntheses were performed by removing the boat from the glove box in a plastic bag and quickly transferring to the furnace tube in air. Later syntheses were loaded through a specially constructed glove box antechamber directly connected to the furnace tube.

Using this method the reactant mass was never in contact with air. The system was sealed and a flow of high purity nitrogen (ultra-high purity nitrogen further purified using a Cr2+ column) gas was maintained. The system was heated to 1500 ˚C at a

heating rate of 100 ˚C/h and held at temperature for two hours then cooled to room

temperature at a rate of 100 ˚C/h. Upon removal from the furnace, the product was

washed with distilled water and characterized.

Refractive Index

The refractive index of the crystalline material was determined using an

optical microscope and Cargille Refractive Index Liquids.

Characterization

A cubic crystal of approximately 0.2 mm in each dimension was obtained

from the synthesis using Ca3N2 and mounted on a glass fiber for X-ray analysis.

Single crystal X-ray intensity data were measured at 293 K on a Bruker SMART

1000 CCD-based X-ray diffractometer equipped with a Mo-target X-ray tube (λ =

0.71703 Å) operated at 2.25 kW. The detector was placed at a distance of 4.986 cm from the crystal.

The structure was refined using Bruker SHELXTL software (version 5.3) package23. The final anisotropic, full-matrix, least-squares refinement on F2 converged at R1 = 3.97%. The final cell constants are listed in Table 3.2 together with other relevant crystallographic data.

The powders were characterized by X-ray diffraction (XRD) using a PANalytical

X’pert Pro powder X-ray diffractometer with a copper source (Cu Kα λ = 1.5418 Å).

Spectra were collected from 5 to 100 degrees two theta using 0.03 degree intervals

and a 1 second dwell time. The patterns obtained were then analyzed and interpreted

using the PANalytical X’pert Highscore Plus software and the Joint Committee on

Powder Diffraction Standards (JCPDS) database.

Results

Colorless, transparent crystals of Ca16Si17N34 (Figure 3-1) with a refractive

index n25 of 1.620 were synthesized by the reaction of Si with Ca3N2 in a 1:1 ratio

under flowing nitrogen in a graphite boat. The product was rinsed with water to

remove any water reactive by-products, such as Ca2Si, and dried for 1 hour at 180 ˚C.

Upon washing, the distinct odor of acetylene was noted, indicating that some of the

excess Ca3N2 reacted with the graphite boat. Crystals were isolated by density using a

mixture of CH2I2 (density = 3.3254 g/ml) and CHBr3 (density = 2.8859 g/ml) and/or physical separation. Figure 3-2 illustrates the material separated by a 170 (90 µm opening) mesh screen.

Figure 3–1. Optical microscope images of the transparent Ca16Si17N34 crystals produced in this study. The left image is under 50x magnification and the right is under 500x.

Figure 3–2. Scanning electron microscope images of Ca16Si17N34 separated from the reaction mass using a 170-mesh screen. This nitride exhibits cubic crystals often in the form of cuboctahedra (right). The role of the boat material, oxygen level, and calcium source were examined by

using an inert boat (binderless boron nitride), adding oxygen in the form of CaO, and

substituting CaCN2 for Ca3N2 (Table 3-1). In most reactions, the Ca16Si17N34 phase

formed. The reaction 17CaCN2 + 17Si + CaO produced a crystalline material that

could not be matched to a known compound in the JCPDS file.

Table 3–1. Ca16Si17N34, forming reactions. The reaction marked with a (*) produced the most crystalline product.

Reaction Boat Material Phases Present

Ca16Si17N34 + o- Si + Ca3N2 boron nitride CaSiN2

*Si + Ca3N2 graphite Ca16Si17N34

CaCN2 + Si graphite Ca16Si17N34 Ca16Si17N34 + 6 CaCN2 +12 Si + Ca3N2 graphite Ca3N2Si2O4

17 CaCN2 + 17 Si +CaO graphite Unknown phase

102 CaCN2 + 204 Si +17 Ca3N2 + 6 CaO graphite Ca16Si17N34 and CaO

17 Si +17 Ca3N2 + CaO graphite Ca16Si17N34 and CaO

The product generally consists of crystals that exhibit octahedral and

cuboctahedral morphologies (Figures 3-1 and 3-2). Single crystals large enough for

X-ray analysis were produced by this method. The unit cell was determined to be cubic (a = 14.8882 Å) with the space group F-43m (216). The atomic positions are summarized in Table 3-3. Figure 3-3 shows the difference in connectivity between

the cubic phase and orthorhombic CaSiN2.

Table 3–2. Crystal data and structural refinement parameters for cubic Ca16Si17N34.

empirical formula Ca16Si17N34 radiation MoKα, λ = 0.71073 Å instrument Bruker SMART 1000 temperature 293(2) K space group F-43m (216) fw 93.79 a/Å 14.8882(3) V/Å3 3300.10(12) Z 48 -3 -3 ρcalc (Mg m ) 3.211 g cm crystal size 0.20 x 0.20 x 0.18 mm independent reflections 453 [R(int) = 0.0224] R1 0.0397 wR2 0.1078

Table 3–3. Atomic coordinates and equivalent isotropic displacement parameters 3 (x10 ) for cubic Ca16Si17N34. U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

site x y z U(eq) Ca(1) 24f ½ 0.7920(1) 0 9(1) Ca(2) 24g 0.5009(2) ¾ ¾ 10(1) Ca(3) 16e 0.5799(2) 0.4201(2) 0.9201(2) 39(1) Si(1) 4b ½ 0 0 5(1) Si(2) 16e 0.3657(1) 0.8657(1) 0.8657(1) 6(1) Si(3) 16e 0.3194(2) 0.8194(2) 0.6806(2) 19(1) Si(4) 16e 0.6372(1) 0.6372(1) 0.8628(1) 8(1) Si(5) 16e 0.6769(1) 0.6769(1) 0.6769(1) 9(1) N(1) 16e 0.4319(3) 0.9319(3) 0.9319(3) 6(2) N(2) 48h 0.3861(3) 0.8861(3) 0.7532(4) 9(1) N(3) 4c ¼ ¾ ¾ 4(4) N(4) 4d ¾ ¾ ¾ 10(5) N(5) 48h 0.6146(3) 0.6146(3) 0.7521(4) 9(1) N(6) 16e 0.5733(4) 0.5733(4) 0.9267(4) 10(2)

Figure 3–3. The unit cell for cubic Ca16Si17N34 (left) consists of clusters of SiN4 units (gray tetrahedra), while that of orthorhombic CaSiN2 (right) consists of SiN4 units in rings. The calcium atoms have been removed to highlight the connectivity of the

SiN4 units.

Discussion

The structure of the Ca16Si17N34 unit cell is better understood by viewing it without the calcium atoms (Figure 3-3 left). The unit cell consists of 8 clusters of SiN4 tetrahedra arranged around a central SiN4 tetrahedron. Each cluster, in turn, consists of 8 SiN4

tetrahedra. There are three distinct Ca sites (Figure 3-4). Ca1 and Ca3 are distorted

octahedral sites, while Ca2 is an elongated cubic site. This structure has been

observed previously in other calcium nitridosilicates 22. Le Toquin and Cheetham’s

cerium substituted structure4 gave a red luminescence which would be consistent with

24 a spectral shift associated with cerium in a cubic site . The open framework of SiN4

tetrahedra may also help to explain the unusually low thermal conductivity measured

-1 -1 -1 -1 16 for the cubic phase (2.4 Wm K versus 17 Wm K for MgSiN2) as well as the low

refractive index for a nitride (n25=1.620). This structure is markedly different from

orthorhombic CaSiN2 (Figure 3-3 right) that has a structure analogous to other alkaline earth silicon nitrides like MgSiN2. Previous attempts to synthesize CaSiN2

have produced the orthorhombic phase. These syntheses were carried out at 1400 ˚C

19 and 1000 ˚C 20.

Figure 3–4. The Ca1 and Ca3 positions are 6-coordinate with the Ca1 position in a severely distorted octahedron. The Ca2 position is 8-coordinate in an elongated cubic site. The simulated powder pattern matches the experimental data well (Figure 3-5)

and matches the previously deleted JC-PDS pattern for CaSiN2 (40-1151).

Figure 3–5. The powder X-ray diffraction pattern simulated from the structural data (top) matches the measured pattern (bottom) and the deleted JC-PDS file 40-1151 (middle).

This pattern was replaced by CaSi(O,N)2 (45-1215) and illustrates a particular problem when dealing with ternary nitrides – determination of the oxygen content.

Every effort was made in this work to ensure low levels of oxygen contamination.

21 However, commercially obtainable Ca3N2 can contain oxide impurities . If site occupancies of 100% are assumed, then the unit cell, which has 268 atoms, will have

a charge imbalance of -8. This could be remedied by 1.9 at% nitrogen vacancies or

random substitution of 8 of the 136 nitrogens with oxygen, which would represent an

oxygen level in the lattice of 2 mass%. These values are difficult to determine with

the current methods available, especially in the highly symmetric space group F-43m.

The Si-N bond lengths range from 1.653 Å to 1.885 Å. Oxygen substitution should

result in reduced bond lengths suggesting that oxygen replacement is most likely on

the N6 site. Substitution of half the N6 sites with oxygen would produce a charge-

balanced lattice. This would also explain the high U(eq) for Ca3, which sits in a

distorted octahedral site with vertices occupied by N3 and N6 nitrogens. The Ca1

and Ca2 sites do not directly interact with the N6 site. Alternatively, 12.5 at% 3+ ion

substitution (such as in Le Toquin and Cheetham’s Ce3+-doped nitride)4 for Ca would

also provide unit cell neutrality.

18 Previous efforts to prepare CaSiN2 have used oxide precursors , nitride precursors16, 17, or silicides19. Although Gal, et al. utilized silicon, their syntheses

were performed at a much lower temperature (1000 ˚C). We have been able to

obtain single crystals of the cubic ternary nitride, Ca16Si17N34, by heating a mixture of

Ca3N2 and Si to 1500˚C for a relatively short period of time (2 hours). A ratio of 1.0

to 1.0 was found to produce the largest crystals. This may be due to the fact that

some of the Ca3N2 evaporates from the reaction zone and that molten Ca3N2 acts as a flux.

The reactions summarized in Table 3-1 show that the nitride phase forms with or without carbon, and from calcium cyanamide as well. More crystalline products

were formed from reactions utilizing Ca3N2, suggesting that the nitride acts as a flux.

Calcium nitride melts at 1195 ˚C and silicon melts at 1410 ˚C. At 1500 ˚C, calcium nitride is molten and evaporation is a major concern.16 Performing the reaction in a

nitrogen atmosphere and using an excess of Ca3N2 reduces the impact of Ca3N2

evaporation. Some of the calcium present reacted with the graphite in the boat to

produce CaC2; this was confirmed by the production of acetylene upon washing the

reaction product with water. The reaction of CaC2 with nitrogen forms calcium

cyanamide, CaCN2. Calcium cyanamide has a melting point of 1340 ˚C. A calcium

nitride/calcium cyanamide mixture may be the flux in which the CaSiN2 crystals

actually grow. In fact, Ca16Si17N34 powders can be realized by heating a mixture of

CaCN2 and Si in stoichiometric quantities (Figure 3-6).

Figure 3–6. The powder X-ray diffractogram of the product produced from the

reaction 16CaCN2 + 17Si + N2→ Ca16Si17N34 + 16C shows that the sample contains silicon (*) and graphite (o) impurities with the remaining peaks matching the pattern expected for Ca16Si17N34.

These nitrides are not indefinitely stable in air. Samples kept in air for longer than 3 months lose much of their crystallinity.

Conclusions

Single crystals of cubic Ca16Si17N34 were readily grown from a mixture of

Ca3N2 and Si heated in a furnace to 1500 ˚C. The lattice consists of an open framework of SiN4 tetrahedra with Ca ions in the voids; the structure has some uncertainty due to potential oxygen substitutions. Powders of cubic Ca16Si17N34 were

synthesized from calcium cyanamide and silicon. The compound’s most highly

oxidized formula may actually be Ca16Si17N32O2, but current methods cannot readily determine this subtle compositional discrepancy. One of the calcium sites in this nitride is cubic. Partial substitution with other metals could produce interesting new phosphors.

Acknowledgements

The authors would like to thank Dr. Saeed Khan at the UCLA Molecular

Instrumentation Center for gathering the X-ray data and solving the crystal structure of the nitride. The authors would also like to thank Saint-Gobain Ceramic Materials for the generous donation of binderless grade AX05 boron nitride.

Supporting Information

Crystal structure file. This material is available free of charge via the Internet at http://pubs.acs.org.

References

(1) Xie, R.-J., N. Hirosaki, Y. Li, and T. Takeda, Rare-earth activated nitride

phosphors: synthesis, luminescence and applications. Materials, 2010.

Reserved.): p. 3777-3793.

(2) He, X.-H., N. Lian, J.-H. Sun, and M.-Y. Guan, Dependence of luminescence

properties on composition of rare-earth activated (oxy)nitrides phosphors for

(3) Li, Y.Q. and H.T. Hintzen, Luminescence properties of novel rare-earth

doped silicon nitride based phosphors. Oyo Butsuri, 2007. 76(Copyright (C)

(4) Le Toquin, R. and A.K. Cheetham, Red-emitting cerium-based phosphor

materials for solid-state lighting applications. Chemical Physics Letters,

2006. 423(4-6): p. 352-356.

(5) Uheda, K., N. Hirosaki, Y. Yamamoto, A. Naito, T. Nakajima, and H.

Yamamoto, Luminescence properties of a red phosphor, CaAlSiN3 : Eu2+,

for white light-emitting diodes. Electrochemical and Solid State Letters, 2006.

9(4): p. H22-H25.

(6) Uheda, K., N. Hirosaki, and H. Yamamoto, Host lattice materials in the

system Ca3N2-AlN-Si3N4 for white light emitting diode. Physica Status Solidi a-

Applications and Materials Science, 2006. 203(11): p. 2712-2717.

(7) Li, Y.Q., N. Hirosaki, R.-J. Xie, T. Takada, Y. Yamamoto, M. Mitomo, and

K. Shioi, Synthesis, crystal and local electronic structures, and

2+ photoluminescence properties of red-emitting CaAlzSiN2+z:Eu with

orthorhombic structure. Int. J. Appl. Ceram. Technol., 2010. 7(Copyright (C)

(8) Chiu, Y.-C., C.-H. Huang, T.-J. Lee, W.-R. Liu, Y.-T. Yeh, S.-M. Jang, and

2+ R.-S. Liu, Eu -activated silicon-oxynitride Ca3Si2O4N2: a green-emitting

phosphor for white LEDs. Opt. Express, 2011. 19(S3): p. A331-A339.

(9) Ambacher, O., Growth and applications of group III nitrides. Journal of

Physics D: Applied Physics, 1998. 31(20): p. 2653-2710.

(10) Yano, M., M. Okamoto, Y.K. Yap, M. Yoshimura, Y. Mori, and T. Sasaki,

Growth of nitride crystals, BN, AlN and GaN by using a Na flux. Diamond

and Related Materials, 2000. 9(3-6): p. 512-515.

(11) Iwahashi, T., F. Kawamura, M. Morishita, Y. Kai, M. Yoshimura, Y. Mori,

and T. Sasaki, Effects of ammonia gas on threshold pressure and seed growth

for bulk GaN single crystals by Na flux method. Journal of Crystal Growth,

2003. 253(1-4): p. 1-5.

(12) Liu, B., D. Zhuang, and J.H. Edgar, AlN grown by sublimation, in Vacuum

Science and Technology: Nitrides as Seen by the Technology. 2002. p. 43-58.

(13) Singh, N.B., A. Berghmans, H. Zhang, T. Wait, R.C. Clarke, J. Zingaro, and

J.C. Golombeck, Physical vapor transport growth of large AlN crystals.

Journal of Crystal Growth, 2003. 250(1-2): p. 107-112.

(14) Porowski, S., I. Grzegory, and A. Morawski, Crystal growth of

semiconductors and superconductors under high gas pressure. Metals,

Materials and Processes, 1992. 3(4): p. 277-83.

(15) Otani, S., Crystal growth at high temperature. Oyo Butsuri, 1986. 55(4): p.

373-6.

(16) Groen, W.A., M.J. Kraan, and G. Dewith, New Ternary Nitride Ceramics -

CaSiN2. Journal of Materials Science, 1994. 29(12): p. 3161-3166.

(17) Laurent, Y., Calcium-Silicon-Nitrogen Systems. Rev. Chim. Miner., 1968.

5(5): p. 1019-50.

(18) Pugar, E.A., J.H. Kennedy, P.E.D. Morgan, and J.R. Porter, New Cubic

Structure in the Ca-Si-O-N System. Journal of the American Ceramic Society,

1988. 71(6): p. C288-C291.

(19) Ottinger, F. and R. Nesper, Synthesis and structure of CaSiN2. Acta Cryst.,

2002. A58: p. C337.

(20) Gal, Z.A., P.M. Mallinson, H.J. Orchard, and S.J. Clarke, Synthesis and

structure of alkaline earth silicon nitrides: BaSiN2, SrSiN2, and CaSiN2.

Inorganic Chemistry, 2004. 43(13): p. 3998-4006.

(21) Chern, M.Y., D.A. Vennos, and F.J. DiSalvo, Synthesis, structure, and

properties of anti-perovskite nitrides Ca3MN, M = phosphorus, arsenic,

antimony, bismuth, germanium, tin and lead. Journal of Solid State

Chemistry, 1992. 96(2): p. 415-25.

(22) Schlieper, T. and W. Schnick, Nitrido silicates. I. High temperature synthesis

and crystal structure of Ca2Si5N8. Zeitschrift fuer Anorganische und

Allgemeine Chemie, 1995. 621(6): p. 1037-41.

(23) Bruker Analytical X-ray Systems, M., WI, 1997.

(24) Blasse, G. and A. Bril, Investigation of Some Ce3+‐Activated Phosphors.

Journal of Chemical Physics, 1967. 47: p. 5139-5145.

Chapter 4: Synthesis of Si3N4 via a

Polysilazane Intermediate

Introduction

Silicon nitride is an important high-temperature, refractory ceramic that is prized for its corrosion resistance and low relative density. It is widely used in bearings, cutting tools, molten metal sheaths, gas nozzles and welding apparatus.

Some applications, such as bearings, require a balance of α and β silicon nitride in

order to take advantage of the hardness of one phase (α) and the toughness of another

1 (β). Despite its high cost, Si3N4 is often used because it exhibits better thermal shock resistance than other, more easily synthesized ceramics.

Silicon nitride was first synthesized in the late 1850s2. In 1879, the reaction

product was analyzed to show three silicons for every 4 nitrogens.3 Around the turn of the 20th century, numerous patents for synthetic routes to silicon nitride were granted.4 By the 1950s, it began being used as a refractory material in various

applications including a sheath for high-temperature thermocouples and gas-turbine

engines. Naturally occurring α-silicon nitride, nierite, was found in particles of

meteoritic rock.5 The uses of silicon nitride have continued to develop and include

amorphous films, some of which can be formulated through mechanical means such

as ball-milling,6 used in the electronics field as passivation layers and insulators.

Recently, Si3N4 has been investigated for use in medical devices, specifically

replacement joints and spinal implants.7-9

While a number of factors have been responsible for the difficulty fabricating a turbine or gas engine entirely out of silicon nitride,10 one of the obstacles has been the cost of silicon nitride powder. At $400-3000 per kilogram, applications have quite often been limited to applications where financial considerations are outweighed by technical considerations. As the uses for silicon nitride and production volumes increase, the price should decrease.

A number of synthetic routes to silicon nitride have been developed over the years. One of the earliest approaches, the carbothermal process11 is still being used

with a few modifications. In this synthesis, carbon is reacted with silicon dioxide and

a nitrogen source, such as ammonia, at 1400-1500 °C. Unfortunately, this reaction

tends to also produce SiC if there is any free silicon present in the reaction mixture,

so it is imperative to avoid this if pure silicon nitride is desired.12 The process produces mostly α-silicon nitride (~95%) and involves an amorphous Si-O-C intermediate; for any β-silicon nitride formed, an Si-O-N intermediate is involved.13

In general, this process tends to produce a product with significant carbon

contamination.

Another method for silicon nitride production is the direct nitridation of

silicon. This reaction is very exothermic and occurs between 1300-1400 °C. This method requires nitriding times of 20 to 100 hours and requires a porous silicon

substrate which, unfortunately, degrades the strength of the material produced.14 One reason for the long reaction times is the fact that silicon tends to form protective nitride layers. The process often requires the use of iron or halide catalysts.15

A third method of synthesizing silicon nitride is the diimide process, so named

because an intermediate diimide is formed which is then annealed to produce silicon

nitride according to Equations 4-1 and 4-2.

0 ˚C (4-1) SiCl4 + 6 NH3 Si(NH)2 + 4 NH4Cl

1400 - 1500 ˚C 3 Si(NH)2 Si3N4 + N2 + 3 H2 (4-2)

This method results in an agglomerated powder product that must be milled and has chloride contamination.

Another method for silicon nitride production is a vapor phase reaction between silicon tetrachloride and ammonia at elevated temperatures (Equation 4-3 below). The use of silicon tetrachloride as a source of silicon in many of these processes is due in part to its availability as a by-product of several industrial refining processes, in particular the extraction of Ti and Zr from their ores.15 1400 ˚C 3 SiCl4 + 4 NH3 Si3N4 + 12 HCl (4-3)

As with the other methods that use a halide precursor, chloride retention is a problem.

Additionally, the product must be milled to reduce agglomeration. Any time a milling step is required it introduces the likelihood of adding additional contaminants, from the grinding media, to the product.

Several other production methods have been devised. A reaction involving

sulfur has been shown to produce silicon nitride (see Equation 4-4).15 900 ˚C (4-4) 3 SiS2 + 4 NH3 Si3N4 + 6 H2S This reaction requires the use of extremely pure silicon sulfide; commercially available grades may not be acceptably pure and the laboratory synthesis is quite challenging.15 A low-temperature process involving magnesium silicide is also possible but involves the use of an autoclave and relatively high pressures (30-40

MPa).16 Equation 4-5 illustrates this reaction that produces predominantly α-silicon nitride.

3 MgSi2 + 12 NH4Cl !-Si3N4 + 6 MgCl2 + 8 NH3 + 12 H2 (4-5)

A method for producing β-silicon nitride has also been developed (see

Equation 4-6).17 1300-1500 ˚C 3 SiO + 4 NH !-Si N + 6 H O 2 3 3 4 2 (4-6)

Additionally, the decomposition of magnesium silicon nitride at elevated temperatures can produce β-silicon nitride (see Equations 4-7 and 4-8).17 >1400 ˚C (4-7) 3 MgSiN2 !-Si3N4 + Mg3N2

3 SiO2 + Mg 3N2 + 4 NH3 3 MgSiN2 + 6 H2O (4-8)

While there are numerous methods of preparing silicon nitride, the processes are time, temperature and/or pressure intensive. This leads, in part, to the relatively high cost of pure grade silicon nitride.

After the powder is produced, there is the still the question of how to fabricate the finished product. The direct nitridation of silicon allows for compacting the

silicon substrate into the shape of the final product and then nitriding it. However, as

mentioned earlier, this is not without drawbacks such as porous product whose

hardness has been compromised.14

There have been attempts at modifying the silicon nitride powders to aid in

fabrication of the final product. A process involving spark plasma sintering of silicon

nitride with titanium nitride has resulted in a composite of relatively low resistivity

(1x10-4 ohm-cm) which can be machined using Electrical Discharge Machining

(EDM).18 Another method that converts brittle silicon nitride into a ductile superplastic form by heating it to temperatures above 1600 °C and using strain rates

<10-5/sec. This helps significantly with net shape fabrication.19

For the most part, however, the fabrication of silicon nitride powder into its final shape involves mixing α-silicon nitride with a sintering aid and annealing the composite to form a new β-silicon nitride composite. Typical sintering aids are

oxides, such as Y2O3 or Nd2O3, that react with oxides on the surface of the powder and also a slightly with the underlying silicon nitride to form an oxynitride liquid phase. After cooling, a glassy or crystalline phase is formed at the grain boundaries.20

Ostwald ripening is used to control these intergranular phases in order to form tough, interlocking microstructures.21

Some processes are designed to produce silicon nitride in a form that lends it to its ultimate use. Formation of α-silicon nitride whiskers has been accomplished by a reaction involving silicon nitride, silicon and sodium azide; this morphology is used to reinforce metals.22 Additionally, silicon nitride rods have been prepared using

carbon nanotubes as templates.23 Another process synthesizes pure β-silicon nitride

rods without a template by reacting silicon tetrachloride, sodium amide and

ammonium chloride.24

While there are many methods of producing silicon nitride and forming the final products, there still remains the issue of producing a silicon nitride in a manner that will significantly reduce its cost. Presented here is a method has the potential of producing high quality silicon nitride at an affordable price by utilizing an amide as the nitrogen source.

Experimental

Chemicals

The reagents sodium amide (Alfa-Aesar, 96%), silicon (Alfa-Aesar, 325 mesh, 99.5%), potassium (Aldrich, 95.5%), and ammonia (Matheson Gas Products) were used as received.

Method

Elemental silicon was reacted with molten sodium amide by using a 1:5 mass

ratio of Si to NaNH2. Silicon and NaNH2 were loaded into an iron crucible in a He-

filled glove box (see Figure 4-1). The crucible was heated to 320 °C to melt the

NaNH2. The crucible was held at temperature until the melt solidified, typically 1

hour. (A similar reaction was run using KNH2 in lieu of NaNH2). A portion of the

product was washed with liquid ammonia to remove sodium (or potassium) metal that

was a reaction by-product (see Figure 4-2). This washed product was heated to

400 °C for 1 to 2 hours in an iron crucible.

The materials were converted to Si3N4 by compressing 0.1 to 0.3 grams of precursor in a ¼” die using a press in a helium-filled glove box. These pellets were placed in a graphite boat that was put inside a 1” diameter alumina tube. Titanium powder was placed around the boat to act as an oxygen scavenger. The ends of the tube were covered in Parafilm and the whole assembly was then removed from the glove box.

The alumina tube was transferred to a glove bag attached to the end of a 2” alumina tube that was continuously purged with ultra-dry, ultra-pure helium. This helium was prepared by passing helium gas through a Cr2+ scavenger column, a molecular sieve condenser in liquid nitrogen, and finally a Mn2+ scavenger column

Figure 4–1. The furnace used to melt the NaNH2 and subsequently react it with silicon. The assembly is in a glove box under an inert atmosphere of helium. (see Figure 4-3).25 In the glove bag, the Parafilm was removed from the ends of the

tube and the alumina tube was inserted into the center of a high temperature furnace.

The ends were closed and ultra-dry, ultra-pure helium was passed over the sample at

a rate of 1 liter/hr. Figure 4-3 illustrates the furnace set-up. The system was heated

to 1500 °C at a heating rate of 100 °C/hr and held at temperature for 2 hours. The

system was then cooled to room temperature at 100 °C/hr.

Figure 4–2. The assembly used to wash the product from the reaction of silicon with

NaNH2. The blue color in the liquid ammonia is due to solvated electrons.

Figure 4–3. The furnace assembly for the annealing of the intermediate product from

the reaction of silicon with NaNH2 under a stream of ultrapure helium. Any traces of water are removed from the helium stream by flowing the gas through the Cr2+ scavenging column, a cryo-trap filled with 4 Å molecular sieves, and finally a Mn2+ scavenging column. The sample is held in a graphite boat loaded into an alumina tube while under helium. This tube is transferred to the furnace via a glove bag placed around the outlet.

Characterization

The products of the reactions were characterized by powder X-ray diffraction

(XRD), Fourier Transform Infrared Analysis (FTIR), and Solid State Nuclear

Magnetic Resonance (SS NMR) Spectroscopy.

The powder X-ray diffraction was collected using a Panalytical X'Pert Pro X-

ray Powder Diffractometer with a copper source (Cu K∝ λ = 1.5418). Spectra were collected from 0 to 100 degrees two theta using 0.01 degree steps and 1 s dwell time.

The infrared spectrum was collected using a Nicolet Nexus 470 FTIR. A pellet was formed by mixing a sample of the intermediate with KBr powder (stored in a 100 °C oven).

Solid State NMR spectra were collected on a Bruker DSX300 broad band FT

NMR Spectrometer for Solid-State samples with magic angle spinning.

Results

Figure 4-4 shows the results of the XRD pattern for the final product of the

reaction between silicon and NaNH2 after ammonia treatment and subsequent annealing under ultra-pure helium at 1500 °C. The product is primarily α -silicon nitride.

Figure 4-5 is an FTIR spectrum of the intermediate from the reaction between

-1 -1 silicon and NaNH2. FTIR (KBr; cm ): 3183 cm absorption peak corresponding to

N-H stretching, a peak at 1935 cm-1 from the Si-H stretch, and 923 cm-1 peak arising

from the Si-N stretch.

Figure 4–4. The bottom XRD pattern is the result of a reaction between Si and

NaNH2 and subsequent washing in liquid NH3 followed by annealing under dry He atmosphere at 1500 ˚C. The top XRD pattern is a simulated pattern for alpha Si3N4 (JCPDS 41-0360). The results of the solid state MAS NMR for the reaction between silicon and

29 KNH2 can be seen in Figure 4-6. Figure 4-6 shows the NMR pattern for Si with and without 1H coupling.

Figure 4–5. The FTIR spectrum for the intermediate formed by the reaction of silicon with NaNH2. The Si-H stretching band along with the Si-N and N-H stretches provide evidence for a possible polysilazane polymer intermediate.

Figure 4–6. Solid-state MAS-NMR spectra of the intermediate formed by the 29 reaction between Si and KNH2. The coupled spectrum for Si is dashed and the decoupled spectrum is solid. In the uncoupled spectrum, the asymmetric peak at ~ -43 ppm is evidence of more than one type of Si environment. The peak at ~ -79 ppm is due to elemental Si that was an impurity in the product. The coupled spectrum, shows interaction between 29Si and 1H. The predominant asymmetric peak indicates the presence of more than one type of Si environment. The elemental Si peak is not present since there would be no 1H interaction.

Discussion

The purpose of this experiment was to develop a process for synthesizing

α-silicon nitride in a manner that would result in an affordable product. Although

there have been numerous methods developed over the past century to produce silicon

nitride, each has its own drawbacks. They either involve expensive starting

materials,14, 15 relatively high pressures,16 long reaction times,14 products that have

significant contamination requiring purification,11-13, 15 or an agglomerated product

that needs extensive grinding and may introduce additional contaminants.15

When considering the expense of a process, the most obvious factor is the cost

of the starting materials. In this experiment, the reactants silicon and NaNH2 are

readily available and relatively inexpensive. KNH2 is easily synthesized from

potassium metal and liquid ammonia.26

The reaction was run at 320 °C for approximately 1 hour. The temperature and time both compare quite favorably with existing synthetic methods. While utilizing a dry, inert atmosphere is somewhat inconvenient; it does not pose a significant cost factor for most industrial concerns. The intermediate does require washing in liquid ammonia, but again this should not be a major expense for commercial companies. Following the washing process, the intermediate was annealed; excluding the ramp-up and ramp-down times, the actual annealing soak took place over a period of 2 hours at a temperature of 1500 °C. It may even be possible to anneal the intermediate at lower temperatures since the exact annealing temperature required for the conversion to α-silicon nitride has not been determined.

Since no halides were used and the reactions takes place in an inert environment, the issue of chloride retention or oxide contamination has been virtually eliminated or at least minimized. While the process is by no means a final solution to the problem of prohibitively priced silicon nitride, it is certainly a step in the right direction.

Cost considerations aside, the XRD pattern for the silicon and NaNH2 reaction shows a relatively pure α-silicon nitride was produced. Depending on whether the intermediate was washed in liquid ammonia (to remove remaining alkali metal) and the temperatures for the subsequent heating, combinations of α and β-silicon nitride or only β-silicon nitride are possible.

Silicon nitride exists in four different phases, i.e. α (P31c, a = 0.775 nm, c =

0.562 nm), β (P63/m, a = 0.760 nm, c = 0.291 nm), γ (Fd3m, a = 0.780 nm) and a new hexagonal phase (P62c, a = 0.737 nm, c = 0.536 nm).27 It is the α and β forms

that are generally discussed, however, when considering engineering fabrication

materials. They are both hexagonal systems but β silicon nitride has a higher aspect ratio than α silicon nitride and, therefore, is more needle-like. The symmetry that these structures exhibit is a common thread that ties silicon nitride to other ultra-hard materials such as diamond and cubic boron nitride.28

The nature of the intermediate is interesting. It was first assumed that the

structure was a diimide, i.e. Si(NH)2. However, the FTIR spectrum showed evidence of a Si-H bond, which would not be present in a diimide structure. The SS NMR spectra indicated the presence of silicon atoms with different proton environments.

These results are all consistent with a probable polysilazane structure for the

intermediate with the formula (SiH2NH)n. It would account for the Si-H, Si-N, and

N-H bonds that were indicated by the FTIR spectrum and also the multiple Si proton environments inferred from the SS NMR spectra.

There is considerable work discussing polysilazanes as precursors to silicon

nitride but most of these follow a primarily organic synthetic route. For example, a

reaction between dichlorosilane, methyldichlorosilane, and ammonia produced a

polysilazane precursor for silicon nitride formation (see Equation 4-9).29

NH3 3 H2SiCl2 + H3CSiHCl2 [(SiH2-NH)3(H3CSiH-NH)]n (4-9)

The product was then annealed at 1400 °C and gave a composite of Si3N4/SiC.

An insight into possible multiple intermediates is given by Mazdiyasni and

Cooke for the formation and subsequent polymerization of [Si(NH)2]n according to

Equations 4-10 through 4-15).30 n-hexane (4-10) SiCl4 + 6 NH3 Si(NH)2 + 4 NH4Cl 0 ˚C

360 ˚C n Si(NH)2 [Si(NH)2]n NH4Cl (4-11) in vacuo

400 ˚C (4-12) 6 [Si(NH)2]n 2 [Si3(NH)3N2]n -NH3

650 ˚C 2 [Si3(NH)3N2]n 3 [Si2(NH)N2]n (4-13) -NH3

1200 ˚C (4-14) 3 [Si2(NH)N2]n 2 Si3N4 (amorphous) -NH3

1300-1400 ˚C (4-15) !-Si3N4

Si3N4 (amorphous)

>1400 ˚C !-Si3N4 and "-Si3N4 The amide route to α-silicon nitride presented in this chapter is quite unique in creating a polysilazane intermediate that does not involve organics. The intermediate itself, although air and moisture sensitive, could conceivably be stored in an air-free atmosphere until needed as a source of silicon nitride.

Conclusions

A cost effective method for synthesizing α-silicon nitride has been developed that utilizes a polysilazane intermediate. This is most interesting since this is the first polysilazane synthesis to be reported without use of an organic precursor. Further research is needed to determine optimum annealing times and possible multiple intermediates.

References

(1) Liu, M. and S. Nemat-Nasser, Microstructure of a bearing-grade silicon

nitride. Journal of Materials Research, 1999. 14(12): p. 4621-4629.

(2) Deville, H. and F. Wohler, Ueber das Stickstoffsilicium. Liebigs Ann. Chem.,

1857. 104(2): p. 1.

(3) Schutzenberger, P., Sur l'azoture de silicium. Compt. Rend. Acad. Sci., 1879.

89: p. 644-6.

(4) Riley, F.L., Silicon Nitride and Related Materials. Journal of the American

Ceramic Society, 2000. 83(2): p. 245-265.

(5) Lee, M.R., S.S. Russell, J.W. Arden, and C.T. Pillinger, Nierite (Si3N4), a new

mineral from ordinary and enstatite chondrites. Meteoritics, 1995. 30(4): p.

387-398.

(6) Li, Z.L., J.S. Williams, D.J. Llewellyn, and M. Giersig, Mechanochemical

reaction and formation of an amorphous nitride phase during ball milling of

Si in NH[sub 3]. Appl. Phys. Lett., 1999. 75(20): p. 3111-3113.

(7) Krstic, Z. and V. Krstic, Silicon nitride: the engineering material of the future.

Journal of Materials Science, 2012. 47(2): p. 535-552.

(8) Bal, B.S. and M.N. Rahaman, Orthopedic applications of silicon nitride

ceramics. Acta Biomaterialia, 2012. 8(8): p. 2889-2898.

(9) Mazzocchi, M. and A. Bellosi, On the possibility of silicon nitride as a

ceramic for structural orthopaedic implants. Part I: processing,

microstructure, mechanical properties, cytotoxicity. J Mater Sci: Mater Med,

2008. 19: p. 2881-2887.

(10) Wiederhorn, S.M. and M.K. Ferber, Silicon nitride for gas turbines. Current

Opinion in Solid State & Materials Science, 2001. 5(4): p. 311-316.

(11) Schwier, G., G. Nietfeld, and G. Franz, Production and Characterization of

Silicon Nitride Powders. Materials Science Forum, 1989. 47: p. 1-20.

(12) Weimer, A.W., Effect of Si metal addition to C--SiO2 precursors for

synthesizing Si3N4 by carbothermal nitridation. Journal of Materials Science

Letters, 1998. 17(2): p. 123-125.

(13) Weimer, A.W., G.A. Eisman, D.W. Susnitzky, D.R. Beaman, and J.W.

McCoy, Mechanism and Kinetics of the Carbothermal Nitridation Synthesis of

alpha-Silicon Nitride. Journal of the American Ceramic Society, 1997. 80(11):

p. 2853-2863.

(14) Rao, R.R. and T.S. Kannan, Nitride-bonded silicon nitride from slip-cast Si +

Si3N4 compacts. Journal of Materials Research, 2002. 17(2): p. 386-395.

(15) Morgan, P.E.D. and E.A. Pugar, Synthesis of Si3N4 with Emphasis on Si-S-N

Chemistry. Journal of the American Ceramic Society, 1985. 68(12): p. 699-

703.

(16) Gu, Y., L. Chen, and Y. Qian, Low-Temperature Synthesis of Nanocrystalline

alpha-Si3N4 Powders by the Reaction of Mg2Si with NH4Cl. Journal of the

American Ceramic Society, 2004. 87(9): p. 1810-1813.

(17) Pawelec, A., B. Strojek, G. Weisbrod, and S. Podsiadlo, Preparation of silicon

nitride powder from silica and ammonia. Ceramics International, 2002. 28(5):

p. 495-501.

(18) Kawano, S., J. Takahashi, and S. Shimada, Highly electroconductive

TiN/Si3N4 composite ceramics fabricated by spark plasma sintering of Si3N4

particles with a nano-sized TiN coating. J. Mater. Chem., 2002. 12: p. 361-

365.

(19) Zhan, G.-D., M. Mitomo, R.-J. Xie, and K. Kurashima, Ductile-to-brittle

transition in superplastic silicon nitride ceramics. J. Mater. Res., 2002. 17(1):

p. 149-155.

(20) Hirosaki, N., T. Saito, F. Munakata, Y. Akimune, and Y. Ikuhara,

Transmission electron microscopy observation of second-phase particles in

β–Si3N4 grains. J. Mater. Res., 1999. 14(7): p. 2959-2965.

(21) Shen, Z., Z. Zhao, H. Peng, and M. Nygren, Formation of tough interlocking

microstructures in silicon nitride ceramics by dynamic ripening. Nature, 2002.

417: p. 266-269.

(22) Cao, Y.G., H. Chen, J.T. Li, C.C. Ge, S.Y. Tang, J.X. Tang, and X. Chen,

Formation of α-Si3N4 whiskers with addition of NaN3 as catalyst. Journal of

Crystal Growth, 2002. 234(1): p. 9-11.

(23) Han, W., S. Fan, Q. Li, B. Gu, X. Zhang, and D. Yu, Synthesis of silicon

nitride nanorods using carbon nanotube as a template. Applied Physics

Letters, 1997. 71(16): p. 2271-2273.

(24) Xu, F., X. Zhang, W. Xi, J. Hong, and Y. Xie, Growth of Pure β-Si3N4

Nanorods from the Synergic Nitrogen Sources. Chemistry Letters, 2003.

32(7): p. 600-601.

(25) Shriver, D.F. and M.A. Drezden, The Manipulation of Air-Sensitive

Compounds. 2nd ed. 1974, New York: John Wiley & Sons. 687-704.

(26) Brauer, G., Potassium Amide, in Handbook of Preparative Inorganic

Chemistry, R.F. Riley, Editor. 1963, Academic Press: New York. p. 1044.

(27) Cai, Y., A. Zimmermann, S. Prinz, and F. Aldinger, Space Group

Determination of a Novel Silicon Nitride Phase. physica status solidi (b),

2002. 231(1): p. R4-R6.

(28) Lowther, J.E., Symmetric Structures of Ultrahard Materials. Journal of the

American Ceramic Society, 2002. 85(1): p. 55-58.

(29) Hörz, M., A. Zern, F. Berger, J. Haug, K. Müller, F. Aldinger, and M.

Weinmann, Novel polysilazanes as precursors for silicon nitride/silicon

carbide composites without “free” carbon. Journal of the European Ceramic

Society, 2005. 25(2-3): p. 99-110.

(30) Mazdiyasni, K.S. and C.M. Cooke, Synthesis, Characterization, and

Consolidation of Si3N4 Obtained from Ammonolysis of SiCl4. Journal of the

American Ceramic Society, 1973. 56(12): p. 628-633.

Chapter 5: Luminescence in Ternary

Silicon Nitrides

Introduction

Photoluminescence has been exploited for fluorescent lamps since prior to

World War II.1 The technology of fluorescent lamp tubes has not progressed much in the last seventy years. Many of the bulbs currently in use are still filled with a low- pressure of mercury vapor which when excited generates ultraviolet radiation, which in turn excites a phosphor coating the walls of the tube; that phosphor then produces white light. Phosphors are essentially an emitting species doped into a host lattice.

The coordination environment in the host lattice gives rise to specific crystal field splitting that produces the desired color. Despite the use of mercury, fluorescent bulbs are more energy efficient than traditional incandescent bulbs and their use has

increased with the mandated phase out of incandescent lighting.

The phosphors in fluorescent lighting are typically oxide-based hosts lattices

doped with lanthanides. Commonly used hosts are aluminates and phosphates. An

example of such a phosphor is BaMgAl10O17:Eu:Mn (also known as BAM and Philips phosphor U788); it is a commonly used phosphor that emits blue-green light. These materials can be synthesized using traditional solid-state methods, hydrothermal syntheses, and pyrolytic approaches.2, 3 Despite the variety of synthetic methods,

these phosphors are degraded due to the formation of oxygen vacancies and thermal

degradation.4, 5 Non-oxide phosphor hosts offer the potential for longer-lived

phosphors and products with lower lifetime costs.

Display and lighting technologies require the three primary luminescent

colors: red, green and blue.6 Researchers have begun to look for new materials to

fulfill these roles,7 the addition of dopants into metathesis products offers another means of producing such phosphors. The dopants necessary to create red, green, and blue emitters are typically lanthanides such as Eu2+, Ce3+, Gd3+, Tb3+, Y3+, and Eu3+.6-8

The structure of the host lattice greatly affects the resulting color and

intensity.6 Both the local coordination environment and the efficiency of energy

transfer to the emissive species are important factors when selecting a host. Efficient

energy transfer to the host can result in non-radiative processes and reduce the overall

luminescence efficiency. Unique coordination environments (such as the cubic sites

9 3+ 10 in Ca16Si17N34) can allow new colors to be realized. In this case, red from Ce .

Alkali and alkaline earth silicon nitrides (nitride phosphors) offer stability, relative

inertness, and the potential for new coordination environments.11

Self-propagating solid-state metathesis (SSM) reactions offer a route to compounds and phases of materials that are often difficult to produce using conventional methods. SSM has been used to successfully synthesize refractory ceramics,11-13 other inorganic solids,14-19 solid-solutions and ceramic composites.20

These self-propagating reactions can also be used to control crystallite size.21 SSM reactions have been used to produce main group nitrides22 and transition metal

nitrides.23-26 The synthesis of doped-ternary silicon nitrides through SSM reactions

could provide an elegant and inexpensive route to luminescent materials. In a slight

departure from traditional SSM reactions, depending on the host ternary silicon

nitride, some of these reactions require significant heat input (an oxy-methane torch) for initiation.27 The ternary silicon nitride phosphors presented in this chapter are characterized using X-ray powder diffraction and fluorimetry.

Experimental

Chemicals

The following reagents were used as received: lithium nitride (Li3N, Cerac;

99.5%, 60 mesh), magnesium nitride (Mg3N2, Alfa Aesar; 99.6%, 325 mesh),

elemental silicon (Alfa Aeser, -325 mesh, 99.5%), red phosphorus (Cerac; 99.5%, -

325 mesh), calcium nitride (Ca3N, Cerac; 99%, 200 mesh), europium oxide (Eu2O3,

Strem, 99.99%), terbium oxide (Tb4O7, Strem, 99.9%), cerium oxide (CeO2, Strem,

99.9%). Fumed silica (Cabot Corporation) was dried in a drying oven at 180 °C and

stored in a helium-filled glove box. Misch metal oxide was made by heating a striker flint (Fisher Scientific) in air at 500 °C to form a mixed oxide. Hydrochloric acid

(Fisher) was diluted to an approximately 6 M concentration by mixing equal parts of acid and water. All other reagents were used as received.

Synthesis

LiSi2N3 Phosphor Synthesis

The general reaction for the formation of the lithium silicon nitride (LiSi2N3) is as follows (Equation 5-1):

6 Si + 12 P + 13 Li3N → 3 LiSi2N3 + 12 Li3P + 2 N2 (5-1)

Since many of the reactants are oxygen or water sensitive, all reactions were prepared and performed in a helium-filled glove box. The reactants and dopants (0.01 to

0.20 mol) were weighed in a helium-filled glove box as described by the following

charts. The synthesis of lanthanide-doped LiSi2N3 was carried out by weighing stoichiometric amounts of elemental silicon, phosphorus, lithium nitride, and a lanthanide oxide in a helium-filled glove box. Approximately 0.65 g of material was used for each reaction. The reactants were ground into a fine mixture with a mortar and pestle. Warning: Solid-state metathesis reactions are very exothermic and highly energetic. Reactions of this type should be performed on a small scale using precautions afforded to similar reactions. While in the glove box, the reaction mixture was transferred to a 0.5-inch diameter die and pressed into a pellet under a load of 3000 lbs. The pellet was then placed into a steel reaction vessel modeled after a bomb calorimeter.21 A resistively heated nichrome wire touching the top of the pellet was used to initiate the reaction. The reaction was completed in a matter of seconds.

After the vessel cooled, the product was removed and taken out of the glove box. The product was ground with a ceramic mortar and pestle, washed with water

and stirred with 6.0 M hydrochloric acid. Caution: The initial washing should be performed in a fume hood. The addition of aqueous reagents to this solid mixture can

initially produce flames, pops, and release toxic gases (PH3). Adequate precautions should be taken. To ensure that the product has been successfully doped, the unincorporated lanthanide oxides were removed by several washes with 6.0 M hydrochloric acid. Subsequent water washes removed any other byproducts and salts.

Between each wash, the product suspension was centrifuged (four cycles of 5 minutes each) to facilitate removal of the supernatant liquid and wash away excess dopant and byproducts. After isolation, the samples were left to dry in a 180 °C oven. These powders were then used to make X-ray and luminescence slides.

MgSiN2 Phosphor Synthesis

The synthesis of MgSiN2 (Equation 5-2) was carried out as outlined in

reference 5-X27. The reactants and dopants (0.01 to 0.20 mol) were weighed in a helium-filled glove box. This reaction requires a more significant input of energy to initiate; this was accomplished using an oxy-methane torch with a flame temperature of ≥2500 °C.

SiO2 + Mg3N2 → MgSiN2 + 2 MgO (5-2)

The addition of Li3N and Ca3N2 to the reaction was shown to increase crystallinity.

XRD Analysis

The products were analyzed using an X’Pert PANalytical powder X-ray

diffractometer with Cu Kα radiation (λ = 1.5418 Å). Product samples were scanned

from 10-100° 2θ at 0.1-degree intervals with a 3 second count time per interval. The

patterns obtained were analyzed and interpreted using the Macintosh software

MacDiff (http://www.geol.uni-erlangen.de/html/software/Macdiff.html) and CMPR.

Luminescence

The products were analyzed using a Jobin-Yvon-Spex Tau3 Fluorolog

spectrofluorometer in reflection mode. Steady-state emission and excitation spectra were collected. The luminescence slides were prepared by placing the powder samples onto plastic slides (~1 cm wide) and then using a drop of methanol to disperse them. After the slides were dry, aerosol hairspray (Rave) was used as a fixant.

Results and Discussion

X-ray powder diffraction of the doped and undoped alkali and alkaline earth silicon nitrides (Figures 5-1 and 5-2) shows no change in the lattice as dopants are added.

Figure 5–1. Powder X-ray diffraction pattern of LiSi2N3 as synthesized (bottom trace), 0.20 mol Eu doped into the LiSi2N3 lattice (middle), and 0.20 mol Tb doped into the LiSi2N3 lattice (top trace).

Figure 5–2. Powder X-ray diffraction pattern of MgSiN2 as synthesized (bottom trace), 0.10 mol Eu doped into the MgSiN2 lattice (middle), and simulated pattern of MgSiN2 (JC-PDS 25-0530) (top trace).

Fluorescence

Europium ions incorporated into the alkali and alkaline earth silicon nitride

lattices emit in the red (MgSiN2) and the blue (LiSi2N3) as seen in Figure 5-3. The

3+ red emission at 618 nm was achieved with 0.1 mol of Eu2O3 and is due to the Eu

2+ ion. The blue emission at 475 nm was due to the Eu emission; 0.1 mol of Eu2O3

was added to the sample at synthesis.

Terbium ions (Figure 5-4), incorporated into the LiSi2N3 lattice through the addition of 0.1 mol of terbium oxide, were shown to have a broad emission in the green (centered on 546 nm, excitation wavelength 340 nm).

Violet emission was obtained from the addition of 0.1 mol of thulium oxide to the lithium silicon nitride reaction; a broad emission peak was seen at 410 nm.

Unexpectedly, the LiSi2N3 lattice itself emits violet as well (broad emission, centered around 429 nm) when excited with 300 nm light (Figure 5-5). A broad violet emission was also seen at 412 nm in lithium silicon nitride doped with 0.1 mol of misch metal oxide (excitation at 326 nm); this emission is from cerium ions incorporated into the lattice.

Figure 5–3. Emission spectra of Eu doped LiSi2N3 (emission at 475 nm, excitation wavelength of 360 nm), MgSiN2 doped with Eu (emission at 618 nm, excitation wavelength of 380 nm) and europium oxide (excitation wavelength of 380 nm).

Figure 5–4. Emission spectra of Tb doped LiSi2N3 (broad emission centered around 546 nm, excitation wavelength of 340 nm), LiSi2N3 and terbium oxide.

Figure 5–5. Emission spectra of Tm doped LiSi2N3 (broad emission centered around 410 nm, excitation wavelength of 300 nm). Unexpectedly, LiSi2N3 (429 nm) and thullium oxide (374 nm) also appear to emit.

Conclusions

Alkali and alkaline earth silicon nitrides can be synthesized via solid-state

metathesis reactions. During these reactions, it is possible to dope the lattice with

metal oxides and produce luminescent materials; it should also be noted that

europium has been used as a dopant in the Eu3+ oxidation state and been reduced to

2+ Eu within the LiSi2N3 lattice. This is a possible route to producing large quantities of a doped nitride suitable for phosphors; a difficult task using traditional methods given the propensity for decomposition when thermally processing nitrides. Red, blue, green and violet emitting phosphors have been produced.

Additional research is needed to determine the optimum doping level for maximum emission as well as the level of dopant incorporation. It would be interesting to study the impact of enhanced product crystallinity on luminescence or if

using different size starting materials (e.g. –100 and –325 mesh SiO2) could produce

more crystalline products.

References

(1) Blasse, G. and B.C. Grabmaier, Luminescent Materials. 1994, Berlin:

Springer-Verlag.

(2) Chen, L.-T., I.L. Sun, C.-S. Hwang, and S.-J. Chang, Luminescence properties

of BAM phosphor synthesized by TEA coprecipitation method. Journal of

Luminescence, 2006. 118(2): p. 293-300.

(3) Li, H.-L., N. Hirosaki, R.-J. Xie, T. Suehiro, and M. Mitomo, Fine yellow a-

SiAlON:Eu phosphors for white LEDs prepared by the gas-reduction–

nitridation method. Science and Technology of Advanced Materials, 2007. 8:

p. 601-606.

(4) Bizarri, G. and B. Moine, On phosphor degradation mechanism: thermal

treatment effects. Journal of Luminescence, 2005. 113(3–4): p. 199-213.

(5) Klaassen, D.B.M., L.D.M. De, and T. Welker, Degradation of phosphors

under cathode-ray excitation. J. Lumin., 1987. 37(1): p. 21-8.

(6) West, A.R., Basic Solid State Chemistry. 2nd ed. 1999, Chichester, West

Sussex, England: John Wiley & Sons Ltd. 480.

(7) Neeraj, S., N. Kijima, and A.K. Cheetham, Novel red phosphors for solid-

3+ state lighting: the system NaM(WO4)2-x(MoO4)x:Eu (M = Gd, Y, Bi).

Chemical Physics Letters, 2004. 387: p. 2-6.

(8) Hao, J. and M. Cocivera, Cathodoluminescence of Sr2B5O9Cl thin films doped

with Tm3+, Tb3+ and Mn2+. Journal of Physics: Condensed Matter, 2002. 14: p.

925-933.

(9) Hick, S.M., M.I. Miller, R.B. Kaner, and R.G. Blair, Synthesis and Crystal

Structure of Cubic Ca16Si17N34. Inorganic Chemistry, 2012. 51(23): p. 12626-

12629.

(10) Toquin, R.L. and A.K. Cheetham, Red-emitting cerium-based phosphor

materials for solid-state lighting applications. Chemical Physics Letters,

2006. 423: p. 352-356.

(11) Gillan, E.G. and R.B. Kaner, Synthesis of Refractory Ceramics via Rapid

Metathesis Reactions between Solid-State Precursors. Chemistry of Materials,

1996. 8(2): p. 333-43.

(12) Gillan, E.G. and R.B. Kaner, Rapid, energetic metathesis routes to crystalline

metastable phases of zirconium and hafnium dioxide. Journal of Materials

Chemistry, 2001. 11(7): p. 1951-1956.

(13) Parkin, I.P. and A.T. Nartowski, Fast Metathesis Routes to Tungsten and

Molybdenum Carbides. J. Mater. Sci. Lett., 1999. 18: p. 267-268.

(14) Kaner, R.B., S.K. Bux, J.-P. Fleurial, and M. Rodriguez, Rapid solid-state

metathesis routes to nanostructured silicon-germanium,Patent

US20110318250A1, 2011

(15) Bux, S.K., M. Rodriguez, M.T. Yeung, C. Yang, A. Makhluf, R.G. Blair, J.-P.

Fleurial, and R.B. Kaner, Rapid solid-state synthesis of nanostructured

silicon. Chem. Mater., 2010. 22(8): p. 2534-2540.

(16) Cumberland, R.W., R.G. Blair, C.H. Wallace, T.K. Reynolds, and R.B. Kaner,

Thermal Control of Metathesis Reactions Producing GaN and InN. Journal of

Physical Chemistry B, 2001. 105(47): p. 11922-11927.

(17) O'Loughlin, J.L., C.H. Wallace, M.S. Knox, and R.B. Kaner, Rapid Solid-

State Synthesis of Tantalum, Chromium, and Molybdenum Nitrides. Inorganic

Chemistry, 2001. 40(10): p. 2240-2245.

(18) Jarvis, R.F., Jr., R.M. Jacubinas, and R.B. Kaner, Self Propagating Metahtesis

Routes to Metastable Group 4 Phosphides. Inorganic Chemistry, 2000. 39: p.

3243-3246.

(19) Bonneau, P.R., J.B. Wiley, and R.B. Kaner, Rapid Solid-State Synthesis of

Materials from Molybdenum Disulfide to Refractories. Inorganic Synthesis,

1995. 30: p. 33-37.

(20) Rao, L., E.G. Gillan, and R.B. Kaner, Rapid Synthesis of Transition Metal

Borides by Solid-State Metathesis. J. Mater. Res., 1995. 10: p. 353-361.

(21) Bonneau, P.R., R.F. Jarvis, Jr., and R.B. Kaner, Metathetical Precursor Route

to Molybdenum Disulfide. Nature (London), 1991. 349: p. 510-512.

(22) Janes, R.A., M. Aldissi, and R.B. Kaner, Controlling Surface Area of

Titanium Nitride Using Metathesis Reactions. Chemistry of Materials, 2003.

15(23): p. 4431-4435.

(23) Cordova, J. and J. Laine, Physical and chemical characterization of national

diatomaceus earth: Laghuna Brava, Estado Merida. Acta Cientifica

Venezolana, 1989. 40(2): p. 155-6.

(24) Hector, A.L. and I.P. Parkin, Magnesium and Calcium Nitrides as Nitrogen

Sources in Metathetical Reactions to Produce Metal Nitrides. Chemistry of

Materials, 1995. 7(9): p. 1728-33.

(25) Ali, S., M.D. Aguas, A.L. Hector, G. Henshaw, and I.P. Parkin, Solid state

metathesis routes to metal nitrides; use of strontium and barium nitrides as

reagents and dilution effects. Polyhedron, 1997. 16(20): p. 3635-3640.

(26) Parkin, I.P. and A.M. Nartowski, Solid state metathesis routes to Group IIIa

nitrides: comparison of Li3N, NaN3, Ca3N2 and Mg3N2 as nitriding agents.

Polyhedron, 1998. 17(16): p. 2617-2622.

(27) Blair, R.G., A.J. Anderson, and R.B. Kaner, A Solid-State Metathesis Route to

MgSiN2. Chemistry of Materials, 2005. 17(8): p. 2155-2161.

Chapter 6: Solid-State Metathesis Routes to

Refractory Carbides

Introduction

Diamond is the hardest form of carbon and one of the hardest materials known

to man. Although widely used in cutting tools and abrasives, diamond is expensive

and reacts with metals such as iron during cutting operations to form carbides. Metal

carbides can be very hard as well. Although not as hard as diamond, they can be

produced in a more cost effective manner and do not adversely interact with iron.

These hard carbides have found uses as diamond substitutes in applications such as

cutting tools and abrasives. Boron carbide (B4C) and zirconium carbide (ZrC) are

both ultra-hard, refractory ceramics used in many applications.1-4

5 Boron carbide (B4C), first synthesized by H. Moissan in 1899. , is one of the

hardest known bulk materials after diamond, cubic boron nitride (c-BN), and certain

transition metal borides. Because of its hardness (~3000 kg/mm2) and excellent wear resistance, it is widely used as an abrasive.6 Because it is light and hard, it has found

applications in tank armor and bulletproof vests.7, 8 Its high cross-section for neutron

absorption makes it a useful material for use in nuclear applications as power plant control rods.

Figure 6–1. Vicker’s hardness of common materials.

Zirconium carbide (ZrC), an excellent material for engineering applications, is

not widely used because it is difficult to sinter into a compact billet. It is of interest because of its high hardness (2700 kg/mm2), high strength (990 MPa), high melting

point (3540 °C), and relatively good electrical conductivity (2.0 x 10-8 ohm-1cm-1).9

These properties make ZrC an excellent candidate for strengthening brittle

thermoelectric compounds. For example, cobalt antimonide (CoSb3) softens at its

desired temperature of operation (~700 ºC) and has poor machinability. This leads to

cracks and poor surface morphologies. Carbide reinforcement using nanoparticles

offers the potential for improved machinability and crack resistance. This comes about through crack pinning. Crack pinning (see Figure 6-2) occurs when crack propagation in a material stops after a crack encounters a grain in the composite with a different bulk modulus than the matrix.

Figure 6–2. Crack pinning in a nanoparticle reinforced material. Crack propagation stops after a crack encounters a nanoparticle. The synthesis of carbonaceous materials by solid-state metathesis has been

previously reported10 and typically involves prolonged heating of the elements or a

mixture of their oxides and a carbon source.11 An alternative to these very energy- intensive approaches could lower the production costs associated with refractory carbides. One approach is to form the carbides though metathesis reactions that are driven by the enthalpy of formation of a salt or oxide by-product. These solid-state metathesis (SSM) reactions can be performed in the absence of a solvent and have been used to produce a variety of refractory inorganic compounds including a range of carbides,12 borides,13 and nitrides.14-18 A metathesis reaction is a double displacement reaction (see Equation 6.1); in the solid-state, these reactions typically involve a metal halide (AB, in equation 6.1 below) and a reactive carbon, boron, or nitrogen source (CD, below) to form a binary (or even a ternary) product.

AB(s) + CD(s) → AD(s) + CB(s) (6.1)

This approach requires an initial input of energy to produce a self-propagating reaction. Typically this energy is supplied with a resistively heated wire. But, any

heat source can be used even an open flame or the application of microwave energy.

Rapid, solid-state metathesis reactions initiated by an open flame are known to

produce transition metal carbides in good yields.10 Solid-state metathesis reactions typically produce highly crystalline, nano-scale materials via microwave initiation with a dramatically reduced completion time.19 Microwave initiated reactions are of interest because through their rapid temperature increase followed by a rapid decrease in temperature, particle size remains low, producing nano- and micron-sized particles.

This rapid cooling prevents the clumping and growth of particles during propagation of the reaction, essentially stopping Ostwald ripening, and produces a higher number of particles will within the nano-scale regime. By combining previously studied

reactions with microwave initiation, nano-sized B4C and ZrC can be accessed efficiently and inexpensively. Such a preparation method could be ideal for industry and may also access improved mechanical properties. In this chapter, we investigate solid-state metathesis (double displacement) reactions between metal halides and reactive carbon sources as an inexpensive route to form boron carbide and zirconium carbide.

Experimental

Chemicals

The reagents used were zirconium chloride (Alfa Aesar, 99.5%), calcium carbide (Fluka; 80%), magnesium boride (Alfa Aesar), hexachloroethane (Aldrich,

99%), boron oxide (Cerac, 99.9%, 200 mesh), and boron (Cerac, 99.9+%). The

calcium carbide (Fluka) was loaded into a ball mill (SPEX Mixer Mill 8000D) under

argon atmosphere, ground for 30 minutes, and sieved through a 400 mesh sieve. All

solid reagents were stored in a helium-filled glove box. Hydrochloric acid (Fisher)

was diluted to an approximately 6 M concentration by mixing equal parts of acid and

water. All other reagents were used as received.

6 MgB2 + 2 C2Cl6 → 3 B4C + 6 MgCl2 + C (6.2)

ZrCl4 + 2 CaC2 → ZrC + 2 CaCl2 + 3 C (6.3)

Synthetic Preparation

Since many of the reactants are oxygen or water sensitive, all of the reactions were prepared in an inert atmosphere (helium) glove box. A total reaction mix of

0.65 g (used for most reactions) was weighed on an analytical balance inside the glove box. The reactions were balanced so that all of the alkali metal and halogen is consumed in the reaction. Using an agate mortar and pestle, the reactant powders were ground into a fine mixture.

B4C synthesis was conducted using steel reaction vessel modeled after a bomb calorimeter.20 The reactions were initiated using torch,18 microwave,19 or carbon arc methods. ZrC synthesis was initiated using the bomb calorimeter-type reactor, torch, and microwave. Table 6.1 lists a partial set of reactions investigated. Bomb reactor and carbon arc reactions were initiated in a glove box. Microwave and torch reactions were removed from the glove box prior to initiation.

Torch Synthesis

The reaction mixture was pressed into a pellet with the use of a hydraulic press located inside the glove box. The pellet was then placed into a tightly capped

20 mL borosilicate glass scintillation vial and removed from the glove box. The flame of an oxy-methane torch was then brought up to the pellet in the vial and quickly removed once initiated.

The reacted pellet was ground under 6 M HCl and stirred. The product was isolated by two subsequent washes with distilled water, followed by centrifugation, and drying in a hot oven (180 ˚C).

Microwave Synthesis

The reaction mixture was prepared as above, but left as a loose powder rather than pressed into a pellet. The reaction mixture was then placed into a tightly capped

20 mL borosilicate glass scintillation vial and removed from the glove box.

Reactions were initiated using microwave energy in an 1100 watt Daewoo household microwave oven (2.45 GHz) with a firebrick plate as a reaction platform in place of the standard glass turntable. Caution! These reactions reach high temperatures rapidly. Care should be taken in the choice of a reaction vessel and when increasing the reaction scale. The reaction vessel can become extremely hot and may soften.

Larger scale reactions may also lead to catastrophic vessel failure. Microwave energy was applied to the reaction mixture at the 100% power setting.

XRD Analysis

The products were analyzed using a PANalytical X’Pert Pro diffractometer

with Cu Kα radiation (λ = 1.5418 Å). Product samples were scanned from 10-100˚ 2θ at 0.03-degree intervals with a 25 second count time per interval. The patterns obtained were then analyzed and interpreted using the PANalytical X’Pert Highscore

Plus software and the Joint Committee on Powder Diffraction Standards (JC-PDS) database.

Crystallite size and lattice strain of ZrC products were determined using the method of integral line breadths.21 In accordance with this method, a plot of the area under each diffraction peak versus its d-spacing gives a linear relationship. The slope of this line contains information on the crystallite size, while the y-intercept indicates the lattice strain. Size broadening is assumed to follow a Gaussian profile and strain broadening is modeled as a Cauchy profile.

Results

The initial reaction of magnesium boride (MgB2) and hexachloroethane

(C2Cl6) in a bomb reactor produced B4C with amorphous carbon impurities. The

reaction between MgB2 and C2Cl6 by carbon arc produced boron carbide with fewer

side products, although the yield was low. Boron carbide synthesis attempted by

microwave initiation most often resulted in the production of carbon compounds

rather than the desired product. The reaction between magnesium boride,

hexachloroethane, and sulfur resulted in the highest yield of B4C (70-80%) with trace

graphite impurities (Figure 6-3). The average crystallite size when initiated in a

bomb reactor was 23.3 nm and 25.0 nm when initiated by an oxy-methane torch.

Figure 6–3. Powder X-ray diffraction pattern of the washed B4C product formed from the reaction MgB2 + C2Cl6 + S that was initiated by a resistively heated nichrome wire in a bomb-type reactor. A simulated pattern for B4C is provided as a reference (JC-PDS 35-0798). The remaining peaks can be attributed to graphite (JC-PDS 08-0415) and an unknown impurity.

Table 6–1. Results of attempted B4C forming reactions broken down by initiation method.

Initiation Reaction Result Method

MgB2 + C2Cl6 Bomb Amorphous Carbon + B4C

Torch Graphite + B4C Microwave Amorphous Carbon

Carbon Arc B4C w/ Graphite Impurities

MgB2 + C2Cl6 + S Bomb B4C w/ Graphite Impurities

Torch B4C w/ Graphite Impurities Microwave Amorphous Carbon

MgB2 + C2Cl6 + P Bomb Graphite Torch Graphite

MgB2 + Poly Vinyl Chloride Bomb No Reaction Torch Graphite

B2O3 + C2Cl6 Bomb No Reaction Torch Amorphous Carbon

B + S + CaC2 Bomb Graphite Torch Graphite

B + P + CaC2 Bomb Graphite Torch Amorphous Carbon

Powder X-ray diffraction data indicated that both the microwave and torch initiated products from the reaction of zirconium chloride and calcium carbide are

made up of zirconium carbide and zirconium oxide (ZrO2), as seen in Figure 6-4 and

6-5. Although contaminated, the microwave and torch initiations successfully produced zirconium carbide. The crystallite size of ZrC was calculated to be 29.0 nm

(strain = 0.2 ± 0.2%) for a microwave initiated reaction and 22.9 nm (strain = 0.21 ±

0.03%) for the torch-initiated reaction.

_ O Figure 6–4. Powder X-ray diffraction pattern of the washed ZrC product formed

from the microwave initiated reaction of ZrCl4 + 2 CaC2 overlaid with a simulated pattern of ZrC (JC-PDS 35-0784). The remaining peaks can be attributed to two different phases of ZrO2 (•, monoclinic, JC-PDS 37-1484 and *, tetragonal, JC-PDS 17-0923) and graphite ( JC-PDS 26-1079).

_ O Figure 6–5. Powder X-ray diffraction pattern of the washed ZrC formed from the reaction of ZrCl4 + 2 CaC2 and initiated by an oxy-methane torch. The simulated pattern of ZrC (JC-PDS 35-0784) is overlaid. Two different phases of ZrO2 (•, monoclinic, JC-PDS 37-1484 and *, tetragonal, JC-PDS 17-0923) and graphite ( JC-PDS 26-1079) account for the remaining peaks.

Table 6–2. Results of ZrC Forming Reactions Broken Down by Initiation Method

ZrCl4 + 2 CaC2 Bomb ZrC + Graphite + ZrO2

Torch ZrC w/ ZrO2 Impurities

Microwave ZrC w/ ZrO2 Impurities

ZrCl4 + 2 CaC2 + S Bomb Amorphous Carbon Torch Graphite

ZrCl4 + 2 CaC2 + Mg Bomb Graphite Torch Graphite

Initially, excess zirconium tetrachloride (ZrCl4) was thought to have caused the ZrO2 in the products. In an effort to remove the excess ZrCl4, a mixture of reactants was sealed in a glass tube and the reaction was initiated with an oxy-

methane torch. The excess ZrCl4 was then transported away from the products by placing the reaction end of the tube in a 500 °C furnace, while exposing the other

sealed end to room temperature; the excess ZrCl4 collected at the cold end of the tube.

Powder X-ray diffraction showed that, after the mixture was washed, the product still

contained ZrO2. A subsequent search of the literature revealed that zirconium tetrachloride decomposes in water forming zirconium oxychloride and hydrochloric acid.11 Zirconium oxychloride is, in turn, soluble in water and, therefore, not the

source of ZrO2.

After determining that the ZrO2 was not produced from ZrCl4, the CaC2 was critically analyzed and found to be the reactant leading to the ZrO2. The CaC2,

purchased commercially through Fluka, is only 80% pure; the other 20% was initially

thought to be excess carbon. A quantitative analysis of a sample of CaC2 confirmed

the presence of graphite as well as calcium oxide. During the reaction, the CaO was

free to react with zirconium metal to form ZrO2 and contaminate the final product.

Quantitative Analysis of Calcium Carbide (CaC2) for Carbon and Calcium Content

To determine the amount of carbon and other impurities in the CaC2 (Fluka),

it was necessary to examine the starting material (400 mesh, ball milled) by

quantitative analysis.

Exactly 0.3613 g of calcium carbide (400 mesh) was left exposed to air in the

back of a hood for three days. (Calcium carbide is highly reactive; it degrades and

produces acetylene gas.) After three days, hydrochloric acid was added to the

remaining powder. The mixture was filtered using 0.1 µm filter paper. The filter

paper and resulting solids were then dried under vacuum. These solids were

equivalent to the amount of elemental carbon present in the 80% CaC2.

The filtrate was treated with 100 mL of a 10 wt.% oxalic acid solution. The pH was adjusted to neutral with ammonium hydroxide. The solution, after the solids were filtered away, was heated to 1000 °C for four hours. After heating, the

remaining solids can be attributed to the mass of CaO present in the Fluka CaC2.

Thermoelectric Composite

A sample of boron carbide was successfully tested as a reinforcement material

for skutterudite thermoelectrics. B4C was blended with a sample of the

thermoelectric material, CoSb3, and hot pressed. The B4C/CoSb3 composite shows sharp edges and is easily cut; in comparison, the CoSb3 is brittle and difficult to machine (Figure 6-6). The inclusion of refractory carbide particles in the CoSb3

matrix dramatically improves the mechanical properties of the material. Through

crack pinning, the B4C particles successfully halt crack propagation. The resulting composite shows an improved resistance to cracking and better processing characteristics.

Figure 6–6. Comparison of newly machined CoSb3 (top) and B4C/CoSb3 composite

(bottom). The CoSb3 sample shows tooling marks and broken edges from the machining process, while the B4C/CoSb3 sample shows no such damage.

Discussion

SSM reactions can be used as a fast and inexpensive route to refractory

carbides. Successful solid-state metathesis routes were found to both B4C and ZrC.

Boron carbide, though contaminated with carbon impurities, was synthesized

via two reaction pathways utilizing several initiation methods; from magnesium

diboride and hexachloroethane initiating with a nichrome wire, oxy-methane torch, or

a carbon arc; or, using a nichrome wire or a torch to initiate the reaction between

MgB2, C2Cl6 and sulfur. Though the products generated were not pure, it is possible

to remove the carbon impurities.22, 23 However, carbon contamination of boron carbide is not entirely undesirable. Due to regular use as nuclear control rods,

mixtures of B4C and graphite are well studied; moderate amounts of graphite only

24-26 slightly reduce the strength of B4C. In addition, because boron carbide is extremely difficult to compact completely and requires high temperatures to do so,

27, 28 nearly all B4C products contain small amounts of carbon (or silicon).

A successful solid-state metathesis route to ZrC was found using a microwave

or torch to initiate the reaction; however, the product was contaminated with ZrO2

(Figures 6-4 and 6-5). In order to synthesize a pure sample of ZrC, pure calcium

10, 29 carbide must be used. A 98% pure sample of CaC2 can be created by reacting elemental calcium and carbon in a high-energy ball mill.30

Conclusions

Nanoparticles of B4C and ZrC were successfully synthesized in a matter of

seconds using solid-state metathesis reactions. The short duration of these reactions

offers an appealing route to the production of refractory carbides for commercial

applications. The oxide contamination of the ZrC product was directly related to the

purity of the CaC2 used in the reaction. Addition of B4C to CoSb3 thermoelectrics

noticeably improved the machinability of the material. In future work, the

composition of the ZrC byproduct could be controlled allowing for its use as a

reinforcement material.

References

(1) Lugscheider, E., O. Knotek, K. Bobzin, M. Maes, and K. Arniz, Investigations

of multifunctional graded zirconium carbide PVD-coatings for the application

on machine parts. Materials Research Society Symposium Proceedings, 2003.

750: p. 399-405.

(2) Mackie, W.A., T. Xie, and P.R. Davis, Transition metal carbide field emitters

for field-emitter array devices and high current applications. Journal of

Vacuum Science & Technology, B: Microelectronics and Nanometer

Structures., 1999. 17: p. 613-619.

(3) Montag, B.W., K.A. Nelson, J.I. Brand, N.M. Boag, and N.P. Platt. Effects of

Purified Metacarborane in Semiconducting Boron Carbide Devices. in 42nd

Midwest Regional Meeting of the American Chemical Society. 2007. Kansas

City, MO, United States.

(4) Wang, Z.-j., Research progress of boron carbon for the bulletproof ceramic.

Guisuanyan Tongbao, 2008. 27: p. 132-135.

(5) Greenwood, N.N. and A. Earnshaw, Chemistry of the Elements. 1984, New

York: Pergamon Press.

(6) Cotton, J. Boron Carbide. [website] 2013 June 11, 2013 [cited 2013

08/10/2013]; Available from:

http://www.azom.com/article.aspx?ArticleID=75.

(7) Vargas-Gonzalez, L., R.F. Speyer, and J. Campbell, Flexural strength,

fracture toughness, and hardness of silicon carbide and boron carbide armor

ceramics. Int. J. Appl. Ceram. Technol., 2010. 7(5): p. 643-651.

(8) Aghajanian, M.K. and A.L. McCormick, Manufacture of ceramic composite

armor materials containing boron carbide,Patent US8128861B1, 2012

(9) Cotton, J. Zirconium Carbide. [website] 2013 June 11, 2013 [cited 2005

10/04/05]; Available from: http://www.azom.com/details.asp?ArticleID=261.

(10) Nartowski, A.M., I.P. Parkin, M. MacKenzie, A.J. Craven, and I. MacLeod,

Solid state metathesis routes to transition metal carbides. Journal of Materials

Chemistry, 1999. 9(6): p. 1275-1281.

(11) Patnaik, P., Zirconium, in Handbook of Inorganic Chemicals. 2003, McGraw-

Hill: New York, NY. p. 995-1006.

(12) Holm, S.R. and R.B. Kaner, Synthesis of carbides and carbide/nitride solid

solutions using displacement reactions. Book of Abstracts, 213th ACS

National Meeting, San Francisco, April 13-17, 1997: p. INOR-850.

(13) Rao, L., E.G. Gillan, and R.B. Kaner, Rapid synthesis of transition-metal

borides by solid-state metathesis. Journal of Materials Research, 1995. 10(2):

p. 353-61.

(14) Janes, R.A., M.A. Low, and R.B. Kaner, Rapid Solid-State Metathesis Routes

to Aluminum Nitride. Inorganic Chemistry, 2003. 42(8): p. 2714-2719.

(15) Cumberland, R.W., R.G. Blair, C.H. Wallace, T.K. Reynolds, and R.B. Kaner,

Thermal Control of Metathesis Reactions Producing GaN and InN. Journal of

Physical Chemistry B, 2001. 105(47): p. 11922-11927.

(16) O'Loughlin, J.L., C.H. Wallace, M.S. Knox, and R.B. Kaner, Rapid Solid-

State Synthesis of Tantalum, Chromium, and Molybdenum Nitrides. Inorganic

Chemistry, 2001. 40(10): p. 2240-2245.

(17) Wallace, C.H., J.L. O'Loughlin, L. Rao, S.-H. Kim, J.R. Heath, M. Nicol, and

R.B. Kaner, Synthesis of crystalline gallium nitride using solid-state

metathesis reactions under high pressures. Koatsuryoku no Kagaku to

Gijutsu, 1998. 7(Proceedings of International Conference--AIRAPT-16 and

HPCJ-38--on High Pressure Science and Technology, 1997): p. 1040-1042.

(18) Blair, R.G., A. Anderson, and R.B. Kaner, A Solid-State Metathesis Route to

MgSiN2. Chemistry of Materials, 2005. 17(8): p. 2155-2161.

(19) Anderson, A.J., R.G. Blair, S.M. Hick, and B. Kaner Richard, Microwave

Initiated Solid-State Metathesis Routes to Li2SiN2. Journal of Materials

Chemistry, 2006. 16(14): p. 1318-1322.

(20) Bonneau, P.R., R.F. Jarvis, Jr., and R.B. Kaner, Metathetical Precursor Route

to Molybdenum Disulfide. Nature (London), 1991. 349: p. 510-512.

(21) Klug, H.P. and L.E. Alexander, X-ray diffraction procedures for

polycrystalline and amorphous materials. 1974, New York: John Wiley &

Sons.

(22) Kurbatov, G.M., A.N. Skachkov, and G.F. Sosnina, Selective elimination of

carbon impurities from boron carbide in the gas phase reaction with

hydrogen. Khimiya Vysokikh Energii, 1987. 21: p. 380-381.

(23) Boyer, J.A. and C.G. Rose, Removal of intergranular graphite from B

carbide., T.C.C. US., Editor. 1943.

(24) Goeddel, W.V., H.K. Lonsdale, and R.A. Meyer, Boron carbide-graphite

nuclear control-rod material: preparation, thermal stability, and irradiation

evaluation., G.D. Corporation, Editor. 1962: San Diego, California.

(25) Shirokane, M., M. Miyauchi, H. Nishizawa, and T. Araki, Compatibility of

Hastelloy XR with boron-carbide and graphite mixture. Koon Gakkaishi,

1985. 11: p. 73-80.

(26) Valin, F., J.D. Lulewicz, and M. Boncoeur, Boron-carbon composites by HIP.

Ceramic Engineering and Science Proceedings, 1993. 14: p. 1184-1189.

(27) Ficai, C., Sintering of boron carbide. Ceramica (Florence), 1958. 13: p. 53-58.

(28) Glasson, D.R. and J.A. Jones, Formation and reactivity of borides, carbides,

and silicides. II. Production and sintering of boron carbide. Journal of

Applied Chemistry, 1969. 19: p. 137-140.

(29) Nartowski, A.M., I.P. Parkin, A.J. Craven, and M. MacKenzie, Rapid, solid-

state metathesis routes to metal carbides. Advanced Materials (Weinheim,

Germany), 1998. 10(10): p. 805-808.

(30) Hick, S.M., C. Griebel, and R.G. Blair, Mechanochemical Synthesis of

Alkaline Earth Carbides and Intercalation Compounds. Inorg. Chem., 2009.

48(5): p. 2333-2338.

Chapter 7: Future Work

Introduction

The preceding chapters outline reactions that provide new routes to nitrides

and carbides for electrical, lighting, and mechanical applications. Some of the

investigated reaction pathways led to unexpected results a few of which are outlined

in this chapter. There are also suggestions for materials properties that can be

explored in the future.

Calcium Silicon Nitride

Cubic-Ca16Si17N34 may have interesting catalytic properties. It has a large unit cell and may be useful as a zeolite-like structure. Oxide-type structures are acidic

(Lewis). Nitride structures would have interesting Lewis base properties. Fewer structures have Lewis base motifs.

Owing to its unique crystal field splitting, Ca16Si17N34 is a potentially interesting host for new phosphors.

Refractive Carbides

The reaction of zirconium chloride and calcium carbide should be reexamined using 98% pure calcium carbide that is easy to prepare in a ball mill. Hopefully, the

amount of ZrO2 will be reduced in the final product and the resulting ZrC would be suitable for use as a reinforcement material.

Luminescence in Ternary Nitrides

The study in Chapter 5 would benefit from complete fluorescence

characterization. The quantum yield and lifetime of the phosphors could be explored

in detail. The electroluminescence of these materials could be studied with

cathodoluminescence.

In addition to the reactants presented in Chapter 5, the following metal oxides

were explored as dopants: cerium oxide (CeO2, Strem, 99.9%), lanthanum oxide

(La2O3, Research Chemicals, A Division of NUCOA Corporation, 99.997%),

praseodymium oxide (Pr6O11, Strem, 99.9%), silver oxide (Ag2O). Misch metal oxide

was made by heating a striker flint (Fisher Scientific) in air at 500 °C to form a mixed

oxide. All of these dopants were used when synthesizing LiSi2N3 and MgSiN2 in an effort to produce a blue emitter. The excitation and emission spectra should be revisited. Even though they did not emit in the blue, the information should not be

discounted. It is also possible that other host lattices (such as Ca16Si17N34) will produce a better phosphor.

During the reaction of 6 Si + 12 P + 13 Li3N + Tm2O3, it was noticed that the reactions with lower amounts of Tm2O3 (0.05 mol and 0.01 mol) produced thulium silicon oxynitride. At increased levels of doping (0.20 mol and 0.10 mol),

LiSi2N3:Tm was favored. The rare-earth silicon oxynitrides are of interest as phosphors as well.1

Additional research is needed to determine the optimum doping level for

maximum emission as well as the level of dopant incorporation. It would be interesting to study the impact of enhanced product crystallinity on luminescence or if

using different size starting materials (e.g. –100 and –325 mesh SiO2) would produce

more crystalline products.

It could also be important to gauge the level of doping after washing and the

synthetic reproducibility. This will be particularly important with low levels of

dopant loading. This could be accomplished using Particle-Induced X-ray Emission

spectrometry (PIXE). This analytical method allows for elemental sensitivity down

to several ppm for lanthanide elements. This type of X-ray analysis is much more

sensitive to the heavier elements (Z >20) than the more commonly used SEM/EDS

analysis, because of the increased ionization potential of an accelerated proton beam

and the increased X-ray production cross-sections. For example, while SEM/EDS

uses the M X-ray lines for identification and analysis of the lanthanides, PIXE has

enough yield of the L X-ray lines to make less ambiguous measurements with lower

limits of detection. Information about the amount of rare earth incorporated into the

lattice could also be gained through Mössbaur Spectroscopy.

EPR studies were conducted during this research, but provided no information

about the oxidation state of the rare earth dopant. Annealing to remove defects in the

lattice may reveal the oxidation state.

100

Li2SiN2 is a known fast ion conductor, by analogy, other alkali and alkaline

earth nitrides (e.g. LiSi2N3 and MgSiN2) most likely are too. A complex impedance

study of the matrix, doped and undoped, could yield interesting properties.

Terbium Silicide

Terbium silicide (TbSi2) is a gray solid first described in detail in the late

2 1950’s. The metallic resistivity and low Schottky barrier of TbSi2 (on n-type doped

silicon) make it a potential candidate for applications such as infrared detectors,

ohmic contacts, magnetoresistive devices, and thermoelectric devices.3-8 The initial

synthesis using a metathesis pathway was discovered serendipitously in an attempt to

produce terbium (Tb) doped magnesium silicon nitride (MgSiN2) by oxy-methane torch initiation.

The following metathesis reaction (Equation 7-1) was chosen as a route to

9, 10 produce Tb doped MgSiN2. However, the reaction could not be initiated with a

nichrome wire; when initiated with an oxy-methane torch, the reaction produced

TbSi2 (Equation 7-1).

SiO2 + Mg3N2 + Li3N + Ca3N2 + Tb4O7 → MgSiN2:Tb + oxides (7-1)

Further research discovered additional routes to TbSi2 utilizing

mechanochemical reaction in a ball mill and a simplified reaction that could be

11 initiated with the oxy-methane torch (Equations 7-2 and 7-3). The reaction of CaSi2

and magnesium silicide (Mg2Si) had previously been used to synthesize transition

12 metal silicides. The addition of magnesium silicide (Mg2Si) to the mixture shown in

101

Equation 7-2, highly crystalline TbSi2 can be produced by torch initiation (Equation

7-3). Reactions initiated with a torch produce the hexagonal phase of TbSi2 and

reactions initiated by ball milling produce the orthorhombic phase of TbSi2.

CaSi2 + Tb4O7 →TbSi2 + CaO (7-2)

CaSi2 + Mg2Si + Tb4O7 → TbSi2 + CaO + MgO (7-3)

Experimental

The following reagents were used in the synthesis of TbSi2: silica (Cab-O-Sil),

magnesium nitride (Afa Aesar, 99.6%), lithium nitride (Cerac, 99.5%, 60 mesh),

calcium nitride (Cerac, 99%, 200 mesh), calcium silicide (Strem, 95%), magnesium

silicide (Aldrich, 99+%, 20 mesh), and terbium oxide (Strem, 99.9%).

Reactions initiated with a torch were performed on relatively small scales (< 1

g amounts). Ball milled reactions were executed on a larger scale (> 5 g). All reactions were prepared in a helium-filled glove box. The reactions were balanced such that all of the alkali or alkali earth metal and oxygen should be consumed in the reaction. Reactions using Equation 7-1 and Equation 7-3 were pressed into a pellet and removed from the glove box; they were then initiated with an oxy-methane flame for 20-120 seconds. Reactions carried out using Equation 7-2 were loaded into an air-tight steel vessel (sealed under a helium atmosphere), loaded into a SPEX 8000D mixer mill, and milled for a total of 8 hours. All products were washed with 6 M HCl to remove the by-products, then washed with water and finally dried in an oven.

The dried products were examined by powder X-ray diffraction (XRD) using a PANalytical Powder X-ray diffractometer using Cu Kα (λ = 1.54 Å) radiation.

102

Scans were conducted with a range of 10° ≤ θ ≤ 100° at 0.0330 degree intervals and

25.1300 second count times.

Results

Powder X-ray diffraction shows that the initial reaction in Equation 7-1

resulted in the high purity production of hexagonal TbSi2. However, after the Li3N on

hand was consumed, the same lot number was received from Cerac at which point all

further experiments were unsuccessful in producing TbSi2. XRD analysis of a small

sample of the initial Li3N revealed a significant amount of lithium oxide (Li2O) impurities. Therefore, some of the Li3N used in the initial reaction (Equation 7-1)

was replaced with Li2O (40% by weight); this mixture (Equation 7-4) successfully produced TbSi2 (Figure 7-1). The reaction with added Li2O3 produces TbSi2 after

heating with the torch for only about 25 seconds; a longer application of heat from the

torch produces Li2SiO3 rather than the desired product. This was unlike the initial

reaction in which TbSi2 production was observed over a much larger range, typically

40-90 seconds.

SiO2 + Mg3N2 + Li3N + Li2O + Ca3N2 + Tb7O4 → TbSi2 + oxides + N2 (7-4)

The salt-balance reaction in Equation 7-2, when initiated using a ball mill,

produced orthorhombic TbSi2 as confirmed by XRD (Figure 7-2). Silicon and silicon dioxide impurities were also present in the product. When Equation 7-2 was product-

balanced, TbSi2 was produced, but contained terbium and silicon dioxide impurities.

The CaSi2 used in this experiment had a 5% iron impurity; it was important to verify that iron silicide impurities were not present in the final product. Powder XRD

103

confirmed that there were no such impurities present. The reaction was also

attempted with the mono-silicide (CaSi); TbSi2 was produced only as a minor product amongst several other compounds (Figure 7-3). The iron impurity in CaSi2 could

potentially be acting as a catalyst for the production of TbSi2.

Figure 7–1. Powder XRD pattern of the product of SiO2 + Mg3N2 + Li3N + Li2O +

Ca3N2 + Tb4O7 initiated by an oxy-methane torch and heated for approximately 25 seconds (overlaid with the JCPDS diffraction pattern number 01-072-2000 for hexagonal TbSi2). The reaction SiO2 + Mg3N2 + Li3N (w/ Li2O impurities) + Ca3N2 +

Tb4O7 produced a similar pattern when heated between a range of 40-90 seconds. The remaining peaks match with the JCPDS diffraction pattern number 00-017-0901 for cubic silicon.

104

Figure 7–2. Powder XRD pattern of the product of CaSi2 + Tb4O7 reacted by mechanochemical initiation in a ball mill (overlaid with the JCPDS diffraction pattern

number 00-013-0383 for orthorhombic TbSi2). The remaining peaks match with the JCPDS diffraction pattern number 01-078-2500 for cubic silicon.

105

Figure 7–3. Powder XRD pattern of the product of CaSi + Tb4O7 initiated by an oxy- methane torch (overlaid with the JCPDS diffraction pattern number 01-072-2000 for

TbSi2). The reaction was unsuccessful in producing TbSi2. The product corresponds to calcium silicate (Ca2SiO4) and unreacted CaSi.

The mixture of CaSi2 + Mg2Si +Tb4O7 (Equation 7-3) was also initiated by a

torch successfully produced hexagonal TbSi2. Samples were heated for

approximately 40 seconds. A salt balanced reaction with a CaSi2:Mg2Si ratio of 1:2

produced the highest purity TbSi2. The XRD of the 1:2 CaSi2:Mg2Si (Figure 7-4) shows a highly crystalline product with minimal silicon impurities.

106

Figure 7–4. Powder XRD pattern of the product of CaSi2 + MgSi2 + Tb4O7 initiated by an oxy-methane torch and heated for 40 seconds (overlaid with the JCPDS

diffraction pattern number 01-072-2000 for hexagonal TbSi2). The remaining peaks match with the JCPDS diffraction pattern numbers 00-017-0901 and 03-065-0387 for cubic silicon and silicon dioxide, respectively. Solid-state metathesis reactions have been shown to be an efficient route to

the production of both the orthorhombic and the hexagonal phases of TbSi2. The

orthorhombic phase can be efficiently produced on large scales by mechanochemical

initiation in a ball mill. A highly crystalline hexagonal phase can be produced by

torch initiation. This is the first report of TbSi2 production by solid-state metathesis reactions.

Future Work

Further optimization of the torch-initiated reactions could result in products of higher purity and crystallinity. The mechanochemical reaction could be optimized to

produced higher purity orthorhombic TbSi2 on a large scale.

107

With the discovery of SSM routes to TbSi2, detailed characterization is needed

to verify the properties of the material. The electrical and magnetic measurements

should be conducted on the metathesis-produced products and compared to literature

values. The materials should be characterized by SQUID (Superconducting Quantum

Interference Device) magnetometry. The thermoelectric properties should also be

examined. The differences between the orthorhombic and hexagonal phases of TbSi2

are poorly defined in the literature. It would be worthwhile to investigate the

properties of the alternate lattice structures.

5 Other rare earth disilicides have similar potential applications to TbSi2. By replacing Tb4O7 in Equations 7-1 and 7-2 with other rare earth oxides, it may be

possible to produce a variety of rare earth disilicides.

Silicon Nitride

Further research is needed to determine optimum annealing times and possible

multiple intermediates. Use of this material for the formation of green bodies and

subsequent conversion to silicon nitride would facilitate study of the mechanical

properties of ceramics derived from this unique precursor.

References

(1) Jeong, H., S.-Y. Seo, and J.H. Shin, Excitation mechanism of visible, Tb3+

photoluminescence from Tb-doped silicon oxynitride. Applied Physics Letters,

2006. 88(16): p. 1910-1912.

108

(2) Perri, J.A., I. Binder, and B. Post, Rare earth metal disilicides. Journal of

Physical Chemistry, 1959. 63: p. 616-619.

(3) Sekizawa, K. and K. Yasukochi, Antiferromagnetism of disilicides of heavy

rare earth metals. Journal of the Physical Society of Japan, 1966. 21: p. 274-

278.

(4) Knapp, J.A. and S.T. Picraux, Epitaxial growth of rare-earth silicides on

(111) silicon. Applied Physics Letters, 1986. 48: p. 466-468.

(5) Pierre, J., E. Siaud, and D. Frachon, Magnetic properties of rare earth

disilicides RSi2. Journal of the Less-Common Metals, 1988. 139: p. 321-329.

(6) Labroo, S. and N. Ali, Magnetism of rare-earth disilicides. Journal of Applied

Physics, 1990. 67: p. 4811-4813.

(7) Sagnes, I., Y. Campidelli, and P.A. Badoz, The Ir/Si/ErSi2 tunable infrared

photoemission sensor. Journal of Electronic Materials, 1994. 23: p. 497-501.

(8) Properties of Metal Silicides, in EMIS Datareview, K. Maex and M. Van

Rossum, Editors. 1995.

(9) Anderson, A.J., R.G. Blair, S.M. Hick, and R.B. Kaner, Metathesis routes to

binary and ternary silicon nitrides. Abstracts of Papers, 229th ACS National

Meeting, San Diego, CA, United States, March 13-17, 2005, 2005: p. INOR-

586.

(10) Blair, R.G., A. Anderson, and R.B. Kaner, A Solid-State Metathesis Route to

MgSiN2. Chemistry of Materials, 2005. 17(8): p. 2155-2161.

109

(11) Rodriguez, M.A., The Synthesis of Refractory Carbides, Rare Earth Silicides,

Metal Nitrides, and Nanocrystalline Silicon through Rapid Solid State

Metathesis, in Chemistry & Biochemistry. 2008, University of California Los

Angeles: Los Angeles, CA. p. 82.

(12) Nartowski, A.M. and I.P. Parkin, Solid state metathesis synthesis of metal

silicides; reactions of calcium and magnesium silicide with metal oxides.

Polyhedron, 2002. 21: p. 187-191.

110