; 0.. IOII|‘ .

$33.51
, A ‘ A 19303.9(. 5.2312?

at

..

3x .

.«N... n“
3 5

mar

.... .5:

Anti}:

99...“... .6.“ 3

., 91;: 7911!.»
'x‘?.\ial}.n
L23... c .
LI}.

 

. .\t||. 5-
:... .53! .( $9.115};
3 ‘ ! .titfnlxxii.
. \1

OK? 1..

ill-.51]

1| Ii.) , A‘t-th
(£3.32 «3:.

51.91.1159:

O ,1}... $1,

it‘l‘ll I:

.\l

.1-

1.. :lvs
’
:rnt .
3:1: . . x ‘ $1
. 4 t. [11.x :n . ‘ . m 15.11)”:
, . . ‘ ..v

.. . a. 3.. .

. 51......

 

 

TPESIS

 

3 LIBRARY
2. O 03 Michigan State
University

 

 

This is to certify that the
dissertation entitled

EXPLORATORY SYNTHESIS IN MOLTEN SALTS:
CHARACTERIZATION, NONLINEAR OPTICAL AND PHASE-
CHANGE PROPERTIES OF NEW CHALCOPHOSPHATE
COMPOUNDS

presented by

IN CHUNG

has been accepted towards fulfillment
of the requirements for the

 

 

Doctoral degree in Chemistry
/
/~
.__{ j /

 

 

Major Pro ssor’s Signature
Xfi ’2 j 20 08

Date

 

MSU is an afiimrative-action, equal-opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:lProj/Acc&Pres/ClRC/DateDue.indd

 

EXPLORATORY SYNTHESIS IN MOLTEN SALTS: CHARACTERIZATION,
NONLINEAR OPTICAL AND PHASE-CHANGE PROPERTIES OF NEW
CHALCOPHOSPHATE COMPOUNDS

By

IN CHUNG

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

CHEMISTRY

2008

ABSTRACT

EXPLORATORY SYNTHESIS IN MOLTEN SALTS: CHARACTERIZATION,
NONLINEAR OPTICAL AND PHASE-CHANGE PROPERTIES OF NEW
CHALCOPHOSPHATE COMPOUNDS

By

IN CHUNG

The polychalcophosphate flux technique has played an important role in
discovery of new chalcophosphate compounds via access to low and intermediate
temperature of 160 — 600 °C. Chalcophosphates are compounds that possess phosphorus
and chalcogen atoms with IP-Q bond, where Q = S, Se, or Te. The structural diversity
within the class of metal chalcophosphates is extensive, and members of this family can
exhibit technologically important ion-exchange, intercalation, magnetic, electrical, and
optical properties.

In the present work exploratory synthesis of chalcophosphate compounds using
polychalcophosphate molten salt method and characterization of physicochemical
properties that are mainly concentrate upon nonlinear optical properties and crystal-glass
phase-change behavior are described. Chapters 2-6 focus on the effort of systematic study
of alkali metal selenophosphate ternary compounds. The first family of this class is one-
dimensional compounds, APSe6 (A = K, Rb, Cs) and A2P28e6 (A =K, Rb). The
compounds adopt noncentrosymmetric polar space group and exhibited remarkably

strong second harmonic generation response in both crystalline and glassy phases. They

also show a reversible crystal-glass phase-change behavior. By coupling
noncentrosymmetry in crystal structure and phase-change behavior, we proposed general
fabrication strategy for optical glassy fibers that yield strong, intrinsic, second-order
nonlinear optical properties. The APSe6 (A=K, Rb) glassy optical fiber exhibited
waveguided second harmonic and difference frequency generation. The second family of
this class is phosphorus-rich, novel molecular complex salts of Rb4P6Sen, Cs4P6Se12,
and CssPsselz. All compounds feature low valent P in two different formal oxidation
states. We attempted rational synthetic conditions to stabilize less oxidized phosphorus
compounds by utilizing excess P in the flux. The polychalcophosphate flux also produced
rare phosphorus telluride compound. The new compound K4P8Te4 featuring P-Te
bonding exhibited reversible crystal-glass phase-change behavior, dissolution process to
photoluminescent solution with hydrazine, and evolution of nanosphere by precipitating
the resulting solution with alcohol. Finally stabilizing Bi metal with selenophosphate flux
at high temperature gave naturally growing flexible nanowires of CssBiP4Se12 that can be
dispersed in alcohol. The compound that adopts the noncentrosymmetric polar space

group exhibited strong near-IR second harmonic generation response.

Copyright by
In Chung
2008

To my wife

Hyun Young

ACKNOWLEDGEMENTS

First of all, I would like to thank my advisor Prof. Kanatzidis for incredible
supports and never giving me up. His research philosophy deeply impressed me to study
solid-state and materials chemistry. Scientific discussion with him was priceless. He
frequently provided deeper insight into my findings and let me progress to higher
standards. One valuable lessons from him was how important interdisciplinary
collaborations are in solid-state chemistry. Thanks to Prof. Kanatzidis and his supports,
he could successfully collaborated with many coworkers in Physics and Electrical
Engineering.

I determined to study solid-state chemistry when I listened to Prof. Jin-Ho Choy’s
class in undergraduate school. His endless passion for new chemistry has been a role
model to me. Dr. Jung-Hwan Do, a former postdoc in our group, greatly helped and
advised me to start my research work and everyday life when I first joined this group. I
learned every basic lab skill and fundamental knowledge including how to solve single
crystal structure. My doctoral degree work could not be finish without my excellent
collaborators. Dr. Jung-Hwan Song in Prof. Freeman group in Northwestern University
never stopped to solve challenging problems for electronic structure calculations. Dr.
Joon I. Jang in Prof. Ketterson group in Northwestern University helped me greatly to
measure nonlinear optical properties. They provided a kind explanation concerting their

fields. I also have to thank M. Gave, C. Canlas, and A. Karst in Prof. Weliky group in

vi

Michigan State University for solid-state NMR measurements. Christos, a labmate,
helped PDF analysis study. I cannot forget all helps from the former and present
members of Prof. Kanatzidis Lab for their valuable discussion.

Finally, I deeply thank my family’s all supports and love. And this work is

dedicated to my dearest wife, Hyun Young.

vii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. xii
LIST OF FIGURES ............................................................................... xv
KEY TO ABBREVIATION ..................................................................... xxiii
Chapter 1

Introduction to Chalcophosphate Compounds and Their Representative Properties. . . .1

1. Introduction to Chalcophosphates Compounds .......................................... l
2. Representation Properties of Chalcophosphate Compounds for Technological
Application ................................................................................... 1 3
2.1 Second Harmonic Generation Response ................................................ 13
2.2 Glass Formation/Reversible Crystal-Glass Phase Transition ......................... 17
3. Characterization of Amorphous Chalcophosphate Compounds ..................... 18
3.1 Solid-state 31P Nuclear Magnetic Resonance(NMR) Spectroscopy ............... 18
3.2 Atomic Pair Distribution Function (PDF) Analysis .................................... 20
Chapter 2
APSe6 (A=K, Rb, and Cs): Polymeric Selenophosphates with Strong Second Harmonic
Generation Response and Reversible Phase-Change Properties .............................. 29
1 . Introduction .................................................................................. 29
2. Experimental Section ...................................................................... 30
2.1 Reagents ..................................................................................... 30
2.2 Synthesis ..................................................................................... 3O
3. Physical Measurements .................................................................... 31
X-ray Powder Diffraction ................................................................. 31
Electron Microscopy ....................................................................... 31
Solid-State UV-vis Spectroscopy ........................................................ 31
Raman Spectroscopy ...................................................................... 31
Infrared (IR) Spectroscopy ................................................................ 31
Differential Thermal Analysis (DTA) ................................................... 32
Solid-state Nuclear Magnetic Resonance (NMR) Spectroscopy ..................... 32
X-ray Crystallography ..................................................................... 32
Nonlinear Optical Property Measurements ............................................. 33

viii

4. Results and Discussion .................................................................... 38
Crystal Structure ........................................................................... 38
Phase-Change Behavior ................................................................... 46
Spectroscopy ................................................................................ 47
Infrared Transmission ..................................................................... 51
Second Harmonic Generation (SHG) Response ....................................... 51

5. Concluding Remarks ....................................................................... 53

Chapter 3

Nonlinear Optical Semiconducting Glassy Fibre Using Noncentrosymmetric Crystal-
Glass Phase-Change Alkali Selenophosphate Materials APSe6 (A=K, Rb).............57

2. Introduction ................................................................................. 58
3. Experimental Section ...................................................................... 60
3.1 Reagents ...................................................................................... 60
3.2 Synthesis .................................................................................... 6O
4. Physical Measurements .................................................................... 61
X-ray Powder Diffraction ................................................................. 61
Electron Microscopy ....................................................................... 61
Solid-State UV-vis Spectroscopy ........................................................ 61
Raman Spectroscopy ...................................................................... 61
Infrared (IR) Spectroscopy ............................................................... 61
Differential Thermal Analysis (DTA) ................................................... 62
Nonlinear Optical Property Measurements .............................................. 62
5. Results and Discussion .................................................................... 63
Second Harmonic Generation Response of APSe6 (A=K, Rb) Bulk Crystal and
Glass .......................................................................................... 65
Glassy Fibre Fabrication ................................................................... 67
Second Harmonic Generation Response and Optical Loss ........................... 72
Difference Frequency Generation Response .......................................... 76
6. Concluding Remarks ....................................................................... 77
Chapter 4
The Helical Polymer 1/00[P28e62']: Strong Second Harmonic Generation Response and
Phase-Change Properties of Its K and Rb Salts .................................................. 81
1 . Introduction .................................................................................. 8 1

ix

Experimental Sectron83

2.1 Reagents ..................................................................................... 83
2.2 Synthesis ..................................................................................... 83
3 Physical Measurements .................................................................... 84
X-ray Powder Diffraction and Pair Distribution Function(PDF) Analysis. . . .84
Electron Microscopy ....................................................................... 84
Solid-State UV-vis Spectroscopy ........................................................ 85
Raman Spectroscopy ....................................................................... 85
Infi'ared (IR) Spectroscopy ................................................................ 85
Differential Thermal Analysis (DTA) ................................................... 85
Solid-state Nuclear Magnetic Resonance (NMR) SpectrOSCOpy ..................... 86
X-ray Crystallography ..................................................................... 86
Nonlinear Optical Property Measurements ............................................. 88
4. Results and Discussion .................................................................... 96
Crystal Structure ............................................................................ 96
Low Temperature Structure .............................................................. 97
Synthesis, Reaction Chemistry and Characterization ................................ 107
Glass Formation, Phase-Change Behavior and Local Structure .................... 111
Infrared Transmission and Nonlinear Optical Properties ............................ 118
Second Harmonic Generation Response of Glassy K2P2$e6 ........................ 120
Thermal Properties of K2P28e6 ......................................................... 122
5. Concluding Remarks ..................................................................... 124
Chapter 5

[P6Se12]4': A Phosphorus-rich Selenophosphate with Low-Valent P Centers. . . . . ..........129

Introduction ................................................................................. 129
Experimental Section ..................................................................... 130
Reagents .................................................................................... 130
Synthesis ................................................................................... 130
Physical Measurements ................................................................... 1 3 1
X-ray Powder Diffraction .............................................................. 131
Electron Microscopy ..................................................................... 131
Solid-State UV-vis Spectroscopy ...................................................... 131
Raman Spectroscopy ..................................................................... 131
Infi'ared (IR) Spectroscopy .............................................................. 131
Differential Thermal Analysis (DTA) ................................................. 132
Solid-state Nuclear Magnetic Resonance (NMR) Spectroscopy .................. 132
X-ray Crystallography ................................................................... 132

4. Results and Discussion ................................................................... 141
Crystal Structure .......................................................................... 141
5. Concluding Remarks ..................................................................... 144
Chapter 6
Low Valent Phosphorus in the Molecular Anions [P536312]; and ,B-[P6Se12]4'. Phase
Change Behavior and Near Infrared Second Harmonic Generation ................................ 147
1 . Introduction ................................................................................ 147
2. Experimental Section ..................................................................... 148
2.1 Reagent .................................................................................... 148
2.2 Synthesis ................................................................................... 148
3. Physical Measurements ................................................................... 149
X-ray Powder Diffraction ............................................................... 149
Electron Microscopy ..................................................................... 149
Solid-State UV-vis Spectroscopy ...................................................... 149
Raman Spectroscopy ..................................................................... 150
Infrared (IR) Spectroscopy .............................................................. 150
Differential Thermal Analysis (DTA) ................................................. 150
X-ray Photoelectron Spectroscopy ..................................................... 151
Solid-state Nuclear Magnetic Resonance (NMR) Spectroscopy ................... 151
X-ray Crystallography ................................................................... 151
Nonlinear Optical Property Measurements ............................................ 152
4. Results and Discussion ................................................................... 161
Crystal Structure. ............................................................................................... 161
Phase-Change Behavior .................................................................. 162
X-ray Photoelectron Spectroscopy .................................................... 166
31P Solid-state NMR Spectroscopy ..................................................... 169
Second Harmonic Generation Response of Crystalline CssPSSelz ................. 171
Second Harmonic Generation Response of Glassy CssPSSelz ..l72
5. Concluding Remarks ..................................................................... 173
Chapter 7
K4P3Te4: A New Phase-Change Compound with P-Te Bonding, Exfoliation, and
Conversion to Photoluminescent Solution ...................................................... 177
1 . Introduction ................................................................................ 177
2. Experimental Section ..................................................................... 178

xi

2. 1 Reagents .................................................................................... 1 78
2.2 Synthesis ................................................................................... 179
3. Physical Measurements .................................................................. 179
X-ray Powder Diffraction ............................................................... 179
Scanning Electron Microscopy .......................................................... 179
Transmission Electron Microscopy .................................................... 180
Solid-State UV-vis Spectroscpoy ...................................................... 180
Therrnogravimetric Analysis (TGA) ................................................... 180
Differential Thermal Analysis (DTA) ................................................. 180
3 IP Solid-State NMR spectroscopy ..................................................... 181
31P Solution-State NMR spectroscopy ................................................. 181
X-ray Crystallography ................................................................... 182
Atomic Pair Distribution Function (PDF) Analysis ............................................. 182
4. Results and Discussions .................................................................. 187
5. Crystal Structure ........................................................................ 187
Phase-Change Behavior .................................................................. 190
Pair Distribution Function Analysis .................................................... 192
Electronic Structure Calculations and Bondings ..................................... 194
Solution Phase Chemistry ............................................................... 201
6. Concluding Remarks ..................................................................... 204
Chapter 8

Emergent Behavior from Weak Interaction between Coordination Complex: Flexible
Nanowires of CssBiP4Se12 with Strong Near IR Second Harmonic Generation

Response ............................................................................................ 209
1. Introduction ................................................................................ 209
2. Experimental Section ..................................................................... 210
2.1 Reagents .................................................................................... 210
2.2 Synthesis ................................................................................... 211
3. Physical Measurements .................................................................. 211
X-ray Powder Diffraction ............................................................... 211
Scanning Electron Microscopy .......................................................... 212
Transmission Electron Microscopy and High Resolution TEM .................... 212
Solid-State UV-vis Spectroscopy ...................................................... 212
Raman Spectroscopy ..................................................................... 212
Infrared (IR) Spectroscopy .............................................................. 213
Differential Thermal Analysis (DTA) ................................................. 213
X-ray Crystallography ................................................................... 213

xii

Electronic Band Structure Calculations ................................................ 214

Nonlinear Optical Property Measurements ............................................ 214
4. Results and Discussion .................................................................. 222
Synthesis and Crystal Structure ......................................................... 222
Nanowires ................................................................................. 223
Electronic Structure Calculations ....................................................... 228
Spectroscopy and Nonlinear Optical Response ....................................... 236
5. Concluding Remarks ..................................................................... 240
Chapter 9
Outlooks ............................................................................................. 244
l. Chalcophosphate compounds as nonlinear optical materials ........................ 245
2. Chalcophosphate compounds as phase-change materials and rich source for glass
compounds .......................................................................................... 245

xiii

LIST OF TABLES

Table 2-1. Crystallographic Refinement Details for KPSeé, RbPSe6, and CsPSeb. . . . . ....35

Table 2-2. Atomic Coordinates (x 104) and Equivalent isotropic displacement parameters
(A2 X 103) for RbPSe6 and CsPSeé. U(eq) Is Defined as One Third of the Trace of the
Orthogonalized Uij tensor ........................................................................ 36

Table 2-3. Anisotropic Displacement Parameters (A2 X 103) for RbPSe6 and CsPSeé.
U(eq) Is Defined as One Third of the Trace of the Orthogonalized Uij- Tensor ......... 37

Table 2-4. Selected Bond Distances (A), Nonbonding Interaction Distances (A), and
Angles (deg) for KPSe6 and RbPSe6 .............................................................. 42

Table 2-5. Selected Bond Distances (A), Nonbonding Interaction Distances (A), and
Angles (deg) for CsPSe6 ......................................................................... 43

Table 4-1. Crystallographic Data and Refinement Details for K2P2Se6 and
szPzSC¢ .............................................................................................. 89

Table 4-2. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (><103 A2) for
K2P28e6 at 298(2)K .................................................................................. 90

Table 4-3. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (><103 A2) for
K2P28e6 at 173(2)K .................................................................................. 91

Table 4-4. Atomic Coordinates (><104) and Equivalent Isotropic Parameters (><103 A2) for
szPZSe6 at 101(2) K ............................................................................... 92

Table 4-5. Selected Bond Distances (A) and Angles (°) for K2P28e6 at 298(2) K. . .....100
Table 4—6. Selected Bond Distances (A) and Angles (°) for K2P28e6 at 173(2) K ........ 101

Table 4-7. K-Se Distances in K2P2Se6 at 173(2) K with Standard Deviation in
Parentheses. The Maximtun Threshold for bond distances is 3.750A. The Coordination
Number of K Atoms Was Determined to be Seven ........................................... 102

Table 4—8. K-Se distances in K2P28e6 at 273(2) K with Standard Deviation in Parentheses.
The Maximum Threshold for bond distances is 3.750A. The Coordination Number of K
Atoms Was Determined to be Six ................................................................ 104

Table 4-9. Rb-Se Distances in szPZSec, at 100 (2) K with Standard Deviation in
parentheses. The Maximum Threshold for Bond Distances is 4.000 A. Rb-Se Distance
Was Not Found up to 4.500 A over the Threshold. The Coordination Number of Rb
Atoms Was Determined to be Ten ............................................................. 105

xiv

Table 4-10. Rb-Se Distances in szPZSeG at 298(2) K with Standard Deviation in
parentheses. The Maximum Threshold for Bond Distances is 4.000 A. Rb-Se Distance
Was Not Found up to 4.500 A over the Threshold. The Coordination Number of Rb
Atoms Was Determined to be Ten ................................................................ 106

Table 5-1. Crystal Data and Structure Refinement for Rb4P6Se12 .......................... 134
Table 5-2. Atomic Coordinates (X 104) and Equivalent Isotropic Displacement Parameters

(A2 X 103) for Rb4P6Se12. U(eq) is Defined as One Third of the Trace of the
Orthogonalized Uij tensor ......................................................................... 135

Table 5-3. Anisotropic Displacement Parameters (A2 X 103) for Rb4PGSe12. The
Anisotropic Displacement Factor Exponent Takes the Form: -27r2[ h2 a*2 U“ +...+2 h k
a*b*U12] .............................................................................................. 136

Table 5-4. Bond Lengths (A) and Angles (°) for Rb4P6Se12 ............................... 138

Table 5-5. Relationship between Observed Selenophosphate Species and the A:P:Se
Ratio ............................................................................................... 139

Table 6-1. Crystallographic Data and Refinement Details for C35P5Se12 and
CS4P68612 ............................................................................................ 154

Table 6-2. Atomic Coordinates (X 104) and Equivalent Isotropic Displacement
Parameters (A2 X 103) for CssPsselz and Cs4P6Se12. U(eq) Is Defined as One Third of the

Trace of the Orthogonalized Uij Tensor ........................................................ 155
Table 6-3. Anisotropic Displacement Parameters (A2 X 103) for CssPsselz and Cs4POSelz.
U(eq) is Defined as One Third of the Trace of the Orthogonalized Uij tensor ........... 156
Table 6-4. Selected Bond Distances (A) and Angles (deg) for CssP5Se12 ................ 159
Table 6-5. Selected Bond Distances (A) and Angles (deg) for CS4P68612 ................ 160
Table 7-1. Crystallographic Data and Refinement Details for K4P8Te4 ................... 184

Table 7-2. Atomic Coordinates (X 104) and Equivalent Isotropic Displacement Parameters
(A2 X 104) for K4P3Te4 at 173(2) K. U(eq) Is Defined as One Third of the Trace of the
Orthogonalized Tensor ............................................................................ 185
Table 7-3. Selected Bond Distances (A) and Angles (°) for K4P3Te4 at 173(2) K. . . . . ...189

Table 8-1. Crystallographic Data and Refinement Details for CsSBiP4Se12 ............... 218

XV

Table 8-2. Atomic Coordinates (X 104) and Equivalent Isotropic Parameters (X 103 A2) for

CssBiP4Se12 at 293(2) K .......................................................................... 219
Table 8-3. Selected Bond Distances (A) for CssBiP4Se12 at 298(2) K ..................... 220
Table 8-4. Selected Angles (°) for CssBiP4Se12 at 298(2) K ................................. 221

xvi

LIST OF FIGURES

Figure 1-1. Typical structure of layered MPQ3 compound. Grey circle denotes M, black
circle phosphorus, and white circle Q atom ....................................................... 2

Figure 1-2. Remarkable structural variety found in the Bi chalcophosphate compounds
obtained by polychalcophosphate flux method. (a) KBIPzS7, (b) K3Bi(PS4)2, (c)
Cs3Bi2(PS4)3, (d) Nao,6Bi1,23PZS6, (e) KBiP2Se6, and (f) ,B-KBiPZSe6. Grey circles
represent alkali metal, crossed circles Bi, black circles P, and open circles S or Se. For ((1),
the Bi-Se bond and Na atom are omitted for clarity ............................................. 7

Figure 1-3. Structural diversity seen in the Sn selenophosphate compounds. (a)
RbsSl’l(PSCs)3, (b) Rb68nZSe4(PSe5)2, (C) Rb3SII(PSC§)(P2866), and (d) Rb4Sn2Ag4(PZSe6)3.
Crossed circles represent Sn atoms, black circles P, open circles Se, large grey circles Rb,
and small grey for (d) Ag ............................................................................ 8

Figure 1-4. (a) View of a long filament of the infinite [CuP3S9],,2"' chain showing the
helical conformation. (b) Local geometry of [Ru(PZSe6)(P38e4)]n3"' showing Ru-P bond.

(c) View of [szpzsrzl' showing its helicity. (d) Tetranuclear cluster of
[M4(Se2)2(PSe4)4]8' (M = Hg, Cd) with a stellar-like core ....................................... 9

Figure 1-5. (a) Interlocked structure of Kbe3(PS4)5 showing the two-dimensional
interpenetrating sublattices (upper) and a single l/00[Yb3(PS4)56'] sublattice (lower)

showing the organization of its pore system. (b) Polyhedral representation of [U4P4Se26]4'
down [100] direction showing the intersecting narrow tunnels. The tunnels in its Rb salt
are accessible to smaller cations via ion exchange. (0) Polyhedral representation of
[U7(PS4)13]l 1' helix in K; 1U(PS4)13. (d) Schematic view of the three equivalent
interpenetrating diamond frameworks in UP4812. The P236 groups are represented by rods
connecting the P atoms of the P286 ligands ...................................................... 11

Figure 1-6. Interrelationships of noncentrosymmetric crystal classes in both Hermann-
Mauguin and Schoenflies symbols ............................................................... 14

Figure 2-1. Structure of RbPSe6 viewed down the b-axis. The Rb+ ions are coordinated
by 12 Se atoms from four selenophosphate chains. Dashed lines indicate weak Sen-Se
interactions (A): Se(3)---Se(5), 3.350(1); Se(3)---Se(6), 3.217(1) ........................... 39

Figure 2-2. (a) Segment of the l/00[PSe6'] anion showing the short Se---Se interactions.
All thermal ellipsoids are presented with 90% probability. (b) View of a l/00[PSe6'] chain
looking down the a-axis, clearly showing its polar character. ................................ 40

xvii

Figure 2-3. (a) Structure of CsPSe6 viewed down the a-axis. Both Cs(l) and Cs(2) are
surrounded by four different [PSe6'] chains. Dashed lines indicate weak Se---Se
interactions (A): Se(2) ---Se(5), 3.280(7); Se(3)---Se(6), 3.160(2); Se(2)-~Se(5'), 3.248(2).
(b) View of a 1/00[PSe6'] chain looking down [1 0 -1] axis. (c) Local geometry of the
l/00[PSe6'] anion in CsPSe6 showing the short Sen-Se interactions. All thermal ellipsoids
are presented with 90% probability ............................................................... 41

Figure 2-4. Differential thermal analysis diagrams of RbPSe6 representing (a) melting in
the first cycle with no crystallization upon cooling and (b) subsequent recrystallization
and complete melting upon heating in the second cycle. * mark represents the
vitrification upon cooling. RbPSe6 is a pristine crystal at A, glass at B and restored

crystal at C ........................................................................................... 44
Figure 2-5. X-ray powder diffraction patterns of pristine (A), glassy (B) and
recrystallized crystal 9(C) .......................................................................... 45
Figure 2-6. Far-IR spectra of RbPSe6 crystal (upper) and glass (below) .................... 48

Figure 2-7. 31P NMR spectra of RbPSe6. Spectra (a) and (c) are for the crystalline and
glassy materials, respectively, and were obtained at ambient temperature and with MAS
frequency of ~11 kHz. Spectrum (b) is for the melt at 350 °C and was obtained under
static conditions ..................................................................................... 49

Figure 2-8. Mid-IR spectrum of glassy KPSe6 showing its excellent IR transmissivity. *
denotes the artifact ................................................................................... 50

Figure 2-9. (a) SHG response of KPSe6 (triangle) and RbPSe6 (square) to AgGaSez over
a wide range of wavelengths. (b) Particle size to SHG intensities diagram of crystalline
RbPSe6 showing type—I phase-matching ......................................................... 52

Figure 3-1. The unit cell viewed down the b-axis ........................................... _. ..64

Figure 3-2. (a) Far-IR/mid-IR/visible absorption spectra of KPSe6 bulk glass showing
wide transparency range. (b) Relative SHG intensities of KPSe6 crystal (triangle),
AgGaSez (circle), and KPSe6 glass (square) ..................................................... 65

Figure 3-3. (a) Representative photograph of an optical fibre showing remarkable
flexibility. (b) A representative SEM image of an fibre showing thickness uniformity at
50.0 pm and surface smoothness .................................................................. 66

Figure 3-4. (a) X-ray diffraction ring patterns of pristine glassy (left) and annealed fibre
(right) confirming their amorphous and crystalline nature, respectively. (b) X-ray
diffraction patterns of the pristine glassy (upper) and annealed fibres (bottom).
Diffraction profiles were regenerated from the ring patterns presented in (a), collected by
STOE 11 single crystal diffractometer (Ag Ka). Note that the Bragg peaks from the

xviii

annealed fibre are successfully indexed, indicative of the restoration of crystal structure
on the fibre. (th) index on the major peak is presented ....................................... 68

Figure 3-5. Raman spectra of KPSe6 crystal, bulk glass, pristine glassy fibre, and
annealed fibre ........................................................................................ 69

Figure 3-6. Pair distribution function (PDF) analysis for the glassy fibre, bulk glass, and
crystalline powders. Fibres with different thickness at d ~ 50 ,um and d ~ 200 ,um were
examined for comparison. Theoretical fit based upon single crystal structure refinement is
plotted as black circles. The first peak at 2.3 A corresponds to interatomic correlations of
P-Se and Se-Se bonds, and the second peak at 3.7 A K---Se and second neighbouring
Se-"Se. Note that PDFs of bulk glass and glassy fibres are very close to that of crystalline
powder, indicating local structural order are significantly preserved in the amorphous
state of bulk powder and fibre, but only lost long-range crystallographic order. This
observation plausibly supports the presence of the intrinsic second-order nonlinear optic
properties on the bulk glass and glassy fibre .................................................... 71

Figure 3- 7. The waveguided SHG response transmitted through 10. 0 mm long KPSeo
glassy fibre 1n a wide range of vis/near IR region... ..73

Figure 3-8. Optical loss of KPSeb glassy fibre at 711.5 nm, estimated by measuring SHG
intensities for various lengths of fibres. Inset: optical loss for 640 to 788 nm SHG signals
at 10.0 (circle), 5.0 (triangle), and 2.5 mm (square) long fibres ................................ 74.

Figure 3-9. The DFG response as a function of Aidlera demonstrating wave-mixing
capability over a wide range of wavelengths ................................................... 75

Figure 3-10. The relative SHG intensities measured from 620 to 805 nm for the pristine
glassy and annealed fibres, representing remarkable enhancement of the SHG response
after heat treatment at 260 °C for 3 min ......................................................... 75

Figure 4-1. Structure of K2P28e6 at 298(2) K. (a) The unit cell viewed down the c-axis.
The thermal ellipsoids with 30% probability are shown. (b) View of a 1/00[P2Se62'] chain

looking down the a-axis. A helix forms by three [P28e6] units and repeats itself at every
18.872 A .............................................................................................. 93

Figure 4-2. (a) Projection view of a 1/,,(,[ste62'] chain slightly slanted in order to show

the staggered anti-conformation of [P2Se6]. The role of short SemSe nonbonding
interactions (dashed line) along a helix segment is shown. Se(1)-~Se(2), 3.559(1) A. (b)
Projection through P-P bonding showing the relationship between short Se-"Se
nonbonding interaction and chain propagation .................................................. 94

Figure 4-3. Structure of K2P28e6 at 173(2) K. (a) The unit cell viewed down the c-axis
showing the superstructure. Crystallographically unique chains are differentiated as A

xix

and B. K atoms are labeled. (b) View of 1/00[P2Se62'] chains with labeling down the a-
axis. K atoms are omitted for clarity .............................................................. 95

Figure 4-4. The coordination environment of K atoms in K2P28e6 is six at 298(2) K (a),
and expands to seven at 173(2) K (b). The coordination environment of Rb atoms in
szPZSe6 at (0) 298(2) K and (d) 100(2) K. It remains unchanged at both temperatures
where Rb is coordinated to ten Se atoms. Large spheres are K or Rb atoms and small ones
are Se atoms. P atoms are omitted for clarity .................................................... 99

Figure 4-5. Solid state 31P MAS NMR spectrum of crystalline K2P28e6 at room
temperature. The asterisk (*) indicates the isotropic peak .................................... 109

Figure 4-6. Differential thermal analysis diagram of a sample of K2P28e6. (a) Heating
curve showing melting at 387 °C in the first heating cycle (solid line) with no
crystallization upon cooling. (b) Exotherrnic crystallization followed by melting upon
heating in the second cycle (dotted line). Asterisk indicates the vitrification event upon
cooling ............................................................................................... 110

Figure 4-7. Solid state UV-vis optical absorption spectra of crystalline and glassy
K2P2$e6 and Rb2P28e6 showing the red shift in absorption edge in the glass samples. The
band gaps are 2.09, 1.98 eV, for the K and 2.32 and 2.10 eV for Rb analogs
respectively ......................................................................................... 1 13

Figure 4-8. Raman spectra of crystalline (upper line) and glassy (lower line) K2P28e6 at
room temperature. The similar but broader features in the spectrum of the glass suggest
the local structure is preserved but long range order is lost ................................... 114

Figure 4-9. Pair distribution function G(r) of the crystalline and glassy K2P2Se6. . . . 1 16

Figure 4-10. Far-IR (line with X)/mid IR (simple line)/vis (line with 0) absorption
spectra of crystalline KzPZSeG. Wide transparent range of crystalline K2P28e6 above the
absorption band at 19.8 pm at far-IR region through mid- IR to 0.596 pm at visible region
is shown” . ..117

Figure 4-11. (a) Particle size to SHG intensities diagram of crystalline K2P28e6 showing
type-I phase-matching. (b) SHG response of KzPZSe6 relative to AgGaSez over a wide
range of wavelengths .............................................................................. 121

Figure 4-12. Temperature variation of the lattice parameters and cell volume for K2P28e6
from 100 to 400 K.............. ....................................................................................... 123

Figure 5-1. (a) [P68e12]4' anion. (b) Structure of Rb4P68e12. (c) Pseudo-one-dimensional
chain of [P6Se12]4' via Se-"Se nonbonding contacts. The thermal ellipsoids are shown

with 50% probability ............................................................................... 137

Figure 5-2. Raman Spectrum of Rb4P68e12 at room temperature ........................... 140

XX

Figure 5-3. 31P Bloch decay NMR spectrum of Rb4POSe12 taken at ambient temperature
on a 9.4 T NMR spectrometer. Acquisition parameters included 5 ps 70’2 pulse, 8000 s
relaxation delay, and 13 kHz magic-angle-spinning frequency. The chemical shift
reference was 85% H3PO4 (0 ppm). Peaks with the same letter are J-coupled. The average
chemical shifts of peaks A-E are in order: 23.9, 47.5, 52.6, 57.6, 58.8 ppm. The J-
couplings of doublets B-E are: 309, 292, 295, 309 Hz ........................................ 140

Figure 6-1. (a) The noncentrosymmetric structure of CssPsselz. The thermal ellipsoids
are shown with 60% probability. (b) [P58612]5' anion. White lines denote long P-Se
bonding at P(3)-Se(1), 2.6606(8) A. Dashed lines indicate short SemSe nonbonding
interaction at P(3)---Se(4), 3.104(2) A. * is defined as equivalent position (-x, l-y,
z) ...................................................................................................... 157

Figure 6-2. Structure of Cs4P68e12 viewed down the (a) c-axis and (b) b-axis. The
thermal ellipsoids are shown with 60% probability. (c) The fl-[P6Se12]4’ anion and (d) a-
[P68e12]4' anion for comparison. ................................................................ 158

Figure 6-3. Differential thermal analysis diagrams of (a) Cs4P6$e12 and (c) CssPSSelz
showing melting in the lst cycle with no crystallization on cooling (upper line). Glass
crystallization is observed in the 2nd heating cycle. Cs4P6$e12 and CssPssen are a pristine
crystal at A, glass at B and restored crystal at C, respectively. X-ray powder diffraction
patterns of (b) Cs4P(,Se12 (d) Cs5P5$e12 pristine (below), glassy (upper) and recrystallized
crystal(middle) ..................................................................................... 163

Figure 6—4. Raman spectra of crystalline (upper trace) and glassy (lower trace) of (a)
CSsP58612 and (b) (3841368612. .................................................................... 164

Figure 6-5. The X-ray photoelectron spectrum, peak fitting, and deconvolution profiles
in the P 2p region of (a) C55P58e12 and (b) Cs4P6Se12 ........................................ 165

Figure 6-6. 31P solid-state NMR spectra of Cs5P5Se12 and Cs4P6$e12 at a 14 kHz MAS
frequency. * denotes spinning side bands and # Cs4P6$elz impurity ....................... 167

Figure 6-7. The electronic absorption spectra of crystalline and glassy (a) Cs5PSSe12 and
(b) (3841368812 ........................................................................................ 168

Figure 6-8. (a) Representative SEM image of CssPsselz glass fiber. (b) X-ray powder
diffraction pattern of CssPSSelz glass fiber showing its amorphous nature .............. 171

Figure 7-1. (a) View of 1/C,O[P3Te44'] chain. P and Te atoms are labeled. Structure of
K4P3Te4 viewed down the (b) a- and (c) b-axes. Dashed line denotes short Te---Te
interaction at Te(1)---Te(2), 3.8016(7) A. Darker large circles represent K atoms. The
thermal ellipsoids are shown with 90 % in (a) and 50 % probability in (b) and
(c) .................................................................................................... 186

xxi

Figure 7-2. Local coordination environment of K(l) and K(2) atoms. (a) K( 1) atom
sandwiched by two 1/00[P3Te44'] chains, and K(2) atom surrounded by four l/...[P8Te44']
chains are shown. Blue broken line denotes ionic bonding of K atoms to P and Te atoms

in (a) and (b). (c) A K(l) and K(2) dimer is shown with K, P, and Te atoms labeled.
Thermal ellipsoids are shown with 90 % probability ......................................... 188

Figure 7-3. DTA diagrams of K4P3Te4 showing melting in the lst cycle with no
crystallization on cooling (upper diagram). Glass crystallization is observed in the 2nd
heating cycle (lower diagram). K4P3Te4 is a pristine crystal at A, glass at B and restored
crystal at C .......................................................................................... 190

Figure 7-4. X-ray powder diffraction patterns of theoretical simulation(A), glass(B) and
recrystallized crystal(C) ........................................................................... 191

Figure 7-5. UV-vis absorption spectra of glass, pristine crystal, and nanosphere
precipitated by MeOH ............................................................................. 191

Figure 7-6. Experimental pair distribution function G(r) of the crystalline (upper) and
glassy (bottom) K4P3TC4. The calculated PDF using the crystallographic coordinates is
shown as red line. Selected atomic correlation distances are indicated. Atomic distances
in the crystallographic model are also shown (inset) .......................................... 193

Figure 7-6. (a) The band structure of K4P8Te4. (b) The projected density of states for p-
orbitals of individual elements (Te and P), and (c) the contour plot of the total charge
density of K4P3Te4 calculated with the sX-LDA method .................................... 195

Figure 7-7. 31P solid-state MAS NMR spectrum of K4P3Te4. * denotes spinning side
bands ................................................................................................ 198

Figure 7-8. (a) TEM image of dispersed K4P3Te4 nanosphere. (b) Selected area electron
diffraction patterns of K4P3Te4 nanosphere. (c) Lattice image on single K4P3Te4
nanosphere by high resolution TEM, indicative of single crystalline nature in nanoscale.
Lattice planes of (012) and (102) are indexed. (d) The representative EDS spectrum of
single K4P3Te4 nanosphere giving a composition of “K4,3P3Te3,5”, reasonably close to
K4P3Te4 ............................................................................................. 199

Figure 7-9. X-ray powder diffraction patterns of (A) isolated K4P3Te4 nanosphere by
methanol and (B) theoretical simulation of bulk K4P3Te4 ..................................... 200

xxii

Figure 7-10. (a) Normalized UV absorbance spectra with respect to the variable
concentration of dissolved K4P3TC4 in hydrazine solution. Solution was consecutively
diluted by 50 % from I to II to III .................................................................................... 200

Figure 7-11. 3'F solution-state NMR spectrum of K4P3Te4/hydrazine solution. * denotes
spinning side bands ................................................................................ 201

Figure 7-12. Room-temperature PL (blue line) under excitation at 380 nm (dashed line)
and absorption (red line) spectra obtained from K4P3Te4/hydrazine. The solid black line
corresponds to the predicted PL afier correcting the reabsorption effect, superimposed by
a symmetric Gaussian fit. The inset exhibits the white emission from K4P3Te4/hydrazine
solution in a cuvet when 355 nm Nd:YAG laser was introduced at the point marked by an
arrow. The nearly saturated deep reddish solution was much diluted to measure
PL .................................................................................................... 203

Figure 8-1. Structure of two crystallographically independent [BiP4Se‘2]5' molecules. Bi,
P, and Se atoms are labeled. Bond distances (A ) are represented .......................... 216

Figure 8-2. (a) Structure of CssBiP4Se12 viewed down the a-axis (down the fiber
direction). Bi, P, and Se atoms are labeled. Large circles are Cs atoms. (b) Polyhedral and
(c) ball and stick representation of a pseudo—one-dimensional [BiP4Selz]5’ chain viewed
the [0 -2 1] direction. Dashed line depicts short nonbonding interaction between Se(5

and Se(16) at 3.478(3) A. Light polyhedron depicts [BiSeo] octahedron and dark [P2866] ’
unit with Cs atoms omitted in (b) ................................................................ 217

Figure 8-3. (a) Representative SEM image of Cs5BiP4Se12 microfibers demonstrating
their flexible texture. (b) TEM image of a bundle of nanowires showing uniform
alignment of individual nanowires. The interval of each nanowire was measured at 2.9
nm, close to crystallographic c-axis of 28.0807(8) A. Inset: selected area electron
diffraction pattern of CssBiP4Se12 nanowires. (c) TEM image, (inset: selected area
electron diffraction pattern of nanowires. (d) EDS analysis (blue dots denote Cs atoms,
orange Bi, grey P, and red Se), and (e) HRTEM image of individual CssBiP4Selz
nanowire dispersed in EtOH. (f) Tangled texture of CssBiP4Se12 nanowire depicting its
striking flexibility. The scale bar corresponds to (a) 500 ,um, (b) 10 nm, (e) 20 nm, and

(e) 5 nm, and (t) 20 nm, respectively .......................................................... 224
Figure 8-4. (a) Differential thermal analysis (DTA) and (b) X-ray powder diffraction
patterns of pristine material and sample obtained after DTA ................................ 225
Figure 8-5. Solid-state optical absorption spectrum of CssBiP4Se1 2 ....................... 227
Figure 8-6. The calculated band structures using LDA (Eg = 1.45 eV). . . . . . .....229

xxiii

Figure 8-7. The calculated band structures using LDA with spin-orbit coupling (SOC)
(Eg = 1.15 eV). .................................................................................... 230

Figure 8-8. The calculated band structures sX-LDA with SOC (Eg = 2.0 eV), shown for
only a small part of the BZ, namely Z-T-Y-I‘) ................................................. 231

Figure 8-9. The projected density of states for p-orbitals of individual elements (Bi, P,
Se) calculated with sX-LDA and SOC ......................................................... 232

Figure 8-10. Three different models, (a), (b), and (0), taken from the original crystal
structure, for the binding energy calculations (see text). The coordinates of two molecular
units are taken from the original crystal structure (Figure 1a) in the b-c plane ((a), (b))
and along the a-axis (c). The neighbor atoms of three molecular [BiP4Se12]5' units (1, II,

and III) around the Cs atom within 4.7 A are indicated in blue circles (d).... . ........234
Figure 8-11. Raman spectrum of CssBiP4Se12 ................................................ 236
Figure 8-12. Far-IR spectrum of CssBiP4Se12 ................................................. 237

Figure 8-13. Far IR/mid IR/vis absorption spectra of CssBiP4Se12. A wide transparency
range of CssBiP4Se12 between 18.7 pm at far-IR region and 0.67 pm at visible region is
shown ................................................................................................ 238

Figure 8-14. SHG response of CssBiP4Se12 relative to AgGaSez at 1004 nm. . . . . . ..238

xxiv

CCD
DTA
TGA
EDS
SEM
TEM

UV/vis

PDF

NLO

SHG

DFG

FLAPW

GGA

sx-LDA

DFT

DOS

CBM

SOC

LIST OF ABBREAVIATIONS

Charge Coupled Device

Differential Thermal Analysis
Therrnogravimetric Analysis

Energy Dispersive Spectroscopy

Scanning Electron Microscopy

Transmission Electron Microscopy
Ultraviolet/visible

Infrared

Pair Distribution Function

Nonlinear Optical

Second Harmonic Generation

Difference Frequency Generation

Full-potential Linearized Augmented Plane Wave
Generalized Gradient Approximation

screened exchange Local Density Approximation
Density Functional Theory

Density of States

Valence Band Maximum

Conduction Band Maximum

Spin Orbital Coupling

XXV

32 Brillouin Zone
)F T Density Functional Theory

’L Photoluminescence

xxvi

Chapter 1

Introduction to Chalcophosphate Compounds and Their Representative

Properties.

1. Introduction to Chalcophosphate Compounds

Solid state materials have provided many technologically important properties
such as superconductivity, giant magnetoresistance, ferroelectricity, photovoltaic energy

conversion and nonlinear optics. Much attention has been paid on the exploration of new

materials as well as studies on existing materials to discover enhanced or novel properties.

To achieve this, exploratory synthesis is a powerful approach because it is almost
impossible to predict the final product of solid state compounds. Phosphorus, in general,
has been known to combine with nearly all elements including chalcogenides. Indeed, the
structural and compositional diversity in known chalcophosphates is prominent and they
show attractive physicochemical properties as described in this chapter. In this regard, the
chalcophosphate class of compounds is the great target for exploratory synthesis to
discover new compound and study their physical and chemical properties.

The chalcophosphate family MxPsz (M = metal; Q = S, Se) has been an
important class because of its interesting properties and great structural variety. In 1960’s,
the first example of chalcophosphate family was synthesized and characterized. The
crystal structure of AIPS4 and BPS4 was determined by Weiss and Schéifer.l Transition

metal thiophosphate MPS; family was discovered by Hahn and Klingen five years later.2

r-‘rrr-rq

The analogous Se chemistry began and showed similar crystal structures and physical
properties in the early 70’s.3 During this era, new chalcophosphates were synthesized by
traditional solid state synthetic methods, that is, direct stoichiometric reaction of elements
or binaries as starting materials at high temperature of ~500 — 900 °C. The compounds
mostly possess the ethane—like [P2Q6]4' or the tetrahedral [PQ4]3' anionic building block,
which are currently regarded as classical chalcophosphate anions. The MPQ3 (Q = S, Se)
familyl‘3’4’5’6’7 shows two dimensional layered structure (Cdlz or CdClz-type structure)

with the [P2Q6]4' anion.

 
 
 

     
 

       
   
 
 

W 33-3 .33.
/I\I .I,|\\/\3II\ |\\II \\I
AIIIS‘QIIKW ijmngll.

 
    

‘0 .1
\II>\<cI

0“

/I"§/\‘I

 
  
  

 

{7|

   

.\7/ 73

    
     
 
 

II /”I\I\’

X; M\\ \
0W 0)
IINI/ \"5‘4 I/\\II/€|WI§\I/\\I7€m|s\ ,
/

Is\<I\ AIIIIXW I\ "Mulls/NW 1i)

    
   

         

 

 

 
     

 

 

Figure 1-1. Typical structure of layered MPQ3 compound. Grey circle denotes M, black

circle phosphorus, and white circle Q atom.

Monovalent (Li, Ag), divalent (Cd, Sn, Hg, Fe, Pb, Fe, V, etc) and trivalent (In)
cationsl 3’4’5‘6’7 have been incorporated in the compounds. Many compounds showed
important properties. For example, SnzPZS6 crystallizes in the noncentrosymmetric (NCS)
space group Pc and undergoes a second order, exothermic phase transition from
ferroelectric to paraelectric (P21/c) at 60 "CS?!1 The compound is promising ferroelectric
material for potential use in nonvolatile memory device.8b SnPZS6 crystallizes in the
acentric R-3 space group represents significant second harmonic generation (SHG)
efficiency.9 CulnPZS6 displays a paraelectric-ferroelectric phase transition near room
temperature.10 In1,33P28e6 shows photoconductivity and is potentially useful for solar
energy conversion.ll Due to layered structure of the MPQ3 family, rich intercalation
chemistry and properties of the resulting compounds have been extensively studied. For
example, anPZSI,12 intercalated with an organic dye displays both strong nonlinear
optical (NLO) properties and spontaneous magnetization up to a high temperature (T 6 ~
40 K). Some members have been studied specifically for their use as cathode materials in
3

secondary lithium batteries.l

The chalcophosphates containing the [PQ4]3' anionic building unit have been
relatively less studied but show notable properties. At 261 °C a—NagPS4 undergoes a
phase transition to ,B-Na3PS4. Above 490 °C there is evidence for a second high
temperature phase; however, this phase has not been fully studied. Both a-NagPS4 and ,6-
Na3PS4 have shown to be good ion conductors.14 CU3PS4 and CU3PS3SC were found to be
active as cathodes for the photoelectrolysis of water.15 InPS4‘6 crystallizes in the NCS
tetragonal space group I-4 which gives rise to nonlinear optical susceptibility and piezo-

coefficients.l7 SHG and optical activity have been detected in Pb3(PS.I)2,18 which

crystallizes in the NCS cubic space group P213. Members of the rare earth metal
compounds, LnPS4 (Ln = Y, La, Ce, Pr, Nd, Sm, Gd, Tb, Er, Tm and Yb),19 have been
studied for luminescence under UV-excitation. TI3PSC420 is a promising material for use
in acoustooptic devices such as optical filters, laser modulators and signal processors.

To that date, the ternary chalcophosphate compounds MPQ3 (M = divalent),
A3PQ4 and MPQ4 (M = trivalent metal) were predominant. They were typically
synthesized by direct combination of the elements and/or binaries at high temperature of
500-800 °C, mainly giving thermodynamic phases. Although a number of interesting
compounds were synthesized using this conventional solid state synthetic method, the
number of compounds and their structural types was still very limited.

It was the introduction of the polychalcophosphate flux method that flourished
the chalcophosphate class to quaternary phase A/M/P/Q (A = alkali metal) in the early
1990’s.2| The polychalcophosphate fluxes are formed by simple in situ fusion of
AzQ/PzQs/Q. When molten [Pny]"' species are solubilized in excess of
polychalcogenide melt, which also continues to serve as the oxidant to dissolve metallic
elements into the flux. The conventional solid state synthetic method requires soaking at
high temperature in order to provide sufficient diffusion between solids for a reaction to
occur. This critically limits the kinetic control or the use of well-defined molecular
building block in tact during reaction. In this regard, the polychalcophosphate flux
method allows intermediate reaction temperature of 300-500 °C which makes possible
the synthesis of not only thermodynamically stable but also metastable or kinetically
stable phases.

In a low temperature flux it is possible to use intact molecular assemblies as

building blocks for incorporation into extended solids just as one does in conventional
solution chemistry. Another important advantage of flux synthesis is the flexibility of the
method. Changes in the composition of the flux can greatly alter the Lewis acid/base
character and in turn affect the outcome of the reaction. A:P:Se ratio determines the
basicity of polychalcophosphate flux AxPsz. Those with a high A:P or AzSe ratio
provide more basic flux (in a Lewis acid/base sense) and the resulting products tend to be
shorter structural fragments or highly oxidized species. Smaller chalcogen-rich flux is
more acidic than bigger chalcogen-rich flux. The trend is similar in alkali metal. For
example, fully oxidized [PS4]3’ is a dominating ligand in thiophosphate compounds while
[P2Se6]4' that possesses a P-P bonding and formally 4+ on P center is widely stabilized in
selenophosphate compounds. Therefore this synthetic technique is best suited for
exploratory synthesis of new phases with complex and unpredictable structural motifs
and composition.

The first compounds prepared from a polychalcophosphate flux were the layered
ABiPZS7 (A = K, Rb)”, A3M(PS4)2 (A = K, Rb, Cs; M = Sb, Bi), Cs3Bi2(PS4)3 and the
nonstoichiometric Nao,6Bil_23P286. 23 The great majority of new chalcophosphates,
reported since this initial discovery, were synthesized via the polychalcophosphate flux
technique. It was shown that with the modest changes in the composition of the AxPyQZ
flux, the structural diversity in the Bi system ranges from the pseudo-helical one-
dimensional NCS of A3Bi(PS4)2 and the layered compounds ABiP2$7 and C33Bi2(PS4)3 to
the dense three dimensional framework of NaofiBimngSa The great multitude of
coordination modes of chalcophosphates ligands gives rise to many possibilities for novel

lattice construction. The thermodynamically metastable acentric compounds ,B-KMPQSeé

(M = Sb, Bi) showed rare phase transition behavior. 24 It originally formed as an
amorphous phase and transformed to metastable ,B-KMPZSe6 phases upon glass transition
temperature. Extended exposure to elevated temperature gave a thermodynamically stable
phase. These compounds are possible candidates for phase-change optical storage and a
three-stage logic gate. SbPS4_xSex grows naturally as a single wall nanotube and its band
gap could by tunable by substitution of S by Se.25

Application of this method to Sn reveals complex chemistry with new structure-
types such as the molecular ASSn(PSe5)3 (A = K, Rb),26 A6SnZSe4(PSe5)2 (A = Rb, Cs),
the one-dimensional Rb38n(PSe5)(P2Se6). 27 The quinary Rb4Sn2Ag4(P2Se6)3 is an
unusual layered compounds with formally Sn2+ and Ag atoms held together by [P28e6]4'
units to feature the rare Sz-dlo interaction between Sn and Ag.28 Other main group metal
chalcophosphate compounds found by flux method included APbPSe4,29 A4Pb(PSe4)2 (A
= Rb, Cs), a- and ,B-Na6Pb3(PS4)4,3° K4In2(PSe5)2(PZSe6),26 and AZZnPZSeé.“

Transition metal chalcophosphate compounds showed interesting properties.
KMPS4 (M = Ni”, Pd”) and NaV0,337(6)PzS634 exfoliated in polar solvent to give infinite
anionic 1/00[MPS4']35 and 1/00[V0,337(6)PZS6'I chains, respectively. Whereas the 3..[NiPS4']
chains undergo autoframentation under the action of the solvent to give rise to
unprecedented discrete crown-shaped [Ni3P3SIZ]3' anions, the 1/00[PdPS4'] and
1/00[V0,337(6)P286'] chains are maintained in solution including complex fluid behavior and
gel formation, respectively. The exfoliated vanadium compound showed induced
transient birefringence and formed sol and gel, which is rare in nonoxidic system such as
chalcogenides. NaszPSlo is soluble in N,N-dimethylformamide and N-methylformamide

to generate 1/00[Nb2PSlo'] nanotube.

 

Figure 1-2. Remarkable structural variety found in the Bi chalcophosphate compounds
obtained by polychalcophosphate flux method. (a) KBiP287, (b) K3BI(PS4)2, (c)
C53Bi2(PS4)3, (d) Nao_5Bil_23P286, (e) KBiP28e6, and (f) fl'KBiPZSCfi. Grey circles
represent alkali metal, crossed circles Bi, black circles P, and open circles S or Se. For ((1),

the Bi—Se bond and Na atom are omitted for clarity.

 

Figure 1-3. Structural diversity seen in the Sn selenophosphate compounds. (a)
RbsSn(PSes)3, (b) Rb63n2364(PSes)2. (C) Rb3Sn(PSe5)(PZSe6), and (d) Rb4Sn2Ag4(P2$ec)3-
Crossed circles represent Sn atoms, black circles P, open circles Se, large grey circles Rb,

and small grey for ((1) Ag.

(a) (b)
Ru-P bond

 

1‘} ‘ i ‘
ll .
II T
II 'I' ,
/* (AA/v I ‘i
I 7 A.” ‘39 I
I.“ I
]l
i
#

 

 

r1941..-
)4—
A
£1

 

 

Figure 1-4. (a) View of a long filament of the infinite [CuP389],,2"' Chain showing the
helical conformation. (b) Local geometry of [Ru(PZSes)(P3Se4)]n3"' showing Ru—P bond.
(c) View of [NbgPZSnT showing its helicity. (d) Tetranuclear cluster of
[M4(Se2)2(PSe4)4]8' (M = Hg, Cd) with a stellar-like core.

The polychalcophosphate flux served as an excellent tool to access to Ru and Os
chemistry, which was elusive for a long time. A3MP58elo (M = Ru, Os) shows significant
departure from the chemistry of any other metal studied in the P/Q systems and adds a
new perspective in the coordination chemistry of [Pny]"' anion.36 There are two
unprecedented features in these compounds: first, the mixed P/Se coordination of the Ru
centers, and second, the stabilization of a novel polymeric [P3Se4]n"‘ anion with P atoms
in two different formal oxidation states of P2+ and PM. The anion binds the Ru2+
exclusively via P atoms which is unique because all other [Pny]"' units coordinate to
the metal centers via the Q atoms.

ATiPs5 (A = K, Rb),37 double helix-containing Aszpzslz,38 K3Cr2(PS4)3,39
A2MP2$e6 (A = K, Rb, Cs; M = Pd, Cd, Hg),3O a novel, tetranuclear cluster
Rb3[M4(Se2)2(PSe4)4I (M = Cd, Hg),4° K2Cu21)..sb,o,41 A2CuP389 (A = K, Rb),
CSzCU2P256, K3CuP287, 42 unusual chiral CszCuP3Sg, 43 KzAuPS4, KAu5P283, 4“
AzAuPZSer, (A = K, Rb),45 K4Pd(PS4)2, Cs4Pd(PSe4)4, CsloPd(PSe4)4, KPdPS4, and
KgPdPZS6323 were the product of chalcophosphate flux reaction with transition metals.

For lanthanide chemistry it is worth mentioning that KYb3(PS4)3 surprisingly
formed an interwoven framework, which is of considerable interest in the field of porous
materials.46 This architecture provides a new design to be targeted for synthesis of so-

called isoreticular materials by those involved in the construction of metal organic
framework and coordination solids. 47 K4Eu(PSe4)z, highly anionic complex
Rb9[Ce(PSe4)4],48 K(RE)PZSe6 (RE = La, Ce, Pr, Gd, Tb),49 K3C6P253,50 NaYPZS6,
NaSszsé, KSmPZS7,51 K3NdP287,52 Cs3Pr5[PS4]653 were also synthesized by flux

method and characterized.

IO

 

Figure 1-5. (a) Interlocked structure of K6Yb3(PS4)5 showing the two—dimensional

interpenetrating sublattices (upper) and a single 1/00[Yb3(PS4)5(*] sublattice (lower)
showing the organization of its pore system. (b) Polyhedral representation of [U4P4Se26]4'
down [100] direction showing the intersecting narrow tunnels. The tunnels in its Rb salt
are accessible to smaller cations via ion exchange. (0) Polyhedral representation of
[U7(PS4)13]"' helix in K11U(PS4)I3. (d) Schematic view of the three equivalent
interpenetrating diamond frameworks in UP4SIZ. The P286 groups are represented by rods

connecting the P atoms of the P286 ligands.

The chemistry of actinide elements with polychalcophosphate flux gave the
unprecedented structure of KzUP3Se9,54 Rb4U4P4Se26,55 UP4SIZ,56 A11U7(PS4),3 (A = K,
Rb) 57, K3Pu(PS4)3, APuPZS7 (A = K, Rb, cs),58 Germs”, and Rb7ThP6$21.59
Rb4U4P4Se26 is a novel three-dimensional, ion-exchange material containing the scarce
U5+ ion. UP4Stz adopts three interwoven polymeric diamond-type frameworks of U“. In
addition to materials described above, polychalcophosphate fluxes have provided many
unusual compounds with main group,60 transition metals,“ lanthanide and actinide
elements.62.

Although many chalcophosphate quaternary compounds have been prepared
thanks to the polychalc0phosphate flux method, ternary compounds AxPyQZ, where A =
alkali element, are limited. These compounds are fundamental interest because they
contain [Pny]"' anions which are not coordinated to main group, transition metals of f-
elements. Examples of these “free” chalcophosphate anionic ligands [Pny]”' (Q = S, Se)
that have been isolated and characterized as alkali metal salts include [PQ4]3',63
IPSC’v4I3"2'5€t3,64 [P25612365 IP2Q6I4',66 [13236912367 [P88e18]6-:68 and [1581014169 They are
all simple discrete molecular anions mainly dependent on the classical [PQ4]3' or [P2Q6]4'
anion. Systematic studies of anionic building block are important. Well-defined discrete
anionic fragments are potential building blocks for new three-dimensional frames or
extended structural motifs. The alkali salts of [Pny n- (Q = S, Se) can also be useful
starting materials for solid-state reactions or coordination chemistry in solution. In
addition, investigations of new [Pny]"' building blocks could help understand the
stability ranges of these in the flux and how the observed solid-state structure motifs

arise.70

12

2. Representative Properties of Chalcophosphate Compounds for Technological

Application

2.1. Second Harmonic Generation Response

Nonlinear optical techniques to produce light of new wavelengths progressed fast
thanks to the discovery of the laser by Maiman in 1960,71 which provided the high-
intensity coherent light essential for such processes. In 1962, Giordamaine72 and Maker

1.73 simultaneously proposed an ingenious method of matching the phase velocities of

et a
the waves at the fundamental and doubled frequencies. In 1965, Wang and Racette
reported the nonlinear process in which two light beams at different wavelengths were
mixed through a nonlinear crystal to produce the difference frequency of the two. This
observation of difference frequency mixing stimulated the research on optical parametric
generation. Giordmaine and Miller introduced a pulsed optical parametric oscillator
(OPO) in the same year.74 Byer et al.75 and Smith et al.76 demonstrated continuous-
wave OPOs independently in 1968. The development of suitable pump lasers in the late
19803 enabled to extend these techniques to shorter time domains of pico- and
femtoseconds. Despite the early introduction of basic principle of optical parametric
generation,77 full potential application should wait over 30 years for the advances of new

optically nonlinear materials and their fabrication at the microstructure level. Optical

parametric oscillators are rapidly extending the range of applications in, for example,

8 9

spectroscopy,7 microscopy,7 medical diagnostics and therapies,80 physical, chemical,
biophysical and biomolecular research,81 environmental monitoring, 8’2 and optical

communications.83

13

Noncentrosymmetric Crystal Classes

Optical Activity (Circular Dichroism)

432
(O)

Enantiomorphic _ Polar (Pyroelectric)
4(84)

422(D4) ”‘(Csi
222(02) 1(01)
2(02)
622(06) 3(c3)

3210.) gigéi

 

 

2 C
23(T) mm ( 2v)
4—2m(DZd)
6(0311) Z3m('rd) ighiaé) )

 

 

 

Piezoelectric, Second-Harmonic Generation

Figure 1-6. Interrelationships of noncentrosymmetric crystal classes in both Hermann-

Mauguin and Schoenflies symbols.84

Much research attention is focused on the development of materials that can
convert nonlinear optical (N LO) frequency from the near- to the mid-infrared regions for
technological application. 85 NLO properties are symmetry-dependent and
noncentrosymmetry of a medium is necessary. For example, all noncentrosymmetric

materials except for 432(0) class can exhibit second harmonic generation (SHG)

l4

materials except for 432(0) class can exhibit second harmonic generation (SHG)
response.86 Figure 1-6 demonstrates a graphical representation of crystallographic inter-
relationships between physical properties and noncentrosymmetric crystal classes.87

When light passes through a medium, its wavelength is unchanged. Although the
light goes through the colored glass and if different wavelengths are already present in
the light, they may be differently absorbed to change the overall color effect. In this
process, however, light that is initially of one wavelength is not converted into light of
another wavelength. This is the realm of linear optics.

Second harmonic generation is a nonlinear optic process. When light of

frequency a) (fundamental frequency) is irradiated onto a nonlinear crystal that adopts

noncentrosymmetric space group, the vibrating electric field of the laser beam excites

oscillation of the electrons bound in the atoms of the nonlinear medium. The electrons re-
emit light at the fundamental frequency but also at a frequency that is double, 2(1), of

incident light. This second harmonic occurs due to the anharmonicity and the asymmetry
of the electric field seen by the electrons in the crystal. Optical frequency conversion in
nonlinear optical crystal by second harmonic generation is an effective means of
generating coherent light in wide spectral range, especially at frequencies where lasers
perform poorly or are unavailable.

The most important characteristics of nonlinear crystals are (a) large SHG
susceptibility )6”, which is a measure of the strength of the nonlinear interaction and is
related to the degree of asymmetry of the electronic potential at the microscopic level,
and (b) ‘phase-matchability’ because nonlinear frequency conversion is efficient only if

the second harmonic waves generated by different atoms interfere constructively, or, at

15

least, do not extinguish each other. Following conditions are also desired for practical use.
(c) The target crystal needs to be grown to a suitable size easily. ((1) They are thermally
stable because very strong laser is used for SHG (e) Hygroscopic materials are less
attractive for commercial use.

The representative nonlinear optic (NLO) crystals for frequency conversion
include perovskite oxide LiNbO3,84a KTiOPO4 (KTP), chalcopyrite AgGan (Q=S, Se)
and ZnGePz, and GaAs.88 Most of the best known and extensively studied NLO
materials are oxides, which are suitable for UV-vis applications but are severely limited
in IR applications due to poor transparency issues caused by M-O stretchings and their
overtones. Compared to oxides, chalcogenide compounds for NLO application are still a
few, despite great industrial demands for IR NLO applications, especially from the
communications industry.

The chalcophosphates are an attractive class in which to search for new nonlinear
optical (NLO) materials because the building units of [PQ4]3' or [P2Q6]4' frequently form
noncentrosymmetric arrangements, which is a structural prerequisite for NLO property.
In addition, chalcogen atoms, S and Se, are much more polarizable than oxygen,
consequently to give bigger second harmonic generation susceptibility. Examples are
include ANDszSn (A=K, Rb, Cs),38 AntI7(1>s.)l3 (A=K, Rb),57 CSZCuP389,43
Introspbljspsr,89 and A3AuPZSe3,90 CssPSSe12,9' APSe6 (A=K, Rb),92 and Azpzsa.93
Their crystalline and glassy phase typically showed good optical transparency in the mid-

IR region. Selenophosphates are transparent above ~600 cm’1 (16.7 11111) through mid-IR
to their band edges, which are typically 1.5~2.5 eV. (0.5~0.8 pm). Currently, the number

of mid-IR crystals with transparency extending well beyond 5 11m and a band edge

16

below 1 pm is very limited and only a few of them have really become commercially

available.94
2.2. Glass F ormation/Reversible Crystal-Glass Phase Transition

Alkali chalcophosphate compounds have shown great tendency to form glassy
phase. In fact, the component elements of this class, e.g. alkali metal, phosphorus and
chalcogen, are known as those which occur most commonly in non-oxidic glass system.95
Such glasses have attracted much industrial interest due to the application for infrared-
transparent optical fibers, 96 reversible conductivity switching devices, 97
semiconductors,98 photoconductors, photoresists 99 and solid electrolytes for battery
applications. '00 Generally glasses are often preferred over crystalline compounds
because of their favorable mechanical and interfacing characteristics, and because their
materials properties can be tailored for specific applications by continuously varying their
composition or processing conditions. These glasses are commonly show moisture/air
sensitivity.

Recently, interest in crystal-glass phase-change materials is growing, as a result
of emerging technologies including commercially available rewritable optical media and
the development of nonvolatile phase—change memory.‘°' The most interesting materials
are those with stoichiometric composition because they can switch between the two states
without complications due to compositional changes. Several crystal-glass reversible
phase-change chalcophosphates are introduced recently and examples are KMP2Se6 (Sb,
Bi),24 APSe6 (A=K, Rb, Cs),76 A2P28e6 (A = K, Rb),77 Cs4P68e12 and CssPSSe12.75
Telluropolyphosphide K4P3Te4 that composed of polyhexaphosphide with P-Te bonding

showed a similar behavior. The details of these materials will be described in chapters 2,

l7

3, 5, and 6 in this Dissertation.

3. Characterization of Amorphous Chalcophosphate Compounds.

Recent development of polychalcophosphate flux method provided new ternary
and quaternary compounds in both crystalline and glass phases. Crystal-glass phase
change materials that can be used for non-volatile memory application84 exhibited
crystalline to glassy states transformation or vice-versa by laser irradiation or thermal
treatment. Chalcogenide glasses have been important in infrared optics and
semiconductor technology. Structural characterization of glassy phase is essential but the
lack of crystallographic long-range order makes it very difficult. For example, X-ray
diffraction method, the most powerful technique for structural analysis of inorganic solids,
only provides the average structure, so very limited information about local structure is
available. Similarly, most of the spectroscopic techniques used are not inherently

quantitative but focus on the ordered system.

3.1. Solid-state 31P Nuclear Magnetic Resonance (N MR) Spectroscopy.

Because solid-state NMR is an elemental selective and inherently quantitative
technique, it is uniquely suited for the structural analysis of glasses. Eckert et al.
conducted the 31P, 6Li, and 7Li magic angle spinning (MAS) NMR study on crystalline
and glassy compositions in the (Li2$)x(Pst)1_x system'oz Various compositions were
obtained by heating a mixture of LiZS:P2$5 = le-x under vacuum at 800-900 °C and

quenched to ice water. NMR spectra were measured for crystalline and glassy products.

18

See scheme for analyzed NMR results. For glassy and crystalline phase, as x varied
different types of anionic ligands were identified. In the crystalline form, [P236]2' (x =
0.50), [P286j4’ (x = 0.67), and 7:3 mixture of [PS4]3' and [Peat-s2 (x = 0.75) were
detected. The results demonstrated that high concentration of LiZS (more basic condition)
favors the simpler tetrahedral [PS4]3' unit while less basic conditions stabilize more
complex, dimeric thiophosphate species. The NMR spectra revealed the tendency of the
different melts to stabilize particular thiophosphate anions. Infinite chains of [PS3]n"',
which is not found in the crystalline state, were formed at x = 0.5. [P287]4' were stabilized
at x=0.67 and when its Li salt crystallized elemental S was released.
Li4P237 (glass) -) Li4PZS6 (crystal) + 1/8 83

Only glass at x = 0.75 agrees with crystal counterpart which possesses [PS4]3', indicating
the preference of tetrahedral unit at basic condition. The results from this study showed
that solid-state NMR can be powerful tool to probe structural relationship between glassy
and crystalline phase, which would be very helpful to develop new phase-change
materials.

3|P solid-state MAS NMR was an important analytical tool in this Dissertation.
In chapter 2, crystal-glass phase change compound RbPSe6 was analyzed by this
technique for its crystalline and glassy forms. High temperature 31F solid-state static
NMR was conducted for its melt at 350 °C. 3 IP solid-state MAS NMR spectroscopy was
also used to analyze K2P28e6, Rb4P6Se12, Cs4P6Se12, CssP5Se12, and K4P8Te4 compounds

in Chapter 3-6.

19

Scheme 1

 

System Glass Crystal
. S\ ’8 - S\ IS\ IS 2
(L'zs)o.s(P285) \ ,P\ / ,P\ ,P\
S S S
. 3‘ [S s\ S 4‘ 3‘ IS 4-
(L'28)0.67(P285)0.33 ’ \ , \/ Sip—R—S
s S s S S
. s\ ,S 3 8‘ ’S 3"
(LIZS)O.75(PZSS)0.25 ’P\ ’P\
S S S S

3.2 Atomic Pair Distribution Function (PDF) Analysis.

A prerequisite to a successful structural refinement is the availability of high
quality single crystals or polycrystalline powders. Because glasses are not crystalline,
they are not fully crystallographically periodic. The PDF analysis has been used for
studying with no long-range order. The PDF technique allow the total drfiiaction, that is
both the Bragg and diffuse scattering, to be analyzed together without bias, revealing the
short and intermediate range order of the material regardless of the degree of disorder.‘03
The Fourier relationship between measurable diffraction intensities and the real-space
arrangement of pairs of atoms are used. Therefore, the technique is both a local and an
extended structure probe and thus a powerful tool that can be used to identify differences
in structural features between the crystalline and glassy version of phase-change
materials. The pair distribution function can be defined directly in real-space in terms of

atomic coordination. It can be also written as a Fourier transform of scattered X-ray or

20

neutron intensities. Since peaks in the PDF come directly from pairs of atoms in the solid,
it is highly intuitive.
The PDF, G(r) is obtained from the experimentally determined total-scattering

structure function, S(Q), by a sine Fourier transform

60) = $1: Q[S(Q)- llsin(Qr)dQ = 4erp<r)— p01

where r is the distance between two atoms, Q is the magnitude of the scattering vector,
S(Q) is the corrected and properly normalized powder diffraction pattern of the material
(also called the structure function), p(r) and p0 are the local and average atomic number
densities, respectively, and r is the radial distance.

In this Dissertation, most compounds synthesized were crystal-glass phase-
change materials that include APSe6 (A=K, Rb, and Cs), A2P28e6 (A=K, Rb), Cs4P68e12,
CssPSSelz, and K4P3Te4. PDF analysis was used to probe the relationship between
crystalline and glassy forms and suggested that the close relationship between two states

possibly explains the facile restoration of crystal structure from amorphous state.

21

References

 

(l) (a) Weiss, A.; Schafer, H. Naturwissenschaften 1960, 47, 495. (b) Weiss, A.;
Schéifer, H. Z. Naturforsch. 1963, 18b, 81.

(2) (a) Hahn, H.; Klingen, W. Naturwissenschaften 1965, 52, 494. (b) Hahn, H.; Ott,
R.; Klingen, W. Z. Anorg. Allg. Chem. 1973, 396, 271.

(3) (a) Garin, J.; Parthe, E. Acta Crystallogr. 1972, 128, 3672 (b) Klingen, W.;
Eulenberger, J.; Hahn, H. Z. Anorg. Allg. Chem. 1973, 401, 97.

(4) (a) Toffoli, P.; Khodadad, P.; Rodier, N. Acta Crystallogr. 1978, B34, 1779. (b)
Jandali, G.; Eulenberger, J.; Hahn, H. Z. Anorg. Allg. Chem. 1978, 447, 105.

(5) (a) Ouvard, G.; Brec, R.; Rouxel, J. Mater. Res. Bull. 1985, 20, 1181 (b) Lee, S.;
Colombet, P.; Ouvard, G.; Brec, R. Inorg. Chem. 1988, 27, 1291. (c) Lee, S.; Colombet,
P.; Ouvard, G.; Brec, R. Mater. Res. Bull. 1986, 21, 917. (d) Durand, E.; Ouvard, G.;
Evain, M.; Bree, R. Inorg. Chem. 1990, 29, 4916.

(6) Carpentier, C. D.; Nitsche, R. Mater. Res. Bull. 1974, 9, 401.

(7) Diehl, R.; Carpentier, C. D. Acta Crystallogr. Sect. B. 1978, 34, 1097.

(8) (a) Scott, B.; Pressprich, M.; Willet, R. D.; Clearly, D. A. J. Solid State Chem.
1992, 96, 294. (b) Amautova, E.; Sviridov, E.; Rogach, E.; Savchenko, E.; Grecov, A.
Integrated F erroelectrics, 1992, I, 147.

(9) Wang, Z.; Willet, R. D.; Laitinen, R. A.; Clearly, D. A. Chem. Mater. 1995, 7,
856.

(10) Simon, A.; Ravez, J.; Maisonneuve, V.; Payen, C.; Cajipe, V. B. Chem. Mater.
1994, 6, 1575.

(11) (a) Etman, M.; Katty, A.; Levy—Clement, C.; Lemasson, P. Mater. Res. Bull.
1982, 17, 579. (b) Katty, A.; Soled, S.; Wold, A. Mater. Res. Bull. 1977, 12, 663.

(12) Lacroix, P. G.; Clement, R.; Nakatani, K.; Zyss, J.; Ledoux, I. Science, 1994, 263,
658.

(13) (a) Thomson, A. H.; Whittingham, M. S. US. Patent 4,049,879 1977. (b) Brec,
R.; Le Mehaute, A. Fr. Patent 7,704,519 1977. (0) Thomson, A. H.; Whittingham, M. S.
Mater. Res. Bull. 1977, 12, 741. (d) Evain, M.; Brec, R.; Whangbo, M.-H. J. Solid State
Chem. 1987, 71, 244.

(14) Jansen, M.; Henseler, U. J. Solid State Chem. 1992, 99, l 10.

22

 

(15) Marzik, J. V.; Hsieh, A. K.; Dwight, K.; Wold, A. J. Solid State Chem. 1983, 49,
43.

(16) Nitsche, R.; Wild, P. Mater. Res. Bull. 1970, 5, 419.

(17) Bridenbaugh, P. M. Mater. Res. Bull. 1973, 8, 1055.

(18) Post, E.; Kramer, V. Mater. Res. Bull. 1984, 19, 1607.

(19) (a) Yampol'akaya, V. V.; Serebrennikov, V. V. Russ. J. Inorg. Chem. 1972, 17,
1771. (b) Wibbelmann, C.; Brockner, W.; Eisenmann, B.; Schafer, H. Z. Naturforsch.
1984, 39a, 190. (c) Palkina, K. K.; Maksimova, S. 1.; Chibiskova, N. T.; Kuvshinova, T.
B.; Volodina, A. N. Inorg. Mater. Engl. Tr. 1984, 20, 1557. (d) Volodina, A. N.;
Koubchinova, T. B.; Maximova, S. 1.; Mouraviev, E. N.; Niazov, C. A.; Tchibiskova, N. 1.
Zh. Neorg. Khim. SSSR 1987, 32, 2899. (e) Le Rolland, B.; Molinie, P.; Colombet, P. C.
R. Acad. Sci. Paris Serie II 1990, 310, 1201. (1) Le Rolland, B.; McMillan, P.; Molinie,
P.; Colombet, P. Eur. J. Solid State Inorg. Chem. 1990, 27, 715. (g) Palkina, K. K.;
Kuvshinova, T. B.; Maksimova, S. 1.; Chibiskova, N. T.; Tirpolskaya, T. A. Inorg. Mater.
Engl. Tr. 1989, 25, 1555. (h) Huang, Z. L.; Victoria, C.; Pilippe M. J. Rare Earth 1999,
17, 6. (i) Huang, Z. L. J. Rare Earth 1998, 16, 167. (j) Huang, Z. L.; Cajipe, V. B.;
LeRolland, B.; Colombet, P.; Schipper, W. J .; Blasse, G. Eur. J. Solid State Inorg. Chem.
1992, 29, 1133.

(20) (a) Gottlieb, M.; Isaacs, T. J.; Feichtner, J. D.; Roland, G. W.; J. Appl. Phys.
1974, 45, 5145. (b) Alkire, R. W.; Vergamini, P. J .; Larson, A. C. Acta Crystallogr. 1984,
C40, 1502. (c) Alkire, R. W.; Larson, A. C.; Vergamini, P. J. Acta Crystallogr. 1985,
C41, 1709.

(21) Kanatzidis, M. G. Curr. Opin. Solid State Mater. Sci. 1997, 2, 139.

(22) McCarthy, T.; Kanatzidis, M. G. Chem. Mater. 1993, 5, 1061.

(23) McCarthy, T.; Kanatzidis, M. G. J. Alloys Compds. 1996, 236, 70.

(24) Breshears, J. D.; Kanatzidis, M. G. J. Am. Chem. Soc. 2000, 122, 7839.

(25) Malliakas, C. D.; Kanatzidis, M. G. J. Am. Chem. Soc. 2006, 128, 6002.

(26) Chondroudis, K.; Kanatzidis, M. G. J. Chem. Soc., Chem. Commun. 1996, 1371.
(27) Chondroudis, K.; Kanatzidis, M. G. J. Solid State Chem. 1998, I36, 79.

(28) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 37, 2848.

(29) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1996, 35, 840.

23

 

 

(30) Aitken, J. A.; Kanatzidis, M. G. Inorg. Chem. 2001, 40, 2938.
(31) Chondroudis, K.; Kanatzidis, M. G. J. Solid State Chem. 1998, 138, 321.

(32) Elder, S. H.; Van der Lee, A.; Bree. R.; Canadell, E. J. Solid State Chem. 1995,
116, 107.

(33) (a) Chondroudis, K.; Kanatzidis, M. G.; Sayettat, J.; Jobic, S.; Brec, R. Inorg.
Chem. 1997, 36, 5859. (b) Coste, S.; Hanko, J.; Bujoli-Doeuff, M.; Louarn, G.; Evain,
M.; Brec, R.; Alonso, B.; Jobic, S, Kanatzidis, M. G. J. Solid State Chem. 2003, 175, 133.

(34) Coste, S.; Gautier, B.; Evain, M.; Bujoli-Doeuff, M.; Brec, R.; Jobic, S.;
Kanatzidis, M. G. Chem. Mater. 2003, I5, 2323.

(35) Sayettat, J.; Bull, L. M.; Gabriel, J.-C. P.; Jobic, S.; Camerel, F .; Marie, A.-M.;
Fourimigue’, M.; Batail, P.; Brec, R.; Inglebert, R.-L. Angew. Chem, Int. Ed. Engl. 1998,
37, 1711.

(36) Chondroudis, K.; Kanatzidis, M. G. Angew. Chem. Int. Ed. Engl. 1997, 36, 1324.
(37) Do. J.; Lee, K.; Yun, H. J. Solid State Chem. 1996, 125, 30.

(38) Gieck, C.; Derstrofi‘, V.; Block, T.; Felser, C.; Regelsky, G.; Jepsen, 0.;
Ksenofontov, V.; Gutlich, P.; Eckert, H.; Tremel, W., Chem. -Eun J. 2004, 10, 382-391.

(39) Derstroff, V.; Ensling, J.; Ksenofontov, V.; Gfitlich, P.; Tremel, W. Z. Anorg.
Allg. Chem. 2002, 628, 1346.

(40) Chondroudis, K.; Kanatzidis, M. G. Chem. Commun. 1997, 401.
(41) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 37, 2098.

(42) Hanko, J. A.; Sayettat, J.; Jobic, S.; Brec. R.; Kanatzidis, M. G. Chem. Mater.
1998, 10, 3040.

(43) Hanko, J. A.; Kanatzidis, M. G. J. Solid State Chem. 2000, 151, 326.
(44) Loken, S; Tremel, W. Eur. J. Inorg. Chem. 1998, 283.

(45) Chondroudis, K.; McCarthy, T. J.; Kanatzidis, M. G. Inorg. Chem. 1996, 35,
3451.

(46) Aitken, J. A.; Kanatzidis, M. G. J. Am. Chem. Soc. 2004, 126, 1 1780.
(47) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J .; O’Keefe, M.; Yaghi, O.
M. Science, 2002, 295 , 469.

24

 

(48) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. Commun. 1998, 1, 55.
(49) Chen, J. H.; Dorhout, P. K.; Ostenson, J. E. Inorg. Chem. 1996, 35, 5627.
(50) Gauthier, G.; Jobic, S.; Brec, R.; Rouxel, J. Inorg. Chem. 1998, 3 7, 2332.
(51) Goh, E.-Y.; Kim, E.-J.; Kim, S.-J. J. Solid State Chem. 2001, 160, 195.
(52) Schleid, T.; Hartenbach, 1.; Komm, T. Z Anorg. Allg. Chem. 2002, 628, 7.
(53) Komm, T.; Schleid, T. Z. Anorg. Allg. Chem. 2004, 630, 712.

(54) Chondroudis, K.; Kanatzidis, M. G. C. R. Acad. Sci. Paris, 1996, 322, 887.
(55) Chondroudis, K.; Kanatzidis, M. G. J. Am. Chem. Soc. 1997, 119, 2574.

(56) Gieck, C.; Rocker, F.; Ksenofontov, V.; Gfitlich, P. Tremel, W. Angew. Chem.
Int. Ed. 2001, 40, 908.

(57) Gieck, C.; Tremel, W. Chem. -Eur. J. 2002, 8, 2980.

(58) Hess, R. F.;Gordon, P. L.; Tait, C. D.; Abney, K. D.; Dorhout, P. K. J. Am. Chem.
Soc. 2002, 124, 1327.

(59) Chan, B. C.; Hess, R. P.; Feng, P. L.; Abney, K. D.; Dorhout, P. K. Inorg. Chem.
2005, 44, 2106.

(60) (a) McCarthy, T. J.; Kanatzidis, M. G. J. Chem. Soc., Chem. Commun. 1994,
1089. (b) McCarthy, T. J .; Hogan, T.; Kannewurf, C. R.; Kanatzidis, M. G. Chem. Mater.
1994, 6, 1072. (c) Chondroudis, K.; McCarthy, T. J.; Kanatzidis, M. G. Inorg. Chem.
1996, 35, 840.

(61) (a) Do. J .; Yun, H. Inorg. Chem. 1996, 35, 3729. (b) Ldken, S.; Tremel, W. Eur.
J. Inorg. Chem. 1998, 283. (c) Caremel, F.; Gabriel, J.-C. P.; Batail, P.; Davidson, P.;
Lemaire, B.; Schmutz, M.; Gulik-Krywicki, T.; Bourgaux, C. Nano Lett. 2002, 2, 403.

(62) (a) Chen, J. H.; Dorhout, P. K. Inorg. Chem. 1995, 34, 5627. (b) Chen, J. H.;
Dorhout, P. K. Ostenson, J. E. Inorg. Chem. 1996, 35, 5627. (c) Gauithier, G.; Jobic, S.;
Brec, R.; Rouxel, J. Inorg. Chem. 1998, 37, 2332.

(63) (a) Mercier, R.; Malugani, J.-P.; Fahys, B.; Robert, G. Acta Crystallogr. Sect. B
1982, 38, 1887. (b) Schafer, H.; Schafer, G.; Weiss, A. Z. Naturforsch. B: Anorg. Chem.
Org. Chem. 1965, 20, 811. (c) Jansen, M.; Henseler, U. J. Solid State Chem. 1992, 99,
110.

25

 

 

(64) Dickerson, C. A.; Fisher, M. J.; Sykora, R. E.; Albrecht-Schmitt, T. E.; Cody, J.
A. Inorg. Chem. 2002, 41, 640.

(65) Brockner, W.; Becker, R.; Eisenmann, B.; Schafer, H. Z. Anorg. Allg. Chem.
1985, 520, 51

(66) Mercier, R.; Malugani, J.-P.; Fahys, B.; Douglade, J.; Robert, G. J. Solid State.
Chem. 1982, 43, 151.

(67) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1995, 34, 5401.
(68) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 3 7, 2582.

(69) Aitken, J. A.; Canlas, C.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem. 2001,
40, 6496.

(70) Canlas, C. G.; Kanatzidis, M. G.; Weliky, D. P. Inorg. Chem. 2003, 42, 3399.
(71) Maiman, T. H. Nature, 1960, 187, 493.
(72) Giordmaine, J. A. Phys. Rev. Lett. 1962, 8, 19.

(73) Maker, P. D.; Terhune, R. W.; Nicenoff, M.; Savage, C. M. Phys. Rev. Lett. 1962,
8, 21.

(74) Giordmaine, J. A.; Miller, R. C. Phys. Rev. Lett. 1965, 14, 973.

(75) Byer, R. L.; Oshman, M. K.; Young, J. F.; Harris, S. E. Appl. Phys. Lett. 1968,
13, 109.

(76) Smith, R. G.; Geusic, J. E.; Levinste, H. J.; Singh, S.; Vanuiter, L. G. J. Appl.
Phys. 1968, 39, 4030.

(77) For early review: Byer, R. L. in Treatise in Quantum Electronics, Rabin, H. and
Tang, C. L. Eds, Academic Press, New York, 1973.

(78) Shen, Y. R. Nature, 1989, 337, 519.

(79) (a) Tang, 8.; Krasieva, T. B.; Chen, Z.; Tempea, G; Tromberg, B. J. J. Biomed.
Opt. 2006, 11, 020501. (b) Denk, W.; Strickler, J. H.; Webb, W. W. Science, 1990, 248,
73.

(80) Filippidis, G.; Kouloumentas, C.; Kapsokalyvas, D.; Voglis, G.; Tavernarakis,
N.; Papazoglou, T. G. J. Phys. D-Appl. Phys. 2005, 38, 2625.

26

 

(81) (a) Zipfel, W. R.; Williams, R. M.; Christie, R.; Nikitin, A. Y.; Hyman, B. T.;
Webb, W. W. Proc. Natl. Acad. Sci. USA 2003, 100, 7075. (b) Campagnola, P. J.; Wei,
M. D.; Lewis, A.; Loew, L. M. Biophys. J. 1999, 77, 3341. (c) Freund, 1.; Deutsch, M.
Opt. Commun. 2001, 196, 325.

(82) (a) Gibbs-Davis, J. M.; Hayes, P. L.; Scheidt, K. A.; Geiger, F. M. J. Am. Chem.
Soc. 2007, 129, 7175.

(83) S. J. B. Yoo, J. Lightwave T echnol. 1996, I4, 955.

(84) Halasyamani, P.; Poeppelmeier, K. R. Chem. Mater. 1998, 10, 2753.

(85) (a) Fejer, M. M. Phys. Today 1994, 40, 25. (b) Ebrahimzadeh, M.; Dunn, M. H.
Science, 1999, 286, 1513. (c) F iore, A.; Berger, V.; Rosencher, B.; Bravetti, R.; Nagle, J.
Nature, 1998, 391, 463.

(86) Glazer, A. M.; Stadnicka, K. Acta Crystallogr. 1989, A45, 234.

(87) Hahn, T. International T ables for Crystallography, D. Reidel Publishing
Company: Drodrecht, 1983.

(88) Rosencher E, Fiore A, Vinter B, Berger V, Bois P, Nagle J, Science 1996, 271 ,
168.

(89) Aitken, J. A.; Marking, G. A.; Evain, M.; lordanidis, L.; Kanatzidis, M. G. J.
Solid State Chem. 2000, 153, 158.

(90) Chondroudis, K.; Hanko, J. A.; Kanatzidis, M. G. Inorg. Chem. 1997, 36, 2623.

(91) Chung, 1.; Jang, J. 1.; Gave, M. A.; Weliky, D. A.; Kanatzidis, M. G. Chem.
Commun. 2007, 4998.

(92) Chung, 1.; Do. J .; Canlas, C. G.; Weliky, D. A.; Kanatzidis, M. G. Inorg. Chem.
2004, 43, 2762.

(93) Chung, 1.; Malliakas, C. D.; Jang, J. 1.; Canlas, C. G.; Weliky, D. P.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2007, 129, 14996.

(94) (a) Petrov, V.; Rotermund, F .; Noack, F. J. Opt. A.; Pure Appl. Opt. 2001, 3, R1.
(b) Isaenko, L.; Yelisseyev, A.; Lobanov, S.; Petrov, V.; Rotermund, F.; Slekys, G.;
Zondy, J.-J. J. App]. Phys. 2002, 91, 9475.

(95) (a) Hilton, A. R.; Jones, C. B.; Btau, M. Phys. Chem. Glasses 1966,7, 108. (b)

Kolomiets, B. T. Phys. Starus Solid; 1964, 7, 359 and 713. (c) Borisova, Z. U.
Semiconductors. Plenum Press, New York 1981, and references therein.

27

 

 

 

(96) Taylor, P. C. Mater. Res. Soc. Bull. 1987, 36.
(97) Ovshinsky, S. R. Phys. Rev. Lett. 1968, 21 , 1450.

(98) (a) Kastner, M. J. Non-Cryst. Solids, 1980, 35-36, 807. (b) Adler, D. J. Non-
Cryst. Solids, 1985, 73, 205. and references therein.

(99) Brandes, R. G.; Laming, F. P.; Pearson, A. D. Appl. Opt. 1970, 9, 1712.

(100) (a) Sahami, 8.; Shea, S. W.; Kennedy, J. H. J. Electrochem. Soc. 1985, 132, 985.
(b) Kennedy, J. H.; Sahami, S.; Shea, S. W.; Zhang, Z. Solid State Ionics, 1986, 18/19,
368.

(101) (a) Katoh, K.; Tsutsumi, T.; Yamada, K.; Terui, G; Niide, Y.; Ochiai, A. Physica
B 2006, 373, 111. (b) Ohta, T. J. Optoelectr. Adv. Mater. 2001, 3, 609. (c) Ohta, T.;
Nishiuchi, K.; Narumi, K.; Kitaoka, Y.; Ishibashi, 11.; Yamada, N.; Kozaki, T. Jpn. J. App.
Phys. Part 1. 2000, 39, 770. (d) We’mic, W.; Pamungkas, A.; Detemple, R.; Steimer, C.;
Blugel, S.; Wuttig, M., Nat. Mater. 2006, 5, 56. (e) Lee, H.; Kim, Y. K.; Kim, D.; Kang, D.
H. IEEE Trans. Magn. 2005, 41, 1034. (i) Kolobov, A. V.; Fons, P.; Frenkel, A. 1.;
Ankudinov, A. L.; Tominaga, J .; Uruga, T. Nat. Mater. 2004, 3, 703.

(102) Francisco, R. H. P.; Tepe, T.; Eckert, H. J. Solid State Chem. 1993, 107, 452.

(103) For review: Billinge, S. J. L.; Kanatzidis, M. G. Chem. Commun. 2004, 749.

28

Chapter2
APSe6 (A=K, Rb, and Cs): Polymeric Selenophosphates with Strong
Second Harmonic Generation Response and Reversible Phase-Change

Properties

1. Introduction

Because complex metal chalcophosphates form readily in an alkali chalcogenide
flux environment,1 it is important to know more about what species are present during
these reaction conditions. A number of free chalcophosphate anionic ligands [Pny]”' (Q
= S, Se) have been isolated and characterized as alkali metal salts, including [PQ4]3',2
[PSe4]3'-28e6,3 [P2361234 [P2Q6]4’,5 [P28e9]2',6 [Pgsmr'] and [stmrzs They are all
discrete molecular anions mainly as they do not coordinate to metals in the structure.
Investigations of new [Pny ’7' building blocks could help understand the stability ranges
of these in the flux and how the observed solid-state structure motifs arise.9 The alkali
salts of [Pny]"' (Q = S, Se) can also be useful starting materials for solid-state reactions
or coordination chemistry in solution.

Here we report on the isolation of a new polymeric [PSe6'] anion crystallized as
alkali salts APSe6 (A = K, Rb, and Cs). These materials undergo an interesting reversible
crystal-glass transition. Polar structural K and Rb salts exhibited remarkably strong
second harmonic generation response over a wide wavelength range. In the infrared
region, those materials are highly transmissive. Their SHG intensities are 60 fold stronger

than that of AgGaSez, which is the top 1R NLO material.

29

2. Experimental Section

2.1. Reagents. The reagents mentioned in this work were used as obtained: K metal
(analytical reagent, Aldrich Chemical Co., Milwaukee, W1); Rb metal (analytical reagent,
Johnson Matthey/AESAR Group, Seabrook, NH); Cs metal (analytical reagent, Johnson
Matthey/AESAR Group, Seabrook, NH); red phosphorus powder, -100 mesh, Morton
Thiokol, 1nc., Danvers, MA; Se (99.9999%; Noranda Advanced Materials, Quebec,
Canada); N,N—dimethylformamide (Spectrum Chemicals, ACS reagent grade); diethyl
ether (Columbus Chemical Industries, Columbus WI, ACS reagent grade, anhydrous).
AZSe (A = K and Rb) starting materials were prepared by reacting stoichiometric amounts
of the elements in liquid ammonia. PZSes was prepared by heating the mixture of P and
Se with a stoichiometric ratio scaling in an evacuated silica tube at 460 °C for 24h.

2.2. Synthesis. Pure APSe6 (A=K, Rb, Cs) were obtained in quantitative yield by heating
a stoichiometric mixture of AZSezP2$e5:Se=l :1 :6 in an evacuated and sealed silica tube at
350 °C for 10 days followed by cooling at a rate of 5°C h'1 to 250 °C, respectively. After
washing with acetonitrile and ether, we obtained pure orange rod-typed single crystals.
Energy dispersive spectroscopy (EDS) analysis of the crystals showed an average
composition of “KPSC62”, “RbngSem”, and “CsPSe6_2”. The glassy phase of APSe6 was
prepared from a stoichiometric mixture of AzsezPZSe5:Se=1:l:6 placed in a silica tube
and melted at 800-900 °C for 1-2 min and subsequent quenching to room temperature.
All compounds are air-sensitive and begin to decompose in a week and are unstable

under NJV-dimethylformamide and H20 but relatively stable under acetonitrile.

30

3. Physical Measurements.

X-ray Powder Diffraction. Analyses were performed using a calibrated CPS 120 INEL
X-ray powder difractometer (Cu Ka radiation) operating at 40 kV/20 mA and equipped
with a position-sensitive detector with a flat sample geometry.

Electron Microscopy. Semiquantitative analyses of the compounds were performed with
a JEOL J SM—35C scanning electron microscope (SEM) equipped with a Tracor Northern
energy dispersive spectroscopy (EDS) detector.

Solid-state UV-vis Spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200—2500 nm region. The instrument
is equipped with an integrating sphere and controlled by a personal computer. BaSO4 was
used as a 100% reflectance standard. The sample was prepared by grinding the crystals to
a powder and spreading it on a compacted surface of the powdered standard material,
preloaded into a sample holder. The reflectance versus wavelength data generated were
used to estimate the band gap of the material by converting reflectance to absorption
data.‘0
Raman Spectroscopy. Raman spectra were recorded on a Holoprobe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation from a HeNe
laser for excitation and a resolution of 4 cm'l. Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 pm. A total of 64 scans
was sufficient to obtain good quality spectra.

Infrared Spectroscopy. FT—lR spectra were recorded as solids in a C51 or KBr matrix.

31

The samples were ground with dry Csl or KBr into a fine powder and pressed into
translucent pellets. The spectra were recorded in the far-1R region (600-100 cm'l, 4 cm'l
resolution) and mid-1R region (500-4000 cm", 4 cm'1 resolution) with the use of a
Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon beam
splitter.

Differential Thermal Analysis (DTA). Experiments were performed on Shimadzu DTA-
50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuwn. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. Samples were heated to 500 °C at
10°C min'l, and after 1 min it was cooled at a rate of —10 °C min'1 to 50 °C. Residues of
the DTA experiments were examined by X-ray powder diffraction. Reproducibility of the
results was checked by running multiple heating/cooling cycles. The melting and
crystallization points were measured at a minimum of the endothermic peak and a
maximum of the exothermic peak.

Solid-state Nuclear Magnetic Resonance (N MR) Spectroscopy. The room temperature
3 IP Solid State NMR spectra for both crystalline and glass RbPSe6 were obtained on a 9.4
T NMR Spectrometer (Varian Infinity Plus) using a double resonance magic angle
spinning (MAS) probe with spinning speeds of 11 and 12 kHz using zirconia rotors of
4mm outer diameter. Bloch decay spectra were taken with a 3.75 microsecond pulse, and
a relaxation delay up to 12,000 3. The spin lattice relaxation time for the crystalline phase
RbPSe6 is 3610 s. The chemical shifts are referenced to 85 % H3PO4 (0 ppm).

X-ray Crystallography. The crystal structure was determined by single-crystal X-ray

diffraction methods. Preliminary examination and data collection were performed on a

32

SMARTl1 platform diffractometer equipped with a 1K CCD area detector using graphite
monochromatized Mo Ka radiation at room temperature. A hemisphere of data was
collected using a narrow-frame method with scan widths of 0.300 in a) and an exposure
time of 30 5 frame". The data were integrated using the SAINT” program. An analytical
absorption correction and empirical absorption correction using the program SADABSll
were performed. The initial positions for all atoms were obtained using direct methods,
and the structures were refined with the full-matrix least-square techniques of the
SHELXTLH crystallographic software package. Crystallographic data are given in Table
2-1. Satisfactory refinements were obtained with the noncentrosymmetric space group,
Pca2. for KPSe6 and RbPSe6 and centrosymmetric space group P2/n. For CsPSeG, Cs(2)
site originally assigned to special position (0 0.5 -0.5) from direct methods. Afier the
absorption correction with SADABS and final refinement followed by weight correction
Cs(2) showed still large x-axis displacement of 0.09228. With the consideration of
relation between C51 and C52 in centrosymmetric space group P2/n, Cs(2) was
introduced into general position. As a result, Cs(2) location was converged into general
position (-0.972682 0.494929 -0.501068) and stabilized at the refined position with small
thermal parameter. But Cs(2) still exhibits the disorder near refined position leading to
relatively high standard deviation of bond lengths between Cs(2) and P and Se atoms.

Nonlinear Optical Property Measurements. We used the frequency-tripled output of a
passive-active mode-locked Nd:YAG laser with a pulse width of about 15 ps and a
repetition rate of 10 Hz to pump an optical parametric amplifier (OPA). The CPA
generates vertically polarized pulses in the ranges 400 ~ 685 nm and 737 ~ 3156 nm. In

order to check the SHG efficiency as a function of the excitation energy, we tuned the

33

wavelength of the incident light from 1000 ~ 2000 nm. In this range, the spectral
bandwidth of the linearly polarized light from the CPA is rather broad, about 2 meV full
width at half maximum. However, the phase space compression phenomena ensures
effective SHG where lower energy portions are exactly compensated by higher parts
thereby satisfying both energy and momentum conservation. The incident laser pulse of
300 ,uJ was focused onto a spot 500 pm in diameter using a 3 cm focal-length lens. The
corresponding incident photon flux is about 10 GW cm'z. The SHG signal was collected
in a reflection geometry from the excitation surface and focused onto a fiber optic bundle.
The output of the fiber optic bundle is coupled to the entrance slit of a Spex Spec-One
500 M spectrometer and detected using a nitrogen-cooled CCD camera. The data

collection time is 20 s.

34

.3 ”ENQENAN $83053 .d u £2 a ._o..~_w\__cm_ -5: w u é a

$2 536
48; :3
Sea; S33
45:3 madam

Nm :0 33:33.32 xEmEésm
Q. \ $3 :53 mn\vow_ \ mam

Noam .. 84 «SN 1 SN
€23 \ 50% £23 \ Sam
3.8m 5.8.6.

33 £3.

:3 x :3 x on; Sod x $3 x £3
a. \ 6;. :3 v \ 528
65.8 8qu
gag: 6:83.:
$2.2 8:830
EEG €9.35:

A2 .2; SE am .2; SE
3.55 8.8m

oommmu comma/m

886
RS
3.92.0
8.936

4: Emfizow
8.8 . 2:“

£23 \ 32:
Sam

:3.

mad x on; x 986
v \ 2:8.me

Soda
$32.:
@883
63%. :
A3 55 i5
and

comma

55588 oSHoEm 82ch<

Nm no 5.3-805500

£23 .38 :3 8065 x

Ao\cv a N-§\u_-

.86 NA: 8065 m .85

@0508 Evans—mom

€225ch \ 28:00—98 2&3: \ :38.
wow .owqfim

< .w \ M K

.-Eo .3.

.22» figs
BE .36 .8960
N \ m< S

as a

< ..o

< .Q

d. .3

95cm 3QO
Emma «seam
«BE—om

m

 

828 Ba ..ommpm swam .é £58 .5558 Ba 23 osaewozsmbo .3 as;

35

Table 2-2. Atomic coordinates (x 104) and equivalent isotropic displacement parameters
(A2 X 103) for RbPSe6 and CsPSe6. U(eq) is defined as one third of the trace of the

orthogonalized Uij tensor.

 

 

x y z U(eq)
RbPSe6
Rb 1322(1) 6900(1) 1 1248(1) 39(1)
P 1660(2) 2753(2) 8181(1) 21(1)
Se(l) -1291(1) 15638(1) 10133(1) 23(1)
Se(2) -1461( 1) 17622( 1) 8467(1) 26(1)
Se(3) -1030(1) 12440(1) 9359(1) 26(1)
Se(4) 851(1) 11790(1) 9914(1) 28(1)
Se(5) 1260(1) 5779(1) 7896(1) 27(1)
Se(6) 1344(1) 655(1) 6822(1) 29(1)
CsPSe6
Cs(l) 2500 273(1) 2500 39(1)
Cs(2) -9727(17) 4949(15) -501 1(18) 42(2)
P(l) -2500 -1517(3) 2500 18(1)
P(2) -2500 3390(3) 2500 20(1)
Se( 1) -124(2) 4257(1) 1829(1) 36(1)
Se(2) -2292(2) -647( 1) 4125(1) 31(1)
Se(3) -88(2) 3238(1) -753(1) 23( 1)
Se(4) 138(2) -2548(1) 2699(1) 23(1)
Se(5) -3 862(2) 1704(1) 4753(1) 22( 1)
Se(6) -1031(2) 2356(1) 3912(1) 22(1)

 

36

Table 2-3. Anisotropic displacement parameters (A2 X 103) for RbPSe6 and CsPSeé.
U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

 

 

U11 U22 U33 U23 U13 U12
RbPSe6
Rb 28(1) 49(1) 39(1) -7(1) -3(1) 7(1)
P 24(1) 18(1) 21(1) 1(1) 0(1) 0(1)
Se(l) 21(1) 26(1) 22(1) 1(1) -1(1) 0(1)
Se(2) 24(1) 25(1) 28(1) 5(1) -1(1) —4(1)
Se(3) 29(1) 23(1) 27(1) -1(1) 4(1) -4(1)
Se(4) 32(1) 28(1) 25(1) 8(1) 4(1) 5(1)
Se(5) 31(1) 20(1) 29(1) 4(1) 6(1) 5(1)
Se(6) 36(1) 23(1) 29(1) -4(1) -7(1) 1(1)
CsPSe6
Cs(l) 35(1) 42(1) 39(1) 0 2(1) 0
Cs(2) 52(6) 37(3) 37(1) 42(2) -6(4) -8(4)
P(l) 16(2) 17(2) 23(2) 0 4(2) 0
P(2) 21(2) 13(2) 20(2) 0 4(2) 0
Se(l) 40(1) 37(1) 31(1) -4(1) 10(1) -20(1)
Se(2) 40(1) 27(1) 27(1) -9(1) 6(1) -2(1)
Se(3) 24(1) 20(1) 26(1) 0(1) 8(1) -2(1)
Se(4) 16(1) 31(1) 22(1) 1(1) 2(1) 4(1)
Se(5) 26(1) 19(1) 22(1) -1(1) 5(1) 0(1)
Se(6) 21(1) 22(1) 24(1) 2(1) 3(1) 3(1)

 

37

4. Results and Discussion

Crystal Structure. KPSe6 (l) and RbPSe6 (2) are isostructural and adopt the
polar space group Pca21. This discussion will concentrate mainly on the Rb+ salt which
consists of infinite one-dimensional chains of 1/(,O[PSe(«,'1 along a-axis and separated by
Rb+ ions, Figure 2. The crystallographically unique anion has PSe4 tetrahedra condensed
with diselenide linkages to give the polymeric l/OO[PSe(5'] chain. The PSe4 tetrahedron is
only slightly distorted from the ideal one with the Se-P-Se angles ranging from 101.46(7)
to 119.79(8)°. The P-Se distances are normal at 2.1529(18) to 2.2994(17) A. The Se-Se
distances range from 2.3534(10) to 2.3859(10) A. The dihedral angles around the bonds
Se(1)-Se(2), Se(2)-Se(3), and Se(3)-Se(4) are -93.53(5)°, 112.41(4)°, and 87.35(5)°. The
largest deviation from 900 gives the longest internal Se(2)-Se(3) distance as reported for
disulfides12 and leads to a bond distance alternation in the Se42' unit. This trend is
generally shown in unligated Se4z' ions.l3 Unusually short SemSe nonbonding
interactions between adjacent chains along the c-axis are observed at 3.217(1) to 3.350
(1) A. Short intra- and interlayer Se---Se interactions much shorter in distance than the
van der Waals radii sum of 3.80 A'4 play a crucial role in influencing the crystal structure.

In fact, other low-dimensional compounds such as NbSe; with its charge density wave

38

phenomenon exhibit similar Se---Se interactions of 3.30 A.'5 In KZSes, short contacts
between the Se52' ions are also observed.'6 The shorter Se---Se interactions in 3 affect the
local geometry of the 1/00[PSe6‘] chains to form pseudo-lamellar packing and relatively

long Se-Se bonds due to the delocalization of electrons through short Se---Se contacts,

b[——>a
C

Figure 2-2b.

 

 

O

g

I

l

I ' O
I

s
...—‘-

 

 

 

 

 

 

 

Figure 2-1. Structure of RbPSe6 viewed down the b-axis. The Rb+ ions are coordinated
by 12 Se atoms from four selenophosphate chains. Dashed lines indicate weak Sen-Se
interactions (A): Se(3)---Se(5), 3.350(1); Se(3)-Se(6), 3.217(1).

39

 

Figure 2-2. (a) Segment of the l/00[PSe6'] anion showing the short Se~--Se interactions.
All thermal ellipsoids are presented with 90% probability. (b) View of a l/(,0[PSe(5’] chain

looking down the a-axis, clearly showing its polar character.

40

 

 

 

 

 

 

Figure 2-3. (a) Structure of CsPSe6 viewed down the a-axis. Both Cs(l) and Cs(2) are
surrounded by four different [PSe6'] chains. Dashed lines indicate weak Se-~-Se
interactions (A): Se(2) ---Se(5), 3.280(7); Se(3)---Se(6), 3.160(2); Se(2)---Se(5'), 3.248(2).
(b) View of a 1/00[PSe6'] chain looking down [1 0 —1] axis. (c) Local geometry of the
l/OO[PSe(5'] anion in CsPSe6 showing the short Se---Se interactions. A11 thermal ellipsoids
are presented with 90% probability

41

Table 2-4. Selected bond distances (A), nonbonding interaction distances (A), and angles

(deg) for KPSe6 and RbPSe6.

 

 

KPSe6 RbPSe6

P-Se(1) 2.296(3) P-Se(1) 2.2994(17)
P-Se(4) 2.244(3) x2 P-Se(4) 2.2521(19) x2
P-Se(5) 2.162(3) x2 P-Se(5) 2.1529(18) x2
P-Se(6) 2.161(3) P-Se(6) 2.1530(17)
Se(1)-Se(2) 2.3473(10) x2 Se(1)—Se(2) 2.3473(10) x2
Se(2)-Se(3) 2.3859(10) Se(2)-Se(3) 2.3859(10)
Se(3)-Se(4) 2.3534(10) Se(3)-Se(4) 2.3534(10)
Se(3)-"Se(5) 3.312(1) Se(3)~’Se(5) 3.350(1)
Se(3)-Se(6) 3.195(1) Se(3)---Se(6) 3.3217(1)
Se(1)-P-Se(4) 105.70(1 1) Se(l )-P-Se(4) 104.41(7)
Se(1)-P-Se(5) 108.68(1 1) Se(] )-P-Se(5) 108.49(7)
Se(1)-P-Se(6) 1 10.62(1 1) Se(1)-P-Se(6) 11 119(8)
Se(4)-P-Se(5) 109.06(1 1) Se(4)-P-Se(5) 1 1030(8)
Se(4)-P-Se(6) 102.02(11) Se(4)-P-Se(6) 101.46(7)
Se(5)-P-Se(6) 1 1970(12) Se(5)-P-Se(6) 119.79(8)
P-Se(l)-Se(2) 9446(8) P-Se(l)-Se(2) 9589(5)
P-Se(4)-Se(3) 9559(8) P-Se(4)-Se(3) 9779(5)
Se(1)-Se(2)-Se(3) 9960(5) Se(l)-Se(2)-Se(3) 101.27(3)
Se(2)-Se(3)-Se(4) 102.54(5) Se(2)-Se(3)-Se(4) 103.92(4)

 

42

Table 2-5. Selected bond distances (A), nonbonding interaction distances (A), and angles

 

(deg) for CSPSB6.
P(1)-Se(2) 2.135(2) X2 Se(2)-P(l)-Se(2*) 117.54(l7)
P(1)-Se(4) 2.240(2) X2 Se(2)-P(1)-Se(4*) 114.18(4)
P(2)-Se(1) 2.138(2) x2 Se(2*)-P(1)—Se(4*) 101.35(4)
P(2)-Se( 1) 2.13 8(2) Se(2)-P(1)-Se(4) 101.35(4)
P(2)-Se(6) 2.262(2) X2 Se(2*)—P(1)—Se(4) 114.17(4)
Se(3)-Se(5) 2.3798(15) Se(4*)-P(1)—Se(4) 108.43(15)
Se(5)-Se(6) 2.3549(16) Se( 1 *)-P(2)~Se( 1) 1 17.92(17)
Se(1*)-P(2)—Se(6*) 102.95(5)
Se(2)---Se(5) 3.280(7) Se(1)—P(2)-Se(6*) 1 1206(5)
Se(3)-"Se(6) 3.160(2) Se(1*)-P(2)«Se(6) 1 1206(5)
Se(l)—P(2)-Se(6) 102.95(5)
Se(6*)—P(2)—Se(6) 108.91(15)
Se(4**)-Se(3)—Se(5*) 9820(5)
P(1)-Se(4)—Se(3 * *) 9856(5)
Se(6)-Se(5)-Se(3*) 100.20(5)
P(2)-Se(6)-Se(5) 9781(6)

 

Symmetry transformations used to generate equivalent atoms:

* (-x-l/2, y, -z+1/2)
** (x+l,y, 2+1)

43

 

 

 

 

1 (a)

3% exo f
A 20 r endo 1
> A
3 10': - r 4
< “..-,

5 °€ , .
.10 L ‘ -_‘ ~--».._.: ‘11ng ‘1 :
.20; 257 315 ‘ :
100_7 200 300 460 500
Temperature (°C)
4W“‘ J“_-+.
. (b) 237 1

301 ’11 exo f

20 1‘ ’1' end0$ i
’>‘ i ‘ c
510 B
< ‘1’" .- -v 1 ; L
la 01 ‘ " 1 1

LB i‘ 1 i
'101 w . il l t
1 ‘7 “~41 .- w
out 257 321 l
100 200 300 400 500

Temperature (’6)

Figure 2-4. Differential thermal analysis diagrams of RbPSe6 representing (a) melting in
the first cycle with no crystallization upon cooling and (b) subsequent recrystallization
and complete melting upon heating in the second cycle. * mark represents the

vitrification upon cooling. RbPSe6 is a pristine crystal at A, glass at B and restored crystal
at C.

44

 

03

O

O

O
1

 

Intensity (a. u.)

 

 

 

 

 

10 ' i ‘20 ‘30 l l 1401 i ‘ 150‘ l 160
2theta, deg

Figure 2-5. X-ray powder diffraction patterns of pristine (A), glassy (B) and
recrystallized crystal 9(C)

45

CsPSe6 (3) is a centrosymmetric compound also with l/00[PSe6'] chains but of a
different conformation than in the K+ and Rb+ analogs, Figure 2-3a. The differences in
conformation are seen clearly in projection of the two chains in Figure 2-2b and 2-3b.
Adjacent l/(,O[PSe6’] chains along the b-axis are related by a 2-fold symmetry operation
along the a-axis. P-Se distances range from 2.135(2) to 2.262(2) A. The Se-Se distances
with alteration are similar to the Rb+ analog at 2.3549(16) to 2.3798(15) A. The dihedral
angle around the Se(2)-Se(3) bond of -121.06(1)° contributes to its longer distance than
the external ones. Short Se---Se interactions are observed at 3.160(2) to 3.280(7) A, even
shorter than those of 2. Interchain interactions generate a pseudo-lamellar packing and

influence the conformation of Se42' linkage to form straight lines of Se(l)-Se(2)---Se(5)

and Se(4)-Se(3)---Se(6) connections and possibly maximize the p7: orbital overlap, Figure
2-3c.

The [PSe2(Se4)'] anion is a rare example of a free standing polymeric
chalcophosphate with no coordinating metals. The only other reported polymeric anion is
l/OO[P3Se4'] ion which was found bound to Ru or Os atoms in K3RuP58elo,l7
Rb3RuPSSe10, KgOsPSSelo, and 126303135860.”

Phase-Change Behavior. According to differential thermal analysis (DTA)
performed at a rate of 10 °C min'l, RbPSe6 (2)19 and CsPSe6 (3) melt congruently at 315
and 303 °C, respectively, Figure 2-4, and form dark red glasses rather than crystals upon
cooling.20 Crystallization is only achieved on heating. The glasses recrystallize
exothermically at 237 °C for 2 and 206 °C for 3 upon subsequent heating followed by
melting at 321 and 310 °C for 2 and for 3, respectively. RbPSe6 showed vitrification at

257 °C on cooling to room temperature. The XRD patterns afier recrystallization are the

46'

same as those of pristine 2 and 3 indicating full recovery of the original crystal structure.
Recrystallization and vitrification were repeatedly observed by repeating the DTA cycles.
This clearly suggests the reversible glass-crystalline transition behavior. We may
speculate that the facile crystallization of these glasses in the solid state suggest that the
glass structure is somewhat related to the crystal structure, and given the polar character
of the Rb salt, we may expect a similar polar nature for the precursor glass.

Spectroscopy. The solid-state diffuse reflectance UV/vis spectra of both
crystalline and glassy APSe6 (A = K, Rb, and Cs) reveal sharp absorption edges. The
band gaps of crystalline phases 1, 2, and 3 revealed 2.16, 2.18, and 2.16 eV while those
of glassy counterparts 4, 5, and 6 showed 1.82, 1.91, and 1.71 eV, respectively, which are
consistent with their respective orange and dark red colors. It is noteworthy that red shifts
in the absorption edge are typical in glassy phases. Glass formation in compounds with
extended solid-state structures generally produces a large number of defects and mid-gap
states (often called band tailing), which leads to lower band gaps compared to
corresponding crystalline phases.21 These materials show congruent melting and
reversible glass-crystalline transitions accompanied with band gap shifts. This
phenomenon is a key aspect of optical storage systems based on phase-change
materials.22 Furthermore, the polar K+ and Rb+ salts enable them to be explored for non-
linear Optical properties.

The far IR spectra of RbPSe6 display absorption peaks at 482(5), 424(3), 386(s),
356(m), 247(m), 228(w) and 164(w) cm", Figure 2-6. The peaks at 247 cm" are
attributed to Se-Se stretchings'23 Those at higher energies are diagnostic of P-Se

vibrations.24 The far IR spectra of glassy RbPSe6 display much broader and weaker peaks

47

at 507(w), 420(w), 385(sh) and 252(sh) cm]. A comparison with crystalline RbPS€6
suggests that the building block [PSe6'] unit is substantially intact in the glass but lacks
long-range order. The bridging 8e42' chains seem to be severely disordered in the glassy

phase according to the very weak and broad Se-Se stretching vibrations.

 

 

 

 

100
90-
E? 80-
(U. .
j 70-
o\ .
05 60-
0
C .
g 50?
1% 40-
g .
204
10 .

 

400 i 300 ' 200
Wavelength, nm

Figure 2-6. Far-1R spectra of RbPSe6 crystal (upper) and glass (below).

31P NMR provided further insight into bonding and structure, Figure 2-7. For
crystalline RbPSe6 under magic angle spinning (MAS), the major peak has an isotropic
chemical shift of 4.3 ppm and a full-width at half-maximum (F WHM) linewidth of 0.5
ppm. This shifi is ~40 ppm higher than a typical shift of [P28e9]4’, a comparable discrete

unit with bridging Se.9 In addition, for discrete selenophosphate units, the 511 -533

48

chemical shift anisotropy principal value difference correlates inversely with the degree
of local symmetry about P, with ~65 ppm values observed for discrete tetrahedral [PSe4]3'
units and ~200 ppm values observed for [P2Se9]4' units.9 The ~115 ppm value observed
for RbPSe6 fits within this correlation, with the pseudo C2 symmetry of P in l/00[PSe(,']
chains intermediate between the high and low symmetries of P in [PSe4]3' and [P28e9]4'

anions, respectively.

(a)

 

 

 

 

I I 1

ppm 0 -50

Figure 2-7. 3'P NMR spectra of RbPSe6. Spectra (a) and (c) are for the crystalline and
glassy materials, respectively, and were obtained at ambient temperature and with MAS
frequency of ~11 kHz. Spectrum (b) is for the melt at 350 °C and was obtained under

static conditions.

49

In a 350 0C melt, the static 3 1P spectrum has a single resonance with -21 ppm
peak chemical shift and 4 ppm FWHM linewidth. The shift is typical of PSe4-type
bonding,9 and the narrow linewidth is diagnostic of rapidly tumbling small molecular
species such as [PSedmn’ ring-based molecules which would result from de-
polymerization of 1/00[PSe6'] chains. In the RbPSe6 glassy phase, there is a broad 31P
MAS signal centered at —25 ppm. The similarity of the average shifts of the melt and
glass spectra and the broad width of the glass spectrum are consistent with a frozen melt
model for the glass with associated conformational, packing, and possibly molecular
heterogeneity. Overall, the NMR, crystallographic, and DTA data suggest that

crystallization and 1/00[PSe6'] chain formation are coupled processes.

 

 

 

 

100 . 1 u
80 - _
o\°
8
t: 60 - ~
.9
g 40 ~ 3
C
Q
|._
20 - -
O 1 1 1
4000 3000 2000 1000

Wavenumber, cm'1

Figure 2-8. Mid-IR spectrum of glassy KPSe6 showing its excellent IR transmissivity. *
denotes the artifact.

50

Infrared Transmission. KPSe6 and RbPSe6 exhibit wide optical transparency
ranging from long wave 1R to near IR/visible light, Figure 2-8. The mid-1R transmittance
spectrum showed little absorption from 4000 cm'I (2 pm) to 500 cm"1 (20 pm). There is
no light absorption below the band-gap transition suggesting uninterrupted light
transmission in the compounds. Above 20 pm in the far-1R region, the compounds

exhibited a complex set of absorptions, consistent with their far-1R spectrum. Optical
transparency is a key feature for materials aimed at NLO applications. For example, the
important NLO material for IR applications, AgGaSez,25 shows LWIR transmission up to
17 pm.

Second Harmonic Generation Response. The polar, noncentrosymmetric chain
structure of l/OO[PSe(,‘j composed of easily polarizable P and Se atoms linked by covalent
bonding and (Se4)2' chain can produce large optical nonlinearity. SHG measurements

(126’27 with an IR light source

were performed using a modified Kurtz powder metho
ranging from 1000 nm to 2000 nm. SHG intensities of crystalline KPSe6 and RbPSe6

were directly compared with that of AgGaSez powder. All samples were prepared in a
similar fashion and the same particle size range of 45.5:t7.5 pm was measured and
compared. KPSe6 and RbPSe6 generated strong double frequency signals from the
fundamental idler beam. The SHG intensity of KPSe6 and RbPSe6 showed a maximum at
771 and 790nm, respectively, which is ~60 times larger than that of in the same
wavelength. At shorter wavelengths the KPSe6 and RbPSe6 outperforms the chalcopyrite
material by over 200 fold, Figure 2-9a. These results demonstrate that the crystalline
KPSe6 and RbPSe6 are very promising in IR NLO application. For comparison, the

absolute nonlinear optical susceptibility at 2.12pm of AgGaSeg and LiNbO3 is 67.7i13

51

and 29.1:t5.2 pm/V, respectively.28

 

 

 

 

 

 

a
( ) 250~ . . RbPSe6
g ' . v KPSe5
,9 200 - v '
9
a:
a)
0 150 .
I
(D
g 100 ~ I
as .
— l
a) 0
o: 50 - ' :
660 680 700 720 740 760 780 800
Wavelength, nm
(b) 80000
>.. . l—I——1
‘c’ 70000L
.92
2
a: .
LL]
0 60000 -
I I—I—-1
U)
(D
.5 50000 -
LU 1—I—1—I—1
<1)
0: 1
40000 1 1 . 1 . 1 1 1 4 a

 

 

 

20 40 60 80 100
Particle Size, um

Figure 2-9. (a) SHG response of KPSe6 (triangle) and RbPSe6 (square) to AgGaSez over
a wide range of wavelengths. (b) Particle size to SHG intensities diagram of crystalline

RbPSe6 showing type-1 phase-matching.

5T?

The SHG intensity of RbPSe6 increased with the particle size and reached a
plateau, Figure 2-9b. In principle, phase-matchable samples reach maximum intensity
and then for larger average particle sizes the intensity is size-independent because of the
existence of a phase-matching direction in the sample.27 In this regard, crystalline RbPSe6
is type 1 phase-matchable and eminently suitable for consideration in applications. KPSe6
also exhibited similar behavior. These results suggest that APSe6 (A=K, Rb) is of special
interest for the middle and deep infrared (IR) applications due to its large nonlinear

optical coefficients and high transmission in the 1R region.

5. Concluding Remarks.

New compounds APSe6 (A=K, Rb, Cs) composed of a rare one dimensional
chain of l/OO[PSe(5']. When molten the compounds can be quenched to a glassy state and
exhibit reversible crystal-glass phase-change behavior. The optical absorption edge for
the glassy phase is red shifted. IR spectroscopy analysis showed that the main building
units are largely intact in the glassy form. This explains the rapid restoration of crystal
structure from the corresponding amorphous phase. The glass forming behavior of the
compounds makes them potentially valuable for producing 1R optical fiber.

Polar one-dimensional compounds APSe6 (A=K, Rb) showed a strong SHG

response with wide optical transmittivity from 2 11m to 20 pm. The compounds are type-

] phase-matchable with a response that is over 60 times larger than that of top performing
NLO material AgGaSez. The remarkably stronger response is attributed to the highly

polar one dimensional structure consisting of Se42' unit that bonds to P atoms in contrast

53

to the weakly polar chalcopyrite structure of AGaSez.

54

References

(1) Kanatzidis, M. G. Curr. Opin. Solid State & Mater. Sci. 1997, 2, 139-149.

(2) (a) Jansen, M.; Henseler, U. J. Solid State Chem. 1992, 99, 1 10-119. (b) Mercier,
R.; Malugani, J. P.; Fahys, B.; Robert, G.; Douglade, J. Acta Crystallogr. B 1982, 38,
1887-1890. (c) Schafer, H.; Schafer, G.; Weiss, A. Z. Naturforsch., B: Chem. Sci. 1965, B
20, 81 1.

(3) Dickerson, C. A.; Fisher, M. J .; Sykora, R. E.; Albrecht-Schmitt, T. E.; Cody, J.
A. Inorg. Chem. 2002, 41, 640—642.

(4) Brockner, W.; Becker, R.; Eisenmann, B.; Schafer, H. Z Anorg. Allg. Chem.
1985, 520, 51-58.

(5) Mercier, R.; Malugani, J. P.; Fahys, B.; Douglade, J.; Robert, G. J. Solid State
Chem. 1982, 43, 151-162.

(6) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1995, 34, 5401-5402.

(7) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 37, 2582-2584.

(8) Aitken, J. A.; Canlas, C.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem. 2001,
40, 6496-6498.

(9) Canlas, C. G.; Kanatzidis, M. G.; Weliky, D. P. Inorg. Chem. 2003, 42, 3399-
3405.

(10) (a) Tandon, S. P.; Gupta, J. P. Phys. Status Solidi 1970, 38, 363-367. (b)
Wendlandt, W. W.; Hecht, H. G. Reflectance spectroscopy; Interscience Publishers: New
York, 1966. (c) Kortfim, G. Reflectance spectroscopy. Principles, methods, applications;
Springer: Berlin, Heidelberg, New York, 1969.

(11) Sll/IART, SAINT, SHELXT L: Data Collection and Processing Software for the
SA/MRT—CCD System; Siemens Analytical X-ray Instruments 1nc.: Madison, WI, 1997.

(12) Hordvik, A. Acta Chem. Scand. 1966, 20, 1885-1891.
(13) Brese, N. B.; Randall, C. R.; Ibers, J. A. Inorg. Chem. 1988, 27, 940-943.
(14) Bondi, A. J. Phys. Chem. 1964, 68, 441-451.

(15) (a) Canadell, B.; Rachidi, I. E. 1.; Pouget, J. P.; Gressier, P.; Meerschaut, A.;
Rouxel, J .; Jung, D.; Evain, M.; Whangbo, M. H. Inorg. Chem. 1990, 29, 1401-1407. (b)

Monceau, P. Electronic Properties of Inorganic Quasi-one-dimensional Compounds; D.
Reidel; Dordrecht, Holland; Boston, MA, U.S.A., 1985.

55

(l6) Miiller, V.; Frenzen, G.; Dehnicke, K.; Fenske, D. Z. Naturforsch., B: Chem. Sci.
1992, 47, 205-210.

(17) Chondroudis, K.; Kanatzidis, M. G. Angew. Chem, Int. Ed. Eng]. 1997, 36,
1324-1326.

(18) Chung, I; Kanatzidis, M. G. Unpulished results. Rb3RuPSSe10, K3OsP58elo, and
RbgOsPSSelo are isostructural to Rb3RuP5Se10.

(19) The thermal behavior of KPSe6 showed the same patterns as that of RbPSe6.

(20) The powder X-ray diffraction (XRD) patterns after each DTA cycle showed that
amorphous glasses had formed.

(21) Dhingra, S.; Kanatzidis, M. G. Science 1992, 258, 1769-1772.

(22) (a) Feinleib, J.; DeneufviJ; Moss, S. C.; R., O. S. App]. Phys. Lett. 1971, 18,
254-257. (b) Maeda, Y.; Andoh, H.; lkuta, 1.; Minemura, 1H. J. App]. Phys. 1988, 64,
1715-1719.

(23) Wachhold, M.; Kanatzidis, M. G. J. Am. Chem. Soc. 1999, 121, 4189-4195.
(24) Chondroudis, K.; Kanatzidis, M. G. Chem. Commun. 1996, 1371-1372.

(25) Nikogosyan, D. N. Nonlinear optical crystals: a complete survey; Springer-
Science: New York, 2005.

(26) Dougherty, 1. P.; Kurtz, s. K. J. App]. Crystallogr. 1976, 9, 145-158.
(27) Kurtz, s. K.; Perry, T. T. .1. Appl. Phys. 1968,39, 3798-3813.
(28) Choy, M. M.; Byer, R. L. Phys. Rev. B 1976, 14, 1693-1706.

56

Chapter 3

Nonlinear Optical Semiconducting Glassy Fibre Using
N oncentrosym metric Crystal-Glass Phase-Change Alkali

Selenophosphate Materials APSe6 (A = K, Rb)

1. Summary

Second harmonic generation (SHG) is usually forbidden in glass due to the
presence of inversion symmetry at the macroscopic level. There have been numerous
efforts1 to induce SHG in glasses via poling using thermalz, optical3, and electron beam4
irradiation; however the procedures are complex and/or expensive, and the resulting SHG
is too small for practical applications and often non-permanent. Here we report general
fabrication strategy for glassy optical fibres that yields a strong, intrinsic, second-order
nonlinear optical (NLO) response based on materials which undergo a crystal-glass
phase-change behaviour and adopt noncentrosymmetric space group. APSe6 (A = K, Rb)
compounds exhibited a strong SHG response in both the bulk crystalline and glassy
phases. The optical glassy fibres of APSe6 studied here were manually drawn at low
temperature of 230-280 °C, to lengths of 10 or more centimetres. These as-prepared
fibres exhibit coherent, continuously tuneable visible/near infrared SHG and difference
frequency generation (DFG) responses over a wide range of wavelengths in centimetre

long fibres. In addition they exhibit low-losses, have smooth surfaces and show

57

considerable mechanical flexibility. The observed SHG response was significantly
enhanced simply by annealing the glassy fibre at 260 °C for a few minutes. We believe
that our approach can be widely applied to prepare NLO glassy fibres by utilizing

materials that undergo a phase-transition.

2. Introduction

Obtaining new coherent light sources involving different frequencies with the
potential for tuneability is of great importance. Relatively few lasers prove to be practical
and commercially viable and they typically generate a single or at best a few optical
frequencies. Frequency conversion by a nonlinear optical (NLO) crystal is an effective
way of producing coherent light at frequencies where lasers perform poorly or are
unavailable. For example, when two incoming frequencies co] and (02 are introduced in an
NLO medium, they interact to produce four distinct frequencies: 20)] and 2602 by second
harmonic generation (SHG), together with (an i 602) by sum and difference frequency
generation (SFG and DFGS). Demand for widely tuneable, coherent 1R laser sources is
emerging. Examples include: high rate (broadband) information transfer for
telecommunications and intemet6 (1.3 — 1.6 pm via wavelength-division-multiplexed
(WDM) all-optical networks7); sensing for organic and inorganic molecules8 (including
chemical warfare agentsg, biohazardslo, explosives” and polluteslz); and medical
applications (in the range 2 - 12 pm)”.

To be maximally useful NLO materials should possess: phase-matchability, high

second-order nonlinearity, wide optical transparency, and thermal stability; in addition,

58

some applications require fibre or thin film forms. Many inorganic oxides and polymer
NLO materials strongly absorb mid-1R light; in addition polymers show poor thermal
stability and low damage thresholds. A crucial challenge facing many inorganic NLO
crystals is the difficulty of fabricating fibres and films. Thanks to their 1R transparency,
high index of refraction (2.2~3.5)'4, and the excellent forrnability, chalcogenide glass is a
promising contender for low-loss infrared optical fibre or planar waveguides; however it
ordinarily lacks a second-order optical nonlinearity since a glass has inversion symmetry
on a macroscopic scale. This latter fact restricts the application of glassy silica fibre, the
backbone of modern telecommunication systems, to passive devices.

Our pair distribution function (PDF) analysis and Raman spectroscopic studies on
the crystal-glass phase-change materials K2P28e615 and K1.xbeSb588l6 revealed that their
glassy phases still largely preserved the basic building blocks that define the crystal
structure but only lost long-range crystallographic order, in contrast to a common glass
like silica. For the noncentrosymmetric compounds in this class, e.g. KZPZSeé15 and
CssPSSen”, we surprisingly observed significant innate SHG response from the as-
prepared bulk glassy powders, plausibly by virtue of the noncentrosymmetric fragments
partially intact in the glassy form of the phase-change materials. If true, optical glassy
fibre exhibiting intrinsic SHG response can be drawn from the melt of
noncentrosymmetric crystal-glass phase-change materials, keeping the advantages of a
glassy fibre, such as mechanical flexibility, optical transparency, and low optical loss.
Here we report an efficient and inexpensive way to produce nonlinear optical glassy fibre

using crystal-glass phase-change materials that adopt noncentrosymmetric space group.

59

3. Experimental Section

3.1. Reagents. The reagents mentioned in this work were used as obtained: K metal
(analytical reagent, Aldrich Chemical Co., Milwaukee, WI); Rb metal (analytical reagent,
Johnson Matthey/AESAR Group, Seabrook, NH); red phosphorus powder, -100 mesh,
Morton Thiokol, Inc., Danvers, MA; Se (99.9999%; Noranda Advanced Materials,
Quebec, Canada); N,N-dimethylformamide (Spectrum Chemicals, ACS reagent grade);
diethyl ether (Columbus Chemical Industries, Columbus W1, ACS reagent grade,
anhydrous). Agse (A = K and Rb) starting materials were prepared by reacting
stoichiometric amounts of the elements in liquid ammonia. P28e5 was prepared by
heating the mixture of P and Se with a stoichiometric ratio sealing in an evacuated silica
tube at 460 °c for 24h

3.2 Synthesis. Pure APSe6 (A = K, Rb) was achieved by a stoichiometric mixture of
Azse: P: Se = 1: 2: 11 under vacuum in a silica tube at 350 °C for 2 d followed by
cooling at a rate of 5 °C h'1 to 250 °C. Energy dispersive spectroscopy analysis of the
crystals showed an average composition of "KPSC62" and "Rb.,2PSe6_1", respectively, for
the orange rods-typed single crystals. For high purity, glassy phases of APSe6 were
obtained by quenching the melts of corresponding single crystals to room temperature. A
glassy fibre of APSe6 was drawn from the melt near melting point of each material.
Crystal data for KPSe6: orthorhombic Pca21, Z = 4; a = 11.5052(7) A, b = 6.7939(6) A, c
= 11.2328(7) A, V = 878.01(11) A3 at 100 K. Crystal data for RbPSe6: orthorhombic
Pca21, Z: 4; a = 11.7764(17) A, b = 6.8580(10) A, c = 11.7764(17) A, V= 925.5(2) A3

at 293 K. KPSe6 m.p.: ~320 °C, RbPSe6 m.p.: ~315 °C.

60

4. Physical Measurements.

X-ray Powder Diffraction. Analyses were performed using a calibrated CPS 120 INEL
X-ray powder difractometer (Cu Ka radiation) operating at 40 kV/20 mA and equipped
with a position-sensitive detector with a flat sample geometry.

Electron Microscopy. Semiquantitative analyses of the compounds were performed with
a J EOL J SM-3 5C scanning electron microscope (SEM) equipped with a Tracor Northern
energy dispersive spectroscopy (EDS) detector.

Solid-State UV—vis spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region. The instrument
is equipped with an integrating sphere and controlled by a personal computer. Details are
described in Chapter 2.

Raman Spectroscopy. Raman spectra were recorded on a Holoprobe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation from a HeNe
laser for excitation and a resolution of 4 cm'l. Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 pm. A total of 64 scans
was sufficient to obtain good quality spectra.

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CS] or KBr matrix.
The samples were ground with dry CsI or KBr into a fine powder and pressed into
translucent pellets. The spectra were recorded in the far-1R region (600-100 cm", 4 cm'1
resolution) and mid-1R region (500-4000 cm", 4 cm'1 resolution) with the use of a

Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon beam

61

splitter.

Differential Thermal Analysis (DTA). Experiments were performed on Shimadzu DTA-
50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. Samples were heated to 500°C at
10°C min", and after 1 min it was cooled at a rate of ~10°C min'l to 50 °C. Residues of
the DTA experiments were examined by X-ray powder diffraction. Reproducibility of the
results was checked by running multiple heating /cooling cycles. The melting and
crystallization points were measured at a minimum of the endothermic peak and a
maximum of the exothermic peak.

Nonlinear Optical Property Measurements. We used the frequency-tripled output (355
nm) of a passive-active mode-locked Nd:YAG laser with a pulse width of about 15 ps and
a repetition rate of 10 Hz to pump an optical parametric amplifier (OPA). The OPA
generates vertically polarized pulses in the range 400 ~ 3,156 nm. In order to study the
waveguided SHG response from our glass fibre, we used the idler beam (Amer = 1,240 —
1,610 nm) from the same OPA setting as above. The incident laser pulse of 0.2 m] was
focused onto the proximal surface of a fibre with a spot 300 pm in diameter using a 3 cm
focal-length parabolic lens. The diameter of this fibre was about 122 i 2 pm and its
length is 10 mm. By selectively focusing the imaging lens on the opposite distal end of
the fibre, the SHG signals were collected in a waveguide mode and dispersed with a Spex
Spec-One 500 M spectrometer coupled to a nitrogen-cooled CCD camera. Since our
monitoring range in the wavelength is rather wide, we did not use any filter but made

sure that other optical components did not generate any extra SHG signals. The SHG

ix)

(1

response from powder samples were measured using a reflection geometry under similar

conditions. The detailed experimental setup is described elsewhere”.

5. Results and Discussion.

In order to test our strategy, we chose the one-dimensional selenophosphate
compounds APSe6 (A=K, Rb)19 which crystallize in the noncentrosymmetric polar space
group of Pca21. Discussion here will be mainly concentrated on the K+ salt because the
Rb+ analogue is isostructural and has similar physicochemical properties including
nonlinear Optical properties to K+ salt. A one-dimensional infinite chain l/00[PSe(,'],
consisting of the [PSe4j tetrahedral unit condensed with Se; linkage, runs along the a-
axis, and is separated by alkali metal cations, Figure 3-1. The easily polarizable P and Se;
chains, linked by covalent bonding, can produce a large optical nonlinearity. The
compounds exhibited a reversible crystal-glass phase-change behaviour with optical
contrast between the phases. APSe6 exhibits wide optical transparency ranging from
long-wave 1R (LWIR) to near lR/visible light: 19.0 ,um - 574 nm, crystalline KPSe6; 19.5
pm - 681 nm, glassy KPSe6; 20.2 pm - 568 nm, crystalline RbPSe6; 18.9 ,um - 649 nm,
glassy RbPSe6, Figure 3-2a. Above the complex P-Se absorptions in the far-1R region,
uninterrupted light transmission continues through the mid-1R region. The optical
transparency extends to an absorption edge of the compound in the visible region. The
observation is consistent with characteristic 1R transmittance of chalcogenide
compounds. An important NLO material for IR application AgGaSeg shows LWIR

transmission up to 17 ,um2°.

63

A
G)
v

(b

V

 

 

 

 

 

 

 

 

 

 

5.
40 —UV—vis 10
* ,1 -—-—-—-Mid-IR 3 _ j ‘ A .
’1‘ 11"” Far-IR 30%-1045. A ‘ ‘ A
3, 30-5. at, . ‘
g .9". I E 3' ‘
.9 'l . (D10 - o ' . .
.éZO- ' Bandgap % i 0 _ . '
go) 1 1 ‘1’ . '
.0 E 2 e i .
< 10- I \ g 10 - o : '
I g F . I
0 L I 1 1 1 . 1 1 101 -. 1 1 1 3. 1 . 1
20 15 10 5 O 600 650 700 750 800
Wavelength, pm Wavelength, nm

Figure 3-2. (a) Far-[R/mid—IR/visible absorption spectra of KPSe6 bulk glass showing
wide transparency range. (b) Relative SHG intensities of KPSe6 crystal (triangle),
AgGaSez (circle), and KPSe6 glass (square).

Second Harmonic Generation Response of APSe6 (A=K, Rb) Bulk Crystal and
Glass. ZnGePz, GaSe, and AgGan (Q = S, Se) are mainstream IR NLO materials.
Despite an exceptional SHG coefficient, CdZnAsz (362) ~ 434 pm V") is made of toxic
elements and is only transparent beyond 4 pm, while optically isotropic GaAs (x0) ~ 240
pm V") is non-birefringent and non-phase-matchable2 1. We examined the SHG response
of crystalline and bulk glassy powders of APSe6 compounds from 1240 to 1610 nm using
a modified Kurtz powder methodn. SHG intensities of crystalline KPSe6 were directly
compared with that of AgGaSez powder. Samples prepared in a similar fashion and
having the same particle size range (45.5 :t 7.5 pm) were measured and compared.

Crystalline APSeb (A=K, Rb) powder generated very strong second harmonic signals

65

over a wide range in the visible/near infrared region. For example, KPSe6 and RbPSe(,
crystalline powder converted 1.5 pm fimdamental (lying mid way in the
telecommunication band) to 0.75 mm near-infrared light with ~63 and 68 times larger
intensities than that of AgGaSez powder, Figure 3-2b. On the basis of the electronic
structure, x0) is estimated to be 151.3 pm V'1 and 149.4 pm V'1 for K+ and Rb+ salts”.
SHG intensities of APSeb compounds increase with the particle sizes, indicating type-I
phase-matchability. It should be noted that the calculated )6” value of APSe6 compounds
are the highest among phase—matchable inorganic NLO materials with band gaps over 1.0
eV. Growing large single crystals of these compounds is in process to measure precise
x0) value. As-prepared KPSe6 bulk glassy powder exhibited SHG intensities comparable

to AgGaSez, Figure 3-2b.

   

Figure 3-3. (a) Representative photograph of an optical fibre showing remarkable
flexibility. (b) A representative SEM image of a fibre showing thickness uniformity at

50.0 um and surface smoothness.

66

Glassy Fibre Fabrication. Based upon the strong SHG response and wide optical
transparency in the IR region of both crystalline and glassy phases of APSe6 (A = K, Rb)
we fabricated glassy optical fibres. The Fibre drawing process was based upon the
reversible thermal behaviour of crystal-glass phase-change materials: upon heating glassy
phase crystallizes followed by subsequent melting, but upon cooling only vitrification
occurs, instead of recrystallization. We drew APSe6 glassy fibre from a viscous melt at
~230-280 °C; on cooling down from the liquid phase, and between vitrification and
melting point, a continuous viscosity-temperature dependence exists which make high-
speed drawing possible”. We note the processing temperature of the chalcogenide fibres
is considerably lower than that of oxide competitors. For example, silica fibre requires
approximately 2,000 K for softening”. Fibres with thickness ranging from a few to a
hundred micrometers, having remarkable flexibility, could be prepared ‘by hand’ with
lengths approaching a metre, Figure 3-3a. As seen in the seaming electron microscope
(SEM) image of Figure 3-3b, a representative d = 50.0 m fibre displays a high degree of
thickness uniformity and surface smoothness; the cross—section of a fibre was continuous
with no bubbles or cracks. KPSe6 glassy fibre could recover its crystallinity by annealing
it at 260 °C for 3 min. Crystallized fibre was slightly distorted in some regions but
preserves its fibre form. An X-ray diffraction study of pristine fibre on a single crystal
diffractometer showed faintly diffused scattering, confirming its amorphous nature,

whereas that of annealed fibre showed remarkable crystallinity, Figure 3-4.

67

 

 

80.000 ....,.........,.... 4

70,000 —

(b)

Glassy fibre

  
 

  

60,000

   

50,000

(221, 122)

(110), (200)

 

40.000
1

(601)

(311), (113)

30.000 -

lntensities (a. u.)

01
(211,112)

20,000 b

(225, 206)

(231,132)
(026)

Annealed fibre

10.000 —

 

 

0.11.1.111144-llnnnlllu.

 

20. deg

Figure 3—4. (a) X-ray diffraction ring patterns of pristine glassy (left) and annealed fibre
(right) confirming their amorphous and crystalline nature, respectively. (b) X-ray
diffraction patterns of the pristine glassy (upper) and annealed fibres (bottom).
Diffraction profiles were regenerated from the ring patterns presented in (a), collected by
STOE 11 single crystal diffractometer (Ag K0, ). Note that the Bragg peaks from the
annealed fibre are successfully indexed, indicative of the restoration of crystal structure

on the fibre. (th) index on the major peak is presented.

68

 

   

 
    

 

  
  
 

 

 

    

 

 

 

 

l : Annealed Fibre
a 3 - 5 5 firm
.a :
C l
g E E Glassy Fibre
'5 2 - a a .... eve-M
1D 1
.5 . '
g l : Bulk Glassy Powder
'5 1 l l 11 _A_A—i
z ' 5
E 1 Crystal
0 - 1 i . 1.. 1 . W'—
200 300 400 500

Raman shift, cm'1

Figure 3-5. Raman spectra of KPSe6 crystal, bulk glass, pristine glassy fibre, and

annealed fibre.

69

The Raman spectrum of crystalline KPSe6 at room temperature shows major
shifts at 220 (s), 231 (m), and 246 (m) cm'l, Figure 3-5. The shift at 220 cm'1 is
unambiguously assigned to the PSe4 stretching mode by comparison in the Ag stretching
mode of Td symmetry of [PSe4]3' ligand. The shifts at 231 and 246 cm'l can be assigned
to antisymmetric and symmetric Se-Se stretching vibration modes of the diselenide
group, respectively15 . The Raman spectra of KPSe6 bulk glassy powder and glassy fibre
are identical, showing the broader and weaker peaks at 220 (bm) and 259 (bm) cm",
whereas the overall peak pattern is similar to that of the crystals. This suggests that the
building blocks of [PSe4] and Se-Se bonds are still intact and local structural motifs are
largely preserved in the bulk glass and glassy fibre but long-range crystallographic order
is lost. The Raman spectrum of annealed fibre is same as that of bulk crystal, confirming
the recovery of crystalline structure in the fibre form, consistent with the X-ray powder
diffraction results.

The PDF analysis of KPSe6 crystalline, bulk glass, and glassy fibre is also in
agreement with the Raman spectroscopic data, Figure 3-6. The PDF of bulk glass and
glassy fibre shows well-defined correlations up to ~8 A with the first two at 2.3 A (P-Se
and Se—Se bonds) and 3.7 A (K---Se and the second neighbour Se---Se distances) being
very close to those of the crystalline phase. Above ~8 A, the PDFs decays rapidly to zero,
indicating the loss of the long-range order. The PDF result is consistent with that of
K2P2866'5. Those observations support the facile restoration of the crystal structure from

the amorphous state at the reversible crystal-glass phase transition.

70

 

10
' Thick fibre

 

5 _ Thin fibre

 

 
 
 

Bulk glass powder

 

 

PDF G(r), A2-

Crystalline powder

 

 

l I l 1 J l l l I l l l l l l l l l l l l L l I l I L 1

2 4 6 8 10 12 14 16 18 20
Radial distance, A

 

Figure 3-6. Pair distribution function (PDF) analysis for the glassy fibre, bulk glass, and
crystalline powders. Fibres with different thickness at (1 ~ 50 pm and d ~ 200 pm were
examined for comparison. Theoretical fit based upon single crystal structure refinement is
plotted as black circles. The first peak at 2.3 A corresponds to interatomic correlations of
P-Se and Se-Se bonds, and the second peak at 3.7 A KmSe and second neighbouring
Se- - °Se. Note that PDFs of bulk glass and glassy fibres are very close to that of
crystalline powder, indicating local structural order are significantly preserved in the
amorphous state of bulk powder and fibre, but only lost long-range crystallographic order.
This observation plausibly supports the presence of the intrinsic second-order nonlinear

optic properties on the bulk glass and glassy fibre.

71

Second Harmonic Generation Response and Optical Loss. Since the bulk glass
powder showed significant SHG response, we examined the corresponding generation
and guiding of NLO light in glassy fibres. Being carefiil to precisely align the fibre with
the laser path, we focused the tuneable incident beam (2 = 1,240 — 1,610 nm) onto the
proximal end of the fibre (d = 122 pm, I = 10.0 mm); wave-guided outgoing light signal
was collected from a distal end. Laser power was carefully adjusted to avoid
crystallization or any laser-induced poling effect on a glassy fibre. As-prepared APSe6
glassy fibre acted as a frequency convertor in a waveguide mode. It produced
continuously tuneable SHG signal over a wide range of wavelengths (640 - 805 nm),
Figure 3-7. The decreased SHG signal below 700 nm is due to the two—photon-induced
absorption beyond the band gap. The observation of intrinsic SHG response from KPSe6
glassy fibre is consistent with that found in the corresponding bulk glass powder. Note
the SHG signal is generated continuously along the path, but the full path represents a
macroscopic distance of 10.0 mm. Although we confirmed the amorphous nature of
glassy fibre by X-ray powder diffraction, it cannot be totally ruled out that the presence
of nanocrystals embedded induces the SHG response. We estimated the optical loss of
KPSe6 glassy fibre by measuring transmitted SHG intensities for various fibre lengths. A
10.0 mm fibre (d = 122 ,um) was consecutively cut to 5.0 and 2.5 mm and the
waveguided SHG intensities shown in Figure 3-8 were recorded. Rather than observing
an exponential decay in the spectral region where no band gap absorption is present, the
optical loss was linearly proportional to the fibre length, approximately 0.66 dB mm},
Figure 3-9. Details concerning the overall behaviour of the propagating SHG signal will

require further study. Various effects including scattering from impurity, grain

72

boundaries, and the fibre end, as well as dispersion and inhomogeneity-induced mode
conversion, all need to be considered. Regarding the materials themselves, fiirther
understanding of the temperature, surface tension and viscosity of the melt, the use of
high purity starting materials, and a more sophisticated mechanical drawing process, all
promise higher quality of the fibres. We note that commercial chalcogenide ASzSC3
optical fibre has an attenuation of 10'2 dB mm'1 at 10.6 pm“. It should be noted that low
optical loss is a prerequisite for optical fibre. Polyester sulphone (PBS) showed
comparable thermal properties to chalcogenide optical fibre, but the optical loss reached

30 dB mm'1 in the mid-IR at 10.6 gm.

 

120

100 —

00
O
l

SHG intensity (arb. units)
-l=- O)
O O
l l

N
O
l

 

 

 

 

 

 

 

 

 

 

 

 

 

Milk“

600 650 700 750 800 850
Converted Wavelength (nm)

0
l

 

Figure 3-7. The waveguided SHG response transmitted through 10.0 mm long KPSeb

glassy fibre in a wide range of vis/near IR region.

73

 

  
 

 

 

 

 

 

 

5000 _ 31035
a i
3’ @102:- - 2 '
. 5 A ‘
g 4000 - 5 : - . ' '
a) 0 3
$33 $101.? 2
U) a, ; O
8 3000 - .5 '
E moo.-----.-,
5 g 640 680 720 760 800
:I: Converted Wavelength, nm
<3 2000 l.
.3 —2.5 mm
% -—--—— 5.0 mm
‘1 1000 ~ -—---10.0 mm
0 l I 1 I n I 1 I

698 700 702 704 706 708 710 712 714 716
Converted wavelength, nm

Figure 3-8. Optical loss of KPSe6 glassy fibre at 711.5 nm, estimated by measuring SHG
intensities for various lengths of fibres. Inset: optical loss for 640 to 788 nm SHG signals
at 10.0 (circle), 5.0 (triangle), and 2.5 mm (square) long fibres.

74

 

 

 

 

 

 

 

 

 

 

A: 1575.0 nm
8000 i B: 1497.4 nm
t A B lidler used: c: 1420.5 nm
,5 D: 1350.1 nm
3 D E: 1282.1 nm
10' 6000 - c
8 1
:5 1 E
3 4000
9 ' i l
(L2 i il l
D 2000 - ' ii i
. L l
0 “L 1 H J 4 JL 1 r A . 1 . r .1 L
700

640 660 680 720 740 760 780 800
Converted Wavelength, nm

Figure 3-9. The DFG response as a function of Aidler, demonstrating wave-mixing

capability over a wide range of wavelengths.

 

 

 

 

 

 

 

1800
,3 1600 —
3 .
g; 1400 ~
m .
.g 1200 -
“a

g, 1000 -
.s -
(9 800 -
I

‘0 600 -
9 l
3 400 -
°’ i
‘1 200 - L1

 

 

 

 

 

O . l . . .
620 640 660 680 700 720 740 760 780 800 820
Converted wavelength, nm

Figure 3-10. The relative SHG intensities measured from 620 to 805 nm for the pristine
glassy and annealed fibres. Darker profiles represent SHG signals of the annealed fibre,
and lighter profiles the pristine fibre.

75

Difference Frequency Generation Response. We also performed DFG
experiments; this process is especially important for generating mid-IR light, and
facilitates the multichannel conversion, low noise, high speed, and good transparency
required for the wavelength division multiplexing (WDM) architecture26. Here we used
both the signal and idler beams generated by an optical parametric amplifier driven by a
Nd:YAG pulsed laser at 355 nm. The energy conservation among these two beams
requires l/Asignal + l/llidler = 1/355 nm, and by definition of DFG, l/Asignal - l/Aidler =

l/ADFG- Thus, for a given Aidlers the expected wavelength for DFG was,

 

xi.
idler
AD = 355nm 1
PG [/1 —710nm] ()

Idler

By introducing different combinations of idler and signal beams, the KPSe6 glassy
fibre successfully generated continuously tuneable near-IR light by DFG, Figure 3-9.
Deviation among DFG intensities arose from the signal beam, which was beyond the
band gap. Although our detection limit (<1 ,um) prohibited observing DFG at mid-IR,
APSe6 (A = K, Rb) should also produce tuneable coherent light throughout mid-IR region
because the compounds are optically transparent there, a region where few NLO
materials are available”. We cannot study SFG because our experimental set up produces
constant ASFG at 355 nm, light that would be strongly absorbed because it is beyond the
band gap.

Phase-change materials are of great interest as emerging technologies including
rewritable optical media and the development of nonvolatile phase-change memory”.
Conversion between the crystalline and glassy states can be driven by applying a voltage

or heat, or irradiating appropriate laser. When they have a pronounced optical or

76

resistivity contrast, one state can be differentiated from the other. The stoichiometric
compounds APSe6 (A = K, Rb) can switch between the crystalline and glassy states
without complications arising from compositional changes. One might exploit this
property to switch the NLO properties of APSe6 fibres. To explore this we annealed
KPSe6 glassy fibre at low temperature of 260 °C for 3 min and measured the waveguided
SHG response. Remarkably, the annealed fibre exhibited over 10 times larger SHG
intensities compared to the pristine glassy fibre in a wide range of wavelengths, Figure 3-
10. This phase switching property of APSe6 fibre can be utilized to periodically pole a
fibre or film for quasi-phase-matchingzg. Various poling techniques developed to induce
SHG on common glasses can be applied to APSe6 glassy fibresl’3’4. By writing a grating
onto a fibre with an optical standing wave the transmission of selected frequencies could

be suppressed, a kind of latching optical switch.

6. Concluding Remarks

APSe6 glassy fibre is in a totally different class from the conventional optical
glassy fibre systems, in that it possesses strong, intrinsic, switchable, second-order NLO
properties. It opens up the possibility of creating active, all-optical, broadband networks
that independently modulate frequency, with no additional NLO or electronic devices.
The fabrication concept suggested here is an example of combining apparently irrelevant
properties (NLO + phase-change behaviour) to explore new functional materials. The
discoveries and approaches described here are expected to stimulate the use of phase-

change materials for NLO glassy fibres, and lead to further studies of the local structure

77

of glassy states in phase-change materials in order to explain the origin of their NLO

properties.

References

(l) Margulis, W.; Garcia, F. C.; Hering, E. N.; Valente, L. C. G.; Lesche, B.;
Laurell, P.; Carvalho, l. C. S. MRS Bull. 1998, 23, 31-35.

(2) Myers, R. A.; Mukherjee, N.; Brueck, S. R. J. Opt. Lett. 1991, 16, 1732-1734.

(3) Corbari, C.; Kazansky, P. G.; Slattery, S. A.; Nikogosyan, D. N. App]. Phys. Lett.
2005, 86, 071106; Kityk, I. V. J. Phys. Chem. 32003, 107, 10083-10087.

(4) Kazansky, P. G.; Dong, L.; Russell, P. S. Opt. Lett. 1994, 19, 701-703.

(5) Bloembergen, N. Nonlinear optics; 4th ed.; World Scientific: Singapore ; River
Edge, NJ, 1996.

(6) Islam, M. N. Phys. Today 1994, 47, 34-40.

(7) Chou, M. H.; Brener, 1.; Fejer, M. M.; Chaban, E. B.; Christman, S. B. IEEE
Photon. Techno]. Lett. 1999, 11, 653-655; Yoo, S. J. B. J. Lightwave Techno]. 1996, 14,
955-966.

(8) Tittel, F. K.; Richter, D.; Fried, A. Solid-State Mid-Infrared Laser Sources 2003,
89, 445-510.

(9) Pushkarsky, M. B.; Webber, M. B.; Macdonald, T.; Patel, C. K. N. App]. Phys.
Lett. 2006, 88, 044103.

(10) Pestov, D.; Wang, X.; Ariunbold, G. O.; Murawski, R. K.; Sautenkov, V. A.;
Dogariu, A.; Sokolov, A. V.; Scully, M. 0. Proc. Nat]. Acad. Sci. U. S. A. 2008, 105,
422—427.

(11) Pushkarsky, M. B.; Dunayevskiy, I. G.; Prasanna, M.; Tsekoun, A. G.; Go, R.;
Patel, C. K. N. Proc. Nat]. Acad. Sci. U. S. A. 2006, 103, 19630-19634.

(12) Pushkarsky, M.; Tsekoun, A.; Dunayevskiy, I. G.; Go, R.; Patel, C. K. N. Proc.
Natl. Acad. Sci. U. S. A. 2006, 103, 10846-10849.

(1 3) Jean, B.; Bende, T. Solid-State Mid-Infiared Laser Sources 2003, 89, 51 1-544.
(14) Seddon, A. B. J. Non-Cryst. Solids 1995, 184, 44-50.

(15) Chung, 1.; Malliakas, C. D.; Jang, J. 1.; Canlas, C. G.; Weliky, D. P.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2007, 129, 14996-15006.

79

(16) Wachter, J. B.; Chrissafis, K.; Petkov, V.; Malliakas, C. D.; Bilc, D.; Kyratsi, T.;
Paraskevopoulos, K. M.; Mahanti, S. D.; Torbrugge, T.; Eckert, H.; Kanatzidis, M. G. J.
Solid State Chem. 2007, 180, 420-431.

(17) Chung, 1.; Jang, J. 1.; Gave, M. A.; Weliky, D. P.; Kanatzidis, M. G. Chem.
Commun. 2007, 4998-5000.

(18) Jang, J. I.; Ketterson, J. B. Phys. Rev. 82007, 76, 155210.

(19) Chung, 1.; Do, J.; Canlas, C. G.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem.
2004, 43, 2762-2764.

(20) Nikogosyan, D. N. Nonlinear optical crystals: a complete survey; Springer-
Science: New York, 2005.

(21) Fiore, A.; Berger, V.; Rosencher, E.; Bravetti, P.; Nagle, J. Nature 1998, 391,
463-466.

(22) Kurtz, S. K.; Perry, T. T. J. Appl. Phys. 1968, 39, 3798-3813.

(23) Song, J.-H.; Freeman, A. J.; Bera, T. K.; Chung, 1.; Kanatzidis, M. G. Nat.
Mater. 2008, Submitted.

(24) Abouraddy, A. F.; Bayindir, M.; Benoit, G.; Hart, S. D.; Kuriki, K.; Orf, N.;
Shapira, O.; Sorin, F.; Temelkuran, B.; Fink, Y. Nat. Mater. 2007, 6, 336-347.

(25) Tong, L. M.; Gattass, R. R.; Ashcom, J. B.; He, S. L.; Lou, J. Y.; Shen, M. Y.;
Maxwell, 1.; Mazur, E. Nature 2003, 426, 816-819.

(26) Park, S. K.; Do, J. Y.; Ju, J. J.; Park, S.; Kim, M. S.; Lee, M. H. Mater. Lett.
2005, 59, 2872-2875.

(27) Wuttig, M.; Yamada, N. Nat. Mater. 2007, 6, 824-832.

(28) Houé, M.; Townsend, P. D. J. Phys. D: Appl. Phys. 1995, 28, 1747-1763.

ES

Chapter 4

The Helical Polymer 1/00[PZSe62']: Strong Second Harmonic Generation

Response and Phase-Change Properties of its K and Rh Salts

1. Introduction

The chalcophosphate anions show great structural diversity and are predominantly
discrete molecular species as for example the [PSe4]3',l [PzSe¢5]4',2 [PZSe9]4',3 [PgSelg]6'4
and the trans-decalin-like [P6S012]4’.5 The number of infinite anions is very small and
includes 1/OO[PSe6']6 and l/OO[P_<,SemS']. The latter is found bonded to transition metal
atoms in A3MPsSem (A=K, Rb; M=Ru, Os).7 The polymeric anions show unusual
structural moieties and coordination chemistry compared to the classical [PSe4]3' and
[steé]4-. All chalcophosphates are potential ligands to metal ions for building finite or
extended structures.8 The anions can be easily stabilized in polychalcogenide fluxes of
suitable composition. Our experimental investigations of alkali selenophosphates
provided useful insights on the relationship between structure and flux composition
(A:P:Se ratio in the composition) in this class.5 The diversity of these anions seems
greater than that of the oxo-phosphate anions which are responsible for defining the
enormous class of metal phosphates.9 Therefore, the chalcogenide analogs have the
potential to produce an even greater family of compounds with metal ions. One striking
contrast between phosphates and chalcophosphates is the ability of the latter to couple

through Q-Q bonds (Q = S, Se) forming catenated species. In principle, every simple

81

molecular anion [PnyF' could be oxidatively coupled through its terminal P-Q bonds to
form lager species linked with P-Q-Q-P moieties.

Here we describe the unique helical anion 1/00[P28e62'] in the form of K+ and Rb+
salts. Although it was not formed via actual oxidative coupling chemistry, the chiral
structure derives from coupling of the ethane-like [PZS%]4' anion. Both K2P2866 and
Rb2P28e6 can be quenched from the melt to a glassy state and can undergo reversible
phase-change crystal-glass transitions. Interest in phase-change materials is growing, as a
result of emerging technologies including commercially available rewritable optical
media and the development of nonvolatile phase-change memorylo’” The most
interesting materials are those with stoichiometric composition because they can switch
between the two states without complications due to compositional changes.

At low temperature K2P2866 (but not szPZSe6) undergoes a phase transition to a
structure of lower symmetry while retaining the helical structure. Polar structure K2P2866
exhibits remarkably strong second harmonic generation (SHG) intensity. In the infrared
region the material is highly transmissive over a wide wavelength range. Compared to the
SHG intensity of the chalcopyrite compound AgGaSez, which is the top infrared NLO
material used commercially, K2P28e6 exhibited 50-fold stronger response as well as type-
I phase matching property making it a potential contender material for applications in the
IR region. To our best knowledge this is one of the largest NLO SHG responses ever
reported for this near and mid infrared region of the spectrum. Glassy K2P2Se6 also
exhibited SHG response. Since homogeneous glasses generally do not show SHG due to
the macroscopically present inversion center, this observation is not only a rare example

of amorphous NLO response with no specific treatment such as poling, it is also of

82

interest for further investigations in optical devices.

2. Experimental Section

2.1. Reagents.

The reagents mentioned in this work were used as obtained: K metal, analytical
reagent, Aldrich Chemical Co., Milwaukee, WI; Rb metal, analytical reagent, Johnson
Matthey/AESAR Group, Seabrook, NH; red phosphorus powder, -100 mesh, Morton
Thiokol, Inc., Danvers, MA; Se, 99.9999%, Noranda Advanced Materials, Quebec,
Canada; N,N-dimethylformamide, ACS reagent grade, Mallinckrodt Backer Inc., Paris,
KY; diethyl ether, ACS reagent grade, anhydrous, Columbus Chemical Industries,
Columbus WI. AzSe (A = K and Rb) starting materials were prepared by reacting
stoichiometric amounts of the elements in liquid ammonia under N2. P2565 was prepared
by heating the mixture of P and Se with a stoichiometric ratio sealing in an evacuated

silica tube at 460 °C for 24h.

2.2. Synthesis

Pure K2P28e6 and RbgPZSe6 were obtained in quantitative yield by heating a
mixture of KZSezPZSe5=lzl and Rb;Se:P:Se=l:2.4:5 in an evacuated and sealed silica
tube at 450 °C for 3 (I followed by cooling at a rate of 5°C h'1 to 250 °C, respectively.
After washing with N,N-dimethylformamide (DMF) and ether, we obtained pure
red/orange thick plate-typed single crystals. Energy dispersive spectroscopy (EDS)

analysis of the crystals showed an average composition of “KngSesg” and

83

“Rb2P1,9Ses,9”. The glassy phases of K2P2866 and Rb2P2886 were prepared from a mixture
of KZSe1PZSes=1 :l and Rb2P2S86 crystals respectively, placed in a silica tube and melted
at 800-900 °C for 1-2 min and subsequent quenching in ice water. The crystalline
compounds were air-stable for at least a week and stable under polar solvents such as
DMF, N—methylformamide, methyl and ethyl alcohol and H20. Glassy K2P2866 and
szPZSeé were soluble in DMF and methyl alcohol without stirring to give clear light

orange solution.

3. Physical Measurements

X-ray Powder Diffraction and Pair Distribution Function (PDF) Analysis. Phase
purity x-ray diffraction analyses were performed using a calibrated CPS 120 INEL X-ray
powder diffractometer (Cu Ka graphite monochromatized radiation) operating at 40
kV/20 mA and equipped with a position-sensitive detector with flat sample geometry.

The powder X-ray diffraction data for PDF analysis of the crystalline and glassy KZPZSeg,
were collected at room temperature using an Inel CPS 120 diffractometer. A graphite
monochromatized and Rh filtered Ag Ka (0.56080 A) radiation was used for the
crystalline and glassy K2P28e6, respectively. The samples were ground to a fine powder
under nitrogen atmosphere, loaded into 0.5 mm capillaries and flame sealed.
Fluorescence due to selenium atoms was successfully filtered out by using a 0.150mm
thick bronze foil in front of the CPS detector. Data reduction and the calculation of the
PDF were performed as described elsewherel2 with the PDFGetX2 software.l3

Electron Microscopy. Semiquantitative analyses of the compounds were performed with

84

a J EOL J SM—3 5C seaming electron microscope (SEM) equipped with a Tracor Northern
energy dispersive spectroscopy (EDS) detector.

Solid-State UV-vis spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region. The instrument
is equipped with an integrating sphere and controlled by a personal computer. BaSO4 was
used as a 100% reflectance standard. The sample was prepared by grinding the crystals to
a powder and spreading it on a compacted surface of the powdered standard material,
preloaded into a sample holder. The reflectance versus wavelength data generated were
used to estimate the band gap of the material by converting reflectance to absorption
data.l4
Raman Spectroscopy. Raman spectra were recorded on a Holoprobe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation from a HeNe
laser for excitation and a resolution of 4 cm". Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 pm. A total of 128 scans
was sufficient to obtain good quality spectra.

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CsI or KBr matrix.
The samples were ground with dry CsI or KBr into a fine powder and pressed into
translucent pellets. The spectra were recorded in the far-IR region (600-100 cm", 4 cm'1
resolution) and mid-IR region (500—4000 cm", 4 cm‘1 resolution) with the use of a
Nicolet 740 FT—IR spectrometer equipped with a TGS/PE detector and silicon beam
splitter.

Differential Thermal Analysis (DTA) Experiments were performed on Shimadzu DTA-

50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. The sample was heated to 600°C
at 10°C/min, and after 1 min it was cooled at a rate of —10 °C min'1 to 50 °C. The residues
of the DTA experiments were examined by X-ray powder diffraction. Reproducibility of
the results was confirmed by running multiple heating/cooling cycles. The melting and
crystallization points were measured at a minimum of endothermic peak and a maximum
of exothermic peak.

Solid-State Nuclear Magnetic Resonance (N MR) Spectroscopy. The room temperature
solid-state NMR measurement was taken on a 9.4 T NMR Spectrometer (Varian Infinity
Plus) using a double resonance magic angle spinning (MAS) probe. The sample was spun
at 8.0 kHz using zirconia rotors of 6 mm outer diameter. Bloch decay spectra were taken
with the excitation/detection channel tuned to 31P at 161.39 MHz, a 4.5 ,us 90° pulse, and
a relaxation delay of 3,000 3. The spectrum was referenced using 85% H3PO4 at 0 ppm.
X-ray Crystallography. The crystal structure was determined by single-crystal X-ray
diffraction methods. For KngSe6, preliminary examination and data collection were
performed on a SMART platform diffractometer equipped with a 1K CCD area detector
using graphite monochromatized Mo Ka radiation at 173(2) and 298(2) K. A
hemisphere of data was collected at 298(2) K using a narrow-frame method with scan
widths of 0.300 in a) and an exposure time of 305 frame". For the data collected at
173(2)K, 60$ frame'l was used. The data were integrated using the SAINT program. An
analytical absorption and empirical absorption correction using the program SADABS

were performed for both sets. The initial positions for all atoms were obtained using

86

direct methods and the structures were refined with the full-matrix least-squares
techniques of the SHELXTL'5 crystallographic software package. Crystallographic data
are given in Table 4-1. Satisfactory refinement was obtained with the chiral space group,
P3121 at 298(2) K and P31 at 173(2) K. The crystals used for structure refinement were
racemically and merohedrally twinned, and the twinning fractions refined to 49.5, 25.8,
23.1 and 1.6 % at 173(2) K. The racemic and merohedral twinning behavior was typical
for K2P28e6 and their relative ratios determined for several crystals were very similar.
Interestingly, flux synthesis using the mixture of KZSe:P:Se = 1:2.4:5 significantly
increased the portion of one enantiomorph compared to the crystals obtained with a direct
combination reaction of KZSe and P2865. The racemic ratio improved from 1:1 to 3:1.

Intensity data for Rb2P2386 were collected at 100(2) K on a STOE IPDS II diffractometer
with Mo K0. radiation operating at 50 kV and 40 mA with a 34 cm image plate. Individual
frames were collected with a 5 min exposure time and a 1.0 a) rotation. The X-AREA, X-
RED and X-SHAPE software package was used for data extraction and integration and to
apply empirical and analytical absorption corrections. The SHELXTL software package
was used to solve and refine the structure. The most satisfactory refinement was obtained
with the chiral space group, P3121 at 100(2) K rather than P3221. The refined Flack x
parameter was 0.901(96). After inversion to P3121, the Flack parameter x was 0 and the
BASF value was 0.89. Following refinement for racemic and merohedral twinning done
at the same time, the twinning fractions refined to 89.43, 5.83, 5.59 and 2.15 % and x
parameter 0.0. szPZSe6 did not show a phase transition with temperature change and its
crystal structure is the same as that of K2P2886 at room temperature (RT). The parameters

for data collection and the details of the structural refinement are given in Table 4-1.

87

Fractional atomic coordinates and displacement parameters for each structure are given in
Tables 4-2 - 4-4. In all cases the atoms were refined to full occupancy.

Nonlinear Optical Property Measurements. We used the frequency-tripled output of a
passive-active mode-locked Nd:YAG laser with a pulse width of about 15 ps and a
repetition rate of 10 Hz to pump an optical parametric amplifier (OPA). The OPA
generates vertically polarized pulses in the ranges 400 ~ 685 nm and 737 ~ 3156 nm. In
order to check the SHG efficiency as a function of the excitation energy, we tuned the
wavelength of the incident light from 1000 ~ 2000 nm. In this range, the spectral
bandwidth of the linearly polarized light from the OPA is rather broad, about 2 meV full
width at half maximum. However, the phase space compression phenomena ensures
effective SHG where lower energy portions are exactly compensated by higher parts
thereby satisfying both energy and momentum conservation. The incident laser pulse of
300 ,uJ was focused onto a spot 500 pm in diameter using a 3 cm focal-length lens. The
corresponding incident photon flux is about 10 GW cm'z. The SHG signal was collected
in a reflection geometry from the excitation surface and focused onto a fiber optic bundle.
The output of the fiber optic bundle is coupled to the entrance slit of a Spex Spec-One
500 M spectrometer and detected using a nitrogen-cooled CCD camera. The data
collection time is 20 5. Single crystal or glassy sample was ground and separated by

various size ranges using sieves. Samples were placed in a capillary tube and measured.

88

Table 4-1. Crystallographic Data and Refinement Details for KZPZSe6 and szPZSeé.

 

Formula

Space group

a, b, A

c, A

a, .6 (deg)
7(deg)

Z

V, A3

(1 (calculated), mg/m3
Temperature, K
2., A

,1], mm"1
F(000)

Bmax, deg

Total reflections

Total unique reflections

No. parameters

Refinement method

Final R indices [I>2o(1)],

Rla/wRZ b (%)

R indices (all data), R1/wR2

Goodness-of-fit on F2

Absolute structure parameter

K2P28€6

P3121 (no. 152)

7.2746(8)
18.870(4)
90

120

3
864.8(2)
3.536
298(2)
0.71073
19.974
816
28.27
7444
1382

47

Full-matrix least-squares on F2

2.69/5.89
3.10/6.04
1.078
027(2)

K2P2866

P31 (no. 144)
14.4916(7)

18.800(2)
90

120

12
3419.2(4)
3.578
173(2)
0.71073
20.209
3264
28.29
21703
10294
365

461/817

14.92/11.43

0.748
0.25(])

Rb2P2886
P3121 (no. 152)
7.2982(5)
19.002(2)
90

120

3
876.5(1)
4.016
101(2)
0.71073
27.272
924
30.99
9968
1866

49

387/4.56

474/470
0.976

008(3)

 

a R1 = 2 “For IFcll/ZIFol- b wR2 = {z [W(F02-Fc2) 210:1w7F0212111/2

89

Table 4-2. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (><103 A2)

 

 

for K2P28e6 at 298(2)K.

atom x y z U(eq)a
Se(l) 757(1) 2584(1) 2233(1) 27(1)
Se(2) 2769(1) 2649(1) 3957(1) 44(1)
Se(3) 4572(1) 2495(1) 801(1) 34(1)
K 5542(2) 2697(2) 2538(1) 49(1)
P 1285(2) 592(2) 548(1) 20(1)

 

a U(eq) is defined as one-third of the trace of the orthogonalized Uij tensor.

90

Table 4-3. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (X103 A2) for

K2P25€6 at 173(2)K.

 

 

atom x y z U(eq)
Se( 1) 3688(5) 9676(5) 645(4) 14(1)
Se(2) 6125(5) 2292(5) 332(4) 17(1)
Se(3) 6208(6) 9822(6) 147(4) 26(2)
Se(4) 1357(5) 4907(6) 2239(3) 20(1)
Se(5) 8603(6) 265 1(5) 2076(4) 19(1)
Se(6) 9132(6) 5372(5) 1757(4) 21(2)
86(7) 8756(5) 4608(5) 627(4) 14(1)
Se(8) 1483(5) 5153(6) 166(4) 29(2)
88(9) 967(6) 7344(6) 324(4) 21(2)
Se( 1 0) 6399(5) 5185(5) 2246(3) 19(1)
Se(] 1) 3 824(6) 2740(5) 2041(4) 18(2)
Se( 12) 4013(6) 5368(6) 1759(3) 17(1)
Se(13) 3654(5) 4607(5) 622(4) 14(1)
Se(14) 6081(6) 7255(5) 350(4) 17( 1)
Se(15) 6201(5) 4810(5) 146(4) 22(2)
Se(16) 3 598(6) 7670(5) 2049(4) 17( 1)
Se( 1 7) 6374(5) 9920(5) 2247(3) 18(1)
Se(18) 4164(6) 0437(5) 1775(4) 18(1)
Se( 19) 8701(5) 9560(5) 639(4) 15(1)
Se(20) 925(6) 2310(6) 355(4) 22(2)
Se(21) 9879(6) 1305(5) 3494(4) 30(2)
Se(22) 1402(5) 0149(5) 2256(3) 19(1)
Se(23) 8845(6) 7708(5) 2070(4) ' 19(1)
Se(24) 8997(6) 0310(5) 1781(4) 17(1)
K(l) 6431(12) 7647(11) 2114(8) 26(3)
K(2) 1398(12) 2618(12) 2087(8) 30(4)
K(3) 1393(11) 7869(11) 2066(8) 28(3)
K(4) 8517(11) 7101(11) 339(7) 26(3)
K(5) 7876(12) 6401(11) 3684(7) 25(3)
K(6) 3789(11) 2345(10) 306(7) 19(3)
K(7) 6404(1 l) 2918(9) 2050(7) 28(3)
K(8) 3815(11) 7381(10) 314(7) 25(3)
P(l) 5302(13) 629(12) 79(9) 11(3)
P(2) 9693(13) 4295(13) 2324(9) 12(3)
P(3) 0372(12) 5712(13) 91(8) 13(3)
P(4) 4711(13) 4389(13) 2312(8) 12(3)
P(5) 5276(12) 5597(13) 86(9) 11(3)
P(6) 4701 (12) 9309(11) 2316(8) 11(3)
P(7) 0319(13) 677(12) 105(9) 15(3)
P(8) 9724(12) 9355(12) 2344(8) 12(3)

 

a U(eq) is defined as one-third of the trace of the orthogonalized Uij tensor.

91

Table 4-4. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (><103 A2)

for Rb2P2366 at 101(2) K.

 

 

atom x y z U(eq)a
Rb 5848(1) 2708(1) 2479(1) 9(1)
Se(l) 857(1) 2488(1) 2207(1) 6(1)
Se(2) 3106(1) 2546(1) 3963(1) 10(1)
Se(3) 4272(2) 2035(1) 749(1) 7(1)
P 979(3) 238(3) 536(1) 5(1)

 

a U(eq) is defined as one-third of the trace of the orthogonalized Uij tensor.

92

 

 

 

Se1

1\ 1 . I
' Se3 C

Helix repeat distance: 18.872A

Figure 4-1. Structure of K2P28e6 at 298(2) K. (a) The unit cell viewed down the c-axis.

The thermal ellipsoids with 30% probability are shown. (b) View of a 1/00[P28662'] chain

looking down the a-axis. A helix forms by three [P2Se(,] units and repeats itself at every
18.872 A.

93

 

Figure 4-2. (8) Projection view of a l/00[P2$e.52'] chain slightly slanted in order to show
the staggered anti-conformation of [PZSeG]. The role of short Se-"Se nonbonding
interactions (dashed line) along a helix segment is shown. Se(l)---Se(2), 3.559(1) A. (b)
Projection through P-P bonding showing the relationship between short Se-"Se

nonbonding interaction and chain propagation.

94

 

 

}J\/.63 1 3

$617
$811 8313 83168

p3 ; SS7 P3
. ’Se12 :5 p5 P1 366 ’ A
$6 -
8614 4 1 p2 Sea
5913 8615 $618 $62

   
 
 

e19 885
9 se22Se23
Pugs

Sezqir’

8621

 

   

 

 

Figure 4-3. Structure of K2P2566 at 173(2) K. (a) The unit cell viewed down the c-axis
showing the superstructure. Crystallographically unique chains are differentiated as A and
B. K atoms are labeled. (b) View of 1/,,O[P2Se(52'] chains with labeling down the a-axis. K

atoms are omitted for clarity.

95

4. Results and Discussion

Crystal Structure. Because at RT KszSe6 and RbgPZSe6 are isostructural, the
description that follows will be concentrated on the K salt. The structure has infinitely
extended helical chains parallel to the c-axis, see Figure 4-1a. The chains are composed
of ethane-like [P2566] repeat units that are linked via a terminal Se-Se linkages to give
infinite helices of l/(,O[PzSe62'], Figure 4-1b. The chirality of K2P28e6 is preserved in the
crystal and at RT the compound adopts the chiral trigonal space group, P3121. Helical
features have been described in the quaternary thiophosphates AszPlez (A = K, Rb,
Cs)16 and AHU7(PS4).3 (A = K, Rb)” but K2P2$66 is the first example to display an
infinite helix structure of free standing chalcophosphate anions.

The unit cell contains a single crystallographically unique strand of the chain.
Each helix runs with identical handedness along the c-axis. Because the helix is generated
by a crystallographic 31 screw axis along the c-direction, each coil of the helix in the unit
cell includes three P28e6 units. There is a single crystallographic P site in the structure
and pitch of the helix is 18.872(4) A, corresponding to the unit cell repeat distance along
the c-axis. The inter-spiral distance is equal to the a-lattice constant of the unit cell.
Projected along its axis, a chiral 1D channel is seen with a cross-section diameter of
~1 .17 A. The P-P bond distance is normal at 2.242(2) A. P-Se bond distances range from
2.126(1) to 2.329(1) A comparable to known values. The P-Se-Se-P dihedral angle of
88.55(1)°, almost 900 causes a relatively short Se-Se bond distance of 2.337(1) A
compared to those in KPSCé.6 The ethane-like P2866 repeat unit adopts a local anti-
conforrnation to minimize the steric hindrance around the P-P bond created by the helical

geometry, Figure 4-2a. The unit is also distorted from the ideal geometry with the Se-P-

Se angles ranging from 111.88(6) to 117.95(6)°.

It is noteworthy to point out the unusually short Se-"Se nonbonding interaction is
observed at 3.559(1) A for Se1---SeZ, Figure 4-2a. This is shorter than the van der Waals
radii sum of 3.80 A and probably meaningful in stabilizing the crystal structure. Because
the P2366 fragment is intrinsically centrosymmetric, the helical architecture results from
the conformation of selenophosphate chain. In this regard, the short Se---Se contacts
likely play a critical role in forming the coil by maximizing Se p1: orbital overlap through
Se2--°Sel-Sel~-Se2 connections, Figure 4-2. To minimize the repulsion force between
them, each [P2865] subunit rotates along the Se2---Se1-Sel---SeZ connection as an axis
and consequently achieves the helical form. In fact, the Se3-P-P-Se3 torsion angle is
severely compressed to only 36.23(l)°. Similarly short Se~--Se interactions are well
known in other chalcogenide systems. In Rb4P686125 and APSe66, for example, these
interactions act to form pseudo one- or two-dimensional structures, respectively. The low-

dimensional charge density wave compound NbSe318 also has similar interactions.

Low Temperature Structure. We discovered that when cooled K2P2Se6
undergoes a displacive phase transition to lower symmetry with no bond breaking at
lower temperature. The P3121 space group at room temperature was lowered to P31 at
173 K by losing two-fold symmetry perpendicular to a- and b-axes, Figure 4a. The new
cell was enlarged to a 2a x 2b x c supercell. P31 is a maximal non-isomorphic subspace
group of P3121. As a result, the crystallographically unique K and P atoms and three
independent Se atoms in K2P2366 split to eight, eight and twenty four independent atoms,

respectively. The resulting four helixes in the unit cell are differentiated

97

crystallographically in two groups as denoted by A and B in Figure 4-4a and 4-4b with a
ratio of 3:1. It is noted that K7 and K8 still generate themselves by 3-fold screw axis
surrounded by three identical strands of l/OO[PZSe.52'] of type A. On the other hand, K1-K6
are crystallographically unique and generated by 31 screw and 2-fold rotation operation
perpendicular to a— and b-axes. They are surrounded by two A and one B strands.
szPZSe6 did not show such a structural transition on cooling down to 100 K. The
different behavior can be understood by the alkali metal size effect. The most notable
change in going from the P3121 to P31 for the K analog is found in the coordination
environment of the K atoms. At room temperature K atoms are coordinated by six closely
lying Se atoms whereas this expands to seven at low temperature, Figure 4-4. This is
reminiscent of a pressure effect imposed on the structure as the unit cell contracts with
falling temperature. Therefore the driving force for the transition seems to be the
formation of the extra K---Se bonds. This efiect seems to be absent in the Rb analog as the
coordination environment of these atoms is already high at room temperature involving
ten Se atoms. Tables 4-7 — 4-10 show a comparison of the K---Se and Rb---Se distances in

the two structures for both room and low temperature.

98

 

Figure 4-4. The coordination environment of K atoms in K2P2Se6 is six at 298(2) K (a),
and expands to seven at 173(2) K (b). The coordination environment of Rb atoms in
szPZSe6 at (c) 298(2) K and ((1) 100(2) K. It remains unchanged at both temperatures
where Rb is coordinated to ten Se atoms. Large spheres are K or Rb atoms and small ones

are Se atoms. P atoms are omitted for clarity.

99

Table 4-5. Selected Bond Distances (A) and Angles (°) for KZPZSe6 at 298(2) K.

 

P-P 2.242(2)
P-Se(l) 2.329(1)
P—Se(2) 2.131(2)
P-Se(3) 2.128(1)
Se(l)-Se(l) 2.337(1)
Se(1)-P-Se(2) 1 1 188(6)
Se(1)-P-Se(3) 1 1229(6)
Se(2)-P-Se(3) 1 1795(6)
P-Se(1)—Se(l) 100.55(4)
P-P-Se(l) 9385(7)
P-P-Se(2) 1 1 142(4)
P—P-Se(3) 106.69(5)

100

Table 4-6. Selected Bond Distances (A) and Angles (°) for K2P28e6 at 173(2) K.

 

P( 1 )-P(2)
P(3)-P(4)
P(5)-P(6)
P(7)-P(8)
P(1)-Se(l)
P(1)-Se(2)
P(l)-Se(3)
P(2)-Se(4)
P(2)-Se(5)
P(2)-Se(6)
P(3)-Se(7)
P(3)-Se(8)
P(3)-36(9)
P(4)-Se(10)
P(4)-Se(] 1)
P(4)-Se(12)
P(5)-Se(13)
P(5)-Se(l4)
P(5)-Se(15)
P(6)-Se(16)

P(6)-Se(17)

2.227(17)
227(3)

2.236(18)
223(3)

2.297(17)
2.141(17)
2.154(17)
2.118(17)
2.151(17)
2.346(17)
2.305(17)
2.138(17)
2.119(18)
2.123(17)
2.133(18)
2.356(17)
2.286(17)
2.140(17)
2.155(16)
2.157(15)

2.129(16)

101

P(6)-Se(18)
P(7)-Se(19)
P(7)-Se(20)
P(7)-Se(21)
P(8)-Se(22)
P(8)-Se(23)
P(8)-Se(24)
Se(l)—Se(18)
Se(6)-Se(7)
Se(12)-Se(13)
Se(l9)—Se(24)
Se(1)-P(l)-Se(2)
Se(1)-P(1)-Se(3)
Se(2)-P(1)-Se(3)
Se(4)-P(2)-Se(5)
Se(4)-P(2)-Se(6)
Se(5)-P(2)-Se(6)
Se(7)-P(3)-Se(8)
Se(7)-P(3)-Se(9)

Se(8)-P(3)-Se(9)

2.363(17)
2.309(17)
2.125(17)
2.126(17)
2.113(16)
2.131(16)
2.368(17)
2.334(13)
2.330(8)
2.342(7)
2.347(12)
113.6(7)
111.1(7)
116.8(8)
120.0(7)
112.8(8)
109.4(7)
113.9(7)
112.9(7)

116.5(8)

Table 4-7. K—Se distances in K2P28e6 at 173(2) K with standard deviation in parentheses.
The maximum threshold for bond distances is 3.750A. The coordination number of K

atoms was determined to be seven.

 

K(l)-Se(l7) 3.345(14) K(3)-Se(22) 3.317(16)
K(1)-Se(14) 3.361(17) K(3)-Se(3)#5 3.317(17)
K(l)-Se(8)#12 3.363(17) K(3)-88(9) 3.348(17)
K(1)-Se(20)#7 3.391(16) K(3)-Se(16) 3.348(15)
K(l)—Se(23) 3.456(16) K(3)-Se(14)#5 3.505(16)
K(1)-Se(12)#9 3.473(16) K(3)-Se(6) 3.509(16)
K(l) Se(lO) 3.554(14) K(3)-Se(23) 3.582(15)
K(l)-Se(24) 3.842(17) K(3)-Se(18) 3.915(16)
K(2)-Se(20)#2 3.3 10(17) K(4) Se(23) 3.341(14)
K(2)-Se(4)#3 3.360(16) K(4)-Se(22)#4 3.345(14)
K(2)-Se(11) 3.433(17) K(4) Se(15) 3.359(15)
K(2)-Se(9)#5 3.393(17) K(4) Se(9) 3.388(15)
K(2)-Se(21) 3.358(17) K(4) Se(19) 3.483(15)
K(2)-Se(22) 3.595(16) K(4)-Se(] 1)#4 3.509(16)
K(2)-Se(24) 3.462(16) K(4) Se(l4) 3.646(15)
K(2)-Se(12) 3.940(16) K(4)-86(7) 3.836(15)

Table 4.7. (Cont’d) K-Se distances in K2P2$e6 at 173(2) K with standard deviation in

parentheses. The maximum threshold for bond distances is 3.750A. The coordination

number of K atoms was determined to be seven.

 

K(5)-Se(5)#5 3.321(15) K(7)-Se(10) 3.309(14)
K(5)-Se(10) 3.351(15) K(7)-Se(2)#6 3.325(15)
K(5)-Se(20)#7 3.352(15) K(7)-Se(]5)#12 3.363(15)
K(5)-Se(3)#5 3.367(15) K(7)-Se(5)#6 3.398(15)
K(5)-Se(23) 3.481(15) K(7)-Se(2)#5 3.470(15)
K(5)-Se(7)#12 3.484(15) K(7)-Se(18)#9 3.473(14)
K(5)-Se(2)#5 3.606(15) K(7)-Se(l 1)#9 3.616(15)
K(5)-Se(]9)#5 3.888(16) K(7)-Se(6) 3.811(14)
K(6)-Se(] 1)#9 3.307(15) K(8)-Se(8)#10 3.318(14)
K(6)-Se( 1 7)#1 1 3.329(16) K(8)-Se(]6)# 10 3.325( 14)
K(6)-Se(21)#1 1 3.339(15) K(8)-Se(4)#4 3.346(14)
K(6)-Se(5)#1 1 3.344(15) K(8)-Se(14) 3.380(15)
K(6)-Se(2)#6 3.423(16) K(8)-Se(16)#4 3.420(15)
K(6)-Se(13) 3.433(15) K(8)-Se(1)#10 3.478(14)
K(6)-Se(15) 3.547(14) K(8)-Se(3) 3.518(15)
K(6)-Se(1)#9 3.849(15)

 

Symmetry transformations used to generate equivalent atoms:

#lx+l,y,z #2 x+l,y+l,z #3x,y+l,z #4 -x+yl-1,-x+2,z-l/3

#5 -y+2,x-y+l,z+l/3 #6x,y—l,z #7 -y+l,x-yl~l,z+l/3

#8 -r+y+1,-r+l,z-l/3 #9x-l,y—l,z #le-l,y,z

#11-x+y,-x+1,z-1/3 #12-y+l,x-y,z+l/3

103

Table 4-8. K-Se distances in K2P2S66 at 273(2) K with standard deviation in parentheses.
The maximum threshold for bond distances is 3.750 A. The coordination number of K

atoms was determined to be six.

 

K-Se(2)
K-Se(3)#6
K-Se(3)#2
K-Se(2)#4
K-Se(2)#7
K-Se(1)#1
K-Se(2)#2
K-Se(1)#4

K-Se(3)#1

3.3367(18)
3.3388(18)
3.3695(19)
3.4263(18)
3.4537(17)
3.4834(17)
3.8395(19)
3.8731(18)

3.9244(19)

 

Symmetry transformations used to generate equivalent atoms:

#1 -x, -x+y, -z+1/3

#4 -x-1, -x+y, -z+1/3

#7 -y, x-y+2, z+1/3

#2 -x-1, -x+y-1, -z+1/3 #3 -x+y-2, -x, z-1/3
#5 -x+y-l, -x+l,z-1/3 #6 -y+l,x-y+2,z+l/3

#8 x-y+1, -y+2, -z+2/3 #9 y-l, x+1, -z

104

Table 4-9. Rb-Se distances in Rb2P2866 at 100 (2) K with standard deviation in
parentheses. The maximum threshold for bond distances is 4.000 A. Rb-Se distance was
not found up to 4.500 A over the threshold. The coordination number of Rb atoms was

determined to be ten.

 

Rb-Se(3) 3.4778(12)
Rb-Se(2) 3.5715(13)
Rb-Se(1) 3.6011(13)
Rb-Se(3)#1 3.4248(12)
Rb-Se(2)#4 3.4357(11)
Rb-Se(2)#7 3.5101(12)
Rb-Se(2)#10 3.7393(13)
Rb-Se(2)#8 3.7754(13)
Rb-Se(l)#5 3.8599(12)

Rb-Se(3)#l 3.8706(13)

 

Symmetry transformations used to generate equivalent atoms:

#1 -x+l, —x+y, -z+l/3 #2 y, x-l, -z #3 -x+2, -x+ytl, -z+1/3
#4x-1,y-l,z #5x,yt1,z #6y+l,x,-z #7x+1,y,z
#8yl'1,x-l,-z #9x,y—l,z #10x+l,ytl,z #llx-1,y,z

#12-x+1,-x+y+1,-z+1/3

105

Table 4-10. Rb-Se distances in RbngSe6 at 298(2) K with standard deviation in
parentheses. The maximum threshold for bond distances is 4.000 A. Rb-Se distance was
not found up to 4.500 A over the threshold. The coordination number of Rb atoms was

determined to be ten.

 

Rb-Se(3) 3.4714(23)
Rb-Se(2) 3.4735(25)
Rb-Se(3)#ll 3.5172(22)
Rb-Se(2)#9 3.5481(23)
le-Se(2)#5 3.6271(23)
Rb-Se(l) 3.6545(22)
Rb-Se(2)#6 3.7805(23)
Rb-Se(1)#7 3.8185(22)
Rb-Se(l)#5 3.9055(21)
Rb-Se(3)#3 3.9302(23)

 

Symmetry transformations used to generate equivalent atoms:

#1 -x+l, -x+y, -z+1/3 #2 y, x-l, -z #3 -x+2, -x+y+l, -z+1/3
#4x—1,y-1,z #5x,y+l,z #6yl-1,x,-z #7x+1,y,z
#8y+1,x-1,-z #9x,y—l,z #10x+l,y+l,z #11x-1,y,z

#12 -x+l, -x+y+1, -z+l/3

106

Synthesis, Reaction Chemistry and Characterization. The alkali metal
chalcophosphate ternary system is highly attractive as a model system to probe how flux
conditions influence the synthetic outcome and the structure of a resulting compounds A
reaction medium forms by simple in situ fission of AzQ/PZQS or P/Q and this flux is
conceptually defined as AxPy z. The basicity of the flux is determined by the A:P:Q
ratio. Higher A:P ratios impart stronger basicity to the flux. Lower A:Q ratios produce
lower basicity, etc. Alternatively, for the same A:P ratio the basicity can be increased in
going from the smaller to the larger alkali atoms. Therefore both the A : P ratio and the
nature of the A atoms could be conveniently used as a key reaction parameter to control
the formation of a product from being a simple molecular species, to being a more
complex discrete or extended structure. For example to obtain the classical molecular
salts A3PSe4 and A4PZSe6 (A = alkali metal), high ratios of A2Se:P2Ses (>2:l) are
required.

The polymeric K2P2886 is realized under less basic conditions, namely with
KZSezPZSes = 1:1, than the typical ternary alkali selenophosphates with molecular
structures. In fact, less basic conditions or lower flux temperatures tend to generate
longer or extended fragments.19 This could explain why more basic Rb or Cs analogues
to KZPZSe6 could not be obtained from identical reaction conditions. The slightly higher
basic character of Rb and Cs changes the reaction path to a more oxidized
chalcophosphates (i.e. the P5+ species RbPSe6 and CsPSe6). Rb2P2$e6 could be obtained
however by adjusting/correcting the flux basicity by adding 0.4 mole of P to the
Rb28ezP2885 = 1:1 ratio. This prevents the oxidation for P from 4+ to 5+. Thus it would

appear that extended structural motifs for chalcophosphates are more likely to be

107

produced by weak basic fluxes with smaller less Lewis basic alkali metals such as
potassium, sodium and lithium.

Conceptually, l/OO[P2Se62'] can be regarded as deriving from oxidative
polymerization of the [P2Seé]4' anion, see Eq (1). Although it was not synthesized in this
fashion, we speculate that an actual oxidation in solution, with e.g. 12 as the oxidant, may

in fact be successful if run under proper conditions.

 

2-
4- Se Se
Se Se
\ / 212 \P P Eq (1)
P —— Pa, 7 e" "”0 ‘-
/;‘ ‘04, Se( \ I’Se
Se § Se Sc? Se
Se Se n

The polymeric chain is stabilized with no help of coordinating to a metal and
represents a rare chalcophosphate anion. The only other example which is stabilized by
alkali cations is l/OO[PSe6'].6 The latter has PSe4 tetrahedra condensed with diselenide
(Se2z') groups.

Finally, the existence of a single crystallographic P site in the RT structure of
K2P2Se6 is supported by solid state 3'P NMR spectroscopy. Under MAS at RT, crystalline
K2P2Se6 gave a single isotropic chemical shift (CS) at 54.6 ppm, Figure 4-5. The result
supports the crystal structure of K2P2Se6 at room temperature which indicates only one
crystallographically unique P atom. The chemical shift value is close to that of A2CdP2886

(A = K, Rb) and K2Cu2P4Se10, all of which include the P2Se6-type ligand.20

108

l
A i 1":
1, ..: 1'1

1 1‘
i A . 1." in .‘ 1A A“
1' V) '- l 1 .".A ”‘1“ *“WA’ASA ‘ “11") 1‘1“
V AV! 1 . 1 J“ ‘1 711:.) 1 ’jr '2.” 7., , A f ‘ r‘r l ~ In“. ' .
_ 1- ..w g‘. _,i ‘1..,71‘ .,1
. ,1 1' . 1.. .',4.,.g,-.,»,f.1’,v
1

lililllll‘lllllAllllll|Alllllflllllllllllllll
300 200 100 0 -100
ppm

Figure 4-5. Solid state 3|P MAS NMR spectrum of crystalline K2P2Se6 at room

 

temperature. The asterisk (*) indicates the isotropic peak.

109

 

 

 

 

 

_30 1 1 1 1 ~
250 300 350 400 450 500

Temperature, °C

Figure 4-6. Differential thermal analysis diagram of a sample of K2P2Seo. (a) Heating
curve showing melting at 387 °C in the first heating cycle (solid line) with no
crystallization upon cooling. (b) Exothermic crystallization followed by melting upon
heating in the second cycle (dotted line). Asterisk indicates the vitrification event upon

cooling.

110

Glass Formation, Phase-change Behavior and Local Structure. Both title
compounds described here exhibit reversible phase change behavior. Differential thermal
analysis of K2P2Se6 performed at a rate of 10 °C min'l indicates that the compound melts
congruently at 387 °C, Figure 4-6. Upon cooling it forms a dark red glass at 270 °C.
Crystallization is only observed on heating where the glass recrystallizes exothermically
at 297 °C. The vitrification is observed again on cooling the melt to room temperature.
The powder X-ray diffraction patterns (PXRD) of the pristine and recrystallized sample
from DTA were identical confirming the structure recovering ability of this material.

The thermal behavior of Rb2P2Se6 is different from that of K2P2Se6. The DTA
carried out at a rate of 10°C/min revealed the melting at 364 °C on heating and gave a
dark red glass at 255 °C on cooling during a first cycle. On a second cycle, no
crystallization was observed as the glass could not recover the crystal structure. The
melting was observed again at 369 °C followed by vitrification at 255°C on cooling. The
PXRD confirmed the amorphous nature of the glassy sample after the first and second
cycles. The amorphous sample recrystallized to pristine structure only when the heating
rate was lowered to less than 5°C/min. The DTA performed at a rate of 5 °C min'l showed
melting at 366 °C on heating and subsequent crystallization at 305 °C on cooling. The
PXRD of pristine and the sample obtained after each cycle matched perfectly. These
results are consistent with a greater glass forming ability of the Rb+-salt compared to the
K+-salt. This is in agreement with a similar trend identified recently in the series K1-
xbeSbsSg.2'c

The solid-state optical absorption UV-vis spectra of crystalline and glassy

A2P2866 show sharp absorption edges, Figure 4-7. The band gap of crystalline and glassy

111

phases was measured at 2.08 and 1.97 eV for K+ salt and 2.32 and 2.10 eV for Rb+ salt,
respectively. The energy gaps are consistent with dark orange/dark red color for K+ salt
and orange/dark orange color for Rb+ salt, respectively. A red shift in absorption edge in
the glassy phases is a common phenomenon as glass formation generally induces
substantial defects and mid-gap states to give a lower band gap than the crystalline
counterpart.” This creates optical contrast and can be a useful feature for optical storage
systems based on phase-change materials.23

The results of Raman spectroscopy and PDF analysis (see below) shed light on
the local structure in the glass and how these materials can readily recover their crystal
structure from the amorphous state. The Raman spectrum of crystalline K2P2Se6 at room
temperature shows shifts at 154(bw), 221(5), 236(m), 259(m), 388(w), 496(w), 514(vw)
cm", Figure 4-8. The shift at 221 cm'1 is unambiguously assigned to the P2Se6 stretching
mode by comparison in the Ag stretching mode of D3d symmetry of [P2Se6]4' ligand.24
Other peaks at 154, 496 and 514 cm'1 are also related with the P2866 fragment.25 By
analogy, shifts at 236 and 259 cm'1 can be assigned to antisymmetric and symmetric Se-
Se stretching vibration modes of the diselenide group, respectively.26

The Raman spectrum of glassy K2P2S86 shows broader and weaker peaks at 162
(bw), 218 (bm) and 258 (bw) are observed whereas the overall peak pattern is similar to
that of the crystals. This suggests that the [P2Se6] unit and Se—Se bonds are still intact and
the local structural motifs are largely preserved in the glass but crystallographic long-
range order is lost. This situation makes it easy to restore the crystal structure from the

amorphous state for the reversible crystal-glass phase transition.

112

3'5 l l I T l l l

 

   

 

 

 

 

a"

3 - _
A 2.5 - 4
3'
16'
v 2 _ —o—K2P2Se6 Crystal -
g —e— Rb2PZSe6 Crystal
-;_-, —I— K2P28e6 Glass
9. — Rb2P2$e6 Glass
0 1.5 r —
(D
.Q
<

1 t _

0.5 - -
O 1 1 1 I

 

 

 

1 1.52.5 3 3.5 4 4.5 5
Energy, eV

Figure 4-7. Solid state UV-vis optical absorption spectra of crystalline and glassy
K2P2Se6 and Rb2P2Se6 showing the red shift in absorption edge in the glass samples. The
band gaps are 2.09, 1.98 eV, for the K and 2.32 and 2.10 eV for Rb analogs respectively.

113

 

 

 

 

 

 

 

 

 

 

 

 

 

I l l l
221
30 r —
236
3. 20 T / 259 "—1
9'.
2? 218 Crystalline K2P2Se6
m 10 — -
0:) 496
4.)
E 154 388 514
0 . ._
258
162
JV Glassy K2P2586
-10 ~ A"‘— a
1 l 1 l
100 200 300 400 500

Raman Shift, cm-1

Figure 4-8. Raman spectra of crystalline (upper line) and glassy (lower line) K2P2386 at
room temperature. The similar but broader features in the spectrum of the glass suggest

the local structure is preserved but long range order is lost.

114

To probe the local structure of both the glassy and crystalline forms of K2P2866
we performed pair distribution function (PDF) analysis. This technique is emerging as a
useful tool for the analysis of the local structure of crystalline and non-crystalline
compounds.27 The PDFs are shown in Figure 4-9. Also shown in the middle profile is the
calculated PDF based on the room temperature single crystal structure model of K2P2Se6.
There appears to be good agreement between the crystal structure model and the
experimentally determined PDF of K2P2Se6. This in fact validates the correctness of the
crystal structure.

The PDF of the glass shows well defined correlations up to ~8 A with the first
two at 2.2 and 3.6 A being very similar to those of the crystalline form. The interatomic
correlations in the structure disappear above ~8 A indicating the lack of long range order
periodicity. The first strong correlation at 2.2 A is assigned to Se-Se, P-Se and P-P bonds
in the structure. The second strong peak at 3.6 A is assigned to K---Se and second
neighbor Se-~-Se distances. The PDF data suggest that the crystalline and glassy forms of
K2P2Se6 are structurally similar which implies that the [P2Se6] units remain intact in the
glass form. This is also in agreement with the Raman spectroscopic data discussed above.
The close structural relationship between crystal and glass, coupled with the
stoichiometric nature of the composition, accounts for the facile congruent crystallization
of the glass. Similar conclusions were reached from the PDF analysis of KSbsSg, another
interesting phase-change material, which shows that its local structure within a radius of

~5 A in the glass phase is clearly refined even though long range order is not observed.2|

115

 

 

PDF G(r)
; E: -

l.. .
'00. ....O

1’1 r ..'°°°Ooo. .°"-ouo
"on '00-
...... ....................... c.....-...................
\. ... .ll ......

O. o.

......I... .....O.
n

._ .m .1311,"
A‘ ’
1

I
.1:- on
O 01
E§ I
(A;
W‘I n lrL {L
11111“
I g I

9'»
O
I l V

 

 

I I I I I I I I I I I I I I I J J L I I I 1 I4 I

1'. 1 AI L1 1 I. 1
12534567891011121314151617181920
Radial distance (A)

 

Figure 4-9. Pair distribution function G(r) of the crystalline and glassy K2P2Seb. The
calculated PDF based on the room temperature single crystal structure model of K2P2Se6

is presented for comparison.

116

 

100 | l I l

 

 

 

 

 

 

 

 

80 — —
9 l
g 60 —
C
.2
‘8
8 4O -
.0
<1:

20 Band-gap -

\O
1.. ad
0 1 1 1 1
20 15 10 5 ' O

Wavelength, m

Figure 4-10. Far-IR (line with X)/mid IR (simple 1ine)/visible (line with 0) absorption
spectra of crystalline K2P2Se6. Wide transparent range of crystalline K2P2Se6 above the
absorption band at 19.8 pm at far-IR region through mid-1R to 0.596 pm at visible region

is shown.

117

Infrared Transmission and Nonlinear Optical Properties. Materials with large NLO
susceptibilities for IR application are highly sought. AgGaQ2 (Q = S, Se), ZnGeP2 and
G888 are main infrared materials28 and we recently introduced the promising long wave
IR NLO compound ,8-K2Hg30e2Sg.29 The chalcophosphates are an attractive class in
which to search for new NLO materials in that basic building units of [PQ4]3' or [P2Q6]4'
frequently form noncentrosymmetric arrangements by coordination to central metals or
polychalcogenide fragments [Qn]2'. Examples include APSe6 (A=K, Rb),5 ANb2P2S.2
(A=K, Rb, Cs),'6 AHU7(PS4)13 (A=K, Rb),'7 C52CuP389,3° Naojpbmps.“ and
A3AuP2Seg (A=K, Rb, Cs).32 In addition, chalcogenide compounds demonstrate better
polarizability than oxide compounds, which have been predominantly studied for NLO
application and wide IR transmission.

K2P2Se6 exhibits wide optical transparency ranging from long wave IR (LWIR) to
near IR (NIR)/visible light, Figure 4-10. The mid-IR transmittance spectrum showed little
absorption from 505 cm‘1 (19.8 m) to 4000 cm'1 (2 pm). There is no light absorption
below the band gap transition suggesting uninterrupted light transmission in the
compound. The optical transparency extends over to its absorption edge of 2.08 eV (596
nm) in the visible region. Above 19.8,um in the far-IR region, the compound exhibited a
complex set of absorptions, consistent with its Raman and far-IR spectra. Optical
transparency is a key feature for materials aimed at NLO applications. For example, the
important NLO material for IR applications, AgGaSe2,28 shows LWIR transmission up to
17 pm.

The polar, noncentrosymmetric helical chain structure of l/C,O[P2Se(r,2'] composed

of easily polarizable P and Se atoms linked by covalent bonding can produce large optical

118

nonlinearity. SHG measurements were performed using a modified Kurtz powder
method}3 with an IR light source ranging from 1000 nm to 2000 nm. SHG intensities of
crystalline K2P2Se6 were directly compared with that of AgGaSe2 powder. All samples

were prepared in a similar fashion and the same particle size range of 45.5i7.5 ,um was

measured and compared. Crystalline K2P2Se6 generated strong double frequency signals
from the fundamental idler beam. The SHG intensity of K2P2566 showed a maximum at
789 nm which is ~50 times larger than that of in the same wavelength. Under the same
experimental conditions AgGaSe2 showed a SHG maximum at 890 nm and in this
wavelength the corresponding response of K2P2Se6 was 20-fold higher. At shorter
wavelengths the K2P2866 outperforms the chalcopyrite material by over 100 fold, Figure
4-lla. These results demonstrate that the crystalline K2P2Se6 is very promising in IR
NLO application. For comparison, the absolute nonlinear optical susceptibility at 2.12pm
of AgGaSe2 and LiNbO3 is 67.7:l:13 and 29.11252 pm V", respectively.34

The SHG intensity of the crystalline phase increased with the particle size and
reached a plateau, Figure 4-11b. In principle, phase-matchable samples reach maximum
intensity and then for larger average particle sizes the intensity is size-independent
because of the existence of a phase-matching direction in the sample.33 In this regard,
crystalline K2P2Se6 is type 1 phase-matchable and eminently suitable for consideration in
applications.

These results suggest that K2P2Se(, is of special interest for the middle and deep
infrared (IR) applications due to its large nonlinear optical coefficients and high
transmission in the 1R region. The phase matching and transmission characteristics of

K2P2Se6 should allow 3-wave interactions in the mid and near IR, particularly for optical

119

parametric oscillator (OPO) devices pumped with Nd:YAG laser, frequency mixing of
OPO outputs pumped by Ti:Sapphire or Nd:YAG laser, as well as frequency mixing
Nd:YAG laser with dye and Ti:Sapphire or other laser sources. K2P2866 is also promising
as an efficient frequency doubling crystal for infrared radiation such as 10.6 um output of

CO2 lasers.

Second Harmonic Generation Response of Glassy K2P2Se6. Surprisingly, the
glassy K2P2Se6 powder exhibited SHG response and its intensity is 38 % that of AgGaSe2.
Because of great optical transparency and potential formability, there have been

tremendous efforts to induce SHG in glasses35

for use in optical fibers in
telecommunications. The observation of significant NLO activity in K2P2S86 glass could
be a rare example of this property found in an amorphous material with no specific
treatment such as thermal poling, electron beam irradiation and so on.36 Since K2P2Se6 is
a phase-change material and retains its local structural motif in the glassy phase, we
expect the noncentrosymmetric arrangement to be partially preserved and consequently to
exhibit some SHG response. Indeed, the SHG response of the K2P2Se6 glass started to be
observed at longer wavelength than that of the crystal. This suggests that the SHG signal
was generated from the glass of which the bad-gap is red-shifted relative to the crystal,
and probably did not originate from traces of the recrystallized phase embedded in the
glass matrix. X—ray powder diffraction patterns after the SHG measurements did not show
evidence of crystallization. It cannot be completely ruled out, however, that idler beam-

induced crystallization of glass may be occurring. Additional work will be necessary to

better understand the NLO properties of the glass.

120

65000

60000

55000

50000

Relative SHG Efficiency

45000

 

i

I

I

l

 

(a)

1—H

I

 

I I

 

 

40

I 60

80

‘100 I120 .140 A160

Relative SHG efficiency

Particle Size, pm

250

 

200 -

150 -

100

0|
0
, .
I
I
I

0.

 

L I l I I I I I

 

 

600 650 700 750 800 850 900 950

1000

Wavelenth, nm

Figure 4-11. (a) Particle size to SHG intensities diagram of crystalline K2P2Se6 showing
type-I phase-matching. (b) SHG response of K2P2Se6 relative to AgGaSe2 over a wide

range of wavelengths.

121

4.7. Thermal properties of K2P2Se6. The thermal properties such as thermal
expansion, are important properties relevant to crystal growing for NLO applications. The
thermal expansion of K2P2Se6 crystal was determined by a single crystal X-ray diffraction
study in the range of 100 — 400 K, Figure 4-12. The linear (Eg. 4-2) and thermal
expansion coefficients (Eg. 4-3) are defined as

1 dL,

a: E.2

’Lth g()
ldV

=—— E.3

where L, is the unit cell dimension along t-axis and V is the unit cell volume. Because
K2P2Se6 exhibited 2a >< 2b X c supercell, a-axis distance was divided by 2 for the
calculation. The thermal expansion coefficients of K2P2Se6 were calculated to be ad =
1.46 X 10'5 K'1 and ac = 2.41 x10"5 K", indicating somewhat isotropic expansion. By
comparison corresponding values for other relevant infrared NLO materials such as
AgGaSe2 (a .. c = -0.81 x 10'5 K", a“: 1.98 x 10'5 K", 298-423 K),37 AgGaS2 (a .I c = -
1.32 x 10'5 K", a“ = 1.27 x 105/K, 298-523 K)38 and ZnGeP2 (a .. c = 1.59 x 10'5 K",
a“ = 1.75 x 10‘5 K", 293—573 K).39 The aof stainless steel is 1.73 x 10'5 K" at 293 K.
Most importantly, K2P2Se6 does not exhibit the anomalous behavior of AgGaQ2 (Q = S,
Se) which contracts along the a- and b-axes and expands along the c-direction. The latter
is an anomaly which can cause thermomechanical stress in high-power applications and

. 40
failure.

122

 

a-axis

 

 

 

XI
AAA—IA
9090909090
QCOCOCOCD
(DON->0)
Trill

 

 

 

861

Cell Volume

858

 

 

 

855 ‘

J 4 I l I I I I I I I

1 00 1 50 200 250 300 350 400

 

Temperature, K

Figure 4-12. Temperature variation of the lattice parameters and cell volume for K2P2Se6
from 100 to 400 K.

123

The volumetric thermal expansion coefficient of K2P2Se6 at ,6 = 5.35 X 10'5 K"
also indicates its moderate thermal expansion. It is noteworthy that crystals of K2P2Se6
did not show any signs of cracking or deformation when exposed to abrupt temperature

change (200 K h") or soaked at high temperature of 500 K for several days.

5. Concluding Remarks

The compounds A2P2Se6 (A = K, Rb) have novel chiral structures composed of
helices of 1/00[P2Se62']. The chirality combined with its phase-change behavior could
enable A2P2Se(, to be explored as a multi functional switchable material. On cooling
K2P2Se6 is susceptible to a displacive phase transition to a lower space group that seems
to be driven by the coordination sphere expansion of the alkali ions. If correct, the phase
transition could also be effected by the application of high pressure as a means of
achieving coordination sphere expansion. When molten the compounds can be quenched
to a glassy state and exhibit reversible crystal-glass phase change behavior. The optical
absorption edge for the glassy phase is red shifted and on the basis of Raman
spectroscopy and PDF analysis it appears that the main building motifs are largely intact
in the glassy form. This accounts for the facile restoration of the crystal structure from the
corresponding amorphous phase.

A2P2Se6 (A = K, Rb) and the previously described AP566 (A = K, Rb, Cs) are the
only reported extended structure compounds in the alkali chalcophosphate family to
exhibit easy glass-formation and phase-change properties. This suggests that polymeric

chalcophosphates have broad glass-forming tendencies and may constitute a fertile source

124

for stoichiometric chalcogenide glasses. The glass forming property of these materials
makes them potentially valuable for producing IR optical glass fibers. Finally, K2P2Se6
widely transparent in the mid-IR up to 19.2 ,um that is coupled with a very large SHG
response. The material is type-I phase-matchable with a response that is over 50 times
larger than that of top performing NLO material AgGaSe2. The remarkably stronger
response is attributed to the helical structure of the selenophosphate which is highly polar,
in contrast to the weakly polar chalcopyrite structure of AgGaSe2. This makes K2P2Se6
one of the best NLO materials known for non-resonant SHG Finally, glassy form of

K2P2Se6 exhibited surprising SHG response with no poling treatment.

References

(1) Garin, J .; Parthe, E. Acta Crystallogr. B 1972, 28, 3672-3674.

(2) Francisco, R. H. P.; Tepe, T.; Eckert, H. J. Solid State Chem. 1993, 107, 452-459.
(3) Chondroudis, K.; Kanatzidis, M. G Inorg. Chem. 1995, 34, 5401-5402.

(4) Chondroudis, K.; Kanatzidis, M. G Inorg. Chem. 1998, 3 7, 2582-2584.

(5) Chung, 1.; Karst, A. L.; Weliky, D. P.; Kanatzidis, M. G Inorg. Chem. 2006, 45,
2785-2787.

(6) Chung, 1.; Do, J .; Canlas, C. G.; Weliky, D. P.; Kanatzidis, M. G Inorg. Chem.
2004, 43, 2762-2764.

(7) (a) Chondroudis, K.; Kanatzidis, M. G Angew. Chem, Int. Ed. Engl. 1997, 36,
1324-1326. (b) Chung, 1.; Kanatzidis, M. G Unpublished work.

(8) Kanatzidis, M. G Curr: Opin. Solid State & Mater: Sci. 1997, 2, 139-149.

(9) The chalcophosphate, for example, are capable of condensing via Q-Q
polychalcogenide bonds, whereas the phosphates are not known to do so.

(10) (a) Oh, H.-R.; Cho, B.-H.; Cho, W. Y.; Kang, S.; Choi, B.-G; Kim, H.-J.; Kim,
K.-S.; Kim, D.-E.; Kwak, C.-K.; Byun, H.-G; Jeong, G-T.; Jeong, H.-S.; Kim, K. IEEE J.
Solid-State Circuits 2006, 41, 122-126. (b) Ohta, T. J. Optoelectron. Adv. Mat. 2001, 3,
609-626. (c) Ohta, T.; Nishiuchi, K.; Narumi, K.; Kitaoka, Y.; Ishibashi, H.; Yamada, N.;
Kozaki, T. Jpn. J. Appl. Phys, Part I 2000, 39, 770-774.

(11) (a) Welnic, W.; Pamungkas, A.; Detemple, R.; Steimer, C.; Blugel, S.; Wuttig, M.
Nat. Mater: 2006, 5, 56-62. (b) Lee, H.; Kim, Y. K.; Kim, D.; Kang, D. H. IEEE Trans.
Magn. 2005, 41 , 1034-1036. (c) Kolobov, A. V.; Fons, P.; Frenkel, A. 1.; Ankudinov, A.
L.; Tominaga, J .; Uruga, T. Nat. Mater: 2004, 3, 703-708.

(12) Hwang, S. J.; Petkov, V.; Rangan, K. K.; Shastri, S.; Kanatzidis, M. G J. Phys.
Chem. 82002, 106, 12453-12458.

(13) Qiu, X. Y.; Bozin, E. S.; Juhas, P.; Proffen, T.; Billinge, S. J. L. J. Appl.
Crystallogr: 2004, 37, 110-116.

(14) (a) Kortiim, G Reflectance spectroscopy. Principles, methods, applications;
Springer: Berlin, Heidelberg, New York,, 1969. (b) Tandon, S. P.; Gupta, J. P. Phys Status
Solidi 1970, 38, 363-367. (c) Wendlandt, W. W.; Hecht, H. G Reflectance spectroscopy;
Interscience Publishers: New York, 1966.

1'16

(15) SMART: SAINT SHELXT L: Data Collection and Processing Software for the
SMART-CCD System; Siemens Analytical X-ray Instruments Inc.: Madison, WI, 1997.

(16) Gieck, C.; Derstroff, V.; Block, T.; Felser, C.; Regelsky, G; Jepsen, O.;
Ksenofontov, V.; Gutlich, P.; Eckert, H.;Treme1, W. Chem. Eur. J. 2004, 10, 382-391.

(17) Gieck, C.; Tremel, W. Chem. Eur: J. 2002, 8, 2980-2987.

(18) (a) Canadell, E.; Rachidi, 1. E. 1.; Pouget, J. P.; Gressier, P.; Meerschaut, A.;
Rouxel, J .; Jung, D.; Evain, M.; Whangbo, M. H. Inorg. Chem. 1990, 29, 1401-1407. (b)
Electronic Properties of Inorganic Quasi-One-Dimensional Compounds; Monceau, P.
ed.; D. Reidel Publishing Company: Dordrecht, 1985.

(19) Kanatzidis, M. G; Sutorik, A. C. Prog. Inorg. Chem. 1995, 43, 151-265.
(20) Canlas, C. G; Kanatzidis, M. G; Weliky, D. P. Inorg. Chem. 2003, 42, 3399-3405.

(21) (a) Chrissafis, K.; Kyratsi, T.; Paraskevopoulos, K. M.; Kanatzidis, M. G Chem.
Mater. 2004, 16, 1932-1937. (b) Kyratsi, T.; Chrissafis, K.; Wachter, J.; Paraskevopoulos,
K. M.; Kanatzidis, M. G Adv. Mater. 2003, 15, 1428-1431. (c) Wachter, J. B.; Chrissafis,
K.; Petkov, V.; Malliakas, C. D.; Bilc, D.; Kyratsi, T.; Paraskevopoulos, K. M.; Mahanti,
S. D.; Torbrugge, T.; Eckert, H.; Kanatzidis, M. G J. Solid State Chem. 2007, 180, 420-
431.

(22) Dhingra, S.; Kanatzidis, M. G Science 1992, 258, 1769-1772.

(23) (a) Maeda, Y.; Andoh, H.; lkuta, I.; Minemura, H. J. App]. Phys. 1988, 64, 1715-
1719. (b) Feinleib, J.; DeneufviJ; Moss, S. C.; Ovshinsk.Sr App]. Phys. Lett. 1971, 18,
254-257.

(24) (a) Piitzmann, U.; Brockner, W. Z. Naturforsch., A 1987, 42, 515-516. (b)
Brockner, W.; Ohse, L.; Péitzmann, U.; Eisenmann, B.; Schéifer, H. Z. Naturforsch., A
1985, 40, 1248-1252.

(25) Aitken, J. A.; Evain, M.; lordanidis, L.; Kanatzidis, M. G Inorg. Chem. 2002, 41,
180-191.

(26) (a) Choi, K. S.; Kanatzidis, M. G Chem. Mater: 1999, 11, 2613-2618. (b) Evain,
M.; Queignec, M.; Brec, R.; Sourisseau, C. J. Solid State Chem. 1988, 75, 413-431. (c)
Butler, 1. 3.; Harvey, P. D.; Mccall, J. M.; Shaver, A. J. Raman Spectrosc. 1986, 17, 221-
228.

(27) (a) Billinge, S. J. L.; Kanatzidis, M. G Chem. Commun. 2004, 749-760. (b)
Petkov, V.; Billinge, S. J. L.; Shastri, S. D.; Himmel, B. Phys. Rev. Lett. 2000, 85, 3436.

(28) Nikogosyan, D. N. Nonlinear optical crystals: a complete survey; Springer-

127

Science: New York, 2005.

(29) L180, J. H.; Marking, G M.; Hsu, K. P.; Matsushita, Y.; Ewbank, M. D.; Borwick,
R.; Cunningham, P.; Rosker, M. J .; Kanatzidis, M. G J. Am. Chem. Soc. 2003, 125, 9484-
9493.

(30) Hanko, J. A.; Kanatzidis, V. G J. Solid State Chem. 2000, 151, 326-329.

(31) Aitken, J. A.; Marking, G A.; Evain, M.; lordanidis, L.; Kanatzidis, M. G J.
Solid State Chem. 2000, 153, 158-169.

(32) Chondroudis, K.; Hanko, J. A.; Kanatzidis, M. G. Inorg. Chem. 1997, 36, 2623-
2632.

(33) (a) Dougherty, J. P.; Kurtz, S. K. J. Appl. Crystallogr. 1976, 9, 145-158. (b) Kurtz,
S. K.; Perry, T. T. J. App]. Phys. 1968, 39, 3798-3813.

(34) Choy, M. M.; Byer, R. L. Phys. Rev. B 1976, 14, 1693-1706.

(35) Antonyuk, B. P.; Antonyuk, V. B. Opt. Commun. 1998, 14 7, 143-147.

(36) (a)Fokine, M.; Saito, K.; Ikushima, A. J. Appl. Phys. Lett. 2005, 87, 171907. (b)
Kityk, I. V. J. Phys. Chem. B 2003, 107, 10083-10087. (c) Qiu, M. X.; Pi, F.; Orriols, G
Appl. Phys. Lett. 1998, 73, 3040-3042. (c) Fujiwara, T.; Takahashi, M.; Ikushima, A. J.
Appl. Phys. Lett. 1997, 71, 1032-1034.

(37) Iseler, G W. J. Cryst. Growth 1977, 41, 146-150.

(38) Korczak, P.; Staff, C. B. J. Cryst. Growth 1974, 24, 386-389.

(39) Kozhina, I. I.; Borshchevskii, A. S. Vestnik Leningradskogo Universiteta, Seriya
4: F izika, Khimiya 1971, 87-92.

(40) Feigelson, R. 8.; Route, R. K. Optical Engineering 1987, 26, 113-19.

128

ChapterS
[P6Se12]4': A Phosphorus-rich Selenophosphate with Low-valent P

Centers

1. Introduction

The diversity found within the class of metal chalcophosphates is extensive.l
Typically, these are ternary (M/P/Q) and quaternary (A/M/P/Q) compounds with [PnyF'
anions in their structure, where M is a metal, A is an alkali metal, and Q is sulfur or
selenium. The selenophosphate anions that are structurally characterized include [PSe4]3'
,2 [PSe4]3'-2Se(,,3 {P2Se6]4‘,4 [stegr'f [P2Se10]4',6 [PgSe13]6‘,7 [9288,1238 and the one-
dimensional polymer 1/00[PSe6'].9 All of these are P5+ species except for [P2Se6]4' and
[PgSe13]6', which are formally P“ and P4+/P3+ compounds with P-P bonds, respectively.
Each anion is capable of coordinating to a metal (M) and giving rise to extended
structures."10 Core questions in this chemistry include the limits of structural diversity
and the stabilization of species with P in even lower oxidation states. Consequent
phosphorus-rich phases have been rare in this chemistry. With this in mind, we conducted
experiments aimed at stabilizing alkali salts of chalcophosphate anions by employing
new synthetic conditions. In the present case, we attempted reductive reactions using
P5+[Pny]z' species as starting materials with an excess of elemental P. This chemistry
was carried out at ~400 °C with RbPSe6 and red P and resulted in the new compound
Rb4P68612.Here we describe the new polar chalcophosphate anion [P6Se12]4', featuring P

in two different oxidation states of 2+ and 4+. This is manifested in three parallel P-P

bonds in the molecule. Direct combination reactions of Rb2Se, P, and Se with the correct
stoichiometric ratio could not produce this compound. The result suggests a unique
suitability of alkali chalcophosphates as starting materials to explore new chemistry. This

is partly due to their low melting points (300-400 °C) and high reactivity.

2. Experimental Section

2.1 Reagents

The reagents mentioned in this work were used as obtained: Rb metal (analytical
reagent, Johnson Matthey/AESAR Group, Seabrook, NH); red phosphorus powder, -100
mesh, Morton Thiokol, Inc., Danvers, MA; Se (99.9999%; Noranda Advanced Materials,
Quebec, Canada); N,N-dimethylformamide (Spectrum Chemicals, ACS reagent grade);
diethyl ether (Columbus Chemical Industries, Columbus WI, ACS reagent grade,
anhydrous). Rb2Se starting material was prepared by reacting stoichiometric amounts of

the elements in liquid ammonia.

2.2. Synthesis

Pure Rb4P6Se12 was achieved by a mixture of RbPSe6zP=l :4 in an evacuated and
sealed silica tube at 400 °C for 3 days followed by cooling at a rate of 5 °C h" to 250 °C.
After washing with degassed N,N-dimethylformamide (DMF) and ether under N2
atmosphere, we obtained pure orange irregular-shaped single crystals. Energy dispersive
spectroscopy (EDS) analysis of the crystals showed an average composition of

“Rb3,9P6Se”,7” for five single crystals. The single crystals are stable in acetonitrile, DMF,

130

and deionized water, and they are air-stable for over 1 week.

3. Physical Measurements

X-ray Powder Diffraction. Phase purity x-ray diffraction analyses were performed using
a calibrated CPS 120 INEL X-ray powder diffractometer (Cu Ka graphite
monochromatized radiation) operating at 40 kV/20 mA and equipped with a position-
sensitive detector with flat sample geometry.

Electron Microscopy. Semiquantitative analyses of the compounds were performed with
a J EOL J SM-35C scanning electron microscope (SEM) equipped with a Tracor Northern
energy dispersive spectroscopy (EDS) detector.

Solid-State UV-Vis spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region using a
procedure described in detail in Chapter 3.

Raman Spectroscopy. Raman spectra were recorded on a Holoprobe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation from a HeNe
laser for excitation and a resolution of 4 cm". Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 pm. A total of 1.28 scans
was sufficient to obtain good quality spectra.

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CsI matrix. The
sample was ground with dry CsI into a fine powder and pressed into translucent pellets.

The spectra were recorded in the far-IR region (600-100 cm", 4 cm'l resolution) with the

131

use of a Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon

beam splitter.

Differential Thermal Analysis (DTA) Experiments were performed on Shimadzu DTA-
50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. The sample was heated to 600 °C
at 10 °C min", and after 1 min it was cooled at a rate of -10 °C min" to 50 °C. The
residues of the DTA experiments were examined by X-ray powder diffraction.
Reproducibility of the results was confirmed by running multiple heating/cooling cycles.
The melting and crystallization points were measured at a minimum of endothermic peak
and a maximum of exothermic peak.

Solid-State Nuclear Magnetic Resonance (NMR) Spectroscopy. The room temperature
solid-state NMR measurement was taken on a 9.4 T NMR Spectrometer (Varian Infinity
Plus) using a double resonance magic angle spinning (MAS) probe. The sample was spun
at 8.0 kHz using zirconia rotors of 6 mm outer diameter. Bloch decay spectra were taken
with the excitation/detection channel tuned to 31P at 161.39 MHz, a 4.5 ,us 900 pulse, and
a relaxation delay of 3,000 s. The spectrum was referenced using 85% H3PO4 at 0 ppm.
X-ray Crystallography. The crystal structure was determined by single-crystal X-ray
diffraction methods. Preliminary examination and data collection were performed on a
SMART platform diffractometer equipped with a 1K CCD area detector using graphite
monochromatized Mo Ka radiation at 293(2) K. A hemisphere of data was collected at
293(2) K using a narrow-frame method with scan widths of 0.300 in a) and an exposure

time of 305 frame". The data were integrated using the SAINT program. An analytical

132

absorption using the program SADABS was performed. The initial positions for all atoms
were obtained using direct methods and the structures were refined with the full-matrix
least-squares techniques of the SHELXTLll crystallographic software package.
Crystallographic data are given in Table 1. Satisfactory refinement was obtained with the
chiral space group, Pca21 at 293(2) K. Absolute structure parameter was refined at
0.07(3). The PLAT ON program12 could not suggest additional symmetry, and the

structure could not be solved in a centrosymmetric space group such as Pbcm

133

Table 5-1. Crystallographic Data and Structure Refinement for Rb4P6Sel2.

 

Formula

Formula weight, g/mol
Temperature, K
Wavelength, A

Crystal system

Space Group

Unit cell dimensions, A
Volume, A3

Z, Density (calculated)
Absorption coefficient

F (000)

Theta range for data collection, deg
Index ranges

Reflections collected/unique
Completeness to Hmax, %
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2

Final R indices [1>20(1)]

R indices (all data)

Absolute structure parameter

Largest diff. peak and hole

Rb4P68e12

1475.22

293(2)

0.71073

Orthorhombic

Pca21

a = 16.409(3), b = 10.640(2), c = 15.105(3)
2637.1(9)

4, 3.716 g cm"3

24.296 mm"

2584

1.91 to 28.29

-21<=h<=20, -l 4<=k<= 14, -19<=l<=19
22687, 6163

96.8

Full-matrix least-squares on F2

6163 / 1 / 199

0.854

R1 0 = 0.0463, wR2 b = 0.0769
R1 = 0.1726, wR2 = 0.1019
007(3)

1.451 and -1.380 e.A'3

 

8 RI = z llFol- chll/ZlFol- b wR2 = {2 [warez-Fez) 2121w(F02)21}’/2.

134

Table 5-2. Atomic coordinates (X 104) and equivalent isotropic displacement parameters

(A2 X 103) for Rb4P6Se12. U(eq) is defined as one third of the trace of the orthogonalized

 

 

U ij tensor.

x y z U(eq)
Rb(l) 1635(1) 7610(2) 10196(1) 33(1)
Rb(2) -890(1) 7356(2) 8418(2) 37(1)
Rb(3) 4744(1) 7632(2) 8022(2) 37( 1)
Rb(4) -2229(1) 721 1(2) 5438(2) 47(1)
Se(l) 1993(1) 2746(2) 7859(2) 33(1)
Se(2) 790(2) 5152(2) 8992(2) 33(1)
Se(3) 1 100(2) 51 19(2) 6601(2) 25(1)
Se(4) 79(1) 7329(2) 5537(1) 24(1)
Se(5) 3933(2) 10218(2) 6735(2) 32(1)
Se(6) -878(2) 9625(2) 4333(2) 40(1)
86(7) 506(1) 2393(2) 5698(2) 26( 1)
Se(8) 3394(2) 9857(2) 9535(2) 34(1)
Se(9) 1386(2) 10007(2) 6939(2) 27(1)
Se(10) 2371(1) 7809(2) 7998(1) 24(1)
Se(l l) 3550(2) 5297(2) 6905(2) 31(1)
Se(12) 3257(1) 5577(2) 9314(2) 29(1)
P(l) 1614(3) 4653(5) 7961(4) 19(1)
P(2) 646(3) 7102(5) 6915(4) 26(1)
P(3) -344(3) 9376(4) 561 1(4) 22( 1)
P(4) 837(3) 10461(5) 5586(4) 22(1)
P(5) 1807(3) 8019(5) 6613(3) 24(1)
P(6) 2776(3) 5757(4) 8006(3) 20(1)

 

135

Table 5-3. Anisotropic displacement parameters (A2 X 103)) for Rb4P68e12. The
anisotropic displacement factor exponent takes the form: -27r2[ h2 a"‘2 U” +...+2 h k

3*b* U12].

 

 

U11 U22 U33 U23 U13 U12
Rb(l) 34(1) 33(1) 33(1) -2(1) -3(1) 5(1)
Rb(2) 36(1) 37(1) 37(1) 4(1) 5(1) 1(1)
Rb(3) 32(1) 36(1) 44(1) 2(1) -7(1) -4(1)
Rb(4) 36(2) 52(2) 53(2) 3(1) -7(1) -3(1)
Se(l) 34(2) 14(1) 50(2) 5(1) 40(1) -1(1)
Se(2) 30(2) 36(2) 34(2) -2(1) 8(1) -6(1)
Se(3) 30(2) 15(2) 30(2) -1(1) -5(2) 0(1)
Se(4) 34(2) 13(1) 26(1) -3(1) -7(1) 5(1)
Se(5) 30(2) 3 1(2) 34(2) 2( 1) 9(1) -4( 1)
Se(6) 46(2) 42(2) 31(2) -1 1(1) -15(1) 18(1)
Se(7) 33(1) 12(1) 34(1) 1(1) -1(1) 2(1)
Se(8) 33(2) 32(2) 37(2) 6(1) 45(1) -9(1)
Se(9) 35(2) 17(2) 28(2) -2(1) -9(2) 4(1)
Se(10) 29(1) 14(1) 28(1) 0(1) -5(1) 0(1)
Se(ll) 30(2) 31(1) 32(2) -1(1) 11(1) -3(1)
Se(12) 34(1) 23(1) 29(1) 2(1) -9(1) 1(1)
P(l) 19(3) 15(3) 21(3) 4(3) -3(3) -4(2)
P(2) 29(3) 17(3) 30(3) -3(3) 1(3) 6(2)
P(3) 24(3) 12(3) 31(3) 2(3) 0(3) 3(2)
P(4) 24(3) 13(3) 30(3) 0(3) 1(3) 4(3)
P(5) 31(3) 15(3) 24(3) 5(2) -1(3) 7(2)
P(6) 21(3) 12(3) 26(3) 0(2) -1(3) 5(2)

 

136

 

 

 

 

 

Figure 5-1. (a) [P68e12]4' anion. (b) Structure of Rb4P6Sel2. Open circles are Cs, octants P,
and black Se atoms. (0) Pseudo-one-dimensional chain of [P6S61214' via Se---Se
nonbonding contacts. The thermal ellipsoids are shown with 50% probability. P and Se

atoms are labeled in (a) and (c)

137

Table 5-4. Bond lengths (A) and angles (°) for Rb4P68e12.

 

 

2.300(6)

Se( 1)-P( 1) 2.128(5) Se(10)-P(5)
Se(2)-P( 1) 2. 130(6) Se(10)-P(6) 2.282(5)
Se(3)-P(1) 2.275(6) Se(l 1)-P(6) 2.148(6)
Se(3)-P(2) 2.288(5) Se(12)-P(6) 2.136(6)
Se(4)-P(3) 2.288(5) P(1)-P(6) 2.241(7)
Se(4)-P(2) 2.294(6) P(2)-P(5) 2. 189(6)
Se(6)-P(3) 2. 136(6) P(3)-P(4) 2.257(7)
Se(8)-P(4)#1 2.128(6) P(3)-Se(5)#5 2.1 16(6)
Se(9)-P(5) 2.279(5) P(4)-Se(7)#10 2.218(6)
Se(9)-P(4) 2.285(7) P(4)-Se(7)# 1 3 2. 133(5)
P(5)-Se(9)-P(4) 97.0(2) Se(8)#10-P(4)-Se(7)#13 102. 1(3)
P(6)-Se(10)-P(5) 102.4(2) Se(8)#10-P(4)-P(3) 1 1 1.6(3)
Se(1)-P(1)-Se(2) l 18.4(2) Se(7)#13-P(4)-P(3) 105.8(2)
Se(1)-P(1)-P(6) 104.7(2) Se(8)#10-P(4)-Se(9) 1 1 1 .7(2)
Se(2)-P(1)-P(6) 1 12.8(3) Se(7)#13-P(4)-Se(9) 103.5(2)
Se(1)-P(1)-Se(3) 104.5(2) P(3)-P(4)-Se(9) 102.5(3)
Se(2)-P(1)-Se(3) 1 1 1.7(2) P(2)-P(5)—Se(9) 96.0(2)
P(6)-P(1)-Se(3) 103.2(2) P(2)-P(5)-Se( 1 0) 96.7(2)
P(5)-P.(2)-Se(3) 94.8(2) Se(9)-P(5)-Se(10) 90.9(2)
P(5)-P(2)-Se(4) 96.7(2) Se(12)-P(6)-Se(11) 118.5(2)
Se(3)-P(2)-Se(4) 92.3(2) Se(12)-P(6)-P(l) 107.2(3)
Se(5)#5-P(3)—Se(6) 1 1 8.0(2) Se(l 1)-P(6)-P(1) l 1 l . 1(3)
Se(5)#5-P(3)-P(4) 1 13.0(3) Se( 1 2)-P(6)-Se(10) 101 .4(2)
Se(6)-P(3)-P(4) 105.9(3) Se(l 1)-P(6)-Se(10) l 12.7(2)
P(4)-P(3)—Se(4) 103.0(2) P(1)-P(6)-Se(10) 104.7(2)

Symmetry transformations used to generate equivalent atoms:

#1 -x+1/2, y, z+1/2 #2 -x, -y+2, z+1/2 #3 -x, -y+l, z+l/2

#4 x-l/2, -yl-l , z #5 x-1/2, -y+2, z #6 -x-1/2,y, z+1/2

#7 x+1/2, -y+l, z #8 -x, -y+1, z-1/2 #9 -x, -y+2, z-1/2

#10-x+l/2,y,z-1/2 #11x+1/2,-y+2,z #12x,y-1,z #13 x,y+l,z

138

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6880 e n _ H _ Coma 8: 8 .8: a: 1 $ .85..
Eda—o anmmfloamfl 0:0

.653... 6880 m H _ H _ reemé 8: as: a: u 3 833
we: 68388-5. 36:; .1 _ L 48m: 88:86

38.5.:
8:80 m H 2 : 4 A28 .8: 1 $ 288.3.

ma: 8:68:68 -me 63:09 endowmeaé
8568885 B: m H mm: H : 326mm: pep: no: u .5 some}.
6880 «an _ H _ seem“: 2:88:28
8386: B: m H _ n N - 6m”: A2: .52 u 3 6mm}.
8: m e _ n N 48%: so .8: a: 1 $ 838
28: B: m n _ n N 426?: 228:8:
B: e H : n m 4.6.6.: :28 .0: 82 .1. 3 amen...

as:
862 8:680 :32 8.00 :25".~ mugafiou :Bo:v_
8 U m u <

 

6:8 amine. on: :5 866% Baggage—om wotomno :ooszon “€825.23” .m1m 635—.

139

350 T T l

 

 

 

 

 

218
3‘ 300 e —
'7)
c
3
E 250 ~ -
E
E 371
(U 166
200 _ 487 414 293 m
150

 

500 400 300 200 100
Raman Shift, cm'1

Figure 5-2. Raman Spectrum of Rb4P6Se12 at room temperature.

 

 

 

40 3'0 6pm

Figure 5-3. 3 lP Bloch decay NMR spectrum of Rb4P68e12 taken at ambient temperature
on a 9.4 T NMR spectrometer. Acquisition parameters included 5 ps n/2 pulse, 8000 s
relaxation delay, and 13 kHz magic-angle-spinning frequency. The chemical shifi
reference was 85% H3PO4 (0 ppm). Peaks with the same letter are J-coupled. The average
chemical shifts of peaks A-E are in order: 23.9, 47.5, 52.6, 57.6, 58.8 ppm. The J-
couplings of doublets B-E are: 309, 292, 295, 309 Hz.

140

4. Results and Discussion

Crystal Structure. The new structure-type of Rb4P68e|2 adopts the noncentrosymmetric
space group Pca21. The compound features discrete [P6Se12]4' molecules, Figure 5-la.
The centrosymmetric molecule generates itself through a 21-screw axis along the c axis
with no mirror plane perpendicular to the c axis and, consequently, crystallizes in a
noncentrosymmetric fashion, Figure 5-1b. The most unusual feature of the structure is its
bicyclic nature and the presence of three P-P bonds with two types of formal charges of
P2+ and P“. The divalent formal charge is found on P2 and P5. The structure of the
[P6Se12]4' molecule is reminiscent of the bicycloalkane trans-decalin. The central P2 unit
is condensed with two ethane-like [P28e6]4' fragments to form the decalin-like skeleton.

The [P68612]4' molecule adopts the C2,, point group, so that the central [P28e4]
subunit has a trans-CZh-type configuration around the P2+ centers. The same trans-C2},-
type configuration is found in P2X4 (X = F, Cl, 1).13 The Se-P-Se angles around P2 and P5
are nearly 90° [Se3-P2-Se4, 92.3(2)°; Se9-P5-Se10, 90.9(2)°]. The dihedral angles of Se3-
P2-P5-Se9 and Se4-P2-P5-Se10 are 178.76(20) and -179.74(20)°, respectively. As a result,
the P2-P5 vector defines the intersection of two planes. The outer [P28e6] residues in the
molecule represent typical anti-type conformations but are distorted. (See relevant angles
in Table 5-4.) The P-Se distances are normal at 2.116(6)-2.300(6) A. The Se-P-Se angles
range from 90.9(2) to 120.1(3)°. The P2-P5 distance is 2.189(6) A, which is only slightly
shorter than the typical P-P distance of ~2.2 A in selenophosphates. Rhombohedral black
P shows a P-P distance of 2. l 3 A.”

To the best of our knowledge, this is a unique anion. The only other anion with

141

formally P2+ centers is the polymeric l/OO[PSSeAmS'], which is strongly bound to transition-
metal atoms in A3MPSSe10 (A = K, Rb; M = Ru, Os).'5’l6 In these compounds, the P”
centers are coordinated to Ru2+ or 052+ metal centers through the P atoms, forming P-M
bonds (M = Ru, Os).

Interestingly, the atoms Se1 - - ~Se3- - -Se4- - -Se6 are collinear, which allows
maximum overlap of their pn orbitals. The Se7- - -Se9- - -Se10- - -Se12 atoms are similarly
collinear. The intramolecular Se3---Se4, Se4~-Se6, Se9~-Se10, and Se10---Se12
distances range from 3.273(3) to 3.416(3) A and indicate both nonbonding interactions
and severely distorted Se-P-Se angles in the central [P28e4] subunit. Similar
intramolecular interactions have been observed in the [PSe3] pyramidal fragment of the
[ngSelgfs'7 anion and the l/C,0[PSe(,'] chain in CsPSeé.9

Another notable feature is the unusually short intermolecular Se1---Se9 and
Se3-~Se7 distances of 3.384(3) and 3.340(3) A, respectively. These distances indicate
nonbonding interactions, but they are much shorter than the 3.80 A sum of the van der
Waals radii.20 They enable the molecules of [P68e12]4' to organize to an infinite pseudo-
one-dimensional structure, Figure 5-lc. Low-dimensional compounds such as NbSe321
and APSe6 (A = K, Rb, Cs) display similar Se~~Se interactions. These intermolecular
interactions may contribute to the compound's stability in air and polar solvents such as
water and N,N-dimethylforrnamide.The synthesis of Rb4PGSe12 adds further insight in the
close relationship between the structure and the flux condition (or A:P:Se ratio) in the
alkali selenophosphate ternary system. More basic fluxes (i.e., those with a high A:P
ratio) or higher reaction temperatures tend to give shorter structural fragments2| and P in

the 5+ oxidation state (Table 5-5). All simple anions, e.g., [PSe4]3' and [P28e9]4', were

142

prepared in strongly basic fluxes (Table 5-5). As the basicity decreases, more complex
species emerge, such as [P28e6]4', [PgSe13]4’, and [P28e3]2'. Under even less basic flux
conditions, the one-dimensional polyselenide chain l/OQ[PSe6'] is stabilized with K, Rb,
and Cs. Rb4P6Se12 is made in intermediate acidic/basic conditions, and it contains P2+
centers because of excess P.

Differential thermal analysis (DTA) on Rb4P6Se12 at a rate of 10 °C min'l showed
melting at ~431 °C and crystallization upon cooling at ~384 °C. X-ray diffraction patterns
for samples before and after DTA were identical. The results suggest that Rb4P6Se12 could
be a promising precursor in synthesis for further reaction chemistry. The UV—vis spectrum
reveals a sharp absorption edge and a band gap of 2.25 eV, which is in good agreement
with its orange color. By comparison, the one-dimensional RbPSe6 with P5+ showed a gap
of 2. 1 8 eV. The Raman spectrum of Rb4P68e12 shows shifts at 218 (vs), 298 (vw), 348(w),
371 (w), 414 (w), 487 (w), and 516 (vw) cm", Figure 5-2. The peak at 218 cm" is
unambiguously assigned to the locally Alg symmetric stretching mode of PSe3.23 The
second shift at 298 cm”1 resembles the V.;; (Bg) mode in szPzSC6 having C2}, site
symmetry.24 The other vibrations can be attributed to PSe3 stretching modes.”4 The far-
IR spectrum is rather complex, showing peaks at 214 (vs), 238 (bw), 301 (vw), 319 (vw),
353 (vw), 374 (vw), 392 (vw), 408 (w), 464 (vw), 486 (vw), 514 (w), and 524 (w) cm".
The peak at 319 cm'1 is assigned to a P-P vibration and was not observed in the Raman
spectrum. Other peaks are well matched with those of szPZSe6 and other compounds that
contain the [P2Se6]4‘ anion.26

Figure 5-3 displays the 3 IP NMR spectrum of Rb4P6$e12. The narrow line widths

afford resolution of most of the P sites. The ratio of the integrated intensity of the B-E

143

cluster of peaks to the intensity of peak A is ~2 and is generally consistent with the
assignment of P1, P3, P4, and P6 to the B-E peaks and P2 and P5 to the A peak. Pl/P6
and P3/P4 J couplings are expected, and analysis of the spectrum yielded the same J
coupling for doublets B and E and the same J coupling for doublets C and D. The
isotropic chemical shifis and J couplings were confirmed by the analysis of spectra taken
on a 7-T spectrometer. It was not possible to make a more detailed assignment of Pl/P6
and P3/P4 to the C/D and B/E doublets, and the P2/P5 shifts were not resolved from one
another. The 3'P NMR chemical shifis are all >0 ppm and are consistent with the positive

chemical shifts of other metal selenophosphates with P-P bonding.27
5. Concluding Remarks

The new selenophosphate compound Rb4P6Se12 was synthesized by employing
new synthetic conditions, e.g., reductive reactions using P5+ [PnyF‘ species as starting
materials with an excess of elemental P. This chemistry showed the high reactivity of
alkali chalcophosphate ternary compounds that melt at low temperature (300-500 °C) as
useful starting materials, to give new compounds. The discovery of the molecular salt
Rb4P6Se12 with its unique bicycloselenophosphate anion and rare combination of P2+ and
P“ centers suggests a more extensive compositional diversity in alkali chalcophosphates.
The unraveling relationship between basicity and the final structure enhances the
understanding of flux chemistry and the prospects of the future discovery of new

materials in this class of solids.

144

Reference

(1) Kanatzidis, M. G. Curr. Opin. Solid State Mater. Sci. 1997, 2, 139.
(2) Garin, J .; Parthe, E. Acta Crystallogr. 1972, B28, 3672.

(3) Dickerson, C. A.; Fisher, M. J.; Sykora, R. B.; Albrecht-Schmitt, T. 13.; Cody, J.
A. Inorg. Chem. 2002, 41, 640.

(4) (a) Toffoli, P. P.; Khodadad, P.; Rodier, N. Acta Crystallogr. 1973, B34, 1779.
(b) Francisco, R. H. P.; Tepe, T.; Eckert, H. J. Solid State Chem. 1993, 107, 452.

(5) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1995, 34, 5401.

(6) Gave, M. A.; Canlas, C. G.; Chung, 1.; lyer, R. G.; Kanatzidis, M. G.; Weliky, D.
P. J. Solid State Chem. 2007, 180, 2877.

(7) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 3 7, 2582.

(8) (a) Chung, 1.; Kanatzidis, M. G. Manuscript in preparation. (b) Zhao, J.;
Pennington, W. T.; Kolis, J. W. J. Chem. Soc., Chem. Commun. 1992, 265.

(9) Chung, 1.; Canlas, C. G.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem. 2004, 43,
2762.

(10) (a) Do, J.; Yun, I-l. Inorg. Chem. 1996, 35, 3729. (b) Chen, J. H.; Dorhout, P.
K.; Ostenson, J. E. Inorg. Chem. 1996, 35, 5627. (c) Gauthier, G.; Jobic, S.; Brec, R.;
Rouxel, J. Inorg. Chem. 1998, 37, 2332. (d) Gieck, C.; Tremel, W. Chem. -Eur. J. 2002, 8,
2980.

(ll) SHELXTL: Data collection and Processing Software for the SMART-CCD
system; Siemens Analytical X-ray Instrument Inc.: Madison, WI, 1997.

(12) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7.

(l3) Corbridge, D. E. C. Phosphorus, An Outline of its Chemistry, Biochemistry and
Uses, 5th ed.; Elsevier: New York, 1995; p 143.

(14) Corbridge, D. E. C. The Structural Chemistry of Phosphorus; Elsevier: New
York, 1974; p 16.

(15) Chondroudis, K.; Kanatzidis, M. G. Angew. Chem, Int. Ed. Engl. 1997, 36,
1324.

(16) Chung, 1.; Kanatzidis, M. G. Unpublished results.

145

(l7) Gave, M. A.; Canlas, C. G.; Chung, 1.; lyer, R. G.; Kanatzidis, M. G. Weliky, D.
P. J. Solid State Chem. 2007, 180, 2877.

(18) Chung, 1.; Jang, J. 1.; Gave, M. A.; Weliky, D. P.; Kanatzidis, M. G. Chem.
Commun. 2007,4998.

(19) Chung, 1.; Malliakas, C. D.; Jang, J. 1.; Canlas, C. G.; Weliky, D. P.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2007, 129, 14996.

(20) Shannon, R. D. Acta Crystallogr. 1964, A32, 751.

(21) (a) Electronic Properties of Inorganic Quasi-One-Dimensional Compounds;
Monceau, P., Ed.; D. Reidel Publishing Co.: Dordrecht, The Netherlands, 1985; Part 11.
(b) Canadell, B.; Rachidi, E.-l.; Pouget, J. P.; Gressier, P.; Meerschaut, A.; Rouxel, J.;
Jung, D.; Evain, M.; Whangbo, M.-H. Inorg. Chem. 1990, 29, 1401.

(22) Kanatzidis, M. G.; Sutorik, A. Prog. Inorg. Chem. 1995, 43, 151.

(23) Mathey, Y.; Clement, R.; Sourisseau, C.; Lucazeau, G. Inorg. Chem. 1980, 19,
2773.

(24) Becker, R.; Brockner, W. Z. Naturforsch. 1984, 39a, 357.

(25) Chondroudis, K.; Charkrabarty, D.; Axtell, E. A.; Kanatzidis, M. G. Z. Anorg.
Allg. Chem. 1998, 624, 975.

(26) Parensen, M.; Brockner, W.; Cyvin, B. N.; Cyvin, S. J. Z. Naturforsch. 1986, 41a,
1233.

(27) Canlas, C. G; Kanatzidis, M. G; Weliky, D. P. Inorg. Chem. 2003, 42, 3399.

146

Chapter6
Low Valent Phosphorus in the Molecular Anions [PSSetz]5’ and #-
[P6Senl4’: Phase Change Behavior and Near Infrared Second Harmonic

Generation

1. Introduction

The structural diversity found in the chalcophosphate family is extensive.
Various selenophosphate anions [PnyF' (Q = S, Se) have been isolated and structurally
characterized, for example, [PSe4]3‘,' [star'F [P28e9]4’,3 [PgSe13]6',4 a—[P68e12]4',5
1/00[PSe(,']6 and 1/00[P28e62’].7 The most common oxidation state of P in selenophosphates

is P4+ and P“. Member of this family can exhibit technologically important properties of
ferroelectric,8 nonlinear optical,7’9 reversible redox chemistry relevant to secondary
batteries,IO photoluminescence,ll and phase-change properties.6’7’l2 Our experimental
investigations of alkali chalcophosphate compounds provided new insights on the
relationship between the structure and the flux composition (A:P:Se ratio in the
composition).5 We also found that excess phosphorus in the flux helps to produce less
oxidized P2+/3+/4+ species such as Rb4P6Se125 and A6P38613(A=K, Rb, Cs).‘ With this in
mind, we focused on devising rational synthetic conditions to obtain P-rich species rather
than the simple classical [PSea]3' or [P28e6]4' anions. Here we describe the novel
molecular [P58e12]5' and fl—[PbSetzr' anions. The former includes P3+ and P4+ centers and

4-5

features octahedrally coordinating P. The latter is a structural isomer of a—[Pbselz] " and

147

it contains P2+ and P4+ centers. Cs5P5Se12 and CsaP6Se12 exhibit reversible phase-change

behavior. Both crystalline and glassy Cs5P5$e12 exhibit second harmonic generation non-

linear optical response.

2. Experimental Section

2.1. Reagents.

The reagents mentioned in this work were used as obtained: Cs metal (analytical
reagent, Johnson Matthey/AESAR Group, Seabrook, NH); red phosphorus powder, -100
mesh, Morton Thiokol, Inc., Danvers, MA; Se (99.9999%; Noranda Advanced Materials,
Quebec, Canada); N,N-dimethylformamide (Spectrum Chemicals, ACS reagent grade);
diethyl ether (Columbus Chemical Industries, Columbus WI, ACS reagent grade,
anhydrous). AZSe (A = K and Rb) starting materials were prepared by reacting
stoichiometric amounts of the elements in liquid ammonia. P2Se5 was prepared by
heating the mixture of P and Se with a stoichiometric ratio sealing in an evacuated silica
tube at 460 °C for 24h.

2.2. Synthesis

Pure orange block-shaped crystals of CssPsselz were obtained by heating a
mixture of CszSezP:Se=1:2.524 under vacuum in a silica tube at 400 °C for 3 (1, followed
by washing the product with degassed N,N-dimethylformamide (DMF) under a N2
atmosphere to remove residual flux. The compound could be also obtained by direct
combination reaction of starting materials at 400 °C for 3d. Energy-dispersive

spectroscopic (EDS) microprobe analysis showed an average composition of

148

“Cs4,gPSSen_6”. The single crystals are stable in DMF and alcohol and in air for several
days. Pure orange block—shaped crystals Cs4P6$e12 were obtained by heating a mixture of
CSZSe:P:Se=1:4:5 under the same conditions described above. EDS microprobe analysis
showed an average composition of “CS3,3P(,Sen,6”. The glassy phase of C84P63612 was
prepared from melting single crystals of CsaPGSetz under vacuum in a quartz tube at 800-
900 °C for 1-2 min and quenching in ice water. The crystals are stable in DMF, alcohol

and water and in air for several days.

3. Physical Measurements

X-ray Powder Diffraction. Phase purity x-ray diffraction analyses were performed using

a calibrated CPS 120 INEL X-ray powder diffractometer (Cu Ka graphite

monochromatized radiation) operating at 40 kV/20 mA and equipped with a position-
sensitive detector with flat sample geometry.

Electron Microscopy. Semiquantitative analyses of the compounds were performed with
a J EOL J SM-35C scanning electron microscope (SEM) equipped with a Tracor Northern
energy dispersive spectroscopy (EDS) detector.

Solid-State UV-vis spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region. The instrument
is equipped with an integrating sphere and controlled by a personal computer. 321804 was
used as a 100 % reflectance standard. The sample was prepared by grinding the crystals

to a powder and spreading it on a compacted surface of the powdered standard material,

149

preloaded into a sample holder. The reflectance versus wavelength data generated were
used to estimate the band gap of the material by converting reflectance to absorption
data.l3

Raman Spectroscopy. Raman spectra were recorded on a Holoprobe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation fi'om a HeNe
laser for excitation and a resolution of 4 cm". Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 pm. A total of 128 scans
was sufiicient to obtain good quality spectra.

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CsI or KBr matrix.
The samples were ground with dry CsI or KBr into a fine powder and pressed into
translucent pellets. The spectra were recorded in the far-IR region (600-100 cm], 4 cm'1
resolution) and mid-IR region (500-4000 cm", 4 cm"1 resolution) with the use of a
Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon beam
splitter.

Differential Thermal Analysis (DTA). Experiments were performed on Shimadzu DTA-
50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. The sample was heated to 600°C
at 10 °C min", and after 1 min it was cooled at a rate of —10 °C min'l to 50 °C. The
residues of the DTA experiments were examined by X-ray powder diffraction.
Reproducibility of the results was confirmed by running multiple heating /cooling cycles.
The melting and crystallization points were measured at a minimum of endothermic peak

and a maximum of exothermic peak.

150

X-ray Photoelectron Spectroscopy. Single crystals of CssPsselz and CsaP68e12 were
ground to powders and pressed to pellets. X-ray photoelectron spectroscopy (XPS)
analyses were performed on the pellets with Omicron ESCA probe equipped with EA

125 energy analyzer. Photoemission was stimulated by a monochromated Al Ka

radiation (1486.6 eV). Binding energies were normalized to the C Is binding energy set
at 285.0 eV. To analyze the XPS results, linear background correction was applied and
photolines were fit by Gaussian-Lorentzian curves. The fitted curve was deconvoluted to
reveal multiple valence states of phosphorus and their 2p1 ,2 and 2pm spectra. The
quantitative analysis of P in Cs5P53e12 and Cs4P6$e12 were carried out by XPS using the
integrated areas of the core-level peaks of P 2p.

Solid-State Nuclear Magnetic Resonance (N MR) Spectroscopy. Room temperature 3 IP
NMR spectra were collected on a 9.4 T spectrometer (Varian Infinity Plus) using a Varian
Chemagnetics double resonance magic angle spinning (MAS) probe. The sample volume
in the 4 mm diameter zirconia rotors was ~50 ,uL and the MAS frequency was 14 KHz.
Each sample was ground to a fine powder before being packed in the rotor. The 7t/2 pulse
width was 5.55 as and was calibrated using H3POa. Spectra were processed using 50 Hz
line broadening and up to a 10th order polynomial baseline correction. Longitudinal
relaxation times (T 1) were determined as described elsewhere.14

X-ray Crystallography. The crystal structure was determined by single-crystal X-ray
diffraction methods. For Cs5P5$e1 2, preliminary examination and data collection were
performed on a SMART” platform diffractometer equipped with a 1K CCD area detector

using graphite monochromatized Mo Ka radiation at 292(2) K. A hemisphere of data was

collected at 292(2) K using a narrow-frame method with scan widths of 030° in a) and an

151

exposure time of 30s frame". The data were integrated using the SAINT ‘5 program. An
empirical absorption correction using the program SADABS were performed. The initial
positions for all atoms were obtained using direct methods and the structures were refined
with the full-matrix least-squares techniques of the SHELXTL]5 crystallographic
software package. Satisfactory refinement was obtained with the noncentrosymmetric
space group, P-4. Flack parameter was refined and absolute structure parameter is 006(2).
Intensity data for CS4P58612 were collected at 100(2) K on a STOE IPDS II diffractometer
with Mo K.I radiation operating at 50 kV and 40 mA with a 34 cm image plate. Individual
frames were collected with a 5 min exposure time and a 1.0 a) rotation. The X-AREA, X-
RED and X-SHAPE software package was used for data extraction and integration and to
apply empirical and analytical absorption corrections. The SHELXTL software package
was used to solve and refine the structure. The most satisfactory refinement was obtained
with the space group, P21/n. The parameters for data collection and the details of the
structural refinement for both compounds are given in Table 6-1. Fractional atomic
coordinates and displacement parameters for each structure are given in Tables 6-2 — 6-4.
In all cases the atoms were refined to full occupancy.

Nonlinear Optical Property Measurements. We used the frequency-tripled output of a
passive-active mode-locked Nd:YAG laser with a pulse width of about 15 ps and a
repetition rate of 10 Hz to pump an optical parametric amplifier (OPA). The OPA
generates vertically polarized pulses in the ranges 400 ~ 685 nm and 737 ~ 3156 nm. In
order to check the SHG efficiency as a fimction of the excitation energy, we tuned the
wavelength of the incident light from 1000 ~ 2000 nm. In this range, the spectral

bandwidth of the linearly polarized light from the OPA is rather broad, about 2 meV full

152

width at half maximum. However, the phase space compression phenomena ensures
effective SHG where lower energy portions are exactly compensated by higher parts
thereby satisfying both energy and momentum conservation. The incident laser pulse of
300 p] was focused onto a spot 500 pm in diameter using a 3 cm focal-length lens. The
corresponding incident photon flux is about 10 GW cm'z. The SHG signal was collected
in a reflection geometry from the excitation surface and focused onto a fiber optic bundle.
The output of the fiber optic bundle is coupled to the entrance slit of a Spex Spec-One
500 M spectrometer and detected using a nitrogen-cooled CCD camera. The data

collection time is 20 s.

153

Table 6-1. Crystallographic Data and Refinement Details for Cs5P5Se12 and CsaP6Se12.

 

Formula

F ormula Weight
Space group

as
b, A

c, A

,6, deg.

V, A3

Z

crystal size, mm3

Pcalcd: g cm-B

,u, cm'l

T, K

A, A

Prange, deg

Total reflections

Total unique reflections
No. parameters
Refinement method

Final R indices [I>2 0(1)],

Rla/wRZ b (%)

R indices (all data), R1/wR2

Goodness-of-fit on F2

Absolute structure parameter

CssP5Se12
1766.92
P-4 (no. 29)
13.968(1)
13.968(1)
7.546(1)
90.000
1472.2(3)
2

N/A
3.986

212.29
292(2)
0.71073
1.46 — 28.25
12701

3424

73

CS4P58612.
1664.98
P21/n (no. 13)
10.836(1)
10.5437(8)
12.273(1)
98.661(8)
1386.3(2)

2

0.065 x 0.055 x 0.031
3.989

213.09
100(2)
0.71073
2.33 — 29.99
8320

3961

101

Full-matrix least-squares on F2

3.00/7.27
3.70/8.55
1.159

006(2)

429/ 10.36
7.52/ 14.03
1.077

N/A

 

0 R1 = 2 “1701- IFen/ZIFOI. b wR2 = {z [W(F02'Fcz)21/ZIW(F02)2]} ”2

154

Table 6-2. Atomic coordinates (X 104) and equivalent isotropic displacement parameters
(A2 X 103) for CssP5Se12 and CsaP6Se12. U(eq) is defined as one third of the trace of the

orthogonalized Uij tensor.

 

 

x y 2 U(PCI)
CS5P58€12
Cs(1) 3124(1) 6547(1) 2712(1) 46(1)
Cs(2) 3240(1) 355(1) 2544(1) 33(1)
Cs(3) 0 0 2477(1) 39(1)
Se(l) -355(1) 3129(1) 2521(1) 30(1)
Se(2) 1492(1) 1856(1) 90(1) 30(1)
Se(3) 1473(1) 1780(1) 4911(1) 28(1)
Se(4) 1651(1) 4644(1) 5133(1) 32(1)
Se(5) 1275(1) 4778(1) 474(1) 26(1)
Se(6) 3466(1) 3698(1) 2132(1) 44(1)
13(1) 1125(1) 2564(1) 2537(3) 20(1)
P(2) 1994(1) 3924(1) 2678(3) 22(1)
P(3) 0 5000 2460(4) 26(1)
(3841363612
Cs(1) 4905(1) -9(1) 2464(1) 13(1)
Cs(2) 5239(1) 4995(1) 1774(1) 15(1)
Se(5) 2385(1) 2359(1) 2759(1) 16(1)
Se(2) 2552(1) 2540(1) 7250(1) 14(1)
Se(3) -16(1) 2876(1) 4993(1) 12(1)
Se(4) 2871(1) -117(1) 4903(1) 14(1)
Se(6) 4914(1) 2553(1) 5048(1) 14(1)
Se(l) 2594(1) 4999(1) 5144(1) 14(1)
P(l) 1948(3) 2999(3) 5557(2) 11(1)
P(2) 2954(3) 1819(3) 4429(2) 12(1)
P(3) 4334(3) 4502(3) 4359(2) 14(1)

 

155

Table 6-3. Anisotropic displacement parameters (A2 X 103) for CssPssetz and Cs4P6Se12.
U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

 

 

U11 U22 U33 U23 U13 U12
CssPSSelz
Cs(1) 59(1) 41(1) 39(1) -8(1) 2(1) -13(1)
Cs(2) 30(1) 41(1) 28(1) 0(1) 4(1) 4(1)
Cs(3) 41(1) 48(1) 29(1) 0 0 -11(1)
Se(l) 22(1) 37(1) 30(1) 2(1) 0(1) 2(1)
Se(2) 42(1) 27(1) 20(1) -6(1) 2(1) 3(1)
Se(3) 40(1) 25(1) 20(1) 4(1) -2(1) 4(1)
Se(4) 41(1) 29(1) 26(1) -7(1) 1(1) 2(1)
Se(5) 29(1) 26(1) 22(1) 5(1) 2(1) 2(1)
Se(6) 23(1) 64(1) 46(1) -9(1) 2(1) 4(1)
P(l) 24(1) 18(1) 16(1) -1(1) 1(1) 2(1)
P(2) 23(1) 21(1) 22(1) -1(1) 0(1) 2(1)
P(3) 24(1) 36(1) 19(1) 0 0 5(1)
Cs4P6Se12
Cs(1) 30(1) 32(1) 32(1) -3(1) 5(1) 0(1)
Cs(2) 38(1) 37(1) 30(1) 1(1) 2(1) -7(1)
Se(l) 16(1) 28(1) 43(2) 4(1) 4(1) 3(1)
Se(2) 41(2) 37(2) 25(2) 4(1) 6(1) 12(1)
Se(3) 33(2) 17(1) 38(2) 1(1) 1(1) 2(1)
Se(4) 16(1) 31(1) 33(1) -1(1) 2(1) -4(1)
Se(5) 33(2) 34(2) 22(1) -3(1) -3(1) 1(1)
Se(6) 24(1) 18(1) 49(2) -3(1) 12(1) 4(1)
P( 1) 19(3) 20(3) 25(3) 0(3) 4(3) 3(3)
P(2) 15(3) 18(3) 25(3) 0(3) 3(2) -1(2)
P(3) 20(3) 22(3) 32(4) -6(3) 7(3) -2(3)

 

156

                        

 

ice/Eu“. .«u, .~ 9. .

.. .....auu..mur.. its ,. .
new. «we...

..v. . over:

   

Alana... :, in»...

..Jaaykma...‘ ,
1.1.1 ”Duh”... .
. ....u.....%.x2.........s

5 And
.....a...
.54! mag».
Viva“. Wei.“ 5
\lwmhc

......

< 4.90.
nhbuvi. l _
_

  
         

¢ii
v;

. a
.025
5&1 .1.

. ...... r ././3..

1am... w:
a...
r)
9‘.- rlkn: 5 A
4 a.
D5,.
6‘.
w l
... .
n. C
o
b «o W

 

Figure 6-1. (a) The noncentrosymmetric structure of CssPSSen. The thermal ellipsoids
bonding at P(3)-Se(1), 2.6606(8) A. Dashed lines indicate short Se---Se nonbonding
157

are shown with 60% probability. (b) [P58e12]5' anion. White lines denote long P-Se
interaction at P(3)---Se(4), 3.104(2) A. * is defined as equivalent position (~x, l-y, z).

 

(b)

 

(d)

   

Figure 6-2. Structure of CsaP6Se12 viewed down the (a) c-axis and (b) b-axis. The
thermal ellipsoids are shown with 60% probability. (c) The ,B-[PoSetzr' anion and (d) a-

[P6Se12]4' anion for comparison.

158

Table 6-4. Selected bond distances (A) and angles (deg) for CssPssetz.

 

P(l)-Se(1)
P(1)-Se(2)
P(1)-Se(3)
P(2)-Se(4)
P(2)-Se(5)

Se(1)-P(1)-P(2)
Se(2)-P(1)-P(2)
Se(3)-P(l)-P(2)
Se(3)-P(1)-P(2)
Se(3)-P(1)-P(1)
Se(2)-P(1)-P(1)
Se(6)-P(2)-P(1)
Se(6)-P(2)-P(5)
Se(4)-P(2)—P(5)

2.2127(18)
2.156(2)
2.156(2)
2.162(2)
2.280(2)

101.72(8)
107.36(9)
105.49(9)
115.14(8)
113.35(9)
112.41(9)
116.78(9)
111.29(9)
106.52(8)

P(2)-Se(6)
P(3)-Se(1)
P(3)-Se(5)
P(1)-P(2)

Se(5*)-P(3)-Se(5)
Se(5)-P(3)-Se(l *)
Se(5*)-P(3)-Se(1)
Se(5)-P(3)-Se(1)
Se(l *)-P(3)-Se(1)
P(2)-P(5)-Se(3)
P(1)-P(2)-Se(4)
P(1)-P(2)-Se(5)
P( l )-P(2)-Se(6)

2.1207(19)
2.6606(8) x2
2.349(2) x2
2.257(2)

100.67(12)
8997(5)
8997(5)
9129(5)
178.0203)
8642(7)
108.22(9)
9970(9)

1 1280(9)

 

* denotes the crystallographically equivalent position by symmetry transformation of (-x,

-y+l, z).

159

Table 6-5. Selected bond distances (A) and angles (deg) for CS4Pfiselz.

 

P(1)-Se(1)
P(1)-Se(2)
P(1)-Se(3)
P(2)-Se(4)
P(2)-Se(5)

Se(l)-P(1)-Se(2)
Se(1)-P(1)-Se(3)
Se(2)-P(1)-Se(3)
Se(4)-P(2)-Se(5)
Se(4)-P(2)-Se(6)
Se(5)-P(2)-Se(6)
Se(1)-P(3)-Se(6)
P(1)-P(2)-Se(4)

P(1)-P(2)-Se(5)

2.303(2)
2.137(2)
2.140(2)
2.128(2)
2.127(2)

1 1126(9)
107.740)
1 1642(8)
120.15(8)
108.07(8)
1 1052(8)
104.36(8)
108.43(8)
110.26(8)

P(2)-Se(6)
P(3)-Se( 1)
P(3)-Se(6)
P(1)-P(2)

P(I)-P(1*)

P(1)-P(2)-Se(6)
P(2)—P(l)—Se(1)
P(2)-P(1)-Se(2)
P(2)-P(1)-Se(3)
P(3*)-P(3)-Se(1)
P(3)-P(3)-Se(6)
P(1)-Se(1)-P(3)
P(2)-Se(6)—P(3)

2.281(2)
2.303(2)
2.274(2)
2.260(2)
2.232(2)

9690(8)
100.29(8)
111.43(8)
108.43(8)
9464(8)
9284(8)
100.29(8)
8950(8)

 

1 denotes the crystallographically equivalent position by symmetry transformation (-x+1,

-y+l, -z+1).

160

4. Results and Discussion

Crystal Structure. The new compound Cs5P5Setz crystallizes in the
noncentrosymmetric nonpolar space group P-4, Figure 6-1a. It features the discrete
molecular [PSSe12]5' anion with two types of formal charge 3+ and 4+ on P, Figure 6-1b.
The trivalent formal charge is found on P(3) which is a central P atom chelated with two
ethane-like [PZSe6]4' units to form a novel octahedral complex. The octahedral
coordination of the trivalent P(3) atom is very unusual and features long bond distances
e. g. P(3)-Se(l) at 2.6606(8) A and an even longer interaction of P(3)-"Se(4) at 3.104(2) A.
The latter distance is much shorter than the 3.76 A of the van der Waals radii sum.'° The
P(3)-Se(5) distance is normal at 2.349(2) A. If the anion is to be viewed as a coordination
complex, the P(3) atom plays the role of the metal and the formula can be expressed as
{P[PZSe6]2}5'. P-Se bond distances in the chelating [PZSe6]4' ligands range from 2.121(2)
to 2.280(2) A. The [PZSeo]4' ligands in the molecule represent typical anti-type
conformation but are distorted (See relevant angles in Table 6-4.).

It is noted that the bond angles for P(3) are close to 90° or 180°, Figure 5-1.
Deviation from 90° [Se(4)-P(3)-Se(4), 9894(2); Se(4)-P(3)-Se(5), 8022(3); Se(5)-P(3)-
Se(5), 100.67(3)] is caused by the 112-P(2). As a result, the octahedron around P(3) has
excellent definition. The dihedral angle of Se(5)-Se(4)-Se(4)-Se(5) is only 2.35(2)° and
indicates that the equatorial Se(2), Se(5), Se(5) and Se(2) atoms are nearly coplanar.

It is noteworthy that Cs5[PSSetz] is the chalcophosphate counterpart of
Css[In(PZSe(,)2]l7 where P(3) occupies the In3+ site. Generally, phosphorus is not regarded
as a coordinating metal but the isolation of {P[P2Se6]2}5' suggests that P itself can be such

a center for [Pny]"-, [Gery]"- or [Snny]n- ligands to produce novel complexes.

161

Cs4P68et2 crystallizes in the monoclinic space group P21/n, Figure 6-2a, b. The
compound features the new molecule ,B-[P6Se12]°', Figure 6-2c. The central P2 dimer is
coordinated with two [P2Se6] units to form a firlvalene-like skeleton. Both isomers a—
and fl-[P6Set2]4‘ anions consist of the central [P2] and two [P2Se6] residues but they bind
in a different fashion, Figure 5-2. The fl-species has a center of symmetry at the middle of

the P(3)-P(3*) bond and adopts the C,- point group. The formal charge on the P(3) atoms

in the central [P2] is 2+. The P(3)-P(3*) distance at 2.232(2) A but slightly longer than the
corresponding distance of 2.189(6) A found in a-[P6Se12]4'. a— and fl-[P6Se12]4' were
isolated as Rb+ and Cs+ salts, respectively, but it is unclear if the different structures are
due to differences in packing forces associated with the alkali metal size.

Phase-Change Behavior. Differential thermal analysis (DTA) of CsaPsset 2
performed at a rate of 10 °C min'l showed reversible crystal-glass phase transition, Figure
6-3a. Upon heating the compound melted at 424 °C and upon cooling the melt solidified
to a red glass. Only subsequent heating restored the crystals with an exothermic event at
263 °C followed by melting at 424 °C. Glass transition was observed at 204 °C for each
cooling step. The amorphous nature of the glassy phase was confirmed with X-ray
powder diffraction. The X-ray powder patterns after recrystallization were the same as
those of pristine Cs4P65e12 indicating full recovery of the original crystal structure, Figure
6-3b. Recrystallization and vitrification were reversible as they could be repeatedly
observed over many cycles. At cooling rates of 10 °C min'l Rb4P6Set2 did not glassify
suggesting that its glass crystallizes faster than that of CsaP68e12. DTA of C35P53e12
performed at a rate of 10 °C min'l exhibited the similar reversible phase-transition

behavior but it melted at 424 °C, formed a red glass at 250 °C and recrystallized at

162

243 °C, Figure 6-3c. The X-ray powder patterns before and after DTA of CsaP6Se1 2 is

shown in Figure 6-4d.
a b
(1,0. '-....()
0 .
_10 Iendo. ' ‘

-20 .
20 -

 

   
  

 

DTA, pm

 

Intensity (a. u.)

 

 

 

 

 

100 200 300 400 500 600 130 2‘0 3'0 4'0 5'0 6'0 7'0 80
Temperature, °C 2 theta, deg

 

 

 

 

DTA, 1.1m
Intensity (a. u.)

 

 

 

1-

 

 

150250350450589001003040506070
Temperature, °C 2 theta, deg

Figure 6-3. Differential thermal analysis diagrams of (a) CsaP6$e12 and (c) Cs5P58e12
showing melting in the 1st cycle with no crystallization on cooling (upper line). Glass
crystallization is observed in the 2nd heating cycle. CsaP5Set2 and CssPSSe12 are a pristine
crystal at A, glass at B and restored crystal at C, respectively. X-ray powder diffraction
patterns of (b) CsaPGSet2 (d) CssPSSen pristine (below), glassy (upper) and recrystallized
crystal(middle).

163

(a) 100 . .

 

 

 

 

 

 

 

 

 

 

 

  

I I I 219
80 - -
5: 60 - 2
s?
0)
g 40
5
20
0 I I I I I I 7
550 500 450 400 350 300 250 200
Raman Shift, cm“1
(b) 120 . 217 ..
100 r .
3: 80 h ..
5!;
if 60 - 368 351 .
8 486447 391
e 4, M 21
217
368
20 L- ..
349

 

 

 

0 I I
550 500 450 400 350 300 250 200
Raman Shift, cm“1

Figure 6-4. Raman spectra of crystalline (upper trace) and glassy (lower trace) of (a)
CS5P5SC12 and (1')) (3841368612.

164

 

0 exp. datild f
fitting
—— P“+ 2p3/2=139.9
----- P4+ 2p1/2=140.8
—— 193+ 2p3/2=137.7
__ ----- P3+ 2p1/2=138.6
” , ratio P4+IP3+= 4.1

  
   
   

 

 

 

Photoelectron Intensity (a.u.)

J

144 142 140 138 136 134 132

 

 

 

 

Binding energy, eV
(b) 0 exp. data
° " fitting
—— P4+ 2p3/2=139.0

----- P4+ 2p1/2=140.1
—— 132+ 2p3/2=137.5
. - - - = 132* 2p1/2=138.3
ratio P‘WP2+ :19

I
- - r'
\\“:: ""
\ .

‘Q.
I ‘ -
““;:o--
\ ‘ ‘ '

.
.5-‘,4""
--

'I
M
I
""'t-Il|to.c.;

Photoelectron Intensity (a.u.)

 

 

142 140 138 136 134 132
Binding energy, eV

 

Figure 6-5. The X-ray photoelectron spectrum, peak fitting, and deconvolution profiles

in the P 2p region of (a) CssPSSe12 and (b) CsaPGSe12.

165

The Raman spectra of CssPSSet 2 display shifts at 219(s), 268(5), 349(w), 370(w),
478(w) and for CsaP68e12 221(3), 351(m), 368(m), 391(w), 447(w), 486(w) and 519(w)
cm", Figure 6-4. The spectra for both compounds were very similar. For CsaPGSet2, the

shift at 221 cm”I can be assigned to the P2Se6 stretching mode by comparison in the Ag
stretching mode of D3“; symmetry of [P2Seé]4' ligand.1° Other peaks at 486 and 519 cm'1

are also related to the P2Se6 fragment.19 Peaks at 351 and 368 cm’] were similarly
observed in the spectrum of Rb4P63612. The Raman spectra of glassy Cs5P58e12 and
Cs4P6Set2 showed broader and weaker peaks at 218 and 358 cm‘1 for the former and 217,
349, and 368 cm'1 for the latter. These are at similar positions to those of the crystalline
phase suggesting that local structural motif is preserved in the glass, but crystallographic
long range order is lost as seen in the X-ray powder diffraction patterns of glassy
CssPSSe12 and C54P6Set 2.

The X-ray Photoelectron Spectra of Cs5P58e12 and C54P6Se12 confirmed the
presence of two different oxidation states in the compounds, Figure 6-5. Since the higher
binding energy results from the higher oxidation state of a corresponding element in
principle, for CssP5Se12 peaks at higher energy (140.8, 139.9 eV) are assigned to P4+
centers and peaks at 138.6 and 137.7 eV are assigned to P”. The XPS analysis revealed
ratio of lower/higher oxidation state of P to be 4.1 and 1.9 for CssPSSen and CsaP6Se12,
respectively, which supported the structural analysis: Css[P3+{P4+2Se(,}2] and

C9102“). 0433.212].

166

 

 

 

 

 

03513589.: 'P(1)l I I
”1(2)
P(3)
W 11W

os,P,Se,, P(1), P(2)
P(3)

* * * *
M --
' I I

200 160 120 80 4O 0 -4O -80
Chemical Shift, ppm

 

 

 

Figure 6-6. 31P solid-state NMR spectra of CssP5Se12 and Cs4P6Set2 at a 14 kHz MAS

frequency. * denotes spinning side bands and # Cs4P6$et2 impurity.

167

(b)

Figure 6-7. The electronic absorption spectra of crystalline and glassy (a) CssPSSe12 and

(b) CS4P68612.

Absorption, a/S

0.1

p
.15.

Absorption, 0118
_o o
N 00

0.1

 

 

 

 

 

 

 

Crystal
- / 2.17 eV
1 2 3 4 5
Energy, eV

     

_. Crystal
2.17 eV

 

 

   

1 2 3 4 5 6

Energy, eV

168

 

3'P Solid-state NMR Spectroscopy. Room temperature solid-state 3'P NMR
spectra of CssPSSe12 and CsaP6Set2 were collected, Figure 6-6. The spectrum of CsaP68e12
contained two doublets centered at 95.5 ppm and 65.5 ppm with coupling constants of J =
250 and J = 270 Hz, respectively, with an approximately 1:2 intensity ratio. Assignment
of these peaks is to the P2+ and PM atoms, respectively. The P(l) and P(2) chemical
environments are very similar, and the bond lengths and angles for each nearly agree
within the error of the crystallographic refinement. It is therefore reasonable that the
chemical shifts corresponding to these atoms overlap, thus accounting for the observed
1:2 ratio of the spectral peaks. One possible assignment of the observed splittings is two
bond P-P J-coupling. Two bond P-P scalar coupling in the solution phase have been
observed with coupling constants of ~300 Hz.2° However, other compounds such as
KBiP2S72' and Rb4P6Set25 contained inequivalent 31P atoms separated by two bonds and
two bond P-P coupling was not observed. Another possible assignment is that the 95.5
and 65.5 ppm peaks correspond to the P(l) and P(2) atoms and the observed splitting is
one-bond P(l)-P(2) J coupling. In this assignment, the P(3) signal was not detected. It is
noted that the 95.5 and 65.5 ppm peaks were the only significant isotropic peaks observed
under the conditions of a 5000 s delay between acquisitions and a 1200 ppm spectral
window. One weakness in this assignment is that the 95.5/65.5 ppm intensity ratio is <1.

The spectrum of CssPSSen showed two doublets centered at 63.1 ppm and 48.7
ppm with coupling constant of J = 200 Hz and singlet centered at 16.1 ppm, Figure 5-6.
The difference in bond lengths and angles between the P(l) and P(2) coordination
environments is significantly outside of the range of the error in the crystallographic

refinement, and therefore distinct chemical shifts resulting from each of these sites are

169

predicted. Including the signal from the P(3) atom, there should therefore be three peaks
in the NMR spectrum. One possible assignment is that the two doublets centered at 63.1
ppm and 48.7 ppm can be attributed to the P(l) and P(2) nuclei and the singlet at 16.1
ppm can be attributed to the P(3) nucleus. There are twice as many P(l) and P(2) atoms
per formula unit than P(3) atoms, so the spectral intensity of the peaks corresponding to
P(l) and P(2) should be twice that of P(3). The spectral intensity of the 63.1 ppm and
48.7 ppm peaks is ~3.5 and ~2.5-fold greater than the 16.1 ppm peak, respectively. The
peak at 94.9 ppm is likely due to Cs4P6Se12 impurity in the sample and this impurity
would also result in a 65.5 ppm peak which would be unresolved from the 63.1 ppm peak
of C55P58e12. The CsaPsset 2 impurity could therefore account for some of the
discrepancy between the observed and predicted intensities of this assignment. The
splitting of the 63.1 and 48.7 ppm peaks is attributed to one-bond P-P coupling between
the crystallographically inequivalent P(l) and P(2) atoms. The two bond P(1)-P(3) or
P(2)-P(3) scalar coupling was not observed. Previously, P4+ in selenophosphate anions
with P-P bonds were observed to have positive 3‘IP chemical shifts, and P” in
selenophosphate anions without P-P bonds were observed to have negative 3|P chemical
shifts.22 The chemical shift of the non P-P bonded P3+ atom in this compound, namely
P(3), was assigned to a spectral feature with a positive chemical shift. The longitudinal
relaxation times were ~600(100), 70(10), and 10(3) 5 for the 63.1, 48.7, and 16.1 ppm
features, respectively. The latter two values are somewhat shorter than those of typical
selenophosphates.l4 Phosphorus-containing compounds with multiple
crystallographically unique 31P sites typically do not have such a wide range of

longitudinal relaxation times for the 3 IP sites.

170

The solid state electronic absorption spectra revealed sharp absorption edges for
CssPSSel2 and CS4P68612 at approximately the same energy of 2.17 eV, consistent with
their orange color, Figure 6-7. The energy gap of the glassy phases was measured at 2.07
and 1.98 eV, respectively. The lower energy gaps of the glasses can be attributed to
structural defects that create mid-gap states and band tailing.23 A spectral shift in the

absorption edge is a key feature in creating nonvolatile memory devices.24

 

 

Figure 6-8. (a) Representative SEM image of CssPssen glass fiber. (b) X-ray powder

diffraction pattern of CssP5Se12 glass fiber showing its amorphous nature.

Second Harmonic Generation Response of Crystalline CSSPSSen. Because of
the non—centrosymmetric crystal structure of CssPsscn and its good optical transparency

from the edge of the energy gap to the mid-IR region, we investigated its second

171

harmonic generation (SHG) response at room temperature, Figure 3d. Using a modified
Kurtz powder technique we measured polycrystalline samples of 45-63 pm size using
1200-2000 nm fundamental idler radiation from a trmable laser.” The SHG intensities
obtained were compared with those of LiNbO3 and AgGaSe2, which is a representative
NLO material for [R applications.26 All samples were similarly prepared, and the same
particle size range and identical laser setting were used. The SHG intensity of CssP58e12
was approximately equal to that of LiNb03 and 25% that of AgGaSe2.24 As the particle
size of C55P5Se12 increased, the SHG intensity continuously decreased. Generally, when a
powder sample is non-phase-matchable, SHG sensitivity peaks near the coherence length,
which is typically 1~20 ,um, and it starts to diminish through larger particle sizes. In this
regard, Cs5P5Se12 is type-I non-phase—matchable.25 Despite this such materials can be
useful through ‘random’ quasi-phase-matching.28

Second Harmonic Generation Response of Glassy CSsPSSeu. Surprisingly,
glassy CssPSSet2 also exhibited significant SHG response, at ~ 5% that of AgGaSe2.
There have been tremendous efforts to induce SHG in glasses because of technological
importance by its great transparency and formability, e. g. fabricating optical fibers.29 It is
possible that this is a rare example of observation of SHG response from an amorphous
material with no specific treatment such as thermal poling, electron beam irradiation and
so on.30 Since CssP58e12 is a phase-change material and retains its local structural motif
in the glassy state, as shown by Raman spectroscopy, we can expect the
noncentrosymmetric arrangement to be partially preserved and plausibly give rise to
some SHG response. Indeed, SHG of glass sample was observed and X-ray powder

diffraction patterns taken after the measurements still showed a predominantly

172

amorphous nature for the sample. It cannot be completely ruled out, however, that idler
beam-induced crystallization of glass may be occurring or the existence of nanocrystals
embedded in glass matrix due to its phase-change property. Further investigation of the
glass thin film is in progress.

CssP5$e12 glass fibers were easily fabricated by rapid cooling extraction from the
corresponding hot melt at 600 °C. Representative SEM image showed the relatively
uniform formation of glass fiber and its amorphous nature was confirmed by X-ray
powder diffraction patterns, Figure 6-8. When the glass fiber was heated up and annealed
at 320 °C, crystallinity was fully recovered. This suggests that phase-change material
with NLO property can be a potential source for facile fabrication of NLO glass fiber for

IR application.

5. Concluding Remarks

The results of this study demonstrate that novel species in the chalcophosphate family
can be stabilized with reduced P atoms in their structure. The low valent phosphorus in
CssPSSen and Cs4P6Se12 is expressed with the central P" and P2“ centers playing the
role of a metal center to form the unusual coordination complexes {P[P2Se6]}°' and
{P2[P2Se6]}4'. Both compounds exhibit phase-change behavior by forming glasses. The
crystalline and glassy Cs5P5Se12 showed significant SHG response in the near infrared
region of the spectrum. Glass fiber of CssP5Set2 was easily obtained by simple rapid
cooling extraction from the corresponding hot melt and its amorphous nature was

confirmed.

173

References.

(l) Garin, J .; Parthe, E. Acta Crystallogr., Sect. B, 1972, 28, 3672-3674.

(2) Francisco, R. H. P.; Tepe, T.; Eckert, H. J. Solid State Chem. 1993, 107, 452—459.
(3) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1995, 34, 5401-5402.

(4) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem, 1998, 37, 2582-2584.

(5) Chung, 1.; Karst, A. L.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem, 2006, 45,
2785-2787.

(6) Chung, 1.; Do, J .; Canlas, C. G.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem,
2004, 43, 2762-2764.

(7) Chung, 1.; Christos, C. D.; Jang, J. 1.; Canlas, C. G.; Weliky D. P.; Kanatzidis, M.
G. J. Am. Chem. S0c.2007, 129, 14996-15006.

(8) (a) Scott, B.; Pressprich, M.; Willet, R. D.; Cleary, D. A. J. Solid State Chem.
1992, 96, 294-300. (b) Carpentier, C. D.; Nitsche, R. Mater. Res. Bull. 1974, 9, 1097-
1100.

(9) (a) Clement, R.; Lacroix, P. G.; Ohare, D.; Evans, J. Adv. Mater. 1994, 6, 794-
797. (b) Lacroix, P. G.; Clement, R.; Nakatani, K.; Zyss, J.; Ledoux, 1. Science, 1994,
263, 658-660.

(10) (a) Lemehaute, A.; Ouvrard, G.; Brec, R.; Rouxel, J. Mater. Res. Bull. 1977, 12,
1191-1197. (b) Thompson, A. H.; Whittingham, M. S. US Pat, 4 049 879, 1977.

(11) Huang, Z. L.; Cajipe, V. B.; Lerolland, B.; Colombet, P.; Schipper, W. J .; Blasse,
G. Eur. J. Solid State Inorg. Chem. 1992, 29, 1133-1144.

(12) Breshears, J. D.; Kanatzidis, M. G. J. Am. Chem. Soc. 2000, 122, 7839-7840.
(13) (a) Reflectance Spectroscopy, Wendlandt, W. W. and Hecht, H.G.; Interscience
Publishers: New York, 1966. (b) Reflectance Spectroscopy, Kotiirn, G.; Springer-Verlag:
New York, 1969. (c) Tandon, S. P.; Gupta, J. P. Phys. Status Solidi, 1970, 38, 363-367.

(14) Canlas, C. G.; Muthukumaran, R. B.; Kanatzidis, M. G.; Weliky, D. P. Solid
State Nucl. Magn. Reson. 2003, 24, 110-122.

(15) SMART, SAINT, SHELXTL: Data Collection and Processing Software for the
SMART—CCD system; Siemens Analytical X-ray Instruments Inc.: Madison, WI, 1997.

(16) Bondi, A. J. Phys. Chem, 1964, 68, 441-451.

174

(17) Chondroudis, K.; Chakrabarty, D.; Axtell, E. A.; Kanatzidis, M. G. Z. Anorg.
Allg. Chem, 1998, 624, 975-979.

(18) (a) Brockner, W.; Ohse, L.; Pfitzmann, U.; Eisenmann, B.; Schafer, H. Z.
Naturforsch., A, 1985, 40, 1248-1252. (b) Pfitzmann, U.; Brockner, W. Z. Naturforsch., A.
1987, 42, 515-516.

(19) Aitken, J. A.; Evain, M.; lordanidis, L.; Kanatzidis, M. G. Inorg. Chem. 2002,
41, 180-191.

(20) (a) Pregosin, P. s.; Kunz, R. w. 3119 and 13C NMR of Transition Metal Phosphine
Complexes. Springer-Verlag: Berlin ; New York, 1979; p 156 p. (b) Perera, S. D.; Shaw,
B. L. J. Chem. Soc. Chem. Commun. 1995, 865-866.

(21) (a) Gave, M.; Malliakas, C. D.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem.
2007, 46, 3632-3644. (b) McCarthy, T.; Kanatzidis, M. G. J. Alloys Compd. 1996, 236,
70-85.

(22) Canlas, C. G.; Kanatzidis, M. G.; Weliky, D. P. Inorg. Chem. 2003, 42, 3399-
3405.

(23) (a) Marking, G. A.; Hanko, J. A.; Kanatzidis, M. G. Chem. Mater. 1998, 10,
1191-1199. (b) Dhingra, S.; Kanatzidis, M. G. Science 1992, 258, 1769-1772.

(24) (a) Kyratsi, T.; Chrissafis, K.; Wachter, J. B.; Paraskevopoulos, K. M.;
Kanatzidis, M. G. Adv. Mater. 2003, 15, 1428-1431. (b) Maeda, Y.; Andoh, H.; lkuta, 1.;
Minemura, H. J Appl. Phys. 1988, 64, 1715-1719. (c) Maior, M. M.; Rasing, T.; Eijt, S.
W. H.; Vanloosdrecht, P. H. M.; Vankempen, H.; Molnar, S. B.; Vysochanskii, Y. M.;
Motrij, S. P.; Slivka, V. Y. J. Phys: Condens. Matter, 1994, 6, 11211-11220. (0)
Wachter, J. B.; Chrissafis, K.; Petkov, V.; Malliakas, C. D.; Bilc, D.; Kyratsi, T.;
Paraskevopoulos, K. M.; Mahanti, S. D.; Torbrugge, T.; Eckert, H.; Kanatzidis, M. G. J.
Solid State Chem. 2007, 180, 420—431.

(25) (a) Kurtz, S. K.; Perry, T. T. J Appl. Phys. 1968, 39, 3798-3813. (b) Dougherty,
J. P.; Kurtz, S. K. J. Appl. Crystallogr. 1976, 9, 145-158.

(26) Nikogosyan, D. N. Nonlinear optical crystals: 0 complete survey. Springer-
Science: New York, 2005.

(27) Choy, M. M.; Byer, R. L. Phys. Rev. B 1976, 14, 1693-1706.

(28) Baudrier-Raybaut, M.; Haidar, R.; Kupecek, Ph.; Lemasson, Ph.; Rosencher, E.
Nature, 2004, 432, 374-376.

(29) Antonyuk, B. P.; Antonyuk, V. B. Opt. Commun. 1998, 147, 143-147.

175

(30) (a) Fokine, M.; Saito, K.; Ikushima, A. J. Appl. Phys. Lett. 2005, 87, 171907. (b)
Fujiwara, T.; Takahashi, M.; Ikushima, A. J. Appl. Phys. Lett. 1997, 71 , 1032-1034. (c)
Qiu, M. X.; Pi, F.; Orriols, G Appl. Phys. Lett. 1998, 73, 3040-3042.

176

Chapter 7
K4P3Te4: A New Phase-Change Compound with P-Te Bonding, Exfoliation, and

Conversion to Photoluminescent Solution

1. Introduction

Solid-state materials with P-Te bonding have long been elusive although
phosphorus, in general, has been known to combine with nearly all elements.l This is in

striking contrast to the well-established metal chalcophosphate class involving the ternary
(M/P/Q) and quaternary (A/M/P/Q) compounds with [PnyF' anions in their structure,

where M is a metal, A is an alkali metal, and Q is sulfur or selenium. The only reported
P/Te containing inorganic compounds are MPTe (M = Ru, Os,2 Ir3) and BanTe2.4 The
latter features P-Te bonding while only unit cell dimensions and spectroscopic data have
been reported for the former. On the other hand, organometallic compounds provide
some examples of P-Te bonded species, e.g. Et3PTeX2 (X = Cl, Br, I),5 TePPh2Ch3° and
Ph3PTe(Ph)I.7 The larger elements in group 15, As, Sb, and Bi, show a wealth of
chemistry with Te to produce a wealth of solid-state inorganic materials.8

Recently, we reported rational synthetic conditions to stabilize P—rich species rather than
the simple classical [PSe4]3' and [P2Se6]4' anions by using excess phosphorus in the flux.
Examples are mpgsel8 (A = K, Rb, Cs),9 111549.362,” Cs.P,Se,2, and Cs5PSSe12,” yet
with the exception of BaP4Te2 there is a paucity of species containing P-Te bonds.

Here we report on the new alkali telluropolyphosphate compound KanTea

featuring P-Te bonding and the unique infinite anion l/00[P3Te.t°']. The compound was

177

prepared with molten salt flux method at intermediate temperature and it shows reversible

phase-change behavior. KanTea was found to be soluble in hydrazine to give rise to

colloidal solution of 1/Oo[PgTe44’] species. Generally it is very difficult to disperse solids

in polar solvents. Because of this, very little is known about the behavior of inorganic
polymers in solution, in some case these solutions have shown mesogenic liquid crystal
properties e.g. 1/(,O[PdPS.{]12 and 1/C,O[Mo3Se3’].l3 The possibility to exfoliate or even
dissolve mineral compounds is of major importance because it then allows modification
of the solids by solution chemistry methods and paves the way to the synthesis of new
organic-inorganic hybrid and nanocomposite materials. The tellurophosphate anion can
be dispersed in hydrazine solution. Nanospheres of the compound were obtained by
precipitation with ethanol at room temperature. The resulting solution exhibited slightly

blue-tinted white photoluminescence at room temperature when excited above the energy

gap.

2. Experimental Section

2.1. Reagents. The reagents mentioned in this work were used as obtained unless noted
otherwise. The reagents mentioned in this work were used as obtained: K metal
(analytical reagent, Aldrich Chemical Co., Milwaukee, WI); red phosphorus powder, -
100 mesh, Morton Thiokol, Inc., Danvers, MA; Te (99.999%; Noranda Advanced
Materials, Quebec, Canada); N,N—dimethylformamide (Spectrum Chemicals, ACS
reagent grade); diethyl ether (Columbus Chemical Industries, Columbus WI, ACS

reagent grade, anhydrous). K2Te starting materials was prepared by reacting

178

stoichiometric amounts of the elements in liquid ammonia. Anhydrous hydrazine (98 %,
Aldrich) was distilled before use. CAUTION: Hydrazine is highly toxic and should be
handled proper protective equipments with special care to prevent contact with
either the vapors or liquid.

2.2. Synthesis. All sample preparation processes were carried out under inter atmosphere.
Pure KanTe4 was achieved by heating a mixture of K2TezP:Te = 3:2:5 under vacuum in a
silica tube at 450 °C for 6 days, followed by cooling to 250 °C at 2 °C h". The excess flux
was dissolved with degassed N,N-dimethylfonnamide (DMF) under a N2 atmosphere to
reveal deep red-tinted black needle crystals. Energy-dispersive spectroscopy microprobe
analysis on five crystals showed an average composition of “K3,3P4Te3,7”. The single

crystals are stable in DMF, N-methylformamide and deionized H20 and air.

3. Physical Measurements.

X-ray Powder Diffraction. Phase purity x-ray diffraction analyses were performed using

a calibrated CPS 120 INEL X-ray powder diffractometer (Cu Ka graphite

monochromatized radiation) operating at 40 kV/20 mA and equipped with a position-
sensitive detector with flat sample geometry.

Scanning Electron Microscopy. Semiquantitative analyses of the compounds were
performed with a JEOL J SM-35C scanning electron microscope (SEM) equipped with a

Tracor Northern energy dispersive spectroscopy (EDS) detector.

179

Transmission Electron Microscopy (TEM) and High Resolution TEM. TEM sample
was diluted with ethanol and irradiated by ultrasonification. TEM and HRTEM images
were obtained with J EOL J EM 2200 F S Field emission TEM.

Solid-State UV-vis Spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region. The details of
the energy gap from these measurements have been discussed elsewhere.14

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CsI matrix. The
samples were ground with dry CsI into a fine powder and pressed into translucent pellets.
The spectra were recorded in the far-IR region (600-100 cm], 4 cm"l resolution) with the
use of a Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon
beam splitter.

Thermogravimetric Analysis. Experiments were performed on Shimadzu TGA-50
thermal analyzer by heating the samples up to 500 °C at a rate of 10 °C min'1 under N2
flow of ca. 20 mL min".
Differential Thermal Analysis (DTA). Experiments were performed on Shimadzu
DTA-50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed
in a silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. The sample was heated to 600 °C
at 10 °C min", and after 1 min it was cooled at a rate of —10 °C min'1 to 50 °C. The

residues of the DTA experiments were examined by X-ray powder diffraction.

Reproducibility of the results was confirmed by running multiple heating/cooling cycles.

180

The melting and crystallization points were measured at a minimum of endothermic peak
and a maximum of exothermic peak.

31P Solid-State NMR Spectroscopy. Room temperature 31P NMR measurements were
taken on a 9.4 T 400 MHz Varian Infinity Plus NMR spectrometer using a double-
resonance magic angle spinning (MAS) probe using a 4 mm (outer) diameter zirconia
rotor. Bloch decay spectra were taken with the excitation/detection charmel tuned to 3‘P
at 161.82 MHz with a 4.5 ,us, 90° pulse (calibrated to $0.5 as), a relaxation delay of 20 to
13,000 5, and samples were spun at frequencies between 6 and 13 kHz. All spectra were
processed with up to 100 Hz of line broadening, up to a tenth-order polynomial baseline
correction, and the chemical shifts (CS) were externally referenced to 85 % H3PO4 at 0
ppm. The spin lattice relaxation time (T 1) of each chemical shift was estimated fi'om the
exponential buildup of the peak intensity area as a function of the relaxation delay at 13
kHz between 20 and 13,000 3. The chemical shift anisotropy (CSA) principle value of
each chemical shift was derived from the average of the Herzfeld-Berger fitting‘5 of the
MAS isotropic and sideband peak intensity areas at spinning frequencies of 8, 8.5, and 10
kHz for the peak at -4.3 ppm, at 6, 8, 8.5, 10, and 12 kHz for the peak at -l6.9 ppm and at
8, 8.5, 10, 12, and 13 kHz for the peak at -34.6 ppm (outliers were excluded). 1

3'P Solution-State NMR Spectroscopy. Room temperature 31P NMR measurements
were taken on a 300 MHz Varian Mercoury NMR spectrometer. Spectra were taken with
a 675° pulse and a relaxation delay of 0.6 s at a frequency of 121.53 MHz. The chemical
shifts were externally refernced to 85 % H3PO4 at 0 ppm. Solution at a concentration of
ca. 1.25 X 10'2 mol L'1 was used for measurements. Preparation of K4P3Te4/hydrazine

solution will be described in Results and Discussion.

181

X-ray Crystallography. The crystal structure was determined by single-crystal X-ray
diffraction methods. Preliminary examination and data collection was performed on a
SMART'° platform diffractometer equipped with a 1K CCD area detector using graphite

monochromatized Mo Ka radiation at 173(2) K. A full sphere of data was collected at

173(2) K using narrow—frame method with scan widths of 030° in (o and exposure time
of 50 s flame". The SAINT software12 was used for data extraction and integration.
Semi-empirical absorption correction using the program SADABS12 was performed. The
initial positions for all atoms were obtained using direct methods, and the structure was
refined with the full-matrix least-squares techniques of the SHELXTL crystallographic
software package.l2 Satisfactory refinement was obtained with the centrosymmetric
P21/m. All atoms were refined to full occupancy and anisotropically. ADDSYM from
PLATON program17 was used for checking out higher symmetry. The parameters for
data collection, the details of the structural refinement, and fractional atomic coordinates
and displacement parameters are given in Table 7-1 and 7-2.

Atomic Pair Distribution Function Analysis. Fine powder (< 40 um) of the crystalline
and glassy KanTea was packed in a Kapton capillary with diameter of 1.0 mm.
Diffraction data were collected at room temperature using the rapid acquisition pair
distribution function technique.18 Data were collected using an MAR345 image plate
detector and ~ 60 keV energy X-rays (21 = 0.2128 A) at the ll-ID-B beam line at the
Advanced Photon Source. Measurements were repeated 4-5 times per sample to improve
counting statistics. The data were combined and integrated using the program FITZD.l9
Various corrections were made to the data, such as subtraction of background, Compton

and fluorescence scattering, geometric corrections, absorption, and so on, as described in

182

reference.2° Corrections were made using the program PDFgetXZ.2| Structural model was

fit to the data using the program PDFFIT.22

183

Table 7-1. Crystallographic Data and Refinement Details for KanTea

 

Formula

Crystal system
Space Group

Unit cell dimensions

Z

V, A3

(1 (calculated), gr cm'3
Crystal dimensions, mm3
Temperature, K

A, A

p, mm'
F(000)
amax (deg)

Total / unique reflections

Rint

No. Parameters

Refinement method

Final R indices [1 > 20(1)],
R la/ WRz b

R indices (all data), RI/wR2
Goodness-of-fit on F2

l

K4P3Te4
Monoclinic
P21/m (no.11)
a = 6.946(1), b= 6.555(1), c = 9.955(2),
B = 90.420(3)°
1

453.2(2)

3.351

0.114 x 0.010 x 0.008
173(2)

0.71073

7.969

404

28.30

4713 / 1167
0.0443
46

Full-matrix least-squares on F2
0.0397 / 0.0889

0.0560 / 0.0958
1.099

a R1 = z 11F01- chll/ZIFOI. b sz = {z IW(F02‘Fc2) 212180022211 ”2

184

Table 7-2. Atomic coordinates (X104) and equivalent isotropic displacement parameters

(A x 104) for KanTea at 173(2) K. U(eq) is defined as one third of the trace of the
orthogonalized Uij tensor.

 

 

x y 2' U(eq)
K(l) 261(3) 2500 3890(2) 18(1)
K(2) 2211(3) 2500 68(2) 19(1)
P(l) 3906(2) 52(3) 5802(2) 10(1)
P(2) 4918(4) 2500 2978(3) 10(1)
P(3) 4927(4) 2500 7169(3) 11(1)
Te(1) 7054(1) 2500 1014(1) 13(1)
Te(2) 8481(1) 2500 7335(1) 15(1)

 

185

(a)

P(1)

a an 0 HM! an a

P(1)/' \ a

2.463(3)A Te“) Te(2) 2-472(3)A

 

 

 

 

 

 

 

 

 

 

 

\ 11 _ {:71
-\___1 ‘3 1
at \.__1‘___::

\ " ~O‘

- || - ||
0- O- 1

 

 

 

1

3.8016(7)A

 

 

 

 

 

 

 

 

 

 

 

Figure 7-1. (a) View of l/oo[PgTe.t°‘] chain. P and Te atoms are labeled. Structure of

K4P3TC4 viewed down the (b) a- and (c) b-axes. Dashed line denotes short TemTe
interaction at Te(1)---Te(2), 3.8016(7) A. Darker large circles represent K atoms. The
thermal ellipsoids are shown with 90 % in (a) and 50 % probability in (b) and (c).

186

4. Results and Discussion

Crystal Structure. The K4P3Te4 adopts the P21/m space group and features
parallel infinite one-dimensional straight chains of 1/.,0[P3Te44'], Figure 7-1. This anion is
unique in that it has no sulfur or selenium analog. The only other compound which
features this species is the orthorhombic BaP4Te2. Due to the different number and size of
counter-cation of K+ and Ba”, the chains adopt different crystal packing. The P3
backbone of the chain is made of six-membered P.;-rings in a chair conformation
equatorially condensed via opposite edges of the hexagon defined by the P(l) atoms. This
creates 3 coordinate P atoms. The equatorial and axial Te(l) and Te(2) atoms are bonded
to P(2) and P(3) with distances of 2.463(3) and 2.472(3) A, respectively. This is
comparable to the sum of covalent radii at 2.47 A.23 Those in organometallic compounds
are also similar, for example, 2.473 A in Et3PTeBr2.5 The dihedral angles in the of
1/00[P3Te44'], of P(3)-P(1)-P(1)-P(3) and P(3)-P(l)-~-P(2)-P(l) are -180.0(1)° and -l.8(1)°,
respectively, showing the P3 backbone chain is nearly ideal. The P-P distances at 2.214(3),
2,217(2) and 2.220(2) A are comparable to those found in low valent chalcophosphate
compounds. Selected bond distances and angles are represented in Table 7-3.

Inter-chain Te---Te nonbonding interactions, which are shorter than the sum of

van der Waals radii of 4.40 A,24 are observed for Te(1)-"Te(2) at 3.8016(7) A, see Figure

7-1b and 1c. This weak interaction helps to organize the 1/00[P3Te44‘] chains arranged side
by side to form layers. K(l) atom, sandwiched by the 1/00[P3Te44'] chain, coordinates to

six P and four Te atoms, whereas K(2) atom, residing at the comer of four l/00[P3Te44']

chains, to two P and 5 Te atoms in a distorted pentagonal bipyramidal geometry, Figure 7-

187

2a and 2b. The distances of K to P atoms range from 3.359(3) to 3.814(3) A, and those to
Te atoms are at 3.487(2) to 3.744(3) A, Table 3. Note that K(1)-Te(1)-K(2)-P(2) are

coplanar and distance of K(1)'--K(2) is 4.0483 A, Figure 7-2c.

 

(C) Te(1)

1111’. 1

     
 

K(2) P(3)

    

P(1)
P(1)

Te(1) Te(2)

P(2) Te(2) Te(1)

Figure 7-2. Local coordination environment of K(l) and K(2) atoms. (a) K(l) atom
sandwiched by two 1/00[P3Te44‘] chains, and K(2) atom surrounded by four 1/00[P3Te44']
chains are shown. Blue broken line denotes ionic bonding of K atoms to P and Te atoms

in (a) and (b). (c) A K(l) and K(2) dimer is shown with K, P, and Te atoms labeled.
Thermal ellipsoids are shown with 90 % probability.

188

Table 7-3. Selected Bond Distances (A) and Angles (°) for K4P3Te4 at 173(2) K.

 

P(l)-P(1) 2.214(3) K(2)-P(2) 3.442(3)
P(1)-P(2) 2.220(2) x3 K(2)-P(3)vii 3.460(4)
P(1)-P(3) 2.217(2) x2 K(2)-Te(1) 3.487(2)
P(2)-Te(1) 2.463(3) K(2)-Te(1)viii 3.489(1)
P(3)-Te(2) 2.472(3) K(2)-Te(1))“ 3.488(1)
K(2)-P(1)iv 3.710(3)
K(l)-P(1) 3.541(3) K(2)-P(2)x 3.744(3)
K(l)—P(1)iaii 3.359(3)
K(1)-P(1)iii 3.541(3) p(1)v -p(1)-p(2)v 96.860 2)
K(1)-P(2) 3.366(3) P(1)v-P(1)-P(3) 104.2903)
1<(1)-P(2)iv 3.814(3) P0)iii-P(3)-P(2) 92.7203)
K(1)-Te(1)iv 3.615(3) p(3)-p(1)-p(2)v 95.44(9)
K(1)- Te(2)V.Vi 3.607(1) p(1)v-p(2)-Te(1) 102.29(9)
K(1)- Te(2)iv 3.653(3) P(1)-P(3)-Te(2) 110.84(9)

 

Symmetry transformations used to generate equivalent atoms:

(i) -x, y+1/2, -z+1; (ii) -x, -y, -z+1; (iii) x, -y+1/2, 2; (iv) x-l, y, 2; (v) -x+l, -y, -z+l;

(vi) -x+1, -y+1, -z+l; (vii) x, y, z-l; (viii) -x+l, -y, -z; (ix) -x+l, -y+1, -z; (x)x-1, y, z-l;
(xi) -x, -y+l, -z+l; (xii) -x, -y, -z

189

Phase-Change Behavior. Differential thermal analysis (DTA) of K4P3Te4
performed in a closed container at a rate of 5 °C min'l showed melting at 488 °C. Upon
cooling, the melt formed a black glass, Figure 7-3. On subsequent heating the glass
recrystallized at 455 °C and again melted at 488 °C to form a glass upon cooling. The
amorphous nature of the glassy phase was confirmed with X-ray powder diffraction. The
powder diffraction patterns after recrystallization were identical to those of the pristine
K4P8Te4 indicating full recovery of the original crystal structure, Figure 7-4. The
reversible crystal-glass transition was repeatedly observed in multiple DTA cycles

indicating congruent melting of the substance.

 

I I I I

1st cycle

0120. -
><
LLI

>1

B C O '
2nd cycle

_20 l l l I
100 200 300 400 500 600
Temperature, °C

 

 

 

 

Figure 7-3. DTA diagrams of K4P3Te4 showing melting in the lst cycle with no
crystallization on cooling (upper diagram). Glass crystallization is observed in the 2nd
heating cycle (lower diagram). K4P3Te4 is a pristine crystal at A, glass at B and restored
crystal at C.

190

140000
1 20000

1 00000

Intensity (a. u.)

20000

0' 'M mum A
20 30 4O 50 60

80000
60000

40000 :" ‘~ .- . h

 

l

 

l

l
r

 

l

 

 

10 70

2 theta, deg

Figure 7-4. X-ray powder diffraction patterns of theoretical simulation(A), glass(B) and

recrystallized crystal(C).

7

Absorption
0 —¥ N 00 ~l>~ ()1 O)

 

I I I I

Nanosphere

 

 

 

 

Energy, eV

Figure 7-5. UV-vis absorption spectra of glass, pristine crystal, and nanosphere
precipitated by MeOH.

191

Thermogravimetric analysis (TGA) at a rate of 10 °C min'l showed that the compound
decomposed under N2 atmosphere from ~3OO — 430 °C with a weight loss of 27.3 %. The
value is in excellent agreement with the theoretical P content in the compound of 27.09 %.
This suggested that the polyhexaphosphorus backbone was destroyed by heating at
>300 °C. X-ray powder patterns of the black residue after TGA revealed a mixture of (X-
K2Te2, K2Te3 and K5Te3 along with a small amount of amorphous phase. The DTA and
TGA results suggest that above the melting point the following equilibrium is in effect:
K4P3Te4 :2 2K2Te2 + 2P4T Eg.(1)

The solid-state UV-vis spectra of the crystalline and glassy phases show
absorption edges at 1.40 eV and 0.91 eV, respectively, Figure 7-5. The lower energy gaps
of the glass can be attributed to structural defects that create midgap states and band
tailing.25 The band gap value is consistent with the black color of the materials. The
infrared spectrum displays absorption peaks at 262(w), 306(5), 322(m), 417(m), 431(5),
and 469(m) cm". The peaks at 417 and 431 cm'1 can be assigned to P-Te by analogy to
MPT e (Ru, Os),2 and those at 323 and 469 cm'1 to a P-P vibration in comparison to
Ba3Pl4.26

Pair Distribution Function (PDF) Analysis. To probe the local structure of both
the crystalline and glassy K4P3Te4 we performed atomic pair distribution fimction (PDF)
analysis. This technique analyzes both the Bragg and diffuse scatterings to reveal the
short and intermediate range order of a solid regardless of the degree of disorder. Peaks in
the PDF directly represent the quantitative real-space interatomic distance correlation of
pairs of atoms in the structure in high resolution, thus the technique is a powerful

analytical tool for studying disorder and amorphous phases.27 The experimental PDF of

192

crystalline K4P3Te4 is in excellent agreement with the calculated PDF based upon the
crystal structure model, validating the correctness of crystal structure refinement, Figure
7-6. The first strong correlation at 2.4 A is assigned to the P-P and P-Te covalent bonds in
the one-dimensional chain. The second peak at 3.7 A corresponds to the distance of short
Te-"Te nonbonding interchain contact and that from K atom to the first neighboring P and
Te atoms, defining the connectedness of [PgTe44'] chain to K atom and the nearest chain.
The third peak at 4.8 A relates the second neighboring interchain Te---Te and KmP
correlations. The peak at 7.0 A results from the PP distances from two neighboring P

hexagons.

 

 

 

 

 

 

 

 

2 4 6 8‘101214‘1618 20
Radial Distance, A
Figure 7-6. Experimental pair distribution function G(r) of the crystalline (upper) and
glassy (bottom) K4P3Te4. The calculated PDF using the crystallographic coordinates is

shown as red line. Selected atomic correlation distances are indicated. Atomic distances
in the crystallographic model are also shown (inset).

193

The experimental PDF of the glass shows well-defmed correlations but up to ~14
A. The first two at 2.4 and 3.7 A and broad feature at 7 A are very similar to those in the
crystalline form. Note that the peak at 7 A defines the relation of two neighboring P
hexagons. Beyond ~14 A all correlations are lost reflecting the lack of long range
periodicity. The structural coherence is much shorter than the crystalline phase but
significantly longer than those of the conventional glasses such as silica. The PDF data
suggest that the crystalline and glassy phases of K4P3Te4 are structurally similar with
respect to their local structure which is defined by a relatively large invariant fragment.
This similarity indicates the rigidity of local structural unit [PgTe44], intact in the glassy
form. K4P3Te4 is a member of a new growing class of phase change materials28’29’30 with
stoichiometric composition and mixed ionic/covalent bonding in the structure. The vast
majority of chalcogenide glasses have nearly continuous compositions and all-covalent
networks. Although the mechanism of switching from glass to crystalline K4P3Te4 is not
yet understood, it is likely involves the two different types of binding found in the
structure, ionic (K--°Te) and covalent (P-P and P-Te). Such bonding anisotropy determines
glass formation tendency and in principle it could be used to control it. Detailed studies
are needed to understand how the coexistence of both ionic and covalent bonding affects
the phase-change and other physical properties in these materials.

Electronic Structure Calculations and Bonding. In order to explore the unusual
P-Te bonding in K4P3Te4 at a deeper level, we performed, ab-initio density functional
calculations with the fiJll-potential linearized augmented plane wave (FLAPW) method.31

These calculations can also determine the nature of the band gap.

194

 

 

(a) 2.0\ -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b) 2 0
//\J \.// ( _ Te (p)
/\ .\ E
l/j ”V3” 0, —P(1,p> F
0.5 A $315 —:(§.P)
> [Eg=146eV\ £3 _ (,p)
a) ....
é-1°;é::g§< 21°
C) (gbV/ng‘t- a .
LE 25;“i;::ib>v E06 ‘
- . A7»? ' a
i‘~<>\/Y>— ' ‘ ‘
7 J 2: / o ' . .
-4.0 5V V /V 0 2 4
F X S Y F B D Y Energy, eV
X
Y_El]_z
ChargeDensity
(electrons/an?)
. . 5
i 0.5
“ 0.2
7‘ z 0.1
F 1‘ 0.08
5' 0.06
m; 0.04
0.02
0.01

 

 

 

 

Figure 7-6. (a) The band structure of K4PgTe4. (b) The projected density of states for p-
orbitals of individual elements (Te and P), and (c) the contour plot of the total charge
density of K4P3Te4 calculated with the sX-LDA method.

195

The electronic structure of K4P3Te4 was calculated with the fully first-principles
screened-exchange local density functional approximation (sX-LDA) scheme32 and spin-
orbit coupling (SOC). The sX-LDA method is known to yield great improvements of the
excited electronic states such as band gaps and band dispersion due to its better long-
range description of the exchange-correlation hole compared to the local density
approximation (LDA). sX-LDA calculation predicted a band gap of 1.46 eV as shown in
Figure 7-7a, which is close to the experimental band gap at 1.4 eV. The valence band
maximum (V BM) occurs along the B-D direction, Figure 7-7a, while the conduction
band minimum (CBM) is located at the X point, which leads to an indirect band gap. The
electronic states between -1.5 eV and the VBM in the band structure are less dispersed
compared to the lower energy levels, which can be attributed to the localized character of
these orbitals, as explained below.

The angular-momentum—resolved density of states (DOS) for Te and P p-orbitals
are shown in Figure 5b. The DOS show a strong p-p mixing effect, which is very similar
to that of BaP4Te2[4]. The p-p mixing in K4P3Te4 makes for roughly three distinct regions
in the DOS for the energy between -6.5 eV and the VBM. The strong contribution of Te
p—orbitals to the energy states between -1.5 eV and the VBM are mainly due to the lone-
pair states of Te atoms. The small dispersion shown in this energy range of the band
structure is caused by these localized lone-pair states. The most interesting energy levels
in K4P3Te4 are located between -4.2 eV and -1.5 eV, which are derived mainly from P-Te
ppa-bonding. The covalent bonding character between Te and P can also be seen in the
contour plots of the total charge density, Figure 5c. The charge densities which are

relatively high along the Te-P and P-P bond axes clearly suggest a strong Te-P covalent

196

bonding character. The lower energy levels [at -6.5 eV ~ -4.2 eV] are mainly composed
of P p-orbitals due to the strong covalent interactions in P-P bondings.

The 31F solid-state magic angle spinning (MAS) NMR spectrum of K4P3Te4
revealed three isotropic chemical shifts at -4.3, -16.9, and -34.6 ppm, Figure 7-7,
consistent with the crystallographic analysis. Because the ratio of the integrated intensity
is l.1:2.0:1.0, the peak at -l6.9 ppm can be assigned to P(l). The P(1)-P(1) fiagment is
shared bond between the fused cyclohexane chairs of the P3 backbone and has no
bonding with Te. The peaks at -4.3 and -34.6 ppm can be attributed to atoms P(2) or P(3).
The 033 - all chemical shifi anisotropy (CSA) principal value differences were ~188, 116,
and 187 ppm, respectively, and also support the assignment. Because P(l) only has
homoatomic bonding in contrast to P(2) and P(3) which bond to Te, it can be considered
more crystallographically and magnetically symmetric than the others. For
selenophosphate anions, the CSA principal value difference correlates inversely with the
degree of local symmetry about R33 P(2) and P(3) have very similar local geometric
environment and could not be individually assigned.

The 31P solid-state magic angle spinning (MAS) NMR spectrum of K4PgTe4
revealed three isotropic chemical shifts at -4.3, -16.9, and -34.6 ppm, Figure 7-7,
consistent with the crystallographic analysis. Because the ratio of the integrated intensity
is 1.1:2.0:1.0, the peak at -16.9 ppm can be assigned to P(l). The P(1)-P(1) fragment is
shared bond between the fused cyclohexane chairs of the P8 backbone and has no
bonding with Te. The peaks at -4.3 and -34.6 ppm can be attributed to atoms P(2) or P(3).
The 033 - a“ chemical shift anisotropy (CSA) principal value differences were ~188, 116,

and 187 ppm, respectively, and also support the assignment. Because P(l) only has

197

homoatomic bonding in contrast to P(2) and P(3) which bond to Te, it can be considered
more crystallographically and magnetically symmetric than the others. For
selenophosphate anions, the CSA principal value difference correlates inversely with the
degree of local symmetry about R33 P(2) and P(3) have very similar local geometric

environment and could not be individually assigned.

13(1)

I3(2). P(3)

 

 

130 1'00 50 0 -50 -1'00
Chemical shift, ppm

Figure 7-7. 31P solid-state MAS NMR spectrum of K4P3Te4. * denotes spinning side
bands.

K4P3Te4/hydrazine solution (10 mL) could be evaporated to dryness by blowing
N2 for 2 days to yield a sticky amorphous black solid. TGA performed at a rate of 2 °C
min'l demonstrated that hydrazine trapped in the structure was removed by 250 °C. The
highly crystalline nanospheres were isolated by precipitation with anhydrous methanol,
centrifuged, filtered and dried. X-ray powder diffraction pattern of the precipitated
residue matched well with that of theoretical calculation, Figure 7-9. The band gap of the
recovered K4P3Te4 was measured at 1.62 eV, which is significantly blue shifted from

1.40 eV of the bulk pristine phase, Figure 7-5.

198

 

Figure 7-8. (a) TEM image of dispersed K4P3Te4 nanosphere. (b) Selected area electron
diffraction patterns of K4P3Te4 nanosphere. (c) Lattice image on single K4P8Te4
nanosphere by high resolution TEM, indicative of single crystalline nature in nanoscale.
Lattice planes of (012) and (102) are indexed. (d) The representative EDS spectrum of
single K4P3Te4 nanosphere giving a composition of “K4_3P3T63_5”, reasonably close to

K4P3Te4.

199

 

003)
(103)

 

(013)

Intensity (a. u.)
(100)
111)

 

 

 

 

 

20 3O 40 50
2theta, deg

Figure 7—9. X-ray powder diffraction patterns of (A) isolated K4P3Te4 nanosphere by
methanol and (B) theoretical simulation of bulk K4P3Te4.

4

 

Weaker concentration

N 00

.—\
l

Intensities (a. u.)

)-

 

 

 

 

800‘700‘600‘600‘400‘
Wavelength, nm

Figure 7-10. (a) Normalized UV absorbance spectra with respect to the variable
concentration of dissolved K4P3Te4 in hydrazine solution. Solution was consecutively
diluted by 50 % from I to II to III.

200

Solution Phase Chemistry of K4P3Te4. In order to figure out the dissolution
mechanism of K4P3Te4, we performed UV-vis absorbance in a varied concentration and
3'P solution-state NMR for K4P3Te4/hydrazine solution. Amax of UV-vis absorbance is
dramatically blue shifted (~3.0 eV) from the band bap of the bulk crystalline phase at
1.40 eV, Figure 7-5 and Figure 7-10. Successive dilution of the solution by hydrazine to a
half and a quarter concentration shifted the 11mm to the deeper UV region. It is understood
that the observed blue shift resulted fiom the separation of one-dimensional chain

1/oo[PgTe44'] due to the exfoliation/dissolution process of K4P3Te4 compound to give K+
cations and 1/00[P3Te44'] anions when loaded in highly polar hydrazine. Further lmax shift

resulted from the chain being further separated from one another upon dilution. Broad
feature at ~500 to 650 nm possibly resulted from the hydrazine shell effect as shown in

the nanoparticle/organic conjugate.34

 

 

 

j l l l l l l l l l I

60 4O 20 0 -20 -40
Chemical shift, ppm

 

Figure 7-11. 3'P solution-state NMR spectrum of K4P3Te4/hydrazine solution. * denotes

spinning side bands.

201

31P solution-state NMR spectrum of K4P3Te4/hydrazine solution at room
temperature displayed three isotropic chemical shifts at 9.6, 14.6, and 36.2 ppm with the
ratio of the integrated intensity of ~1 :12, Figure 11. Note that the NMR peak positions of
the solution is close to that of black phosphorus (22.2 ppm)35 that consists of two
dimensional network of phosphorus hexagons in a chair conformation. The observed
NMR peak positions of solution are different from those of the pristine solid by 31P solid-
state NMR. 31P solution-state NMR and UV-vis absorbance results with TEM images

showing nanospheres suggested that the 1/00[P3Te44'] chain was exfoliated and

fragmented during solvation as found in KNiPS4.12 If the dissociated 1/00[P3Te44'] chain

were intact, one-dimensional species must have been observed as previously found in
KPdPS412 and NaszPSm.36 No major structural change or decomposition during
solvation, however, is likely to have occurred because (1) “K4P3Te4” composition on the
nanosphere was confirmed by EDS and scanning TEM, and (2) rapid recovery of
crystalline K4P3Te4 phase from solution by precipitation with MeOH. Because we could
not obtain crystals that possess the structure of fragmented [PsTC4]n ' species, no
structural information or assignment of 31P solution-state NMR spectrum was available.
Additional work will be necessary to figure out the more detailed dissolution and
nanosphere formation mechanism of K4P3Te4 and solution NMR spectrum.

The K4P3Te4/hydrazine solution exhibited photoluminescence (PL) at room
temperature, Figure 7-12. It is noteworthy that the PL entirely spans the visible region
(~420 — 690 nm) with unusually large Stokes shift (~81 run). When a 355nm Nd:YAG
laser with pulse width of 15ps was introduced, white light with slight blue-tint was

observed (see inset). Statistically, the PL peak should be symmetric, but the observed

202

peak shape is asymmetric due to reabsorption of the PL. We can approximately reproduce
a symmetric peak by multiplying the observed absorption by exp[a(E)d], where d ~ 1cm
is the thickness of a cuvet. As confirmed by the superimposed Gaussian fit, the predicted
peak distribution is symmetric around 2.46 eV. The PL profiles of the K4P3Te4
nanospheres remained unchanged when hydrazine solution was diluted, suggesting the
observed PL may originate from HOMO-LUMO electronic transitions in the solubilized

[PsTe4],,"' fragment rather than from impurities.

 

IIIIIIIIIIIIIIIII‘Il—I—I—rIITTjI'IIII

Intensities (a. u.)

 

 

 

600 400 200 0
Wavelength, nm

Figure 7-12. Room-temperature PL (A) under excitation at 380 nm (B) and absorption
(C) spectra obtained from K4P3Te4/hydrazine. The solid line (D) corresponds to the
predicted PL after correcting the reabsorption effect, superimposed by a symmetric
Gaussian fit(1ine with O). The inset exhibits the white emission from K4P3Te4/hydrazine
solution in a cuvet when 355 nm Nd:YAG laser was introduced at the point marked by an
arrow. The nearly saturated deep reddish solution was much diluted to measure PL.

203

5. Concluding Remarks

Molten salt flux method was a rich source of new synthetic thio- and
selenophosphate compounds. Synthesis of K4P3Te4 exemplifies the availability of this

method to phosphorus tellurides that has long elusive. The compound features a unique

1/00[P3Te44'] chain with low valent P and rare P-Te bonding. Remarkably, K4P3Te4

compound readily forms various states of crystalline and glassy solid, nanosphere, and
solution, and each state is differentiated by significant optical absorption contrast. Glassy
phase is easily accessible by quenching out molten crystals to room temperature, and is
switchable to the corresponding crystalline phase by heat treatment. The optical
absorption edge for the glassy phase is red-shifted from the crystalline counterpart, and
on the basis of the PDF analysis it appears that the basic building blocks are largely intact
in the glassy state. This accounts for the facile restoration of the crystal structure from the
corresponding amorphous phase. K4P3Te4 is a member of the emerging class of phase
change materials that are compositionally stoichiometric and possess alternating ionic
and covalent bondings, in contrast to the conventional class that is mostly
nonstoichiometric with covalent bonding networks.

Solution-state K4P8Te4 is obtained simply by dissolving solid in hydrazine
solution at room temperature. This rare behavior can open up the possibility of new
chemistry: synthesis of wide range of metal tellurophosphate compound via a chimie
deuce route, functional organic-inorganic hybrids, and nanocomposite compounds; study
of gel-formation and liquid crystalline behavior. We suggested that solvation of K4P3Te4
was driven by an exfoliation/fragmentation process by using 3'P NMR, UV-vis

absorbance and TEM studies. The K4P8Te4 solution exhibited highly desirable white

204

emission with slight blue-tint by Nd:YAG laser irradiation. Highly crystalline K4P3Te4
nanosphere was obtained by precipitation with alcohol, giving rise to full recovery of
crystal structure from solution state. In this regard, precipitation reaction with various

salts is expected to produce new phosphorus tellurides in a crystalline form.

205

References
(1) von Schnering, H. G.; Honle, W. Chem. Rev. 1988, 88, 243-273.

(2) Lutz, H. D.; Schmidt, T.; Waschenbach, G. Z. Anorg. Allg. Chem. 1988, 562, 7-
16.

(3) Kliche, G. Z. Naturforsch., B 1986, 41 , 130-131.
(4) Jorgens, S.; Johrendt, D.; Mewis, A. Chem. Eur. J. 2003, 9, 2405-2410.
(5) Konu, J .; Chivers, T. Dalton Trans. 2006, 3941-3946.

(6) Domhaus, F.; Bolte, M.; Lerner, H. W.; Wagner, M. Eur. J. Inorg. Chem. 2006,
1777-1785.

(7) Boyle, P. D.; Cross, W. 1.; Godfrey, S. M.; McAuliffe, C. A.; Pritchard, R. G.;
Sarwar, 8.; Sheffield, J. M. Angew. Chem, Int. Ed. 2000, 39, 1796-1798.

(8) Poudeu, P. F. P.; Kanatzidis, M. G. Chem. Commun. 2005, 2672-2674; Hsu, K.
F .; La], 8.; Hogan, T.; Kanatzidis, M. G. Chem. Commun. 2002, 1380-1381; Sheldrick,
W. S.; Wachhold, M. Coord. Chem. Rev. 1998, 176, 211-322.

(9) Chondroudis, K.; Kanatzidis, M. G. Inorg. Chem. 1998, 37, 2582-2584.

(10) Chung, L; Karst, A. L.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem. 2006, 45,
2785-2787.

(11) Chung, I.; Jang, J. 1.; Gave, M. A.; Weliky, D. P.; Kanatzidis, M. G. Chem.
Commun. 2007, 4998-5000.

(12) Sayettat, J .; Bull, L. M.; Gabriel, J. C. P.; Jobic, S.; Camerel, F.; Marie, A. M.;
Fourmigue, M.; Batail, P.; Brec, R.; Inglebert, R. L. Angew. Chem, Int. Ed. 1998, 37,
1711-1714.

(13) Davidson, P.; Gabriel, J. C.; Levelut, A. M.; Batail, P. Europhys. Lett. 1993,21,
317-322.

(14) Mccarthy, T. J .; Ngeyi, S. P.; Liao, J. H.; Degroot, D. C.; Hogan, T.; Kannewurf,
C. R.; Kanatzidis, M. G. Chem. Mater. 1993, 5, 331-340.

(15) Herzfeld, J.; Berger, A. E. J. Chem. Phys. 1980, 73, 6021-6030.

(16) SADABS, SHELXTL, version 6; Bruker Analytical X-ray Instrumentals, Inc.:
Madison, WI, 2000.

206

(17) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7-13.

(18) Chupas, P. J.; Qiu, X. Y.; Hanson, J. C.; Lee, P. L.; Grey, C. P.; Billinge, S. J. L.
J. Appl. Crystallogr. 2003, 36, 1342-1347.

(19) Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D.
High Pressure Research 1996, 14, 235-248.

(20) Egarni, T.; Billinge, S. J. L. Underneath the Bragg peaks : structural analysis of
complex materials; 1st ed.; Pergamon: Amsterdam ; Boston, 2003.

(21) Qiu, X. Y.; Bozin, E. S.; Juhas, P.; Proffen, T.; Billinge, S. J. L. J. Appl.
Crystallogr. 2004, 37, 110-116.

(22) Proffen, T.; Billinge, S. J. L. J. Appl. Crystallogr. 1999, 32, 572-575.

(23) Pauling, L. The nature of the chemical bond and the structure of molecules and
crystals; an introduction to modern structural chemistry; 3d ed.; Cornell University Press:
Ithaca, N. Y., 1960.

(24) Cherin, P.; Unger, P. Acta Crystallogr. 1967, 23, 670-671.

(25) Marking, G. A.; Hanko, J. A.; Kanatzidis, M. G. Chem. Mater. 1998, 10, 1191-
1199; Dhingra, S.; Kanatzidis, M. G. Science 1992, 258, 1769-1772.

(26) Dahlmann, W.; von Schnering, H. G. Naturwissenschafien 1973, 60, 429-429.
(27) Billinge, S. J. L.; Kanatzidis, M. G. Chem. Commun. 2004, 749-760.

(28) Chung, 1.; Malliakas, C. D.; Jang, J. 1.; Canlas, C. G.; Weliky, D. P.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2007, 129, 14996-15006.

(29) Wachter, J. B.; Chrissafis, K.; Petkov, V.; Malliakas, C. D.; Bilc, D.; Kyratsi, T.;
Paraskevopoulos, K. M.; Mahanti, S. D.; Torbrugge, T .; Eckert, H.; Kanatzidis, M. G. J.
Solid State Chem. 2007, 180, 420—431.

(30) Kaidatzis, A.; Wachter, J. B.; Chrissafis, K.; Paraskevopoulos, K. M.; Kanatzidis,
M. G. J. Non-Cryst. Solids 2008, 354, 3643-3648.

(31) Wimmer, B.; Krakauer, H.; Weinert, M.; Freeman, A. J. Phys. Rev. B 1981, 24,
864-875.

(32) Asahi, R.; Mannstadt, W.; Freeman, A. J. Phys. Rev. B 1999, 59, 7486-7492;
Bylander, D. M.; Kleinman, L. Phys. Rev. B 1990, 41 , 7868-7871.

207

(33) Canlas, C. G.; Kanatzidis, M. G.; Weliky, D. P. Inorg. Chem. 2003, 42, 3399-
3405.

(34) Thakar, R.; Chen, Y. C.; Snee, P. T. Nano Lett. 2007, 7, 3429-3432.
(35) Lange, S.; Schmidt, P.; Nilges, T. Inorg. Chem. 2007, 46, 4028-4035.

(36) Camerel, F.; Gabriel, J. C. P.; Batail, P.; Davidson, P.; Lemaire, B.; Schmutz,
M.; Guilk-Krzywicki, T.; Bourgaux, C. Nano Lett. 2002, 2, 403-407.

208

Chapter8
Flexible Polar Nanowires of Cs5BiP4Se12 from Weak Interactions
between Coordination Complexes and Their Strong Second Harmonic

Generation Nonlinear Optical Response.

1. Introduction

Since the discovery of WSzl and M08; nanotubes,2 the synthesis of other one-
dimensional inorganic nanostructures is of great interest because of promising new
optical and electronic properties3 that can be useful in technological applications.
Synthetic efforts to create inorganic nanowires involve either the exploration of new
compounds with nanofeatures or downsizing existing materials to the nano-scale.
Examples of the former include SbPS43 and LiM03Se3.4 Examples of the latter include
MS; (M=Ti,5 Nb, Ta,6 Re7), BiZS3,8 BN,9 V205,10 and V8;11 and the synthetic procedures
to obtain such nanowires require special non-equilibrium conditions such as discharging,
chemical vaporization, laser vaporization, or hydrocarbon pyrolysis. Consequently, these
are limited in terms of yield, phase purity, crystallinity, and uniformity. Most of the
known nanowires/-tubes are simple atomic or binary phases. More complex multinary
inorganic solids in this class are still rare, yet such materials can exhibit important
properties such superconductivity,” giant magnetoresistance,13 ferroelectric,l4 liquid
crystallinity15 and optical nonlinearities.16

Here we describe the new compound CssBiP4Se12 which intrinsically grows as

nanowires. The compound is a semiconductor and shows a wide optical transparency

209

range through mid/near IR region and strong second harmonic generation (SHG)
response at 1 am. Integrated photonic networks with high wavelength conversion

efficiency in the IR region (especially l.3~1.5 ,um) are in high demand because of the
dramatic development of the broadband intemet communication industry.'7 Since optical
nano-waveguidance system can offer superior light confinement due a to strong index
contrast, it is ideal for nonlinear optical applications.'8 In this regard, Cs5BiP4Se12 can be
a potential candidate for such purposes. In spite of the nanowire microstructure and
fibrous nature of the material which made the selection of single crystals very
challenging we succeeded in determining the structure of the material using single crystal
X-ray crystallography on very thin fibers. The rather surprising finding is that the
repeating unit giving rise to the nanofibers is the simple coordination complex
[Bi(PZSe6)2]5'. Using the crystallographic information, ab-initio density functional
calculations with the full-potential linearized augmetned plane wave (F LAPW) method]9
were performed to confirm the experimental band gap and to explain the nanowire nature
of CssBiP4Se12. To the best of our knowledge, the CssBiP4Se12 presents a rare example of

polar nanowires.

2. Experimental Section

2.1. Reagents. The reagents mentioned in this work were used as obtained: Cs metal,
analytical reagent, Johnson Matthey/AESAR Group, Seabrook, NH; red phosphorus

powder, -100 mesh, Morton Thiokol, Inc., Danvers, MA; Se, 99.9999%, Noranda

Advanced Materials, Quebec, Canada; N,N-dimethylformamide, ACS reagent grade,

210

Mallinckrodt Backer Inc., Paris, KY; diethyl ether, ACS reagent grade, anhydrous,
Columbus Chemical Industries, Columbus WI. The CS2Se starting material was prepared

by reacting stoichiometric amounts of the elements in liquid ammonia under N2.

2.2. Synthesis. The synthesis of pure CssBiP4Se12 microfibers was achieved by reacting a
mixture of CS2SezBizP:Se=2.5:1:4:9.5 under vacuum in a fused silica tube at 850 °C for 5
(1, followed by washing with degassed N,N-dimethylformamide (DMF) under a N2
atmosphere (~80 % yield based on Bi). Alternatively, the title compound could be
obtained by a stoichiometric mixture of the same reagents in a fused silica tube at 850 °C
for 3h. Shorter reaction times produced typical tiny needles instead of microfibers.
Energy dispersive spectroscopic analysis showed an average composition of
"Cs4,9BiP3,gSe) 1,6" and "Cs4,gBiP3,9Sen,g" for the respective synthetic conditions. The
single crystals are stable in DMF, alcohol, and air at least over a week. They are soluble
in NMF to give deep orange solutions. CssBiP4Se12 nanowires were obtained by
ultrasonic irradiation of the microwire-shaped compound in EtOH for 10 min. For TEM
studies, the resulting colloidal solution was centrifuged at 4000 rpm for 1 hr and filtered

by a 0.2 pm syringe filter.

3. Physical Measurements.

X-ray Powder Diffraction. Phase purity X-ray diffraction analyses were performed

using a calibrated CPS 120 INEL X-ray powder diffractometer (Cu Ka graphite

monochromatized radiation) operating at 40 kV/20 mA and equipped with a position-

211

sensitive detector with flat sample geometry.

Scanning Electron Microscopy. Semiquantitative analyses and taking morphology
images of the compounds were performed with a JEOL ISM-35C scanning electron
microscope (SEM) equipped 'with a Tracor Northern energy dispersive spectroscopy
(EDS) detector.

Transmission Electron Microscopy (T EM) and High Resolution TEM. TEM sample
was diluted with ethanol and irradiated by ultrasonification. TEM and HRTEM images
were obtained with J EOL J EM 2200 PS (Field emission TEM).

Solid-State UV-vis Spectroscopy. Optical diffuse reflectance measurements were
performed at room temperature using a Shimadzu UV-3101 PC double-beam, double-
monochromator spectrophotometer operating in the 200-2500 nm region. The instrument
is equipped with an integrating sphere and controlled by a personal computer. BaSO4 was
used as a 100% reflectance standard. The sample was prepared by grinding the crystals to
a powder and spreading it on a compacted surface of the powdered standard material,
preloaded into a sample holder. The reflectance versus wavelength data generated, were
used to estimate the band gap of the material by converting reflectance to absorption
data20 according to Kubelka-Munk equations: a/S = (1 - R)2/(2R), where R is the
reflectance and a and S are the absorption and scattering coefficients, respectively.
Raman Spectroscopy. Raman spectra were recorded on a H010probe Raman
spectrograph equipped with a CCD camera detector using 633 nm radiation from a HeNe
laser for excitation and a resolution of 4 cm]. Laser power at the sample was estimated to
be about 5 mW, and the focused laser beam diameter was ca. 10 m. A total of 128 scans

were sufficient to obtain good quality spectra.

212

Infrared Spectroscopy. FT-IR spectra were recorded as solids in a CsI or KBr matrix.
The samples were ground with dry CsI or KBr into a fine powder and pressed into
translucent pellets. The spectra were recorded in the far-IR region (600-100 cm", 4 cm'1
resolution) and mid-IR region (500-4000 cm", 4 cm'1 resolution) with the use of a
Nicolet 740 FT-IR spectrometer equipped with a TGS/PE detector and silicon beam

splitter.

Differential Thermal Analysis (DTA) Experiments were performed on Shimadzu DTA-
50 thermal analyzer. A sample (~30 mg) of ground crystalline material was sealed in a
silica ampoule under vacuum. A similar ampoule of equal mass filled with A1203 was
sealed and placed on the reference side of the detector. The sample was heated to 800°C
at °C min", and after 1 min it was cooled at a rate of —10 °C min'1 to 50 °C. The residues
of the DTA experiments were examined by X-ray powder diffraction. Reproducibility of
the results was confirmed by running multiple heating/cooling cycles. The melting and
crystallization points were measured at a minimum of endothermic peak and a maximum
of exothermic peak.

X-ray Crystallography. Since CssBiP4Se12 crystallized as a very thin microwire form
that is a bundle of individual nanowires, it was extremely hard to find suitable single
crystals for X-ray diffraction studies. Most of single crystals chosen showed substantial
diffused peaks. Intensity data for CssBiP4Sel2 were collected at 293(2) K on a STOE

IPDS II diffractometer with Mo Ka radiation operating at 50 kV and 40 mA with a 34 cm

diameter imaging plate. Data collection at lower temperature was not favorable because
N2 stream affected the very thin wire-shaped single crystal. Individual frames were

collected with a 15 min exposure time and a 1.0 a) rotation. The X-AREA, X-RED and

213

X-SHAPE software package was used for data extraction and integration and to apply
empirical and analytical absorption corrections. The SHELXTL software package was
used to solve and refine the structure. The most satisfactory refinement was obtained with
the polar space group, Pmc21. The Cs(8) atom was modeled as a split site to Cs(8A) and
Cs(8B). Their occupancy ratio was refined to 95:5. The Flack parameter x refined to 0.
The parameters for data collection and the details of the structural refinement are given in
Table 8-1. Fractional atomic coordinates and displacement parameters for each structure
are given in Tables 8-2 — 8-4. In all cases the atoms refined to full occupancy.

Electronic Band Structure Calculations. The electronic structure calculations were
performed using the full potential linearized augmented plane wave (FLAPW) method19
within density functional scheme. The experimental lattice parameters and atomic
coordinates were employed for the calculations. The core states and the valence states
were treated fully relativistically and scalar relativistically, respectively. The spin-orbit
interaction was also included self-consistently by a second variational method.21 A 5X3><1
mesh of special k-points was used in the irreducible Brillouin zone and the energy cutoffs
for the interstitial plane-wave basis and the star functions were 13.0 and 144.0 Ry,
respectively. The muffin-tin radii of P, Se, Cs, and Bi were 1.9, 2.0, 2.7, and 2.8 bohrs,
respectively.

Nonlinear Optical Property Measurements. We used the frequency-tripled output of a
passive-active mode-locked Nd:YAG laser with a pulse width of about 15 ps and a
repetition rate of 10 Hz to pump an optical parametric amplifier (OPA). The OPA
generates vertically polarized pulses in the ranges 400 ~ 685 nm and 737 ~ 3156 nm. In

order to check the SHG efficiency as a firnction of the excitation energy, we tuned the

214

wavelength of the incident light from 1000 ~ 2000 nm. In this range, the spectral
bandwidth of the linearly polarized light from the OPA is rather broad, about 2 meV full
width at half maximum. However, the phase space compression phenomena ensures
effective SHG where lower energy portions are exactly compensated by higher parts
thereby satisfying both energy and momentum conservation. The incident laser pulse of
300 pl was focused onto a spot 500 am in diameter using a 3 cm focal-length lens. The
corresponding incident photon flux is about 10 GW cm'z. The SHG signal was collected
in reflection geometry from the excitation surface and focused onto a fiber optic bundle.
The output of the fiber optic bundle was coupled to the entrance slit of a Spex Spec-One
500 M spectrometer and detected using a nitrogen-cooled CCD camera. The data
collection time was 20 3. Single crystal CssBiP4Se12 sample and AgGaSe2 were ground
and separated by various size ranges using sieves. Samples were placed in capillary tubes
and measured. The SHG intensity of CssBiP4Sel2 in a wide range of 500 — 1000 nm was

compared to that of AgGaSe2.

215

Se3 Se1 Se9
(a Se1

   
     
 

    
 

P2 Se10 g”
82e ’ Se 12 Se12
Se4 )2.8(82
2.986(3)A 2 927(3 / 2.953(3)A
= Bi2
2914(3)A
2.832 3 A
( ) 2.937(3)A \2:17(3)A
Se6 Se8
Se14 ‘ 8616

Se7

$95 $913 $615 8815

Figure 8-1. Structure of two crystallographically independent [BiP4Se12]5' molecules. Bi,
P, and Se atoms are labeled. Bond distances (A ) are represented.

216

 

V'X'V‘V‘V.

r/I\.o ¢§o./¢\o/:\Q(I\o
o o 0
>30 \oonoXm
0.9;...0 0.4;.

0%.:0 43127::

Figure 8-2. (a) Structure of Cs5BiP4Se12 viewed down the a-axis (down the fiber
direction). Bi, P, and Se atoms are labeled. Large circles are Cs atoms. (b) 5Polyhedral and
(0) ball and stick representation of a pseudo-one-dimensional [B1P4Se12] chain viewed
the [0 -2 1] direction. Dashed line depicts short nonbonding interaction between Se(5)
and Se(l6) at 3. 478(3) A. Light polyhedron depicts [BiSe6] octahedron and dark [P2Se5]
unit with Cs atoms omitted in (b).

 
 

217

Table 8-1. Crystallographic Data and Refinement Details for CssBiP4Se12

 

Formula

Crystal system
Space group

Unit cell dimensions

Z

V, A3

d (calculated), gr cm'3
Crystal dimensions, mm3
Temperature, K

,1, A

a. m’
F (000)
9max (deg)

Total / unique reflections
Rim

No. Parameters
Refinement method

Final R indices [1> 26(1)], Rla/wR2 b
R indices (all data), R l/wR2
Goodness-of-fit on F2

Largest difl‘. peak and hole, e. A'3

l

(385311348612
Monoclinic
Pmc2] (no. 26)
a = 7.5357(2)
b= 13.7783(6)
c =28.0807(8)
4

2915.59(17)
4.431

0.1189 x 0.0006 x 0.0006
293(2)
0.71073
27.392

3304

27.392

18838 / 6306

0.0443

246

Full-matrix least-squares on F2
0.0626 / 0.1151

0.0790 / 0.1207
1.077
2.198/-3.339

 

a R1 = 2 “For chll/ZIFOI- b wR2 = {z [waif-F02) 21/th(F02)21} ”2

218

Table 8-2. Atomic Coordinates (X104) and Equivalent Isotropic Parameters (X103 A2) for

CssBiP4Se12 at 293(2) K. ‘1

 

 

Atom x y z U(eq)
Cs(1) 5000 1495(2) 1862(1) 10(1)
Cs(2) 5000 1511(2) 85(1) 17(1)
Cs(3) 5000 1680(2) 6808(1) 22(1)
Cs(4) 5000 5132(2) 3542(1) 20(1)
Cs(5) 5000 5112(2) 1653(1) 21(1)
Cs(6) 0 1356(1) 5139(1) 9(1)

Cs(7) 0 1675(2) 3247(1) 20(1)
Cs(8A) 0 1680(2) 6780(1) 21 (1)
Cs(8B) -1080(60) 1710(2) 6823(1) 21(1)
Cs(9) 0 4953(2) 209(1) 71(1)
Cs(IO) 0 7879(2) 3559(1) 22(1)
Bi(l) 5000 1681(1) 4219(1) 6(1)

Bi(2) 0 3406(1) 1777(1) 10(1)
P(l) 5000 50(7) 3042(1) 15(2)
P(2) 5000 594(6) 8775(1) 12(2)
P(3) 5000 2792(6) 5492(1) 7(1)

P(4) 5000 3750(6) 4835(1) 9(2)

P(5) 0 871(6) 1 187(1) 8(2)

P(6) 0 2200(6) 737(1) 7(2)

P(7) 0 4521(6) 2829(1) 18(2)
P(8) 0 5922(7) 2417(1) 16(2)
Se(l) 2540(3) 382(2) 7718(1) 19(1)
Se(2) 5000 1658(3) 3155(1) 12(1)
Se(3) 5000 7873(3) 3753(1) 21(1)
Se(4) 2597(3) 24(2) 4122(1) 15(1)
Se(5) 2594(3) 3137(2) 5873(1) 12(1)
Se(6) 5000 1308(2) 5210(1) 9(1)

Se(7) 5000 4745(2) 36(2) 21(1)

 

219

Table 8-2. (Cont’d) Atomic Coordinates (X104) and Equivalent Isotropic Parameters
(><103 A2) for CssBiP48e12 at 293(2) K.“

 

Se(8) 2633(3) 3322(2) 4423(1) 12(1)
Se(9) 2406(3) 83(2) 1005(1) 11(1)
Se(lO) 0 1339(2) 1933(1) 10(1)
Se(] 1) 0 1846(3) 3(1) 22(1)
Se(12) 2378(3) 3025(2) 962(1) 13(1)
Se(13) 0 4821(4) 3564(1) 29(1)
Se(14) 2390(3) 3734(2) 2576(1) 20(1)
Se(15) 2467(3) 6631(2) 2605(1) 17(1)
Se(l6) 0 5509(2) 1653(1) 11(1)

 

“ The occupancy of Cs(8A) and Cs(8B) is 95 and 5 %. All other sites are fully occupied.
b U(eq) is defined as one-third of the trace of the orthogonalized U i]- tensor.

Table 8-3. Selected Bond Distances (A) for CssBiP4Se12 at 298(2) K.

 

Bi(1)— Se(2) 2.986(3) Bi(2) — Se(lO) 2.882(4)
Bi(1)— Se(4) 2.927(3) x2 Bi(2) — Se(12) 2.953(3) x2
131(1) — Se(6) 2.832(3) Bi(2) — Se(14) 2.914(3) x2
Bi(1)— Se(8) 2.937(3) x2 Bi(2) — Se(l6) 2.917(3)
P(1)— Se(l) 2.149(5) x3 P(5) — Se(9) 2.175(5) x2
P(1)- Se(2) 2.238(10) P(5) — Se(10) 2.192(8) x2
P(2) — Se(3) 2.114(9) x2 P(6) — Se(] 1) 2.118(8)
P(2) — Se(4) 2.226(5) x3 P(6) — Se(12) 2.214(5) x2
P(3) — Se(5) 2.159(5) x2 P(7) — Se(13) 2.105(10)
P(3) — Se(6) 2.193(8) P(7) — Se(l4) 2.219(7) x2
P(4) — Se(7) 2.148(8) x2 P(8) — Se(lS) 2.165(5) x2
P(4) — Se(8) 2.206(5) x2 P(8) — Se(l6) 2.220(9)
P(1) —P(2) 2.241(12) x2 P(S) — P(6) 2.224(11)
P(3) - P(4) 2.268(12) P(7) - P(8) 2.252(13)

 

220

Table 8-4. Selected Angles (°) for CssBiP4Se12 at 298(2) K.

 

Se(2)-Bi(l)-Se(4) 8421(7) P(l)#7-P(2)-Se(3)#5 111.6(4)
Se(2)-Bi(1)-Se(6) 168.91(10) Se(10)-Bi(2)-Se(12) 8667(7)
Se(2)-Bi(1)-Se(8) 101.78(7) Se(10)-Bi(2)-Se(14) 9205(8)
Se(4)-Bi(1)-Se(4)#l 76.45(10) Se(10)-Bi(2)-Se(16) 178.10(11)
Se(4)-Bi(1)-Se(6) 87.09(7) Se(12)-Bi(2)-Se(12)#2 74.74(10)
Se(4)-Bi(l)-Se(8) 104.06(6) Se(12)-Bi(2)-Se(14) 104.44(6)
Se(4)#1-Bi(1)-Se(8) l74.01(7) Se(12)-Bi(2)-Se(14)#2 178.51(9)
Se(6)-Bi(l)-Se(8) 8698(7) Se(12)-Bi(2)-Se(16) 9484(8)
Se(8)-Bi(l)-Se(8) 7479(9) Se(l4)—Bi(2)-Se(14)#2 76.36(11)
Se(1)#3-P(1)-Se(1)#4 119.2(4) Se(14)-Bi(2)-Se(16) 8645(8)
Se(1)#3-P(1)-Se(2) 109.5(3) P(1)#7-P(2)-Se(4)#7 104.5(3)
Se(3)#5-P(2)-Se(4)#6 113.3(3) P(2)#3-P(1)-Se(2) 105.2(4)
Se(4)#6-P(2)-Se(4)#7 108.9(4) P(3)-P(4)-Se(7)#5 110.4(4)
Se(5)-P(3)-Se(5)#1 114.3(4) P(3)-P(4)-Se(8) 105.7(3)
Se(5)-P(3)-P(6) 112.6(2) P(4)-P(3)-Se(5) 106.0(3)
Se(7)#5-P(4)-Se(8)#l 113.3(3) P(4)-P(3)-Se(6) 104.5(4)
Se(8)-P(4)-Se(8)#1 107.9(3) P(5)-P(6)-Se(11) 111.3(4)
Se(9)-P(5)-Se(9)#2 113.0(4) P(5)—P(6)-Se(12) 105.1(3)
Se(9)-P(5)-Se(10) 111.8(3) P(6)-P(5)-Se(9) 106.1(3)
Se(11)-P(6)-Se(12) 113.3(2) P(6)-P(5)-Se(10) 107.5(4)
Se(12)-P(6)-Se(12)#2 108.1(4) P(7)-P(8)-Se(15) 105.2(3)
Se(13)-P(7)-Se(14) 114.2(3) P(7)-P(8)-Se(16) 106.1(4)
Se(14)-P(7)-Se(14)#2 108.5(5) P(8)-P(7)-Se(13) 109.6(5)
Se(15)-P(8)-Se(15)#2 118.3(4) P(8)-P(7)-Se(14) 104.7(3)
Se(15)-P(8)-Se(16) 110.6(3)

 

Symmetry transformation used to generate equivalent atoms: (#1)-x+1, y, z; (#2)-x, y, z;
(#3) -x+1, -y, z-1/2; (#4) x, -y, z-1/2; (#5) —x+1, -y+1, z+1/2; (#6) x, -y, z+1/2; (#7) —x+l,
-y, z+1/2.

221

4. Results and Discussion

Synthesis and Crystal Structure. The new compound CssBiP48e12 adopts the
polar space group Pmc21. It features the discrete molecular [Bi(P2Se6)2]5' anion, Figure 8-
1, which is isolated by Cs+ cation, Figure 8-2a. The molecule is isostructural to
[In(P2Se6)2]5'22 and [P(P2Se(;)2]5'23 anions. There are two crystallographically unique Bi
atoms and each octahedrally coordinating Bi3+ center is capped by two chelating
tridentate [P2Se6]4' units from opposite directions. Although the respective [Bi(P2Se6)2]5'
molecules are acentric due to the distorted [BiSe6] octahedra, the polar structure arises
from the noncentrosymmetric packing of Cs+ ions. The absence of mirror plane
perpendicular to the 21 screw axis creates the polar structural arrangement of the
compound. The Bi-Se distances are normal ranging fiom 2.832(3) to 2.986(3) A. The P-
Se distances range from 2.114(9) to 2.241(12) A. The P-P distances are typical at
2.224(11) to 2.252(11) A. Noteworthy is the unusually short intermolecular nonbonding
interaction of Se(5)-"Se(l6) at 3.478 A, which is much shorter than the van der Waals
radii sum of 3.8 A.24 This weak interaction is important in enabling the compound to
organize in an infinite pseudo-one-dimensional structure along a-axis, Figure 8-2b. The
ionic interaction of Cs+ with [Bi(P2Se6)2]5' is, however, also prominent as we will argue
later. Low-dimensional compounds such as NbSe3, APSC6 (A=K, Rb, Cs),25 K2P2Se6,26
and the molecular Rb4P68e1227 display similar Se---Se interactions. The dipole associated
with the noncentrosymmetric structure is parallel to the c-axis which is perpendicular to
the nanowire direction. Other polar nanowire systems include the II-IV semiconductors

such as ZnO, CdS, CdSe, and the III-V semiconductors GaAs, InP, etc. In these systems,

222

the dipole is aligned along the nanowire axis. The CssBiP4Se12 is the first example where
the dipole is perpendicular to the long axis.

Comparing polychalcophosphate flux reaction conditions that favor
CssBi(P2Se6)2 vis-a-vis CSgBi4(ste6)528 we can gain deeper insight into the evolution of
structure. The latter was synthesized in a more basic alkali polyselenophosphate flux of
CS2SezP2Se5:Se=1:1.5:2 at 460 °C, giving a remarkably complicated layered structure of

2/00[Bi4(P2Se5)3']. In contrast, the simpler molecular complex [Bi(P2Se6)2]5' of the title

compound was obtained by less basic flux of CS2SezP2Se5:Se=1:1.6:3.8 at 850 °C. In
general, less basic conditions or lower flux temperatures tend to generate longer or
extended fragments.” Hence, we assume that the higher temperature plays a more
dominant role in stabilizing CssBi(P2Se6)2 than the flux basicity. Thus, we can expect for
a given flux composition different products to form at high reaction temperature of 800-
1000 °C fi'om those formed at intermediate reaction temperatures (300-600 °C).
Nanowires. The fibrous nature of CssBiP4Se12 is apparent by visual inspection
and in scanning electron microscopic (SEM) images, Figure 8-3a. As-prepared
microfibers are typically ~mm in length and submicron in thickness. The microfibers
readily split to thinner fibers by physical contact. Transmission electron microscopic
(TEM) images of single microfibers confirmed that they consist of individual nanowires
with uniform thickness and alignment with the interval at 2.9 nm, close to the dimension
of the crystallographic c-axis, Figure 8-3b. This creates the extreme difliculty in finding
single crystals suitable for X-ray diffraction studies. Most crystals we examined showed
extensive diffuse scattering normal to the a*-axis consistent with the nanosize

dimensions of the coherence length along the b- and c- axes (i.e. perpendicular to the

223

fiber).

 

 

 

Figure 8-3. (a) Representative SEM image of CssBiP4Se12 microfibers demonstrating
their flexible texture. (b) TEM image of a bundle of nanowires showing uniform
alignment of individual nanowires. The interval of each nanowire was measured at 2.9
nm, close to the crystallographic c-axis of 28.0807(8) A. Inset: selected area electron
diffraction pattern of CssBiP48e12 nanowires. (c) TEM image, (inset: selected area
electron diffraction pattern of nanowires, (d) EDS analysis, and (e) HRTEM image of
individual CssBiP4Se12 nanowire dispersed in EtOH. (f) Tangled texture of CssBiP48e12
nanowire depicting its striking flexibility. The scale bar corresponds to (a) 500 um, (b) 10
nm, (c) 20 nm, (e) 5 11m, and (f) 20 nm, respectively.

224

 

'559'

 

 

 

 

 

'8q00 200 300 400 500 600 700 800

Temperature, °C

 

 

  
  
 

 

 

 

 

(b) 5 1 t r . 1 . .

s“ 4- -

c6

‘6: 3' ~

3L 2. -

3‘

'5 AfierDTA

C 1 1

B .

5 Pristine
10 20 3O 4O 50 60 70 80

2theta,deg

Figure 8-4. (a) Differential thermal analysis (DTA) and (b) X-ray powder diffraction
patterns of pristine material and sample obtained afier DTA.

225

The CssBiP48e12 microfibers can be dispersed in EtOH under N2 atmosphere and
ultrasonificated for ~10 min to form a colloidal suspension of the corresponding
individual nanowires, Figure 8-3c. Selected area electron diffraction pattern showed
crystalline nature of the nanowires, inset in Figure 30. Energy dispersive spectroscopic
(EDS) analysis on the sample examined by TEM showed all Cs, Bi, P, and Se to be
present in the expected ratio, Figure 8-3d. The distribution in nanowire diameter is xx-yy
nm. Typical long CssBiP4Se12 nanowires show a striking flexibility and tangled texture,
Figure 8-3e-f. This is extraordinary considering the single crystalline nature of the fibers.
The flexible character likely arises from the weak intrachain and interchain interactions
created by long nonbonding Sen-Se contacts between [Bi(P2Se6)2]5' molecules which
allows them to slide past one another and bend. These weak non-bonding interactions
play a key role in determining the emergent behavior in this material which is the long
flexible fiber morphology from the packing of simple coordinating complexes of
[131053602153

Differential thermal analysis (DTA) of CssBiP4Se12 at a rate of 10 °C min'l
showed melting at 590 0C and crystallization upon cooling at 559 °C, Figure 8-4a. The
powder X-ray diffraction patterns before and after melting/recrystallization were identical
indicating congruent melting, Figure 8-4b. The solid-state optical absorption spectrum
revealed a sharp absorption edge and a band gap of 1.85 eV, which is in good agreement
with its dark red color, Figure 5. The nearly vertical rise of the absorption coefficient
suggests a direct bang-gap material. In comparison, CssBi4(P2Se6)5 compound showed a

sharp optical gap at 1.44 eV.

226

1.5 . . . i

 

—‘L
I

Absorption, a/S
_o
01

 

 

 

 

J Eg=1.85eV
O r r r r I

0 1 2 3 4 5
Energy, eV

Figure 8-5. Solid-state optical absorption spectrum of CssBiP48e12.

227

Electronic Structure Calculations. To better understand the experimental
findings and in particular to explore the origin and nature of the band gap result, the
electronic structure was calculated using the FLAPW method with the screened-exchange
local density approximation (sX-LDA) as well as the Hedin-Lundqvist30 (LDA) forms of
the exchange-correlation potential. The sX-LDA method has been successfully applied to
a wide range of semiconductors, and shows great improvements of the excited electronic
states in terms of correctly determining the band gaps and band dispersion.3'1 While the
LDA calculations yield a band gap of 1.48 eV without spin orbital coupling (SOC),
Figure 8-6, and 1.15 eV with SOC included, Figure 8-7, the electrOnic structure
calculated by the sX-LDA method with SOC determines the band gap to be 2.0 eV, which
is close to the experimental band gap, 1.85 eV, Figure 8-8. The valence band maximum
(V BM) and the conduction band minimum (CBM) occur along the T-Y direction and give
rise to an indirect band gap, Figure 8-9. However, the almost plateau-like band dispersion
(S 5 meV) along the T-Y direction causes a large joint density of states at 2.0 eV, which is
responsible for the very sharp absorption edge observed in the experiment, Figure 8-5.
The narrow band dispersions (e. g. S 0.05 eV for the valence top band along the T-Z-T-Y)
indicate the molecular nature of the compound. The relatively strong dispersion (~0.24
eV for the valence top band) along the T-X direction arises from the Se---Se
intermolecular interactions along the nanowire direction (a-axis in Figure 8-2a) compared

to other directions.

228

2.0

1.0

Energy (eV)

.1.
c

 

-2.0
FZ TY l" X S R U

Figure 8-6. The calculated band structures using LDA (Eg = 1.45 eV).

229

2.0

1.0

Energy (eV)

-l.0

 

-2.0
FZ TY l" X S R U

Figure 8-7. The calculated band structures using LDA with spin-orbit coupling (SOC)
(13g =1.15 eV).

230

2.0

1.0

Energy (eV)

-l.0

 

Z T Y 1"

Figure 8-8. The calculated band structures sX-LDA with SOC (Eg = 2.0 eV), shown for
only a small part of the Brillouin zone, namely Z-T-Y-l').

231

 

 

 

 

 

 

 

 

 

 

— Bi (P) E F
—— P (P)
______ Se (4. P)
-— Se (7. p)

‘6’

E 2 ‘

U)

c...,

O

.Q‘

a ,1

a 1 ' 1H1

1'1 11
0
—6 —4 —2 0 2
Energy (eV)

Figure 8-9. The projected density of states for p-orbitals of individual elements (Bi, P,
Se) calculated with sX-LDA and SOC.

232

The angular-momentum-resolved density of states (DOS) for the individual
atoms reveals that the p-p mixing has a strong effect on the electronic structure and on the
energy level ordering. The strong covalent interactions between mainly P and Se p-
orbitals forms the lower energy levels [at -6 eV ~ -4 eV] in the valence band. The
energy states derived from the relatively weak hybridization between the Bi and Se p-
orbitals are located between -4 eV and -2 eV. The dominating Se p-character fiom -2 eV
to the VBM can be attributed to the lone pair states of Se and to the ionic bonding
between Cs and Se. In this energy range, the p orbital contribution of other elements are
negligibly small as shown in Figure 8-9. The conduction band minimum shows strong Bi
p-character, which is highly affected, by the spin-orbit coupling (SOC), Figures 8-6 and
8-7. Therefore, the band gap excitation observed in the electronic spectrum of the
compound shown in the Figure 5 is due to direct transitions from Se p to Bi p states.

It is unusual for discrete molecules to form nanofibers. Careful inspection of the
crystal structure, Figures 8-2a and 8-10, shows that Cs atoms along the a-axis have more
anionic Se atoms available to interact with, which gives stronger ionic interactions along
the a-axis than those in the other directions. For instance, Figure 8-10d shows the
neighbor atoms of three different molecular units (1, II, and III) around the Cs atom
within 4.7 A. Due to the larger number of neighbors in Figure 8-10d the molecular units
(1) and (11) along the a-axis have much stronger ionic interactions with the Cs than the
molecular unit (III) in the b-c plane. As a result, ionic interactions can predominantly
propagate along the a-axis, which in turn would give better chance of crystal seeds

growing along the a-axis, e.g. apparent nanofiber growing direction.

233

(a) (b)

  

Figure 8-10. Three different models, (a), (b), and (c), taken from the original crystal
structure, for the binding energy calculations (see text). The coordinates of two molecular
units are taken from the original crystal structure (Figure 1a) in the b-c plane ((a), (b))
and along the a-axis (c). The neighbor atoms of three molecular [BiP48e12]5' units (1, II,
and III) around the Cs atom within 4.7 A are indicated in blue circles (d).

234

To better understand how the discrete molecules give rise to the nanowire nature
of CssBiP4Se12, the binding energies of the molecular [BiP4Selz]5' units in each
crystallographic direction were investigated with the LDA and the gradient-corrected
Perdew, Burke, and Emzerhof32 (GGA) forms of the exchange-correlation potential. We
prepared three different models in which the coordinates of two molecular units were
taken from the original crystal structure (Figure 28) in such a way that they have inter-
molecular interactions for different orientations in the b-c plane, Figures 7a and 7b, and
along the a-axis, Figure 7c. For these models, the binding energies were estimated by
examining the total energy behavior in terms of the distance between two molecular units.

The LDA calculations yielded binding energies of 1.90 eV, 2.07 eV, and 4.13 eV
for the 1st, 2nd, and 3rd models in Figure 8-10, respectively while the GGA calculations
showed significantly smaller binding energies, namely 0.84 eV, 0.87 eV, and 3.35 eV for
the lst, 2nd, and 3rd models, respectively. The difference in the binding energies between
the LDA and the GGA comes mainly from the van der Waals interactions. While the
strongest interactions are the ionic interactions between the molecular units in the
compound, there are also van der Waals closed-shell interactions, for which the LDA
shows in general largely overestimated binding energies of the bonds. The GGA, by
contrast, may give over- or underestimated or even no binding energies depending on the
different GGA forms (i.e. BLYP, PBE) employed.33 Even with the deficiency of these
exchange-correlation approximations for the van der Waals interactions, both the LDA
and GGA strongly showed significantly stronger bonds in the 3rd model, which is the
interaction along the fiber direction. This is, therefore, the fastest growing direction

which is consistent with the nanofiber nature found experimentally. The electronic

235

structure calculations of CssBiP4Se12 clearly demonstrate the existence of dominating
interactions along specific directions. This is indicated by the significant differences in
binding energies along different crystallographic directions and explains how the
ostensible molecular compound grows naturally to be nanofibers. The result suggests the
possibility of theoretical modeling and prediction of what structural species can
intrinsically produce nanofibers without complex preparation processes.

Spectroscopy and Nonlinear Optical Response. The Raman spectrum of
Cs5BiP4$e12 is very active with shifts at 118, 143, 172, 221, 300, 433 and 467 cm],
Figure 8-11. The peak at 221 cm'1 is unambiguously assigned to the locally Alg
symmetric stretching mode of the PSe3 unit (half of the [P2Se6]4' ligand).21 The peaks
below 200 cm'1 are possibly related with Bi-Se stretching excitations. The other shifts can

be attributed to PSe3 stretching modes.

 

16 . 221

.3

N
l
A

 

 

 

>.
.2 118
g 8 - 1431 -
E.

4 - 433 172 d

467/
300
0» .

 

 

 

600 500 400 300 200 100
Raman Shift, cm‘1

Figure 8-11. Raman spectrum of Cs5BiP4Se12.

236

The far-IR spectrum showed peaks at 150, 181, 221, 300, 415, 440 and 502 cm",
similar to those of KBiP2Se6 which include [P2Se6]4' anion, Figure 8-12.34 CssBiP4Se12 is
optically transparent in the far-IR through the mid-IR region to below the band edge in
the visible region from 18.7 am to 0.67 am, corresponding to the region above the
[P2Se6] stretching absorptions, Figure 8-13. Wide optical transparency is a critical
property for application of nonlinear optical (NLO) materials and optical fibers in the
infrared region. NLO materials with high second harmonic generation (SHG) response is
highly desirable as for example in telecommunications (near IR at l.3-1.5 ,um), as
pollutant detection in the molecular fingerprint region (mid IR),35 and medical surgery

(mid IR at 10.6 ,um) applications.36

 

 

10 I I l I
A8. .
:5
a",
m6- -
0
C
(U
SE
5 4-

U)
C
E
l- 2_

  

 

 

  
  

181
502 440415 300 150

(800 500 400 300 200 100
Wavenumber, cm'1

 

 

Figure 8-12. F ar-IR spectrum of CssBiP4Se12.

237

 

100 ................fi-...

 

 

 

 

 

 

 

 

90 - .
T‘T
; 80 ' / ‘
j; 70 r- Band-gap -
E
g 60 n ‘ )
E 50 Transparent '
cu
; 40 -
3O ' ‘— Far-IR absorption
20 . I 1 n . . I . n . I . n . 1 I . . . n
20 15 10 5 0

Wavelength, ,um

Figure 8-13. Far IR/mid IR/vis absorption spectra of CssBiP48e12. A wide transparency
range of CssBiP48e12 between 18.7 pm at far-IR region and 0.67 pm at visible region is

shown.

 

oss'Bi‘P,‘se,,'
-AgGaSe2 1

  

 

 

Relative SHG intensities (a. u.)

 

 

996 1000 1004 1008 1012
Converted Wavelength, nm

Figure 8-14. SHG response of Cs5BiP4Se12 relative to AgGaSe2 at 1004 nm.

238

Because of the polar crystal structure of CssBiPaSen, we investigated its NLO
properties. Using a modified Kurtz powder method,37 we measured SHG response for
polycrystalline samples of 45-63 pm size with 1.2-2 ,um fundamental idler radiation from
a tunable laser. The second harmonic generation (SHG) response could be observed only
above 0.79 am because of the absorption beyond the band edge, and its intensity
increased with wavelength up to 1 pm. The SHG intensities obtained were compared
with those of AgGaSe2, which is a representative NLO material for IR applications.38 All
samples were similarly prepared and the same particle size range and identical laser
settings were used. The SHG intensity of CssBiP4Se12 is ~2 times larger than that of
AgGaSe2 at 2 um idler beam showing good performance in converting short wave IR,
Figure 8-14. The SHG intensity reached maximum at 32-45 ,um particle size and the
decreased with larger sizes. Thus, Cs5BiP48el2 is type-I non-phase-matchable and the
corresponding particle size of 38.5 :1: 6.5 pm could be regarded as the coherence length of
CssBiPaSe12. Despite this, such materials can be useful through ‘random’ quasi-phase-
matching.39 It should be noted that a modified Kurtz powder method may not be suitable
to determine SHG intensity for samples with microwire morphology. The Kurtz powder
method deals with the relationship of SHG intensity and particle size. However, typical
diameter of as prepared Cs5BiP48e. 2 is less than 1 pm so that sieves to differentiate the
size distribution do not work well for this case. For this reason the SHG measured could
be underestimated. As a result, to better understand the SHG of CssBiPaseiz, the use of

larger single crystal samples is required.

239

Concluding remarks

The new compound CssBiP4Se12 forms naturally long flexible fibers. The
compound reflects the rich structural chemistry of Bi with [Pny]"' chalcophosphate

ligands. The packing mode of the [Bi(P28e6)2]5' molecules and the weak Se---Se
interactions between molecules is responsible for the self-formed long flexible nanowires
which organize into fibers. Therefore, because of this natural tendency, to obtain
Cs5BiP4$e12 long nanowires in high yield does not require complex chemical processes.
This is a rare example of a simple phase material giving long crystalline fibers. The fiber
nanowire morphology emerges from the specific intermolecular secondary interactions
which could not be predicted a priori. This discovery implies that innate one- or two-
dimensional nanostructures may be rationally approached under consideration of crystal
structures. Cs5BiP48e12 is widely transparent in the near-/mid IR ranging from 18.8 to
0.67 pm, and it exhibits a relatively strong SHG response, which is ~2 times larger than
that of AgGaSe2. The compound is nearly direct band gap semiconductor with a very
sharp absorption edge and melts congruently. Inherent formation of nanowires coupled
with strong near IR SHG response and optical transparency enables this material to be
promising for sensing noncentrosymmetric biomolecules by utilizing IR NLO signal,40

. . . . 14
nanowrre mlcroscopy,4| and lntegrated photonlc networks.

240

References

(1) Tenne, R.; Margulis, L.; Genut, M.; Hodes, G. Nature 1992, 360, 444-446.

(2) Feldman, Y.; Wasserman, B.; Srolovitz, D. J .; Tenne, R. Science 1995, 267, 222-
225.

(3) Malliakas, C. D.; Kanatzidis, M. G. J. Am. Chem. Soc. 2006, 128, 6538-6539.

(4) Davidson, P.; Gabriel, J. C.; Levelut, A. M.; Batail, P. Europhys. Lett. 1993, 21 ,
317-322.

(5) Chen, J .; Li, S. L.; Tao, Z. L.; Gao, F. Chem. Commun. 2003, 980-981.
(6) Nath, M.; Rao, C. N. R. J. Am. Chem. Soc. 2001, 123, 4841-4842.

(7) Brorson, M.; Hansen, T. W.; Jacobson, C. J. H. J. Am. Chem. Soc. 2002, 124,
11582-11583.

(8) Ye, C. H.; Meng, G. W.; Jiang, Z.; Wang, Y. H.; Wang, G. 2.; Zhang, L. D. J.
Am. Chem. Soc. 2002, 124, 15180-15181.

(9) Chopra, N. G.; Luyken, R. J.; Cherrey, K.; Crespi, V. H.; Cohen, M. L.; Louie, S.
G.; Zettl, A. Science 1995, 269, 966-967.

(10) Spahr, M. E.; Bitterli, P.; Nesper, R.; Muller, M.; Krumeich, F.; Nissen, H. U.
Angew. Chem, Int. Ed. 1998, 3 7, 1263-1265.

(11) Therese, H. A.; Rocker, F.; Reiber, A.; Li, J. X.; Stepputat, M.; Glasser, G.;
Kolb, U.; Tremel, W. Angew. Chem, Int. Ed. 2005, 44, 262-265.

(12) Cava, R. J. Science 1990, 24 7, 656-662; Sleight, A. W. Science 1988, 242, 1519-
1527.

(13) J in, S.; Tiefel, T. H.; Mcconnack, M.; Fastnacht, R. A.; Rarnesh, R.; Chen, L. H.
Science 1994, 264, 413-415; Moritomo, Y.; Asamitsu, A.; Kuwahara, H.; Tokura, Y.
Nature 1996, 380, 141-144.

(14) Scott, J. F.; Dearaujo, C. A. P. Science 1989, 246, 1400-1405; Dearaujo, C. A.
P.; Cuchiaro, J. D.; Mcmillan, L. D.; Scott, M. C.; Scott, J. F. Nature 1995, 3 74, 627-629;
Hemberger, J .; Lunkenheimer, P.; Fichtl, R.; von Nidda, H. A. K.; Tsurkan, V.; Loidl, A.
Nature 2005, 434, 364-367.

(15) Gabriel, J. C. P.; Camerel, F.; Lemaire, B. J .; Desvaux, H.; Davidson, P.; Batail,

P. Nature 2001, 413, 504-508; Miyarnoto, N.; Nakato, T. Adv. Mater. 2002, 14, 1267-
1270; Michot, L. J.; Bihannic, I.; Maddi, S.; Funari, S. S.; Baravian, C.; Levitz, P.;

241

Davidson, P. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 16101-16104; Miyarnoto, N.;
Yamada, Y.; Koizumi, S.; Nakato, T. Angew. Chem, Int. Ed. 2007, 46, 4123-4127.

(16) Bordui, P. F .; Fejer, M. M. Annu. Rev. Mater. Sci. 1993, 23, 321-379.

(1 7) Dom, R.; Baums, D.; Kersten, P.; Regener, R. Adv. Mater. 1992, 4, 464-473; Ma,
H.; Jen, A. K. Y.; Dalton, L. R. Adv. Mater. 2002, 14, 1339-1365; Yamada, H.; Shirane,
M.; Chu, T.; Yokoyama, H.; Ishida, S.; Arakawa, Y. Jpn. J. Appl. Phys, Part 1 2005, 44,
6541-6545.

(18) Knight, J. C. Nature 2003, 424, 847-851.

(19) Wimmer, B.; Krakauer, H.; Weinert, M.; Freeman, A. J. Phys. Rev. B 1981, 24,
864-875.

(20) Wendlandt, W. W.; Hecht, H. G. Reflectance spectroscoPJ’; Interscience
Publishers: New York,, 1966; Kortitm, G. Reflectance spectroscopy. Principles, methods,
applications; Springer: Berlin, Heidelberg, New York,, 1969; Tandon, S. P.; Gupta, J. P.
Phys Status Solidi 1970, 38, 363-367; Chung, D. Y.; Choi, K. S.; lordanidis, L.;
Schindler, J. L.; Brazis, P. W.; Kannewurf, C. R.; Chen, B. X.; Hu, S. Q.; Uher, C.;
Kanatzidis, M. G. Chem. Mater. 1997, 9, 3060-3071.

(21) MacDonald, A. 11.; Pickett, W. B.; Koelling, D. D. Journal of Physics C-Solid
State Physics 1980, 13, 2675-2683.

(22) Chondroudis, K.; Chakrabarty, D.; Axtell, E. A.; Kanatzidis, M. G. Z. Anorg.
Allg. Chem. 1998, 624, 975-979.

(23) Chung, 1.; Jang, J. I.; Gave, M. A.; Weliky, D. P.; Kanatzidis, M. G. Chem.
Commun. 2007, 4998-5000.

(24) Bondi, A. J. Phys. Chem. 1964, 68, 441-451.

(25) Chung, 1.; Do, J.; Canlas, C. G.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem.
2004, 43, 2762-2764.

(26) Chung, 1.; Malliakas, C. D.; Jang, J. I.; Canlas, C. G.; Weliky, D. P.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2007, 129, 14996-15006.

(27) Chung, 1.; Karst, A. L.; Weliky, D. P.; Kanatzidis, M. G. Inorg. Chem. 2006, 45,
2785-2787.

(28) McCarthy, T. J .; Hogan, T.; Kannewurf, C. R.; Kanatzidis, M. G. Chem. Mater.
1994, 6, 1072-1079.

(29) Kanatzidis, M. G.; Sutorik, A. C. Prog. Inorg. Chem. 1995, 43, 151-265.

242

(30) Hedin, L.; Lundqvist, B. I. J. Phys. C 1971, 4, 2064-2083.

(31) Asahi, R.; Mannstadt, W.; Freeman, A. J. Phys. Rev. B 1999, 59, 7486-7492;
Bylander, D. M.; Kleinman, L. Phys. Rev. B 1990, 41 , 7868-7871.

(32) Perdew, J. P.; Burke, K.; Emzerhof, M. Phys. Rev. Lett. 1996, 7 7, 3865-3868.

(33) Montanari, B.; Ballone, P.; Jones, R. O. J. Chem. Phys. 1998, 108, 6947-6951;
Wu, Q.; Yang, W. T. J. Chem. Phys. 2002, 116, 515-524.

(34) Breshears, J. D.; Kanatzidis, M. G. J. Am. Chem. Soc. 2000, 122, 7839-7840.

(35) Tittel, F. K.; Richter, D.; Fried, A. Solid-State Mid-Infrared Laser Sources 2003,
89, 445-510.

(36) Jean, B.; Bende, T. Solid-State Mid-Infrared Laser Sources 2003, 89, 511-544.

(37) Dougherty, J. P.; Kurtz, S. K. J. Appl. Crystallogr. 1976, 9, 145-158; Kurtz, S.
K.; Perry, T. T. J. Appl. Phys. 1968, 39, 3798-3813.

(38) Nikogosyan, D. N. Nonlinear optical crystals .' a complete survey; Springer-
Science: New York, 2005.

(39) Baudrier-Raybaut, M.; Haidar, R.; Kupecek, P.; Lemasson, P.; Rosencher, E.
Nature 2004, 432, 374-376.

(40) Zhang, D. B.; Qi, L. M.; Ma, J. M.; Cheng, H. M. Adv. Mater. 2002, 14, 1499-
1502; Zipfel, W. R.; Williams, R. M.; Christie, R.; Nikitin, A. Y.; Hyman, B. T.; Webb,
W. W. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 7075-7080.

(41) Johnson, J. C.; Yan, H. Q.; Schaller, R. D.; Petersen, P. B.; Yang, P. D.;

Saykally, R. J. Nano Lett. 2002, 2, 279-283; Kim, J .; Song, K. B. Micron 2007, 38, 409-
426.

243

Chapter 9

Outlooks

Chalcophosphate class is a worthy target for exploratory synthesis because,
coupled with the remarkable structural and compositional diversity, they showed
technologically important magnetic, electrical, and optical properties. The
polychalcophosphate flux method is a proven synthetic tool during last decade.
Chalcophosphate anionic building unit of [Pny]"' was known to combine nearly all
metals to form metal chalcogenide compounds. By employing this approach, researchers
produced many new ternary M/P/Q or quaternary A/M/P/Q (A=Li-Cs, M=metal, Q=S,
Se) compounds.

Despite expanded registry of chalcophosphate compounds, the number of anionic
ligand species [Pny]"' that had been structurally characterized was just a few and they
are mostly simple molecular anions such as [PQ4]3' and [P2Q6]4‘. Published works on
chalcophosphate compounds were mostly regarding on synthesis of new compounds with
structural characterization, but studies on their potential technological applications were
left behind. Accordingly, this work started with questions of how many chalcophosphate
ligands exist and which flux condition can stabilize the specific ligand, hoping to find
rational synthetic conditions for desired structural species. As a result, our investigations
of alkali chalcophosphate compounds provided new insights into the relationship between
the structure and the flux condition (A:P:Se ratio in the composition). New one-
dimensional and molecular complex chalcophosphate compounds were found based upon

those understandings during this dissertation work.

244

1. Chalcophosphate compounds as nonlinear optical materials.

With less basic flux condition (A:Se S 2), one-dimensional chalc0phosphate
compounds APSe6, and A2P2Se6 (A = K, Rb) were obtained. Both compounds crystallized
in the noncentrosymmetric polar space group, and showed remarkably strong second
harmonic generation. (Se)x chains are condensed with [PSe4]3' or [P2Se6]4' units to form a
highly anisotropic one-dimensional structure. Note that chalcogens (S, Se, Te) are more
polarizable than oxygen atom, giving larger second-order nonlinearity. Thus, exploratory
synthesis of structurally anisotropic, one- or two-dimensional compounds should be
interesting for such purpose.

2. Chalcophosphate compounds as phase-change materials and rich source
for glass compounds.

Prominent observation during this dissertation work was great tendency to form
glassy phase of chalcophosphate compounds. Examples are APSe6 (A = K, Rb, Cs),
A2P2866 (A = K, Rb), Cs4P6Se12, Cs5P58e12, and KanTea. Glassy phases of those
compounds readily recover their crystal structure upon heating near crystallization points.
Note that those materials are stoichiometric and possess mixed ionic/covalent bondings,
in contrast to the vast majority of chalcogenide glasses having nearly continuous
compositions and all-covalent networks. Thus, switching between crystalline and glassy
phases are not complicated by compositional changes. As a result, chalcophosphate
compounds can be rich source for stoichiometric glasses and crystal-glass phase-change
materials. It should be noted that chalcophosphate compounds synthesized in the
intermediate and less basic condition frequently exhibited phase-change behavior.

In this dissertation, in order to probe the relationship between crystalline and

glassy phases of phase-change materials, we performed Raman spectroscopic and pair

245

distribution function (PDF) analysis. However, to explain mechanism of reversible
crystal-glass phase-change, firrther work should be necessary. This study is very
important because we observed a significant second harmonic generation response in
glassy phases of phase-change materials and this observation is technologically very
important.

Note that chalcogenide glasses are interesting for applications for infrared
transparent optical fibers, reversible conductivity switching devices, semiconductors,
photoconductors, and solid electrolytes for battery materials. In this regard, physical and
chemical property characterizations of glassy phases are required. Extensive exploratory
synthesis of chalcophosphate compounds can provide a map of which flux conditions
prefer phase-change behavior, crystalline, or glassy phases.

Finally, studies on KanTea found the dissolution behavior of K4P3Te4 compound
in hydrazine. Since hydrazine solution can dissolve various metals with or without help
of additional Te, this rare behavior can open up the possibility of new chemistry:
synthesis of wide range of metal tellurophosphate compound via a chimie deuce route,
functional organic-inorganic hybrids, and nanocomposite compounds; study of gel-
formation and liquid crystalline behavior. It is also noteworthy that precipitating

hydrazine/ KanTe4 solution with alcohol produced nanosphere.

246

        

YLIB A

T R RIES

4690

 

 

 

 

Mllllelllllllllllll 1111111111111
3 12 9 3 0 3 0 6 2