copyright by gaurav gupta 2008

Copyright

by

Gaurav Gupta

2008

The Dissertation Committee for Gaurav Gupta

certifies that this is the approved version of the following dissertation:

Design of Novel Catalysts by Infusion of Presynthesized

Nanocrystals into Mesoporous Supports

Committee:

_________________________________ Keith P. Johnston, Supervisor _________________________________ Brian A. Korgel _________________________________ Arumugam Manthiram _________________________________ Richard M. Crooks _________________________________ Miguel Jose Yacaman



by

Gaurav Gupta, B.Tech.

Dissertation

Presented to the Faculty of the Graduate School of

the University of Texas at Austin

in Partial Fulfillment

of the Requirements

for the Degree of

Doctor of Philosophy

The University of Texas at Austin

August 2008

Dedicated to my parents, Ramesh Chandra and Archana Gupta

v

Acknowledgements

This thesis is the result of interactions and discussions with several people,

and I would like to take this opportunity to thank them individually. First I would like

to thank my advisor, Dr. Keith Johnston. In addition to the scientific guidance he

provided, he also encouraged exploring new and novel research ideas. I would also

like to thank my dissertation committee for their contribution towards this thesis.

This thesis is written in collaboration with many individuals and I would like

to thank them starting with Dr. Keith Stevenson who provided valuable guidance as

well as use of the lab equipments for the fuel cell work. I would like to thank Mehul

Patel for having intellectually stimulating discussions which significantly enriched

this thesis. Thanks to Dan Slanac, Pavan Kumar, Andrew Heitsch, Cynthia Stowell,

Dr. Aaron Saunders, Dr. Domingo Ferrer, Dr. Xiaoxia Gao, Dr. Parag Shah, Dr.

Steven Swinnea and Jaclyn Wiggins for their invaluable help in the several key

experiments. I also had the pleasure of having insightful discussions with Arindam

Sarkar, Dr. Hiroshi Uchida, Dr. Won Ryoo, and Corinne Atkinson.

Special thanks for the support and friendship of the lunch group involving

Manas Shah, Gaurav Goel, Mehul Patel, Amogh Prabhu, John Keagy, Dale Simpson

and Landry Khounlavong.

Exceptional thanks to my parents and my brother for their support, patience

and understanding throughout these years.

vi



Publication No. ___________

Gaurav Gupta, Ph.D.

The University of Texas at Austin, 2008

Supervisor: Keith P. Johnston

Traditionally, supported metal catalysts have been synthesized by reduction of

precursors directly over the support. In these techniques, it is challenging to control

the metal cluster size, composition and crystal structure. Herein, we have developed a

novel approach to design catalysts with controlled morphologies by infusing

presynthesized nanocrystals into the supports. High surface area mesoporous

materials, including graphitic carbons, have been utilized for obtaining a high degree

of metal dispersion to enhance catalyst stabilities and activities. Gold and iridium

nanocrystals have been infused in mesoporous silica with loadings up to 2 wt % using

supercritical CO2 as an antisolvent in toluene to enhance the van der Waals

interactions between nanocrystals and the silica. The iridium catalysts show high

catalytic activity and do not require high temperature annealing for ligand removal, as

ligands bind weakly to the iridium surface. To further enhance metal loadings to >10

% in the catalysts, short-ranged interactions between the metal nanocrystals and the

support are further strengthened with weakly binding ligands to expose more of the

metal surface to the support. For pre-synthesized FePt nanocrystals, coated with oleic

acid and oleylamine ligands, high loadings >10 wt % in mesoporous silica are

achieved, without using CO2. The strong metal-support interactions favor FePt

vii

adsorption on the support and also enhance stability against sintering at high

temperatures. High resistance to sintering favors formation of the FePt intermetallic

crystal structure with <4 nm size upon thermal annealing at 700 °C. The fundamental

understanding of the metal-support interactions gained from these studies is then

utilized in the design of highly stable Pt and Pt-Cu electrocatalysts with controlled

size, composition and alloy structure supported on graphitized mesoporous carbons

for oxygen reduction. The resistance of the graphitic carbons to oxidation coupled

with strong metal-support interactions mitigate nanoparticle isolation from the

support, nanoparticle coalescence, Pt dissolution and subsequent Ostwald ripening

and thus enhance catalyst stability. The control of the Pt nanocrystal morphology with

high concentrations of highly active (111) surface leads to 25% higher activities than

commercial Pt catalysts. Furthermore, the catalyst activities obtained for Pt-Cu

catalysts are 4-fold higher than Pt catalysts due to strained Pt shell generated from

electrochemical dealloying of copper from the nanoparticle surface.

viii

Table of Contents

List of Tables………………………………………………………………………...xii

List of Figures………......….………………………………………………………xiii

Chapter 1: Introduction ...………………………………………………………… 1 1.1 Nanocrystal Synthesis ……………………………….…………………. 3 1.2 Mesoporous Supports ………………………………………………….. 5 1.3 Infusion of Presynthesized Nanocrystals into Mesoporous Supports ...... 9 1.4 Electrocatalysts for Oxygen Reduction Reaction: Activity and Stability ……………………………………………………………….. 13 1.5 Objectives ……………………………………………………………... 14 1.6 Dissertation Outline………….…………………………………………15 1.7 References………………………………………………………………18

Chapter 2: Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with Supercritical Carbon Dioxide ………………………………….27 2.1 Introduction……………………………………………………………..28 2.2 Experimental……………………………………………………………32 2.2.1 Nanocrystal Synthesis………………………………………...32

2.2.2 Mesoporous Silica Synthesis………………………………….32 2.2.3 Infusion Procedure……………………………………………33 2.3 Results…………………………………………………………………..35 2.3.1 Nanoparticle Characterization………………………………..35 2.3.2 Characterization of Mesoporous Silica…………………….....37 2.3.3 Nanoparticle Dispersibility in Toluene-CO2............................43 2.3.4 Effect of CO2 on Loading of Gold Nanocrystals in Silica……44 2.3.5 Effect of Pore Size and Concentration on the Loading……….46 2.3.6 Kinetic Study of the Infusion Process………………………...46

2.3.7 TEM Micrographs of Infused Silica…………………………..49 2.3.6 Effect of Conventional Liquid Anti-Solvent on Infusion……..49

2.4 Discussion………………………………………………………………51 2.4.1 Equilibrium Nanocrystal Loading……….……………………28 2.4.2 Transport Aspects of Nanocrystal Loading…………………..52 2.5 Conclusions……………………………………………………………..58 2.6 References………………………………………………………………59 Chapter 3: Infusion of Presynthesized Iridium Nanocrystals into Mesoporous Silica for High Catalyst Activity……………………………………..65 3.1 Introduction……………………………………………………………..66

ix

3.2 Experimental……………………………………………………………69 3.2.1 Materials……………………………………………………...69 3.2.2 Iridium Nanocrystal Synthesis………………………………..70 3.2.3 Mesoporous Silica Synthesis ………………………………...70 3.2.4 Infusion of Ir into Mesoporous Silica with CO2..…………….71 3.2.5 Determination of Nanocrystal Precipitation Boundary..……...74 3.2.6 Catalysis ……………………………………………………...76 3.3 Characterization………………………………………………………..76 3.3.1 Nitrogen Porosimetry…………………………………………76 3.3.2 Transmission Electron Microscopy………………………..…76 3.3.3 Gas Chromatography/ Mass Spectroscopy…………………...77 3.4 Results and Discussion…………………………………………………77 3.4.1 Nanoparticle Characterization……………………………..…77 3.4.2 Characterization of Mesoporous Silica……………………….80 3.4.3 Nanocrystal Dispersibility in Toluene-CO2 Mixtures………...80 3.4.4 Nanocrystal Loading in Pores using CO2…….………………82 3.4.5 Visualization of Nanocrystals in Mesoporous Silica…………85 3.4.6 Decene Hydrogenation in Ir/MPS Catalysts………………….88 3.4.7 Turnover Number for uncalcined Ir/MPS…………………….91 3.4.8 Effect of Catalyst Pretreatment on Catalytic Activity………..91 3.4.9 Stability of Supported Nanocrystals during the Reaction…….94 3.4.10 Activation Energy of the Decene Hydrogenation…………...95 3.5 Conclusions……………………………………………………………..99 3.6 References……………………………………………………………..100

Chapter 4: Stable Ordered FePt-Mesoporous Silica Catalysts with High Loadings……………………………………………………………...106 4.1 Introduction……………………………………………………………107 4.2 Experimental…………………………………………………………..109 4.2.1 FePt Nanocrystal Synthesis…………………………………109 4.2.2 Mesoporous Silica Synthesis………………………………..110 4.2.3 Infusion of FePt in Mesoporous Silica………………………110 4.2.4 Calcination…………………………………………………..111 4.2.5 Transmission Electron Microscopy…………………………111 4.2.6 X-ray diffraction…………………………………………….111 4.2.7 Catalysis……………………………………………………..112 4.3 Results…………………………………………………………………112 4.3.1 Structure of Nanocomposites after Infusion of FePt………...112 4.3.2 Stability of the FePt Nanocomposites……………………….118 4.3.3 Crystal Structure of the Bimetallic FePt after Annealing…...120 4.3.4 Activity and Stability of Catalyst for Decene Hydrogenation123 4.4 Discussion……………………………………………………………..125 4.4.1 Mechanism for Metal loading and Stability…………………125

x

4.4.2 Design of Bimetallic Nanocrystals…….……………………127 4.4.3 Catalytic Activity of Supported Nanocrystals………………130 4.5 Conclusions……………………………………………………………131 4.6 References……………………………………………………………..132

Chapter 5: Highly Stable ORR Electrocatalysts from Presynthesized Pt Nanocrystals Supported on Graphitic Mesoporous Carbons…….141 5.1 Introduction……………………………………………………………142 5.2 Experimental Section………………………………………………….145 5.3 Results…………………………………………………………………150 5.4 Discussion……………………………………………………………..169 5.4.1 Carbon Oxidation and its Effect on Electrocatalyst Stability.169 5.4.2 Catalyst Synthesis and Stability by Pt-Carbon Interactions...170 5.4.3 Control of Pt Nanoparticle Morphology for High Activity…174 5.5 Conclusions……………………………………………………………175 5.6 References……………………………………………………………..176

Chapter 6: Highly Stable and Active Electrocatalysts for Oxygen Reduction: Presynthesized Bimetallic Nanocrystals on Graphitic Carbon…..183 6.1 Introduction……………………………………………………….…...184 6.2 Experimental…………………………………………………………..186 6.3 Results…………………………………………………………………190 6.3.1 Structure of the Pt-Cu on Mesoporous Carbon Catalysts…...190 6.3.2 Activity and Stability of Pt-Cu Catalysts for ORR………….202 6.4 Discussion……………………………………………………………..205 6.4.1 Carbon Corrosion and Impact on Stability………………….205 6.4.2 Metal-Support Interactions: Catalyst Synthesis and Stability.208 6.4.3 Dealloying – Structure and Activity Enhancement…………211 6.4.4 Composition of Bimetallic Nanocrystals before Dealloying 212 6.5 Conclusions……………………………………………………………214 6.6 References …………………………………………………………….215

Chapter 7: Conclusions and Recommendations………………………………..223 7.1 Conclusions……………………………………………………………223 7.1.1 Infusion of Metal Nanocrystals in Mesoporous Silica by CO2223 7.1.2 Synthesis of Stable and Ordered FePt Bimetallic Catalysts...224 7.1.3 Stable Pt Electrocatalysts Supported on Graphitic Carbon ...224 7.1.4 Highly Stable and Active Bimetallic Pt-Cu for PEMFC……225 7.2 Recommendations……………………………………………………..226 7.2.1 Design of Core-Shell Bimetallic Catalysts for ORR………..226 7.2.2 Design of Active Catalysts for Water-Gas Shift Reaction…..226 7.3 References……………………………………………………………..227

xi

Appendix A: Infusion of FePt Nanocrystals in Mesoporous Silica……………229 A.1 Characterization………………………………………………………229 A.2 Results and Discussion………………………………………………..230 A.3 References……………………………………………………………..241

Appendix B: Highly Stable and Active Pt-Cu Electrocatalysts for ORR……..243 B.1 Characterization………………………………………………………243 B.2 Results and Discussion………………………………………………..244 B.3 References……………………………………………………………..257

Bibliography……………………………………………………………………….258

Vita…………………………………………………………………………………284

xii

List of Tables

Table 2.1 Methods for loading of metal nanoparticles in mesoporous silica…..29

Table 2.2 Properties of mesoporous silica……………………………………...40

Table 2.3 Overall Hamaker constants A132 between silica surface and gold

core mediated by supercritical CO2-toluene solution. Component

1 is gold, 2 is SiO2, and 3 is a mixture of toluene and CO2………….54

Table 2.4 Overall Hamaker constants A132 between silica surface and gold

core mediated by methanol-toluene solution. Component 1 is gold

, 2 is SiO2, and 3 is a mixture of toluene and methanol……………...56

Table 3.1 Activity and turnover frequency for decene hydrogenation of 1.3%

Ir supported on mesoporous silica relative to 5 % Pd on alumina. ….92

Table 5.1 Properties of mesoporous carbon…………………………………...153

Table 5.2 Loadings of infusion of Pt nanoparticles in carbon supports……….156

Table 5.3 Weight loss of the carbon supports at 195°C air for 16 days due to

carbon oxidation catalyzed by platinum……………………………..162

Table 5.4 Electrochemical data for 20% Pt on mesoporous carbons before and

after ADT potential cycling between 0.5 – 1.2 V at 50 mV/s for 1000

cycles in 0.1M HClO4 saturated with argon………………………...163

Table 6.1 Properties of mesoporous carbon…………………………………...191

Table 6.2 Infusion of Pt-Cu nanoparticles in mesoporous carbon supports.

All infusions were done on 50 mg carbon and nanoparticles were

dispersed in 3 mL of toluene………………………………………..196

Table 6.3 Composition and structural analysis of 14% Pt20Cu80/DMC-2000

before and after dealloying…………………………………………201

Table 6.4 Activity and stability of dealloyed Pt-Cu catalysts supported on

mesoporous carbons………………………………………………...206

Table A.1 Loading of FePt nanocrystals in mesoporous silica as a function of

time and concentration……………………………………………...240

xiii

List of Figures

Figure 1.1 Different techniques for the synthesis of heterogeneous catalysts…….2

Figure 1.2 Arrested precipitation technique of nanocrystal synthesis…………….4

Figure 1.3 Synthesis of bimetallic catalysts via wetness-impregnation and

infusion of presynthesized nanocrystals into porous supports………...6

Figure 1.4 Synthesis approaches for mesoporous materials………………………8

Figure 1.5 Influence of CO2 concentration on nanoparticle loading……………..11

Figure 1.6 TEM images of FePt nanocrystals (3.6 nm) after infusion into

mesoporous silica (7 nm). (a) HAADF STEM planar view; (b)

HRTEM………………………………………………………………12

Figure 2.1 Schematic of reaction infusion apparatus……………………………34

Figure 2.2 Calibration curve of absorbance versus concentration measured at

500 nm for gold nanocrystal solutions……………………………….36

Figure 2.3 Intensity I(q) vs q for gold nanocrystals dispersed in toluene

determined by SAXS………………………………………………....38

Figure 2.4 SAXS for calcined mesoporous silica SBA-15 (a), L-MCM41 (b) and

S-MCM41 (c)………………………………………………………...39

Figure 2.5 N2 isotherms and pore size for SBA-15 (a), L-MCM41 (b), and

SBA-15 (c)…………………………………………………………...42

Figure 2.6 Loading of gold nanocrystals versus vol% CO2 for 24 hour

impregnation period in L-MCM41. () 1 atm and 23°C ;() 241 atm

and 35°C……………………………………………………………..45

Figure 2.7 Variation of loading of gold nanocrystals with the pore

diameter of silica……………………………………………………...47

Figure 2.8 Loading versus time for S-MCM41. (a) Loading versus t1/2 for

up to 24 hours) with a correlation coefficient of 99.5%. (b)

xiv

Loading versus time approaching equilibrium……………………....48

Figure 2.9 TEM of gold impregnated SBA-15 (a), L-MCM41 (b), and

S-MCM41 (c)………………………………………………………...50

Figure 3.1 Schematic of the infusion apparatus…………………………………72

Figure 3.2 Calibration curve of UV- Vis absorbance versus concentration for

iridium nanocrystal dispersions in toluene, measured at 500 nm…....73

Figure 3.3 Schematic of apparatus for determining the dispersibility of the

iridium nanocrystals in CO2-toluene solvent………………………...75

Figure 3.4 GC-MS Calibration curve of a mixture of 1-decene and 1-decane….78

Figure 3.5 (a) TEM micrograph of Iridium nanocrystals stabilized by TOAB

and (b) corresponding histogram for Iridium nanocrystals………….79

Figure 3.6 (a) N2 adsorption-desorption isotherms for mesoporous silica MPS

and (b) corresponding pore size distributions for MPS……………...81

Figure 3.7 Dispersibility of Ir-TOAB nanocrystals in CO2-toluene mixed

solvent measured by UV-Vis spectroscopy………………………….83

Figure 3.8 (a) Z contrast STEM (Planar view) of MPS infused with Ir-TOAB;

(b) Out of focus view of (a); (c) top down view of (a); and (d) TEM

Micrograph of MPS infused with Ir-TOAB nanocrystals without any

pre-treatment…………………………………………………………87

Figure 3.9 Comparison of catalyst activities for decene hydrogenation with

Ir-MPS (without any pretreatment) and an industrial catalyst,

Pd-Al2O3……………………………………………………………..89

Figure 3.10 Calibration curve of UV- Vis absorbance versus concentration for

Iridium nanocrystal dispersions in 1-decene, measured at 500 nm….96

Figure 3.11 TEM micrographs of Ir-MPS (without any pretreatment) after 1-

decene hydrogenation………………………………………………..97

Figure 3.12 Arrhenius plot of decene hydrogenation over untreated Ir-MPS…….98

xv

Figure 4.1 (a) HRTEM images of pre-synthesized FePt nanocrystals prior to

infusion. (b) Histogram of nanocrystal size………………………...113

Figure 4.2 Loading of FePt nanocrystals within mesoporous silica in 10 minutes

as a function of initial concentration of nanocrystals………………114

Figure 4.3 TEM images of 7.3 wt % FePt nanocrystals after infusion into

mesoporous silica. (a) HAADF STEM planar view; (b) HRTEM;

(c) Histogram of nanocrystal size using (a)………………………...116

Figure 4.4 7.3 wt % FePt nanocrystals in mesoporous silica calcined at

700°C for 4 h: (a) HAADF- STEM; (b) Histogram of nanocrystals

after calcination determined from (a). (c) Filtered image of random

portion of (a) to eliminate shot noise; (d) Intensity profile of FePt

marked (A) in image (c); (e) Intensity profile of FePt marked

(B) in image (c)…………………………………………………….119

Figure 4.5 (a) XRD diffraction of FePt nanocrystals in silica before and after

annealing at 700 °C. (b) Convergent beam electron diffraction

of FePt nanocrystals in silica……………………………………….121

Figure 4.6 Turn over frequency for 1-decene hydrogenation by 5 wt % FePt

in mesoporous silica………………………………………………..124

Figure 5.1 (a) UV-Vis spectra of Pt nanoparticles dispersed in toluene before

and after infusion in AMC and (b) calibration curve for Pt

nanoparticles using known concentrations and measuring

absorbance at 800 nm………………………………………………148

Figure 5.2 SAXS of mesoporous carbons……………………………………...151

Figure 5.3 Nitrogen porosimetry of the mesoporous carbons and their pore size

distributions in the inset for GMC (A) and AMC (B)……………...152

Figure 5.4 XRD of the mesoporous carbon samples without added Pt………..155

Figure 5.5 TEM micrographs of 20% Pt on (A) GMC uncalcined and (B)

GMC calcined at 300 °C, (C) AMC calcined at 500 °C. (D), (E),

xvi

(F) Histograms for (A), (B) and (C), respectively. More than 50

nanoparticles were counted in each case…………………………...158

Figure 5.6 (A) TGA for the pure (blank) carbons and 20% Pt infused

mesoporous carbons AMC and GMC calcined at 500 and 300 °C

in N2/H2 atmosphere and (B) Differential thermal gravimetry (DTG)

for the Pt on AMC and GMC. Inset shows the magnified section of

(B) where onset of oxidation occurs for AMC…………………….159

Figure 5.7 XRD spectra of 20% Pt on (a) GMC and (b) AMC, before and

after ADT (for conditions in Table 5.4). Calcinations were done in

7% H2-93%N2 environment………………………………………...161

Figure 5.8 XPS spectra before and after ADT for 20% Pt supported on AMC

and GMC. (A),(B): C1s region for AMC and GMC, respectively.

(C),(D): Pt4f region for AMC and GMC, respectively…………….165

Figure 5.9 Electrochemical studies for 20% Pt on mesoporous carbon before

and after ADT cycling for conditions in Table 5.4 (A)CV of GMC

at 20 mV/s (B) CV of AMC at 20 mV/s (C) Linear Scan Voltammetry

(LSV) of GMC at 5 mV/s rotated at 1600 rpm (D) LSV for AMC

under identical conditions………………………………………….166

Figure 5.10 Tafel Slopes for Pt-GMC calcined at 300°C (A) Mass activity (B)

Specific activity and Pt-AMC calcined at 500°C (C) Mass Activity

and (D) Specific Activity…………………………………………...168

Figure 6.1 XRD spectra of the disordered mesoporous carbons annealed at

different temperatures………………………………………………193

Figure 6.2 Raman spectra of the disordered mesoporous carbons annealed at

different temperatures………………………………………………194

Figure 6.3 Cyclic voltammetry of Pt20Cu80 nanoparticles supported on

DMC-2000 during electrochemical dealloying at 20 mV/s in argon

atmosphere………………………………………………………….198

xvii

Figure 6.4 XRD spectra of the Pt20Cu80 nanocrystals supported on DMC-2000

before and after electrochemical dealloying of copper (a); peak

analysis of the XRD spectra of the as-synthesized Pt20Cu80

nanoparticles (b); and electrochemically dealloyed nanoparticles

with a final composition of Pt85Cu15………………………………..199

Figure 6.5 TEM micrographs and histograms of Pt20Cu80 on DMC-2000

(A,C) as-synthesized and calcined at 350 °C (B,C) Electrochemically

dealloyed……………………………………………………………203

Figure 6.6 Linear scan voltammetry of the dealloyed Pt-Cu nanoparticles

supported on DMC at 5 mV/s with rotation rate of 1600 rpm……...204

Figure 6.7 Potential-Activity plots of dealloyed Pt20Cu80 nanoparticles

supported on DMC-2000 and DMC-850 compared to commercial

Pt (ETEK) catalysts (a) Pt mass-based activities and (b) Pt specific

surface area activities……………………………………………….207

Figure A.1 SAXS on mesoporous silica of type SBA-15. Center-to-center

spacing between cylindrical pores is 8.5 nm………………………..231

Figure A.2 (a) Adsorption isotherms for SBA-15 measured by nitrogen

porosimetry. (b) Pore size distribution for SBA-15 measured by

nitrogen porosimetry………………………………………………..232

Figure A.3 HAADF-STEM micrographs of 7.3 wt % FePt nanocrystals in

mesoporous silica calcined at 700 oC in 7%H2/93%N2…………….233

Figure A.4 Total interaction energy as a function of separation distance and

metal nanocrystal size. Inset shows the van der Waals energy as a

function of separation distance for different nanocrystal sizes……..236

Figure A.5 Low resolution TEM of 7.3 wt % FePt in mesoporous silica………238

Figure A.6 Calibration of UV-Vis absorbance of FePt…………………………239

Figure B.1 Nitrogen adsorption-desorption curves (a) and corresponding pore

xviii

size distributions (b) for the disordered mesoporous carbons…….....245

Figure B.2 TEM micrographs (A) Low resolution (B) High-resolution and (C)

Histogram, of as-synthesized Pt20Cu80 nanoparticles. The average

nanoparticle size is 2.85±0.38 nm…………………………………..246

Figure B.3 (a) Effect of infusion of the Pt20Cu80 nanoparticles in DMC-850 on

absorbance as determined by UV-Visible spectrophotometry (b)

Calibration curve of the UV-Visual absorbance vs. concentration of

Pt20Cu80 nanoparticles in toluene, measured at 800 nm…………….247

Figure B.4 First 20 cyclic voltammetry profiles of Pt20Cu80 nanoparticles

supported on DMC-2000 during electrochemical dealloying at 20

mV/s in argon atmosphere. The black arrows indicate the direction in

which the peak shifts as cycling progresses………………………...249

Figure B.5 Cyclic voltammogram of 20% monometallic Pt on DMC-2000

scanned at 100 mV/s in argon atmosphere…………………………250

Figure B.6 XPS of 20% Pt supported on graphitized mesoporous carbon……..251

Figure B.7 TGA of Pt nanoparticles supported on disordered mesoporous

carbons……………………………………………………………...252

Figure B.8 Cycling voltammograms for Pt20Cu80 on DMC 2000 before and after

ADT potential cycling……………………………………………...253

Figure B.9 Peak analysis of the XRD spectra of the as-synthesized Pt75Cu25

nanoparticles………………………………………………………..254

1

Chapter 1

Introduction

A major goal in catalysis is the rational design of active catalyst sites to

produce high turnover numbers, selectivities and stabilities. 1-4 The arrangement of

metal clusters on supports which governs metal-support interactions has a profound

effect on catalytic performance. Metal nanocrystals with controlled morphology offer

high fractions of (1) coordinatively unsaturated surface atoms and (2) highly active

bifunctional perimeter sites at the metal-support interface.5,3 It is thus desirable to

design heterogeneous catalysts with control over the morphology of both the

nanocrystals and the support. Traditionally, heterogeneous catalysts are synthesized

by precursor reduction1,6,7, ion exchange8,9 and chemical vapor/fluid deposition.10-12

In these techniques, it is challenging to control the metal cluster size, alloy

composition, crystal structure and morphology. Metal precursors upon reduction may

sinter to form large nanocrystals with low active surface area and consequently show

poor activities (Figure 1.1). Synthesizing multimetallic catalysts by in-situ reduction

of multiple precursors on the supports may result in high compositionally

polydispersity, which is undesirable.13-15 By-products may be formed during the

reduction reaction and may be detrimental for the activity or selectivity of catalysts.

Another approach for synthesizing supported catalysts is to synthesize the

support structure in the presence of a dispersion of pre-synthesized nanocrystals.16,17

In this technique it can be difficult to achieve high loadings of dispersed nanocrystals

within the pores. In addition, the presence of the particles can interfere with the

synthesis of the desired pore structure and conversely, the synthesis of the

mesoporous material can change the morphology of the nanocrystals. The

nanocrystals may be completely trapped in the support matrix and may become

inaccessible to the reactants during the reaction. The ligands stabilizing the

2

Figure 1.1

Different techniques for the synthesis of heterogeneous catalysts

+ Metal precursor

Reduction

Metal oxide precursor (like TEOS)

Precursor reduction

Support synthesis in nanoparticle solution

+

3

nanocrystals may desorb and adsorb onto the substrate during the reaction, leading to

destabilization of the nanocrystals.

Yet another novel technique for catalyst design is to infusion of the

presynthesized nanocrystals in the catalyst supports (Figure 1.1).3,18,19 The

decoupling of the nanocrystal synthesis step and the infusion step, leads to exquisite

control of the nanocrystal size, morphology and dispersibility within the pores. An

advantage of this approach is that a large number of well-known synthetic methods

may be utilized for the design of the nanocrystals20 as well as that of support

materials.21 Prior to infusion, the particles are easily isolated from unwanted

byproducts, cleaned, separated by size, and the ligands may be exchanged if needed.

1.1 Nanocrystal Synthesis

The ability to control the uniformity of the size, shape, composition, and

surface properties of metal nanocrystals is desirable for optimizing their performance

in the catalysis. Recently, colloidal nanocrystals with controlled structure and

morphology have been synthesized by arrested precipitation technique using

stabilizing ligands (Figure 1.2).18,22-29 A typical synthesis system for colloidal

nanocrystals consists of three components: metal precursors, surfactants (ligands) and

solvents.20 The formation of the nanocrystals involves two steps: nucleation and

growth. In the nucleation step, precursors decompose or react, typically at a relatively

high temperature, to form a supersaturation of monomers followed by a burst of

nucleation of nanocrystals. The adhesion energy of ligands to the surface needs to be

such regions of the nanocrystal surface are transiently accessible for growth, yet

entire crystals are, on average, monolayer- protected to block aggregation.

In arrested precipitation technique size, morphology and composition can be

controlled by adjusting the reaction conditions such as time, temperature, precursor

concentration, surfactant type, and surfactant concentration.30-33 When the

nanocrystals are pre-synthesized, heterogeneous sites on a support are not present,

4

Figure 1.2

Arrested precipitation technique of nanocrystal synthesis.

Metal precursor + Ligand + Solvent

Reduction

Arrested Growth Metal nanoparticle dispersion

3.6 nm FePt nanoparticles

5

which could otherwise lead to uncontrolled composition, resulting from competition

between nucleation in solution and on the surface (Figure 1.3). When the rates of

precursor decomposition and equilibrium solubilities of the two resulting metals are

similar, formation of binary metal nuclei is favored.31 Even when these properties are

somewhat dissimilar, the nuclei formed from the first metal may capture newly

formed atoms of the second metal before it nucleates, which would otherwise lead to

two separate populations of growing nanocrystals.34 As metal from the more rapidly

decaying precursor is incorporated into the nanoparticle, it will be depleted in the

solution relative to the second metal. Thus, the probability of adding the second

component will increase. In essence, this may be considered to be composition

focusing by analogy to size focusing. The temperature is often a key variable to

produce rapid decomposition of both precursors to achieve focusing of both size and

composition, along with sufficient dynamic solvation by the ligand to arrest growth.

Once the particles are pre-synthesized, they may be loaded physically into the pores

of the support at completely different concentrations than those used in the synthesis

of the particles. Thus, both the pre-synthesis and infusion steps may be optimized in

ways not available when precursors are reduced within pores of the support.

1.2 Mesoporous Supports

An emerging frontier in catalysis is the design of extremely active and stable

catalysts by depositing pre-synthesized nanocrystals with desired morphologies onto

high surface area supports with ordered mesopores.34-38 Mesoporous materials of

silica such as MCM-41 and SBA-15 and various metal oxides39 have been

synthesized with surfactant templates to produce periodic ordered cylindrical pores

with diameters in the range of 2 to 20 nm with internal surface areas up to 1000

m2/g.40-42 The pore diameter is ideal for dispersing metal nanocrystals with diameters

in the same size range and for the transport of reactants and products. The high

dispersion of the nanocrystals within the extremely high surface area pores will

6

Figure 1.3

Synthesis of bimetallic catalysts via wetness-impregnation and infusion of

presynthesized nanocrystals into porous supports.

Metal-1 precursor

Metal-2 precursor

Reduction over Supports

Heterogeneous Nucleation

Support

Growth

Support

Support

Homogeneous Nucleation

Growth

Metal-1 precursor

Metal-2 precursor

Reduction in solvent in presence of ligands

Support

Infusion

7

minimize contact between the nanocrystals to prevent sintering. Particles in adjacent

pores are completely isolated. Particles in a single cylindrical pore are separated over

a long pore length. The convex surfaces on the nanocrystals favored contact with the

cylindrical mesopores with concave curvature, enhancing the geometric interfacial

area which favors strong metal-support interactions, thereby facilitating high stability

of nanocrystals towards desorption or sintering.43 These strong interactions can be

used to manipulate the crystal structure of nanoparticles, while controlling the

nanoparticle size. For example, ordered intermetallic FePt alloy with <4 nm size is

achieved by transformation of the disordered face-centered-cubic FePt nanocrystals

supported in mesoporous silica by annealing at 700 oC. However, the surfaces of both

the metal nanocrystal and support are often both convex for traditional supports

leading to a smaller degree of interfacial contact and poor metal-support interactions.

For traditional supports, the nanocrystals are supported on the external surface of

support particles and have a greater tendency to contact other nanocrystals and grow

by sintering.

Typically, mesoporous materials are synthesized by a soft surfactant

templating approach, where long-chained surfactants are used as a template.21 These

surfactants tend to form complex micelles in solution whose morphology can be

controlled by altering surfactant concentration. The morphology of mesoporous

materials can in turn be tuned by controlling the morphology of the surfactant

micelles. When silica/metal-oxide precursors are introduced in the micelles,

mesostructures are formed by a cooperative self-assembly of surfactant/precursor ion

pairs that arrange themselves in various ordered arrays (Figure 1.4). Upon

condensation of the metal-oxide oligomers and subsequent surfactant removal by

calcination or ion exchange, the mesoporous materials with surface areas up to 1000

m2/g and pore diameter up to 20 nm are formed. For example, cylindrical

mesoporous silica of type SBA-15 with pore size ~ 10 nm is synthesized using

amphiphilic surfactant pluronic P123 (EO20PO70EO20) as a template and

tetraethoxysilane as a silica precursor (Figure 1.4).42

8

(C) (D)

Figure 1.4

(A, B) Synthesis approaches for mesoporous materials. The mesostructure can be

formed previously (A) or by a cooperative process between surfactants and metal-

oxide precursors. Figure taken from Soler-Illia et al.21 TEM images of mesoporous

silica of type SBA-15 with 6 nm pores (top view) (C) and planar view (D). Images

taken from Zhao et al.42

9

Another approach for synthesizing mesoporous materials is to cross-link the

metal-oxide precursors within the preformed mesoporous materials like silica which

are used as hard templates, followed by etching away the template.44,45 In this way the

morphology of the mesoporous template is transferred to the new mesoporous metal-

oxide materials. The first self-supported highly ordered mesoporous carbon of type

CMK-1 with 3-d cubic morphology was synthesized by using 3-D highly ordered

mesoporous silicate material, MCM-48, as a hard template.46 Since then several

different morphologies of mesoporous carbons have been synthesized by hard

templating method.47-49

1.3 Infusion of Presynthesized Nanocrystals into Mesoporous Supports

Relatively few studies have examined infusion of pre-synthesized

nanocrystals into porous materials with carrier solvents.18,19,50 There are two primary

challenges in the infusion of pre-synthesized nanocrystals into a nanoporous matrix.

First, the nanocrystal dispersion must be transported through the pores by favorable

capillary wetting, and the nanocrystals must not block small pore entrances. Once the

nanocrystals penetrate the pores, the second challenge is to achieve sufficient

adsorption of the nanocrystals from the solution phase to the substrate surface. As the

surface of the metal is covered with the stabilizing ligands which may bind very

strongly to the metal, like thiol to gold surface, short-ranged metal-support

interactions which facilitate adsorption may be highly screened. To achieve the metal

loadings >0.5 wt %, the long-ranged van der Waals (vdW) interactions be used;

However, the solvent must allow for enough attraction between the nanocrystals and

substrate to achieve sufficient adsorption, while simultaneously preventing

flocculation of nanocrystals in the solvent phase. The nanocrystals must be dispersed

in “good” solvent for the stabilizing ligands to prevent flocculation. However, good

solvents may be expected to cause the interactions between the nanocrystals and the

surface of the nanoporous material to be repulsive or weakly attractive.30 The

10

resulting weak adsorption of the particles inside the pores may produce low loadings.

Thus, the solvent must be tuned to provide a judicious balance between nanocrystal-

nanocrystal versus nanocrystal-substrate interactions.

The use of supercritical CO2 as an antisolvent helps overcome both transport

and thermodynamic challenges in nanoparticle infusion into porous materials (Figure

1.5).18 Supercritical CO2 is added to an organic solvent like toluene to achieve high

levels of infusion of pre-synthesized nanocrystals into mesoporous silica by

enhancing long-ranged interactions. CO2 has an unusually low polarizability/density

(very weak VdW forces) and upon addition to toluene forms a mixture of low

Hamaker constant. This consequently weakens the vdW forces for the intervening

solvent and results in a stronger interaction between the metal and support thus

facilitating nanocrystal adsorption on the porous surface and enhancing the loadings

upto 5 wt %.18,50 Furthermore, the loading may be tuned over a wide range by

controlling the van der Waals interactions by varying the CO2 composition.

Yet another technique to further enhance the loadings over 5 wt % not

typically achieved by tuning the long-ranged van der Waals interactions using CO2, is

to strengthen short-ranged interactions between the metal and support.19 This is done

by strategic selection of weakly binding ligands for nanocrystal stabilization like

amines and fatty acids that known to bind to metals like Pt or Au weakly relative to

ligands like thiols.51 For example, surface coverage of dodecylamine on Pt is 46%52

compared to 75% for dodecanethiol on Ag26. The lower surface coverage exposes

barer metal surface for the formation of specific short-ranged interactions with the

silica. These short-ranged interactions could facilitate the loadings of ~20 wt %,

relative to the much lower loadings for gold capped with dodecanethiol,18 where the

surface coverage is higher. For example, Figure 1.6 shows the FePt nanoparticles

with 8 wt % loadings supported on mesoporous silica. The convex surfaces on the

nanocrystals favored contact with the cylindrical mesopores with concave curvature,

further enhancing metal-support interactions.43

11

Figure 1.5

Influence of CO2 concentration on nanoparticle loading. (A) Optimized addition of

the CO2 in toluene promotes nanocrystal adsorption on the supports while

simultaneously keeping nanocrystals dispersed in the solvent. (B) Too much addition

of CO2 in toluene leads to nanocrystal flocculation and poor loadings.

10 μm

~5 nm

Nanocrystals dispersed in CO2-toluene

10 μm

~5 nm

(A)

(B)

12

(a) (b)

Figure 1.6

TEM images of FePt nanocrystals (3.6 nm) after infusion into mesoporous silica (7

nm). (a) HAADF STEM planar view; (b) HRTEM; Figures taken from Chapter 4.

13

1.4 Electrocatalysts for Oxygen Reduction Reaction: Activity and

Stability

The high dispersion of catalysts on mesoporous supports is beneficial for

electrocatalysis, for example for the oxygen reduction reaction for Pt on C. Polymer

electrolyte membrane fuel cells (PEMFC) are considered a potential alternative

source of energy due to high efficiencies and clean operation.53-55 Commercialization

of PEMFCs is still a challenge due to the high cost of platinum based catalysts as well

as poor catalyst stability.56-58 The performance of the platinum catalysts needs to be

increased to 4-5 times that of current commercial catalysts in order to meet the

performance targets for fuel cell cathodes.53 To achieve activity enhancement,

platinum based bimetallic alloys, Pt-M (M=Co, Ni, Fe, V, Ti), have been extensively

studied and activity enhancements up to 2.5 times that of pure Pt have been

achieved.59-65 Recently, Strasser and co-workers synthesized novel Pt-Cu catalysts on

carbon with a pronounced activity enhancement of >4 time that of pure Pt

catalysts.66,67 The copper was electrochemically etched from the nanoparticle surface

to form a strained Pt shell. However, the catalysts were synthesized by sequential

wetness impregnation and led to catalysts with multiple alloys phases. Another

significantly bigger challenge hindering the fuel cell technology is the poor catalyst

stability.53,56,68-70 For automotive applications, catalyst support stability at 1.2V is

required for an accumulated life time of a vehicle, with a maximum allowable

performance loss of <30 mV at 1.5A/cm2 (< 1 µV/cycle target).56 Traditionally

amorphous carbons, such as Vulcan XC-72 carbon, easily oxidize at potentials of

1.2V, leading to significant activity losses due carbon oxidation, metal nanoparticle

coalescence and metal dissolution.56

The challenges in stability and activity can be ameliorated by the infusion of

presynthesized nanocrystals with controlled size, composition, crystal phases and

morphology in the graphitic mesoporous carbon supports. Graphitized carbons are

more resistant to oxidation than amorphous carbons due to significantly reduced

14

number of oxygen sites on the surface.69,71-73 In addition, the abundant π electron sites

on the surface interact strongly with platinum, thus strengthening the metal-support

interactions, as demonstrated by XPS and electron-spin resonance (ESR).69,71,73

Strong metal-support interactions are expected to mitigate coalescence , dissolution

and subsequent Ostwald ripening of the Pt. Thus, oxidation resistant graphitic

supports with strong Pt-support interactions are expected to enhance the stability of

ORR catalysts.

1.5 Objectives

The objective of this thesis is to synthesize supported catalysts with control

over the morphology of the nanoparticles as well as of the supports for achieving high

catalyst activities and stabilities. To accomplish this, nanocrystals with controlled

size, composition and alloy structure have been presynthesized using stabilizing

ligands and then infused within high surface area mesoporous supports. One of the

key objectives of this thesis is to synthesize the catalysts with high metal loadings.

Thus, the effect of the ligand-metal binding strength on metal loadings was studied.

Nanocrystals stabilized with strongly binding ligands preclude short-ranged metal-

support interactions and lead to poor loadings. For achieving significant metal

loadings of up to 5 wt % for strongly bound nanocrystals, the use of supercritical CO2

as an antisolvent in toluene to enhance the long-ranged metal-support interactions is

explored. Subsequently, to achieve the nanocrystal loadings ≥5 wt %, weakly binding

hydrocarbon ligands that allow short-ranged metal-support interactions are used for

nanocrystal synthesis. These weakly binding hydrocarbon ligands are demonstrated to

be removed from the metal surface by mild thermal treatment without perturbing the

morphology of the metal or the support. The easy ligand removal frees metal sites and

facilitates high catalyst activities.

Another key focus of the thesis is enhancement of the catalyst stability.

Immobilization of the nearly spherical nanocrystals in concave mesopores with

15

similar radii enhances the interfacial area and favors strong metal-support

interactions, relative to other support geometries. The high stabilities resulting from

these strong interactions have been shown to be beneficial in the transformation of the

bimetallic nanoparticle structure from a disordered to an ordered intermetallic alloy

upon annealing at 700 oC without perturbing the size. The fundamental understanding

of the metal-support interactions is then utilized in the design of highly stable Pt

electrocatalysts with controlled size and morphology for oxygen reduction using

graphitized mesoporous carbons as supports. The high surface area and resistance to

oxidation of the graphitic carbon69 in addition to the strong binding to Pt69 has been

demonstrated to mitigate nanoparticle isolation, sintering and dissolution, and hence

enhance catalyst stability. Finally, highly stable Pt-Cu catalysts on mesoporous

graphitic carbons for the oxygen reduction have been synthesized to achieve activities

4-fold that of pure Pt, by selective electrochemical dealloying of copper. The

relationship between the catalyst activity and strain in the Pt shell produced by copper

dealloying has been investigated by controlling the composition of the single-phased

Pt-Cu nanoparticles.

1.6 Dissertation Outline

Chapter 2 represents the first example of infusion of presynthesized gold

nanocrystals into mesoporous silica using sc-CO2 as an antisolvent in toluene. The

van der Waals interactions between gold nanocrystals, stabilized with dodecanethiol,

and silica are strengthened upon addition of CO2 to toluene. The enhanced long

ranged interactions facilitate adsorption of nanocrystals on the silica thus leading to

metal loadings of ~ 2 wt %. Concentration of CO2 in toluene needs to be tuned to

promote nanocrystal adsorption without flocculating them in the solvent. A simple

van der Waals model for calculating the overall Hamaker constant between the metal

nanocrystal and silica wall, with an intervening solvent, is developed and

16

qualitatively predicts the linear increase in loadings with increase of CO2

concentration in toluene, as seen experimentally.

While Chapter 2 examined just the synthesis of the gold catalysts with

significant metal loadings, in Chapter 3 iridium catalysts are synthesized using

supercritical CO2 and investigated for hydrogenation reactions. The

tetraoctylammonium bromide ligands for iridium nanoparticle stabilization are

strategically chosen such that they bind sufficiently weakly to the nanoparticle

surface, so that it becomes unnecessary to pre-treat the catalyst at high temperature to

remove the ligand, for activation of the catalysts. The iridium catalysts are used for 1-

decene hydrogenation and show good catalyst activities which are higher than

homogeneous iridium catalysts and twice better than commercial Pd-Al2O3 catalysts.

The catalyst does not show any enhancement of the activity when pre-treated under

high temperatures or when annealed in supercritical CO2 indicating migration of

weakly binding ligands on to high surface area supports after nanoparticle adsorption.

Whereas CO2 was used to facilitate nanocrystal adsorption on the support in

the above two chapters, it was not needed in Chapters 4 to 6, since the interactions of

the nanocrystals with the support were stronger as a consequence of uncovered metal

on the particle surface. Chapter 4 demonstrates for the first time, a novel concept to

form supported catalysts by infusion of pre-synthesized bimetallic nanocrystals into

ordered mesoporous supports, facilitated by short-ranged metal-support interactions.

For pre-synthesized FePt nanocrystals (<4 nm), coated with oleic acid and oleylamine

ligands in toluene, high loadings above 10 wt % are achieved in order of minutes. The

strong metal-support interactions are favored by the low coverage of the weakly

bound ligands and enhanced interfacial area between spherical nanocrystals and

concave mesopores. These strong interactions facilitate phase transformation of FePt

from a disordered phase (FCC) to ordered intermetallic phase (FCT) by annealing at

temperatures of 700 °C while preventing any significant sintering. The calcined FePt

catalysts are used for 1-decene hydrogenation and exhibit 6-fold higher catalyst

17

activity (TOF =30 s-1) than that of a commercial Pd-alumina catalyst with stability for

multiple reactions.

Whereas Chapters 3-4 examined hydrogenation reactions, Chapters 5-6

examine electrocatalysis with similar types of pre-synthesized nanocrystal catalysts,

but with mesoporous carbon supports. In Chapter 5, highly stable Pt electrocatalysts

have been prepared by the infusion of pre-synthesized nanocrystals with controlled

size (<4 nm) and morphology onto the ordered graphitic mesoporous carbon. Weakly

binding ligands (dodecylamine) stabilized the nanocrystals, while exposing sufficient

amount of bare metal surface to facilitate strong Pt-support interactions, resulting in

nanocrystal loadings of 20 wt % with high metal dispersion. The mass activity for the

oxygen reduction reaction (ORR) at 0.9V reaches 0.14 A/mgPt for Pt on the graphitic

mesoporous carbon, relative to ~0.11 A/mgPt for Pt on commercial amorphous

Vulcan XC-72. The loss in ORR catalyst activity for Pt supported on graphitic

mesoporous carbons under an accelerated durability test of potential cycling between

0.5 to 1.2 V for 1000 cycles, is extremely small, <2%, compared to a 30% loss for Pt

on amorphous mesoporous carbons and 70% typically for commercial Pt on

amorphous carbon catalysts. The highly oxidation resistant graphitized carbon

coupled with the strong Pt interactions with the π sites of graphite, prevents metal

sintering, Pt dissolution, and Ostwald ripening, resulting in high catalyst stability.

Chapter 6 is a follow up study of Chapter 5 where, highly stable and active

electrocatalysts have been prepared by the infusion of pre-synthesized bimetallic Pt-

Cu nanocrystals with controlled size, composition and alloy structure onto graphitic

mesoporous carbon. Electrochemically dealloying via leaching of Cu from the Pt-Cu

nanoparticle surface generates strained Pt rich shell raising the catalyst activity for

ORR 4-folds over commercial Pt catalysts. The low polydispersity in the size and

composition during pre-synthesis is maintained through the dealloying stage,

resulting in a Pt shell with a well controlled level of strain that favors high ORR

activity. The catalysts supported on graphitic mesoporous carbons are stable towards

potential cycling between 0.5- 1.2V for 1000 cycles and show a loss of <2 % activity,

18

compared to a 40% loss for identical nanoparticles on amorphous mesoporous

carbons. The high stability is due to oxidation resistant carbon and strong Pt-

π interactions.

The dissertation conclusion and recommendations for future work are given in

Chapter 7. Supporting information for Chapter 4 can be found in Appendix A, while

Appendix B contains supplementary information for Chapter 7.

1.7 References

(1) Wakayama, H.; Setoyama, N.; Fukushima, Y. Size-Controlled

Synthesis and Catalytic Performance of Pt Nanoparticles in Micro- and Mesoporous

Silica Prepared Using Supercritical Solvents. Adv. Mater. 2003, 15, 742-745.

(2) Rolison, D. R. Catalytic nanoarchitectures - The importance of nothing

and the unimportance of periodicity. Science (Washington, DC, United States) 2003,

299, 1698-1702.

(3) Rioux, R. M.; Song, H.; Hoefelmeyer, J. D.; Yang, P.; Somorjai, G. A.

High-Surface-Area Catalyst Design: Synthesis, Characterization, and Reaction

Studies of Platinum Nanoparticles in Mesoporous SBA-15 Silica. Journal of Physical

Chemistry B 2005, 109, 2192-2202.

(4) Liang, C.; Dai, S. Synthesis of Mesoporous Carbon Materials via

Enhanced Hydrogen-Bonding Interaction. Journal of the American Chemical Society

2006, 128, 5316-5317.

(5) Cho, A. Catalysis news: Connecting the dots to custom catalysts.

Science (Washington, DC, United States) 2003, 299, 1684-1685.

(6) Lensveld, D. J.; Mesu, J. G.; van Dillen, A. J.; de Jong, K. P. The

application of well-dispersed nickel nanoparticles inside the mesopores of MCM-41

by use of a nickel citrate chelate as precursor. Studies in Surface Science and

Catalysis 2002, 143, 647-657.

19

(7) Zong, Y.; Watkins, J. J. Deposition of copper by the H2-assisted

reduction of Cu(tmod)2 in supercritical carbon dioxide: Kinetics and reaction

mechanism. Chemistry of Materials 2005, 17, 560-565.

(8) Singh, U. K.; Albert Vannice, M. Liquid-Phase Hydrogenation of

Citral over Pt/SiO2 Catalysts. Journal of Catalysis 2000, 191, 165-180.

(9) Goodenough, J. B.; Hamnett, A.; Kennedy, B. J.; Manoharan, R.;

Weeks, S. A. Methanol oxidation on unsupported and carbon supported platinum +

ruthenium anodes. Journal of Electroanalytical Chemistry and Interfacial

Electrochemistry 1988, 240, 133-145.

(10) Hayek, K.; Goller, H.; Penner, S.; Rupprechter, G.; Zimmermann, C.

Regular Alumina-Supported Nanoparticles of Iridium, Rhodium and Platinum Under

Hydrogen Reduction: Structure, Morphology and Activity in the Neopentane

Conversion. Catalysis Letters 2004, 92, 1-9.

(11) Serp, P.; Hierso, J. C.; Feurer, R.; Kihn, Y.; Kalck, P.; Faria, J. L.;

Aksoylu, A. E.; Pacheco, A. M. T.; Figueiredo, J. L. Single-step preparation of

activated carbon supported platinum catalysts by fluidized bed organometallic

chemical vapor deposition. Carbon 1999, 37, 527-530.

(12) Kim, K. T.; Kim, Y. G.; Chung, J. S. Effect of surface roughening on

the catalytic activity of Pt-Cr electrocatalysts for the oxygen reduction in phosphoric

acid fuel cell. Journal of the Electrochemical Society 1995, 142, 1531-1538.

(13) Raghuveer, V.; Manthiram, A.; Bard, A. J. Pd-Co-Mo Electrocatalyst

for the Oxygen Reduction Reaction in Proton Exchange Membrane Fuel Cells.

Journal of Physical Chemistry B 2005, 109, 22909-22912.

(14) Antolini, E. Formation of carbon-supported PtM alloys for low

temperature fuel cells: a review. Materials Chemistry and Physics 2003, 78, 563-573.

(15) Koh, S.; Leisch, J.; Toney, M. F.; Strasser, P. Structure-Activity-

Stability Relationships of Pt-Co Alloy Electrocatalysts in Gas-Diffusion Electrode

Layers. Journal of Physical Chemistry C 2007, 111, 3744-3752.

20

(16) Konya, Z.; Puntes, V. F.; Kiricsi, I.; Zhu, J.; Alivisatos, P.; Somorjai,

G. A. Novel Two-Step Synthesis of Controlled Size and Shape Platinum

Nanoparticles Encapsulated in Mesoporous Silica. Catalysis Letters 2002, 81, 137-

140.

(17) Konya, Z.; Puntes, V. F.; Kiricsi, I.; Zhu, J.; Ager, J. W., III; Ko, M.

K.; Frei, H.; Alivisatos, P.; Somorjai, G. A. Synthetic Insertion of Gold Nanoparticles

into Mesoporous Silica. Chemistry of Materials 2003, 15, 1242-1248.

(18) Gupta, G.; Shah, P. S.; Zhang, X.; Saunders, A. E.; Korgel, B. A.;

Johnston, K. P. Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with

Supercritical Carbon Dioxide. Chemistry of Materials 2005, 17, 6728-6738.

(19) Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.;

Johnston, K. P. Stable Ordered FePt Mesoporous Silica Catalysts with High Loadings

Chemistry of Materials 2008.

(20) Yin, Y.; Alivisatos, A. P. Colloidal nanocrystal synthesis and the

organic-inorganic interface. Nature (London, United Kingdom) 2005, 437, 664-670.

(21) de Soler-Illia, G. J.; Sanchez, C.; Lebeau, B.; Patarin, J. Chemical

Strategies To Design Textured Materials: from Microporous and Mesoporous Oxides

to Nanonetworks and Hierarchical Structures. Chemical Reviews (Washington, DC,

United States) 2002, 102, 4093-4138.

(22) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R.

Synthesis of Thiol-Derivatised Gold Nanoparticles in a Two-Phase Liquid-Liquid

System. J. Chem. Soc., Chem. Commun. 1994, 7, 801-802.

(23) Sun, S. Recent advances in chemical synthesis, self-assembly, and

applications of FePt nanoparticles. Advanced Materials (Weinheim, Germany) 2006,

18, 393-403.

(24) Sun, S.; Fullerton, E. E.; Weller, D.; Murray, C. B. Compositionally

controlled FePt nanoparticle materials. IEEE Transactions on Magnetics 2001, 37,

1239-1243.

21

(25) Korgel, B. A.; Fitzmaurice, D. Small-angle x-ray-scattering study of

silver-nanocrystal disorder-order phase transitions. Physical Review B: Condensed

Matter and Materials Physics 1999, 59, 14191-14201.

(26) Korgel, B. A.; Fullam, S.; Connolly, S.; Fitzmaurice, D. Assembly and

Self-Organization of Silver Nanocrystal Superlattices: Ordered \"Soft Spheres\".


(27) Knecht, M. R.; Garcia-Martinez, J. C.; Crooks, R. M. Hydrophobic

Dendrimers as Templates for Au Nanoparticles. Langmuir 2005, 21, 11981-11986.

(28) Kim, Y.-G.; Garcia-Martinez, J. C.; Crooks, R. M. Electrochemical

Properties of Monolayer-Protected Au and Pd Nanoparticles Extracted from within

Dendrimer Templates. Langmuir 2005, 21, 5485-5491.

(29) Ferrer, D.; Blom, D. A.; Allard, L. F.; Mejia, S.; Perej-Tijerina, E.;

Yacaman, M. J. Atomic structure of three-layer Au/Pd nanoparticles revealed by

aberration-corrected scanning transmission electron microscopy. Journal of Materials

Chemistry 2008, 18, 2442.

(30) Israelachvili, J. Intermolecular & Surface Forces; 2nd ed.; Academic

Press, Inc.: San Diego, 1992.

(31) Sun, X.; Kang, S.; Harrell, J. W.; Nikles, D. E.; Dai, Z. R.; Li, J.;

Wang, Z. L. Synthesis, chemical ordering, and magnetic properties of FePtCu

nanoparticle films. Journal of Applied Physics 2003, 93, 7337-7339.

(32) Santiago, E. I.; Varanda, L. C.; Villullas, H. M. Carbon-Supported Pt-

Co Catalysts Prepared by a Modified Polyol Process as Cathodes for PEM Fuel Cells.

Journal of Physical Chemistry C 2007, 111, 3146-3151.

(33) Alden, L. R.; Roychowdhury, C.; Matsumoto, F.; Han, D. K.;

Zeldovich, V. B.; Abruna, H. D.; DiSalvo, F. J. Synthesis, Characterization, and

Electrocatalytic Activity of PtPb Nanoparticles Prepared by Two Synthetic

Approaches. Langmuir 2006, 22, 10465-10471.

22

(34) Chan, K.-Y.; Ding, J.; Ren, J.; Cheng, S.; Tsang, K. Y. Supported

mixed metal nanoparticles as electrocatalysts in low temperature fuel cells. Journal of

Materials Chemistry 2004, 14, 505-516.

(35) Lu, Y. Surfactant-templated mesoporous materials: from inorganic to

hybrid to organic. Angewandte Chemie, International Edition 2006, 45, 7664-7667.

(36) Schueth, F. Non-siliceous Mesostructured and Mesoporous Materials.

Chemistry of Materials 2001, 13, 3184-3195.

(37) Tanev, P. T.; Pinnavaia, T. J. Mesoporous Silica Molecular Sieves

Prepared by Ionic and Neutral Surfactant Templating: A Comparison of Physical

Properties. Chemistry of Materials 1996, 8, 2068-2079.

(38) Tanev, P. T.; Pinnavaia, T. J. A neutral templating route to

mesoporous molecular sieves. Science (Washington, D. C.) 1995, 267, 865-867.

(39) Chen, Y.; Yang, F.; Dai, Y.; Wang, W.; Chen, S. Ni@Pt core-shell

nanoparticles: synthesis, structural and electrochemical properties. Journal of

Physical Chemistry C 2008, 112, 1645-1649.

(40) Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D.

Block Copolymer Templating Syntheses of Mesoporous Metal Oxides with Large

Ordering Lengths and Semicrystalline Framework. Chemistry of Materials 1999, 11,

2813-2826.

(41) Li, Y.; Fu, Q.; Flytzani-Stephanopoulos, M. Low-temperature water-

gas shift reaction over Cu- and Ni-loaded cerium oxide catalysts. Applied Catalysis,

B: Environmental 2000, 27, 179-191.

(42) Yang, H.; Yan, Y.; Liu, Y.; Zhang, F.; Zhang, R.; Meng, Y.; Li, M.;

Xie, S.; Tu, B.; Zhao, D. A Simple Melt Impregnation Method to Synthesize Ordered

Mesoporous Carbon and Carbon Nanofiber Bundles with Graphitized Structure from

Pitches. Journal of Physical Chemistry B 2004, 108, 17320-17328.

(43) Haller, G. L. New catalytic concepts from new materials:

understanding catalysis from a fundamental perspective, past, present, and future.

Journal of Catalysis 2003, 216, 12-22.

23

(44) Liang, C.; Li, Z.; Dai, S. Mesoporous carbon materials: synthesis and

modification. Angewandte Chemie, International Edition 2008, 47, 3696-3717.

(45) Lee, J.; Kim, J.; Hyeon, T. Recent progress in the synthesis of porous

carbon materials. Advanced Materials (Weinheim, Germany) 2006, 18, 2073-2094.

(46) Ryoo, R.; Joo, S. H.; Jun, S. Synthesis of Highly Ordered Carbon

Molecular Sieves via Template-Mediated Structural Transformation. Journal of

Physical Chemistry B 1999, 103, 7743-7746.

(47) Gierszal, K. P.; Kim, T.-W.; Ryoo, R.; Jaroniec, M. Adsorption and

Structural Properties of Ordered Mesoporous Carbons Synthesized by Using Various

Carbon Precursors and Ordered Siliceous P6mm and Ia.hivin.3d Mesostructures as

Templates. Journal of Physical Chemistry B 2005, 109, 23263-23268.

(48) Ryoo, R.; Joo, S. H. Nanostructured carbon materials synthesized from

mesoporous silica crystals by replication. Studies in Surface Science and Catalysis

2004, 148, 241-260.

(49) Joo, S. H.; Choi, S. J.; Oh, I.; Kwak, J.; Liu, Z.; Terasaki, O.; Ryoo, R.

Ordered nanoporous arrays of carbon supporting high dispersions of platinum

nanoparticles. [Erratum to document cited in CA135:184349]. Nature (London,

United Kingdom) 2001, 414, 470.

(50) Gupta, G.; Stowell, C. A.; Patel, M. N.; Gao, X.; Yacaman, M. J.;

Korgel, B. A.; Johnston, K. P. Infusion of Pre-synthesized Iridium Nanocrystals into

Mesoporous Silica for High Catalyst Activity. Chemistry of Materials 2006, 18,

6239-6249.

(51) Jana, N. R.; Peng, X. Single-Phase and Gram-Scale Routes toward

Nearly Monodisperse Au and Other Noble Metal Nanocrystals. Journal of the

American Chemical Society 2003, 125, 14280-14281.

(52) Fu, X.; Wang, Y.; Wu, N.; Gui, L.; Tang, Y. Surface Modification of

Small Platinum Nanoclusters with Alkylamine and Alkylthiol: An XPS Study on the

Influence of Organic Ligands on the Pt 4f Binding Energies of Small Platinum

Nanoclusters. Journal of Colloid and Interface Science 2001, 243, 326-330.

24

(53) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Activity

benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts

for PEMFCs. Applied Catalysis, B: Environmental 2005, 56, 9-35.

(54) Dresselhaus, M. S.; Thomas, I. L. Alternative energy technologies.

Nature (London, United Kingdom) 2001, 414, 332-337.

(55) Steele, B. C. H.; Heinzel, A. Materials for fuel-cell technologies.

Nature (London, United Kingdom) 2001, 414, 345-352.

(56) Mathias, M. F.; Makharia, R.; Gasteiger, H. A.; Conley, J. J.; Fuller, T.

J.; Gittleman, C. I.; Kocha, S. S.; Miller, D. P.; Mittelsteadt, C. K.; Xie, T.; Yan, S.

G.; Yu, P. T. Two fuel cell cars in every garage? Electrochemical Society Interface

2005, 14, 24-35.

(57) Wagner, F. T.; Gasteiger, H. A.; Makharia, R.; Neyerlin, K. C.;

Thompson, E. L.; Yan, S. G. Catalyst development needs and pathways for

automotive PEM fuel cells. ECS Transactions 2006, 3, 19-29.

(58) Gasteiger, H. A. Proton exchange fuel cell materials and R&D needs

for future market success. Electrochemistry (Tokyo, Japan) 2007, 75, 103.

(59) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.;

Lucas, C. A.; Markovic, N. M. Improved Oxygen Reduction Activity on Pt3Ni(111)

via Increased Surface Site Availability. Science (Washington, DC, United States)

2007, 315, 493-497.

(60) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.;

Lucas, C. A.; Wang, G.; Ross, P. N.; Markovic, N. M. Trends in electrocatalysis on

extended and nanoscale Pt-bimetallic alloy surfaces. Nature Materials 2007, 6, 241-

247.

(61) Xiong, L.; Manthiram, A. Nanostructured Pt-M/C (M = Fe and Co)

catalysts prepared by a microemulsion method for oxygen reduction in proton

exchange membrane fuel cells. Electrochimica Acta 2005, 50, 2323-2329.

(62) Xiong, L.; Manthiram, A. Effect of atomic ordering on the catalytic

activity of carbon supported PtM (M = Fe, Co, Ni, and Cu) alloys for oxygen

25

reduction in PEMFCs. Journal of the Electrochemical Society 2005, 152, A697-

A703.

(63) Wakabayashi, N.; Takeichi, M.; Uchida, H.; Watanabe, M.

Temperature Dependence of Oxygen Reduction Activity at Pt-Fe, Pt-Co, and Pt-Ni

Alloy Electrodes. Journal of Physical Chemistry B 2005, 109, 5836-5841.

(64) Yano, H.; Kataoka, M.; Yamashita, H.; Uchida, H.; Watanabe, M.

Oxygen Reduction Activity of Carbon-Supported Pt-M (M = V, Ni, Cr, Co, and Fe)

Alloys Prepared by Nanocapsule Method. Langmuir 2007, 23, 6438-6445.

(65) Ye, H.; Crooks, R. M. Effect of Elemental Composition of PtPd

Bimetallic Nanoparticles Containing an Average of 180 Atoms on the Kinetics of the

Electrochemical Oxygen Reduction Reaction. Journal of the American Chemical

Society 2007, 129, 3627-3633.

(66) Koh, S.; Strasser, P. Electrocatalysis on Bimetallic Surfaces:

Modifying Catalytic Reactivity for Oxygen Reduction by Voltammetric Surface

Dealloying. Journal of the American Chemical Society 2007, 129, 12624-12625.

(67) Mani, P.; Srivastava, R.; Strasser, P. Dealloyed Pt-Cu Core-Shell

Nanoparticle Electrocatalysts for Use in PEM Fuel Cell Cathodes. Journal of


(68) Antolini, E.; Salgado, J. R. C.; Gonzalez, E. R. The stability of Pt-M

(M=first row transition metal) alloy catalysts and its effect on the activity in low

temperature fuel cells. Journal of Power Sources 2006, 160, 957-968.

(69) Yu, X.; Ye, S. Recent advances in activity and durability enhancement

of Pt/C catalytic cathode in PEMFC. Journal of Power Sources 2007, 172, 145-154.

(70) Shao, Y.; Yin, G.; Gao, Y.; Shi, P. Durability study of Pt/C and

Pt/CNTs catalysts under simulated PEM fuel cell conditions. Journal of the

Electrochemical Society 2006, 153, A1093-A1097.

(71) Coloma, F.; Sepulveda-Escribano, A.; Rodriquez-Reinoso, F. Heat-

treated carbon blacks as supports for platinum catalysts. Journal of Catalysis 1995,

154, 299-305.

26



(73) Shao, Y.; Yin, G.; Gao, Y. Understanding and approaches for the

durability issues of Pt-based catalysts for PEM fuel cell. Journal of Power Sources

2007, 171, 558-566.

27

Chapter 2

Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with

Supercritical Carbon Dioxide

Gold nanocrystal dispersions in toluene-CO2 mixtures were infused into cylindrical

pores in mesoporous silica to achieve high loadings over 2 wt % in 24 hours. The

nanocrystals were highly dispersed according to transmission electron microscopy

and the loadings approached equilibrium. In contrast, the loadings were small for

infusion with pure toluene or toluene mixed with an antisolvent, methanol. The

differences in loading were correlated with the long-ranged van der Waals forces

between the gold and silica through the intervening solvent. These van der Waals

forces became stronger as CO2 was added to toluene, as a consequence of a reduction

in the Hamaker constant of the mixed intervening solvent, resulting in stronger

nanocrystal adsorption. The decoupling of the nanocrystal synthesis step and the

infusion step, leads to exquisite control of the nanocrystal size, morphology and

dispersibility within the pores. The simplicity of the method allows for the facile

production of nanocrystal/silica composites for applications such as catalysis and

optoelectronics.

Large portions of this chapter have been previously published as Gupta, G.; Shah, P. S.; Zhang, X.; Saunders, A. E.; Korgel, B. A.; Johnston, K. P. Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with Supercritical Carbon Dioxide. Chemistry of Materials 2005, 17, 6728-6738.

28

2.1 Introduction

Decreasing the size of metals and semiconductors leads to unique size

dependent electronic, optical and catalytic properties, which deviate from the bulk.1-4

One of the key challenges is to control the arrangement of nanocrystals in various

substrates including block copolymer patterned templates,5 nanoscale features

patterned by e-beam lithography6 and nanoporous materials7 as described in Table

2.1. Reactive precursors may be deposited into a porous material from either the gas

phase or via a wetting liquid in order to synthesize nanocrystals with the pores.8-13

Here, the size and the shape of the pores may be used as a template to control the

particle size and shape, to form spherical particles10 or in the case of cylindrical pores,

nanowires.9,11,12,14 The diffusion of precursors and nanocrystals through the solvent

may be undesirably slow for mesopores. It took about 3 weeks for HAuCl4 aqueous

precursor solution to completely infiltrate mesoporous silica,13 perhaps indicating

diffusional resistances of entrapped air in regions of variable curvature.15

Relatively few studies have examined infusion of pre-synthesized

nanocrystals into porous materials with carrier solvents. Such nanocrystals have been

imbibed into the pores by capillary forces, in conjunction with additional driving

forces including vacuum induction and electrophoresis.16 With this technique, it can

be challenging to fill the pores uniformly. Nanocrystals synthesized in the aqueous

core of the reverse micelles have subsequently been infused into nanoporous

matrices.17,18 The ability to isolate the particles from the unreacted precursors and to

keep the particles coated with surfactant can be challenging in these micelles. Yet

another approach is to entrap pre-synthesized nanocrystals during the synthesis of a

nanoporous matrix, although the nanocrystals can in some cases perturb the geometry

of the matrix.7,19

There are two primary challenges in the infusion of pre-synthesized

nanocrystals into a nanoporous matrix. First, the nanocrystal dispersion must be

transported through the pores by favorable capillary wetting, and the nanocrystals

29

Nanoporous Material

Material impreg-nated

Pore diameter

(nm)

Nano- particle

Size (nm)Loading Method of

loading Reference

FSM-16 Pt 2.4 2.3±0.1 0.9 wt % Reduction of precursor

8

FSM-16 Pt 3.5 2.9±0.16 2.9 wt % Reduction of precursor

8

Thiol modified MCM41

CdS 3.8 3.3 -

Impregnation of

nanocrystals from reverse

micelle

18

Planar-like monolithic Au 5.7 5.2 -

Reduction of precursor by sonification

13

HMM-2 Au d100=2.2 3.2±0.5 2 wt % Reduction of Precursor

72

SBA-15 Pt d100= 10.23 2-5

1017

particle/gm

silica

Silica synthesis in presence of nanocrystals

62

Porous Alumina

membrane Au 50 15 -

Impregnation of

nanocrystals by vacuum induction

16

SBA-15 Au 8.8 2.2±0.3 2.3 wt %Impregnation

of nanocrystals

♦

MCM-41 Au 2.9 2.2±0.3 2.0 wt % -do- ♦ MCM-41 Au 2.4 2.2±0.3 1.7 wt % -do- ♦ * Time and temperature are only shown for the infusion of either the precursor or the pre-synthesized silica. ♦ shows results for our work.

Table 2.1

Methods for loading of metal nanoparticles into mesoporous silica.

30

must not block small pore entrances. Once the nanocrystals penetrate the pores, the

second challenge is to achieve sufficient adsorption of the nanocrystals from the

solution phase to the substrate surface. A major dilemma must be addressed. The

solvent must allow for enough attraction between the nanocrystals and substrate to

achieve sufficient adsorption, while simultaneously preventing flocculation of

nanocrystals in the solvent phase. The nanocrystals must be dispersed in “good”

solvent for the stabilizing ligands to prevent flocculation. However, good solvents

may be expected to cause the interactions between the nanocrystals and the surface of

the nanoporous material to be repulsive or weakly attractive.20 The resulting weak

adsorption of the particles inside the pores may produce low loadings. Thus, the

solvent must be tuned to provide a judicious balance between nanocrystal-nanocrystal

versus nanocrystal-substrate interactions.

The penetration of reactive precursors into high aspect ratio features21 and

nanoporous materials such as activated carbon22 is particularly efficient with

supercritical fluids, due to high diffusion coefficients and low viscosities. Examples

include the synthesis of porous silica fibers23 and metal oxide fibers24 within activated

carbon fiber templates. Silicon and germanium nanowires have been synthesized in

mesoporous silica from reactive precursors using sc-hexane and sc-CO2,

respectively.25,26 High diffusivities in Sc-CO2 facilitate transport of the precursors

into the pores. The miscibility between CO2 and air or organic solvents often leads to

favorable contact angles and wetting, as in the case of CO2 enhanced oil recovery in

small sandstone pores27 and in the cleaning of low k nanoporous dielectric insulators. 28

CO2 is used extensively to swell organic solvents to lower the solvent strength

to precipitate polymers from solution29 and to form nanoparticles of many types of

materials30-33 including pharmaceuticals. The influence of CO2 as an antisolvent on

the solubility of solutes34 and on spectral shifts of solvatochromic probes35,36 has been

determined over a wide range in CO2 concentration. CO2-expanded liquids (CXLs)

have been exploited for homogeneous catalytic oxidations37,38 and for hydrogenation

31

and hydroformylation reactions.39,40 The presence of CO2 in the mixed medium

increases the solubility of O2 by 100 times and also improves diffusion rates. Tunable

properties of CXLs has been utilized for the rapid and precise size selection of the

nanocrystals.41

Herein, we show that the use of supercritical CO2 as an antisolvent helps

overcome both transport and thermodynamic challenges in nanoparticle infusion into

porous materials. Our objective is to add CO2 to an organic solvent, toluene, to

achieve high levels of infusion of pre-synthesized nanocrystals into mesoporous silica

within a 24 hour period. The CO2 concentration in toluene is tuned to produce

sufficiently strong nanocrystal-silica interactions for high nanocrystal adsorption in

the pores without flocculating the nanocrystals in the solvent phase. At 35°C and

pressures of this study above 68 bar,42 CO2 forms a homogeneous one-phase solution

with toluene. Dodecanethiol capped metal nanocrystals cannot be dispersed in pure

CO2;4 however they are soluble in our CO2-toluene mixtures. The loading of the gold

particles was determined by measuring spectrophotometrically, the reduction in the

absorbance of gold in the supernatant after separation from the mesoporous silica

phase. As a control, the infusion was also studied (1) without CO2 present and (2)

with methanol as an antisolvent in toluene. The diameters of the porous cylinders in

the mesoporous silica were varied from 2.5 nm to 9 nm. The kinetics of infusion was

studied over a period of 3 to 72 hours, both with and without CO2, to understand the

effect of changes in the transport properties of the nanocrystals. The effects of CO2 on

both the kinetics and the final loading are analyzed relative to non-expanded liquid

solvents, toluene or toluene-methanol mixtures. In the first part of the discussion

section, we analyze the equilibrium loading and then examine the kinetic aspects in

the second section.

A key advantage of impregnating pre-synthesized nanocrystals into pores is

that the nanocrystal synthesis may be controlled with stabilizing ligands to produce

the desired size, crystallinity, shape, and surface properties. Numerous synthetic

schemes have been reported.2 The nanocrystal morphology may be chosen

32

independently of the pore morphology, as long as the crystals are smaller than the

pores. Prior to infusion, the particles may be isolated from unwanted byproducts,

cleaned, separated by size, and the ligands may be exchanged if desired for

alternative ligands. It would be difficult if not impossible to execute these post-

reaction steps if the nanocrystals were synthesized inside the nanoporous material.

Finally, synthesis inside nanoporous materials can be perturbed by surface chemistry

due to the substrate, for example, capping ligands may adsorb to the substrate rather

than the nanocrystals.

2.2 Experimental

2.2.1 Nanocrystal Synthesis

Gold nanocrystals were synthesized by a two-phase (organic-water) arrested

precipitation technique,2 using dodecanethiol (C12H25SH) as a capping agent. Initially,

36 ml of aqueous (0.03 M) hydrogen tetrachloroaurate(III) trihydrate (HAuCl4.3H2O)

was combined with 25 mL of toluene solution containing 2.7 gm of phase transfer

catalyst, tetraoctylammonium bromide. After stirring for 1 h, the organic phase, with

the transferred gold, was collected. The gold salt was then reduced using 15 mL of an

aqueous sodium borohydride (NaBH4) solution (0.44 M) resulting in gold

nanocrystals dispersed in toluene, protected by the phase transfer catalyst. To this

solution, 240 mL of dodecanethiol was added immediately and stirred over night. The

organic phase rich in gold nanocrystals was collected. Methanol was added to the

organic phase and the solution was centrifuged to separate the nanocrystals. The

nanocrystals were washed and redispersed in 20 mL of toluene.

2.2.2 Mesoporous Silica Synthesis

Mesoporous silica SBA-15 was synthesized via a procedure similar to that

reported by Zhao et al.43 A solution of 1.6 g EO20PO70EO20 : 7 mL HCl: 8 mL TEOS:

38 mL H2O was prepared, stirred for 24 h at 50°C and then heated at 100°C for 24 h.

33

The resulting slightly yellow powder was dispersed in ethanol, filtered and then again

heated at 100°C for 24 h. The yellow product was then calcined at 550°C.

Mesoporous silica MCM-41 was synthesized by S+X-I+ assembly via a

procedure similar to that reported by Tanev et al.44 CnH2n+1N(CH3)3Br template

(n=10, 16) was used as a structure directing agent. Depending on the chain length of

the surfactant, MCM-41 silica with different pore diameters was obtained. Small pore

diameter silica (S-MCM41) was obtained with surfactant chain length n=10 while

large pore diameter silica (L-MCM41) was obtained with chain length n=16.

For the synthesis of S-MCM41, a solution of 0.8 g C10H21N(CH3)3Br: 7 mL

HCl: 5 mL TEOS: 38 mL H2O was prepared, stirred overnight at ambient temperature

and then heated at 100°C for 6 h. The ratio of the components in the solution is

different than that reported by Tanev et al.44 The while powder was then dispersed in

ethanol, filtered and dried at 100°C for 24 hours. The collected solid was then

calcined at 600°C. The L-MCM41 was synthesized similarly to S-MCM41 except

that a solution with composition 1.6 g C16H33N(CH3)3Br: 7 mL HCl: 8 mL TEOS: 38

mL H2O was used.

2.2.3 Infusion Procedure

In a typical experiment, 20 mg of mesoporous silica and 2.5 mL of gold nanocrystals

dispersed in toluene, with a concentration of 0.3 mg/ml, were loaded in the reaction

cell with a working volume (volume with piston in the cell) of 27 mL. 0-5 mL of

toluene was added in the cell. The cell was then sealed and then an amount of CO2

equivalent to 2.5 – 7.5 mL at 241 bar (3500 psi) and 35°C was added so as to produce

a total volume of reaction mixture of 10 mL. The cell was then pressurized to 241 bar

with CO2 at the backside of the piston using an ISCO model 100 DX Computer

Controlled syringe pump. The contents of the cell, in a water bath, were stirred using

a magnetic stir bar that was inserted inside the cell (Figure 2.1). The reaction was

maintained at 35°C and 241 bar (3500 psia) and allowed to proceed for 24 h. After

the reaction, the cell was cooled to 0°C and CO2 was then vented as a vapor from the

34

Figure 2.1

Schematic of reaction infusion apparatus.

35

top of the cell, leaving most of the other components in the cell. Lowering the

temperature to 0°C reduces the loss of toluene vapor along with CO2. The supernatant

was collected and filtered, using a polycarbonate 0.05 μm filter, to separate the

mesoporous silica. The volume of the supernatant was found to be slightly less than

the total volume of gold feed solution and toluene added to the cell because of some

toluene loss along with CO2. The loss of toluene was usually about 5-6 vol%. The

extent of incorporation of gold nanocrystals in the silica was determined by

subtracting from the initial mass of the nanocrystals, the final mass of the

nanocrystals recovered in the supernatant. The absorbance of the gold nanocrystals

dispersed in toluene before and after infusion was measured at a wavelength of 500

nm using a Cary 500 UV-Vis-NIR spectrophotometer to determine the nanocrystal

mass. An optical path length of 1 cm was used. A standard calibration curve was

generated at 500 nm using known concentrations of gold nanocrystals in toluene as

shown in Figure 2.2. A typical starting gold concentration was 0.3 mg/ml and after

infusion the value decreased to 0.195 mg/ml for a loading of 2 wt % and to 0.17

mg/ml for a loading of 2.23 wt %. The experiments were repeated several times to

check for the reproducibility, which was on the order of +0.3 in wt % units.

2.3 Results

2.3.1 Nanoparticle Characterization

SAXS was used to characterize mesoporous silica structure as well as the gold

nanocrystal dispersions. The intensity of scattered radiation, I(q), is proportional to a

shape factor and structure factor, P(q) and S(q), respectively: )()()( qSqPqI ∝ . Here,

q is the wave vector, defined as q = 4π/λ sin(θ), where λ is the wavelength (0.154

nm) and 2θ is the scattering angle. The wave vector is also inversely proportional to a

characteristic distance, d, in the scattering system: q = 2π/d. The shape factor is

related to the size and shape of the scatterers, whereas the structure factor contains

information about interactions or ordering within the system.

36

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Abs

orba

nce

at 5

00 n

m

gold concentration (mg/ml)

Figure 2.2

Calibration curve of absorbance versus concentration measured at 500 nm for gold

nanocrystal solutions.

37

For dilute solutions of non-interacting nanocrystals S(q) = 1, and the scattered

intensity relates only the shape factor of the particles dispersed in solution , i.e. I(q) ∝

P(q).48-51 The shape factor for spherical particles is given analytically

by52,53 ( )[ ]23)/()cos()sin(3)( qRqRqRqRqP −= , where R is the nanocrystal radius.

The effects of sample polydispersity on the scattered intensity can be taken into

account using ∫∝ RRqPRNqI d)()()( 6 , where N(R) is a normalized size

distribution.48,54 Here the size distribution is assumed to be Gaussian, where Ravg and

σ are the average radius and standard deviation, respectively. Fitting this model to the

scattering data from the toluene dispersion of gold nanocrystals (Figure 2.3) indicates

that the gold nanocrystals were relatively monodisperse with a mean diameter of

2.2±0.3 nm. The curve in Figure 2.3 does not show any peaks, indicating weak

interactions between the particles due to the good solvation of the ligands, as seen

previously for nanocrystals in pure organic solvents.54 In contrast, strong attractive

interactions of perfluoropolyether coated gold nanoparticles have been observed with

SAXS in CO2 due to the stronger Hamaker interactions between the cores, and the

limited solvation of the ligands, both resulting from the weak solvent strength.55

2.3.2 Characterization of Mesoporous Silica

SAXS was used to characterize the mesoporous silica powders prior to

infusion. Highly oriented pores, arranged in a 2-D hexagonal lattice, give rise to a

series of Bragg peaks in the scattering pattern. Figure 2.4 and Table 2.2 give the small

angle XRD pattern for three mesoporous silicas, SBA-15 (a), L-MCM41 (b) and S-

MCM41 (c). For SBA-15, the four well-resolved peaks are indexable as the (100),

(110), (200), and (210) reflections associated with p6mm hexagonal symmetry. One

additional weak peak is found at q(nm-1) =1.82 and corresponds to a (300) reflection.

Presence of multiple sharp peaks indicates that calcined SBA-15 has a high degree of

hexagonal mesoscopic organization. The first three peaks for SBA-15 are located at

38

0.01

0.1

1

10

0 0.5 1 1.5 2 2.5 3 3.5

I(q)

q (nm-1)

Figure 2.3

Intensity I(q) vs q for gold nanocrystals dispersed in toluene determined by SAXS.

39

10

100

1000

104

105

106

107

0 0.5 1 1.5 2 2.5 3 3.5 4

I(q)

q (nm -1)

a

b

c

100

110200

210

100 110

100

Figure 2.4

SAXS for calcined mesoporous silica SBA-15 (a), L-MCM41 (b) and S-MCM41 (c).

40

* Average length of the pore was considered 10 microns for all three mesoporous silica Table 2.2

Properties of mesoporous silica.

41

q100 = 0.61, q110 = 1.06, q200 = 1.22 nm-1 with the peak position ratio of 1: 3 : 2,

which is consistent with the literature.56-58 The distance d100 between adjacent planes

is given by d100 = 2π/ q100 = 10.29 nm. The unit cell parameter a0 = 3/2 100d = 11.9

nm gives the center-to-center spacings of the pores and agrees fairly with the

literature.43

L-MCM41 shows two indexable peaks corresponding to (100) and (110)

reflections. The peaks are broad and show diffuse scattering centered at q = 1.63 nm-1

corresponding to reflection from the (100) plane, which manifests a lower degree of

order as compared to SBA-15. The distance d100 between adjacent planes is 3.85 nm

and the unit cell parameter a0 = 4.44 nm for L-MCM41. S-MCM41 exhibits single

d100 reflections accompanied by a diffuse scattering centered at ~ 2.95 nm with a unit

cell parameter of 3.4 nm. Pinnavaia et al. have demonstrated that “single-reflection”

MCM-41 still can have short ranged hexagonal symmetry.44,59 They show that MCM-

41 synthesized by acidic S+X-I+ pathway with surfactant chain length <14 result in

poorly ordered products. For S-MCM41, the lower degree of order can be attributed

to the small chain length of only 10 for the structure directing agent.

Figure 2.5 gives the adsorption-desorption isotherms for the three calcined

silica samples. Representative nitrogen adsorption/desorption isotherms and

corresponding pore size distribution (BJH method) are shown in Fig 2.5(a). SBA-15

had a mean pore diameter of 8.8 nm, BET surface area of 700 m2/g and a pore

volume of 1.09 cc/gm (refer Table 2.2), consistent with the literature.43 Three well

distinguished regions of the adsorption-isotherm are evident: (i) monolayer-multilayer

adsorption, (ii) capillary condensation and (iii) multilayer adsorption on the outer

particle surfaces. The approximate pore size calculated using the BJH analysis is

significantly smaller than the center-to-center distance between adjacent pores

determined by SAXS, because the later includes the thickness of the pore wall. From

the difference in these two values, the thickness of the pore wall is estimated to be 3.1

nm for SBA-15 which is in excellent agreement with the values reported by Zhao et

42

0

100

200

300

400

500

600

700

800

0 0.2 0.4 0.6 0.8 1

Vol

(cc/

g)

P/Po

Desorption

Adsorption

0

0.005

0.01

0.015

0.02

0.025

0 50 100 150 200

Dv(

r) [c

c Å

-1 g

-1]

Pore Diameter (Å) (a)

0

50

100

150

200

250

300

350

400

0 0.2 0.4 0.6 0.8 1

Vol

(cc/

g)

P/Po

Adsorption

Desorption

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

15 20 25 30 35 40

Pore

Vol

ume

(cc

Å1

g1 )

Pore Diameter (Å) (b)

140

160

180

200

220

240

260

0 0.2 0.4 0.6 0.8 1

Vol

(cc/

g)

P/Po

Adsorption

Desorption

0

0.005

0.01

0.015

0.02

0.025

0 10 20 30 40 50

Pore

Vol

ume

(cc

Å-1

g-1)

Pore Diameter (Å) (c)

Figure 2.5

N2 isotherms and pore size for SBA-15 (a), L-MCM41 (b) and S-MCM41 (c).

43

al.43 The adsorption-desorption isotherm for L-MCM41 and S-MCM41 and their

corresponding pore size distributions, determined by SF method,46 are shown in

Figures 2.5(b) and Figure 2.5(c) respectively and Table 2.2.

As can be seen in Figure 2.5(a), SBA-15 shows a type-H1 hysteresis loop with

capillary condensation at a relative pressure of P/Po= 0.7.43 The step in the hysteresis

loop is very steep with a jump from 300 cc/gm to 700 cc/gm due to capillary

condensation. The slope and the height of the step are clear indications of very well

defined mesopores with narrow pore size distribtution,44 which is further

substantiated by the multiple peaks arising in SAXS (Figure 2.4).

MCM-41 (Figures 2.5(b) and 2.5(c)) shows a type 4 adsorption isotherm44

with a small degree of hysteresis. This behavior is typical of MCM-41.44 The pore

sizes of 2.5 and 2.9 nm are smaller as compared to that of SBA-15. The adsorption

increases gradually with an increase of P/Po without a steep jump, which suggests that

the pores have a wider pore size distribution than for SBA-15. The surface areas

obtained for L-MCM41 and S-MCM41 were 825 and 738 m2/g, respectively, and are

typical for MCM-41 mesoporous silica. The adsorption volume, for P/Po approaching

unity, becomes the pore volume due to complete condensation of N2. S-MCM41 to

SBA-15 (refer Table 2.2) resulting in increase of the adsorption volumes at STP.

2.3.3 Nanoparticle Dispersibility in Toluene-CO2

To date only fluorinated ligands have been shown to disperse nanocrystals in

pure CO2.3,4,60,61 However, we show that standard alkane capped nanocrystals may be

dispersed in the toluene-CO2 co-solvent system. Nanocrystal dispersions were

analyzed with a Cary 3E UV-Vis spectrophotometer in a vertically mounted variable-

volume high-pressure optical cell with a path length of 2 cm. The temperature was

controlled at 35°C with heating tape and the pressure was 241 bar. At a concentration

of CO2 in the solution of 50 vol%, the absorbance at 500 nm remained 1.92±0.02 for

a period of 8 hrs indicating a stable dispersion.

44

2.3.4 Effect of CO2 Concentration on Loading of Gold Nanocrystals in

Mesoporous Silica

Figure 2.6 shows the loadings of the nanocrystals in mesoporous silica as a

function of CO2 concentration. For nanocrystals dispersed in pure toluene, there is

virtually no infusion into the pores, as loadings were less than 0.1 wt %. The loadings

increase linearly with CO2 concentrations of 25% and 50% and then level off at the

highest value of 75%. It is not possible to go to higher CO2 concentrations, as the

particles become unstable and start to precipitate, indicating insufficient solvation of

the hydrocarbon capping ligands.

To investigate the effect of hydrostatic pressure on trapped air in the pores,

infusion was conducted in pure toluene at 241 bar, in absence of carbon dioxide. Air

may potentially be trapped between two liquid domains at both ends of a cylindrical

pore. A very small loading of 0.3 wt % was observed indicating that trapped air had

little effect. The pores in the silica are interconnected to each other via small diameter

channels. Perhaps the high pressure forced some of the particles into these

interconnects between pores.15

For CO2-toluene mixtures, the relatively large loadings above 2% comparable

to those obtained by the other infusion techniques in Table 2.1. We estimate, based on

the molecular weight of the nanocrystals (50,000 gm/mole of nanocrystals), a loading

of about 1017-1018 nanocrystals per gm of silica, for each of our types of silica. A

rough estimate of the order of number of pores per gm of silica (N) can be made from

the specific surface area. For a surface area (S) of 1000 m2/g and a radius (r) of 1 nm

and pore length (l) of 10 microns (assuming that the pore length is equal to the

particle size), we estimate from N=S/(2πrl) that there are about 1016 -1017 pores per

gram of silica. According to this estimate, our loading was about 10-100 particles per

pore on average. Konya et al62 obtained 1016 particles per gm of mesoporous silica,

roughly equal to 2 particles per pore, which corresponds to 1016 pores per gram,

similar to our value. It is important to note that we have obtained about 10 times

45

0

0.5

1

1.5

2

2.5

3

0 25 50 75 100Wt%

Gol

d Lo

adin

g (m

g A

u/m

g Si

lica)

Vol% CO2

Figure 2.6

Loading of gold nanocrystals versus vol% CO2 for 24 hour impregnation period in L-

MCM41. () 1 atm and 23°C ;() 241 atm and 35°C.

46

higher loadings than Konya et al.62 At higher loadings Konya et al62 observed that the

crystallinity of the silica decreased and worm like channels appeared due to the

presence of the nanocrystals.

2.3.5 Effect of Pore Size and Concentration of Nanocrystals on the Loading

Figure 2.7 shows that the loadings of nanocrystals in the S-MCM41, L-

MCM41 and SBA-15 silicas from pure toluene dispersions were negligible. However,

with the addition of 50% CO2 the loadings were large in all cases and increased

modestly with the pore diameter from 1.7% to 2.4%. The loading of the 2.2 nm

particles into the silica with 2.4 nm pores was well over half the loading for silica

with 8.8 nm pores. Thus the diameters of the pores along the pore length appeared to

be relatively uniform such that the particles did not block many narrow regions. The

surface areas were similar for the three types of silica. The concentration of gold in

the feed solution was increased ten fold for L-MCM41 to 3 mg/ml, as shown in

Figure 2.7. The loading increased a moderate amount from 1.97% to 2.37%.

2.3.6 Kinetic Study of the Infusion Process

For pure toluene, the nanocrystal loadings were negligible even after 7 days.

Detailed kinetic studies were performed for 50% CO2 with the silica with the smallest

pores, S-MCM41. Here the mean diameter of the gold nanocrystals was 1.95 nm with

a standard deviation/mean of 14% in contrast with the somewhat larger nanocrystals

for our other systems. The loadings increased with time but appeared to reach an

asymptote at 72 hours suggesting an approach to equilibrium (Figure 2.8). Over the

first 24 hours, where most of the infusion occurred, the loading was linear in t1/2 as

shown in Figure 2.8(a).

47

0

0.5

1

1.5

2

2.5

3

0 2 4 6 8 10

Wt%

Gol

d L

oadi

ng

Diameter of silica pore (nm)

Figure 2.7

Variation of loading of gold nanocrystals with the pore diameter of silica. The hollow

squares () are for loadings at 50 vol% CO2, solid circles () for 0 vol% CO2 and ♦

for loading at 50 vol% CO2 with the feed at 10 times higher concentration.

48

(a)

0

0.5

1

1.5

2

2.5

3

3.5

0 10 20 30 40 50 60 70 80

Wt%

Gol

d Lo

adin

g

Time (hr)

(b)

Figure 2.8

Loading versus time for S-MCM41. (a) Loading versus t1/2 for up to 24 hours with a

correlation coefficient of 99.5%. (b) Loading versus time approaching equilibrium.

0

0.5

1

1.5

2

2.5

0 1 2 3 4 5

y = 0.45826x R= 0.99541

Wt%

Gol

d L

oadi

ng

sqrt (t) (hr0.5)

49

2.3.7 TEM Micrographs of Infused Silica

Figure 2.9 shows typical images of gold nanocrystals infused in the three

types of silica. The pore sizes and structures appear to be uniform. In each case, large

numbers of highly dispersed nanocrystals are present in each pore, consistent with the

estimate of about 100 particles/pore above. The presence of large numbers of

nanocrystals throughout each pore corroborates the relatively high loadings measured

by spectroscopy in Table 2.1. The presence of large number of particles in the pores

was most likely for SBA-15 where the pore diameter is more than 4 times the

diameter of the nanocrystals. Hence, the adsorption of a nanocrystal in the pore would

be less likely to block the entry of a second nanocrystal. However, in the case of L-

MCM41 and S-MCM41, where the ratio of the diameter of pore to the diameter of the

particle is less than 2, the chances of pore blocking were much larger. Thus, the

presence of large numbers of particles in each pore suggests that the adsorbed

particles are mobile and can move toward the center of the pore by diffusion on the

surface and/or in the bulk.

2.3.8 Effect of a Conventional Liquid Anti-Solvent on Infusion

The effect of a liquid antisolvent, methanol, on the loading was used as a

control experiment to shed further insight into the above experiments for the

supercritical antisolvent, CO2. Polar solvents are utilized as antisolvents for size

selective precipitation of alkanethiol coated nanocrystals.2 The addition of methanol

may influence the nanocrystal-silica and -solvent interactions, but will not lower the

viscosity of the solvent as is the case for CO2. It was observed visually that gold

nanocrystals at a concentration of 0.3 mg/ml dispersed in 50% methanol

concentrations were stable for at least 24 hours. At a level of 55% methanol-45%

toluene, the dispersions became unstable with a visible settling front in less than one

hour. In order to adsorb nanocrystals in the pores without aggregation in the solvent

mixture, we chose a methanol concentration of 50% for the infusion experiments.

50

(a)

(b)

(c)

Figure 2.9

TEM of gold impregnated SBA-15 (a), L-MCM41 (b) and S-MCM41 (c).

51

An amount of 20 mg of S-MCM41 was added to a 10ml stirred solution of 1.95 nm

gold nanocrystals (0.4 mg/ml). After 24 hours, the loading was 0.56±0.50 %. After 7

days it reached 0.80±0.18 %, suggesting that the differences for methanol and CO2

are not due simply to kinetic factors. These two loadings are within the experimental

uncertainty indicating that the system approached equilibrium within 24 hours. These

loadings are less than half of the values obtained with CO2.

2.4 Discussion

We begin by demonstrating that toluene wets the pores rapidly, and thus

capillary wetting does not limit the infusion of the nanocrystals. The capillary forces

become substantial for the cylindrical nanopores as described by the Laplace pressure

RP LV θγ cos2

=Δ (2.1)

where γLV is the surface tension of the liquid, R is the pore radius, and θ is the contact

angle between the surface of the pore and the liquid. Toluene with a surface tension

of 28 dynes/cm, wets silica with a contact angle of about 6º.20 The Laplace pressure

for wetting of a 4 nm pore for a toluene-silica-air interface of 278 bars provides a

strong driving force for capillary wetting. On the basis of the conservative assumption

that the initial air in the pores does not dissolve in toluene (if toluene is already pre-

saturated with air), the air will be compressed from 1 bar to 278 bar and occupy a

negligible volume, even if it is not displaced through one of the pore ends. Thus

almost all of the pore volume will be available to the toluene. The capillary wetting

by each of the other solvent mixtures in this study will also be favorable, as the

contact angles will be well below 90o.

The rate at which the liquid solvent penetrates the pores can be estimated

using the Washburn equation, (combination of Laplace and Poiseuille equations)63

hR

dtdh

ηθγ

4cos

= (2.2)

52

where η is the bulk viscosity. For cylindrical pores that extend the length of the

particle, h is on the order of 20 μm. The time to fill a 4 nm diameter pore, is on the

order of the order of 10-4 seconds. Thus, the diffusion and adsorption of the

nanocrystals in the pores required several orders of magnitude more time than solvent

infusion.

2.4.1 Equilibrium Nanocrystal Loading

In each of the solvents, pure toluene, toluene-CO2, and toluene-methanol, the

nanocrystal loading did not increase significantly after 24 hours, indicating an

approach to equilibrium. For toluene-CO2, the similar loadings for the three types of

mesoporous silica are consistent with the similar surface areas, despite the difference

in pore diameters. The difference in pore diameter did not appear to cause an

appreciable difference in the loading kinetics. The similar loadings for the similar

surface areas suggest that the loadings approached equilibrium. The equilibrium

adsorption of a nanocrystal in a pore onto the silica surface may be expected to be

driven primarily by the chemical potential of the nanocrystals in the solvent and long-

ranged van der Waals forces between the particles and silica surface, since the ligands

are nonpolar. For simplicity, we treat the silica surface as a flat wall. The Hamaker

constant A132, for a gold particle (1) and silica surface (2) across a medium (toluene-

CO2 mixture, 3) may be approximated by20

))(( 33223311132 AAAAA −−= (2.3)

where Aii is the Hamaker constant for pure i, which can be estimated using a

simplification of the Lifshift theory.20

( )( ) 2/322

2222

2163

43

vacuumy

vacuumye

vacuumy

vacuumybyy

nn

nnhTkA

+

−+⎟

⎟⎠

⎞⎜⎜⎝

⎛

+

−=

νεεεε

(2.4)

53

where eν is the maximum electronic ultraviolet adsorption frequency, typically

assumed to be 15103× s-1.4 The A for gold, 19104 −× J,20 the ε and n for silica, 3.91 and

1.45, respectively64,65 and for toluene, 2.38 and 1.496, respectively66,67 were taken

from the literature. The Hamaker constants for silica and toluene calculated from

equation (2.2) are 201097.5 −× J and 201098.6 −× J, respectively.

The A33 was determined for the binary solvent mixture from equation 2.4

along with the Lorenz-Lorenz mixing rule,

∑ +−=+− 2222 )2/()1()2/()1( iiimixmix nnnn φ (2.5)

where φi is the volume fraction of component i. An equivalent expression for the

dielectric constant is obtained by replacing n with ε1/2 in equation 2.5. The n and ε for

CO2 were obtained as a function of density from the following.68

34242

2

10171.310412.107016.021

rrrnn ρρρ −− ×−×+=

+− (2.6)

202602.02386.01 rr ρρε +=− (2.7)

where /r cρ ρ ρ= . At the reaction conditions of 241 Bar and 35°C, the n and ε for

CO2 were determined to be 1.2 and 1.12, respectively. The A for the pure CO2 was

calculated from equation 2.2 to be 201044.1 −× J. The results for A33 are summarized

in the Table 2.3.

The Hamaker constant for the gold-silica interaction across pure toluene, A132, turns

out to be negative, indicating an unfavorable driving force for adsorption of the

54

Vol. % CO2

A33 × 1020 (J) A132 × 1020 (J)

100 1.44 6.38

75 2.81 3.57

50 4.19 1.70

25 5.56 0.34

0 6.93 -0.31

Table 2.3

Overall Hamaker constants A132 between silica surface and gold core mediated by

supercritical CO2-toluene solution. Component 1 is gold, 2 is SiO2, and 3 is a mixture

of toluene and CO2.

55

nanocrystals. The very low loading of nanocrystals from pure toluene is consistent

with this result. With the addition of CO2, A132 becomes increasingly positive, which

would lead to an increase in the adsorption as observed. The loadings are correlated

with the particle-silica van der Waals forces as characterized by A132 as listed in

Table 2.3, which becomes more attractive with an increase in CO2 concentration. In

addition, the solvation of the nanocrystals in the solvent phase decreases with added

CO2 increasing the chemical potential of the nanocrystals. Both of these factors will

produce stronger adsorption resulting in higher loadings, as observed.

This type of analysis was also being applied to methanol as an antisolvent. For

20% methanol, A132 becomes positive, indicating attractive van der Waals forces

(Table 2.4). The A132 value with 50% methanol is much weaker than in the case of

only 25% CO2. Therefore, the much weaker adsorption for methanol as an antisolvent

relative to CO2 is consistent with the A132 values.

Pure CO2, with an exceptionally low polarizability/volume60 has a much lower

A than methanol or just about any other antisolvent. For example, CO2 produces a

greater reduction in A when add to toluene than does methanol. As the A of the

intervening solvent, A33, decreases, A132 always becomes less negative or more

positive in equation 2.1, thus favoring adsorption. Therefore, CO2 provides an

opportunity to achieve a greater range in the tunability of a mixed solvent, which in

our case plays a key role in achieving high nanocrystal loadings.

2.4.2 Transport Aspects of Nanocrystal Loading

Koone and Zerda69 reported that the diffusion coefficient of water in a

nanoporous glass of diameter 2.9 nm is of the order of 10104 −× m2/s, while the

diffusion coefficient of water in the unbound system was 9102.2 −× m2/s. A much

greater reduction in diffusion rate may be expected for the nanocrystals in the pores

due to the much higher aspect ratio or solute size to pore diameter. Vigne-Maeder et

al. performed molecular dynamics simulations of the diffusion of gases in zeolites.70

56

Vol. % methanol A33 × 1020 (J)** A132 × 1020

(J)

50

5.84

0.52

40

6.05

0.34

30

6.27

0.17

20

6.49

0.003

10

6.71

-0.16

0

6.93

-0.23

** Hamaker constant of solvent interacting with itself in vacuum.

Table 2.4

Overall Hamaker constants A132 between silica surface and gold core mediated by

methanol-toluene solution. Component 1 is gold, 2 is SiO2, and 3 is a mixture of

toluene and methanol.

57

The primary mechanism for transport was adsorption followed by surface diffusion to

a pore mouth, rather than direct entrance to the pore. The importance of surface

diffusion increased as the aspect ratio of particle diameter to pore diameter increased.

The nanocrystals diffuse from the entrance of the pore at both ends towards

the center along the concentration gradient. The diffusion will take place both in the

solvent within the pore and on the surface. In Figure 2.8(a), the loading of the

nanocrystals is plotted as a function of t1/2 with a correlation coefficient of 0.995.

According to the Stokes-Einstein relationship the diffusion coefficient of 2 nm gold

nanocrystals in bulk toluene is 10102 −× m2/s. However, the actual diffusion

coefficient in the pores was lower due to confinement in that it took about 24 hours to

approach the equilibrium loading as seen in figure 2.8(b). The effective diffusion

coefficient in the pores of length l, including surface diffusion,71 may be

approximated by the equation Dtl = . For 10 micron long pores in the particles, the

diffusion coefficient was calculated to be 10-15 m2/s.

When the diameter of the pore approaches that of the particle, the mass

transfer resistance for the entry of the particle into the pore increases significantly and

diffusion through the pore becomes considerably difficult. The aspect ratios, or

particle/pore diameters are 0.9, 0.76, and 0.25 for our three silicas. The slight increase

in the loadings with the pore diameter (Figure 2.7) can be attributed to the larger

pores and the increase in the structural ordering. The degree of order was shown to

increase with an increase in pore size. The presence of well-ordered uniform

cylindrical pores would lead to less pore blockage than disordered pores. As seen

from the SAXS data, the degree of long-range ordering is highest for SBA-15

followed by L-MCM41, and finally S-MCM41 with limited order. It is observed that

the Even with relatively little order in S-MCM41, the loadings are significant,

indicating that kinetic limitations were minor.

The similar loadings for the three silica supports with similar surface areas

suggest that blocking of pores by nanocrystals is not significant. In the case of S-

58

MCM41 where the aspect ratio of the particle to the pore is 0.9, surface diffusion

must have been significant to prevent blocking of the pores. For this high aspect ratio,

the polydispersity was rather low, otherwise, many smaller pores would not have

been accessible.

2.5 Conclusions

A novel approach is presented to infuse pre-synthesized gold nanocrystals into pre-

formed mesoporous silica by tuning the solvent quality with CO2 as an antisolvent.

The concept of decoupling the nanocrystal synthesis step and the infusion step,

provides exquisite control of the nanocrystal size, morphology and dispersibility

within the pores, without perturbing the morphology of the mesoporous silica.

High loadings of nanocrystals in mesoporous silica over 2 wt % were obtained

in carbon dioxide-toluene mixtures in 24 hours. It is estimated that roughly 10-100

nanocrystals were infused per pore inside the mesoporous silica. The loading

approached equilibrium and the individual nanocrystals were highly dispersed

according to TEM. In contrast, the loadings were small with pure toluene or toluene

mixed with the antisolvent methanol. The differences in loading were correlated with

the long-ranged van der Waals forces between the gold and silica through the

intervening solvent, as described by the Hamaker constant, A132. As CO2 has an

unusually low polarizability/density (very weak VdW forces), it may be added to a

liquid solvent to form a mixture with an unusually low Hamaker constant. The

weaker vdW forces for the intervening solvent result in a stronger interaction between

the gold and silica and a higher chemical potential for the gold in the mixed solvent.

Both factors raise the nanocrystal adsorption on the surface and thus the equilibrium

loading. Furthermore, the loading may be tuned over a wide range by varying the CO2

composition. The difference in pore size had only a small effect on the loadings, even

when the aspect ratios reached 0.9, indicating pore blockage was minor. Thus it

59

appeared that the adsorbed nanocrystals were mobile, despite the proximity of the

nanocrystals to the surface even at the center of a pore.

The simplicity of this infusion method, in that well-known independent

syntheses may be utilized for both the nanocrystals and silica, allows for the facile

production of nanocrystal/silica composites for applications such as catalysis and

optoelectronics. The approach is general and may be applied to the wide varieties of

nanocrystals and mesoporous materials that are available or are being discovered.

2.6 References

(1) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993,

115, 8706-8715.

(2) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J.

Chem. Soc., Chem. Commun. 1994, 7, 801-802.

(3) Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. J. Phys. Chem.

B 2001, 105, 9433-9440.

(4) Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. J. Phys. Chem.

B 2002, 106, 12178-12185.

(5) Skaff, H.; Lin, Y.; Dinsmore, A.; Russell, T.; Emrick, T. Abstracts of

Papers, 224th ACS National Meeting, Boston, MA, United States, August 18-22, 2002

2002, COLL-026.

(6) Weimann, T.; Geyer, W.; Hinze, P.; Stadler, V.; Eck, W.; Golzhauser,

A. Microelectronic Engineering 2001, 57-58, 903-907.

(7) Konya, Z. P., Victor F.; Kiricsi, Imre; Zhu, Ji; Alivisatos, Paul;

Somorjai, Gabor A. Catalysis Letters 2002, 81, 137-140.

(8) Wakayama, H.; Setoyama, N.; Fukushima, Y. Adv. Mater. 2003, 15,

742-745.

(9) Han, Y.-J.; Kim, J. M.; Stucky, G. D. Chem. Mater. 2000, 12, 2068-

2069.

60

(10) Fukuoka, A.; Araki, H.; Sakamoto, Y.; Sugimoto, N.; Tsukada, H.;

Kumai, Y.; Akimoto, Y.; Ichikawa, M. Nano Letters 2002, 2, 793-795.

(11) Coleman, N. R. B.; Ryan, K. M.; Spalding, T. R.; Holmes, J. D.;

Morris, M. A. Chem. Phys. Lett. 2001, 343, 1.

(12) Coleman, N. R. B.; O'Sullivan, N.; Ryan, K. M.; Crowley, T. A.;

Morris, M. A.; Spalding, T. R.; Steytler, D. C.; Holmes, J. D. J. Am. Chem. Soc.

2001, 123, 7010.

(13) Chen, W.; Cai, W.; Zhang, L.; Wang, G.; Zhang, L. J. Colloid and

Interface Sci. 2001, 238, 291-295.

(14) Ryan, K. M.; Erts, D.; Olin, H.; Morris, M. S.; Holmes, J. D. J. Am.

Chem. Soc. 2003, 125, 6284.

(15) Lake, L. W. Enhanced Oil Recovery; Prentice Hall, PTR, 1996.

(16) Hornyak, G.; Kroll, M.; Pugin, R.; Sawitowski, T.; Schmid, G.; Bovin,

J.-O.; Karsson, G.; Hofmeister, H.; Hopfe, S. Chemistry--A European Journal 1997,

3, 1951-1956.

(17) Hirai, T.; Okubo, H.; Komasawa, I. J. Phys. Chem. B 1999, 103, 4228-

4230.

(18) Hirai, T.; Okubo, H.; Komasawa, I. J. Colloid and Interface Sci. 2001,

235, 358-364.

(19) Konya, Z. P., Victor F.; Kiricsi, Imre; Zhu, Ji; Ager, Joel W., III; Ko,

Moon Kyu; Frei, Heinz; Alivisatos, Paul; Somorjai, Gabor A. Chem. Mater. 2003, 15,

1242-1248.



(21) Watkins, J. J.; Blackburn, J. M.; McCarthy, T. J. Chemistry of

Materials 1999, 11, 213-215.

(22) Wakayama, H.; Fukushima, Y. Ind. Eng. Chem. Res. 2000, 39, 4641-

4645.

61

(23) Fukushima, Y.; Wakayama, H. J. Phys. Chem. B 1999, 103, 3062-

3064.

(24) Wakayama, H.; Itahara, H.; Tatsuda, N.; Inagaki, S.; Fukushima, Y.

Chem. Mater. 2001, 13, 2392-2396.

(25) Coleman, N. R. B.; Morris, M. A.; Spalding, T. R.; Holmes, J. D. J.

Am. Chem. Soc. 2001, 123, 187-188.

(26) Ziegler, K. J.; Polyakov, B.; Kulkarni, J. S.; Crowley, T. A.; Ryan, K.

M.; Morris, M. A.; Erts, D.; Holmes, J. D. Journal of Materials Chemistry 2004, 14,

585-589.

(27) Van Bergen, F.; Gale, J.; Damen, K. J.; Wildenborg, A. F. B. Energy

(Amsterdam, Netherlands) 2004, 29, 1611-1621.

(28) Zhang, X.; Pham, J. Q.; Martinez, H. J.; Wolf, P. J.; Green, P. F.;

Johnston, K. P. Journal of Vacuum Science & Technology, B: Microelectronics and

Nanometer Structures--Processing, Measurement, and Phenomena 2003, 21, 2590-

2598.

(29) McClellan, A. K.; McHugh, M. A. Polymer Engineering and Science

1985, 25, 1088-1092.

(30) Gallagher, P. M., Coffey, M.P., Krukonis, V.J. and Klasuts, N. In

Supercritical Fluid Science and Technology; Johnston, K. P. a. P., J.M.L., Ed.;

American Chemical Society: Washington, D.C., 1989; Vol. 406, pp 334-354.

(31) Dixon, D. J.; Bodmeier, R. A.; Johnston, K. P. AIChE J. 1993, 39, 127.

(32) Randolph, T. W.; Randolph, A. D.; Mebes, M.; Yeung, S.

Biotechnology Progress 1993, 9, 429-435.

(33) Yeo, S.-D.; Debenedetti, P. G.; Radosz, M.; Schmidt, H.-W.

Macromolecules 1993, 26, 6207-6210.

(34) Dixon, D. J.; Johnston, K. P. AIChE Journal 1991, 37, 1441-1449.

(35) Maiwald, M.; Schneider, G. M. Berichte der Bunsen-Gesellschaft

1998, 102, 960-964.

(36) Kelley, S. P.; Lemert, R. M. AIChE Journal 1996, 42, 2047-2056.

62

(37) Wei, M.; Musie, G. T.; Busch, D. H.; Subramanian, B. J. Am. Chem.

Soc. 2002, 124, 2513-2517.

(38) Wei, M.; Musie, G. T.; Busch, D. H.; Subramanian, B. Green Chem.

2004, 6, 387-393.

(39) Xie, X.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem. Res. 2004, 43,

7907-7911.

(40) Sellin, M. F.; Webb, P. B.; Cole-Hamilton, T. J. Chem. Commum.

2001, 781-782.

(41) McLeod, M. C.; Anand, M.; Kitchens, C. L.; Roberts, C. B. Nano

Letters 2005, 5, 461-465.

(42) Ng, H.-J.; Robinson, D. B. Journal of Chemical and Engineering Data

1978, 23, 325-327.

(43) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Frederickson, G. H.;

Chmelka, B. F.; Stucky, G. D. Science (Washington, D. C.) 1998, 279, 548-552.

(44) Tanev, P. T.; Pinnavaia, T. J. Chemistry of Materials 1996, 8, 2068-

2079.

(45) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. Journal of the American

Chemical Society 1951, 73, 373-380.

(46) Saito, A.; Foley, H. C. AIChE Journal 1991, 37, 429-436.

(47) Horvath, G.; Kawazoe, K. J. Chem. Engg. Japan 1983, 16, 470-475.

(48) Lifshin, E. X-ray Characterization of Materials; Wiley-VCH: New

York, 1999.

(49) Korgel, B. A.; Fitzmaurice, D. Physical Review B: Condensed Matter

and Materials Physics 1999, 59, 14191-14201.

(50) Mattoussi, H.; Cumming, A. W.; Murray, C. B.; Bawendi, M. G.;

Ober, R. Physical Review B: Condensed Matter and Materials Physics 1998, 58,

7850-7863.

(51) Korgel, B. A.; Fullam, S.; Connolly, S.; Fitzmaurice, D. Journal of


63

(52) Guinier, A.; Fournet, G. Small-Angle Scattering of X-Rays; Wiley:

New-York, 1955.

(53) Glatter, O.; Kratky, O. Small Angle X-ray Scattering; Academic Press

Inc: New-York, 1982.

(54) Saunders, A. E.; Korgel, B. A. Journal of Physical Chemistry B 2004,

108, 16732-16738.

(55) Saunders, A. E.; Shah, P. S.; Park, E. J.; Lim, K. T.; Johnston, K. P.;

Korgel, B. A. J. Phys. Chem. B 2004, 108, 15969-15975.

(56) Lee, J.; Park, Y.; Kim, P.; Kim, H.; Yi, J. Journal of Materials

Chemistry 2004, 14, 1050-1056.

(57) Ehrburger-Dolle, F.; Morfin, I.; Geissler, E.; Bley, F.; Livet, F.; Vix-

Guterl, C.; Saadallah, S.; Parmentier, J.; Reda, M.; Patarin, J.; Iliescu, M.;

Werckmann, J. Langmuir 2003, 19, 4303-4308.

(58) Kresge, C. T.; Leonowickz, M. E.; Roth, W. J.; Beck, J. S. Nature

1992, 359, 710.

(59) Tanev, P. T.; Pinnavaia, T. J. Science (Washington, D. C.) 1995, 267,

865-867.

(60) Shah, P. S.; Holmes, J. D.; Doty, R. C.; Johnston, K. P.; Korgel, B. A.

J. Am. Chem. Soc. 2000, 122, 4245-4246.

(61) Shah, P. S.; Novick, B. J.; Hwang, H. S.; Lim, K. T.; Carbonell, R. G.;

Johnston, K. P.; Korgel, B. A. Nano Letters 2003, 3, 1671-1675.

(62) Konya, Z. P., Victor F.; Kiricsi, Imre; Zhu, Ji; Alivisatos, A. Paul;

Somorjai, Gabor A. Nano Letters 2002, 2, 907-910.

(63) Hiemenz, P. C. Principles of Colloid and Surface Chemistry.; Marcel

Dekker, Inc.: New York, 1986.

(64) Hourri, A.; St-Arnaud, J. M.; Bose, T. K. Journal of Chemical Physics

1997, 106, 1780-1785.

(65) Burns, R. C.; Graham, C.; Weller, A. R. M. Mol. Phys. 1986, 59, 41.

64

(66) Rubio, J. E. F.; Arsuaga, J. M.; Taravillo, M.; Baonza, V. G.; Caceres,

M. Experimental Thermal and Fluid Science 2004, 28, 887-891.

(67) Williams, J. W.; Krchma, I. J. J. Am. Chem. Soc. 1926, 48, 1888-1896.

(68) Lewis, J. E.; Biswas, R.; Robinson, A. G.; Maroncelli, M. Journal of


(69) Koone, N. D.; Zerda, T. W. Journal of Non-Crystalline solids 1995,

183, 243-251.

(70) Vigne-Maeder, F.; Amrani, S. E.; Gelin, P. J. Catal. 1992, 134, 536.

(71) Zhang, Z.; Lagally, M. G. Science 1997, 276, 377-383.

(72) Fukuoka, A.; Araki, H.; Kimura, J.-i.; Sakamoto, Y.; Higuchi, T.;

Sugimoto, N.; Inagaki, S.; Ichikawa, M. Journal of Materials Chemistry 2004, 14,

752-756.

65

Chapter 3

Infusion of Pre-synthesized Iridium Nanocrystals into Mesoporous

Silica for High Catalyst Activity

Traditionally, finely dispersed metal catalysts have been formed by reduction

of precursors within supports. A new concept for designing catalysts with enhanced

activities and selectivities is to infuse pre-synthesized nanocrystals, with well-defined

morphologies, into ordered mesoporous materials. The decoupling of nanocrystal

synthesis and infusion provides exquisite control of the nanocrystal size, morphology,

and dispersibility within the pores. A dispersion of iridium nanocrystals was infused

into mesoporous silica by expanding the solvent toluene with supercritical CO2. To

achieve high nanocrystal loadings, up to 1.3 wt %, the solvent quality was tuned to

strengthen the interactions of the nanocrystals with the pore walls, but without

precipitating the nanocrystals in the bulk solvent. Z-contrast STEM indicates

conclusively that the iridium nanocrystals were located within the pores and not on

the external silica surface. High catalytic activity was observed for 1-decene

hydrogenation, which is consistent with a high degree of dispersion of the 4.5 nm

nanocrystals throughout the pores, as observed by TEM. A maximum turnover

frequency (TOF) of 16 s-1 was measured, which was higher than the initial TOF for

homogeneous catalysis with the same nanocrystals in 1-decene. The iridium catalysts

do not require pre-treatment to remove the tetraoctylammonium bromide ligands to

achieve activation, as the ligands bind weakly to the iridium surface. Consequently,

the activity was not enhanced when calcined at 500oC in nitrogen.

Large portions of this chapter have been previously published as Gupta, G.; Stowell, C. A.; Patel, M. N.; Gao, X.; Yacaman, M. J.; Korgel, B. A.; Johnston, K. P. Chemistry of Materials 2006, 18, 6239-6249.

66

3.1 Introduction

In catalysis, substantial improvements in turnover rates and selectivities are

possible with metal nanocrystals containing high concentrations of active surface

sites.1,2 Nanocrystal growth may be controlled in solution with stabilizing ligands to

achieve specific size, crystallinity, shape, and surface properties.3 The design of the

surface sites on the nanocrystals may be guided by fundamental studies of reaction

mechanisms on well-defined surfaces on single crystals.4,5 Immobilization of the

nanocrystals within the pores of a mesoporous support helps maintain high surface

area by preventing coalescence of the nanocrystals during reaction.6 The size of the

mesopores may be designed to influence reaction selectivity and catalyst activity by

limiting access to the nanocrystals for certain reactants and products, as a function of

their size. Finally, the immobilization on the support facilitates catalyst recovery and

recycles.

Traditionally, supported catalysts are synthesized by precursor reduction,6-8

ion exchange9,10 and chemical vapor/fluid deposition.11-13 In these techniques, it is

challenging to control the metal cluster size and surface morphology. One approach

to metal nanocrystal encapsulation within a mesoporous support is to form the

mesoporous structure in the presence of a dispersion of pre-synthesized

nanocrystals.14,15 In this technique it can be difficult to achieve high loadings of

dispersed nanocrystals within the pores. In addition, the presence of the particles can

interfere with the synthesis of the desired pore structure and conversely, the synthesis

of the mesoporous material can change the morphology of the nanocrystals. For

example, the ligands stabilizing the nanocrystals may desorb and adsorb onto the

substrate during the reaction, leading to destabilization of the nanocrystals.

Furthermore, the surfactant templates used for the mesoporous material may interact

with the nanocrystal surfaces and modify their morphologies.

An alternative approach is to decouple nanocrystal synthesis from the

synthesis of the mesoporous material.16,17 An advantage of this approach is that a

67

large number of well-known synthetic methods may be utilized for the design of the

nanocrystals18 and of the mesoporous materials.19 Prior to infusion, the particles are

easily isolated from unwanted byproducts, cleaned, separated by size, and the ligands

may be exchanged if desired for alternative ligands. Here it is not necessary to

remove excess nanocrystal precursors and reaction by-products from the pores as in

traditional techniques such as wet impregnation.20 After synthesis and purification,

the nanocrystals may be infused into the pores of a pre-synthesized mesoporous

material by a physical technique without modifying the structure of either material.

Rioux et al16 infused platinum nanocrystals coated with polyvinylpyrrolidone

in an aqueous dispersion into high surface area mesoporous silica. Under sonication,

pore loadings reached approximately 1wt %. Without sonication, the nanocrystals

were not able to infuse into the pores and remained coated to the outside of the silica.

The nanocrystals penetrated the pores for a 1:1 mixture of water and ethanol, but not

with pure water. The observed difference in penetration of the pores was attributed to

the difference in surface tensions of water and ethanol.

In our previous study,17 gold nanocrystals coated with dodecanethiol were

infused into mesoporous silica to achieve highly dispersed nanocrystals with loadings

over 2 wt %. The infusion required the addition of supercritical CO2. CO2 has a low

polarizability relative to hydrocarbon solvents, and serves as an antisolvent for the

alkyl-coated nanocrystals.21-23 (However, nanocrystals dispersions have been obtained

in pure CO2 at high temperatures24 and in very dilute dispersions.25) For toluene

expanded with CO2, it is possible to tune the solvent strength to raise the long-ranged

van der Waals interactions of the nanocrystals with the pore walls, and thus the

nanocrystal loading, without flocculating the nanocrystals in the solution phase.17

This goal was not achieved by adding methanol as an antisolvent to toluene.

The exact location of nanocrystals within the mesopores versus on the

external surface is an important issue in metal/mesoporous silica catalysts. Rioux et

al.16 showed that nanoparticles, upon annealing at 400 °C, formed nanorods with a

conformation related to the geometry of the cylindrical mesopores. However, the

68

possibility of nanoparticles on the outside surface was not ruled out. To the best of

our knowledge, the location of pre-synthesized nanocrystals infused into mesoporous

materials has not been determined conclusively in previous studies. Midgley et al.26

showed that the exact location of nanocrystals within mesoporous silica can be

determined by Z-contrast scanning transmission electron microscopy (STEM) with a

high angle annular dark field (HAADF) detector. The nanocrystals were synthesized

inside the mesoporous silica by thermal reduction of a precursor. The use of a

HAADF detector has the advantage that the images do not suffer from diffraction

contrast. Furthermore, the contrast between heavy and light elements is enhanced as

the signal is proportional to the square of the atomic number.27

In a recent study of Pt nanocrystals infused into mesoporous silica,16 the as-

synthesized supported catalyst was inactive because the strongly bound polymeric

shell blocked the active sites on the metal surface.28 Calcination at 450 °C and above

for 12 hours was necessary to remove the polymeric coating and activate the catalyst.

High temperature annealing can potentially result in oxidation of the nanocrystals,

can significantly change the nanocrystal size and surface morphology, and can also

result in aggregation and coalescence, which reduces the available surface area of the

catalyst.29 The nature of ligand binding also has a large effect on homogeneous

catalysis with nanoparticle solutions. Stowell et al.30 showed that the catalytic activity

of Ir nanocrystals in solution for hydrogenation of 1-decene was influenced markedly

by the nature of the binding of the stabilizing ligands. Strong binding ligands

passivated the surface and inhibited catalytic activity, while nanocrystals with weakly

bound ligands showed much greater catalytic activity. However, as weakly bound

ligands desorbed from the nanocrystals surface into the solvent, the nanocrystals

aggregated and the activity declined.

Herein, we show that pre-synthesized iridium nanocrystals stabilized by

weakly bound tetraoctylammonium bromide (TOAB) ligands may be infused into

pre-synthesized mesoporous silica to produce a highly active catalyst for 1-decene

hydrogenation. Z-contrast STEM with an HAADF detector is used to discriminate

69

between nanocrystals within the mesopores and on the external silica surface. We

chose to study decene hydrogenation, given the previous data for homogeneous

catalysis with Ir nanocrystals in solution stabilized by TOAB.30 The activity is

compared for the original infused catalyst without pre-treatment and for catalysts in

which the ligands were removed by annealing at high temperature in air and nitrogen,

or by extraction with compressed carbon dioxide. The objective was to identify a

ligand with sufficiently weak binding to the surface, such that it becomes unnecessary

to pre-treat the catalyst to remove the ligand. In order to prevent aggregation of the

nanocrystals in solution during infusion, excess ligand was added to the solution. The

effect of CO2 concentration on the stability of the nanocrystal dispersion was

investigated to identify the optimum conditions for the infusion. The stability of the

catalyst was investigated by recovering the catalyst and performing a second reaction.

The activation energy of decene hydrogenation reaction was determined over a

temperature range of 40 °C to 75 °C. The catalytic activities of the Ir nanocrystals

infused into the mesoporous silica supports were compared with a commercial Pd

hydrogenation catalyst supported on alumina used under the same reaction

conditions.

3.2 Experimental

3.2.1 Materials

All chemicals were used as received. Tetraoctylammonium bromide or TOAB

(C32H68N Br, 98 %), decane (99 %), 1-decene (94 %) and 1,2 hexadecanediol (90%)

were purchased from Aldrich. (Methylcyclopentadienyl) (1,5-cyclooctadiene) iridium

( (C6H7)(C8H12)Ir, 99 %) was purchased from Strem. Toluene (99.9%) and

concentrated liquid HCl (normality 12.1) were obtained from Fisher Scientific.

Ethanol (Absolute 200 proof) was obtained from Aaper alcohol. Methanol (purity >

99.8%) was obtained from EM Sciences. Tetraethoxysilane (TEOS), and dioctylether

(97%) were obtained from Fluka Chemika and Poly(ethylene oxide)-b-

70

poly(propylene oxide)-b-poly(ethylene oxide) EO~20PO~70EO~20 (Pluronic P-123)

from BASF Corporation. CO2 (purity > 99.99%) (Matheson Gas Products) was used

as received. Polycarbonate filters (0.05 micron) were obtained from Osmonics. 5%

Palladium on alumina (Pd-Al2O3) commercial catalyst was obtained from Alfa Aesar.

Water was doubly distilled and deionized.

3.2.2 Iridium Nanocrystal Synthesis

Iridium nanocrystals with TOAB ligands were synthesized by a method

developed by Stowell et al.30 using a 25 ml- three neck, round bottom flask. 0.2 g of

hexadecanediol, 7 ml of dioctylether, 0.2 g of (methylcyclopentadienyl) (1,5-

cyclooctadiene) iridium, and 0.76 g of TOAB were measured into the flask. The

solution was freeze-pump-thawed for three cycles using liquid nitrogen, and heated

up to 270 °C for 30 minutes. The product was a black liquid and the particles were

isolated using one rinse in ethanol as an antisolvent. Rinsing the solution multiple

times in ethanol resulted in the removal of capping ligands from the iridium surface,30

so exposure to ethanol was kept at a minimum. The nanocrystals were finally

dispersed in toluene before the infusion procedure. Excess TOAB ligand (10 mg/ml)

was dissolved in the nanocrystal dispersion so as to prevent desorption of the ligands

and hence prevent aggregation of the particles.


Mesoporous silica SBA-15 (referred to as MPS) was synthesized with a block

copolymer template similarly to the reported procedure of Zhao et al.31,32 A solution

of 1.6 g EO20PO70EO20 : 7 mL HCl: 8 mL TEOS: 38 mL H2O was prepared, stirred

for 24 h at 50°C and then heated at 100°C for 24 h. The resulting slightly yellow

powder was dispersed in ethanol, filtered and then again heated at 100 °C for 24 h.

Calcination was carried out by slowly increasing temperature from room temperature

to 500 °C in 4 h and heating at 500 °C for 6 h.

71

3.2.4 Infusion of Ir into Mesoporous Silica with CO2

In a typical experiment, 100 mg of mesoporous silica and 6 mL of iridium

nanocrystals dispersed in toluene, with a concentration of 0.4 mg/ml, were loaded

into a variable-volume reaction cell17 with a working volume (volume with piston in

the cell) of 27 mL. An amount of CO2 equivalent to 4 mL at 241 bar (3500 psia) and

35 °C was added so as to produce a total volume of reaction mixture of 10 mL. The

cell was then pressurized to 241 bar with CO2 at the back side of the piston using an

ISCO model 100 DX computer controlled syringe pump. The contents of the cell, in a

water bath, were stirred with a magnetic stir bar (Figure 3.1). The reaction was

maintained at 35 °C and 241 bar (3500 psia) and allowed to proceed for 24 h. After

the reaction, the cell was cooled to 0°C and CO2 was then vented as a vapor from the

top of the cell at the vapor pressure, 35 bar. The concentrations of organics in the CO2

vapor were small at 0 °C such that losses were minimized. The supernatant was

collected and filtered, to separate the Ir-MPS. Ir-MPS was cleaned by washing with 2

ml of pure toluene to remove the nanocrystals adsorbed on the outer surface of the

silica. Hence the resulting nanocomposite contained Ir nanocrystals immobilized

within the pores.

The extent of incorporation of iridium nanocrystals in the silica was

determined by subtracting from the initial mass of the dispersed nanocrystals in

toluene, the final mass of the nanocrystals recovered in the supernatant after

filtration.17 A correction was not made for the nanocrystals removed by washing with

pure toluene. The absorbance of the iridium nanocrystals dispersed in toluene before

and after infusion was measured at a wavelength of 500 nm using a Cary 500 UV-

Vis-NIR spectrophotometer with an optical path length of 1 cm. A standard

calibration curve was generated at 500 nm using known concentrations of iridium

nanocrystals in toluene as shown in Figure 3.2. The experiments were repeated

several times to check for the reproducibility in the nanocrystal loading, which was

on the order of +0.2 in wt % units, or 15 % of the final Ir mass. The different factors

72

Figure 3.1

Schematic of the infusion apparatus.

CO2 Cylinder ISCO Pump

Variable Volume View Cell PT

Stir Bar Piston

73

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.1 0.2 0.3 0.4 0.5

Abs

orba

nce

at 5

00 n

m

Concentration of Iridium (mg/ml)

Figure 3.2

Calibration curve of UV- Vis absorbance versus concentration for iridium nanocrystal

dispersions in toluene, measured at 500 nm.

74

which cause uncertainty in final iridium loadings are volume measurement (1%),

weight of silica (1%), uncertainty in absorbance (1%) and uncertainty in linearity (2-

3%) of calibration curve. The uncertainty of each of these individual factors results in

uncertainty on nanoparticle loading. Total uncertainty in loadings is reported as the

sum of the uncertainty due to each individual factor, which is approximately close to

10-15%. The uncertainty in this method of loading determination is comparable to

that of elemental analysis (10% error), which makes it equally effective for estimating

the metal loadings. The concentration of nanocrystals washed from the outer surface

of the silica with pure toluene was measured spectrophotometrically. The absorbance

measured was less than 0.01 and within the detection limit indicating a maximum of

0.1 mg of nanocrystals. The low concentration of nanocrystals is consistent with a

ratio of external to internal surface area of mesoporous silica on the order of 10-4

based on the pore diameter (~1 nm) and particle size (~10 μm). Hence if the loading

of nanocrystals is uniform on external and internal areas, then the loading on the

external surface would be negligible (~10-4%).

3.2.5 Determination of Nanocrystal Precipitation Boundary

The effect of CO2 concentration was determined on the stability of the

nanocrystal dispersions in the mixed solvent with toluene. Nanocrystal dispersions

were analyzed at 35 oC and 241 bar with a Cary 3E UV-Vis spectrophotometer in a

vertically mounted variable-volume high-pressure optical cell with a path length of 2

cm as described previously.33 The schematic is shown in the Figure 3.3. 6 mL of 0.4

mg/mL concentration of iridium nanocrystals in toluene with 10 mg/mL of excess

ligands in the solvent was loaded in the front part of the cell. The contents were

stirred using a magnetic stir bar. The computer controlled ISCO syringe pump was

used to introduce a known amount of CO2 in the front part of the cell, and the cell

was stirred for 30 s to mix the solvents. The stirring was stopped to allow the particles

75

Figure 3.3

Schematic of apparatus for determining the dispersibility of the iridium nanocrystals

in CO2-toluene solvent.

Light Source

UV-Vis Spectrophotometer

Detector

CO2 Syringe pump

Variable volume View Cell

76

to settle and the absorbance was measured for 20 min. Another aliquot of CO2 was

added and the procedure was repeated.

3.2.6 Catalysis

1-decene hydrogenation reactions were carried out in a 25 ml round bottom

flask batch reactor. Supported iridium nanocrystals were added to neat 1-decene to

create a 1500:1 decene to iridium mass ratio. Hydrogen gas at 5 psig was bubbled into

the dispersion. The system was stirred using a magnetic stir bar at high speed to raise

mass transfer rates. After the system was well saturated with hydrogen for 20 min. at

23 °C, the temperature was increased to the reaction temperature of 75 °C, and

hydrogen flow was maintained throughout the reaction. Aliquots of the solution (100

μl) were removed every hour for next four hours and stored at room temperature in

sealed vials for a maximum of 4 hours before analysis by gas chromatography.

3.3 Characterization

3.3.1 Nitrogen Porosimetry

N2 porosimetry was used to measure the surface area and the pore size

distribution of the mesoporous silica with a high speed surface area BET analyzer

(NOVA 2000, Quantachrome Instruments, Boynton Beach, FL) at a temperature of

77 K. Silica samples were pretreated at 100 ºC for 24 h in vacuum immediately prior

to data collection. Pore size distributions were analyzed by using the adsorption

branch of the isotherm. The Barrett- Joyner- Halenda (BJH) method34 was used for

pore size distribution.

3.3.2 Transmission Electron Microscopy

Low resolution pictures of mesoporous silica, infused with iridium

nanocrystals, were visualized by TEM on a Phillips EM280 microscope with a 4.5 Å

point-to-point resolution operated with an 80 kV accelerating voltage. High resolution

77

transmission electron microscopy (HRTEM) was performed using a JEOL 2010F

TEM operating at 200 kV. Z- Contrast STEM was performed on FEI TECNAI G2

F20 X-TWIN transmission electron microscope using a HAADF (High-angle annular

dark field) STEM detector with a 2 Å point-to-point resolution. Images were obtained

primarily with a GATAN digital photography system. Silica particles were deposited

from a dilute chloroform solution onto 200 mesh carbon-coated copper TEM grids

(Electron Microscopy Sciences).

3.3.3 Gas Chromatography/ Mass Spectroscopy

The extent of catalytic conversion of 1-decene was determined with a

Finnigan MAT GCQ gas chromatograph with mass spectrometer detector (GC/MS).

The column used was DB5-MS (5% phenyl-methylpolysiloxane, J & W scientific)

with a diameter of 0.32 mm and a length of 30 meter. The GC was used in positive

electron ionization mode with electrons accelerated at 70eV. 0.1μl of the undiluted

sample was injected using a split injection technique with an autosampler. The

retention time of the 1-decene and decane peaks was approximately 30 seconds. The

GCMS was calibrated to determine the relative amounts of 1-decene (MW = 140) and

decane (MW = 142) in each aliquot as shown in Figure 3.4. The curve was highly

linear with a correlation coefficient of 0.9995. 3.4 Results and Discussion

3.4.1 Nanoparticle Characterization

Figure 3.5(a) shows a representative TEM image of TOAB-stabilized Ir

nanocrystals. The particles were spherical in shape. By counting over 200 particles,

the average diameter was found to be 4.5 nm with a standard deviation of 0.6 nm (see

Figure 3.5(b)). Nanoparticles in the image appear slightly larger than the original due

to pixel aberration caused by image magnification.

78

0

0.1

0.2

0.3

0.4

0.5

0 0.2 0.4 0.6 0.8 1

Peak

Are

a (1

-dec

ene)

/ P

eak

Are

a (1

-dec

ene)

+ (d

ecan

e)

Volume % decane

Figure 3.4

GC-MS calibration curve of a mixture of 1-decene and 1-decane.

79

(a)

0

20

40

60

80

100

120

3 3.5 4 4.5 5 5.5

Num

ber

of P

artic

les

Nanoparticle diameter

(b)

Figure 3.5

(a) TEM micrograph of iridium nanocrystals stabilized by TOAB and (b)

corresponding histogram for iridium nanocrystals.

80

3.4.2 Characterization of Mesoporous Silica

Representative nitrogen adsorption/desorption isotherms and corresponding

pore size distributions (BJH method34) are shown in Figure 3.6. MPS has a mean pore

diameter of 6.5 nm, BET surface area of 742 m2/g and a pore volume of 0.98 cm3/g.

As well cited in literature for type SBA-15 mesoporous silica31 three distinct regions

are evident in the adsorption isotherm: (i) monolayer-multilayer adsorption, (ii)

capillary condensation and (iii) multilayer adsorption on the outer particle surfaces.

As can be seen in Figure 3.6, MPS shows a type-H1 hysteresis loop with capillary

condensation at an activity P/Po= 0.7.31 The step in the hysteresis loop is very steep

with a jump from 300 cm3/g to 500 cm3/g indicating capillary condensation. The

slope and the height of the step are clear indications of very well defined mesopores

with narrow pore size distribution.35 In our previous study17 we measured interpore

spacing of mesoporous silica type SBA-15 synthesized exactly as in the current study,

using small angle X-ray scattering (SAXS). SAXS showed multiple peaks indicating

a long range arrangement of pores with a pore diameter of 8.8 nm and a wall

thickness of 3.1 nm, typically associated with SBA-15 silica.31

3.4.3 Nanocrystal Dispersibility in Toluene-CO2 Mixtures

The nanocrystals are stabilized in the dispersion by steric repulsion from the

ligands adsorbed to the surface. The interactions between the ammonium bromide

head group and metal surfaces like iridium30 and gold36 have been found to be

relatively weak. For Ir, weak adsorption of TOAB ligands results in visible

aggregation of the nanocrystals over a period of 12 h. Ligand desorption can be

reduced by adding excess free ligand to the solvent. Mattoussi et al.37 examined this

concept quantitatively with SAXS studies on CdSe nanocrystals stabilized by

trioctylphosphine and trioctylphosphine oxide (TOP/TOPO) dispersed in hexane. In

the case of pure solvent with no excess ligands, the nanocrystals exhibited attractive

81

100

200

300

400

500

600

0 0.2 0.4 0.6 0.8 1

Vol

ume

(cc/

g)

Relative Pressure (P/Po)

(a)

0

0.005

0.01

0.015

0.02

0.025

0.03

0 50 100 150 200

Ads

orpt

ion

Dv(

d) (c

c/A

/g)

Pore diameter (A)

(b)

Figure 3.6

(a) N2 adsorption-desorption isotherms for mesoporous silica MPS and (b)

corresponding pore size distributions for MPS.

82

interactions. However, the addition of free ligands eliminated interparticle attractions

and the nanocrystal interactions eventually became repulsive. In this study, ~10

mg/mL of TOAB was added to the Ir nanocrystal dispersion in toluene with a

concentration of 0.5 mg/ml and nanocrystals were found to be stable for several days

on the basis of visual observation.

Figure 3.7 shows the absorbance at 500 nm of the Ir nanocrystal dispersion as

a function of solvent composition and time. As the total amount of nanocrystals

loaded in the variable-volume cell remains constant, the absorbance decreases as the

concentration of Ir is decreased by the addition of CO2 to the dispersion. It was

determined that the absorbance remained constant with time for less than 50% CO2

(w/w). When the CO2 concentration reached 50%, the absorbance began to decrease

modestly with time. At 55% CO2, absorbance decreased faster, indicating faster

aggregation and settling of particles. CO2 is a poor solvent for the hydrocarbon

ligands, and the insufficiently solvated ligands do not provide enough steric

stabilization to counter the attractive van der Waals forces between the cores. Since

large particles possess greater interparticle van der Waals attractions, they precipitate

first upon lowering the solvent quality.38 On the basis of these results, the loading of

the Ir particles for catalytic studies was performed at 40% CO2, where the absorption

was constant, indicating a stable dispersion.

3.4.4 Nanocrystal Loading in Pores using CO2

The Ir nanocrystals dispersed in toluene-CO2 solvent were infused into the

mesoporous silica over a 24 hour period. In our previous study,17 we showed that 24

hours was sufficient for reaching a maximum loading, with little increase from 24 to

72 hrs. We assumed 24 hours was also sufficient in the current study, since the metal

cores in each study had nearly the same Hamaker constants (gold vs. iridium) and

since both studies used hydrocarbon ligands. The loading of the iridium nanocrystals

within the mesoporous silica measured spectrophotometrically was about 1.3±0.2 wt

% (g Ir/g silica) for a CO2 concentration of 40%. For the nanocrystals dispersed in

83

Figure 3.7

Dispersibility of Ir-TOAB nanocrystals in CO2-toluene mixed solvent measured by

UV-Vis spectroscopy.

84

pure toluene without CO2, the loading was negligible. The high loading suggests that

the nanocrystals are mobile on the silica pore surface with the ability to diffuse deeply

into the pores without causing pore blockage.

The equilibrium adsorption of a nanocrystal onto the silica surface within a

pore is a function of the chemical potential of the nanocrystals in the solvent and the

long-ranged van der Waals forces between the nanocrystals and silica surface,

mediated by the intervening solvent.17 The van der Waals force of attraction between

the particle and the silica wall is directly proportional to the Hamaker constant of the

system, A132, for an iridium particle (1) and silica surface (2) across an intervening

medium (toluene or toluene-CO2 mixture, 3). It may be approximated by39

))(( 33223311132 AAAAA −−= (3.1)

where Aii is the Hamaker constant for pure i, which can be estimated using a

simplification of the Lifshift theory.39

Ayy =34

kbTεy −εvacuum

εy + εvacuum

⎛

⎝ ⎜ ⎜

⎞

⎠ ⎟ ⎟

2

+3hν e

16 2ny

2 − nvacuum2( )2

ny2 + nvacuum

2( )3 / 2 (3.2)

where eν is the maximum electronic ultraviolet adsorption frequency, typically

assumed to be 15103× s-1.23 The A for iridium, 19104 −× J,39 the ε and n for silica, 3.91

and 1.45, respectively40,41 and for toluene, 2.38 and 1.496, respectively42,43 were taken

from the literature. The Hamaker constants for silica and toluene calculated from

equation (3.2) are 201097.5 −× J and 201098.6 −× J, respectively.

For pure toluene without CO2, A132 is -0.31×10-20 J. The negative sign in the

overall Hamaker constant indicates that the interactions between the iridium

nanoparticle and the silica wall were repulsive or unfavorable for particle adsorption.

Hence, even though the nanocrystals can diffuse inside the pores along with the

solvent and fill the pores, the weak adsorption onto the silica surface may be expected

85

to result in low loadings as observed experimentally. If the dispersion of nanocrystals

in the pores contained the bulk concentration of nanocrystals, the calculated loading

would be less than 0.1 wt %, as the internal pore volume was only 1% of the entire

volume in the system.

CO2 has an unusually low polarizability/density (very weak vdW forces), and

may be added to a liquid solvent to form a mixture with an unusually low Hamaker

constant.17 The Hamaker constant for pure CO2 was calculated to be only 201044.1 −×

J.17 The Hamaker constant of the intervening mixed-solvent A33, was determined for

the binary solvent mixture with the Lorenz-Lorenz mixing rule as described

previously.17 The interactions between the nanocrystals and silica become strongly

attractive as the Hamaker force of the mixed solvent is reduced by adding CO2.17 For

40% CO2 in the intervening solvent, A132 was determined to be 4.1×10-20 J, indicating

strong attractive interactions of the Ir nanocrystals with the silica surface. The high Ir

loading of 1.3 wt.% seen experimentally were consistent with this prediction, as was

the low loading for pure toluene, where A132 was weakly repulsive.

The low Hamaker constant of the mixed solvent, and the weaker solvation of

the ligands also strengthens the interactions between the nanocrystals in solution, as

was seen in Figure 3.7. However, iridium nanocrystals form a stable dispersion with

40% CO2 in the solvent mixture. Thus, the solvent strength may be tuned with CO2 to

achieve strong adsorption on the silica surface, and thus high loading of nanocrystals,

without aggregating the nanocrystals in solution. The ability to fine tune the solvent

strength with CO2 to achieve high nanocrystal loadings has also been observed for Au

nanocrystals.17 Given these results for gold, it is to be expected that Ir could also be

infused into silica, given the similar Hamaker constants for Au and Ir, and the similar

hydrocarbon tails in the ligands used in both studies.

3.4.5 Visualization of Nanocrystals in Mesoporous Silica

High resolution Z-contrast STEM was utilized to confirm that most of the

nanocrystals are dispersed within the pores, rather than on the external surface of the

86

particles. The contrast variation in the STEM is proportional to the mass density of

the region multiplied by the atomic number squared (Z).27 Regions of high density are

shown in white. Figure 3.8(a) shows the planar view of the nanocrystals aligned

within the pores. Iridium has higher atomic number and mass density with respect to

the pore walls composed of silica, which in turn are denser than empty pores. Hence,

the iridium nanocrystals are observed as spherical bright dots, silica walls appear as

discrete white lines and empty pores appear as dark spaces between the silica walls.

Most of the nanocrystals are aligned within the empty pores along the pore wall while

a small fraction appears to be in the pore wall. The nanocrystals, which appear to be

on the pore walls, could have been in many cases located inside the secondary pores

(which act as interconnects between the pores) within the pore walls, which can be as

large as 4 nm in diameter for mesoporous silica of type SBA-15.44

Figure 3.8(b) is an out-of-focus image of 3.8(a) showing two different planes

of ordered pores (as indicated by arrows) with iridium nanocrystals inside the pores.

The planes were too close together to come from two different particles, but instead

indicated a fault line in the surfactant template for a single particle. Thus these

nanocrystals were in the interior of a particle in the pores and not on the surface,

further supporting the observations in Figure 3.8(a). Figure 3.8(c) shows the top down

view of the pores where we see the pore opening on the silica surface, with pores

arranged in hexagonal assembly. The arrows indicate the nanocrystals that are located

within the dark empty pores. Almost all the nanocrystals are dispersed within the

pores rather than on the external surface, consistent with the much higher internal

surface area as compared to the external area. The micrographs thus conclusively

prove that almost all the iridium nanocrystals are present within the pores.

Further evidence that the nanocrystals are mostly within the mesopores is

provided by the theoretical estimate of the close-packed monolayer coverage for

nanocrystals on the external silica surface. The particle diameter was measured to be

87

(a)

(b)

(c)

(d)

Figure 3.8

(a) Z contrast STEM (planar view) of MPS infused with Ir-TOAB; (b) Out of focus

view of (a); (c) top down view of (a); and (d) TEM micrograph of MPS infused with

Ir-TOAB nanocrystals without any pre-treatment. Scale bars in (a-c) are 10 nm.

88

greater than 10 micron by laser light scattering (Malvern Mastersizer-S, Malvern

Instruments Inc., Southborough, MA). The average molecular weight of 4.5 nm Ir

nanocrystals was approximated to be about 400,000 g/mol. The external surface area

of 10 μm nonporous silica spheres is 0.3 m2/g. From geometrical calculations, it was

found that in a hexagonal unit lattice 45% of the surface area was occupied by pores.

With this correction for 7 nm pores, the maximum external surface area of the porous

silica particles was determined to be 0.16 m2/g. In addition, the packing density of

0.906 (maximum density considering hexagonal close packing) was assumed for the

monolayer assembly, which arises because of the voids between the spherical

nanocrystals. The 1.3 wt % loading corresponded to 161096.1 × nanocrystals per gm

silica, while only 15103.4 × nanocrystals were required to form a close packed

monolayer on the external silica surface. The 1.3 wt % loading of nanocrystals was

therefore approximately 4.5 times that of monolayer coverage on the external surface.

Thus, even if external surface area was covered by a monolayer of iridium

nanocrystals, 78% of the nanoparticles would still be inside the pores, consistent with

the observations by Z-contrast microscopy. The high loading in the pores suggests

that the adsorbed particles are mobile and can move toward the center of the pore by

diffusion on the surface and/or in the bulk, without blocking the pores.

Figure 3.8(d) shows a typical HRTEM image (without Z contrast) of Ir

nanocrystals infused into mesoporous silica. The nanocrystals are highly dispersed,

although the resolution is considerably lower than in the Z contrast images in Figures

3.8 (a-c).

3.4.6 Decene Hydrogenation in Ir/MPS catalysts

The rates for 1-decene hydrogenation with uncalcined Ir/MPS are shown in

Figure 3.9. The conversion was linear in time. As the reaction was conducted in 1-

decene as a neat solvent, the reaction rate was zero order with respect to 1-decene.

Given that the solvent was the reactant, the mass transfer resistances for 1-decene to

89

0

100

200

300

400

500

600

0 1 2 3 4 5 6

Mol

dec

ene

reac

ted/

g c

atal

yst

Time (hr)

Ir-MPS (No pre-treatment)

Pd-Alumina

Figure 3.9

Comparison of catalyst activities for decene hydrogenation with Ir-MPS (without any

pretreatment) and an industrial catalyst Pd-Al2O3.

90

the surface were negligible. The surface of the catalyst was constantly saturated with

1-decene, eliminating any concentration dependence with time. The linearity also

suggested that the catalyst was not deactivated over the course of the experiment.

Hydrogen was continuously bubbled into the reaction mixture at 760 torr to maintain

a constant H2 concentration in the liquid phase. The rate of 1-decene hydrogenation is

a function of reaction temperature, concentration of hydrogen in the liquid phase and

the 1-decene concentration as follows

constCCfRTE

Ar Hdecenea

decene =⋅⎟⎠⎞

⎜⎝⎛ −

⋅=− ),(exp2

(3.3)

where A is the pre-exponential factor, Ea is the activation energy, Cdecene is the

concentration of neat 1-decene and CH2 is the concentration of dissolved hydrogen in

the liquid phase. In our case, all three factors remain constant; hence the rate of

reaction remains constant.

A turnover frequency (TOF) can be defined as the moles of reactant consumed

per unit time per mole of available active sites of the catalyst30

⎟⎠⎞

⎜⎝⎛

⎟⎟⎠

⎞⎜⎜⎝

⎛=

dtdm

NTOF decene

site

.1 (3.4)

where ( )site p ucN n A /A= , Ap is the surface area of the nanocrystal determined from

TEM, Auc is the area of one unit cell on the surface and n is the number of Ir atoms in

a unit cell face. The number of surface unit cells is simply Ap/Auc. The number of

active sites may only be approximated. Even if the moles of binding sites are

measured for a certain gas via chemisorption, the active sites for the catalytic reaction

may not be exactly the same sites.28 We chose to assume that the TOAB ligands did

not block any active sites as has been assumed in similar studies.28,45 It was assumed

that all of the surface Ir atoms were active, resulting a conservative lower limit for the

TOF via equation 3.4. The number of surface atoms was determined from the total

91

surface area of the nanocrystals, the number of iridium unit cells on the surface, a

lattice length of 0.384 nm, and the number of iridium atoms within the FCC unit cell.

3.4.7 Turnover Number for Uncalcined Ir-MPS

The TOF for uncalcined Ir-MPS in the 1-decene hydrogenation was 16 s-1.

The highly dispersed nanocrystals, with an average diameter of 4.5 nm and minimal

aggregation enabled the high catalyst activity. The use of weakly binding ligands for

nanoparticle stabilization in the bulk solvent, which readily desorb from the surface

of the Ir nanocrystals dispersed on the support, may be expected to produce a large

number of active sites and thus high catalyst activity.

For metal nanocrystals with very high activities, it can be challenging to

prevent poisoning of the surface during reaction resulting in reduced catalytic

activity or in some cases complete deactivation.28 Catalyst deactivation can be

particularly severe for concentrated reactants, in our case pure reactant.46 In order to

determine if Ir-MPS undergoes any surface poisoning during decene hydrogenation,

the Ir-MPS was recovered and used for a second catalytic reaction. Table 3.1

illustrates that the activity only dropped by 7% in the second reaction, within

experimental error, and remained much higher than the industrial catalyst in the first

reaction.

3.4.8 Effect of Catalyst Pretreatment on Catalytic Activity

Although, capping ligands are necessary for dispersing the nanocrystals

during infusion, they potentially may have a negative effect on the catalytic activity if

they block active sites on the Ir surface. Stowell et al.30 showed that iridium

nanocrystals stabilized with strongly bound ligands like oleic acid/oleylamine were

essentially catalytically inactive, whereas those stabilized with weakly bound ligands

like TOAB provided much higher catalytic activity. The TOF in decene

hydrogenation improved significantly with successive reactions, from 5 s-1 to 124 s-1,

92

Catalyst Temp

(oC)

TOF

(mol decene

reacted/ mol

surface atoms

of catalyst /sec)

Activity

(mol of decene

reacted / g of

metal /sec× 104)

Ir-MPS (uncalcined) 75 16.3 1.66

Ir-MPS (2nd run

uncalcined) 75 15.1 1.54

Ir-MPS (calcined in

inert atmosphere,

nitrogen at 500C for 6

hrs)

75 15.4 1.59

Ir-MPS (annealed in

CO2 at 4000psi, 60C

for 12 hrs)

75 17.1 1.75

Ir-MPS (calcined at

250C for 6 hrs in air) 75 11 1.12

Ir-MPS(calcined at

500C for 6 hrs in air) 75 4 0.41



5% Pd on Alumina

(Industrial Catalyst) 75 - 0.92

Table 3.1

Activity and turnover frequency for decene hydrogenation of 1.3% Ir supported on

mesoporous silica relative to 5 % Pd on alumina.

93

as the weakly bound ligands desorbed from the surface. However, the nanocrystals

eventually aggregated from too much ligand desorption resulting in a loss in surface-

to-volume ratio and activity.30 This aggregation may be mitigated by supporting the

nanocrystals within mesoporous silica.

Table 3.1 compares the TOFs, with 760 torr hydrogen pressure at 75 °C unless

otherwise specified, for decene hydrogenation for Ir-MPS catalysts pre-treated in

various ways. The TOF obtained with the untreated Ir-MPS was 16±1 s-1. The TOF

was essentially the same when the catalyst was calcined in an inert nitrogen

atmosphere at 500 °C for 6 hours. This temperature was sufficient to desorb and

volatilize the TOAB ligands. The nanocrystals were immobilized on the support, as

they did not show any deactivation from sintering upon calcination at 500 oC. The

catalysts were also calcined in air at two different temperatures, 250 ˚C and 500˚C.

The TOF decreased from 16.5 to 11 s-1 when calcined at 250˚C, and to 4, less than a

quarter of the original activity, at 500 oC. Given that the TOF did not change after

calcination in N2 at 500 oC, the loss in activity upon calcination in air may be

attributed to oxidation of the surface of the Ir nanocrystals. Annealing in supercritical

CO2 at 275 bar pressure and 60 °C for 12 hours produced a slight increase in the

catalytic activity. The solubility of the ionic molecule TOAB in supercritical CO2

may be expected to be low and hence 0.1 gm of 1.3% Ir-MPS was annealed in

approximately 25 gm of supercritical CO2.

In the case of supported nanocrystals on mesoporous silica, we observe that

high catalytic activity may be obtained without thermal pre-treatment to remove the

ligands from the surface. Furthermore, the high activity remains constant even after

pretreatment by a) thermal calcination in N2 and b) CO2 extraction. Thus any ligands

on the surface prior to these pretreatments did not inhibit catalyst activity, as was the

case for polymeric stabilizers on Pt nanocrystals.28 Further evidence will be provided

to explain this highly beneficial result.

The TOAB ligands will partition between the bulk phase, the nanocrystals and

the mesoporous silica during the infusion process and during the catalytic reaction.

94

The mesoporous silica with a surface area of 742 m2/g has a large number of sites for

ligand adsorption. The total surface area of the adsorbed nanocrystals was only 3.5

m2. Suppose that all of the TOAB adsorbed exclusively on the silica. The area for a

linear dodecanethiol ligand has been reported to be approximately 20 A2 per

molecule.47 If it is assumed that the area of TOAB (highly branched) is twice that of

dodecanethiol, we estimate that the total amount of TOAB required for monolayer

coverage on the mesoporous silica is about 1.7 g TOAB/ g silica. Consequently, even

if all the TOAB in the system (60 mg added) adsorbed exclusively on the mesoporous

surface, only 35% of the silica sites would be covered. Thus, it is likely that a large

fraction of the weakly bound ligands on the dilute nanocrystals within the pores will

migrate to the vacant sites on the silica and reduce the coverage of ligands on the

surface. Ligand desorption from the nanocrystals will increase the number of

catalytically active sites.

TOAB’s weak binding and highly branched architecture will leave more

catalytic sites available relative to ligands such as oleic acid and oleylamine. The

volume occupied by the small ammonium bromide head groups will be relatively

small for TOAB relative to the larger volume of the four hydrocarbon tails. In

essence, the bulky tails will leave gaps between the head groups containing active

catalytic sites for the small hydrogen molecules. Consequently, H2 atoms may still

chemisorb on the surface and react very close to the binding site of the TOAB

underneath the larger cone of the tail groups. The transfer of ligands from the

nanocrystals to the silica surface and the gaps between the headgroups of the ligands

is likely to play an important role in the high activity of the catalysts without pre-

treatment.

3.4.9 Stability of Supported Nanocrystals During the Reaction

It was determined that the metal nanocrystals do not desorb from the

mesoporous silica upon exposure to the 1-decene during 4 hours of reaction at 75 °C.

After the reaction, the mixture was filtered and the supernatant was recovered and

95

analyzed spectrophotometrically at 500 nm. The absorbance was below the detectable

limit within the experimental uncertainty (<0.005). From an extrapolation of the

calibration curve (Figure 3.10), the concentration of the desorbed iridium

nanocrystals was determined to be less than 0.01 mg/ml, or less than 3% of the total

amount of catalyst. The stability of the catalyst against nanocrystal desorption is

further supported by the fact that the TOF for the second catalytic run was almost the

same as that of the original run, as shown in Table 3.1. Figure 3.11 shows the TEM

micrograph of Ir-MPS after reaction of 1-decene with hydrogen for 4 hours. The

micrograph shows that the particles were well dispersed and did not aggregate,

further indicating the stability of the nanocrystals within mesoporous silica. The

expected desorption of the ligand from the adsorbed nanoparticle surface onto silica

would increase the attractive van der Waals attraction between the particle and the

silica, further enhancing its binding to the surface.

3.4.10 Activation Energy of the Decene Hydrogenation

The apparent activation energy for the decene hydrogenation by Ir-MPS was

measured over a temperature range of 40-75 ˚C as shown in Table 3.1. As shown in

the Arrhenius plot for the TOF16 (Figure 3.12), the activation energy was

approximately 13.5 kcal/mol. To further put our results in perspective, the catalytic

activity of Ir-MPS was compared with a typical industrial hydrogenation catalyst (5%

palladium on Alumina). Decene hydrogenation reactions were performed under

identical conditions with the same mass of metal as shown in Figure 3.9. The rate of

decene hydrogenation is constant for Pd-alumina as was observed above for Ir-MPS

for the same reasons. For Ir-MPS, the activity was 117 mole decene reacted/g Ir/hr

versus only 65 for Pd. For the Pd catalyzed reaction of 1-decene, the activation

energy was 17 kcal/mol.48 The higher activity for Ir-MPS vs. Pd-alumina, by a factor

of 1.8, is caused in part by the high surface area on the order of 80 nm2 for 5 nm in

diameter Ir nanocrystals. The greater fraction of surface atoms of metal, as

characterized by the higher surface area, increases the number of active sites.

96

0.04

0.05

0.06

0.07

0.08

0.09

0.1

0.11

0.12

0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

Abs

orba

nce

at 5

00 n

m

Conc of Ir in decene (mg/ml)

Figure 3.10

Calibration curve of UV- Vis absorbance versus concentration for iridium nanocrystal

dispersions in 1-decene, measured at 500 nm.

97

Figure 3.11

TEM micrographs of Ir-MPS (without any pretreatment) after 1-decene

hydrogenation.

98

0.5

1

1.5

2

2.5

3

2.9 2.95 3 3.05 3.1 3.15 3.2

ln (T

OF)

1/T * 1000 (Kelvin-1)

Figure 3.12

Arrhenius plot of decene hydrogenation over untreated Ir-MPS.

99

3.5 Conclusions

Herein, we show a novel approach for designing catalysts with high activities

by infusing pre-synthesized iridium nanocrystals, with well-defined surface

morphologies, into ordered mesoporous silica. Decoupling the synthesis step and the

infusion step enables control of the iridium nanocrystal size, morphology, catalytic

properties and dispersibility within the mesopores. A dispersion of iridium

nanocrystals may be infused into mesoporous silica by expanding the solvent,

toluene, with supercritical CO2. High loadings of the nanocrystals are achieved by

tuning the solvent strength with CO2, in order to strengthen the interactions of the

nanocrystals with the pore walls, without flocculating the nanocrystals in the bulk

solvent. With Z-contrast STEM using a HAADF detector, it was demonstrated clearly

that the iridium nanocrystals are highly dispersed within the pores, with relatively few

particles on the external particle surface. The ability to maintain the pre-designed

nanocrystal size during infusion and to achieve high loadings within the pores, as

characterized by Z-contrast TEM, are important milestones in the development of

highly active and selective catalysts. The high surface area of the nanocrystals

produced high turnover frequencies (TOF) of 16 s-1 for decene hydrogenation, even

without thermal pre-treatment to remove the TOAB ligands. The TOF was higher

than the initial TOF for homogeneous catalysis30 with the same nanocrystals

dissolved in 1-decene. Thermal calcination at 500 °C or CO2 annealing at 275 bars

had no impact on catalyst activity, which is not surprising given the high activity of

the original catalyst.

The nanocrystals remain dispersed in the pores and do not aggregate or desorb

into the solvent during the reaction, as observed by post reaction TEM and UV-Vis

spectrophotometric analysis of the product solution. Weakly bound ligands on the

nanocrystal surface, such as TOAB, may be expected to migrate to the vacant sites on

the silica, significantly reducing the ligand coverage on the surface and hence

increasing the catalyst activity. Additionally, tetraoctylammonium bromide ligands

100

with bulky tails will leave gaps between the head groups allowing small hydrogen

molecules to chemisorb on the surface and react very close to the binding site of the

TOAB. Desorption of ligands from the adsorbed nanoparticle surface, onto the

support and into the decene during reaction, will enhance metal binding to the surface

and aid catalyst stability. This general infusion method for controlling the size,

surface morphology and stabilizing ligands of the nanocrystals, independently of the

synthetic procedure for the mesoporous material, offers broad opportunities in

designing novel highly active catalysts.

3.6 References

(1) Scott, R. W. J.; Wilson, O. M.; Crooks, R. M. Synthesis,

Characterization, and Applications of Dendrimer-Encapsulated Nanoparticles.




299, 1698-1702.

(3) Kanaras, A. G.; Soennichsen, C.; Liu, H.; Alivisatos, A. P. Controlled

Synthesis of Hyperbranched Inorganic Nanocrystals with Rich Three-Dimensional

Structures. Nano Letters 2005, 5, 2164-2167.

(4) Farias, M. H.; Gellman, A. J.; Somorjai, G. A.; Chianelli, R. R.; Liang,

K. S. The coadsorption and reactions of sulfur, hydrogen and oxygen on clean and

sulfided molybdenum(100) and on molybdenum disulfide(0001) crystal faces.

Surface Science 1984, 140, 181-196.

(5) Strongin, D. R.; Carrazza, J.; Bare, S. R.; Somorjai, G. A. The

importance of C7 sites and surface roughness in the ammonia synthesis reaction over

iron. Journal of Catalysis 1987, 103, 213-215.

101




(7) Lensveld, D. J.; Mesu, J. G.; van Dillen, A. J.; de Jong, K. P. The

application of well-dispersed nickel nanoparticles inside the mesopores of MCM-41

by use of a nickel citrate chelate as precursor. Studies in Surface Science and

Catalysis 2002, 143, 647-657.

(8) Zong, Y.; Watkins, J. J. Deposition of copper by the H2-assisted

reduction of Cu(tmod)2 in supercritical carbon dioxide: Kinetics and reaction

mechanism. Chemistry of Materials 2005, 17, 560-565.

(9) Singh, U. K.; Albert Vannice, M. Liquid-Phase Hydrogenation of

Citral over Pt/SiO2 Catalysts. Journal of Catalysis 2000, 191, 165-180.

(10) Haran, N. P.; Bhagade, S. S.; Nageshwar, G. D. Catalysis by ion

exchange resins - alkylation. Chemicals & Petro-Chemicals Journal 1978, 9, 21-28.

(11) Hayek, K.; Goller, H.; Penner, S.; Rupprechter, G.; Zimmermann, C.

Regular Alumina-Supported Nanoparticles of Iridium, Rhodium and Platinum Under

Hydrogen Reduction: Structure, Morphology and Activity in the Neopentane

Conversion. Catalysis Letters 2004, 92, 1-9.

(12) Serp, P.; Hierso, J. C.; Feurer, R.; Kihn, Y.; Kalck, P.; Faria, J. L.;

Aksoylu, A. E.; Pacheco, A. M. T.; Figueiredo, J. L. Single-step preparation of

activated carbon supported platinum catalysts by fluidized bed organometallic

chemical vapor deposition. Carbon 1999, 37, 527-530.

(13) Chu, H. P.; Lei, L.; Hu, X.; Yue, P.-L. Metallo-Organic Chemical

Vapor Deposition (MOCVD) for the Development of Heterogeneous Catalysts.

Energy & Fuels 1998, 12, 1108-1113.

(14) Konya, Z.; Puntes, V. F.; Kiricsi, I.; Zhu, J.; Alivisatos, P.; Somorjai,

G. A. Novel Two-Step Synthesis of Controlled Size and Shape Platinum

Nanoparticles Encapsulated in Mesoporous Silica. Catalysis Letters 2002, 81, 137-

140.

102







Chemistry B 2005, 109, 2192-2202.






(19) de Soler-Illia, G. J.; Sanchez, C.; Lebeau, B.; Patarin, J. Chemical

Strategies To Design Textured Materials: from Microporous and Mesoporous Oxides

to Nanonetworks and Hierarchical Structures. Chemical Reviews (Washington, DC,

United States) 2002, 102, 4093-4138.

(20) Shyue, J.-J.; De Guire, M. R. Single-Step Preparation of Mesoporous,

Anatase-Based Titanium-Vanadium Oxide and Its Application. Journal of the


(21) Shah, P. S.; Holmes, J. D.; Doty, R. C.; Johnston, K. P.; Korgel, B. A.

Steric Stabilization of Nanocrystals in Supercritical CO2 Using Fluorinated Ligands.

J. Am. Chem. Soc. 2000, 122, 4245-4246.

(22) Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. Nanocrystal

Arrested Precipitation in Supercritical Carbon Dioxide. J. Phys. Chem. B 2001, 105,

9433-9440.

(23) Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. Role of Steric

Stabilization on the Arrested Growth of Silver Nanocrystals in Supercritical Carbon

Dioxide. J. Phys. Chem. B 2002, 106, 12178-12185.

103

(24) Lu, X.; Ziegler, K. J.; Ghezelbash, A.; Johnston, K. P.; Korgel, B. A.

Synthesis of Germanium Nanocrystals in High Temperature Supercritical Fluid

Solvents. Nano Letters 2004, 4, 969-974.

(25) Bell, P. W.; Anand, M.; Fan, X.; Enick, R. M.; Roberts, C. B. Stable

Dispersions of Silver Nanoparticles in Carbon Dioxide with Fluorine-Free Ligands.

Langmuir 2005, 21, 11608-11613.

(26) Midgley, P. A.; Weyland, M.; Thomas, J. M.; Johnson, B. F. G. Z-

Contrast tomography: a technique in three-dimensional nanostructural analysis based

on Rutherford scattering. Chemical Communications (Cambridge, United Kingdom)

2001, 907-908.

(27) Midgley, P. A.; Weyland, M. 3D electron microscopy in the physical

sciences: the development of Z-contrast and EFTEM tomography. Ultramicroscopy

2003, 96, 413-431.

(28) Yoo, J. W.; Hathcock, D. J.; El-Sayed, M. A. Propene hydrogenation

over truncated octahedral Pt nanoparticles supported on alumina. Journal of Catalysis

2003, 214, 1-7.

(29) Scott, R. W. J.; Wilson, O. M.; Crooks, R. M. Titania-Supported Au

and Pd Composites Synthesized from Dendrimer-Encapsulated Metal Nanoparticle

Precursors. Chemistry of Materials 2004, 16, 5682-5688.

(30) Stowell, C. A.; Korgel, B. A. Iridium Nanocrystal Synthesis and

Surface Coating-Dependent Catalytic Activity. Nano Letters 2005, 5, 1203-1207.


Chmelka, B. F.; Stucky, G. D. Triblock copolymer syntheses of mesoporous silica

with periodic 50 to 300 angstrom pores. Science (Washington, D. C.) 1998, 279, 548-

552.

(32) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. Nonionic

triblock and star diblock copolymer and oligomeric surfactant syntheses of highly

ordered, hydrothermally stable, mesoporous silica structures. Journal of the American


104

(33) Ryoo, W.; Webber, S. E.; Johnston, K. P. Water-in-Carbon Dioxide

Microemulsions with Methylated Branched Hydrocarbon Surfactants. Industrial &

Engineering Chemistry Research 2003, 42, 6348-6358.

(34) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. The determination of pore

volume and area distributions in porous substances. I. Computations from nitrogen

isotherms. Journal of the American Chemical Society 1951, 73, 373-380.

(35) Tanev, P. T.; Pinnavaia, T. J. Mesoporous Silica Molecular Sieves

Prepared by Ionic and Neutral Surfactant Templating: A Comparison of Physical

Properties. Chemistry of Materials 1996, 8, 2068-2079.

(36) Saunders, A. E.; Sigman, M. B., Jr.; Korgel, B. A. Growth Kinetics

and Metastability of Monodisperse Tetraoctylammonium Bromide Capped Gold

Nanocrystals. Journal of Physical Chemistry B 2004, 108, 193-199.

(37) Mattoussi, H.; Cumming, A. W.; Murray, C. B.; Bawendi, M. G.;

Ober, R. Properties of CdSe nanocrystal dispersions in the dilute regime: Structure

and interparticle interactions. Physical Review B: Condensed Matter and Materials

Physics 1998, 58, 7850-7863.

(38) McLeod, M. C.; Anand, M.; Kitchens, C. L.; Roberts, C. B. Precise

and Rapid Size Selection and Targeted Deposition of Nanoparticle Populations Using

CO2 Gas Expanded Liquids. Nano Letters 2005, 5, 461-465.



(40) Hourri, A.; St-Arnaud, J. M.; Bose, T. K. Dielectric and pressure virial

coefficients of imperfect gases: CO2-SF6 mixtures. Journal of Chemical Physics

1997, 106, 1780-1785.

(41) Burns, R. C.; Graham, C.; Weller, A. R. M. Mol. Phys. 1986, 59, 41.

(42) Rubio, J. E. F.; Arsuaga, J. M.; Taravillo, M.; Baonza, V. G.; Caceres,

M. Refractive index of benzene and methyl derivatives: temperature and wavelength

dependencies. Experimental Thermal and Fluid Science 2004, 28, 887-891.

105

(43) Williams, J. W.; Krchma, I. J. Dielectric constants of binary mixtures.

J. Am. Chem. Soc. 1926, 48, 1888-1896.





(45) Le Bars, J.; Specht, U.; Bradley, J. S.; Blackmond, D. G. A Catalytic

Probe of the Surface of Colloidal Palladium Particles Using Heck Coupling

Reactions. Langmuir 1999, 15, 7621-7625.

(46) Ha, S.; Larsen, R.; Masel, R. I. Performance characterization of Pd/C

nanocatalyst for direct formic acid fuel cells. Journal of Power Sources 2005, 144,

28-34.




(48) Goyal, H. B.; Gupta, A. K. Kinetics of liquid-phase hydrogenation of

C10-C16 olefins on a palladium-alumina catalyst. Indian Journal of Technology

1989, 27, 360-362.

106

Chapter 4

Stable Ordered FePt-Mesoporous Silica Catalysts with High

Loadings

A new concept is presented to form catalysts by infusion of pre-synthesized

bimetallic nanocrystals into ordered mesoporous supports. For pre-synthesized FePt

nanocrystals (<4 nm), coated with oleic acid and oleylamine ligands in toluene, high

loadings above 10 wt % were achieved in 10 minutes. The strong metal-support

interactions were favored by the low coverage of the weakly bound ligands. The

nanocrystals were highly dispersed within the pores as indicated by HAADF-STEM

and X-ray diffraction (XRD) and stable against sintering at 700 °C and desorption

into polar and nonpolar solvents at room temperature. A phase transformation from a

disordered phase (FCC) to ordered phase (FCT) was observed upon thermal

annealing at 700 °C without sintering, as confirmed by convergent beam electron

diffraction and XRD. The calcined FePt catalyst exhibited six fold higher catalyst

activity (TOF =30 s-1) than that of a commercial Pd-alumina catalyst for liquid 1-

decene hydrogenation and was stable for multiple reactions. The decoupling of

nanocrystal synthesis and infusion provides exquisite control of the nanocrystal size,

alloy structure, binding to the support and dispersibility within the pores, offering

broad opportunities for enhanced catalyst activities, selectivities and stabilities.

Large portions of this chapter have been previously published as Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.; Johnston, K. P., Chem. Mater., 2008, ASAP.

107

4.1 Introduction

Highly active and stable catalysts may be designed by controlling the size and

composition of supported metal nanoparticles.1-8 For supported catalysts synthesized

by precursor reduction,1,9 chemical vapor/fluid deposition10 or coprecipitation,5,11 it is

difficult to control the size and surface morphology of the metal clusters as well as

the binding to the surface. An emerging concept in catalyst design is to pre-synthesize

metal nanocrystals coated with stabilizing ligands to control the morphology and then

to infuse the particles onto high surface area ordered mesoporous supports.3,12,13 The

highly uniform and tunable size and curvature of 1-D and 3-D pores in ordered

mesoporous substrates14 facilitates dispersion and stability of the nanoparticles.15-

17,18,19

The metal loading of catalysts formed with pre-synthesized nanocrystals has

been found to be highly dependent on the stabilizing ligands on the particle surface.

Infusion of Pt nanocrystals coated with polyvinylpyrollidone (PVP) ligands into

mesoporous silica with sonication yielded a metal loading of approximately 1 wt %.3

Gold12 nanocrystals capped with dodecanethiol and iridium20 with

tetraoctylammonium bromide were loaded into mesoporous silica up to 2.5 wt %.

The low molecular weight ligands facilitated ligand removal. The high catalytic

activity for 1-decene hydrogenation without calcination suggests that the weakly

bound ligand tetraoctylammonium bromide did not cover the active sites. Recently,

Jung et al21 showed that FePt nanocrystals could be infused in hydrophobic

mesoporous silica, although the metal loading was not given. An alternative approach

is to synthesize the mesoporous support in the presence of a nanocrystal

dispersion.22,23 However, the presence of the nanocrystals can interfere with the

mesopore structure and conversely, the synthesis of the mesoporous material can alter

the morphology and surface composition of the nanocrystals. To date, high metal

loadings >10 wt % have received little attention for the infusion technique, but would

be of interest in various applications.6,24

108

Multimetallic alloy nanocrystals often enable enhanced catalyst activities as a

consequence of synergistic interactions between atoms25,26 including formation of

ordered intermetallic nanocrystals.6,26,27 The high annealing temperatures required to

form these intermetallic alloys often leads to particle sintering.28-30 Traditionally,

bimetallic nanocrystals have been synthesized by reduction of multiple precursors

over a metal support28 or by decomposition of multimetal carbonyl precursors31.

Thomas and coworkers decomposed specially synthesized multi-metal carbonyl

precursors thermally over mesoporous silica to form nanoclusters25,32 composed of

10-20 atoms with a size of ~1 nm. Whereas these catalysts have great potential, larger

metal nanocrystals ≥3 nm are in some cases more catalytically active than small metal

clusters.33

The reduction of multiple metal precursors within porous supports often

produces, high polydispersity and high variability in particle composition.34 Without

the use of stabilizing ligands, the nanocrystal size may exceed 10 nm and even reach

the size of the pores of the support, as has been observed for FePt,30 NiFe28 and

PtCo35 in mesoporous silica. For example, FePt nanoparticles grew to a size of 6 nm

in mesoporous silica with 9 nm pores. Here the nanoparticles may block some of the

pores. The varying rates of nucleation and growth for each metal precursor over the

heterogeneous sites on the supports may lead to highly variable composition.34 For

example, the composition varied for Pt:Co between 44:56 and 97:3, and the size

reached 6-10 nm.35 Potentially, these limitations may be overcome by infusion of pre-

synthesized bimetallic nanocrystals into the porous support, although this approach

has received little attention. For example, FePt nanocrystals have been synthesized

with precise control over the nanocrystal size, shape and composition by arrested

growth precipitation.2,36,37

The primary objectives of this study were to: (1) design active and stable

catalysts at high loadings above 10 wt % by infusion of pre-synthesized FePt

nanocrystals on mesoporous silica and (2) to transform the crystal structure to a face

centered tetragonal (FCT) intermetallic alloy upon annealing at 700 oC without

109

perturbing the size. These objectives were facilitated by designing stabilizing ligands

to strengthen short-ranged interactions between bare Pt and silica, which have been

characterized by XPS,38 EELS39 and wettability studies40 at high temperatures.

Because various types of ligands such as dodecanethiol12 and polyvinylpyrrolidone3

screen these interactions, we chose to investigate weakly bound oleic acid and

oleylamine low molecular weight ligands that expose a greater fraction of the metal

surface domains. The large interfacial area between the spherical nanocrystals and

concave mesopores with similar diameters favors nanocrystal-support interactions,

relative to other support geometries.

The size and composition of the nanocrystals was determined by X-ray

diffraction and high angle annular dark field-scanning transmission electron

microscopy (HAADF-STEM). The kinetics and equilibrium loading of nanocrystals

in mesoporous silica were studied as a function of the initial nanocrystal

concentration over a period from 10 minutes to 16 hours. The nanocrystals within the

pores were differentiated from those on the external silica surface by HAADF-STEM.

Finally, the catalyst activity and stability for liquid phase hydrogenation of 1-decene

were examined both before and after calcination.

4.2 Experimental Section

4.2.1 FePt Nanocrystal Synthesis

3–4 nm spherical FePt nanocrystals were synthesized with the high

temperature arrested precipitation method developed be Sun et al.36 The reaction was

carried out under a N2 environment with a Schlenk Line. The nanocrystals were

prepared by mixing 0.098 g (0.25 mmol) platinum (II) acetylacetonate, 0.195 g (0.77

mmol) 1,2–hexadecanediol, and 10 mL of dioctyl ether in a 50 mL three-neck flask.

The stirred solution was heated to 100 °C under N2 flow, and 0.065 mL (0.50 mmol)

iron (0) pentacarbonyl, 0.085 mL (0.26 mmol) oleylamine, and 0.08 mL (0.25 mmol)

oleic acid were injected. The temperature was raised at a rate of 20 °C/min to the

110

reflux temperature of dioctyl ether (297 °C) and the reaction solution was incubated

for 30 minutes. The solution was allowed to cool to room temperature and excess

ethanol was used to precipitate the FePt nanocrystals which were recovered after

centrifugation (10 min, 8000 rpm). The nanocrystals were washed twice by

redispersing them in hexane and repeating the precipitation procedure. No additional

size selective precipitation was necessary to obtain the monodispersity.


Mesoporous silica SBA-15 was synthesized with a block copolymer template

with the exact procedure of Zhao et al.18,41 A solution of 1.6 g EO20PO70EO20:7 mL

HCl:8 mL TEOS:38 mL H2O was prepared, stirred for 24 h at 50 °C and then heated

at 100 °C for 24 h. The resulting slightly yellow powder was dispersed in ethanol,

filtered and then again heated at 100 °C for 24 h. Calcination was carried out by

slowly increasing the temperature from room temperature to 550 °C in 4 h and

heating at 550 °C for 6 h.

4.2.3 Infusion of FePt into Mesoporous Silica

To prepare the supported nanocrystal composites, 10 mg of mesoporous silica

was mixed with 3 mL of FePt nanocrystal dispersion in toluene of known

concentrations from 0.25 to 2 mg/mL. The contents were then stirred for ~10 min.

After stirring, the mesoporous silica infused with FePt nanocrystals was separated

from the nanocrystal supernatant by filtration with a 0.2 μm PTFE filter. The extent

of incorporation of FePt nanocrystals in the silica was determined by subtracting from

the initial mass of the dispersed nanocrystals in toluene, the final mass of the

nanocrystals recovered in the supernatant after filtration.12 The absorbance of the FePt

nanocrystals dispersed in toluene before and after infusion was measured at a

wavelength of 500 nm using a Cary 500 UV-Vis-NIR spectrophotometer with an

optical path length of 1 cm. A standard calibration curve was generated with known

111

concentrations of FePt nanocrystals in toluene as shown in Appendix A, Figure A.6.

The experiments were repeated three times to check for the reproducibility in the

nanocrystal loading, which was on the order of ± 0.2 in wt % unit. Total uncertainty

in loadings determined with this technique was approximately 12% as explained in

detail previously.20 The uncertainty for this method is comparable to that for

elemental analysis (10% error). The stability of the nanocrystals immobilized within

the pores was studied by suspending the nanocomposite in pure ethanol. The

suspension was sonicated for 2 h after which the nanocomposite was separated by

vacuum filtration, redispersed in ethanol and analyzed by using UV-Vis spectroscopy.

4.2.4 Calcination

To modify the crystal structure of FePt, nanocomposites were calcined at 700

°C in N2/H2 atmosphere for 4 h. Temperature was ramped from ambient to 700 °C in

20 min. To activate the FePt for catalysis, nanocomposites were calcined in stagnant

air at 500 °C for 4 h. Temperature was ramped from ambient to 500 °C in 15 min.

The nanocomposites were washed for 5 min in ethanol.

4.2.5 Transmission Electron Microscopy

Microscopy was performed on a FEI TECNAI G2 F20 X-TWIN TEM using a

high-angle annular dark field detector (HAADF). HAADF images were obtained with

a 2 Å resolution. High resolution transmission electron microscopy (HRTEM) was

performed using a JEOL 2010F TEM operating at 200 kV. Pictures were obtained at

the optimum defocus condition. Nanocomposites were deposited from a dilute

ethanol solution onto 200 mesh carbon-coated copper TEM grids.

4.2.6 X-ray Diffraction

Wide angle X-ray diffraction was performed with samples on a quartz slide

using a Bruker-Nonius D8 Advance diffractometer. Samples were typically scanned

for 12 h at a scan rate of 12 deg/min with 0.02 deg increments. The average

112

nanocrystal size was estimated from the Scherrer equation by fitting the XRD line

shapes using JADE software (by Molecular Diffraction Inc).

4.2.7 Catalysis

1-decene hydrogenation reactions were carried out in a 25 ml round bottom

flask batch reactor as reported previously.20 Supported FePt nanocrystals were added

to neat 1-decene to create a 2000:1 decene to iridium mass ratio. Reaction conversion

was identified using GC-MS.

4.3. Results

4.3.1 Structure of Nanocomposites after Infusion of FePt Nanocrystals

Figure 4.1 shows a representative high resolution TEM of FePt of the original

nanocrystals prior to infusion, along with a histogram of the particle number

distribution. The particles are spherical in shape and crystalline as can be seen from

the lattice fringes in HRTEM. The lattice spacing is estimated to be 2.24 Ǻ, which

matches to the (111) plane as reported by Sun.36 Upon counting 250 particles, the

average diameter was found to be 3.6 ± 0.5 nm as shown in the histogram. The

mesoporous silica as synthesized is of type SBA-15 with hexagonally ordered 1D

cylindrical mesopores. The mesoporous silica had a surface area of 540 m2/g, a mean

pore diameter of 5.5 nm and a wall thickness of 4.3 nm (Appendix A, Figure A.1 and

A.2 in supplementary information).

The pre-synthesized FePt nanocrystals were infused into the mesoporous

silica for 10 min. Figure 4.2 shows the loadings reached 14 wt % as a function of

initial nanocrystal concentration in toluene. The loading markedly exceeds previous

values which was a maximum of 2.5 wt % using scCO2 as an anti-solvent.12 The

loadings may be tuned linearly as a function of the initial FePt concentration. At low

initial concentrations (~0.25 mg/ml), nearly all of the FePt nanocrystals in the bulk

113

(a)

0

20

40

60

80

100

2 2.5 3 3.5 4 4.5 5 5.5 6

Num

ber

of P

artic

les

FePt Diameter (nm)

(b)

Figure 4.1

(a) HRTEM images of pre-synthesized FePt nanocrystals prior to infusion. (b)

Histogram of nanocrystal size. Average diameter is 3.6 ± 0.5 nm. 250 nanocrystals

were counted for the histogram.

114

6

8

10

12

14

16

0 0.5 1 1.5 2 2.5

FePt

Loa

ding

(wt%

)

Initial Concentration FePt (mg/mL)

Figure 4.2

Loading of FePt nanocrystals within mesoporous silica in 10 minutes as a function of

initial concentration of nanocrystals.

115

were adsorbed within the mesoporous silica, leaving a clear supernatant. At a higher

initial concentration of 2 mg/ml, the final concentration of nanocrystals after infusion

was only ~0.37 mg/ml in the supernatant, indicating over ¾ of the original

nanocrystals were adsorbed. The loadings for FePt nanocrystals were measured as a

function of time to determine the kinetics of nanocrystal infusion (Appendix A,

Table A.1). For an initial concentration of 0.7 mg/ml, loadings increased from 8.9 wt

% to 11.4 wt % when the time of infusion was increased from 10 min to 16 h,

suggesting an approach to equilibrium. For higher initial concentrations of 2 mg/ml,

the loadings increased slightly from 13.8 wt % to 14.3 wt % over this time interval,

which is within the experimental reproducibility. Thus, the level of loading

approached a near equilibrium value on the order of minutes.

High resolution HAADF-STEM was used to visualize the FePt nanocrystals

supported on the mesoporous silica prior to thermal annealing. The contrast variation

is proportional to the mass density of the region multiplied by the atomic number Z

squared.32 Regions of high atomic number appear in white. The image of the system

may be understood as a convolution of the probe, the SiO2 support and the metal

nanocrystals.

I (total) = Iprobe θ IMetal θ ISiO2 ………………………(4.1)

where the symbol θ represents the convolution operation. The interpretation of the

image can be complicated. However the SiO2 porous structure is periodic and

relatively straightforward to understand and allows the interpretation of the HAADF

images. On the other hand, the particles are located in both black and white fringes,

with somewhat similar sizes in each domain. Figure 4.3(a) shows a planar view of the

FePt nanocrystals aligned within the pores at 7.3 wt % metal loading after infusion.

FePt nanocrystals are observed as spherical bright dots, silica walls appear as white

lines and empty pores appear as dark spaces between the silica walls, as expected due

to differences in atomic number. The nanocrystals are highly dispersed without any

significant aggregation. The FePt nanocrystals associated more with the dark line

regions are more likely to be in the pores and are estimated to be 65% of the

116

(a) (b)

0

10

20

30

40

50

2 2.5 3 3.5 4 4.5 5 5.5

Num

ber

of P

artic

les

FePt nanocrystal size (nm)

(c)

Figure 4.3

TEM images of 7.3 wt % FePt nanocrystals after infusion into mesoporous silica.

(a) HAADF STEM planar view; (b) HRTEM; (c) Histogram of nanocrystal size using

(a) for 100 nanocrystals.

20 nm

117

total infused nanocrystals, while those in the white regions can be either on the

surface or in the pore just above or below the pore wall or in the secondary

interconnecting pores.19 The secondary pores of size up to 3.6 nm are present in SBA-

15 mesoporous silica as reported by Joo and Ryoo. Most of the nanocrystals are

aligned with the pores along the pore wall and show a definite orientation parallel to

the pores, suggesting that they are in the pores, and not the external surface. In

addition, many of the particles in the white domains are also in the pores. The FePt

nanocrystals in the tail end of the distribution with 5 nm metal cores (and 1.8 nm

longs ligands on the metal surface) will not be able to enter the pores with a diameter

of 5.5 nm. These larger nanocrystals may only be adsorbed on the external surface

and appear in the white regions of the pore walls. As a result, the average size of

some of the nanocrystals that appear on the white lines or pore walls is larger (>4 nm)

than that of the nanocrystals in the dark spaces or pores (~ 3.5 nm).

The average interparticle distance between the nanocrystals in Figure 4.3(a) is

between 5-10 nm. Assuming a uniform distribution of nanocrystals for a 7.3 wt %

loading in the cylindrical mesopores, the average interparticle distance was estimated

to be 35 nm. The STEM micrograph is necessarily obtained near an edge of a particle

where the particle is not too thick. Thus, this difference in interparticle distance

suggests that nanocrystals are more concentrated closer to the pore openings than in

the center of the pore, as they have to diffuse longer distances to reach the center of

the pore. The bright field high resolution image of the FePt nanocrystals in Figure

4.3(b) indicates that the nanocrystals retain their size and crystallinity as confirmed

by the presence of lattice fringes. The lattice spacing of these supported FePt was

estimated to be 2.28 Ǻ, which is comparable to the value of 2.24 Ǻ (for (111) plane)

estimated for unsupported nanocrystals.

A calculation of the surface coverage of the nanocrystals may be used to

provide further evidence that the nanocrystals are in the pores and not merely

adsorbed on the external surface. The external surface area of mesoporous silica is

118

extremely small, ~0.3 m2/g based on a particle size estimate of 10 μm. For an

area/particle of 101057.1 −× m2 on the surface, a closely packed monolayer of FePt

nanocrystals would only produce a loading of <0.5 wt %. The observed loading of 10

wt % would require at least 20 layers of nanocrystals, if all the particles were located

on the external surface. As can be seen in Figure 4.4(a), the thickness of the

nanocrystal layer on the external surface is only a few layers thick, far below 20.

Thus, the STEM–HAADF images and this calculation suggest that most of the

particles were adsorbed in the mesopores and not on the surface.

4.3.2 Stability of the FePt Nanocomposites

The stability of the supported FePt nanocrystals towards desorption in polar

and nonpolar solvents were studied in ethanol and toluene separately. 20 mg of 7.3 wt

% FePt in mesoporous silica was dispersed in 5 mL of solvent and sonicated for 2 h at

ambient temperature. Using UV-Vis spectroscopy, the absorbance was less than 0.01.

Thus, a maximum of less than 2.5 % of the total infused nanocrystals for both

solvents could have desorbed, a very minor amount, within the error limit of the

instrument.

The stability of supported FePt nanocrystals towards sintering and

coalescence at high temperature was characterized by annealing the nanocomposite at

700°C as shown in the STEM-HAADF images in Figure 4.4(a). Most of the

nanocrystals are well dispersed without any noticeable aggregation. From the TEM

image in Fig 4.4(a), pore collapse or structural damage is not present for the

mesoporous silica thermally treated at 700 °C. The size of the larger nanocrystals in

the distribution is still within the range of the as synthesized nanocrystal dispersion

shown in Figure 4.3(c). The size distribution of annealed supported nanocrystals as

seen in the histogram (Figure 4.4(b)) was very similar to that of as synthesized

nanocrystals prior to infusion. However, there are a few regions in the STEM image

which show sintering of nanocrystals, as examined in greater detail in the

119

(a) (b)

(c) (d)

(d) (e)

Figure 4.4

7.3 wt % FePt nanocrystals in mesoporous silica calcined at 700 °C for 4 h: (a)

HAADF- STEM; (b) Histogram of nanocrystals after calcination determined from (a).

116 nanocrystals were counted. Average size is 3.7 ± 0.7 nm.; (c) Filtered image of

random portion of (a) to eliminate shot noise; (d) Intensity profile of FePt marked (A)

in image (c); (e) Intensity profile of FePt marked (B) in image (c).

Inte

nsity

(arb

itrar

y u

Length (arbitrary units)

Inte

nsity

(arb

itrar

y u


Inte

nsity

(arb

itrar

y un

its)


Inte

nsity

(arb

itrar

y un

its)


0

10

20

30

40

50

2 3 4 5 6

Num

ber o

f Par

ticle

s

FePt nanocrystal size (nm)

120

supplementary information (Appendix A, Figure A.3). These regions are likely on the

external surface of the silica and not within the pores. The STEM image in Figure

4.4(a) was filtered to eliminate shot noise, as shown in Fig 4.4(c). Due to the fact that

most of the particles are of the same composition, the STEM-HAADF profile

represents the profile of the size and shape of the nanocrystals. The intensity profile

of larger particles in region A (Figure 4.4(d)) can be compared with the smaller

particles in region B (Figure 4.4(e)). The particles in A are more intense and are less

spherical than the particles in B. The size of the particles in A approaches the pore

size of 5 nm. Therefore, it is reasonable to assume that the particles in A wetted the

external surface of the silica during thermal annealing. Figure 4.4(d) shows two small

nanocrystals near each other. In contrast with the large type A particles, the size of

the type B nanocrystals is ~3 nm, well below the pore size. If the B particles were on

the surface of the silica, it is likely that the particle would sinter at the high annealing

temperature. Thus, it is likely that the B particle was in a pore. From a statistical

analysis of the number of particles of type A and B in a sample of 106 nanocrystals,

85 correspond to type B, based on a size cutoff of 4.5 nm. With the STEM analysis of

the samples before and after calcination, along with the calculation of the particle

surface coverage, it is evident a large fraction (at least 80%, a conservative estimate)

of the particles was in the pores. Furthermore, nanoparticles on the external surface

will sinter on annealing as shown in the supplementary figure (Appendix A, Fig A.3)

and on the top edge of the Fig 4.4(a). The small degree of sintering shown in STEM

images and in the XRD results indicate that a large fraction of the particles were

protected in the pores.

4.3.3 Crystal Structure of the Bimetallic FePt after Annealing

To facilitate the ordering of atoms in the supported FePt nanocrystals, the

samples were annealed at 700 °C. The change in crystal structure of FePt nanocrystals

on thermal annealing was observed with wide-angle XRD and convergent beam

electron diffraction (CBED) studies. Figure 4.5(a) gives the XRD

121

0

100

200

300

400

500

600

10 20 30 40 50 60 70 80 90

2 theta

Inte

nsity

Uncalcined FePt on MPS

calcined FePt MPS 700C

(a)

(b) Figure 4.5

(a) XRD diffraction of FePt nanocrystals in silica before and after annealing at 700

°C. The (111) peak to the right shift after annealing supports the phase transition from

FCC phase of uncalcined nanocrystals to FCT (L10). (b) convergent beam electron

diffraction of FePt nanocrystals in silica before and after annealing also confirms

phase transformation from FCC to FCT phase.

220

111 200

220

111

200

Amorphous Silica

311

311

122

patterns of infused FePt nanocrystals in mesoporous silica, before and after thermal

annealing. In the as-synthesized uncalcined nanocomposite, peaks are observed at

2θ= 40.2°, 46.7° and 67.6° and 83° are assignable to the (111), (200), (220) and (311)

orientations for FePt. A broad peak is found at 2θ= 22.5°, which can be assigned to

amorphous silica. The FePt peaks can be indexed to a FCC phase, which is reported

to be a disordered phase.36 From the XRD spectra according to the Scherrer equation,

the nanocrystal size is estimated to be 4.0 nm and is in close agreement with the FePt

particle size before infusion measured using TEM. The small nanoparticle sizes lead

to weaker and broader peaks than for larger particles.

The XRD spectra for the annealed FePt nanocomposite indicates a (111) peak

shift by 0.5°.36 The full width half maxima of (111) peak of annealed FePt

nanocomposite does not show any significant narrowing, as compared to the as-

synthesized sample, indicating coalescence of nanocrystals was insignificant,

consistent with the STEM results. The nanocrystal size after sintering from peak

fitting was ~4.1 nm, close to uncalcined nanocrystal size. Hence, the FePt

nanocrystals underwent ordering to the FCT phase upon thermal annealing without

any appreciable sintering. The (110) peak for FCT phase is very broad and weak but

we used another technique of convergent beam electron diffraction to further study

the phase transformation.

The technique of convergent beam electron diffraction (CBED) was employed

to confirm the structural changes of nanocrystals loaded in the silica. This is the an

accurate method to determine three-dimensional information about the reciprocal

lattice point and space group symmetry details, lattice parameters from fine regions

and atomic positions within a unit cell.42,43 Electron diffraction patterns recorded from

a single supported FePt nanocrystal are shown in Figure 4.5b. The patterns are

indexed for the single nanocrystal before and after thermal treatment at 700 °C, to

identify the change in crystal structure. CBED of a single uncalcined FePt nanocrystal

supported on mesoporous silica confirms the FCC phase of the nanocrystals

123

undergoes a phase transformation to FCT phase upon thermal annealing (Figure

4.5(b)).

4.3.4 Activity and Stability of Catalysts for 1-decene Hydrogenation

The turnover frequency (TOF) for the FePt supported on mesoporous silica,

calculated from moles of 1-decene reacted per second per mole of active sites on the

nanocrystal surface, as described earlier20 is shown in Figure 4.6. The blank

mesoporous silica did not have any catalytic effect for decene hydrogenation.

Conversion for 1-decene by blank silica was <1% after 4 h (wt of decene: wt of silica

= 10:1), and is within the error limits of the GC-MS. The number of active sites for

FePt was calculated for a particle size of 3.6 nm, as determined from TEM in Figure

4.3(a). To ensure complete ligand removal, the catalyst was calcined at 500 °C. The

temperature was much higher (~100 °C ) than where the ligands are removed, as

determined by TGA.44 The reaction was carried out in neat 1-decene. It was zero

order with respect to decene and with a linear conversion in time. A large increase in

catalyst activity for fresh FePt catalyst was observed after thermal oxidation. In order

to test the stability of the FePt catalyst, it was recovered and used for two more

successive reactions. The TOF doubled in the second hydrogenation reaction to a

high level of 30 s-1, which is twice that observed for the 4.5 nm iridium catalyst,20 and

did not show any significant increase for the third reaction. The corresponding

specific activity was 8.5×10-4 moles/g-s, which was 6 times higher than the industrial

catalyst 5 wt % Pd-alumina; however, the metals were different in each case. As the

size of the Pd nanoparticles is not known, TOF for Pd-alumina cannot be calculated.

The uncalcined as synthesized FePt nanocomposite exhibited significant catalyst

activity (~33% relative to a 5% Pd-alumina industrial catalyst). Thus, the passivating

ligands did not block the entire metal surface. The reproducibility in the TOFs was

124

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

5% FePtcalcined -1st time(Fresh)

5% FePtcalcined-2nd time

(Used once)

5% FePtcalcined-3rd time(Usedtwice)

5% FePtuncalcined

(Fresh)

5% FePtuncalcined-2nd time

Tur

n ov

er fr

eque

ncy

(Sec

-1)

Figure 4.6

Turn over frequency for 1-decene hydrogenation by 5 wt % FePt in mesoporous

silica.

125

determined by synthesizing new batch of 5% FePt-silica catalyst and then repeating

the reactions (fresh, washed once and washed 2nd time). The reproducibility for all

cases was within 5% of the 1-decene conversion and is within the detection error of

GC-MS. During infusion, the weakly bound ligands may desorb and adsorb on the

silica with a much larger surface area.20 The catalyst activity doubled relative to the

first reaction when the uncalcined FePt nanocomposites were used for a second

reaction. The performance of the FePt catalyst after first cycle was three times better

than the commercial catalyst (5 % Pt on alumina). The significant increase in the

activity of the calcined catalyst for the second run may likely be attributed to the

removal of carbonaceous contaminants that block active sites from the metal surface.

The lack of change during the third run suggested that these contaminants were

completely removed.

4.4. Discussion

4.4.1 Mechanism for Metal Loading and Stability

The inherent stability of the nanocrystals against desorption into polar

solvents at low temperature or sintering at 700 oC may be attributed to two key

factors, electronic and geometrical effects. Zheng and Stucky have reported that gold

nanocrystals stabilized with thiol ligands adsorbed on non-porous substrates, such as

silica and titania, adsorb weakly and desorb readily into a polar solvent such as

ethanol.7

Metal-support interactions play a key role in both the activity and the thermal

stability of the nanocrystal catalysts. Yu et al. studied the thermal stability of Pt

nanocrystals supported on silica thin films using in-situ TEM.40 The particles began

to change shape at 350 oC, and spread on the surface above 500 oC. The wetting of

the surface by the nanocrystals is attributed to the interfacial mixing of Pt and silica,

as a result of strong metal-support interactions. Klie et al. characterized Pt-silica

specific short range interactions using electron energy loss spectroscopy.39 Strong

silica charge transfer interactions with Pt were identified using X-ray photoelectron

126

spectroscopy (XPS).38 Similar strong interactions between Pt atoms on the surface of

FePt nanocrystals and silica may be expected to aid both the infusion process and the

stabilization of the particles upon thermal annealing.

Long-ranged electrostatic and van der Waals interactions also contribute to the

interaction between a nanoparticle and the support. FePt nanocrystals did not did not

migrate under a field of 166 V/cm, indicating minimal charging. Silica in nonpolar

media acquires negative charges45 but the surface charge density is very small, with

each charge separated by ~100 nm. Metal nanoparticles are extremely polarizable

and form a large induced dipole in presence of a charge.46 The attractive charge-

induced dipole interactions between charged silica and a single metal nanocrystal

(Appendix A, Figure A.4) are proportional to the cube of the nanocrystal radius,

whereas the long-ranged van der Waals interactions were found to be weakly

repulsive in toluene. The total long-ranged interaction energy (charge-induced dipole

+ van der Waals) is only weakly attractive for the 3.6 nm FePt nanocrystals

(Appendix A, Figure A.4). However, with the sparse charge densities for silica (20-

200 charges/μm2) the total loading would be only 0.5 wt % if one nanocrystal

interacted with each charge on the silica. Consistent with this prediction, we have

shown that 2 nm gold nanocrystals stabilized by dodecanethiol adsorb weakly on

silica with small loadings of <0.3 wt %.20 Therefore, the high levels of adsorption for

both the FePt and Pt nanocrystals (up to 20%) in this study must be driven primarily

by the short-ranged interactions described above.38-40

The hydrocarbon capping ligands on the metal surface must not screen the

strong short-ranged interactions with the silica. Amines and fatty acids are known to

bind to Au more weakly relative to thiols.47 For example, surface coverage of

dodecylamine on Pt is 46%48 compared to 75% for dodecanethiol on Ag.49 The lower

surface coverage exposes barer metal surface for the formation of specific short-

ranged interactions with the silica. These short-ranged interactions could explain the

loadings of ~20 wt %, relative to the much lower loadings for gold capped with

dodecanethiol, where the surface coverage is higher. Furthermore, more of the weakly

127

bound amine ligands will migrate from the nanocrystals to the silica surface and

expose even more surface area than in the case of the strongly bound thiols. The large

thermodynamic driving force for adsorption in the FePt case led to near equilibrium

loadings in only 10 minutes versus 24 h for the dodecanethiol capped gold.20 The

short-ranged specific interactions were sufficient to prevent sintering from

mobilization of nanocrystals at high temperatures up to 700 °C and from desorption

into both polar (ethanol) and nonpolar (toluene) solvents upon sonication. The

controlled particle size and composition were favored by the metal-support

interactions ease of ligand removal at moderate temperatures (<400 °C) and the

removal of byproducts of nanocrystal synthesis prior to infusion. The convex surfaces

on the nanocrystals favored contact with the cylindrical mesopores with concave

curvature.50 In contrast, the surfaces of both the metal nanocrystal and support are

often both convex for traditional supports leading to a smaller degree of interfacial

contact. The beneficial geometrical effect of enhanced interfacial area on

nanocomposite stability has been reported by Rolison and co-workers for 1-10 wt %

gold-titania composites.51,52

Previously, loadings only reached ~2.5 wt % for gold nanocrystals, coated

with strongly bound alkanethiols, infused into silica with CO2 as an antisolvent in

toluene.12 The addition of CO2 lowers the Hamaker constant of the intervening

solvent and strengthens the van der Waals interactions of the particles with the silica,

thus raising the nanocrystal loading.12 In the present study, it was not necessary to add

CO2 to drive the nanocrystal loadings since the particle-support interactions were

strong as a consequence of the weakly bound ligands.

4.4.2 Design of Bimetallic Nanocrystals

The results in Figs. 4.1, 4.3 and 4.5 indicate that the size, composition and

crystal structure of the bimetallic nanocrystals were relatively uniform. These

properties may be shown to be favored by the arrested precipitation in the presence of

stabilizing ligands.53,54 Here, the dynamic solvation of the nanocrystal surface by the

128

ligands may be tuned to produce regions of the nanocrystal surface accessible for

growth, while providing enough steric repulsion to inhibit nanocrystal aggregation.

The temperature may be varied to adjust the precursor decomposition and

consequently the supersaturation, nucleation and growth rates. At sufficiently high

temperature and monomer concentrations, the balance between interfacial energy and

supersaturation produces a small critical size where the growth rate is zero.54 Here,

the nanocrystal diameter may be focused to narrow polydispersity, as seen in Figure

4.1, as the growth rate is higher for the smaller nanocrystals than the larger ones.54

The focusing is optimal if the monomer concentration and dynamic solvation

maintain the average nanocrystal size slightly larger than the critical size.

A variety of previous studies indicate that the size and composition of

bimetallic nanocrystals may be controlled over a significant range in the arrested

precipitation technique by varying the temperature and precursor concentrations.36,55-

57 When the rates of precursor decomposition and equilibrium solubilities of the two

resulting metals are similar, formation of binary metal nuclei is favored.36 Even when

these properties are somewhat dissimilar, the nuclei formed from the first metal may

capture newly formed atoms of the second metal before it nucleates, which would

otherwise lead to two separate populations of growing nanocrystals.2 In essence, this

may be considered to be composition focusing by analogy to size focusing. As metal

from the more rapidly decaying precursor is incorporated into the nanoparticle, it will

be depleted in the solution relative to the second metal. Thus, the probability of

adding the second component will increase. This process will focus the overall

composition as well as the particle size. The temperature is a key variable to produce

rapid decomposition of both precursors to achieve focusing of both size and

composition, along with sufficient dynamic solvation by the ligand to arrest growth.

Because our FePt nanocrystals were pre-synthesized, heterogeneous sites on a support

were not present, which could otherwise lead to uncontrolled composition, resulting

from competition between nucleation in solution and on the surface. Once the

particles are pre-synthesized, they may be loaded physically into the pores of the

129

support at completely different concentrations than those used in the synthesis of the

particles. Thus, both the pre-synthesis and infusion steps may be optimized in ways

not available when precursors are reduced within pores of the support.

A variety of factors complicate control over the size, composition and

morphology of bimetallic nanocrystals synthesized by the traditional technique of

precursor reduction directly on a support.34 The precursors may be reduced either

simultaneously or sequentially.58-61 The diffusional resistance of the precursors

through the pores may prevent the production of large supersaturation values that

favor control of nucleation, growth and size focusing. For simultaneous impregnation,

the metal ions may separate chromatographically as they diffuse into the pores,

leading to large variation in nanocrystal composition.58 For sequential impregnation,

the second metal may form a shell around the first metal, or even nucleate on separate

heterogeneous sites on the support.59 High temperature annealing may promote

alloying between metals, but can also cause sintering of the nanoparticles.58 Without

the use of stabilizing ligands, dynamic solvation is not present to aid size and

composition focusing. The precursor concentrations of the metals must be chosen to

control the particle size and composition, as well as the final metal loadings on the

support. For example, low precursor concentrations for low loadings may produce

insufficient supersaturation and thus defocusing,54 whereas high loadings may require

long reaction times, which may lead to multiple nucleation events resulting in

polydispersity. Another challenge is that the reactants and/or by-products formed

during the reduction step may prevent adhesion of the metal particles on the support,

resulting in limited catalyst stability. In the pre-synthesis technique, the particles may

be isolated from unwanted byproducts, cleaned, and separated by size, and the ligands

may be exchanged if desired prior to infusion.

Thermal annealing above ~550 °C may be utilized to order the atoms in the

FePt nanocrystals to form an intermetallic alloy.36 The reduction in the melting point

of the nanocrystals can cause spreading of the nanocrystals as they wet the support.40

The resulting surface diffusion may lead to coalscence.62 This undesired coalescence

130

appeared to be mitigated with the pre-synthesized nanocrystals, indicating strong

metal-support interactions (Figure 4.4). Ligand removal at moderate temperatures

below the melting point further strengthens interactions between the bare metal and

the support.

High temperature annealing is not always needed to make ordered

multimetallic nanocrystals by the “polyol process”. Ordered intermetallic

nanocrystals have been synthesized in solvents including ethylene glycol63-65 or

tetraethylene glycol66,67 with a control over the size at 100-300 °C. These ordered

nanoparticles may then be infused into the support to form catalysts without the need

for high temperature thermal annealing.

4.4.3 Catalytic Activity of Supported Nanocrystals

FePt catalysts were highly active and stable for 1-decene hydrogenation for at

least three reaction cycles. The enhanced stability may be attributed to the strong

short-ranged metal-support interactions, which prevent desorption or

coalescence/aggregation of the metals during the reaction. For the hydrogenation of

1-hexene, the TOF of bimetallic Pd-Ru nanoparticles exceeded by 10 times that of

individual Pd and Ru nanoparticles indicating synergism between these two types of

atoms.68 Similar effects have been reported for other bimetallic catalysts.69 The

uniform composition of the pre-synthesized bimetallic nanocrystals in the current

study and the ability to control the intermetallic alloy structure may be expected to be

beneficial for achieving these types of synergistic effects on catalytic activity.

The FePt particle size formed by precursor reduction of 6 nm in mesoporous

FePt SBA-15 silica without ligands was much larger than the 3.6 nm particles formed

in the current study by infusion of ligand stabilized particles. Furthermore, upon

ordering the FePt by thermal annealing at 650 °C, the size increased modestly from 6

to 7.2 nm. As the particles approach the pore size of the silica, blocking of pores may

limit mass transfer during catalysis. Furthermore, the surface area/volume is two

times larger for our 3.6 nm particles versus 7.2 nm particles.

131

Reducing the size of the metal nanoparticles below 5 nm significantly

increases the fraction of surface atoms and hence the mass catalytic activity. For the

small 3.6 nm FePt nanocrystals, the morphologies of the crystal facets may be

designed to produce more active surfaces than for large particles.70 For example,

atoms present on the jagged “step edges” of small gold nanoparticles (<5 nm) were

found to be highly catalytically active for CO oxidation.2 Pre-synthesis of the

bimetallic nanocrystals may enable greater control over these types of active sites

versus traditional precursor reduction techniques, with the potential of higher catalyst

activity and selectivity.

4.5 Conclusions

Herein, we have demonstrated a novel method to design highly stable and

active catalysts composed of bimetallic nanocrystals on mesoporous supports with a

controlled intermetallic alloy structure. The catalysts were formed by infusion of pre-

synthesized nanocrystals with well-controlled size and composition. The arrested

growth precipitation favors focusing of the size (<4 nm) as well as the composition of

the nanocrystals. The stabilizing ligands control dynamic solvation, supersaturation,

nucleation and growth, as reflected in the small nanocrystal size. The low surface

coverage of the weakly bound ligands resulted in strong metal-support interactions40

as indicated by high metal loadings above 10 wt % in only 10 min and stability

against sintering upon thermal annealing at 700 °C. These interactions are further

increased by the enhanced interfacial contact area between the metal nanocrystals and

concave mesopores with similar radii. For strongly binding ligands such as

dodecanethiol20 or polymers such as polyvinylpyrrolidone3 with high surface

coverages, high loadings were not achieved. According to HAADF-STEM, more than

80% of the nanocrystals were within the pores, relative to the external surface,

consistent with the resistance to sintering at 700 °C. Finally, the calcined FePt

catalyst was found to exhibit six fold higher catalyst activity than commercial Pd-

132

alumina catalyst for liquid 1-decene hydrogenation and was stable for multiple

reactions. The decoupling of nanocrystal synthesis and infusion provides exquisite

control of the nanocrystal size, alloy structure, binding to the support and

dispersibility within the pores, relative to conventional precursor reduction

techniques. The ability to design catalyst morphology and crystal structure with pre-

synthesized bimetallic nanocrystals offers novel opportunities for enhanced catalyst

activities, selectivities and stabilities.

4.6 References




(2) Chan, K.-Y.; Ding, J.; Ren, J.; Cheng, S.; Tsang, K. Y. Supported

mixed metal nanoparticles as electrocatalysts in low temperature fuel cells. Journal of

Materials Chemistry 2004, 14, 505-516.




Chemistry B 2005, 109, 2192-2202.

(4) Akita, T.; Okumura, M.; Tanaka, K.; Kohyama, M.; Haruta, M.

Analytical TEM observation of Au nano-particles on cerium oxide. Catalysis Today

2006, 117, 62-68.

(5) Meunier, F. C.; Tibiletti, D.; Goguet, A.; Shekhtman, S.; Hardacre, C.;

Burch, R. On the complexity of the water-gas shift reaction mechanism over a

Pt/CeO2 catalyst: Effect of the temperature on the reactivity of formate surface

species studied by operando DRIFT during isotopic transient at chemical steady-state.

Catalysis Today 2007, 126, 143-147.

133




247.

(7) Zheng, N.; Stucky, G. D. A General Synthetic Strategy for Oxide-

Supported Metal Nanoparticle Catalysts. Journal of the American Chemical Society

2006, 128, 14278-14280.

(8) Kang, S.; Jia, Z.; Zoto, I.; Reed, D.; Nikles, D. E.; Harrell, J. W.;

Thompson, G.; Mankey, G.; Krishnamurthy, V. V.; Porcar, L. Sintering behavior of

spin-coated FePt and FePtAu nanoparticles. Journal of Applied Physics 2006, 99,

08N704/701-708N704/703.

(9) Koh, S.; Toney, M. F.; Strasser, P. Activity-stability relationships of

ordered and disordered alloy phases of Pt3Co electrocatalysts for the oxygen

reduction reaction (ORR). Electrochimica Acta 2007, 52, 2765-2774.

(10) Song, Z.; Cai, T.; Hanson, J. C.; Rodriguez, J. A.; Hrbek, J. Structure

and Reactivity of Ru Nanoparticles Supported on Modified Graphite Surfaces: A

Study of the Model Catalysts for Ammonia Synthesis. Journal of the American


(11) Li, Y.; Fu, Q.; Flytzani-Stephanopoulos, M. Low-temperature water-

gas shift reaction over Cu- and Ni-loaded cerium oxide catalysts. Applied Catalysis,

B: Environmental 2000, 27, 179-191.




(13) Tsung, C.-K.; Hong, W.; Shi, Q.; Kou, X.; Yeung, M. H.; Wang, J.;

Stucky, G. D. Shape- and orientation-controlled gold nanoparticles formed within

mesoporous silica nanofibers. Advanced Functional Materials 2006, 16, 2225-2230.

(14) Lee, J.; Kim, J.; Hyeon, T. Recent progress in the synthesis of porous

carbon materials. Advanced Materials (Weinheim, Germany) 2006, 18, 2073-2094.

134

(15) Morey, M. S.; Davidson, A.; Stucky, G. D. Silica-based, cubic

mesostructures: synthesis, characterization and relevance for catalysis. Journal of

Porous Materials 1998, 5, 195-204.

(16) Hanrahan, J. P.; Copley, M. P.; Ziegler, K. J.; Spalding, T. R.; Morris,

M. A.; Steytler, D. C.; Heenan, R. K.; Schweins, R.; Holmes, J. D. Pore Size

Engineering in Mesoporous Silicas Using Supercritical CO2. Langmuir 2005, 21,

4163-4167.





Chmelka, B. F.; Stucky, G. D. Triblock copolymer syntheses of mesoporous silica

with periodic 50 to 300 angstrom pores. Science (Washington, D. C.) 1998, 279, 548-

552.



nanoparticles. Nature (London, United Kingdom) 2001, 412, 169-172.




6239-6249.

(21) Jung, J. S.; Lim, J. H.; Malkinski, L.; Vovk, A.; Choi, K. H.; Oh, S. L.;

Kim, Y. R.; Jun, J. H. Magnetic properties of structurally confined FePt nanoparticles

within mesoporous nanotubes. Journal of Magnetism and Magnetic Materials 2007,

310, 2361-2363.




135

(23) Song, H.; Rioux, R. M.; Hoefelmeyer, J. D.; Komor, R.; Niesz, K.;

Grass, M.; Yang, P.; Somorjai, G. A. Hydrothermal Growth of Mesoporous SBA-15

Silica in the Presence of PVP-Stabilized Pt Nanoparticles: Synthesis,

Characterization, and Catalytic Properties. Journal of the American Chemical Society

2006, 128, 3027-3037.

(24) Raghuveer, V.; Manthiram, A.; Bard, A. J. Pd-Co-Mo Electrocatalyst

for the Oxygen Reduction Reaction in Proton Exchange Membrane Fuel Cells.


(25) Raja, R.; Khimyak, T.; Thomas, J. M.; Hermans, S.; Johnson, B. F. G.

Single-step, highly active, and highly selective nanoparticle catalysts for the

hydrogenation of key organic compounds. Angewandte Chemie, International Edition

2001, 40, 4638-4642.

(26) Casado-Rivera, E.; Volpe, D. J.; Alden, L.; Lind, C.; Downie, C.;

Vazquez-Alvarez, T.; Angelo, A. C. D.; DiSalvo, F. J.; Abruna, H. D. Electrocatalytic

Activity of Ordered Intermetallic Phases for Fuel Cell Applications. Journal of the


(27) Xu, Y.; Ruban, A. V.; Mavrikakis, M. Adsorption and dissociation of

O2 on Pt-Co and Pt-Fe alloys. Journal of the American Chemical Society 2004, 126,

4717-4725.

(28) Folch, B.; Larionova, J.; Guari, Y.; Datas, L.; Guerin, C. A

coordination polymer precursor approach to the synthesis of NiFe bimetallic

nanoparticles within hybrid mesoporous silica. Journal of Materials Chemistry 2006,

16, 4435-4442.

(29) Vondrova, M.; Klimczuk, T.; Miller, V. L.; Kirby, B. W.; Yao, N.;

Cava, R. J.; Bocarsly, A. B. Supported Superparamagnetic Pd/Co Alloy Nanoparticles

Prepared from a Silica/Cyanogel Co-gel. Chemistry of Materials 2005, 17, 6216-

6218.

(30) Kockrick, E.; Krawiec, P.; Schnelle, W.; Geiger, D.; Schappacher, F.

M.; Poettgen, R.; Kaskel, S. Space-confined formation of FePt nanoparticles in

136

ordered mesoporous silica SBA-15. Advanced Materials (Weinheim, Germany) 2007,

19, 3021-3026.

(31) Hermans, S.; Raja, R.; Thomas, J. M.; Johnson, B. F. G.; Sankar, G.;

Gleeson, D. Solvent-free, low-temperature, selective hydrogenation of polyenes using

a bimetallic nanoparticle Ru - Sn catalyst. Angewandte Chemie, International Edition

2001, 40, 1211-1215.

(32) Thomas, J. M.; Johnson, B. F. G.; Raja, R.; Sankar, G.; Midgley, P. A.

High-Performance Nanocatalysts for Single-Step Hydrogenations. Accounts of

Chemical Research 2003, 36, 20-30.

(33) Valden, M.; Lai, X.; Goodman, D. W. Onset of catalytic activity of

gold clusters on titania with the appearance of nonmetallic properties. Science

(Washington, D. C.) 1998, 281, 1647-1650.

(34) Zhou, B.; Hermans, S.; Somorjai, G. A. Nanotechnology in Catalysis;

Kluwer Academic/ Plenum 2004; Vol. 1.

(35) King, N. C.; Blackley, R. A.; Wears, M. L.; Newman, D. M.; Zhou,

W.; Bruce, D. W. The synthesis of mesoporous silicates containing bimetallic

nanoparticles and magnetic properties of PtCo nanoparticles in silica. Chemical

Communications (Cambridge, United Kingdom) 2006, 3414-3416.

(36) Sun, X.; Kang, S.; Harrell, J. W.; Nikles, D. E.; Dai, Z. R.; Li, J.;

Wang, Z. L. Synthesis, chemical ordering, and magnetic properties of FePtCu

nanoparticle films. Journal of Applied Physics 2003, 93, 7337-7339.

(37) Liang, C.; Dai, S. Synthesis of Mesoporous Carbon Materials via

Enhanced Hydrogen-Bonding Interaction. Journal of the American Chemical Society

2006, 128, 5316-5317.

(38) Komiyama, M.; Shimaguchi, T. Partial reduction of Si(IV) in SiO2

thin film by deposited metal particles: an XPS study. Surface and Interface Analysis

2001, 32, 189-192.

137

(39) Klie, R. F.; Disko, M. M.; Browning, N. D. Atomic Scale

Observations of the Chemistry at the Metal-Oxide Interface in Heterogeneous

Catalysts. Journal of Catalysis 2002, 205, 1-6.

(40) Yu, R.; Song, H.; Zhang, X.-F.; Yang, P. Thermal Wetting of Platinum

Nanocrystals on Silica Surface. Journal of Physical Chemistry B 2005, 109, 6940-

6943.

(41) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. Nonionic

triblock and star diblock copolymer and oligomeric surfactant syntheses of highly

ordered, hydrothermally stable, mesoporous silica structures. Journal of the American


(42) Williams, D. B.; Carter, C. B. Transmission Electon Microscopy;

Springer: New York, 1996.

(43) Ferrer, D.; Torres-Castro, A.; Gao, X.; Sepulveda-Guzman, S.; Ortiz-

Mendez, U.; Jose-Yacaman, M. Three-layer core/shell structure in Au-Pd bimetallic

nanoparticles. Nano Letters 2007, 7, 1701-1705.

(44) Zhang, L.; He, R.; Gu, H.-C. Oleic acid coating on the monodisperse

magnetite nanoparticles. Applied Surface Science 2006, 253, 2611-2617.

(45) Papirer, E. Adsorption on Silica Surfaces CRC, 2000; Vol. 90.










138




(50) Haller, G. L. New catalytic concepts from new materials:

understanding catalysis from a fundamental perspective, past, present, and future.

Journal of Catalysis 2003, 216, 12-22.

(51) Pietron, J. J.; Stroud, R. M.; Rolison, D. R. Using Three Dimensions in

Catalytic Meso-porous Nanoarchitectures. Nano Letters 2002, 2, 545-549.



299, 1698-1702.

(53) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R.

Synthesis of Thiol-Derivatised Gold Nanoparticles in a Two-Phase Liquid-Liquid

System. J. Chem. Soc., Chem. Commun. 1994, 7, 801-802.



(55) Sra, A. K.; Ewers, T. D.; Schaak, R. E. Direct Solution Synthesis of

Intermetallic AuCu and AuCu3 Nanocrystals and Nanowire Networks. Chemistry of

Materials 2005, 17, 758-766.

(56) Santiago, E. I.; Varanda, L. C.; Villullas, H. M. Carbon-Supported Pt-

Co Catalysts Prepared by a Modified Polyol Process as Cathodes for PEM Fuel Cells.


(57) Alden, L. R.; Roychowdhury, C.; Matsumoto, F.; Han, D. K.;

Zeldovich, V. B.; Abruna, H. D.; DiSalvo, F. J. Synthesis, Characterization, and

Electrocatalytic Activity of PtPb Nanoparticles Prepared by Two Synthetic

Approaches. Langmuir 2006, 22, 10465-10471.

(58) Ponec, V.; Bond, G. C. Cataysis by Metals and Alloys; Elsevier

Netherlands, 1995.

139

(59) Venezia, A. M.; Liotta, L. F.; Deganello, G.; Schay, Z.; Guczi, L.

Characterization of Pumice-Supported Ag-Pd and Cu-Pd Bimetallic Catalysts by X-

Ray Photoelectron Spectroscopy and X-Ray Diffraction. Journal of Catalysis 1999,

182, 449-455.

(60) Konopny, L. W.; Juan, A.; Damiani, D. E. Preparation and

characterization of g-Al2O3-supported Pd-Mo catalysts. Applied Catalysis, B:

Environmental 1998, 15, 115-127.

(61) Batista, J.; Pintar, A.; Gomilsek, J. P.; Kodre, A.; Bornette, F. On the

structural characteristics of g-alumina-supported Pd-Cu bimetallic catalysts. Applied

Catalysis, A: General 2001, 217, 55-68.

(62) Wang, Z. L.; Petroski, J. M.; Green, T. C.; El-Sayed, M. A. Shape

Transformation and Surface Melting of Cubic and Tetrahedral Platinum Nanocrystals.

Journal of Physical Chemistry B 1998, 102, 6145.

(63) Jeyadevan, B.; Hobo, A.; Urakawa, K.; Chinnasamy, C. N.; Shinoda,

K.; Tohji, K. Towards direct synthesis of fct-FePt nanoparticles by chemical route.

Journal of Applied Physics 2003, 93, 7574-7576.

(64) Iwaki, T.; Kakihara, Y.; Toda, T.; Abdullah, M.; Okuyama, K.

Preparation of high coercivity magnetic FePt nanoparticles by liquid process. Journal

of Applied Physics 2003, 94, 6807-6811.

(65) Roychowdhury, C.; Matsumoto, F.; Mutolo, P. F.; Abruna, H. D.;

DiSalvo, F. J. Synthesis, Characterization, and Electrocatalytic Activity of PtBi

Nanoparticles Prepared by the Polyol Process. Chemistry of Materials 2005, 17,

5871-5876.

(66) Minami, R.; Kitamoto, Y.; Chikata, T.; Kato, S. Direct synthesis of

L10 type Fe-Pt nanoparticles using microwave-polyol method. Electrochimica Acta

2005, 51, 864-866.

(67) Leonard, B. M.; Bhuvanesh, N. S. P.; Schaak, R. E. Low-temperature

polyol synthesis of AuCuSn2 and AuNiSn2: Using solution chemistry to access

140

ternary intermetallic compounds as nanocrystals. Journal of the American Chemical

Society 2005, 127, 7326-7327.

(68) Raja, R.; Hermans, S.; Shephard, D. S.; Johnson, B. F. G.; Raja, R.;

Sankar, G.; Bromley, S.; Thomas, J. M. Preparation and characterisation of a highly

active bimetallic (Pd-Ru) nanoparticle heterogeneous catalyst. Chemical

Communications (Cambridge) 1999, 1571-1572.

(69) Scott, R. W. J.; Sivadinarayana, C.; Wilson, O. M.; Yan, Z.; Goodman,

D. W.; Crooks, R. M. Titania-Supported PdAu Bimetallic Catalysts Prepared from

Dendrimer-Encapsulated Nanoparticle Precursors. Journal of the American Chemical

Society 2005, 127, 1380-1381.

(70) Min, M.-K.; Cho, J.; Cho, K.; Kim, H. Particle size and alloying

effects of Pt-based alloy catalysts for fuel cell applications. Electrochimica Acta

2000, 45, 4211-4217.

141

Chapter 5

Highly Stable ORR Electrocatalysts from Pre-synthesized Pt

Nanocrystals Supported on Graphitic Mesoporous Carbons

Highly stable Pt electrocatalysts for oxygen reduction have been prepared by the

infusion of pre-synthesized nanocrystals with controlled size (<4 nm) and surface

morphology into a highly graphitic, ordered 3d cubic mesoporous carbon with 7 nm

pores. The weakly bound dodecylamine ligands expose a sufficient number of Pt

atoms on the nanoparticle surface that bind strongly to the carbon surface to facilitate

high Pt loading up to 20 wt % with high metal dispersion, as indicated by XRD and

TEM. The mass activity for the oxygen reduction reaction (ORR) at 0.9 V reaches

0.14 A/mgPt for Pt on the graphitic mesoporous carbon, 25% larger than for Pt on

commercial amorphous Vulcan XC-72. The loss in ORR catalyst activity for Pt

supported on graphitic mesoporous carbons under an accelerated durability test of

potential cycling between 0.5 to 1.2 V for 1000 cycles, is extremely small, <2%,

compared to a 30% loss for Pt on amorphous mesoporous carbons and a 70% loss that

is typical for commercial Pt on amorphous carbon catalysts. The high resistance of

the graphitized carbon against oxidation, as characterized by thermo gravimetric

analysis, coupled with the strong and electron transfer Pt-C interactions, inhibit

nanoparticle isolation, nanoparticle coalescence, Pt dissolution and Ostwald ripening,

resulting in high catalyst stability. The large interfacial contact area between the

metal nanocrystals and concave mesopores with comparable radii, along with the high

Pt dispersion on the large surface area (>320 m2/g) supports further lower the

tendency of Pt coalescence. The ability to disperse pre-synthesized metal nanocrystals

with controlled size and surface morphologies onto mesoporous graphitic carbon

supports offers the potential for new classes of highly stable and active catalysts.

The contents of this chapter are being prepared for a manuscript.

142

5.1 Introduction

A significant challenge hindering the development of catalysts for the oxygen

reduction reaction in proton exchange membrane fuel cells, especially in

transportation applications, is the catalyst stability towards potential cycling.1-5 The

catalyst must be stable at 1.2 V for 100 h corresponding to the accumulated life time

of the vehicle, with a maximum allowable performance loss of <30 mV at 1.5A/cm2

(<1 µV/cycle target).4,5 However, Pt catalysts supported on traditional amorphous

carbons, containing oxygenated surface groups, undergo significant losses at 1.2 V

due to carbon oxidation to CO/CO2. Pt is known to catalyze the oxidation of the

support. As the carbon oxidatively degrades, the supported nanoparticles become

isolated, resulting in a loss in activity.1,5 Aside from carbon degradation, potential

cycling between 0.5 and 1.2 V causes Pt sintering as well as dissolution.6 For

example, Pt on Vulcan XC-72 carbon showed a 70% loss in the electrochemical area

upon potential cycling between 0.5 and 1.2 V for 1000 cycles.7 The presence of

oxygen species withdraws electrons from the Pt and weakens the binding of the Pt to

the carbon, thus facilitating Pt sintering and dissolution.8

The stability of Pt catalysts may be enhanced with corrosion resistant

graphitized carbons.5 These carbons are considerably more crystalline than typical

more amorphous carbons such as Vulcan XC-72 and contain very few oxidizable

polar groups on the surface. Consequently graphitized carbons are far more resistant

towards oxidation catalyzed by platinum.6,9,10 The abundant π sites on the surface of

graphitized carbon interact strongly with Pt as demonstrated by XPS and electron-

spin resonance (ESR) studies.3,6,9 The Pt-C interactions are partially covalent, due to

electron delocalization between the p-orbital of the π sites and Pt d-orbital, and are

partially ionic due to electron transfer from Pt to the carbon.11 Strong metal-support

interactions may inhibit Pt oxidation and dissolution during potential cycling and

lower the degree of Ostwald ripening and coalescence.6 Platinum supported on carbon

nanotubes (CNT) with graphitic surfaces is more stable than Pt on amorphous

143

carbon.12-14 For example, 30% loss in ECSA occurred for CNT compared to 70% loss

for VulcanXC-72 during potential cycling between 0.5 – 1.2 V for 1000 cycles. This

advancement is an important step in meeting the stability criteria for viable fuel cell

catalysts.5

The traditional technique of wetness impregnation may be utilized to form Pt

nanoparticles by reduction of Pt precursors on amorphous carbon, however, this

approach is limited for graphitic carbon supports.15 For amorphous carbon, polar

oxygenated surface species are required for adsorption of ionic platinum precursors,

such as hexachloroplatinic acid, to produce dispersed metal nanoparticles on the order

of 2-3 nm.3 However, the limited number of oxygen sites on graphitized carbon is

insufficient to achieve sufficient precursor adsorption. Consequently, reduction of

these weakly binding precursors leads to large nanoparticle sizes (>5 nm).6 To

prevent sintering, graphitic supports typically require surface modification, such as

refluxing in nitric acid, to introduce polar oxygen groups on the carbon surface to

facilitate binding of the precursors.16-19 However, after such treatments, the modified

graphitic supports are more susceptible to oxidation and the surface oxygen groups

weaken the final Pt nanocrystal-support interactions, raising the tendency towards Pt

dissolution and coalescence.14

An emerging concept in catalyst design is to pre-synthesize colloidal metal

nanocrystals coated with stabilizers, and then to infuse them onto high surface area

supports.20-23 The decoupling of nanocrystal synthesis and infusion provides exquisite

control over the nanocrystal size, morphology, and dispersibility within the porous

support. This technique has been reported for the preparation of fuel cell catalysts on

amorphous carbon, where the nanoparticles were stabilized with ligands24-27 or

microemulsions.28,29 High metal loadings have been achieved for nanoparticles

stabilized by weakly bound hydrocarbon capping ligands.21 The high loadings

suggested that the low metal surface coverage of these weakly bound ligands did not

screen the strong short-ranged interactions with the support. Furthermore, the ligands

may be removed at temperatures of ≤500 °C to activate the metal surface prior to

144

catalysis.4,21,23 For example, highly active FePt catalysts with high loadings have been

synthesized by infusing pre-synthesized nanocrystals coated with weakly bound oleic

acid and oleylamine ligands in mesoporous silica.21

The primary objective of this study was to design highly stable and active

ORR Pt catalysts with controlled size and morphology supported on mesoporous

graphitic carbon formed from mesophase pitch. The catalyst stability was tested under

relatively severe conditions, 1000 cycles between 0.5 and 1.2 V at 50 mV/s. Pt

nanocrystals were pre-synthesized by arrested growth precipitation with

dodecylamine ligands to control the size and surface morphology, and then infused

into graphitic mesoporous carbons. The large pore size of the mesoporous carbons (7-

10 nm) was achieved by eliminating the microporosity in the walls of mesoporous

silica template, to prevent formation of interconnecting carbon rods.30 Graphitic

supports are highly resistant to carbon oxidation3 and produce strong interactions with

Pt.6 The strong metal-support interactions will be shown to enhance catalyst stability

by inhibiting Pt dissolution, Ostwald ripening and coalescence. The concave pores in

the carbon were designed to achieve large interfacial area with the 3 nm spherical Pt

nanocrystals to strengthen metal-support interactions. Furthermore, the high

dispersion of the Pt nanoparticles on the high surface area (350 m2/g) graphitic carbon

with the 7 nm mesopores lowers the tendency for Pt coalescence. Weakly binding

ligands (dodecylamine) were chosen to stabilize the Pt nanoparticles to expose a

greater fraction of the metal surface to the graphitic π sites, to promote adsorption of

the nanocrystals at loadings up to 20 wt %. The low molecular weight ligands were

removed by mild thermal calcination to activate the catalyst while attempting to avoid

changing the nanocrystal size and morphology. The catalyst stabilities under potential

cycling are compared with those of previous studies on amorphous mesoporous

carbons.24,31-35 To the best of our knowledge, use of highly graphitic mesoporous

carbon as a support for synthesis of highly stable Pt catalysts for ORR has not been

reported.

145

The porous structure of the mesoporous carbons are characterized by nitrogen

porosimetry and small angle X-ray scattering (SAXS). The degree of graphitization of

the mesoporous carbons is characterized by X-ray diffraction (XRD) and the

oxidative stability by thermal gravimetric analysis. The size of the supported Pt

nanoparticles is characterized by XRD, TEM and electrochemical surface area

(ECSA). Pt-C interactions are quantified by XPS. In addition, the size of

nanoparticles and surface characteristics of the nanoparticles are characterized by

TEM, XRD and XPS to provide insight into the catalyst activity and stability.

5.2 Experimental Section

All chemicals were used as received. Platinum (IV) chloride (99.9%), ,

dodecylamine (98%), perchloric acid 70% (99.999% purity), sodium borohydride

(98%), tetradecylammonium bromide (98%), anhydrous ethanol, Pluronic P123

(EO20PO70EO20), n-butanol (99.9%), 5 wt % nafion solution in lower alcohols were

purchased from Aldrich. Toluene (99.9%) was obtained from Fisher Scientific,

ethanol (Absolute 200 proof) from Aaper alcohol. Polycarbonate filters (0.05 micron)

were obtained from Osmonics. High purity water (resistance ~18 MΩ) was used.

Platinum nanoparticles stabilized by dodecylamine were synthesized using the

phase-transfer method developed by Wikander at al.36 15mL of 60mM aqueous

solution of platinum chloride was combined with 105mL of toluene containing 1.5 of

TDAB (phase transfer catalyst). The mixture was stirred for 2 hours leading to a

platinum transfer from the aqueous phase to an organic phase. 3.4 g of dodecylamine

in 15mL of toluene was added to the organic phase while it stirring. Platinum was

then reduced by the drop wise addition NaBH4 (0.57 g NaBH4 dissolved in 25mL

water). Upon reduction, the solution turned dark brown and was stirred for 2h, before

initiating the cleaning phase. For cleaning, 500mL of ethanol was added to the

toluene phase (reduced to approximately 1/10th the original volume) and then kept

overnight in a freezer (-10 °C). The platinum nanocrystals precipitated at the bottom

146

of the breaker and were collected by decanting and dispersed in toluene. The platinum

nanocrystals were further washed by adding excess ethanol (toluene:ethanol=1:35)

and then dispersed in toluene.

The ordered mesoporous carbons were synthesized by the recipe described in

detail by using mesoporous silica of type KIT-637 as a template.30,38 Briefly, the

mesoporous silica was synthesized by mixing Pluronic P123 and n-butanol in

aqueous HCl solution. After complete dissolution, tetraethyl orthosilicate (TEOS,

98% Acros) was added into the mixture at 308 K. The molar composition of the

starting reaction mixture was 0.017 P123: 1.2 TEOS: 1.31 BuOH: 1.83 HCl: 195

H2O. The mixture was magnetically stirred for 1 day at the same temperature.

Subsequently, the mixture was heated for 1 day in an oven for hydrothermal

treatment. After the hydrothermal treatment, the solid product was recovered by

filtration, washed with ethanol and then calcined at 823K to obtain mesoporous silica.

The mesoporous silica was then mixed with carbon precursor (furfural alcohol or

mesophase pitch) and p-toluene sulfonic acid at room temperature (1g silica: 0.458

mL carbon precursor: 1.97 g p-toluene sulfonic acid). The resultant sample was

heated at 308 K for 1 h, at 373 K for 1h and then for 2 h at 623 K. Carbonization was

performed by heating to 1173 K at 3.7 K/min in argon. The silica template was then

removed using HF to recover mesoporous carbon. Mesoporous carbon synthesized

with furfural alcohol is more amorphous39 and is denoted as AMC and that with

mesophase pitch is more graphitic and is denoted as GMC.

The infusion of the Pt nanoparticles in mesoporous carbon was done by a

similar procedure as previously reported in Gupta et al.21 30 mg of mesoporous

carbon was mixed with ~3 mL of Pt nanocrystal dispersion in toluene of known

concentrations ~4-10 mg/mL. The contents were then stirred for 2 days. After

stirring, the mesoporous carbon infused with Pt nanocrystals was separated from the

nanocrystal supernatant by filtration with a 0.2 μm PTFE filter. The extent of

incorporation of Pt nanocrystals in the carbon was determined by subtracting from the

147

initial mass of the dispersed nanocrystals in toluene, the final mass of the nanocrystals

recovered in the supernatant after filtration.22 The absorbance of the Pt nanocrystals

dispersed in toluene before and after infusion was measured using a Cary 500 UV-

Vis-NIR spectrophotometer with an optical path length of 1 cm. Figure 5.1a shows a

representative absorbance curve for the Pt nanoparticles.

A standard calibration curve with a correlation coefficient of 0.998 was

generated with known concentrations of Pt nanocrystals in toluene for a wavelength

of 800 nm (Figure 5.1b). Typically loadings of 20% metal on mesoporous carbons for

both AMC and GMC were obtained after infusion. Total uncertainty in loadings

determined with this technique was approximately 12% as explained in detail

previously.23 To activate the catalysts, ligands were removed by calcining the catalyst

at 300-500 °C for 2h. The temperature was ramped at 16 °C/min. After calcination,

the catalysts were washed with 100 mL of ethanol.

N2 sorption analysis was performed on a Micromeritics ASAP 2010

volumetric analyzer. The specific surface area was calculated using the BET method

from the nitrogen adsorption data in the relative range (P/P0) of 0.05-0.30. The

samples were outgases at 200 °C in the degas port of the adsorption analyzer. The

total pore volume was determined from the amount of N2 uptake at P/P0 = 0.95. The

pore size distribution was derived from the adsorption branch of the isotherm based

on the BJH model.


using a Bruker Nokius instrument using Cu Kα radiation with wavelength = 1.54 Å.

Samples were scanned for 12 h at a scan rate of 12 °/min with 0.02 deg increments.

The average nanocrystal size was estimated from the Scherrer equation using JADE

software (by Molecular Diffraction Inc). Small angle X-ray diffraction was collected

using Pohang Light Source in the reflection mode.

Thermogravimetric analyses (TGA) were made with a Perkin Elmer TGA 7

instrument with ~2 mg catalyst samples. All samples were initially heated to 50 °C,

148

(a)

Figure 5.1

(a) UV-Vis spectra of Pt nanoparticles dispersed in toluene before and after infusion

in AMC and (b) calibration curve for Pt nanoparticles using known concentrations

and measuring absorbance at 800 nm. Correlation coefficient was 0.9998.

00.5

11.5

22.5

33.5

4

400 500 600 700 800Wavelength (nm)

Abs

orba

nce

Before infusion

After infusion

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.2 0.4 0.6

Conc of Pt‐DDA (mg/ml)

Absorbance at 800

nm

149

held at that temperature for 10 min, and heated to 900 °C at 20 °C/min. Isothermal

weight loss studies were conducted by holding the samples of 100 mg weight at an

elevated temperature of 195 °C for 16 days, similarly to as described by Stevens and

Dahn.40

Low resolution TEM was conducted on a Phillips EM280 microscope with a

4.5Å point-to-point resolution operated with a 80 kV operating voltage. Platinum

catalyst was deposited from a dilute ethanolic solution onto a 200 mesh carbon-coated

copper TEM grid (Electron Microscopy Sciences).

A glassy carbon rotating disk electrode (0.196 cm2, from Pine Instruments)

was used as a substrate for the catalyst and was polished with alumina suspension of

0.3µm followed by 0.05 µm to give a mirror finish. 1.0 mg of Pt catalyst was

suspended by sonicating in 300µL of anhydrous ethanol and 300 µL of 0.15 wt %

nafion solution (diluted from 5% nafion stock solution). A total of 10 µL of Pt

catalyst suspension was pipetted onto the substrate and was dried at room temperature

for 15 min.

Electrochemical measurements were carried out using a three-compartment

electrochemical cell with a Pt wire and Hg/Hg2SO4 serving as the counter and the

reference electrodes, respectively. Electrochemical measurements were recorded

using CHI 832 electrochemical analyzer. All potentials cited in this article are

reported with respect to NHE.

After preparation of the catalyst on the substrate, the catalyst was cleaned by

cycling potential between 0.05 and 1.24 V at 200 mV/s for 100 cycles in argon

saturated 0.1M HClO4 electrolyte. The area under the desorption peak of

underpotentially deposited hydrogen in a CV (scanned at 100 mV/sec between 0.04

and 1.24V), was used to evaluate electrochemically active surface area (ECSA) of the

Pt catalyst on the electrode. Hydrodynamic voltammograms at rotating disk electrode

were recorded by anodically scanning its potential from 0.04 to 1.24 V at 5 mV/s in

oxygen saturated 0.1 M HClO4. Accelerated durability tests (ADT) were employed by

150

cycling the electrode potential between 0.5 and 1.2 V at 50 mV/s for 1000 cycles in

argon saturated atmosphere over a period of 7 h.

5.3 Results

The structure of the ordered mesoporous carbons was studied using small

angle X-ray scattering. For both AMC and GMC, multiple reflections associated with

the I4132 space-group symmetry (3d cubic morphology) are observed (Figure 5.2)

and indicate a long range ordering of pores. The first three reflections appear at 2θ =

0.68°, 1.2° and 1.4° and 2θ = 0.68°, 1.22° and 1.43° for GMC and AMC,

respectively. The d-spacings corresponding to the reflections are calculated using

Bragg’s Law ( )sin(.2 θλ dn = ).The d-spacings are estimated to be 130, 74 and 64 Å

for GMC and 130, 72.3 and 62 Å for AMC for the first three reflections. The ratio of

the d-spacings d1:d2:d3 is thus 3:2:32 for both AMC and GMC, and can be

indexed as 110, 211 and 220 planes of I4132 tetragonal geometry which is consistent

with the literature.30

Figures 5.3(a) and 5.3(b) gives the adsorption-desorption isotherms and

corresponding pore size distributions (in inset) for GMC and AMC, respectively. For

both carbons, three well-distinguished regions of the adsorption isotherm are evident:

(i) monolayer- multilayer adsorption, (ii) capillary condensation, and (iii) multilayer

adsorption on the outer particle surfaces. The adsorption isotherms show a type H1

behavior with capillary condensation at P/Po ~ 0.6 for AMC and of type H2 for GMC

with onset of capillary condensation at P/Po ~ 0.5. The slope and almost vertical step

in the hysteresis loops suggest well-defined mesopores with a narrow pore size

distribution as reflected by the pore size distributions.

The surface area, pore volume and mean pore size for GMC and AMC are

shown in Table 5.1. The surface area for AMC and GMC are 978 and 337 m2/g and

are significantly higher than VulcanXC-72 (~250 m2/g). Lower surface areas and pore

volume for GMC relative to AMC can be attributed to the fact that mesophase pitch

151

Figure 5.2

SAXS of mesoporous carbons.

GMC

AMC

110 211

220

152

(a)

(b)

Figure 5.3

Nitrogen porosimetry of the mesoporous carbons and their pore size distributions in

the inset for GMC (A) and AMC (B).

153

Carbon Surface

Area [m2/g]

Total pore

volume [cc/g]

Pore size

(BJH) [nm]

La (XRD)

[Å]

Lc (XRD)

[Å]

d (XRD)

[Å]

AMC 978.5 1.7 9.5 37 7 3.57 GMC 337.4 0.43 7.1 55 16 3.42

Table 5.1

Properties of mesoporous carbons.

154

contains a considerable amount of large polyaromatic compounds, which after

carbonization, yield mesoporous carbon with little microporosity, as reported in the

literature.41 AMC, on the other hand, is synthesized from the furfural alcohol and

contains significant number of micropores, as indicated by the large uptake of

nitrogen at P/Po → 0, and thus has higher surface area. GMC has a predominant pore

size of 7.1 nm but also shows a slight hump at ~5 nm too while AMC has a pore size

of 9.5 nm. The large pores are facilitated by the lack of microporosity in the

mesoporous silica template which prevents formation of the interconnecting carbon

rods. Upon removal of the silica template, the enantiomeric carbon frameworks join

pair-wise leading to pore enlargement.30

XRD spectra of the carbons at each of the processing temperatures are shown

in Figure 5.4. The peak at ~25° corresponds to the graphitic <002> plane on carbon.

The narrower taller <002> peak is shown for the GMC indicating a larger degree of

graphitization, relative to the broader diffuse peak for the AMC. The layer dimension

parallel to the basal place, La, and perpendicular, Lc, were calculated from the <10>

and <002> reflections respectively (Table 5.1), according to the relations (La =

1.84λ/Bcos(θ) and Lc = 0.89λ/Bcos(θ),42,43 where λ is the wavelength of the X-rays

(1.54Å), B is is the full angular width at half max, and θ is the Bragg angle. The

increased layer dimensions (La and Lc) for GMC relative to AMC indicate a greater

degree of graphitization; there is longer range ordering in GMC due to inherent

graphitic character of the mesophase pitch used as precursor for synthesis of GMC .

The interplanar d-spacings calculated from the <002> reflection using Bragg’s law

(nλ = 2dsin(θ)) are shown in Table 5.1. There is a decrease in d-spacing for GMC

relative to AMC, indicating a tighter packing of the carbon planes, again consistent

with an increase in degree of graphitization.43

The loading of the Pt nanocrystals in mesoporous carbons is given in Table

5.2. For AMC, loadings of 23 wt % are attained after Pt dispersion at concentration of

4 mg/mL is infused in the support for 2 days. The final concentration after infusion is

155

Figure 5.4

XRD of the mesoporous carbon samples without added Pt.

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 20 40 60 80 100

AMC

GMC

2θ

Inte

nsity

002

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 20 40 60 80 100

AMC

GMC

2θ

Inte

nsity

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 20 40 60 80 100

AMC

GMC

2θ

Inte

nsity

002

156

Support Time of infusion (days)

Initial Conc

(mg/ml)

Final Conc.

(mg/ml)

Loading† (wt %)

Loading‡ (wt %)

AMC 2 4 <0.1 23±1 25 GMC 2 4 2.5 10±1 - GMC 3 10 6.8 20±3 22

* All infusions were done on 30 mg carbon and nanoparticles were dispersed in 3 mL of solvents † Loading calculated by UV-Vis ‡ Loading by TGA

Table 5.2

Loadings for infusion of Pt nanoparticles in carbon supports

157

<0.1 mg/ml and is within the experimental error, indicating almost all the

nanocrystals are infused in the mesopores. For GMC, 10 wt % loadings are achieved

under identical conditions of infusion and are lower than AMC. The smaller pore size

and surface area of the GMC relative to AMC may lead to lower loadings. However,

20 wt % loadings are achieved by increasing the starting concentration of Pt

nanocrystals to 10 mg/mL. The final concentration of nanoparticles after infusion was

6.8 mg/mL, indicating that 32% of the initial Pt nanoparticles are infused in the pores.

As-synthesized unsupported Pt nanoparticles showed an average size of

2.3±0.3 nm. Figure 5.5 shows TEM images of the Pt nanoparticles supported on

mesoporous carbons. Pt nanoparticles are highly dispersed on GMC and do not show

any sintering or aggregation upon thermal treatment for ligand removal, as indicated

by the similar histograms (Figure 5.5d, e) for the nanoparticle sizes before and after

thermal calcination. For GMC, the average Pt nanoparticle size for as-synthesized

nanocomposites is 3.05±0.42 nm which upon thermal calcination at 300 °C changes

to 3.37±0.65 nm. The slight increase in the particle size for uncalcined supported Pt

can be attributed to the low image contrast between the Pt and the carbon, which

makes the particles appear larger.23 The marginal increase in the nanoparticle size

during thermal calcination indicates strong Pt interactions with graphitic GMC. For

AMC, the ligands are removed at 500 °C and the average nanoparticle size is

3.55±0.65 nm after annealing, indicating little sintering. Thus the average Pt size on

AMC and GMC after calcination was comparable.

Figure 5.6(a) shows the TGA curves for blank AMC and GMC as well these

carbons with 20% Pt infused into the pores. Each curves exhibits four distinctive

regions; the stable region with no weight loss at low temperatures, a transition region

with gradual weight loss, a steep weight loss region, and a plateau after all the carbon

is oxidized. Figure 5.6(b) shows the derivative of Figure 5.6(a), that is differential

thermal gravimetry (DTG). The onset of carbon oxidation for Pt-AMC occurs at

relatively low temperature of ~320 °C (as shown in inset of 6b), and is slightly better

than the value of ~250 °C for 20% Pt on Vulcan XC-72.44 GMC on the other hand

158

Figure 5.5

TEM micrographs of 20% Pt on (A) GMC uncalcined and (B) GMC calcined at 300 °C, (C) AMC calcined at 500 °C. (D), (E), (F) Histograms for (A), (B) and (C), respectively. More than 50 nanoparticles were counted in each case.

0

5

10

15

20

25

2 2.4 2.8 3.2 3.6 4 4.4 4.8

Cou

nt

Pt nanoparticle size (nm)

0

5

10

15

20

25

1.5 2.2 2.9 3.6 4.3 5

Cou

nt

Pt nanoparticle size (nm)

0

5

10

15

20

25

1.5 2.2 2.9 3.6 4.3 5 5.7 6.4

Cou

nt

Pt Np Size (nm)

A B

C D

E F

159

Figure 5.6

(A) TGA for the pure (blank) carbons and 20% Pt infused mesoporous carbons AMC

and GMC calcined at 500 and 300 °C in N2/H2 atmosphere respectively and (B)

Differential thermal gravimetry (DTG) for the Pt on AMC and GMC. Inset shows the

magnified section of (B) where onset of oxidation occurs for AMC.

0

20

40

60

80

100

200 300 400 500 600 700 800Temp (C)

Wt %

Pt-AMCPt-GMCBlank AMC

Blank GMC

A

B

160

does not show any carbon degradation until a much higher temperature of 430 °C.

The AMC losses most of the carbon steeply at 400 °C (Figure 5.6(b)), which is lower

than the onset oxidation temperature for GMC. GMC looses most of the weight in

two phases at temperatures of 550 and 700 °C. The final stable weight for the AMC

and GMC is 23 and 21 wt % respectively and matches well with that determined from

the UV-Vis spectroscopy (Table 5.2). The onset temperatures for the blank AMC and

GMC are 420 °C and 510 °C respectively, significantly higher than corresponding

carbons with Pt. Similarly, blank GMC and AMC has a steep weight loss at 590 and

630 °C respectively, again higher than with carbons with Pt.

The relative stabilities of the mesoporous carbons and Vulcan XC-72 towards

accelerated thermal oxidation in presence of Pt at elevated temperature of 195 °C is

reported over an extended period of 16 days in Table 5.3. Vulcan XC-72 showed a

significant loss of 8.84 wt % due to carbon oxidation to CO/CO2, consistent with the

previously reported results.40 In contrast, the weight losses for Pt-AMC and GMC

were 2.95 and 0.71 wt %, respectively, much lower than Vulcan carbon. The values

were negligible for the three carbons without Pt, indicating Pt catalyzes the oxidation

as shown previously.40

Figure 5.7 gives the XRD spectra of Pt on mesoporous carbons before and

after ADT cycling. Pt crystalline peaks observed at 2θ= 39.8°, 46.5° and 67.8° and

81.6° may be assigned to the (111), (200), (220) and (311) orientations for an FCC

crystal phase. The nanocrystal size was determined by the Scherrer equation using the

(111) peak. For GMC, the size before ADT was determined to be 3.2 nm in good

agreement with the average size of 3.4 nm determined from TEM. Upon ADT

cycling, the Pt diameter increased slightly to 3.6 nm, an increase of ~10%. All of the

peaks for the orientations in the FCC crystal phase remained. For AMC, upon ADT

cycling the peaks become sharper and narrower, indicating significant increase in Pt

nanoparticle size. An additional peak at 86.2° corresponding to (222) plane appears.

The nanoparticle size for AMC increased from 3.6 nm to 5.6 nm upon ADT, an

increase of over 55%, significantly larger than in the case of GMC.

161

400

600

800

1000

1200

1400

1600

1800

30 40 50 60 70 80 90

before ADTafter ADT

Inte

nsity

2 theta

111

200220

311

(a)

(b)

Figure 5.7

XRD spectra of 20% Pt on (a) GMC and (b) AMC, before and after ADT

(for conditions in Table 5.4). Calcinations were done in 7% H2-93% N2 environment.

0

2000

4000

6000

8000

1 104

30 40 50 60 70 80 90

Before ADTPost ADT

Inte

nsity

2 Theta

111

200

220 311

111

111200

220 311

222

162

Type of Carbon % Loss of Carbon

20% Pt - Vulcan XC-72 8.84

20% Pt – A MC 2.95

20% Pt - GMC 0.71

Blank Vulcan XC-72 <0.5

Blank AMC <0.5

Blank GMC <0.5

Table 5.3

Weight loss of the carbon supports at 195 °C air for 16 days due to carbon oxidation

catalyzed by platinum.

163

Support

Calcination

Temp (°C)

Dxrd†

(nm) CSAxrd

†

(m2/g)

I † (A/mg

Pt)

I † (µA/cm2

)

ECSA†

(m2/g)Dxrd

‡

(nm) CSAxrd

‡

(m2/g)

I ‡ (A/mg

Pt)

I ‡ (µA/cm

2)

ECSA‡ (m2/g)

GMC 300 3.2 87 0.14 208 70 3.6 79 0.14 203 69

GMC 350 - - 0.13 170 74 - - 0.135 171 76

GMC 400 - - 0.1 155 62 - - 0.12 149 71

GMC Uncalcined - - 0.042 82 52 - - 0.046 84 54

AMC 500 3.3 84 0.14 218 65 5.6 51 0.095 202 46

VC-XC 721 - 3 93 0.11 200 69 5.382 52 0.035 ~180 21

Table 5.4

Electrochemical Data for 20% Pt on mesoporous carbons before and after ADT

potential cycling between 0.5 and 1.2 V at 50 mV/s for 1000 cycles in 0.1 M HClO4

saturated with argon.

164

XPS in Figure 5.8a for AMC and 8b for GMC show the C 1s spectra before and after

accelerated durability testing, whereas the Pt 4f spectra are shown in Figures 5.8c for

AMC and 5.8d for GMC. The C 1s spectra for GMC does not show any change after

potential cycling. However, for AMC, two new pronounced peaks appear at higher

binding energies relative to the original peak at 284.4 eV. The peak appearing at

288.5 eV can be attributed to the formation of COOH groups on the surface due to

carbon oxidation.45 The O/C ratio increases by 100% upon potential cycling for AMC

compared to only 25% for GMC. The binding energy for the Pt 4f7/2 peak for GMC

(72.1) is larger than for both unsupported Pt (71.0 eV) and Pt on AMC (71.2).46,47

For both AMC and GMC the Pt 4f7/2 peak does not change significantly upon ADT

cycling.

The electrochemical stabilities of the Pt-GMC and AMC catalysts were

studied for ADT cycling between 0.5 and 1.2V. Figures 5.9a (GMC), and 5.9b

(AMC) give CVs of the Pt catalyst before and after ADT potential cycling. The

ECSA of the platinum nanocrystals was estimated from the area of the anodic peak

corresponding to H-desorption, QH, appearing between 0.04 and 0.4 V

by13

cvmQECSAPt

H

..=

where mpt is the mass of platinum on the electrode, ν is scan rate (V/s) and c is a

constant defined by the charge required to oxidize a monolayer of hydrogen on Pt

(0.21 mC/cm2). The ECSA for 20% Pt on GMC calcined at 300 °C was 70 m2/g and

was comparable to that of commercial Pt catalysts (65-68 m2/g).1 A modestly smaller

value of 52 m2/g was observed for the uncalcined Pt-GMC (Table 5.4). Thus, this

modest decrease in the ECSA indicates that the dodecylamine ligands present on the

Pt nanocrystal surface block only a small fraction of the active Pt sites. For

calcination at 400 °C, the ECSA of 62 m2/g is slightly lower, perhaps indicating a

small degree coalescence of nanoparticles. The ECSA of 20% Pt nanoparticles

supported on AMC calcined at 500 °C was 65 m2/g, comparable to that of 20% Pt on

GMC calcined at 300 °C.

165

Figure 5.8 XPS spectra before and after ADT for 20% Pt supported on AMC and GMC. (A),(B): C1s region for AMC and GMC, respectively. (C),(D): Pt4f region for AMC and GMC, respectively.

85 80 75 70 65

Nor

mal

ized

Inte

nsity

Binding Energy (eV)

GMC GMC post ADT

85 80 75 70 65

Nor

mal

ized

Inte

nsity

Binding Energy (eV)

AMC AMC post ADT

294 292 290 288 286 284 282 280 278 276

Nor

mal

ized

Inte

nsity

Binding Energy (eV)

AMC AMC post ADT

294 292 290 288 286 284 282 280 278 276

Nor

mal

ized

Inte

nsity

Binding Energy (eV)

GMC GMC post ADT

A B

C D

166

Figure 5.9

Electrochemical studies for 20% Pt on mesoporous carbon before and after ADT

cycling for conditions in Table 5.4 (A) CV of GMC at 20 mV/s (B) CV of AMC at

20 mV/s (C) Linear Scan Voltammetry (LSV) of GMC at 5 mV/s rotated at 1600

rpm (D) LSV for AMC at same conditions

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1 1.2 1.4

E (vs NHE)

I (m

A/c

m2

geo)

Before ADT cycling

After ADT cycling

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1 1.2

E (vs NHE)

I (m

A/c

m2)

-6

-5

-4

-3

-2

-1

0

1

0.2 0.4 0.6 0.8 1 1.2

E (vs NHE)

I (m

A/c

m2

elec

trod

e)

Before ADT cycling

After ADT cycling

-7

-6

-5

-4

-3

-2

-1

0

1

0.2 0.4 0.6 0.8 1 1.2

E (vs NHE)

I (m

A/ c

m2

elec

trod

e)

Before ADT cycling

After ADT cycling

C D

167

After ADT cycling, the CV for Pt-GMC overlaps to the curve for the as-synthesized

catalyst, indicating negligible loss in stability with a constant ECSA, consistent with

the XRD results. The 20% Pt on AMC sample however loses substantial ECSA, with

a 30% drop as shown in Figure 5.9(b). Furthermore, 20% Pt on Vulcan-XC 72 shows

a loss of 70% ECSA under identical ADT cycling conditions.7 Thus, the trend in

stability for the catalysts can be established as Pt-GMC >> Pt-AMC > Pt-VulcanXC-

72.

Figure 5.9c (GMC and 9d (AMC) show the representative hydrodynamic

voltammogram obtained for the ORR before and after ADT cycling. The diffusion-

limited current densities at highly reducing potentials were 5.6 and 6 mA/cm2 for Pt

supported on GMC and AMC, respectively, both before and after ADT. These current

densities are comparable to the 6 mA/cm2 for a smooth polycrystalline Pt-disk

electrode, demonstrating negligible mass-transport effects. For Pt-GMC, the

hydrodynamic voltammograms after ADT overlaps with the before ADT, indicating

no change in catalyst activity, consistent with the results of ECSA as shown in Figure

5.9a, b. However, for Pt-AMC the hydrodynamic voltammograms shifts to lower

potentials in the kinetic region between 0.6-1.0 V after ADT relative to as-

synthesized, indicative of activity losses.

Figure 5.10 shows the Tafel slopes of mass activity and specific area activities

for the Pt supported on GMC and AMC, in the kinetically controlled regime, before

and after 0.5 – 1.2 V potential cycling. The activities are reported in Table 5.4 for a

potential of 0.9V. The mass activities for both the Pt-GMC calcined at 300 °C and

calcined Pt-AMC before cycling are 0.14 A/mg Pt, about 30% higher than that of

commercial 20% Pt-Vulcan carbon catalyst.1 Thus the degree of graphitization of the

support did not appear to influence the initial mass activity. Similarly, the specific

area activities for both Pt-GMC and AMC catalysts before cycling were ~200 µA/cm2

comparable to the value for commercial Pt-Vulcan XC-72 catalyst. The mass and area

activities for Pt-GMC before ADT testing decreased modestly with a calcination

temperature of 400 °C, consistent with a decrease in ECSA. For each of the calcined

168

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.01 0.1 1

Mass activity (A/ mg Pt)

E (v

s N

HE)

before stability

After stability

0.84

0.86

0.88

0.9

0.92

0.94

0.96

10 100 1000Specific activity (µ

A/ cm2 Pt)

E (v

s N

HE)

before stability

After stability

(A) (B)

0.84

0.86

0.88

0.9

0.92

0.94

0.96

10 100 1000

Specific Current Activity (µA/cm2 Pt)

E (v

s N

HE)

Before cycling

After cyling

(C) (D)

Figure 5.10

Tafel slopes for Pt-GMC calcined at 300 °C (A) mass activity (B) specific activity

and Pt-AMC calcined at 500 °C (C) mass activity and (D) specific activity

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.01 0.1 1

Mass Current Activity (A/ mg Pt)

E (v

s N

HE)

Before ADT cycling

After ADT cycling

169

Pt-GMC catalysts, the catalyst activity and ECSA did not decrease after ADT cycling,

indicating extremely high stability relative to the AMC and Vulcan catalysts. For Pt-

AMC, a significant ~30% loss in mass activity occurred during ADT, which

correlates well with the 30% loss in ECSA (Figure 5.10). An even more substantial

loss in these properties was observed for the Vulcan carbon catalyst.

5.4 Discussion

5.4.1 Carbon Oxidation and its Effect on Electrocatalyst Stability

The XRD spectra in Figure 5.4 indicates that AMC is significantly more

amorphous (broad <002> peak) relative to GMC (narrow <002> peak). In lieu of this,

AMC is expected to have a significant number of surface oxygen groups (carbonyl,

carboxylate, ester-like and alcohol).6,25 The presence of Pt atoms on amorphous

carbon accelerates carbon oxidation, which is facilitated by the surface oxygen groups

acting as reactive centers. This is confirmed by carbon weight loss under isothermal

treatment at 195 °C for Pt-AMC and Pt-Vulcan XC-72 (Table 5.3).6,48 The weight

loss for Vulcan carbon is significantly higher than that for AMC, suggesting that

Vulcan carbon contains a significantly higher proportion of oxygen groups.49

Similarly, Pt catalyzed oxidation of AMC and Vulcan XC-72 initiated on surface

oxygen sites may occur when subjected to highly oxidative potentials (up to 1.2V) in

corrosive, acidic environments. Carbon oxidation leads to Pt nanoparticle isolation

and is possibly one of the several factors leading to the loss in ECSA during potential

cycling, as seen for AMC and Vulcan carbon. The oxidation of the AMC after

potential cycling between 0.5 and 1.2 V is reflected by the manifestation of peaks

within the 287-293 eV binding energy range for the C 1s XPS spectra (Figure 8).

These peaks are commonly referred to as the carbon “oxide shoulders”.19,50 The

increase in the atomic ratio of O/C for Pt/AMC is 1.6 times that of Pt/GMC,

indicating that the degree of oxidation of AMC is higher than that of GMC, which is

consistent with the previous results.19 As the degree of oxidation/degradation of the

170

carbon increases due to higher number of oxygen groups, the activity of the catalyst

decreases due to a loss of active sites from nanoparticle isolation, as was the case for

Vulcan carbon.

The support degradation and consequent activity loss is alleviated for GMC

due to its higher degree of graphitization, which is expected to reduce the amount of

surface oxygen species (defect sites) as displayed in Table 5.4.5 Graphitization of the

mesoporous carbon enhances the basal layer dimension, as determined from XRD,

and thus leads to the reduction of the number of edge sites. This reduces the number

of dangling bonds (located on the edge sites), which readily react to form surface

oxides, thus mitigating carbon oxidation by Pt.3 This is confirmed by the insignificant

weight loss of carbon, (<0.7 wt %) for 20% Pt on GMC held isothermally at 195 °C

(Figure 5.7). In addition, the onset temperature of oxidation for GMC (~430 °C) is

higher compared to AMC (~310 °C), further demonstrating the high corrosion

resistance of graphitic carbon towards oxidation.. The high thermal stability of

graphitized carbons is expected to lead to high electrochemical stability towards

potential cycling (Table 5.4). After potential cycling, no “oxide shoulder” peaks

between 287-293 eV are observed for GMC, due to inherent resistance to oxidation

afforded to the support from graphitization. Graphitized mesoporous carbons thus

remain intact and do not oxidize upon exposure at 1.2V, preventing the electrical

isolation of the Pt nanoparticles and maintaining stable activities.

5.4.2 Carbon Synthesis and Stability by Pt-Carbon Interactions

The strong binding interaction between the Pt nanoparticles and GMC is

evidenced by the shift in the XPS Pt 4f7/2 peak to a higher binding energy (72.1 eV)

as compared to both unsupported Pt (71.0 eV) and Pt on amorphous AMC (71.2 eV)

(Figure 5.8).46,47 This shift in binding energy can be attributed to a change in the

coordination of the Pt due to the interaction of the metal with the carbon support.

While the exact nature of the Pt interaction with graphite is still a topic of

considerable discussion, it is reported that the delocalized electrons in the p-orbital of

171

the π sites on graphite overlap with the d-orbital of the Pt ([Xe] 4f14 5d9 6s1),

indicating that the nature of the binding is partially covalent.11 This occurs due to π-

backbonding, where the electron in the d-orbital of the metal is shared with a π*

antibonding orbital from the graphitic support. In addition, partial electron transfer

from the Pt to the graphitic surface has also been determined by previous XPS11 and

DFT51 studies. Thus, the strong partially covalent bond due to metal/support orbital

overlap (π backbonding) coupled with partial charge transfer between Pt and

graphitic surfaces, enables strong metal-support interactions and thus facilitates high

Pt loading.

The strategic selection of the weakly binding dodecylamine ligands with low

surface coverage for Pt stabilization (wt. of ligand on Pt was 25% as determined from

TGA) exposes sufficient bare metal to facilitate the strong metal-support

interactions.8,21,52 The low surface coverage of the metal by the ligand is reflected by

an ECSA of 52 m2/g for uncalcined Pt on GMC, which is 75% of the ECSA for Pt

calcined at 300 °C (70 m2/g Pt). All the ligands are expected to be removed at 300 °C

as dodecylamine starts oxidizing at 180-200 °C (as seen in TGA of pure Pt

nanoparticles). These results indicate that the ligand only covers ~25% of the Pt

active sites prior to calcination. The Pt-carbon interactions are further strengthened by

large interfacial contact area between the convex nanoparticles and the concave

mesopores of similar radii.21

The strong Pt interactions with graphitic carbon above, prevent the

coalescence of Pt nanoparticles during potential cycling, as seen by <10% change in

Pt size (ECSA and CSA) for GMC. These interactions restrain migration of the

nanoparticles on the carbon surface that would otherwise produce particle growth via

coalescence. The resistance to oxidative surface degradation of the graphitic carbon

helps keep the nanoparticles anchored to the surface and further prevents particle

migration. Finally, the high dispersion of the Pt on the high surface area mesoporous

supports lowers the tendency of Pt coalescence due to the large interparticle distances

172

and tortuous pathways in the 3D pores. All of these factors contribute to the

exceptional stability during potential cycling for the ORR reaction. In contrast, the

size increases significantly for Pt supported on AMC (>50%). For carbons with

significant amorphous character, the presence of oxygen complexes on the surface

withdraws electrons from the π sites, thus weakening the binding of the Pt to the

carbon surface.9 As the degree of graphitization increases, thereby raising the Pt-C

binding, sintering of metal nanoparticles has been shown to decrease by Coloma and

et al.9

During ADT, the stability in Pt size and ORR activity further indicate limited

dissolution of Pt from the metal surface (Table 5.4). At potentials above 0.7 V, Pt is

oxidized to PtO and forms multilayers of PtO at even higher oxidation potentials

(>1.0 V). PtO reacts with the hydrogen ions of the acidic electrolyte to form Pt2+ (PtO

+ 2H+ = Pt2+ + H2O). These Pt2+ ions can then migrate into the solution and may be

reduced during the reduction cycle of ADT to form Pt clusters which are not

supported on the carbon and hence do not participate in the ORR, thus causing loss in

ECSA and activity. Alternatively, Pt2+ can dissolve in the electrolyte and then

redeposit on other Pt nanoparticles causing 3D Ostwald ripening and consequent loss

in ECSA.

Given the stability in ECSA, CSA and ORR activity for the GMC, as

observed through XRD, Pt dissolution and Ostwald ripening appeared to be

insignificant, as a result of the resistance of the carbon to oxidation and strong metal-

support interactions. Although, the effect of metal-support interactions on the Pt

dissolution is not well studied, it is plausible that the C-Pt binding due to π

backbonding synergistically alters the electronic structure of the Pt metal, causing the

observed shift to higher binding energies for the Pt 4f7/2 XPS spectra for Pt on GMC,

thereby creating a stabilizing effect against Pt oxidation. The delocalization of

electron density in the orbital overlap between the Pt and the π sites on the carbon

lowers the tendency to form a Pt-O bond as the potential becomes more positive,

173

resulting in a decrease in Pt dissolution. A related method to alter the Pt electronic

structure to make it less resistant to dissolution is galvanic deposition of Au clusters

on Pt nanoparticles.53 For the more amorphous AMC carbon, the Pt-C interactions are

weaker (as shown by the binding energy in the Pt 4f7/2 XPS spectra as well as the

catalyst stability in ADT) whereby the weaker Pt- π site overlap makes the Pt more

susceptible to oxidation to PtO.

In Table 5.4 for Pt on Vulcan C, the loss is 70% in ECSA compared to only

55% in CSA. The higher loss for ECSA may be attributed to isolated nanoparticles

upon carbon corrosion that are no longer electrochemically active, as observed

previously.19 For AMC, the smaller and more similar losses in CSA (35%) and ECSA

(30%) indicate a smaller degree of nanoparticle isolation, consistent with the greater

resistance to C oxidation in Table 5.3, as well as the higher degree of graphitization,

which favors strong Pt-C interactions.

The lattice orientation of the graphene layers in GMC and AMC with respect

to the pore walls is difficult to characterize given the complexity of the bicontinuous

cubic mesostructure pore structure as reported previously.54 From the high stability

in the ORR activity and morphology of Pt during ADT, it appears that the graphene

layers were accessible in the pore walls to the Pt particles.

Traditional wetness impregnation techniques pose a variety of challenges in

the synthesis of the fuel cell catalysts with graphitic supports. The lack of polar

oxygenated sites on the graphitic surfaces precludes electrostatic interactions between

ionic metal precursors and the supports.6 Consequently, metal precursors nucleate and

grow to large nanoparticle sizes (>5nm) upon reduction limiting metal dispersion and

thus lowering ORR activities.6 To enhance the electrostatic interactions of the ionic

precursors with the supports, the graphitic carbons may be oxidized, for example,

with nitric acid to generate polar oxygen groups.16-19 Platinum nanoparticles with <3

nm size were synthesized on nitric acid treated graphitized carbon nanotubes (CNT)

by wetness impregnation and showed 15% loss in ECSA upon potential cycling

174

between 0.6 – 1.2 V for 1000 cycles.14 Though graphitization of the CNT partially

ameliorated the catalyst activity relative to Vulcan carbon, (which shows ~70% loss),

oxidation of the graphitized CNT generated corrosion prone sites on the surface and

may have led to the activity loss. The greater stability in ECSA for GMC in the

current study (essentially no loss) may be attributed to the greater dispersion and

separation of Pt particles in the 3D concave internal pores versus the external surface

of convex CNTs, as well as the smaller amount of oxygen species on the surface (no

nitric acid treatment) .

5.4.2 Control of Pt Nanoparticle Morphology for High Activity

Not only does the control of the Pt nanoparticle morphology in the pre-

synthesis technique offer a route to achieve high catalyst stabilities, but also high

activities. Thus, the lattice planes on the Pt particle surface may be designed in the

original arrested growth precipitation and maintained during infusion and calcination,

at relatively low temperatures (300 °C), as was shown by XRD. For ORR in HClO4

electrolyte, the (111) phase is more active than (110), which in turn is more active

than (100).1,55 In Figure 5.4, the intensity of the (111) peak is much stronger than the

(200) peak (which is associated with the (100) peak)55 indicating a higher fraction of

(111) over (100) facets (Figure 5.7). The FCC structure does not include a (110)

surface, further maximizing the (111) surface fraction. Furthermore, for the

polyhedral shape of the Pt particles according to TEM (Figure 5.5) the (111) surface

is dominant, consistent with the XRD results.55 Thus, the ability to design a high

fraction of highly active (111) sites in pre-synthesis, and to maintain these sites

through infusion and calcination, is an important strategy for raising catalyst activity.

This approach was also successful for the AMC catalyst. The activities are

comparable for AMC and GMC, indicating that metal-support interactions do not

appear to influence the catalytic behavior of platinum, consistent with previous

reports.9

175

5.5 Conclusions

Herein, we have designed highly stable Pt electrocatalysts for ORR on ordered

graphitized mesoporous carbon with controlled nanocrystal size and surface

morphology by infusion of pre-synthesized Pt nanoparticles. The loss in ORR catalyst

activity for Pt supported on graphitic mesoporous carbons under an accelerated

durability test of potential cycling between 0.5 to 1.2 V for 1000 cycles, is extremely

small, <2%, consistent with the lack of change in Pt size as observed by XRD, ECSA

and CSA. In comparison, the loss is 30% for Pt on amorphous mesoporous carbons

and 70% on commercial Pt on Vulcan carbon. The highly graphitic supports, with

relatively few surface oxygen sites, as shown by XRD and XPS, are highly resistant

to oxidation. This stability of the support minimizes nanoparticle isolation from the

surface and facilitates anchoring of the nanoparticles to the surface to inhibit

coalescence. The strong metal-support interactions from electron delocalization

between d-p orbitals of Pt and π sites and partial charge transfer from Pt to C as

observed by XPS51 and DFT11 calculations, mitigate nanoparticle migration and

coalescence, as well as Pt dissolution and subsequent Ostwald ripening. The high

dispersion of the Pt on the high surface area mesoporous supports with tortuous

pathways in the 3D pores, as well as the large interfacial contact area between the

metal nanocrystals and 7 nm concave mesopores with relatively similar radii further

lower the tendency of Pt coalescence. For the less graphitic AMC and Vulcan XC-72,

the greater losses activity results from carbon oxidation and weaker Pt-C interactions.

The arrested growth precipitation of Pt nanocrystals to form 3 nm Pt particles

with high concentrations of the highly active (111) surface leads to 25% higher

activities than typically seen for Pt formed on Vulcan carbon. The low surface

coverage of the weakly bound ligands favors strong Pt-C interactions8 as indicated by

high metal loadings of 20 wt % and ECSA of 52 m2/g. The active sites on the pre-

synthesized nanocrystals may be maintained during the facile removal of the weakly

bound ligands at low temperatures. In contrast, it is challenging to achieve high

176

loadings of Pt on highly graphitic carbon with low oxygenation by precursor

reduction, given weak adsorption of precursors on the carbon.3 The ability to control

catalyst size and surface morphology with pre-synthesized nanocrystals on highly

graphitic mesoporous carbons offers novel opportunities for designing highly active

and stable catalysts including electrocatalysts for oxygen reduction in PEMFCs.

5.6 References






(3) Shao, Y.; Yin, G.; Gao, Y. Understanding and approaches for the

durability issues of Pt-based catalysts for PEM fuel cell. Journal of Power Sources

2007, 171, 558-566.

(4) Wagner, F. T.; Gasteiger, H. A.; Makharia, R.; Neyerlin, K. C.;

Thompson, E. L.; Yan, S. G. Catalyst development needs and pathways for

automotive PEM fuel cells. ECS Transactions 2006, 3, 19-29.




2005, 14, 24-35.



(7) Waje, M. M.; Li, W.; Chen, Z.; Larsen, P.; Yan, Y. Effect of scan

range on Pt surface area loss in potential cycling experiments. ECS Transactions

2007, 11, 1227-1233.

177



6943.

(9) Coloma, F.; Sepulveda-Escribano, A.; Rodriquez-Reinoso, F. Heat-

treated carbon blacks as supports for platinum catalysts. Journal of Catalysis 1995,

154, 299-305.

(10) Antolini, E. Formation, microstructural characteristics and stability of

carbon supported platinum catalysts for low temperature fuel cells. Journal of

Materials Science 2003, 38, 2995-3005.

(11) Cuong, N. T.; Chi, D. H.; Kim, Y.-T.; Mitani, T. Structural and

Electronic Properties of Ptn (n=3, 7, 13) clusters on metallic single wall carbon

nanotube. Physica Status Solidi 2006, 13, 3472-3475.

(12) Kongkanand, A.; Kuwabata, S.; Girishkumar, G.; Kamat, P. Single-

Wall Carbon Nanotubes Supported Platinum Nanoparticles with Improved

Electrocatalytic Activity for Oxygen Reduction Reaction. Langmuir 2006, 22, 2392-

2396.

(13) Chen, Z.; Deng, W.; Wang, X.; Yan, Y. Durability and activity study

of single-walled, double-walled and multi-walled carbon nanotubes supported Pt

catalyst for PEMFCs. ECS Transactions 2007, 11, 1289-1299.

(14) Wang, J.; Yin, G.; Shao, Y.; Wang, Z.; Gao, Y. Investigation of

Further Improvement of Platinum Catalyst Durability with Highly Graphitized

Carbon Nanotubes Support. Journal of Physical Chemistry C 2008, 112, 5784-5789.




(16) Rajalakshmi, N.; Ryu, H.; Shaijumon, M. M.; Ramaprabhu, S.

Performance of polymer electrolyte membrane fuel cells with carbon nanotubes as

oxygen reduction catalyst support material. Journal of Power Sources 2005, 140,

250-257.

178

(17) Matsumoto, T.; Komatsu, T.; Nakano, H.; Arai, K.; Nagashima, Y.;

Yoo, E.; Yamazaki, T.; Kijima, M.; Shimizu, H.; Takasawa, Y.; Nakamura, J.

Efficient usage of highly dispersed Pt on carbon nanotubes for electrode catalysts of

polymer electrolyte fuel cells. Catalysis Today 2004, 90, 277-281.

(18) Li, X.; Hsing, I. M. Electrooxidation of formic acid on carbon

supported PtxPd1-x (x = 0-1) nanocatalysts. Electrochimica Acta 2006, 51, 3477-

3483.

(19) Shao, Y.; Yin, G.; Gao, Y.; Shi, P. Durability study of Pt/C and

Pt/CNTs catalysts under simulated PEM fuel cell conditions. Journal of the





Chemistry B 2005, 109, 2192-2202.

(21) Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.;

Johnston, K. P. Stable Ordered FePt Mesoporous Silica Catalysts with High Loadings








6239-6249.

(24) Ding, J.; Chan, K.-Y.; Ren, J.; Xiao, F.-s. Platinum and platinum-

ruthenium nanoparticles supported on ordered mesoporous carbon and their

electrocatalytic performance for fuel cell reactions. Electrochimica Acta 2005, 50,

3131-3141.

179

(25) Vijayaraghavan, G.; Stevenson, K. J. Synergistic Assembly of

Dendrimer-Templated Platinum Catalysts on Nitrogen-Doped Carbon Nanotube

Electrodes for Oxygen Reduction. Langmuir 2007, 23, 5279-5282.

(26) Liu, Z.; Koh, S.; Yu, C.; Strasser, P. Synthesis, dealloying, and ORR-

electrocatalysis of PDDA-stabilized Cu-rich Pt alloy nanoparticles. Journal of the

Electrochemical Society 2007, 154, B1192-B1199.

(27) Tian, Z. Q.; Jiang, S. P.; Liu, Z.; Li, L. Polyelectrolyte-stabilized Pt

nanoparticles as new electrocatalysts for low temperature fuel cells. Electrochemistry

Communications 2007, 9, 1613-1618.

(28) Xiong, L.; Manthiram, A. Nanostructured Pt-M/C (M = Fe and Co)

catalysts prepared by a microemulsion method for oxygen reduction in proton

exchange membrane fuel cells. Electrochimica Acta 2005, 50, 2323-2329.

(29) Raghuveer, V.; Ferreira, P. J.; Manthiram, A. Comparison of Pd-Co-

Au electrocatalysts prepared by conventional borohydride and microemulsion

methods for oxygen reduction in fuel cells. Electrochemistry Communications 2006,

8, 807-814.

(30) Kim, T.-W.; Solovyov, L. A. Synthesis and characterization of large-

pore ordered mesoporous carbons using gyroidal silica template. Journal of Materials

Chemistry 2006, 16, 1445-1455.





(32) Choi, W. C.; Woo, S. I.; Jeon, M. K.; Sohn, J. M.; Kim, M. R.; Jeon,

H. J. Platinum nanoclusters studded in the microporous nanowalls of ordered

mesoporous carbon. Advanced Materials (Weinheim, Germany) 2005, 17, 446-451.

(33) Nam, J.-H.; Jang, Y.-Y.; Kwon, Y.-U.; Nam, J.-D. Direct methanol

fuel cell Pt-carbon catalysts by using SBA-15 nanoporous templates.

Electrochemistry Communications 2004, 6, 737-741.

180

(34) Liu, S.-H.; Lu, R.-F.; Huang, S.-J.; Lo, A.-Y.; Chien, S.-H.; Liu, S.-B.

Controlled synthesis of highly dispersed platinum nanoparticles in ordered

mesoporous carbons. Chemical Communications (Cambridge, United Kingdom)

2006, 3435-3437.

(35) Wikander, K.; Ekstroem, H.; Palmqvist, A. E. C.; Lundblad, A.;

Holmberg, K.; Lindbergh, G. Alternative catalysts and carbon support material for

PEMFC. Fuel Cells (Weinheim, Germany) 2006, 6, 21-25.

(36) Wikander, K.; Petit, C.; Holmberg, K.; Pileni, M.-P. Size Control and

Growth Process of Alkylamine-Stabilized Platinum Nanocrystals: A Comparison

between the Phase Transfer and Reverse Micelles Methods. Langmuir 2006, 22,

4863-4868.

(37) Kim, T.-W.; Kleitz, F.; Paul, B.; Ryoo, R. MCM-48-like Large

Mesoporous Silicas with Tailored Pore Structure: Facile Synthesis Domain in a

Ternary Triblock Copolymer-Butanol-Water System. Journal of the American


(38) Ryoo, R.; Joo, S. H.; Jun, S. Synthesis of Highly Ordered Carbon

Molecular Sieves via Template-Mediated Structural Transformation. Journal of


(39) Kim, T.-W.; Park, I.-S.; Ryoo, R. A synthetic route to ordered

mesoporous carbon materials with graphitic pore walls. Angewandte Chemie,

International Edition 2003, 42, 4375-4379.

(40) Stevens, D. A.; Dahn, J. R. Thermal degradation of the support in

carbon-supported platinum electrocatalysts for PEM fuel cells. Carbon 2004, 43, 179-

188.

(41) Gierszal, K. P.; Kim, T.-W.; Ryoo, R.; Jaroniec, M. Adsorption and

Structural Properties of Ordered Mesoporous Carbons Synthesized by Using Various

Carbon Precursors and Ordered Siliceous P6mm and Ia.hivin.3d Mesostructures as

Templates. Journal of Physical Chemistry B 2005, 109, 23263-23268.

181

(42) Tian, J. M.; Wang, F. B.; Shan, Z. H. Q.; Wang, R. J.; Zhang, J. Y.

Effect of preparation conditions of Pt/C catalysts on oxygen electrode performance in

proton exchange membrane fuel cells. Journal of Applied Electrochemistry 2004, 34,

461-467.

(43) Kinoshita, K. Carbon: Electrochemical and Physicochemical

Properties; John Wiey and Sons, 1988.

(44) Baturina, O. A.; Aubuchon, S. R.; Wynne, K. J. Thermal Stability in

Air of Pt/C Catalysts and PEM Fuel Cell Catalyst Layers. Chemistry of Materials

2006, 18, 1498-1504.

(45) Darmstadt, H.; Roy, C.; Kaliaguine, S.; Choi, S. J.; Ryoo, R. Surface

Chemistry of Ordered Mesoporous Carbons. Carbon 2002, 40, 2673-2683.





(47) Antolini, E.; Giorgi, L.; Cardellini, F.; Passalacqua, E. Physical and

morphological characteristics and electrochemical behavior in PEM fuel cells of

PtRu/C catalysts. Journal of Solid State Electrochemistry 2001, 5, 131-140.

(48) Hodgkins, R. P.; Ahniyaz, A.; Parekh, K.; Belova, L. M.; Bergstroem,

L. Maghemite Nanocrystal Impregnation by Hydrophobic Surface Modification of

Mesoporous Silica. Langmuir 2007, 23, 8838-8844.

(49) Stevens, D. A.; Hicks, M. T.; Haugen, G. M.; Dahn, J. R. Ex Situ and

In Situ Stability Studies of PEMFC Catalysts. Journal of the Electrochemical Society

2005, 152, A2309-A2315.

(50) Kuo, T.-C.; McCreery, R. L. Surface Chemistry and Electron-Transfer

Kinetics of Hydrogen-Modified Glassy Carbon Electrodes. Analytical Chemistry

1999, 71, 1553-1560.

(51) Yang, D. Q.; Zhang, G. X.; Sacher, E.; Jose-Yacaman, M.; Elizondo,

N. Evidence of the Interaction of Evaporated Pt Nanoparticles with Variously Treated

182

Surfaces of Highly Oriented Pyrolytic Graphite. Journal of Physical Chemistry B

2006, 110, 8348-8356.




(53) Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Stabilization of Platinum

Oxygen-Reduction Electrocatalysts Using Gold Clusters. Science (Washington, DC,

United States) 2007, 315, 220-222.

(54) Yang, H.; Yan, Y.; Liu, Y.; Zhang, F.; Zhang, R.; Meng, Y.; Li, M.;

Xie, S.; Tu, B.; Zhao, D. A Simple Melt Impregnation Method to Synthesize Ordered

Mesoporous Carbon and Carbon Nanofiber Bundles with Graphitized Structure from

Pitches. Journal of Physical Chemistry B 2004, 108, 17320-17328.

(55) Wang, C.; Daimon, H.; Onodera, T.; Koda, T.; Sun, S. A general

approach to the size- and shape-controlled synthesis of platinum nanoparticles and

their catalytic reduction of oxygen. Angewandte Chemie, International Edition 2008,

47, 3588-3591.

183

Chapter 6

Highly Stable and Active Electrocatalysts for Oxygen Reduction:

Pre-synthesized Bimetallic Nanocrystals on Graphitic Carbon

Extremely stable and highly active electrocatalysts have been prepared by the

infusion of pre-synthesized bimetallic nanocrystals with controlled size (<4 nm),

composition and alloy structure onto graphitic mesoporous carbon. The low coverage

of weakly bound oleic acid and oleylamine ligands on the nanocrystal surface

facilitates strong metal-support interactions, resulting in nanocrystal loadings up to

~15 wt % with high dispersion as indicated by XRD and TEM. Electrochemically

dealloying via leaching of Cu from the Pt20Cu80 nanoparticle surface generates

Pt85Cu15 with a Pt rich shell. The specific current for the oxygen reduction reaction

(ORR) reaches 0.46 A/mg Pt as a result of the strained Pt shell, relative to 0.11 A/mg

Pt for pure Pt nanocrystals on amorphous carbon. For Pt-Cu supported on graphitic

mesoporous carbons, the loss in ORR catalyst activity over 1000 cycles from 0.5 to

1.2 V was extremely small, <2%, compared to a 40% loss for identical Pt-Cu

nanoparticles on amorphous mesoporous carbons. A typical loss for Pt on a

commercial amorphous carbon is 70%. The high resistance to oxidation of

graphitized carbon and the strong interactions of the π electrons with Pt prevent metal

sintering, Pt dissolution, and Ostwald ripening, resulting in high catalyst stability. The

high activities and stabilities for the ORR approach the levels targeted for practical

proton exchange membrane fuel cells. The ability to design catalysts composed of

pre-synthesized bimetallic nanocrystals with controlled size, composition and surface

properties onto high surface area mesoporous supports including graphitic carbon

offers the potential for new classes of highly stable and active catalysts.

The contents of this chapter are being prepared for a manuscript.

184

6.1 Introduction

The commercialization of polymer electrolyte membrane fuel cells (PEMFC)

depends upon improvement in catalyst activities by about 4-5 times relative to pure Pt

for the oxygen reduction reaction.(ORR).1-5 The activity may be raised up to about

2.5 times that of pure Pt with platinum based bimetallic alloys, Pt-M (M=Co, Ni, Fe,

V, Ti).6-13 Recently, Strasser and co-workers synthesized novel Pt-Cu catalysts on

carbon with a pronounced activity enhancement of >4 time that of pure Pt

catalysts.14,15 The copper was electrochemically etched from the nanoparticle surface

to form a strained Pt shell,15-19 with a reduced Pt-Pt distance, which alters the rate of

oxygen chemisorption and coverage by the intermediate oxygen species, and

enhances the activity for the ORR.

The second major challenge is to increase the catalyst stability upon exposure

to potential cycles encountered, for example, in automotive applications.2,5,20-22 Here,

catalyst stability at 1.2 V is required for an accumulated life time of 100 h, with a

maximum allowable performance loss of <30 mV at 1.5A/cm2 (<1 µV/cycle target).5

Pt supported on traditional primarily amorphous carbons show significant losses at

1.2 V from carbon oxidation as well as Pt dissolution upon potential cycling. For

example, for Pt on Vulcan XC-72 carbon, a 70% loss in electrochemical area was

observed upon potential cycling between 0.5 and 1.2 V for 1000 cycles.23 In highly

acidic environments, polar sites on amorphous carbons are oxidized at potentials ≥1.2

V leading to metal nanoparticle isolation and activity loss.2,5 The oxidation also

weakens the Pt-support interactions resulting in Pt dissolution and metal sintering.24

Graphitized carbons are more resistant to oxidation than amorphous carbons since

they contain much fewer defect sites and polar species on the surface.21,25-27 In

addition, the abundant π electron sites on the surface interact strongly with platinum,

thus strengthening the metal-support interactions, as demonstrated by XPS and

electron-spin resonance (ESR).21,25,27 Strong metal-support interactions reduce the

sintering, and coalescence of Pt. Charge transfer from the graphite to Pt28, inhibits Pt

185

oxidation and subsequent dissolution. Thus, corrosion resistant graphitic supports

with strong Pt-support interactions are expected to enhance the stability for the ORR.

Metals such as Pt are not dispersed efficiently on graphitic carbon upon

reduction of ionic precursor by traditional techniques such as wetness impregnation

.21,26,29 These precursors interact too weakly with graphitized supports, which do not

have a sufficient number of polar sites. For bimetallic catalysts, multiple metal

precursors are reduced on the supports either simultaneously30,31 or sequentially14,32.

This often produces nanoparticles with high polydispersity in size and composition,

with multiple alloy phases.33 The lack of uniformity results from varying rates of

nucleation and growth for each metal precursor in solution and on the support, as well

as non-uniform interactions with the heterogeneous support.15,17 These interactions

will be particularly weak for graphitic carbon. It is critical to control the size as well

as composition of bimetallic nanoparticles to achieve high catalyst activities. For

example, Pt-Co catalysts show enhanced activities over pure Pt at compositions of

Pt3Co but no appreciable enhancement at other compositions.9

An emerging concept in catalyst design is to pre-synthesize metal nanocrystals

with controlled size and composition via arrested growth precipitation with stabilizing

ligands and to infuse these nanocrystals onto high surface area supports.34-40 The

arrested growth precipitation step provides much greater control over the nanocrystal

size, composition, alloy structure than traditional techniques where reduction in the

presence of the support complicates nucleation, growth and the binding of stabilizing

ligands. For example, highly stable FePt catalysts with a controlled size, composition

and intermetallic alloy structure have been synthesized by pre-synthesizing FePt

nanocrystals and then infusing them into mesoporous silica.35 Weakly bound low

molecular weight ligands have been designed to stabilize the nanocrystals while

exposing significant area of the metal to the support for sufficient binding to achieve

loading up to 15 wt %.36,37 The weak binding facilitates ligand removal and catalyst

activation at low temperatures without perturbing the nanocrystal structure

significantly.

186

The primary objective of this study was to design highly stable Pt-Cu

catalysts on mesoporous graphitic carbons for the ORR, even for activities 4-fold the

typical values on Pt. Pt-Cu nanocrystals were pre-synthesized to controlled the

nanocrystal size, composition, and alloy structure, and then infused into the

mesoporous carbon to attempt to achieve high dispersion and nanocrystal stability.

We hypothesize that the high surface area and resistance to oxidation of the graphitic

carbon21 in addition to the strong binding of Pt21 as well as copper,41 will mitigate Pt-

Cu sintering and dissolution during potential cycling. The large interfacial area

between the pre-synthesized relatively spherical nanocrystals and the concave

mesopores further favors strong metal-support interactions. Weakly binding ligands

(oleic acid and oleylamine) were chosen to stabilize the Pt-Cu nanoparticles to expose

a significant fraction of the copper rich surface domains to the graphitic π sites

needed to promote adsorption and high loading of the nanocrystals. The low

molecular weight ligands were removed by calcination at relatively low temperature

to ameliorate changes in nanocrystal structure. The nanoparticles were subsequently

dealloyed electrochemically to selectively remove copper from the surface in order to

strain the platinum shell for the enhancement of the catalyst activity, as reported

previously.14,15 The degree of graphitization of the mesoporous carbons (without

metal) was characterized by XRD and Raman spectroscopy and the oxidative stability

by thermal gravimetric analysis. In addition, the size and composition of Pt-Cu

nanoparticles were characterized before and after electrochemical dealloying by

TEM, XRD and SEM-EDX to provide insight into the activity for the ORR reaction

and the stability under potential cycling between 0.5 and 1.2 V for 1000 cycles at 50

mV/s.

6.2 Experimental

All chemicals were used as received. Platinum (II) acetylacetonate (97%),

Copper bis(2,2,6,6-tetramethyl-3,5-heptanedionate), Oleylamine (70%), Oleic Acid

187

(99%) diphenyl ether (98% purity), perchloric acid 70% (99.999% purity), anhydrous

ethanol, and 5 wt % nafion solution in lower alcohols were purchased from Aldrich.

Toluene (99.9%) was obtained from Fisher Scientific, and ethanol (Absolute 200

proof) from Aaper alcohol. Polycarbonate filters (0.05 micron) were obtained from

Osmonics. High purity water (resistance ~18 MΩ) was used. Oxygen (research grade,

99.999% purity) and argon (research grade, 99.999% purity) were obtained from

Praxair.

Pt-Cu particles were synthesized by employing the modified polyol process

described by Sun et al.42 The synthesis was carried out in a 50 mL three-neck flask

under an argon atmosphere using standard Schlenk line techniques. For the

preparation of Pt20Cu80 nanoparticles, 0.16 mmol Platinum acetylacetonate and 0.50

mmol Copper bis(2,2,6,6-tetramethyl-3,5-heptanedionate) were separately dissolved

in 10mL of diphenyl ether and injected into the flask. The solution was continuously

stirred and heated to 80 °C using a heating mantle. At 80 °C, 0.5 mmol of the

stabilizing agents, oleic acid and oleyl amine, were injected. The temperature was

then raised to the reflux temperature of diphenyl ether (~260 °C). At around 215 °C a

sudden change in the reaction mixture was observed as the solution turned yellow and

then dark brown, indicating the reduction of the precursors. The reaction mixture was

maintained at 260 °C for 30 minutes and then cooled down to room temperature. The

dark dispersion was removed from the flask and the particles were isolated by the

addition of 50mL of acetone and centrifugation. The nanoparticles were further

cleaned by redispersing the Pt20Cu80 particles in 1mL of toluene and excess acetone

and centrifuging again. The other compositions of Pt-Cu were synthesized by

changing the molar ratios of the precursors.

Disordered mesoporous carbons (DMC) were prepared via self-assembly of

block copolymer and phenolic resin under acidic conditions as described in our

previous reports.44,45 Typically, 16.5 g of resorcinol and 16.5 g of F127

(EO106PO70EO106) were dissolved in 67.5 mL of EtOH and 67.5 mL of HCl aqueous

solution (3.0 M). To this solution, 19.5 g of formaldehyde (37 wt % in H2O) was

188

added with stirring at room temperature. As the polymerization reaction of resorcinol

and formaldehyde proceeded, self-assembly of in situ formed phenolic resin and F127

induced a phase separation. The polymer-rich gel phase was collected in the bottom

of centrifuge tubes by centrifugation at 9500 rpm for 4 min after the mixture was

stirred for 30 min. The gel was redissolved in a mixed solvent of 18 g of THF and 12

g of EtOH and the mixture was loaded on a substrate, dried at room temperature

overnight, and cured at 80 °C for 24h. Carbonization was carried out under N2

atmosphere at 400 °C for 2 h with a heating rate of 1 °C/min, which was followed by

further treatment at 850 °C for 3 h with a heating rate of 5 °C/min. To enhance the

degree of graphitization, the heat treatment of mesoporous carbons was carried out in

a high-temperature furnace (Thermal Technology Inc., CA) under helium atmosphere

at desired temperatures for 1 h with a heating ramp of 20 °C/min. The mesoporous

carbons heat treated at temperatures of 850 °C, 2000 °C and 2600 °C will be referred

to as DMC-850, DMC-2000 and DMC-2600 respectively for the rest of the article.

The conductivity of the DMC-2000 and DMC-2600 is ~300 S/cm, about 100 times

higher than that of Vulcan XC-72 carbon46 (3.0 S/cm). The conductivity of the less

graphitic DMC-850 is only 2.0 S/cm. The pores of the mesoporous carbons were

disordered in nature as peaks were not present in small angle X-ray scattering (not

shown).

The infusion of the Pt-Cu nanoparticles into mesoporous carbon was done by

a similar procedure as reported in Gupta et al.38 25 mg of mesoporous carbon was

mixed with 3 mL of a Pt-Cu nanocrystal dispersion of known concentration, ~20

mg/mL, in toluene. The contents were stirred for 4 days. The product was separated

from the nanocrystal supernatant by filtration with a 0.2 μm PTFE filter. The extent

of incorporation of Pt-Cu nanocrystals in the carbon was determined by subtracting

from the initial mass of Pt-Cu in toluene, the final mass recovered in the supernatant

after filtration.36 The masses were determined by absorbance of the Pt-Cu

nanocrystals with a Cary 500 UV-Vis-NIR spectrophotometer with an optical path

189

length of 1 cm (Appendix B, Figure B.2). A standard linear calibration curve was

generated with known concentrations of Pt-Cu nanocrystals in toluene for a

wavelength of 800 nm with correlation coefficient of 0.998 (Appendix B, Figure B.2).

Loadings of ~15% metal on mesoporous carbons were obtained after infusion (Table

6.2). The individual uncertainties in the loadings produced by factors of volume

measurement, weight of carbon, absorbance, and linearity of the calibration curve are

(+2%), (+1%), (+5%) and (+3%) respectively. The total uncertainty is calculated

conservatively by adding the individual uncertainties in loadings and is ~11%, as

explained in detail previously.37 To activate the Pt-Cu catalysts ligands were

removed by calcining the catalyst at 350 °C for 2h. The temperature was ramped at 16

°C/min. After calcination, the catalysts were washed with 100 mL of ethanol.


using a Bruker-Nonius D8 Advance diffractometer. Samples were scanned for 12 h at

a scan rate of 12 deg/min with 0.02 deg increments. The average nanocrystal size was

estimated from the Scherrer equation using JADE software (by Molecular Diffraction

Inc).

Raman spectra were acquired using a Renishaw inVia Microscope using a

514.5 nm Argon laser at 50% power with a 50x aperture. Four acquisitions were

taken for each sample from 500 to 2000 cm-1 for 50 seconds. The spectra were fit

using the PeakFitTM software, following the convention set forth by Cuesta et. al47 of

fitting five Gaussians at 1624, 1583, 1487, 1351, and 1220 cm-1, corresponding to the

D’, G, D”, D, and I bands, respectively. All spectral fits had correlation factors (R2)

greater than 0.98.

A glassy carbon rotating disk electrode (0.196 cm2, from Pine Instruments)

was used as a substrate for the catalyst. It was polished with a 0.3µm followed by a

0.05 µm alumina suspension to give a mirror finish. 1.0 mg of Pt-Cu catalyst was

suspended by sonication in 300µL of anhydrous ethanol and 300 µL of 0.15 wt %

nafion solution (diluted from 5% nafion stock solution). A total of 10 µL of Pt-Cu

190

suspension was pipetted onto the substrate and was dried at room temperature for 15

min.

Electrochemical measurements were carried out using a three-compartment

electrochemical cell with a Pt wire counter electrode and a Hg/Hg2SO4 reference

electrode. Electrochemical measurements were recorded using CHI 832

electrochemical analyzer (CH instruments Inc., Austin). All potentials cited in this

article are normalized with respect to NHE.

After preparation of the catalyst on the substrate, the catalyst was dealloyed to

leach off copper by cycling the potential between 0.04 and 1.24 V at 200 mV/s for

500 cycles in argon saturated 0.1M HClO4 electrolyte.1 The area under the desorption

peak of underpotentially deposited hydrogen in a cyclic voltammogram (scanned at

100 mV/sec between 0.04 and 1.24V), was used to evaluate electrochemically active

surface area (ECSA) of the Pt catalyst on the electrode. Hydrodynamic

voltammograms at rotating disk electrode were recorded by anodically scanning the

potential from 0.04 to 1.24 V at 5 mV/s in oxygen saturated 0.1 M HClO4. The

oxygen pressure was 1 atm. Accelerated durability tests (ADT) were employed by

cycling the electrode potential between 0.5 and 1.2 V at 50 mV/s for 1000 cycles in

an argon saturated atmosphere for a total time of 7h.22

6.3 Results

6.3.1 Structure of the Pt-Cu on Mesoporous Carbon Catalysts

The surface area, pore volume and mean pore size for the three carbons

measured by nitrogen porosimetry (see supplementary information) is given in Table

6.1. As the calcination temperature for the carbons increases from 850 °C to 2600 °C,

both the surface areas and the pore volume decrease as expected. DMC-850 and

DMC-2000 show a surface area of 577 and 344 m2/g respectively, which are

191

Carbon Surface

Area (m2/g)

Pore volume

Pore size (nm)

La (Å)

Lc (Å)

d (Å)

DMC 850 577 0.70 8.1 33.8 16 3.81

DMC 2000 344 0.57 8.0 35.4 22 3.63

DMC 2600 262 0.39 7.9 43.2 30 3.35

Table 6.1

Properties of mesoporous carbon.

192

significantly higher than commercial Vulcan XC-72 (~250 m2/g) supports. The pore

size for all the three mesoporous carbons was ~8 nm and did not show any shrinkage

in the pore diameter or structural collapse with increasing temperature, manifesting

the inherent stability of the supports. The pores of the mesoporous carbons were

disordered in nature as peaks were not present in small angle X-ray scattering.

XRD spectra of the carbons at annealed at varied temperatures are shown in

Figure 6.1. The peak at ~25° corresponds to the graphitic <002> plane on carbon. A

sharp <002> peak is associated with a higher graphitic character as compared to a

broad diffuse peak seen in amorphous carbons.48 With increasing annealing

temperature, the <002> peak becomes sharper and narrower, indicating an increase in

the degree of graphitization. The layer dimension parallel to the basal place, La, and

perpendicular, Lc, were calculated from the <10> and <002> reflections respectively,

according the relations,48,49

La = 1.84λ/Bcos(θ) and

Lc = 0.89λ/Bcos(θ)

where λ is the wavelength of the x-rays (1.54A), B is is the full angular width at half

max, and θ is the Bragg angle (Table 6.1). The monotonic increase in the values of La

and Lc along with increasing temperature indicates progressively enhanced degree of

graphitization of carbon supports. Table 6.1 also lists the interplanar d-spacing

calculated from the <002> reflection using Bragg’s law (nλ = 2dsin(θ)). There is a

monotonic decrease in d-spacing with increasing temperature, indicating a tighter

packing of the carbon planes, again consistent with an increasing graphitization.48

The first order Raman spectra for DMC-850, 2000, and 2600 are shown in

Figure 6.2. There are two peaks apparent in the spectra for all three samples,

corresponding to the D band at 1355 cm-1 and the G band at 1583 cm-1. Increasing the

annealing temperature for the carbons enhances the overlap between the two bands,

making it necessary to deconvolute the FWHM values by fitting five bands at 1624,

1583, 1487, 1355, and 1220 cm-1, corresponding to the D’, G, D”, D, and I bands,

193

0

5000

1 104

1.5 104

2 104

2.5 104

0 20 40 60 80 100

DMC 850 °CDMC 2000 °CDMC 2600 °C

Inte

nsity

2 theta

002

100

004

Figure 6.1

XRD spectra of the disordered mesoporous carbons annealed at different

temperatures.

194

0

2000

4000

6000

8000

1 104

500 1000 1500 2000

DMC 850DMC 2000DMC 2600

Inte

nsity

Raman Shift (cm-1)

Figure 6.2

Raman spectra of the disordered mesoporous carbons annealed at different

temperatures.

195

respectively, following the convention of Cuesta et. al.47 The D band is often referred

to as the “disorder” band and is commonly believed to be a Raman inactive mode that

becomes active as the disorder in the system is increased, or likewise as the symmetry

at or near a crystalline edge is reduced.50,51 The G band is present in all sp2 hybridized

carbon materials and arises from the E2g vibrational mode. By determining the area

under the peak for the D band and the G band, ID/IG (ratio of amorphous to graphitic

character) were estimated and decreased monotonically as 1.30, 1.24 and 1.02 for

DMC-850, 2000 and 2600 respectively. The decreasing ID/IG with increasing carbon

temperature is indicative of improving degree of graphitization. The relative levels of

graphitization can be discerned by estimating the in plane crystalline length, La (=

4.4/(Id/Ig))52 in nanometers. As the ID/IG ratio decreases, La increases (La=3.38, 3.74,

4.32 nm for DMC 850, 2000 and 2600 respectively) as graphitic crystallites grew in

size with increasing carbon treatment temperature.

Two different compositions of Pt-Cu nanoparticles were synthesized by using

the molar precursor ratio of Pt:Cu ~24:76 and 50:50 respectively. After synthesis of

the nanoparticles, the atomic/molar compositions were determined to be Pt20Cu80 and

Pt45Cu55 respectively using SEM-EDX, and matches well with the starting precursor

molar ratio. For DMC-850 where the loading was 17% (Table 6.2), about ~25% of

total initial nanoparticles were infused in the mesopores. To enhance the loadings for

the more graphitic DMC-2000, the initial concentration of Pt20Cu80 was increased to

30 mg/mL and the time of infusion was raised to 5 days. For DMC-2600 with the

highest degree of graphitization, loading were limited for Pt20Cu80, but were higher

for Pt45Cu55 since Pt binds more strongly to graphitic C. In contrast, for amorphous

Vulcan XC-72, almost all the Pt20Cu80 nanoparticles were adsorbed in a relatively

short time span of 2 days and at a low starting concentration of 5 mg/mL, suggesting

that the polar sites on the amorphous carbon facilitated adsorption. However, the

desired loading could be obtained on the DMC-2000 with higher concentrations to

drive the thermodynamics and longer times.

196

Nanoparticles Support

Time of infusion (days)

Initial Conc

(mg/ml)

Final Conc.

(mg/ml)

Loading (wt %)

Pt20Cu80 DMC-

850 3 20 15.5 17

Pt20Cu80 DMC-2000 3 20 18 8

Pt20Cu80 DMC-2000 5 30 26 14

Pt20Cu80 DMC-2600 7 30 29 <5

Pt45Cu55 DMC-2600 7 30 26 14

Pt20Cu80 Vulcan-XC72 2 5 <0.1 19

Table 6.2

Infusion of Pt-Cu nanoparticles in mesoporous carbon supports. All infusions were

done on 50 mg carbon and nanoparticles were dispersed in 3 mL of toluene.

197

The Pt-Cu catalysts were electrochemically dealloyed to leach off copper from

the nanoparticle surface and produce a highly active strained Pt surface.1 Figures 6.3

(and Appendix B, Figure B.4) show voltammograms for Pt20Cu80/SDC-2000 catalyst

during dealloying cycles for slow potential cycling between 0.04 and 1.24 at 20

mV/s,. The first dealloying cycle does not show a characteristic anodic stripping peak

between 0.04 and 0.4V, associated with underpotentially deposited hydrogen on

platinum surfaces. Here, the surface of the Pt-Cu nanoparticles is primarily copper

and contains very little Pt. This result is consistent with the segregation of the copper

on the surface in Cu rich Pt-Cu alloys, given its low surface energy.53 A broad anodic

peak appeared at 0.3 V and extended almost to 0.85 V indicating dissolution of the

copper. After 3 dealloying cycles, two distinct copper dissolution peaks at 0.3 V and

0.7V emerge. With the progression of the cycles, the peak intensity and area of

copper dissolution peaks decreased as the surface became enriched with Pt, indicated

by the appearance and growth of an anodic hydrogen desorption peak between 0.04

and 0.4V. After about 500 cycles, the cyclic voltammograms reached a steady state or

equilibrium behavior with the signature of a pure Pt surface (Appendix B, Figure

B.5), with the familiar H2 ad/desorption range (0.05-0.4 V) as well as the Pt oxide

formation and desorption peaks at 0.83 and 0.72 V on the anodic and cathodic scan,

respectively.

The dealloying mechanism changes significantly upon progressive Cu

removal. After just 20 dealloying cycles, copper dissolution peaks at 0.3 and 0.7 V

essentially disappear, as the surface becomes composed of almost pure Pt. However,

the “strongly” and “weakly” bound hydrogen desorption peaks associated with Pt

<100> and <111> respectively, are not present, indicating an amorphous structure.54

After the completion of 500 dealloying cycles, these peaks become present at 0.16

and 0.21 V (Figure 6.3) indicating crystalline platinum.55 Thus, while the initial

dealloying cycles facilitate Cu removal from the surface, the later dealloying cycles

facilitate the structural reordering of the platinum atoms upon crystallization.

198

-1.20

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

0.000 0.200 0.400 0.600 0.800 1.000 1.200

Potential vs NHE

Cur

rent

(mA

)/ cm

2 ge

o

cycle 1

cycle 20

cycle 500

cycle 2

Figure 6.3

Cyclic voltammetry of Pt20Cu80 nanoparticles supported on DMC-2000 during

electrochemical dealloying at 20 mV/s in argon atmosphere.

Cycle 500 Cycle 1

Cycle 2

Cycle 20

199

1000

1500

2000

2500

3000

20 30 40 50 60 70 80 90 100

before dealloying after dealloying

Inte

nsity

2 theta

111

311220

311

220200

111

(a) (b)

(C)

Figure 6.4

XRD spectra of the Pt20Cu80 nanocrystals supported on DMC-2000 before and after

electrochemical dealloying of copper (a); peak analysis of the XRD spectra of the as-

synthesized Pt20Cu80 nanoparticles (b); and electrochemically dealloyed nanoparticles

with a final composition of Pt85Cu15.

200

Figure 6.4(a) shows XRD spectra of the Pt20Cu80/DMC-2000 before and after

dealloying. Before dealloying, the <111> peak approaching that of pure copper is

present., The peaks at 43.1°, 69° and 83.1°, are associated with the <111>, <220> and

<311> crystal planes of FCC lattice. The phases of the nanocrystal were investigated

by peak fitting using JADE software, as shown in Figure 6.4(b) and Table 6.3. A

single-phase Cu rich alloy with a peak at 42.7° was present, along with a small Pt rich

phase as indicated by the peak at 40.1°. From the relative areas, the Cu rich phase was

~95% with only about ~5% of a Pt rich phase. The peak position can be used to

determine the phase compositions with Vegard’s Law.56,57 The composition of the

FCC-disordered Pt–Cu alloy can be estimated from observed unit cell parameters

according to ac = a0 − k. xCu where a0= 3.923 Å is the unit cell parameter of pure Pt,

ac is the measured cell parameter of a Pt–Cu alloy, xCu is the molar ratio of Cu in the

binary alloys, and k is a constant (~0.308 Å for the Pt-Cu system).58 The unit cell

parameter ac for the Pt-Cu alloy can be related to the lattice spacing for 111 peak and

2θ according to

)sin(2

3/

111

111

θλ CuPtc

CuPt

d

ad−

−

=

= (6.1)

The composition of the copper rich phase was Pt17Cu83, while that of the platinum

rich phase was ~Pt90Cu10. The composition of the copper rich phase agrees well with

that determined by SEM-EDX. The nanoparticle size was estimated to be 3.0 nm

from the Scherrer equation.

After dealloying, the <111> peak shifts to lower 2θ values relative to the

<111> peak before dealloying, as copper is leached from the nanoparticle surface

(Figure 6.4(a)). After dealloying four peaks appear at 40.54°, 46.74°, 68.4° and

81.26°, corresponding to the <111>, <200>, <220>, and <311> FCC lattice planes,

respectively. As shown in Figure 6.4(c), only one platinum-rich alloy phase was

present in the dealloyed nanoparticles. The composition was estimated to be Pt85Cu15

201

Catalyst DTEM (nm)

DXRD (nm)

CSATEM (m2/g)

SEM-EDX

Compo-sition

Bravais lattice

Lattice Parameter

(A)

Composition

(Vegard’s law)

Ratio of 111 peak areas

Before dealloying 3.25 3.0 86 Pt20Cu80

FCC1 FCC2

3.663 3.881

Pt17Cu83 Pt90Cu10

FCC1:FCC2= 19:1

After dealloying 2.95 2.7 95 Pt85Cu15 FCC 3.876 Pt85Cu15 -

Table 6.3

Composition and structural analysis of 14% Pt20Cu80/DMC-2000 before and after

dealloying.

202

using Vegard’s Law (Table 6.3). It was in good agreement with the bulk composition

of Pt82Cu18 determined from SEM-EDX. The nanocrystal size after dealloying from

the Scherrer equation was 2.8 nm indicating a 10% reduction after dealloying.

Figure 6.5 gives the TEM micrographs and the corresponding histograms of

Pt20Cu80 on DMC-2000 before and after electrochemical dealloying. The

nanoparticles are well dispersed on the support with little aggregation or sintering

upon ligand removal or during the dealloying process. The average nanoparticle size

increases slightly from 2.85 nm (Appendix B, Figure B.2) to 3.25 nm (Figure 6.5)

upon thermal calcination for ligand removal. It decreased from 3.25 nm to 2.95 nm

upon dealloying (Table 6.3) in good agreement with the sizes from XRD. The

chemical surface area (CSA) after dealloying was determined from the particle sizes

(CSA= 3/(ρPt.rnp))and was estimated to be ~95-100 m2/g by XRD and TEM (Table

6.3).

6.3.2 Activity and Stability of Pt-Cu Catalysts for ORR

The electrochemical surface area (ECSA=QH/mpt.ν.c) of the platinum enriched

nanocrystals after dealloying was estimated from the area of the anodic H2 desorption

peak between 0.04 and 0.4V,59 where QH is the charge for desorption, mpt is the mass

of platinum on the electrode, ν is scan rate (V/s) and c is a constant defined by the

charge required to oxidize a monolayer of hydrogen on Pt (0.21 mC/cm2). The

estimated ECSA of 83m2/g Pt was comparable or slightly better than that of

commercial Pt catalysts.23 The ECSA was ~85% of the chemical surface area

estimated by XRD and TEM (Table 6.3), indicating that 85% of the total Pt sites

participate in the ORR.

Figure 6.6 shows a representative linear scan voltammogram obtained for the

oxygen reduction reaction with dealloyed Pt-Cu/DMC-2000. The diffusion-limited

current densities were 6.3, 6.0 and 5.6 mA/cm2 for Pt20Cu80 / DMC-2000, Pt20Cu80 /

DMC-850 and Pt45Cu55 / DMC-2000, respectively. These results are similar to the

203

(A)

(B)

(C)

(D)

Figure 6.5

TEM micrographs and histograms of Pt20Cu80 on DMC-2000 (A,C) as-synthesized

and calcined at 350 °C (B,C) Electrochemically dealloyed. Average nanoparticle size

changed from 3.25 nm to 2.95 nm after dealloying.

0

2

4

6

8

10

12

14

2 2.6 3.2 3.8 4.4 5

Cou

nt

PtCu Nanoparticle size (nm)

0

2

4

6

8

10

12

1.75 2.25 2.75 3.25 3.75 4.25 4.75

Cou

nt

Nanoparticle Size (nm)

204

-7

-6

-5

-4

-3

-2

-1

0

1

0 0.2 0.4 0.6 0.8 1 1.2

14% Pt20Cu80 / DMC-200017% Pt20Cu80 / DMC-85014% Pt45Cu55 / DMC-2600

I ( m

A/c

m2 e

lect

rode

)

Potential (vs NHE)

Figure 6.6

Linear scan voltammetry of the dealloyed Pt-Cu nanoparticles supported on DMC at

5 mV/s with rotation rate of 1600 rpm.

205

value of 6 mA/cm2 for a smooth polycrystalline Pt-disk electrode, demonstrating

negligible mass-transport effects.

Figure 6.7 shows the Tafel slopes for the dealloyed Pt20Cu80 on DMC-2000

and DMC-850 catalysts, in the kinetically controlled regime relative to a commercial

platinum catalyst. Both the mass and surface activities for the dealloyed Pt20Cu80 are

significantly better than those of the commercial Pt catalyst, as seen in previous

studies with similar Pt-Cu composition.1,12 These values are also reported in Table 6.4

at 0.9V. The value of 0.45 A/mg Pt for dealloyed Pt20Cu80 on DMC-850 and DMC-

2000 was ~4 times better than the commercial Pt catalyst (~0.11 A/mg Pt). For these

two supports, mass activities did not change with the degree of graphitization.

Similarly, the area activities were more than three times greater than that of the

commercial Pt catalyst. The mass activity was only half as large for Pt45Cu55

supported on DMC-2600. For Pt75Cu25 nanoparticles supported on DMC-850, the

activity was 0.16 A/mg Pt, marginally higher than for pure Pt.

The activities of the catalysts were measured after ADT cycling (Figure 6.7

and Table 6.4). Both Pt20Cu80 / DMC-2000 and Pt45Cu55 / DMC-2600 did not show

any significant losses in mass activity or ECSA after ADT cycling, indicating high

catalyst stabilities for these highly graphitic supports. The loss was significant, about

40%, for Pt20Cu80 on the less graphitic DMC-850. However, the loss in ECSA was

only 15%. The loss was even greater, >50%, for Pt20Cu80 on Vulcan carbon after ADT

cycling. Pt20Cu80 supported on DMC-2600 did not show any loss in catalyst activity,

consistent with the high corrosion resistant of carbon.

6.4 Discussion

6.4.1 Carbon Corrosion and Impact on Stability

Pt-Cu supported on the DMC-850 support showed significant loss in activity

upon ADT potential cycling. DMC-850 is amorphous in nature as shown in Figure

6.1 and 6.2. It is oxidized when subjected to potentials ≥1.2 V in highly corrosive

206

Sample* I (A/mg Pt)

Before ADT†

I (µA/cm2) Before ADT†

ECSA before ADT

(m2/g Pt)

I (A/mg Pt)

after ADT†

I (µA/cm2) after ADT†

ECSA after ADT

(m2/g Pt)17% Pt20Cu80 /

DMC- 850 0.46 630 73 0.27 435 62

14% Pt20Cu80 / DMC - 2000 0.45±0.03 537±25 83±4 0.45±0.03 540±28 84±4

14% Pt45Cu55- DMC-2600 0.23 382 59 0.22 332 66

5% Pt20Cu80 / DMC- 2600 0.16 232 68 0.15 150 66

20% Pt20Cu80 / vulcan-XC72 0.42 495 68 0.20 198 38

* All samples have been dealloyed by potential cycling

† All activities are calculated at 0.9V

Table 6.4

Activity and stability of dealloyed Pt-Cu catalysts supported on mesoporous carbons.

207

(a)

(b)

Figure 6.7

Potential-Activity plots of dealloyed Pt20Cu80 nanoparticles supported on

DMC-2000 and DMC-850 compared to commercial Pt (ETEK) catalysts (a) Pt mass-

based activities and (b) Pt specific surface area activities.

208

acidic environments as reported previously for other amorphous carbons.21,28 The

surface oxygen species on the amorphous carbon acts as reactive centers and

accelerate carbon oxidation.26,27 Pt atoms on the carbon surface (after dealloying Pt-

Cu) catalyze the carbon oxidation.60 We found that 20% Pt on amorphous Vulcan

XC-72 lost about 9% of its weight under isothermal conditions of 195 °C in 16 days

due to carbon loss. Degradation of carbon leads to metal nanoparticle isolation and

partly contributes to activity losses during ADT potential cycling (Figure 6.7).

The support degradation and consequent activity loss were alleviated by

graphitization of the carbon support as seen for DMC-2000 and 2600 (Table 6.4).21

The high temperature graphitization reduces the number of surface oxygen on the

surface and thus increases resistance to oxidation.28 For example, the onset

temperature for 20% Pt on graphitized DMC-2000 is ~180 °C higher than for 20% Pt

on amorphous DMC-850, as seen by TGA (Appendix B, Figure B.7) indicating

significantly higher oxidation resistance. The weight loss of carbon for 20% platinum

on the more graphitic carbons under isothermal conditions at 195 °C for 16 days was

insignificant, (<0.7 wt %, which is within the error limit) also indicating high

oxidation resistance. Graphitized mesoporous carbons remain intact and do not

oxidize even at 1.2V, thus preventing the electrical isolation of the dealloyed Pt-Cu

nanoparticles as indicated by the insignificant loss in ECSA and ORR activity (Table

6.4).

6.4.2 Metal-Support Interactions: Catalyst Synthesis and Stability

The high metal loadings of Pt-Cu nanoparticles infused into the graphitized DMC-

2000 and DMC-2600 were facilitated by the short-ranged interactions between the

metal and the support, as shown previously for metals on silica supports.38 The

surface of the Pt20Cu80 nanoparticles is Cu-rich as observed from the initial dealloying

curves (Figure 6.3). Relatively strong interactions between Cu and graphitic fullerene

due to charge transfer from Cu to fullerene have been identified by means of a high

frequency Raman spectra.42 The peak width of the pentagon-pinch mode becomes

209

wider and the peak position shifts to the lower frequency as the atomic ratio of Cu/C

increases, indicating strong metal-support interactions. These short-ranged

interactions enable the adsorption of Cu-rich Pt-Cu nanoparticles on the mesoporous

carbon supports to render high loadings (Table 6.2).

A key challenge to adsorb high levels of nanoparticles onto carbon is to

overcome screening of the short-ranged interactions by the hydrocarbon capping

ligands. The strategic selection of the weakly binding oleic acid/oleyl amine ligands

with low surface coverage (wt of ligand on Pt-Cu was 20% as determined from

TGA), exposes enough bare metal to facilitate strong metal (primarily Cu)-support

interactions.24,38,61 These short-ranged interactions are further strengthened by the

favorable geometric contact area between the convex nanoparticles and the concave

mesopores.38 In addition, the weak binding allows facile removal of the ligands at a

relatively low temperature of 350 °C, without causing any significant nanoparticle

sintering.

After dealloying, the Pt-rich surface (Figure 6.3) binds even more strongly to

graphite than the Cu-rich surface in the original particles. Pt binds very strongly with

the graphitized carbons, as indicated by the shifting of the binding energy for the Pt

4f7/2 peak to higher energy value (refer XPS data in Appendix B, Figure B.6)

compared to both unsupported Pt and Pt on amorphous carbon.62,63 The exact nature

of the Pt interaction with graphite is still a topic of considerable discussion. However,

it is reported that the delocalized electrons in the p-orbital of the π sites on graphite

overlap with the d-orbital of the Pt, indicating that the nature of the binding is

partially covalent.29 In addition, partial electron transfer from the Pt to the graphitic

surface has also been determined by previous XPS29 and DFT64 studies. Thus, the

strong partially covalent bond due to orbital overlap coupled with partial charge

transfer between Pt and graphitic surfaces, enables strong metal-support interactions

and thus facilitates high catalyst stability (Table 6.4). The metal-support interactions

prevent the coalescence of dealloyed Pt nanoparticles during stability cycling,

210

consistent with no loss in electrochemical surface area after ADT (Figure 6.7),

typically observed for amorphous carbons.20 Similar beneficial effect of strong Pt-

graphite interactions for preventing coalescence of Pt nanoparticles at high

temperatures has been reported by Coloma et al.25

The stability in ECSA, CSA and ORR activity for the dealloyed Pt-Cu on

DMC-2000 and DMC-2600 suggest insignificant Pt dissolution and/or subsequent

Ostwald ripening, as a result of the resistance of the carbon to oxidation and strong

metal-support interactions. The relatively monodisperse size of the dealloyed particles

further limits Ostwald ripening. Although, the effect of metal-support interactions on

Pt dissolution is not well studied, it is plausible that the C-Pt binding synergistically

alters the electronic structure of the Pt, causing the observed shift to higher binding

energies for the Pt 4f7/2 XPS spectra (Appendix B, Figure B.6), thereby creating a

stabilizing effect against Pt oxidation. The delocalization of electron density in the

orbital overlap between the Pt and the π sites on the carbon lowers the tendency to

form a Pt-O bond as the potential becomes more positive, resulting in a decrease in Pt

dissolution. A related method to alter the Pt electronic structure to make it less

resistant to dissolution is galvanic deposition of Au clusters on Pt nanoparticles.17 The

large interfacial contact area between the nanoparticles and concave mesopores with

relatively similar radii, as well as the high dispersion of the nanoparticles on the high

surface area mesoporous supports with tortuous pathways in the 3D pores coupled

with the strong metal-graphite interactions appeared to prevent Pt dissolution,

Ostwald ripening and nanoparticle coalescence.

Although traditional wetness impregnation techniques are applicable for

producing catalysts with relatively amorphous carbons, a variety of challenges must

be addressed for highly graphitic supports. The lack of polar and charged oxygenated

sites on the graphitic surface precludes strong ion-dipole and electrostatic interactions

with ionic metal precursors.26 Consequently, the weakly interacting metal precursors

often lead to undesirably large nanoparticle sizes (>5 nm) upon reduction.26 To

211

enhance precursor interactions, the graphitic carbons may be oxygenated with acids.18

However, these oxygen groups weaken the interactions of the final metal

nanoparticles with graphitic π electrons and make the carbon more susceptible to

oxidation in fuel cells.

6.4.3 Dealloying – Structure and Activity Enhancement

The straining of the Pt shell from the underlying Cu in the dealloyed

nanoparticles has been shown to raise the catalyst activity markedly.1,12 However, the

relationship between strain and catalyst activity can be highly complex for multiphase

binary catalysts. Suppose a mixture of compositions of nanoparticles were dealloyed.

The Pt shells from starting particles with too little copper may not be strained enough

to produce high activity. For particles with too much copper, the final shell may have

too little Pt or the strain be too high. Thus, for multiphase catalysts, it may be difficult

to understand the relationship between activity and strain and to optimize the activity.

The low polydispersities in the composition of the essentially single-phase

nanoparticles in this study are beneficial for probing these phenomena, and

ultimately, for optimizing catalyst activity. In order to control, characterize and

optimize the strain in the Pt shell, it would be beneficial to form bimetallic

nanoparticles with low polydispersity in composition after dealloying. Clearly,

dealloying of single-phase particles would be more likely to produce such particles.

In this section, we discuss the effect of dealloying on particle morphology, and in the

next section we describe the control of the composition prior to dealloying.

The strain upon dealloying will lower the d-band centers relative to the Fermi

level and consequently enhance the ORR activity according to DFT studies.65-67 This

lowering reduces the adsorption energy of oxygen on the Pt-rich surface. The

resulting reduction in the coverage of Pt surface sites by oxygen intermediate species4

enhances the activity.

212

The dealloying of the Pt20Cu80 nanoparticles to Pt85Cu15 results in an almost

pure Pt surface with a copper enriched core (Figure 6.3 and Appendix B, Figure B.5),

as previously hypothesized by Strasser and others.1 A smaller lattice parameter was

observed for the Pt-rich particles after dealloying relative to pure Pt nanoparticles

(Table 6.3). Thus the strain from the copper core reduced Pt-Pt distances on the

particle surface and enhanced the activity. Furthermore, the catalyst activity increases

with increasing strain as reflected by the increase in activity from 0.16 A/ mg Pt to

0.45 A/ mg Pt for essentially single-phased Pt75Cu25 versus Pt20Cu80 nanoparticles,

respectively, prior to dealloying. The lack of any significant increase in ECSA upon

dealloying indicated that the activity increase was not be caused by surface

roughening (see supplementary information for further details) consistent with

previous reports.1

During dealloying at a given potential, the copper will thus be leached at

similar rates from particle to particle when the compositional polydispersity is low.

Hence, the dealloyed particles may be expected to also have low compositional

polydispersity and thus a similar degree of strain in the Pt shell for all nanoparticles,

as was observed. The dealloying process may be controlled more precisely than for

polydisperse nanoparticles with variable leaching rates. With these well-defined and

uniform particles, the relationship between catalyst activity and strain may be more

clearly elucidated. This fundamental understanding is beneficial for targeting the

optimum strain level to maximize catalyst activity.

6.4.4 Composition of Bimetallic Nanocrystals before Dealloying

The ability to achieve low polydispersity in particle composition in the

arrested growth precipitation, as shown by XRD in Figure 6.4, led to the high degree

of control of the composition and morphology of the particles after dealloying. In

contrast, dealloying of multiphase nanoparticles has the potential to produce a range

of different nanoparticle compositions with widely varying catalyst activities.

213

It is difficult to synthesize single-phased bimetallic catalyst with narrow

composition polydispersity and controlled size by the traditional wetness-

impregnation technique.1,12,31,68,69 For sequential impregnation the second metal

precursor is infused and reduced on pre-formed particles of the first metal on the

support. However, the second precursor may nucleate separately on the

heterogeneous support sites, instead of on the first metal resulting in a multi-phased

catalyst,33,34 as was seen for Pt-Cu nanoparticles.1 Three distinct composition phases

of Pt68Cu32, Pt24Cu76 and pure Cu were observed upon reduction of Cu precursor on

supported Pt particles.1 Upon dealloying, a very high activity enhancement of >4-fold

(0.5 A/mg Pt) relative to pure Pt was observed for this multi-phased catalyst. The

relationship between dealloying of the multiple phases and the varying Pt strain

depending upon the initial phase composition can be highly complex. For

simultaneous impregnation, the metal ions may separate chromatographically as they

diffuse into the pores, and produce large compositional polydispersity.31,32,69,70 For

example, PtNi synthesized by simultaneous impregnation produced two different

phases of NiO and Pt94Ni6 for a precursor ratio of Pt(70):Ni(30).71

The formation of essentially single-phase nanoparticles with narrow

compositional polydispersity in this study is favored by the arrested precipitation

technique of synthesis in the presence of stabilizing ligands.39,72,73 Both pure platinum

and copper exhibit FCC crystal structure which is favorable for attempting to prevent

phase separation. The precursor concentrations can be tuned to control nucleation and

growth rates to focus the nanocrystal size to narrow polydispersity, as shown in Table

6.2, facilitated by the dynamic solvation of the nanocrystal surface by the ligands.

The precursors are simultaneously reduced in absence of the carbon supports, which

may otherwise cause heterogeneous nucleation of multiple phases as is often seen in

the wetness impregnation technique.1,12,31,68,69 The composition may be varied by

varying the precursor concentrations to produce essentially single-phase particles (as

shown in Table 6.2 for Pt-Cu). The dealloying of these essentially single-phase

214

particles leads to better control of strain in the Pt shell which is beneficial both for

fundamental understanding and improvement of catalyst activity.

6.5 Conclusions

Herein, we have demonstrated a novel presynthesis/infusion method to design

highly stable and active bimetallic catalysts on graphitized mesoporous carbon. The

unprecedented stability of these highly graphitic ORR catalysts is indicated by a loss

in ECSA of <1% and a loss in catalytic activity of less than 2% after 1000 cycles

between 0.5 and 1.2 V, versus 70% loss for a typical commercial carbon, Vulcan XC-

72. The high resistance of the highly graphitic supports, with very few surface oxygen

sites, to oxidation to CO2 at 1.2 V for 7 hours, prevents nanoparticle isolation from

the surface. The strong covalent and charge transfer interactions between the π

electrons on the graphitic surface and the enriched Pt surface after dealloying mitigate

nanoparticle sintering, Pt dissolution and Ostwald ripening. The strong metal-carbon

interactions are further enhanced by the large interfacial contact area between the

metal nanocrystals and concave mesopores with relatively similar radii. Removal of

90% of the copper from these nanoparticles via dealloying produced essentially

single-phased Pt85Cu15 with large strain in the Pt shell, resulting in a 4-fold activity

enhancement for ORR over pure Pt nanoparticles. This large enhancement

corroborates previous results for Pt-Cu nanoparticles on Vulcan XC-72 carbon.1 The

arrested growth precipitation of Pt-Cu nanoparticles with stabilizing ligands favors

focusing of the size (<3 nm) as well as the composition, each with narrow

polydispersity, indicating essentially single-phase nanoparticles, as shown by XRD

and TEM. The low surface coverage of the weakly bound ligands resulted in strong

Cu-support interactions24 as indicated by high Pt-Cu loadings up to 15 wt %. The

ability to design and control the size and composition of uniform bimetallic

nanocrystals with the pre-synthesis concept and to achieve strong binding to graphitic

mesoporous carbon offers novel opportunities for understanding fundamental

215

catalytic mechanisms. This understanding will be beneficial for designing highly

active and stable catalysts, including electrocatalysts for oxygen reduction in PEM

fuel cells.

6.6 References

(1) Koh, S.; Strasser, P. Electrocatalysis on Bimetallic Surfaces: Modifying

Catalytic Reactivity for Oxygen Reduction by Voltammetric Surface Dealloying.

Journal of the American Chemical Society 2007, 129, 12624-12625.

(2) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.; Lucas, C.

A.; Markovic, N. M. Improved Oxygen Reduction Activity on Pt3Ni(111) via

Increased Surface Site Availability. Science (Washington, DC, United States) 2007,

315, 493-497.

(3) Zhang, J.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R.

Controlling the catalytic activity of platinum-monolayer electrocatalysts for oxygen

reduction with different substrates. Angewandte Chemie, International Edition 2005,

44, 2132-2135.

(4) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.;

Wang, G.; Ross, P. N.; Markovic, N. M. Trends in electrocatalysis on extended and

nanoscale Pt-bimetallic alloy surfaces. Nature Materials 2007, 6, 241-247.

(5) Gasteiger, H. A. Proton exchange fuel cell materials and R&D needs for

future market success. Electrochemistry (Tokyo, Japan) 2007, 75, 103.

(6) Salgado, J. R. C.; Antolini, E.; Gonzalez, E. R. Structure and Activity of

Carbon-Supported Pt-Co Electrocatalysts for Oxygen Reduction. Journal of Physical

Chemistry B 2004, 108, 17767-17774.

(7) Koh, S.; Leisch, J.; Toney, M. F.; Strasser, P. Structure-Activity-Stability

Relationships of Pt-Co Alloy Electrocatalysts in Gas-Diffusion Electrode Layers.


216

(8) Xiong, L.; Manthiram, A. Nanostructured Pt-M/C (M = Fe and Co) catalysts

prepared by a microemulsion method for oxygen reduction in proton exchange

membrane fuel cells. Electrochimica Acta 2005, 50, 2323-2329.

(9) Xiong, L.; Manthiram, A. Effect of atomic ordering on the catalytic activity of

carbon supported PtM (M = Fe, Co, Ni, and Cu) alloys for oxygen reduction in

PEMFCs. Journal of the Electrochemical Society 2005, 152, A697-A703.

(10) Wakabayashi, N.; Takeichi, M.; Uchida, H.; Watanabe, M. Temperature

Dependence of Oxygen Reduction Activity at Pt-Fe, Pt-Co, and Pt-Ni Alloy

Electrodes. Journal of Physical Chemistry B 2005, 109, 5836-5841.

(11) Yano, H.; Kataoka, M.; Yamashita, H.; Uchida, H.; Watanabe, M. Oxygen

Reduction Activity of Carbon-Supported Pt-M (M = V, Ni, Cr, Co, and Fe) Alloys

Prepared by Nanocapsule Method. Langmuir 2007, 23, 6438-6445.

(12) Mani, P.; Srivastava, R.; Strasser, P. Dealloyed Pt-Cu Core-Shell Nanoparticle

Electrocatalysts for Use in PEM Fuel Cell Cathodes. Journal of Physical Chemistry C

2008, 112, 2770-2778.

(13) Koh, S.; Strasser, P. Electrocatalysis on bimetallic surfaces: modifying

catalytic reactivity for oxygen reduction by voltammetric surface dealloying. J Am

Chem Soc FIELD Full Journal Title:Journal of the American Chemical Society

FIELD Publication Date:2007, 129, 12624-12625. FIELD Reference Number:

FIELD Journal Code:7503056 FIELD Call Number:.

(14) Mani, P.; Srivastava, R.; Yu, C.; Strasser, P. In-situ, in-layer de-alloying of Pt-

M intermetallics for high performance PEMFC electrode layers: MEA activity and

durability studies. ECS Transactions 2007, 11, 933-939.

(15) Strasser, P.; Koh, S.; Yu, C. Voltammetric surface dealloying of Pt bimetallic

nanoparticles: a novel synthetic method towards more efficient ORR electrocatalysts.

ECS Transactions 2007, 11, 167-180.

(16) Koh, S.; Yu, C.; Strasser, P. De-alloyed Pt-M nanoparticle electrocatalysts for

efficient electroreduction of oxygen: structural-activity-stability relationship. ECS

Transactions 2007, 11, 205-215.

217

(17) Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Stabilization of Platinum

Oxygen-Reduction Electrocatalysts Using Gold Clusters. Science (Washington, DC,

United States) 2007, 315, 220-222.

(18) Wang, J.; Yin, G.; Shao, Y.; Wang, Z.; Gao, Y. Investigation of Further

Improvement of Platinum Catalyst Durability with Highly Graphitized Carbon

Nanotubes Support. Journal of Physical Chemistry C 2008, 112, 5784-5789.

(19) Shao, Y.; Yin, G.; Gao, Y.; Shi, P. Durability study of Pt/C and Pt/CNTs

catalysts under simulated PEM fuel cell conditions. Journal of the Electrochemical

Society 2006, 153, A1093-A1097.

(20) Ferreira, P. J.; la O, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha,

S.; Gasteiger, H. A. Instability of Pt/C electrocatalysts in proton exchange membrane

fuel cells: A mechanistic investigation. Journal of the Electrochemical Society 2005,

152, A2256-A2271.

(21) Mathias, M. F.; Makharia, R.; Gasteiger, H. A.; Conley, J. J.; Fuller, T. J.;

Gittleman, C. I.; Kocha, S. S.; Miller, D. P.; Mittelsteadt, C. K.; Xie, T.; Yan, S. G.;

Yu, P. T. Two fuel cell cars in every garage? Electrochemical Society Interface 2005,

14, 24-35.

(22) Waje, M. M.; Li, W.; Chen, Z.; Larsen, P.; Yan, Y. Effect of scan range on Pt

surface area loss in potential cycling experiments. ECS Transactions 2007, 11, 1227-

1233.






6943.

(25) Coloma, F.; Sepulveda-Escribano, A.; Rodriquez-Reinoso, F. Heat-treated

carbon blacks as supports for platinum catalysts. Journal of Catalysis 1995, 154, 299-

305.

218

(26) Yu, X.; Ye, S. Recent advances in activity and durability enhancement of Pt/C

catalytic cathode in PEMFC. Journal of Power Sources 2007, 172, 145-154.

(27) Yu, X.; Ye, S. Recent advances in activity and durability enhancement of Pt/C

catalytic cathode in PEMFC. Journal of Power Sources 2007, 172, 133-144.

(28) Shao, Y.; Yin, G.; Gao, Y. Understanding and approaches for the durability

issues of Pt-based catalysts for PEM fuel cell. Journal of Power Sources 2007, 171,

558-566.

(29) Cuong, N. T.; Chi, D. H.; Kim, Y.-T.; Mitani, T. Structural and Electronic

Properties of Ptn (n=3, 7, 13) clusters on metallic single wall carbon nanotube.

Physica Status Solidi 2006, 13, 3472-3475.

(30) Wakayama, H.; Setoyama, N.; Fukushima, Y. Size-Controlled Synthesis and

Catalytic Performance of Pt Nanoparticles in Micro- and Mesoporous Silica Prepared

Using Supercritical Solvents. Adv. Mater. 2003, 15, 742-745.

(31) Raghuveer, V.; Ferreira, P. J.; Manthiram, A. Comparison of Pd-Co-Au

electrocatalysts prepared by conventional borohydride and microemulsion methods

for oxygen reduction in fuel cells. Electrochemistry Communications 2006, 8, 807-

814.

(32) Freund, A.; Lang, J.; Lehmann, T.; Starz, K. A. Improved Pt alloy catalysts

for fuel cells. Catalysis Today 1996, 27, 279-283.

(33) Min, M.-K.; Cho, J.; Cho, K.; Kim, H. Particle size and alloying effects of Pt-

based alloy catalysts for fuel cell applications. Electrochimica Acta 2000, 45, 4211-

4217.

(34) Antolini, E. Formation of carbon-supported PtM alloys for low temperature

fuel cells: a review. Materials Chemistry and Physics 2003, 78, 563-573.

(35) Rioux, R. M.; Song, H.; Hoefelmeyer, J. D.; Yang, P.; Somorjai, G. A. High-

Surface-Area Catalyst Design: Synthesis, Characterization, and Reaction Studies of

Platinum Nanoparticles in Mesoporous SBA-15 Silica. Journal of Physical Chemistry

B 2005, 109, 2192-2202.

219

(36) Gupta, G.; Shah, P. S.; Zhang, X.; Saunders, A. E.; Korgel, B. A.; Johnston,

K. P. Enhanced Infusion of Gold Nanocrystals into Mesoporous Silica with


(37) Gupta, G.; Stowell, C. A.; Patel, M. N.; Gao, X.; Yacaman, M. J.; Korgel, B.

A.; Johnston, K. P. Infusion of Pre-synthesized Iridium Nanocrystals into


6239-6249.

(38) Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.; Johnston, K.

P. Stable Ordered FePt Mesoporous Silica Catalysts with High Loadings Chemistry of

Materials 2008.




(40) Vijayaraghavan, G.; Stevenson, K. J. Synergistic Assembly of Dendrimer-

Templated Platinum Catalysts on Nitrogen-Doped Carbon Nanotube Electrodes for

Oxygen Reduction. Langmuir 2007, 23, 5279-5282.

(41) Ye, H.; Crooks, R. M. Effect of Elemental Composition of PtPd Bimetallic

Nanoparticles Containing an Average of 180 Atoms on the Kinetics of the

Electrochemical Oxygen Reduction Reaction. Journal of the American Chemical

Society 2007, 129, 3627-3633.

(42) Hou, J. G.; Li, X.; Wang, H.; Wang, B. Cu-C60 interaction and nano-

composite structures. Journal of Physics and Chemistry of Solids 2000, 61, 995-998.

(43) Sun, X.; Kang, S.; Harrell, J. W.; Nikles, D. E.; Dai, Z. R.; Li, J.; Wang, Z. L.

Synthesis, chemical ordering, and magnetic properties of FePtCu nanoparticle films.


(44) Liang, C.; Dai, S. Synthesis of Mesoporous Carbon Materials via Enhanced

Hydrogen-Bonding Interaction. Journal of the American Chemical Society 2006, 128,

5316-5317.

(45) Wang, X. Q.; Liang, C.; Dai, S. Langmuir 2008, in press.

220

(46) Pantea, D.; Darmstadt, H.; Kaliaguine, S.; Summchen, L.; Roy, C. Electrical

conductivity of thermal carbon blacks. Influence of surface chemistry. Carbon 2001,

39, 1147-1158.

(47) Cuesta, A.; Dhamelincourt, P.; Laureyns, J.; Martinez-Alonso, A.; Tascon, J.

M. D. Raman microprobe studies on carbon materials. Carbon 1994, 32, 1523-1532.

(48) Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties; John

Wiey and Sons, 1988.

(49) Tian, J. M.; Wang, F. B.; Shan, Z. H. Q.; Wang, R. J.; Zhang, J. Y. Effect of

preparation conditions of Pt/C catalysts on oxygen electrode performance in proton

exchange membrane fuel cells. Journal of Applied Electrochemistry 2004, 34, 461-

467.

(50) Nemanich, R. J.; Solin, S. A. First- and second-order Raman scattering from

finite-size crystals of graphite. Physical Review B: Condensed Matter and Materials

Physics 1979, 20, 392-401.

(51) Maldonado, S.; Morin, S.; Stevenson, K. J. Structure, composition, and

chemical reactivity of carbon nanotubes by selective nitrogen doping. Carbon 2006,

44, 1429-1437.

(52) Tuinstra, F.; Koenig, J. L. Raman spectrum of graphite. Journal of Chemical

Physics 1970, 53, 1126-1130.

(53) Ruban, A. V.; Skriver, H. L.; Norskov, J. K. Surface segregation energies in

transition-metal alloys. Physical Review B: Condensed Matter and Materials Physics

1999, 59, 15990-16000.

(54) Mukerjee, S. Particle size and structural effects in platinum electrocatalysis.

Journal of Applied Electrochemistry 1990, 20, 537-548.

(55) Kinoshita, K. Particle size effects for oxygen reduction on highly dispersed

platinum in acid electrolytes. Journal of the Electrochemical Society 1990, 137, 845-

848.

(56) Vegard, L. Formation of mixed crystals by the contact of solid phases.

Zeitschrift fuer Physik 1921, 5, 393-395.

221

(57) Vegard, L. The constitution of mixed crystals and the space occupied by

atoms. Zeitschrift fuer Physik 1921, 5, 17-26.

(58) Liu, Z.; Koh, S.; Yu, C.; Strasser, P. Synthesis, dealloying, and ORR-

electrocatalysis of PDDA-stabilized Cu-rich Pt alloy nanoparticles. Journal of the

Electrochemical Society 2007, 154, B1192-B1199.

(59) Chen, Z.; Deng, W.; Wang, X.; Yan, Y. Durability and activity study of

single-walled, double-walled and multi-walled carbon nanotubes supported Pt

catalyst for PEMFCs. ECS Transactions 2007, 11, 1289-1299.

(60) Stevens, D. A.; Hicks, M. T.; Haugen, G. M.; Dahn, J. R. Ex Situ and In Situ

Stability Studies of PEMFC Catalysts. Journal of the Electrochemical Society 2005,

152, A2309-A2315.

(61) Jana, N. R.; Peng, X. Single-Phase and Gram-Scale Routes toward Nearly

Monodisperse Au and Other Noble Metal Nanocrystals. Journal of the American


(62) Fu, X.; Wang, Y.; Wu, N.; Gui, L.; Tang, Y. Surface Modification of Small

Platinum Nanoclusters with Alkylamine and Alkylthiol: An XPS Study on the



(63) Antolini, E.; Giorgi, L.; Cardellini, F.; Passalacqua, E. Physical and

morphological characteristics and electrochemical behavior in PEM fuel cells of

PtRu/C catalysts. Journal of Solid State Electrochemistry 2001, 5, 131-140.

(64) Yang, D. Q.; Zhang, G. X.; Sacher, E.; Jose-Yacaman, M.; Elizondo, N.

Evidence of the Interaction of Evaporated Pt Nanoparticles with Variously Treated

Surfaces of Highly Oriented Pyrolytic Graphite. Journal of Physical Chemistry B

2006, 110, 8348-8356.

(65) Xu, Y.; Ruban, A. V.; Mavrikakis, M. Adsorption and dissociation of O2 on

Pt-Co and Pt-Fe alloys. Journal of the American Chemical Society 2004, 126, 4717-

4725.

222

(66) Greeley, J.; Krekelberg, W. P.; Mavrikakis, M. Strain-induced formation of

subsurface species in transition metals. Angewandte Chemie, International Edition

2004, 43, 4296-4300.

(67) Schlapka, A.; Lischka, M.; Gross, A.; Kasberger, U.; Jakob, P. Surface Strain

versus Substrate Interaction in Heteroepitaxial Metal Layers: Pt on Ru(0001).

Physical Review Letters 2003, 91, 016101/016101-016101/016104.

(68) Zhou, B.; Hermans, S.; Somorjai, G. A. Nanotechnology in Catalysis; Kluwer

Academic/ Plenum 2004; Vol. 1.

(69) Fernandez, J. L.; Raghuveer, V.; Manthiram, A.; Bard, A. J. Pd-Ti and Pd-Co-

Au Electrocatalysts as a Replacement for Platinum for Oxygen Reduction in Proton

Exchange Membrane Fuel Cells. Journal of the American Chemical Society 2005,

127, 13100-13101.

(70) Vasiliev, M. A. Surface effects of ordering in binary alloys. Journal of

Physics D: Applied Physics 1997, 30, 3037-3070.

(71) Antolini, E.; Salgado, J. R. C.; dos Santos, A. M.; Gonzalez, E. R. Carbon-

supported Pt-Ni alloys prepared by the borohydride method as electrocatalysts for

DMFCs. Electrochemical and Solid-State Letters 2005, 8, A226-A230.

(72) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. Synthesis of

Thiol-Derivatised Gold Nanoparticles in a Two-Phase Liquid-Liquid System. J.

Chem. Soc., Chem. Commun. 1994, 7, 801-802.

(73) Yin, Y.; Alivisatos, A. P. Colloidal nanocrystal synthesis and the organic-

inorganic interface. Nature (London, United Kingdom) 2005, 437, 664-670.

223

Chapter7

Conclusions and Recommendations

7.1 Conclusions

7.1.1 Infusion of Metal Nanocrystals in Mesoporous Silica using sc-CO2

A novel approach is presented to infuse presynthesized gold nanocrystals into

mesoporous silica by tuning the solvent quality with CO2 as an antisolvent. The

concept of decoupling the nanocrystal synthesis step and the infusion step provides

exquisite control of the nanocrystal size, morphology and dispersibility within the

pores, without perturbing the morphology of the mesoporous silica. High loadings of

the gold and iridium nanocrystals in mesoporous silica (~2 wt %) are achieved by

tuning the solvent strength with CO2, in order to strengthen the van der Waals

interactions of the nanocrystals with the silica pore walls, without flocculating the

nanocrystals in the bulk solvent. The loadings do not show significant dependence on

the pore size of the silica as well as concentration of nanocrystals in the bulk solvent.

The high surface area of the iridium nanocrystals produced high turnover frequency

(TOF) of 16 s-1 for decene hydrogenation and is higher than the TOF for

homogeneous iridium catalysts. Strategic selection of weakly binding ligand

(tetraoctylammonium bromide) for iridium stabilization facilitated catalyst activation

without any thermal calcination. Weakly bound ligands on the nanocrystal surface

may be expected to migrate on to the vacant sites on the silica, significantly reducing

the ligand coverage on the surface and hence increasing the catalyst activity.

Desorption of ligands from the adsorbed nanoparticle surface, onto the support and

into the decene during reaction, will enhance metal binding to the surface and aid

catalyst stability.

224

7.1.2 Synthesis of Stable and Ordered FePt Bimetallic Catalysts

Highly stable and active catalysts composed of bimetallic nanocrystals on

mesoporous supports with a controlled intermetallic alloy structure have been

designed for the first time. Whereas CO2 was used to facilitate gold nanocrystal

adsorption on the support as shown in Chapter 2, it is not needed for FePt coated with

weakly binding ligands, since the interactions of the nanocrystals with the support

were stronger as a consequence of uncovered metal on the particle surface. The

catalysts are formed by infusion of pre-synthesized FePt nanocrystals, with well-

controlled size and composition. The arrested growth precipitation favors focusing of

the size (<4 nm) as well as the composition of the nanocrystals. The stabilizing

ligands control dynamic solvation, supersaturation, nucleation and growth, as

reflected in the small nanocrystal size. The low surface coverage of the weakly bound

ligands resulted in strong metal-support interactions as indicated by high metal

loadings above 10 wt % in only 10 min and stability against sintering upon thermal

annealing at 700 °C. These interactions are further increased by the enhanced

interfacial contact area between the metal nanocrystals and concave mesopores with

similar radii. Finally, the calcined FePt catalyst exhibits 6-fold higher catalyst activity

than commercial Pd-alumina catalyst for liquid 1-decene hydrogenation and was

stable for multiple reactions.

7.1.3 Stable Pt Electrocatalysts Supported on Graphitic Carbon

A novel presynthesis/infusion method to design highly stable Pt

electrocatalysts on ordered graphitized mesoporous carbon with controlled

nanocrystal size and morphology has been demonstrated for the first time. The

graphitic supports with very few surface oxygen sites, offers high resistance to

oxidation to CO2, even at potentials of 1.2 V for 7 hours. Lack of support oxidation

prevents nanoparticle isolation from the surface and consequent activity loss. The

strong metal- support interactions between the π sites on the graphitic surface and the

225

Pt surface mitigate nanoparticle sintering, Pt dissolution and Ostwald ripening, as

indicated by a loss in ECSA of <1% and a loss in catalytic activity of <2% after 1000

cycles between 0.5 and 1.2 V. These metal-carbon interactions are further enhanced

by the large interfacial contact area between the metal nanocrystals and concave

mesopores with relatively similar radii. For amorphous carbons such as Vulcan XC-

72, carbon oxidation and Pt dissolution lead to a loss in ORR activity of 70% upon

potential cycling. The arrested growth precipitation of Pt nanocrystals with 3 nm size

and high concentrations of the active (111) surface leads to 25% higher activities than

typically seen for Pt formed on Vulcan carbon. The low surface coverage of the

weakly bound ligands favors strong Pt-C interactions as indicated by high metal

loadings of 20 wt %.

7.1.4 Highly Stable and Active Bimetallic Pt-Cu for PEMFC

Highly active bimetallic catalysts on graphitized mesoporous carbon with

controlled nanocrystal size, composition and alloy structure and high stability have

been synthesized. Electrochemically leaching of copper from the nanocrystals

generates a strained Pt shell, enhancing the activity for oxygen reduction over 4-fold

relative to Pt catalysts, as observed previously for Vulcan XC-72 carbon. The arrested

growth precipitation of Pt-Cu nanoparticles with stabilizing ligands favors focusing of

the size (<3 nm) as well as the composition of the nanocrystals resulting in a single

crystal phase. The low surface coverage of the weakly bound ligands resulted in

strong Cu-support interactions as indicated by high Pt-Cu loadings up to 15 wt %.

The strong interactions between the π electrons on the graphitic surface and the

enriched Pt surface after dealloying mitigate nanoparticle sintering, Pt dissolution and

Ostwald ripening during potential cycling between 0.5 and 1.2 V, as previously

shown in Chapter 5. The ability to design catalyst morphology and crystal structure

with pre-synthesized bimetallic nanocrystals on graphitic mesoporous carbon offers

226

novel opportunities for highly active and stable catalysts including electrocatalysts for

oxygen reduction in PEMFCs.

7.2 Recommendations

7.2.1 Design of Core-Shell type bimetallic catalysts for ORR

It is shown in Chapter 6 that Pt-Cu nanoparticles after electrochemical

dealloying of copper show activity enhancement of 4x over pure Pt catalyst possibly

due to formation of a core shell structure, with an copper-rich core and a Pt metal

shell.1 However in using the Pt-Cu catalyst as the cathode in the fuel cell stack, a

practical challenge needs to be overcome. The copper ions that are dealloyed will

leach into the proton conducting membrane (like nafion) and may stick to the

negative sites of the membrane. Consequently, the resistance of the membrane for the

transport of the H+ to the cathode significantly increases which limits the performance

of the PEMFC. To overcome this challenge, it would be desirable to design catalysts

with core-shell nanoparticles composed of pure copper core and a pure Pt shell. These

core shell particles may potentially display high catalyst activities due to the strain

generated on the Pt shell by the pure copper core, analogous to the dealloyed Pt-Cu

systems but without any catalyst activation step for copper removal.

The area of core-shell nanoparticles as ORR catalysts is largely unexplored.

Recently, there have been a few advances in the development of core-shell

nanoparticles.2,3 By synthesizing and testing various Cu-Pt core-shell nanoparticles,

relationships between the strain on the Pt surface and the core copper can be

elucidated. In this way, the structure of nanoparticles in relation to their catalytic

activity can be studied in a facile manner.

7.2.2 Design of Active Catalysts for Water-Gas Shift Reaction

227

As per the recommendation of Department of Energy, there is a need for (1)

design and assembly of nanoscale catalysts for energy related reactions, (2)

understanding and control of catalyst activities and selectivities for energy related

reactions, and (3) development of concepts in catalysis to meet major energy research

needs. The key reactions in hydrogen production are the low temperature water gas

shift (WGS) reaction, CO + H2O = CO2 + H2 and preferential oxidation of CO

(PROX) in order to increase H2 conversions.4-6 The CO level must be lowered to

about 1% in the WGS reaction and to 50 ppm in PROX to prevent poisoning of

catalysts in fuel cells. Existing catalysts are not selective and stable enough to meet

these goals. The novel “bottom up design” concept to prepare catalysts that has been

developed for fuel cells as shown in Chapters 4-6 can be used to design WGS

catalysts with much higher performance and lower amounts of expensive noble

metals including Au and Pt. Pre-synthesized metal nanocrystals will be infused into

ordered mesoporous metal oxide supports to achieve extraordinary control of the

morphology of the active sites at the metal-metal oxide interface. The structure and

bonding at the metal-metal oxide interface will be characterized by X-ray absorption

spectroscopy (XAS), electron energy-loss spectroscopy (EELS) and other techniques

to understand the catalytic mechanisms.4,5,7,8 This understanding of structure-activity

relationships for these catalysts with well-defined active sites will guide the discovery

novel generations of highly active, selective and stable catalysts for hydrogen related

reactions.

7.3 References

(1) Koh, S.; Strasser, P. Electrocatalysis on Bimetallic Surfaces:

Modifying Catalytic Reactivity for Oxygen Reduction by Voltammetric Surface

Dealloying. Journal of the American Chemical Society 2007, 129, 12624-12625.

(2) Muraviev, D. N.; Macanas, J.; Parrondo, J.; Munoz, M.; Alonso, A.;

Alegret, S.; Ortueta, M.; Mijangos, F. Cation-exchange membrane as nanoreactor:

228

Intermatrix synthesis of platinum-copper core-shell nanoparticles. Reactive &

Functional Polymers 2007, 67, 1612-1621.







2005, 14, 24-35.








(7) Tibiletti, D.; Fonseca, A. A.; Burch, R.; Chen, Y.; Fisher, J. M.;

Goguet, A.; Hardacre, C.; Hu, P.; Thompsett, D. DFT and in situ EXAFS

investigation of gold/ceria-zirconia low-temperature water gas shift catalysts:

identification of the nature of the active form of gold. J. Phys. Chem. B 2005, 109,

22553-22559. FIELD Reference Number: FIELD Journal Code:101157530 FIELD

Call Number:.

(8) Schwartz, V.; Mullins, D. R.; Yan, W.; Chen, B.; Dai, S.; Overbury, S.

H. XAS Study of Au Supported on TiO2: Influence of Oxidation State and Particle

Size on Catalytic Activity. Journal of Physical Chemistry B 2004, 108, 15782-15790.

229

Appendix A

Infusion of FePt Nanocrystals in Mesoporous Silica

These results relate to FePt nanocrystals infused in mesoporous silica, as

described in Chapter 4.

A.1 Characterization

N2 porosimetry was used to measure the surface area and the pore size

distribution of the mesoporous silica with a high speed surface area BET analyzer

(NOVA 2000, Quantachrome Instruments, Boynton Beach, FL) at a temperature of

77 K. Silica samples were pretreated at 100 ºC for 24 h in vacuum immediately prior

to data collection. Pore size distributions were analyzed by using the adsorption

branch of the isotherm. The Barrett- Joyner- Halenda (BJH) method1 was used for

pore size distribution.

Small angle X-ray scattering (SAXS) was used to determine the long ranged

structure and the center-to-center distance between the pores of synthesized

mesoporous silica. SAXS was performed using a rotating copper anode generator

(Bruker Nonius) operated at 3.0 kW. Scattered photons were collected on a multiwire

gas-filled detector (Molecular Metrology, Inc.). The scattering angle was calibrated

using a silver behenate (CH3(CH2)20COOAg) standard.

Reproduced with permission from Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.; Johnston, K. P.; Chemistry of Materials 2008, 106, 12178-12185, Supplementary Information. Copyright 2008 American Chemical Society.

230

A.2 Results and Discussion

SAXS was used to characterize the mesoporous silica prior to infusion as

shown in Figure A.1. Highly oriented pores, arranged in a 2-D hexagonal lattice, give

rise to a series of Bragg peaks in the scattering pattern. For mesoporous silica, the

three well-resolved peaks are indexable as the (100), (110), (200) reflections

associated with p6mm hexagonal symmetry. Presence of multiple sharp peaks

indicates that calcined mesoporous silica has a high degree of hexagonal mesoscopic

organization. The three peaks for SBA-15 are located at q100 = 0.74, q110 = 1.28, q200 =

1.48 nm-1 with the peak position ratio of 1: 3 :2, which is consistent with the

literature2-4. The distance d100 between adjacent planes is given by d100 = 2π/q100 =

8.5 nm. The unit cell parameter a0 = 3/2 100d = 9.8 nm gives the center-to-center

spacing of the pores.

Representative nitrogen adsorption/desorption isotherms and corresponding

pore size distributions (BJH method1) are shown in Figure A.2. Mesoporous silica has

a mean pore diameter of 5.5 nm, BET surface area of 541 m2/g and a pore volume of

0.62 cm3/g. The pore size distribution is narrow as indicated by the slope and step in

the adsorption isotherms. The wall thickness for the silica is 4.3 nm, which is

determined by subtracting pore diameter from the unit cell parameter as determined

from SAXS.

It is estimated that for 1 gm of mesoporous silica with a loading of 7.3% of

FePt, the surface area of FePt considering there are no ligands on the surface, is just

8.3 m2. Assuming 40% coverage of ligands, the surface area of FePt will be just 5 m2.

A small change in surface area of ~ 5 m2 due to infusion of the nanocrystals will be

difficult to be identified by nitrogen porosimetry.

Figure A.3 shows the HAADF-STEM micrographs of the FePt nanoparticles

infused in mesoporous silica, after calcination at 700 °C in 7%H2/93%N2 atmosphere.

Most of the nanoparticles were well dispersed in the pores and did not show any

significant change in size upon annealing. However, there are small regions in the

231

0.00E+00

5.00E+02

1.00E+03

1.50E+03

2.00E+03

2.50E+03

3.00E+03

3.50E+03

0 0.5 1 1.5 2 2.5 3

2 theta

Inte

nsity

Figure A.1

SAXS on mesoporous silica of type SBA-15. Center-to-center spacing between

cylindrical pores is 8.5 nm.

232

Adsorption Isotherm

0

50

100

150

200

250

300

350

400

450

0.00E+00 2.00E-01 4.00E-01 6.00E-01 8.00E-01 1.00E+00

Relative Pressure (P/P0)

Volu

me

(cc/

g)

Adsorption

Desorption

(a)

Pore size distribution for mesoporous silica

-5.00E-030.00E+005.00E-031.00E-021.50E-022.00E-022.50E-023.00E-023.50E-024.00E-02

-10 10 30 50 70 90 110 130 150

Pore diameter (A)

Ads

orpt

ion

dV(D

) (cc

/A/g

)

(b)

Figure A.2

(a) Adsorption isotherms for SBA-15 measured by nitrogen porosimetry. (b) Pore size

distribution for SBA-15 measured by nitrogen porosimetry.

233

Figure A.3

HAADF-STEM micrographs of 7.3 wt % FePt nanocrystals in mesoporous silica

calcined at 700 oC in 7%H2/93%N2.

Nanoparticles on external surface aggregated at high temperature calcination

Nanoparticles within the pores do not show aggregation

234

nanocomposite, where the nanoparticles grew to 20 nm. These particles were most

likely on the external surface of the silica and not protected by the concave curvature

of the internal pores.

We develop a model where contributions from van der Waals interactions and

charge-induced dipole interactions are used to estimate the total long ranged

interaction potential between the metal nanocrystal and silica. The charge-induced

dipole interaction varies as ~ 1/R4 (where R is the separation distance between the

silica and nanocrystal surface) and decays rapidly with distance.5 Although the charge

on mesoporous silica could not be measured, values from literature for silica particles

dispersed in benzene give a charge density between 20 to 200 charges/μm2,

depending on the size of the particle.6 Assuming the charges are spaced equally apart,

the charge to charge distance is approximately 80 to 250 nm. With a 1/R4 dependence

on distance, it can be assumed that the metal nanocrystal interacts primarily with the

closest point charge at small distances. The interaction energy for charge-induced

dipole is given by5

420

2

dipole induced-charge )4(2q V

Rrεπεα

= (A.1)

where q is the charge, εr is the relative dielectric constant, ε0 is the vacuum

permittivity, and α is the polarizability of the metal and is given by

3

0

.4

arεπεα

= (A.2)

where a is radius of the metal nanocrystal. Hence Vcharge-induced dipole varies as a3.

The van der Waals interaction potential between the ligand bound metal core and

silica surface mediated by toluene is given by7,8

(A.3)

(A.4) 1

ln21

11)( , ++

++=

ij

ij

ijijji x

xxx

xH

( )[ ])()()(121

2,11,1 xHAAxHAV SiltolueneLigSiltolueneMetalSiltolueneLigvdW −−−−−− −+−=

235

(A.5)

where ALig-toluene-Sil and AMetal-toluene-Sil are the mixed Hamaker constants and l is the

ligand length. The Hamaker constant, Axyz, for a medium x and medium z interacting

across solvent y, may be approximated by5,

))(( yyzzyyxxxyz AAAAA −−= (A.6)

where Axx, Ayy and Azz are Hamaker constants for pure components.

The total long interaction potential (VvdW + Vcharge-induced dipole) is plotted as a

function of the separation distance between the metal nanocrystal and silica for

different nanocrystal diameters (Figure A.4). The van der Waals contribution to the

total interaction potential is weakly negative on the order of 10-4 to 10-5 eV for 4.6 to

2.2 nm particle sizes at small separation distances. Hence, the attractive interactions

are caused chiefly by the charge-induced dipole interactions, which are orders of

magnitude stronger than van der Waals. From Figure A.4, 2.2 nm nanocrystals do not

have a high enough attractive potential for silica in the toluene medium. Even at a

separation distance of 2 nm, which is comparable to the ligand length of

dodecanethiol or oleic acid/oleyl amine stabilizing the nanocrystal, the attractive

potential is much smaller than (3/2)kBT (kB is the Boltzmann constant). Hence,

nanocrystal adsorption on silica is very weak and nanocrystals can desorb easily by

thermal energy associated with the 3 degrees of freedom for the nanocrystal. This

result was seen previously with insignificant loadings of less than 0.1 wt.% for 2.2

nm gold nanocrystals in mesoporous silica with a pore diameter of 8.8 nm, as

reported in our earlier publication.9

alR

xa

Rx2

;2 1211

+==

236

-0.25

-0.2

-0.15

-0.1

-0.05

0

0 2 4 6 8 10 12

2.2 nm3.6 nm4.6 nm(3/2)kT

Long

rang

ed In

tera

ctio

n En

ergy

(eV)

Separation Distance (nm)

0 100

5 10-5

1 10-4

1.5 10-4

2 10-4

0 2 4 6 8 10 12

Van

der W

aals

Ene

rgy

(eV)

Figure A.4

Total interaction energy as a function of separation distance and metal nanocrystal

size. Inset shows the van der Waals energy as a function of separation distance for

different nanocrystal sizes.

237

As the size of the nanocrystal increases from 2.2 nm to 3.6 nm for FePt nanocrystals

in this study, the magnitude of the attractive charge-induced dipole interactions

increases significantly, and the net interaction potential becomes larger than (3/2)kBT.

At a separation distance of 2.7 nm, a distance greater than the ligand length, the total

interaction potential is high enough for the FePt nanocrystal to adsorb strongly with

the silica surface and cannot be easily desorbed by Brownian motion.

Figure A.5 shows the low resolution bright field TEM image for FePt in

mesoporous silica. The magnification is low and hence it is difficult to visualize the

FePt nanocomposite morphology. HAADF-STEM on the other hand provides much

better contrast and magnification for visualizing the FePt nanocrystals supported in

mesoporous silica. Figure A.6 shows the calibration curve for the FePt nanocrystals.

Table A.1 shows the loadings as a function of initial FePt concentration as

well as time of infusion. For a given concentration, there is no significant change in

loadings observed when time is increased from 10 min to 960 min. At low

concentration of 0.24 mg/ml, all the nanocrystals are adsorbed in the silica in the first

10 min, resulting in a clear supernatant. For higher concentration of 0.7 mg/ml, the

supernatant is not completely clear. Loadings increase slightly from 9 wt % to 11 wt

% when the time of infusion is increased from 10 min to 960 min. At 2 mg/ml, there

is no change in loadings after 16 h. Thus it can be concluded that the equilibrium

loadings of FePt in mesoporous silica are obtained in the order of minutes.

238

Figure A.5

Low resolution TEM of 7.3 wt % FePt in mesoporous silica.

239

Calibration at 500 nm

y = 7.6529xR2 = 1

0

0.5

1

1.5

2

2.5

3

3.5

0 0.1 0.2 0.3 0.4 0.5

Conc (mg/ml)

Abs

orba

nce

Figure A.6

Calibration of UV-Vis absorbance of FePt.

240

Wt % loading of FePt-Oleic acid/Oleylamine

Initial Nanocrystal Concentration (mg/mL)

Time (min) 0.24 0.70 2.0

10 7.3 ± 0.2 8.9 ± 0.2 13.8 ± 1.3

30 ― 9.0 ± 0.4 ―

960 ― 11.4 ± 0.4 14.3 ± 0.8

Table A.1

Loading of FePt nanocrystals in mesoporous silica as a function of time and

concentration.

241

A.3 References

(1) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. The determination of pore

volume and area distributions in porous substances. I. Computations from nitrogen

isotherms. Journal of the American Chemical Society 1951, 73, 373-380.

(2) Lyu, Y.-Y.; Yi, S. H.; Shon, J. K.; Chang, S.; Pu, L. S.; Lee, S.-Y.;

Yie, J. E.; Char, K.; Stucky, G. D.; Kim, J. M. Highly Stable Mesoporous Metal

Oxides Using Nano-Propping Hybrid Gemini Surfactants. Journal of the American


(3) Ehrburger-Dolle, F.; Morfin, I.; Geissler, E.; Bley, F.; Livet, F.; Vix-

Guterl, C.; Saadallah, S.; Parmentier, J.; Reda, M.; Patarin, J.; Iliescu, M.;

Werckmann, J. Small-Angle X-ray Scattering and Electron Microscopy Investigation

of Silica and Carbon Replicas with Ordered Porosity. Langmuir 2003, 19, 4303-4308.

(4) Kresge, C. T.; Leonowickz, M. E.; Roth, W. J.; Beck, J. S. Nature

1992, 359, 710.



(6) Papirer, E. Adsorption on Silica Surfaces CRC, 2000; Vol. 90.

(7) Nir, S. Van der Waals interactions between surfaces of biological

interest. Progress in Surface Science 1977, 8, 1-58.

(8) Meli, L.; Li, Y.; Lim, K. T.; Johnston, K. P.; Green, P. F. Phase

Inversion in Diblock Copolymer Nanocomposite Films. In Preparation 2007.




(10) Sun, S.; Fullerton, E. E.; Weller, D.; Murray, C. B. Compositionally

controlled FePt nanoparticle materials. IEEE Transactions on Magnetics 2001, 37,

1239-1243.

242

(11) Sun, S.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Monodisperse

FePt nanoparticles and ferromagnetic FePt nanocrystal superlattices. Science

(Washington, D. C.) 2000, 287, 1989-1992.

243

Appendix B

Highly Stable and Active Pt-Cu Electrocatalysts for ORR

The contents of this Appendix are supplementary to the data shown in Chapter 6 on

the synthesis of bimetallic Pt-Cu electrocatalysts.

B.1 Characterization

N2 sorption analysis was performed on a Micromeritics Germini analyzer at -

196 °C (77 K). The specific surface area was calculated using the BET method from

the nitrogen adsorption data in the relative range (P/P0) of 0.06-0.30. The samples

were heated overnight at 100°C for degassing. The total pore volume was determined

from the amount of N2 uptake at P/P0 = 0.95. The pore size distribution was derived

from the adsorption branch of the isotherm based on the BJH model.

Low-resolution pictures of mesoporous silica, infused with gold nanocrystals,

were visualized by TEM on a Phillips EM280 microscope with a 4.5 Å point-to-point

resolution operated with a 80 kV operating voltage. High resolution transmission

electron microscopy (HRTEM) was performed using a JEOL 2010F TEM operating

at 200 kV. Pictures were obtained at the optimum defocus condition.

Nanocomposites were deposited from a dilute ethanol solution onto 200 mesh carbon-

coated copper TEM grids.

XPS was acquired using a Kratos AXIS Ultra DLD spectrometer equipped

with a monochromatic Al X-ray source (Al Kα, 1.4866 keV) and 20 eV path energy

for high resolution elemental scans and 160 eV for survey scans conducted at angle of

45°. The sample is prepared by depositing the catalysts (1-2 mg) on the double sided

copper tape and then entered in the vacuum chamber.

244

TGA was conducted using Perkin–Elmer TGA7 equipment under oxygen

flow. Samples were heated to 750 °C at the rate of 20 °C/min. Samples were held at

50 °C for 10 min before increasing the temperature.

B.2 Results and Discussion

Figure B.1 gives the adsorption-desorption isotherms for the three DMCs and

corresponding pore size distributions. For all the carbons, three well-distinguished

regions of the adsorption isotherm are evident: (i) monolayer- multilayer adsorption,

(ii) capillary condensation, and (iii) multilayer adsorption on the outer particle

surfaces. The adsorption isotherms show a type H1 behavior with capillary

condensation at P/Po ~ 0.6 for DMC-850 and DMC-2000, while DMC-2600 likely

shows a H2 hysteresis behavior. The slope and almost vertical step in the hysteresis

loops suggest well-defined mesopores with a narrow pore size distribution as shown

in Figure B.1b.

Figure B.2 gives the low resolution TEM micrographs of unsupported

Pt20Cu80 nanoparticles. The nanoparticles are relatively monodisperse with a average

diameter of 2.85±0.38 nm as seen in the histogram. The high resolution images show

lattice fringes indicating that the crystalline nature of nanoparticles. The lattice d-

spacing is 2.25Å and can be indexed for a Pt (111) phase.

Figure B.3(a) gives the absorbance of the Pt20Cu80 nanoparticles before and

after infusion after dilution by a common factor. After infusion, the absorbance of the

nanoparticle effluent solution is 23% lower than that before infusion, as the

nanoparticles leave the solution and adsorb onto the mesoporous carbon (DMC-

850).The initial concentration of nanoparticles was 20 mg/mL. The change in

concentration after infusion was determined by using a calibration curve generated by

measuring the absorbances at 800 nm for several dilutions of Pt20Cu80 nanoparticles at

245

0.0 0.2 0.4 0.6 0.8 1.00

50

100

150

200

250

300

350

400

450

500

550

Am

ount

Ads

orbe

d (c

m3

STP

/g)

P/P0

DMC 850 DMC 2000 DMC 2600

0 2 4 6 8 10 12 14 16 18 200

2

PSD

Pore Diameter (nm)

(a) (b)

Figure B.1

Nitrogen adsorption-desorption curves (a) and corresponding pore size distributions

(b) for the disordered mesoporous carbons.

246

(a) (b)

(c)

Figure B.2

TEM micrographs (A) Low resolution (B) High-resolution and (C) Histogram, of as-

synthesized Pt20Cu80 nanoparticles. The average nanoparticle size is 2.85±0.38 nm.

0

5

10

15

20

1.75 2 2.25 2.5 2.75 3 3.25 3.5 3.75

Cou

nt

PtCu Nanoparticle size

247

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

500 550 600 650 700 750 800

Wavelength [nm]

Abs

orba

nce

Before Infusion

After Infusion

(a)

0

0.5

1

1.5

2

2.5

0 0.2 0.4 0.6 0.8 1 1.2Concentration (mg/ml)

Abs

orba

nce

(b)

Figure B.3

(a) Effect of infusion of the Pt20Cu80 nanoparticles in DMC-850 on absorbance as

determined by UV-Visible spectrophotometry (b) Calibration curve of the UV-Visual

absorbance vs. concentration of Pt20Cu80 nanoparticles in toluene, measured at 800

nm.

248

known concentrations and was found to be 15.5 mg/mL. The absorbance-

concentration calibration curve is linear (Figure B.3 (b)) with a correlation coefficient

of 0.998.

Figure B.4 shows the dealloying curves for Pt20Cu80. Cycle 500 in Figure B.4

approaches pure Pt as shown in Figure B.5.

Figure B.6 shows the XPS spectra of the pure Pt nanoparticles of 3 nm size

supported on graphitized carbon. The binding energy for the Pt 4f7/2 peak is 72.1 eV

as seen in Figure B.6. The binding energy is shifted to higher level by 1.1 eV relative

to bulk unsupported Pt (71.0 eV) and by 0.8 eV relative to Pt supported on

commercial amorphous carbon (~71.4 eV. As the nanoparticle size for the Pt is ~ 3

nm, the shift in the binding energy to higher value can not be attributed to the particle

size effects, which are observed for 1-2 nm nanoparticles but become insignificant for

>2 nm nanoparticles. The binding energy shift to higher levels can hence be attributed

to the enhanced Pt-carbon interactions facilitated by the π sites charge transfer to the

Pt metal. It can hence be deduced that the metal-support interactions for Pt on

graphitic supports are stronger than Pt on amorphous carbons and play a key role in

enhancing the catalyst stability towards ADT potential cycling (Figure 6.7).

Figure B.7 shows the TGA curves for 20% Pt on DMC-850 and DMC-2000

mesoporous carbons. The onset of carbon oxidation for DMC-850 is gradual and

occurs at relatively low temperature of ~ 300 °C for DMC 850, and is comparable to

that of 20% Pt on vulcan–XC72 which also starts oxidizing at ~ 300 °C.1 DMC-2000

on the other hand does not show any carbon degradation for DMC 2000 until 600 °C.

Thus graphitic DMC-2000 is significantly more stable than DMC-850 as well as

Vulcan carbon towards carbon oxidation by Pt. High stability of the carbon enables

complete retention in activity upon ADT potential cycling for the dealloyed

Pt20Cu80 on DMC-2000 catalysts (as seen in Figure 6.7).

Figure B.8 shows the cyclic voltammograms of the dealloyed Pt80Cu20 on

DMC-2000 before and after ADT cycling. The cyclic voltammograms show two

249

Figure B.4

First 20 cyclic voltammetry profiles of Pt20Cu80 nanoparticles supported on DMC-

2000 during electrochemical dealloying at 20 mV/s in argon atmosphere. The black

arrows indicate the direction in which the peak shifts as cycling progresses.

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0.000 0.250 0.500 0.750 1.000 1.250E (V) vs SHE

I (m

A/c

m2

geo)

Cycle 5Cycle 3Cycle 2Cycle 1 Cycle 6 Cycle 10

Cycle 20

250

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 0.2 0.4 0.6 0.8 1 1.2

E (vs NHE)

I (A

)

Figure B.5

Cyclic voltammogram of 20% monometallic Pt on DMC-2000 scanned at 100 mV/s

in argon atmosphere.

251

Figure B.6

XPS of 20% Pt supported on graphitized mesoporous carbon.

252

Figure B.7

TGA of Pt nanoparticles supported on disordered mesoporous carbons

Figure B.7

TGA of Pt nanoparticles supported on disordered mesoporous carbons.

0

20

40

60

80

100

120

0 200 400 600

Temperature (C)

Wei

ght % DMC 2000

DMC 850

Onset for DMC-2000

Onset for DMC-850

253

-1.2

-0.7

-0.2

0.3

0.8

1.3

0 0.25 0.5 0.75 1 1.25

E (vs NHE)

I (A)

Before ADT cycling After ADT cycling

Figure B.8

Cycling voltammograms for Pt20Cu80 on DMC 2000 before and after ADT potential

cycling

254

peaks in the underpotentially deposited hydrogen region between 0.05 and 0.3V

associated with the weak and strong desorption of hydrogen. The multiple crystal

facets present on the crystalline dealloyed Pt80Cu20 nanoparticles bind with hydrogen

with varying strength and hence lead to two desorption peaks. After ADT cycling

between potentials of 0.5V and 1.2V for 1000 cycles, the area under the hydrogen

desorption peak remains unchanged, indicating no loss in platinum area due to

sintering, coalescence, Pt dissolution or Ostwald ripening.

The potential range between 0.75- 0.85V where water activation leads to

surface Pt-OH formation according to Pt + H2O = Pt-OH + H+ + e- is shifted to higher

potentials of ~0.85V for dealloyed Pt80Cu20 as compared to that of pure Pt (0.8V,

Figure B.5). Anodic shift of the peak potentials for the dealloyed catalysts compared

to pure Pt suggested a delayed formation of Pt-OH surface species during anodic

sweeps. This upward shift of the Pt-OH onset potential for Pt alloy electrocatalysts is

either considered the cause for enhanced ORR activity2,3 or perceived more of an

effect of a reduced Pt-O chemisorption energy.4

The change in composition from Pt20Cu80 to ~Pt20Cu4 (≡ Pt85Cu15) during

dealloying may be compared with the shrinkage in nanoparticle size, by examining

the relevant lattice parameters. Assuming close packing of atoms, the ratio of final

volume (Vf) to initial volume (Vi), Vf/Vi= (NPt,f.aPt + NCu,f.aCu) /(NPt,i.aPt + NCu,i.aCu =

0.28) where the atomic volumes aPt = 1.51 cc/mole and aCu = 1.18 cc/mole and the

numbers of atoms NPt,i , NCu,i , NPt,f, NCu,f are 20, 80, 20, and 4, respectively.

Consequently, the particle diameter ration df/di = (Vf/Vi)1/3 = (1.245/0.349)1/3 = 0.66

which corresponds to a shrinkage of 34%. However, from XRD and TEM, the

shrinkage in the diameter is 10-12%. This discrepancy can be addressed by including

the effect of lattice restructuring and expansion during copper removal. The lattice

parameters were 3.610 Å for bulk Cu, 3.668 Å for the original Pt-Cu nanoparticles,

3.876 Å for the dealloyed particles and 3.923 Å for bulk Pt (Table 3). The net

expansion in the unit cell lattice is thus estimated to be about 8%, ((3.876/3.668)-

1)*100. Thus, the net estimated reduction of the nanoparticle size was 34% - 8% =

255

26%, whereas the experimentally observed reduction was 12%. The difference

reflects the experimental uncertainty in size measurement by TEM and lattice

parameters by XRD.

256

Figure B.9

Peak analysis of the XRD spectra of the as-synthesized Pt75Cu25 nanoparticles

257

B.3 References

(1) Baturina, O. A.; Aubuchon, S. R.; Wynne, K. J. Thermal Stability in

Air of Pt/C Catalysts and PEM Fuel Cell Catalyst Layers. Chemistry of Materials

2006, 18, 1498-1504.

(2) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.;

Lucas, C. A.; Markovic, N. M. Improved Oxygen Reduction Activity on Pt3Ni(111)

via Increased Surface Site Availability. Science (Washington, DC, United States)

2007, 315, 493-497.




247.

(4) Norskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J.

R.; Bligaard, T.; Jonsson, H. Origin of the Overpotential for Oxygen Reduction at a

Fuel-Cell Cathode. Journal of Physical Chemistry B 2004, 108, 17886-17892.

258

Bibliography Ajayan, P. M.; Ebbesen, T. W. Reports on Progress in Physics 1997, 60, 1025-1062.

Ajayan, P. M.; Iijima, S. Nature (London, United Kingdom) 1993, 361, 333-4.

Akita, T.; Okumura, M.; Tanaka, K.; Kohyama, M.; Haruta, M. Catalysis Today

2006, 117, 62-68.

Alden, L. R.; Roychowdhury, C.; Matsumoto, F.; Han, D. K.; Zeldovich, V. B.; Abruna, H. D.; DiSalvo, F. J. Langmuir 2006, 22, 10465-10471.

Alivisatos, A. P. Science 1996, 271, 933-937.

Allum, K. G.; Hancock, R. D.; Howell, I. V.; Lester, T. E.; McKenzie, S.; Pitkethly,

R. C.; Robinson, P. J. Journal of Organometallic Chemistry 1976, 107, 393-

405.

Allum, K. G.; Hancock, R. D.; Howell, I. V.; Lester, T. E.; McKenzie, S.; Pitkethly,

R. C.; Robinson, P. J. Journal of Catalysis 1976, 43, 331-8.

Antolini, E. Materials Chemistry and Physics 2003, 78, 563-573.

Antolini, E. Journal of Materials Science 2003, 38, 2995-3005.

Antolini, E.; Giorgi, L.; Cardellini, F.; Passalacqua, E. Journal of Solid State


Antolini, E.; Salgado, J. R. C.; dos Santos, A. M.; Gonzalez, E. R. Electrochemical

and Solid-State Letters 2005, 8, A226-A230.

Antolini, E.; Salgado, J. R. C.; Giz, M. J.; Gonzalez, E. R. International Journal of

Hydrogen Energy 2005, 30, 1213-1220.

Antolini, E.; Salgado, J. R. C.; Gonzalez, E. R. Journal of Power Sources 2006, 160,

957-968.

Araki, H.; Fukuoka, A.; Sakamoto, Y.; Inagaki, S.; Sugimoto, N.; Fukushima, Y.;

Ichikawa, M. Journal of Molecular Catalysis A: Chemical 2003, 199, 95-102.

Argo, A. M.; Gates, B. C. Abstracts of Papers, 221st ACS National Meeting, San

Diego, CA, United States, April 1-5, 2001 2001, CATL-053.

259

Argo, A. M.; Odzak, J. F.; Lai, F. S.; Gates, B. C. Nature (London, United Kingdom)

2002, 415, 623-626.

Barrett, E. P.; Joyner, L. G.; Halenda, P. P. Journal of the American Chemical Society

1951, 73, 373-80.

Bartholin, M.; Graillat, C.; Guyot, A. Journal of Molecular Catalysis 1981, 10, 377-

84.

Batista, J.; Pintar, A.; Gomilsek, J. P.; Kodre, A.; Bornette, F. Applied Catalysis, A:

General 2001, 217, 55-68.

Baturina, O. A.; Aubuchon, S. R.; Wynne, K. J. Chemistry of Materials 2006, 18,

1498-1504.

Beard, B. C.; Ross, P. N., Jr. Journal of the Electrochemical Society 1990, 137, 3368-

74.

Beckman, E. J. Journal of Supercritical Fluids 2004, 28, 121-191.

Bell, A. T. Science (Washington, DC, United States) 2003, 299, 1688-1691.

Bell, P. W.; Anand, M.; Fan, X.; Enick, R. M.; Roberts, C. B. Langmuir 2005, 21,

11608-11613.

Bergeron, B. V.; Kelly, C. A.; Meyer, G. J. Langmuir 2003, 19, 8389-8394.

Blackburn, J. M.; Long, D. P.; Cabanas, A.; Watkins, J. J. Science FIELD Publication

Date:2001 Oct 5, 294, 141-5. FIELD Reference Number: FIELD Journal

Code:0404511 FIELD Call Number:.

Bockstaller, M. R.; Thomas, E. L. Journal of Physical Chemistry B 2003, 107, 10017-

10024.

Bond, G. C.; Rank, J. S. Proc. Intern. Congr. Catalysis, 3rd, Amsterdam, 1964 1965,

2, 1225-36, discussion 1236-7.

Bond, G. C.; Wells, P. B. Journal of Catalysis 1966, 5, 419-27.

Borodko, Y.; Ager, J. W., III; Marti, G. E.; Song, H.; Niesz, K.; Somorjai, G. A.


Brehm, A.; Busch, V.; Wilken, J. Chemie Ingenieur Technik 1993, 65, 573-5.

260

Brunauer, S.; Emmett, P. H.; Teller, E. Journal of the American Chemical Society

1938, 60, 309-19.

Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc.,

Chem. Commun. 1994, 7, 801-802.

Burns, R. C.; Graham, C.; Weller, A. R. M. Mol. Phys. 1986, 59, 41.

Cabanas, A.; Long, D. P.; Watkins, J. J. Chemistry of Materials 2004, 16, 2028-2033.

Cable, R. E.; Schaak, R. E. Chemistry of Materials 2005, 17, 6835-6841.

Campbell, C. T.; Parker, S. C.; Starr, D. E. Science (Washington, DC, United States)

2002, 298, 811-814.

Carlock, J. T. 1979, 4 pp.

Casado-Rivera, E.; Volpe, D. J.; Alden, L.; Lind, C.; Downie, C.; Vazquez-Alvarez,

T.; Angelo, A. C. D.; DiSalvo, F. J.; Abruna, H. D. Journal of the American


Casula, M. F.; Corrias, A.; Falqui, A.; Serin, V.; Gatteschi, D.; Sangregorio, C.; de

Fernandez, C.; Battaglin, G. Chemistry of Materials 2003, 15, 2201-2207.

Chan, K.-Y.; Ding, J.; Ren, J.; Cheng, S.; Tsang, K. Y. Journal of Materials

Chemistry 2004, 14, 505-516.

Chao, M.-C.; Lin, H.-P.; Mou, C.-Y.; Cheng, B.-W.; Cheng, C.-F. Catalysis Today

2004, 97, 81-87.

Chen, J. Y.; Pan, F. M.; Chang, L.; Cho, A. T.; Chao, K. J. Journal of Vacuum

Science & Technology, B: Microelectronics and Nanometer Structures--

Processing, Measurement, and Phenomena 2005, 23, 2034-2040.

Chen, M.; Kim, J.; Liu, J. P.; Fan, H.; Sun, S. Journal of the American Chemical

Society 2006, 128, 7132-7133.

Chen, M.; Liu, J. P.; Sun, S. Journal of the American Chemical Society 2004, 126,

8394-8395.

Chen, M. S.; Goodman, D. W. Science (Washington, DC, United States) 2004, 306,

252-255.

261

Chen, W.; Cai, W.; Zhang, L.; Wang, G.; Zhang, L. J. Colloid and Interface Sci.

2001, 238, 291-295.

Chen, Y.; Yang, F.; Dai, Y.; Wang, W.; Chen, S. Journal of Physical Chemistry C

2008, 112, 1645-1649.

Chen, Z.; Deng, W.; Wang, X.; Yan, Y. ECS Transactions 2007, 11, 1289-1299.

Cho, A. Science (Washington, DC, United States) 2003, 299, 1684-1685.

Choi, W. C.; Woo, S. I.; Jeon, M. K.; Sohn, J. M.; Kim, M. R.; Jeon, H. J. Advanced

Materials (Weinheim, Germany) 2005, 17, 446-451.

Chou, N. H.; Schaak, R. E. Journal of the American Chemical Society 2007, 129,

7339-7345.

Chu, H. P.; Lei, L.; Hu, X.; Yue, P.-L. Energy & Fuels 1998, 12, 1108-1113.

Chytil, S.; Glomm, W. R.; Vollebekk, E.; Bergem, H.; Walmsley, J.; Sjoeblom, J.;

Blekkan, E. A. Microporous and Mesoporous Materials 2005, 86, 198-206.

Claus, P.; Brueckner, A.; Mohr, C.; Hofmeister, H. Journal of the American Chemical

Society 2000, 122, 11430-11439.

Coleman, N. R. B.; Morris, M. A.; Spalding, T. R.; Holmes, J. D. J. Am. Chem. Soc.

2001, 123, 187-188.

Coleman, N. R. B.; O'Sullivan, N.; Ryan, K. M.; Crowley, T. A.; Morris, M. A.;

Spalding, T. R.; Steytler, D. C.; Holmes, J. D. J. Am. Chem. Soc. 2001, 123,

7010.

Coleman, N. R. B.; Ryan, K. M.; Spalding, T. R.; Holmes, J. D.; Morris, M. A. Chem.

Phys. Lett. 2001, 343, 1.

Coloma, F.; Sepulveda-Escribano, A.; Fierro, J. L. G.; Rodriguez-Reinoso, F. Applied

Catalysis, A: General 1996, 148, 63-80.

Coloma, F.; Sepulveda-Escribano, A.; Rodriquez-Reinoso, F. Journal of Catalysis

1995, 154, 299-305.

Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354-357.

Cuesta, A.; Dhamelincourt, P.; Laureyns, J.; Martinez-Alonso, A.; Tascon, J. M. D.

Carbon 1994, 32, 1523-32.

262

Cuong, N. T.; Chi, D. H.; Kim, Y.-T.; Mitani, T. Physica Status Solidi 2006, 13,

3472-3475.

Dabbousi, B. O.; Bawendi, M. G.; Onitsuka, O.; Rubner, M. F. Appl. Phys. Lett.

1995, 68, 1316-1318.

Darmstadt, H.; Roy, C.; Kaliaguine, S.; Choi, S. J.; Ryoo, R. Carbon 2002, 40, 2673-

2683.

de Soler-Illia, G. J.; Sanchez, C.; Lebeau, B.; Patarin, J. Chemical Reviews

(Washington, DC, United States) 2002, 102, 4093-4138.

Debenedetti, P. G.; Tom, J. W.; Kwauk, X.; Yeo, S. D. Fluid Phase Equilibria 1993,

82, 311-21.

DeSimone, J. M.; Maury, E. E.; Menceloglu, Y. Z.; McClain, J. B.; Romack, T. J.;

Combes, J. R. Science (Washington, DC, United States) 1994, 265, 356-9.

Diaz, E.; Ordonez, S.; Vega, A.; Coca, J. Microporous and Mesoporous Materials

2005, 77, 245-255.

Dickson, J. L.; Gupta, G.; Horozov, T. S.; Binks, B. P.; Johnston, K. P. Langmuir

2006, 22, 2161-2170.

Ding, J.; Chan, K.-Y.; Ren, J.; Xiao, F.-s. Electrochimica Acta 2005, 50, 3131-3141.

Dixon, D. J.; Bodmeier, R. A.; Johnston, K. P. AIChE J. 1993, 39, 127.

Dixon, D. J.; Johnston, K. P. AIChE Journal 1991, 37, 1441-1449.

Dresselhaus, M. S.; Thomas, I. L. Nature (London, United Kingdom) 2001, 414, 332-

337.

Dumas, J. M.; Rmili, S.; Barbier, J. Journal de Chimie Physique et de Physico-

Chimie Biologique 1998, 95, 1650-1665.

Eckert, C. A.; Liotta, C. L.; Bush, D.; Brown, J. S.; Hallett, J. P. Journal of Physical

Chemistry B 2004, 108, 18108-18118.

Ehrburger-Dolle, F.; Morfin, I.; Geissler, E.; Bley, F.; Livet, F.; Vix-Guterl, C.;

Saadallah, S.; Parmentier, J.; Reda, M.; Patarin, J.; Iliescu, M.; Werckmann, J.

Langmuir 2003, 19, 4303-4308.

Ertl, G.; Knozinger, H.; Weitkamp, J.; Editors Preparation of Solid Catalysts, 1999.

263

Fan, H.; Yang, K.; Boye Daniel, M.; Sigmon, T.; Malloy Kevin, J.; Xu, H.; Lopez

Gabriel, P.; Brinker, C. J. Science FIELD Full Journal Title:Science (New

York, N.Y.) FIELD Publication Date:2004 2004, 304, 567-71. FIELD

Reference Number: FIELD Journal Code:0404511 FIELD Call Number:.

Fan, J.; Boettcher, S. W.; Stucky, G. D. Chemistry of Materials 2006, 18, 6391-6396.

Farias, M. H.; Gellman, A. J.; Somorjai, G. A.; Chianelli, R. R.; Liang, K. S. Surface

Science 1984, 140, 181-96.

Fernandez, J. L.; Raghuveer, V.; Manthiram, A.; Bard, A. J. Journal of the American


Ferreira, P. J.; la O, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.;

Gasteiger, H. A. Journal of the Electrochemical Society 2005, 152, A2256-

A2271.

Ferrer, D.; Blom, D. A.; Allard, L. F.; Mejia, S.; Perej-Tijerina, E.; Yacaman, M. J.

Journal of Materials Chemistry 2008, 18, 2442.

Ferrer, D.; Torres-Castro, A.; Gao, X.; Sepulveda-Guzman, S.; Ortiz-Mendez, U.;

Jose-Yacaman, M. Nano Letters 2007, 7, 1701-1705.

Folch, B.; Larionova, J.; Guari, Y.; Datas, L.; Guerin, C. Journal of Materials

Chemistry 2006, 16, 4435-4442.

Ford, D. M.; Glandt, E. D. J. Mem. Sci. 1995, 107, 47-57.

Freund, A.; Lang, J.; Lehmann, T.; Starz, K. A. Catalysis Today 1996, 27, 279-83.

Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science (Washington, DC, United

States) 2003, 301, 935-938.

Fu, X.; Wang, Y.; Wu, N.; Gui, L.; Tang, Y. Journal of Colloid and Interface Science

2001, 243, 326-330.

Fujdala, K. L.; Tilley, T. D. Journal of Catalysis 2003, 216, 265-275.

Fukuoka, A.; Araki, H.; Kimura, J.-i.; Sakamoto, Y.; Higuchi, T.; Sugimoto, N.;

Inagaki, S.; Ichikawa, M. Journal of Materials Chemistry 2004, 14, 752-756.

Fukuoka, A.; Araki, H.; Sakamoto, Y.; Sugimoto, N.; Tsukada, H.; Kumai, Y.;

Akimoto, Y.; Ichikawa, M. Nano Letters 2002, 2, 793-795.

264

Fukushima, Y.; Wakayama, H. J. Phys. Chem. B 1999, 103, 3062-3064.

Gallagher, P. M., Coffey, M.P., Krukonis, V.J. and Klasuts, N. 1989, 406, 334-354.

Gasteiger, H. A. Electrochemistry (Tokyo, Japan) 2007, 75, 103.

Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Applied Catalysis, B:

Environmental 2005, 56, 9-35.

Ghosh, A.; Ranjan Patra, C.; Mukherjee, P.; Sastry, M.; Kumar, R. Microporous and

Mesoporous Materials 2003, 58, 201-211.

Gierszal, K. P.; Kim, T.-W.; Ryoo, R.; Jaroniec, M. Journal of Physical Chemistry B

2005, 109, 23263-23268.

Glatter, O.; Kratky, O. Small Angle X-ray Scattering, Academic Press Inc: New-

York, 1982.

Gonzalez, O.; Lujano, J.; Pietri, E.; Goldwasser, M. R. Catalysis Today 2005, 107-

108, 436-443.

Goyal, H. B.; Gupta, A. K. Indian Journal of Technology 1989, 27, 360-2.

Greeley, J.; Krekelberg, W. P.; Mavrikakis, M. Angewandte Chemie, International

Edition 2004, 43, 4296-4300.

Grobert, N.; Mayne, M.; Walton, D. R. M.; Kroto, H. W.; Terrones, M.;

Kamalakaran, R.; Seeger, T.; Ruhle, M.; Terrones, H.; Sloan, J.; Dunin-

Borkowski, R. E.; Hutchison, J. L. Chemical Communications (Cambridge,

United Kingdom) 2001, 471-472.

Gross, A. Topics in Catalysis 2006, 37, 29-39.

Guinier, A.; Fournet, G. Small-Angle Scattering of X-Rays, Wiley: New-York, 1955.

Gupta, G.; Patel, M. N.; Ferrer, D.; Heitsch, A. T.; Korgel, B. A.; Johnston, K. P.


Gupta, G.; Shah, P. S.; Zhang, X.; Saunders, A. E.; Korgel, B. A.; Johnston, K. P.


Gupta, G.; Stowell, C. A.; Patel, M. N.; Gao, X.; Yacaman, M. J.; Korgel, B. A.;

Johnston, K. P. Chemistry of Materials 2006, 18, 6239-6249.

265

Ha, S.; Larsen, R.; Masel, R. I. Journal of Power Sources 2005, 144, 28-34.

Hall, S. C.; Subramanian, V.; Teeter, G.; Rambabu, B. Solid State Ionics 2004, 175,

809-813.

Haller, G. L. Journal of Catalysis 2003, 216, 12-22.

Han, Y.-J.; Kim, J. M.; Stucky, G. D. Chem. Mater. 2000, 12, 2068-2069.

Hanrahan, J. P.; Copley, M. P.; Ziegler, K. J.; Spalding, T. R.; Morris, M. A.;

Steytler, D. C.; Heenan, R. K.; Schweins, R.; Holmes, J. D. Langmuir 2005,

21, 4163-4167.

Haran, N. P.; Bhagade, S. S.; Nageshwar, G. D. Chemicals & Petro-Chemicals

Journal 1978, 9, 21-8.

Hayashi, T.; Tanaka, K.; Haruta, M. Journal of Catalysis 1998, 178, 566-575.

Hayek, K.; Goller, H.; Penner, S.; Rupprechter, G.; Zimmermann, C. Catalysis

Letters 2004, 92, 1-9.

Hermans, S.; Raja, R.; Thomas, J. M.; Johnson, B. F. G.; Sankar, G.; Gleeson, D.

Angewandte Chemie, International Edition 2001, 40, 1211-1215.

Hiemenz, P. C. Principles of Colloid and Surface Chemistry, Marcel Dekker, Inc.:

New York, 1986.

Hillenbrand, L. J.; Lacksonen, J. W. Journal of the Electrochemical Society 1965,

112, 245-9.

Hirai, T.; Okubo, H.; Komasawa, I. J. Colloid and Interface Sci. 2001, 235, 358-364.

Hirai, T.; Okubo, H.; Komasawa, I. J. Phys. Chem. B 1999, 103, 4228-4230.

Hirsch, P. B.; et al. Electron Microscopy of Thin Crystals, 1977.

Hodgkins, R. P.; Ahniyaz, A.; Parekh, K.; Belova, L. M.; Bergstroem, L. Langmuir

2007, 23, 8838-8844.

Hornyak, G.; Kroll, M.; Pugin, R.; Sawitowski, T.; Schmid, G.; Bovin, J.-O.;

Karsson, G.; Hofmeister, H.; Hopfe, S. Chemistry--A European Journal 1997,

3, 1951-1956.

Horvath, G.; Kawazoe, K. J. Chem. Engg. Japan 1983, 16, 470-475.

266

Hou, J. G.; Li, X.; Wang, H.; Wang, B. Journal of Physics and Chemistry of Solids

2000, 61, 995-998.

Hourri, A.; St-Arnaud, J. M.; Bose, T. K. Journal of Chemical Physics 1997, 106,

1780-1785.

Huang, M. H.; Choudrey, A.; Yang, P. Chemical Communications (Cambridge) 2000,

1063-1064.

Huang, S.-M.; He, B.-L. Reactive Polymers 1994, 23, 11-18.

Hyun, C.; Lee, D. C.; Israel, C.; Korgel, B. A.; de Lozanne, A. IEEE Transactions on

Magnetics 2006, 42, 3799-3802.

Iijima, S. Nature (London, United Kingdom) 1991, 354, 56-8.

Imagawa, Y.; Kojima, K.; Seto, H. 2005, 17 pp.

Israelachvili, J. Intermolecular & Surface Forces, Academic Press, Inc.: San Diego,

1992.

Iwaki, T.; Kakihara, Y.; Toda, T.; Abdullah, M.; Okuyama, K. Journal of Applied

Physics 2003, 94, 6807-6811.

Jana, N. R.; Peng, X. Journal of the American Chemical Society 2003, 125, 14280-

14281.

Janssen, A. H.; Yang, C. M.; Wang, Y.; Schueth, F.; Koster, A. J.; De Jong, K. P.


Jarmer, D. J.; Lengsfeld, C. S.; Randolph, T. W. Journal of Supercritical Fluids 2003,

27, 317-336.

Jeyadevan, B.; Hobo, A.; Urakawa, K.; Chinnasamy, C. N.; Shinoda, K.; Tohji, K.


Jiang, Y.-X.; Weng, W.-Z.; Si, D.; Sun, S.-G. Journal of Physical Chemistry B 2005,

109, 7637-7642.

Johnston, K. P.; Shah, P. S. Science (Washington, DC, United States) 2004, 303, 482-

483.

267

Jongste, H. C. d.; Kuijers, F. J.; Ponec, V. Preparation of Catalysts, Elsevier:

Amsterdam, 1976.

Joo, S. H.; Choi, S. J.; Oh, I.; Kwak, J.; Liu, Z.; Terasaki, O.; Ryoo, R. Nature

(London, United Kingdom) 2001, 412, 169-172.

Joo, S. H.; Choi, S. J.; Oh, I.; Kwak, J.; Liu, Z.; Terasaki, O.; Ryoo, R. Nature

(London, United Kingdom) 2001, 414, 470.

Jung, J. S.; Lim, J. H.; Malkinski, L.; Vovk, A.; Choi, K. H.; Oh, S. L.; Kim, Y. R.;

Jun, J. H. Journal of Magnetism and Magnetic Materials 2007, 310, 2361-

2363.

Kanaras, A. G.; Soennichsen, C.; Liu, H.; Alivisatos, A. P. Nano Letters 2005, 5,

2164-2167.

Kaneda, M.; Tsubakiyama, T.; Carlsson, A.; Sakamoto, Y.; Ohsuna, T.; Terasaki, O.;

Joo, S. H.; Ryoo, R. Journal of Physical Chemistry B 2002, 106, 1256-1266.

Kang, S.; Jia, Z.; Zoto, I.; Reed, D.; Nikles, D. E.; Harrell, J. W.; Thompson, G.;

Mankey, G.; Krishnamurthy, V. V.; Porcar, L. Journal of Applied Physics

2006, 99, 08N704/1-08N704/3.

Kawi, S.; Gates, B. C. Journal of Catalysis 1994, 149, 317-25.

Kelley, S. P.; Lemert, R. M. AIChE Journal 1996, 42, 2047-2056.

Khlobystov, A. N.; Britz, D. A.; Wang, J.; O'Neil, S. A.; Poliakoff, M.; Briggs, G. A.

D. Journal of Materials Chemistry 2004, 14, 2852-2857.

Kim, K.-M.; Kyu, K.-S.; Park, N.-G.; Park, Y.-J.; Chung, S.-H. 2002, 6 pp.

Kim, K. T.; Kim, Y. G.; Chung, J. S. Journal of the Electrochemical Society 1995,

142, 1531-8.

Kim, T.-W.; Kleitz, F.; Paul, B.; Ryoo, R. Journal of the American Chemical Society

2005, 127, 7601-7610.

Kim, T.-W.; Solovyov, L. A. Journal of Materials Chemistry 2006, 16, 1445-1455.

Kim, Y.-G.; Garcia-Martinez, J. C.; Crooks, R. M. Langmuir 2005, 21, 5485-5491.

268

Kim, Y.-J.; Kang, J. G.; Lim, S.-H.; Song, J. H. Bulletin of the Korean Chemical

Society 2005, 26, 1129-1131.

King, N. C.; Blackley, R. A.; Wears, M. L.; Newman, D. M.; Zhou, W.; Bruce, D. W.

Chemical Communications (Cambridge, United Kingdom) 2006, 3414-3416.

Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties, John Wiey

and Sons: 1988.

Kinoshita, K. Journal of the Electrochemical Society 1990, 137, 845-8.

Klee, D.; Hilgers, C. 2002, 26 pp.

Klie, R. F.; Disko, M. M.; Browning, N. D. Journal of Catalysis 2002, 205, 1-6.

Knecht, M. R.; Garcia-Martinez, J. C.; Crooks, R. M. Langmuir 2005, 21, 11981-

11986.

Kockrick, E.; Krawiec, P.; Schnelle, W.; Geiger, D.; Schappacher, F. M.; Poettgen,

R.; Kaskel, S. Advanced Materials (Weinheim, Germany) 2007, 19, 3021-

3026.

Koh, S.; Leisch, J.; Toney, M. F.; Strasser, P. Journal of Physical Chemistry C 2007,

111, 3744-3752.

Koh, S.; Strasser, P. Journal of the American Chemical Society 2007, 129, 12624-

12625.

Koh, S.; Toney, M. F.; Strasser, P. Electrochimica Acta 2007, 52, 2765-2774.

Koh, S.; Yu, C.; Strasser, P. ECS Transactions 2007, 11, 205-215.

Komiyama, M.; Shimaguchi, T. Surface and Interface Analysis 2001, 32, 189-192.

Kongkanand, A.; Kuwabata, S.; Girishkumar, G.; Kamat, P. Langmuir 2006, 22,

2392-2396.

Kongkanand, A.; Martinez Dominguez, R.; Kamat, P. V. Nano Letters 2007, 7, 676-

680.

Konopny, L. W.; Juan, A.; Damiani, D. E. Applied Catalysis, B: Environmental 1998,

15, 115-127.

269

Konya, Z.; Puntes, V. F.; Kiricsi, I.; Zhu, J.; Ager, J. W., III; Ko, M. K.; Frei, H.;

Alivisatos, P.; Somorjai, G. A. Chemistry of Materials 2003, 15, 1242-1248.

Konya, Z.; Puntes, V. F.; Kiricsi, I.; Zhu, J.; Alivisatos, P.; Somorjai, G. A. Catalysis

Letters 2002, 81, 137-140.

Konya, Z. P., Victor F.; Kiricsi, Imre; Zhu, Ji; Ager, Joel W., III; Ko, Moon Kyu;

Frei, Heinz; Alivisatos, Paul; Somorjai, Gabor A. Chem. Mater. 2003, 15,

1242-1248.

Konya, Z. P., Victor F.; Kiricsi, Imre; Zhu, Ji; Alivisatos, A. Paul; Somorjai, Gabor

A. Nano Letters 2002, 2, 907-910.

Konya, Z. Z., Ji; Szegedi, Agnes; Kiricsi, Imre; Alivisatos, Paul; Somorjai, Gabor A.

Chem. Commun. 2003, 3, 314-315.

Koone, N. D.; Zerda, T. W. Journal of Non-Crystalline solids 1995, 183, 243-251.

Korgel, B. A.; Fitzmaurice, D. Physical Review B: Condensed Matter and Materials

Physics 1999, 59, 14191-14201.

Korgel, B. A.; Fullam, S.; Connolly, S.; Fitzmaurice, D. Journal of Physical

Chemistry B 1998, 102, 8379-8388.

Kresge, C. T.; Leonowickz, M. E.; Roth, W. J.; Beck, J. S. Nature 1992, 359, 710.

Kruk, M.; Jaroniec, M.; Ko, C. H.; Ryoo, R. Chemistry of Materials 2000, 12, 1961-

1968.

Krukonis, V. 1984.

Kuo, T.-C.; McCreery, R. L. Analytical Chemistry 1999, 71, 1553-1560.

Lagerev, S. P.; Babak, S. F. Zhurnal Obshchei Khimii 1937, 7, 1661-3.

Lai, F. S.; Gates, B. C. Nano Letters 2001, 1, 583-587.

Lake, L. W. Enhanced Oil Recovery, Prentice Hall, PTR: 1996.

Lang, H.; May, R. A.; Iversen, B. L.; Chandler, B. D. Journal of the American


Le Bars, J.; Specht, U.; Bradley, J. S.; Blackmond, D. G. Langmuir 1999, 15, 7621-

7625.

270

Lee, D. C.; Mikulec, F. V.; Pelaez, J. M.; Koo, B.; Korgel, B. A. Journal of Physical

Chemistry B 2006, 110, 11160-11166.

Lee, J.; Kim, J.; Hyeon, T. Advanced Materials (Weinheim, Germany) 2006, 18,

2073-2094.

Lee, J.; Park, Y.; Kim, P.; Kim, H.; Yi, J. Journal of Materials Chemistry 2004, 14,

1050-1056.

Lensveld, D. J.; Mesu, J. G.; van Dillen, A. J.; de Jong, K. P. Studies in Surface

Science and Catalysis 2002, 143, 647-657.

Leonard, B. M.; Bhuvanesh, N. S. P.; Schaak, R. E. Journal of the American


Lewis, J. E.; Biswas, R.; Robinson, A. G.; Maroncelli, M. Journal of Physical

Chemistry B 2001, 105, 3306-3318.

Li, H.; Wang, R.; Hong, Q.; Chen, L.; Zhong, Z.; Koltypin, Y.; Calderon-Moreno, J.;

Gedanken, A. Langmuir 2004, 20, 8352-8356.

Li, X.; Hsing, I. M. Electrochimica Acta 2006, 51, 3477-3483.

Li, Y.; Fu, Q.; Flytzani-Stephanopoulos, M. Applied Catalysis, B: Environmental

2000, 27, 179-191.

Liang, C.; Dai, S. Journal of the American Chemical Society 2006, 128, 5316-5317.

Liang, C.; Li, Z.; Dai, S. Angewandte Chemie, International Edition 2008, 47, 3696-

3717.

Lifshin, E. X-ray Characterization of Materials, Wiley-VCH: New York, 1999.

Lin, S.; Yin, C.-y.; Wang, L.-f.; Jiang, D.-z.; Xiao, F.-s. Shandong Daxue Xuebao,

Lixueban 2005, 40, 101-104.

Liu, S.-H.; Lu, R.-F.; Huang, S.-J.; Lo, A.-Y.; Chien, S.-H.; Liu, S.-B. Chemical


Liu, Z.; Koh, S.; Yu, C.; Strasser, P. Journal of the Electrochemical Society 2007,

154, B1192-B1199.

Lu, X.; Ziegler, K. J.; Ghezelbash, A.; Johnston, K. P.; Korgel, B. A. Nano Letters

2004, 4, 969-974.

271

Lu, Y. Angewandte Chemie, International Edition 2006, 45, 7664-7667.

Maiwald, M.; Schneider, G. M. Berichte der Bunsen-Gesellschaft 1998, 102, 960-

964.

Maldonado, S.; Morin, S.; Stevenson, K. J. Carbon 2006, 44, 1429-1437.

Mani, P.; Srivastava, R.; Strasser, P. Journal of Physical Chemistry C 2008, 112,

2770-2778.

Mani, P.; Srivastava, R.; Yu, C.; Strasser, P. ECS Transactions 2007, 11, 933-939.

Mathias, M. F.; Makharia, R.; Gasteiger, H. A.; Conley, J. J.; Fuller, T. J.; Gittleman,

C. I.; Kocha, S. S.; Miller, D. P.; Mittelsteadt, C. K.; Xie, T.; Yan, S. G.; Yu,

P. T. Electrochemical Society Interface 2005, 14, 24-35.

Matsumoto, T.; Komatsu, T.; Nakano, H.; Arai, K.; Nagashima, Y.; Yoo, E.;

Yamazaki, T.; Kijima, M.; Shimizu, H.; Takasawa, Y.; Nakamura, J. Catalysis

Today 2004, 90, 277-281.

Matthey, D.; Wang, J. G.; Wendt, S.; Matthiesen, J.; Schaub, R.; Laegsgaard, E.;

Hammer, B.; Besenbacher, F. Science (Washington, DC, United States) 2007,

315, 1692-1696.

Mattoussi, H.; Cumming, A. W.; Murray, C. B.; Bawendi, M. G.; Ober, R. Journal of

Chemical Physics 1996, 105, 9890-9896.

Mattoussi, H.; Cumming, A. W.; Murray, C. B.; Bawendi, M. G.; Ober, R. Physical

Review B: Condensed Matter and Materials Physics 1998, 58, 7850-7863.

Mattoussi, H.; Radzilowski, L. H.; Dabbousi, B. O.; Thomas, E. L.; Bawendi, M. G.;

Rubner, M. F. J. Appl. Phys. 1998, 83, 7965-7974.

McClellan, A. K.; McHugh, M. A. Polymer Engineering and Science 1985, 25, 1088-

92.

McLeod, M. C.; Anand, M.; Kitchens, C. L.; Roberts, C. B. Nano Letters 2005, 5,

461-465.

Meli, L.; Li, Y.; Lim, K. T.; Johnston, K. P.; Green, P. F. In Preparation 2007.

Meunier, F. C.; Tibiletti, D.; Goguet, A.; Shekhtman, S.; Hardacre, C.; Burch, R.

Catalysis Today 2007, 126, 143-147.

272

Midgley, P. A.; Weyland, M. Ultramicroscopy 2003, 96, 413-431.

Midgley, P. A.; Weyland, M.; Thomas, J. M.; Johnson, B. F. G. Chemical


Min, M.-K.; Cho, J.; Cho, K.; Kim, H. Electrochimica Acta 2000, 45, 4211-4217.

Minami, R.; Kitamoto, Y.; Chikata, T.; Kato, S. Electrochimica Acta 2005, 51, 864-

866.

Mohr, C.; Hofmeister, H.; Radnik, J.; Claus, P. Journal of the American Chemical

Society 2003, 125, 1905-1911.

Molnar, E. v.; Tasi, G.; Konya, Z.; Kiricsi, I. Catalysis Letters 2005, 101, 159-167.

Morey, M. S.; Davidson, A.; Stucky, G. D. Journal of Porous Materials 1998, 5, 195-

204.

Morgenstern, K.; Rosenfeld, G.; Poelsema, B.; Comsa, G. Physical Review Letters

1995, 74, 2058-2061.

Mukerjee, S. Journal of Applied Electrochemistry 1990, 20, 537-48.

Mukerjee, S.; Srinivasan, S. Journal of Electroanalytical Chemistry 1993, 357, 201-

24.

Mukerjee, S.; Srinivasan, S.; Soriaga, M. P. Journal of the Electrochemical Society

1995, 142, 1409-22.

Muraviev, D. N.; Macanas, J.; Parrondo, J.; Munoz, M.; Alonso, A.; Alegret, S.;

Ortueta, M.; Mijangos, F. Reactive & Functional Polymers 2007, 67, 1612-

1621.

Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, 8706-

8715.

Nam, J.-H.; Jang, Y.-Y.; Kwon, Y.-U.; Nam, J.-D. Electrochemistry Communications

2004, 6, 737-741.

Napolsky, K. S.; Eliseev, A. A.; Knotko, A. V.; Lukahsin, A. V.; Vertegel, A. A.;

Tretyakov, Y. D. Materials Science & Engineering, C: Biomimetic and

Supramolecular Systems 2003, C23, 151-154.

273

Nemanich, R. J.; Solin, S. A. Physical Review B: Condensed Matter and Materials

Physics 1979, 20, 392-401.

Ng, H.-J.; Robinson, D. B. Journal of Chemical and Engineering Data 1978, 23, 325-

327.

Niesz, K.; Grass, M.; Somorjai, G. A. Nano Letters 2005, 5, 2238-2240.

Nir, S. Progress in Surface Science 1977, 8, 1-58.

Norskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard,

T.; Jonsson, H. Journal of Physical Chemistry B 2004, 108, 17886-17892.

Oh, S.; Niu, Y.; Crooks, R. M. Langmuir 2005, 21, 10209-10213.

Ozkaya, D.; Zhou, W.; Thomas, J. M.; Midgley, P.; Keast, V. J.; Hermans, S.

Catalysis Letters 1999, 60, 113-120.

Pantea, D.; Darmstadt, H.; Kaliaguine, S.; Summchen, L.; Roy, C. Carbon 2001, 39,

1147-1158.

Papirer, E. Adsorption on Silica Surfaces CRC: 2000.

Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.;

Radmilovic, V.; Markovic, N. M.; Ross, P. N. Journal of Physical Chemistry

B 2002, 106, 4181-4191.

Pietron, J. J.; Stroud, R. M.; Rolison, D. R. Nano Letters 2002, 2, 545-549.

Pitkethly, R. C.; McKenzie, S.; Allum, K. G. 1974, 17 pp.

Platonova, O. A.; Bronstein, L. M.; Solodovnikov, S. P.; Yanovskaya, I. M.;

Obolonkova, E. S.; Valetsky, P. M.; Wenz, E.; Antonietti, M. Colloid and

Polymer Science 1997, 275, 426-431.

Ponec, V.; Bond, G. C. Cataysis by Metals and Alloys, Elsevier

Netherlands, 1995.

Qi, Z.-m.; Honma, I.; Zhou, H. Applied Physics Letters 2006, 88, 053503/1-053503/3.

Raghuveer, V.; Ferreira, P. J.; Manthiram, A. Electrochemistry Communications

2006, 8, 807-814.

274

Raghuveer, V.; Manthiram, A.; Bard, A. J. Journal of Physical Chemistry B 2005,

109, 22909-22912.

Raja, R.; Hermans, S.; Shephard, D. S.; Johnson, B. F. G.; Raja, R.; Sankar, G.;

Bromley, S.; Thomas, J. M. Chemical Communications (Cambridge) 1999,

1571-1572.

Raja, R.; Khimyak, T.; Thomas, J. M.; Hermans, S.; Johnson, B. F. G. Angewandte

Chemie, International Edition 2001, 40, 4638-4642.

Rajalakshmi, N.; Ryu, H.; Shaijumon, M. M.; Ramaprabhu, S. Journal of Power

Sources 2005, 140, 250-257.

Rajca, I. 1982, 4 pp.

Randolph, T. W.; Randolph, A. D.; Mebes, M.; Yeung, S. Biotechnology Progress

1993, 9, 429-435.

Rindfleisch, F.; DiNoia, T. P.; McHugh, M. A. Journal of Physical Chemistry 1996,

100, 15581-15587.

Rioux, R. M.; Song, H.; Hoefelmeyer, J. D.; Yang, P.; Somorjai, G. A. Journal of


Rolison, D. R. Science (Washington, DC, United States) 2003, 299, 1698-1702.

Roychowdhury, C.; Matsumoto, F.; Mutolo, P. F.; Abruna, H. D.; DiSalvo, F. J.


Ruban, A. V.; Skriver, H. L.; Norskov, J. K. Physical Review B: Condensed Matter

and Materials Physics 1999, 59, 15990-16000.

Rubio, J. E. F.; Arsuaga, J. M.; Taravillo, M.; Baonza, V. G.; Caceres, M.

Experimental Thermal and Fluid Science 2004, 28, 887-891.

Ryan, K. M.; Erts, D.; Olin, H.; Morris, M. S.; Holmes, J. D. J. Am. Chem. Soc. 2003,

125, 6284.

Ryoo, R.; Choi, M.; Choi, S. H. 2005, 13 pp.

Ryoo, R.; Joo, S. H. Studies in Surface Science and Catalysis 2004, 148, 241-260.

275

Ryoo, R.; Joo, S. H.; Jun, S. Journal of Physical Chemistry B 1999, 103, 7743-7746.

Ryoo, W.; Webber, S. E.; Johnston, K. P. Industrial & Engineering Chemistry

Research 2003, 42, 6348-6358.

Saito, A.; Foley, H. C. AIChE Journal 1991, 37, 429-36.

Salgado, J. R. C.; Antolini, E.; Gonzalez, E. R. Journal of Physical Chemistry B

2004, 108, 17767-17774.

Santiago, E. I.; Varanda, L. C.; Villullas, H. M. Journal of Physical Chemistry C

2007, 111, 3146-3151.

Saunders, A. E.; Korgel, B. A. Journal of Physical Chemistry B 2004, 108, 16732-

16738.

Saunders, A. E.; Shah, P. S.; Park, E. J.; Lim, K. T.; Johnston, K. P.; Korgel, B. A. J.

Phys. Chem. B 2004, 108, 15969-15975.

Saunders, A. E.; Sigman, M. B., Jr.; Korgel, B. A. Journal of Physical Chemistry B

2004, 108, 193-199.

Schlamp, M. C.; Peng, X.; Alivisatos, A. P. J. Appl. Phys. 1997, 82, 5837-5842.

Schlapka, A.; Lischka, M.; Gross, A.; Kasberger, U.; Jakob, P. Physical Review

Letters 2003, 91, 016101/1-016101/4.

Schlogl, R.; Abd Hamid Sharifah, B. Angew Chem Int Ed Engl 2004, 43, 1628-37.

Schueth, F. Chemistry of Materials 2001, 13, 3184-3195.

Schwartz, V.; Mullins, D. R.; Yan, W.; Chen, B.; Dai, S.; Overbury, S. H. Journal of


Scott, R. W. J.; Sivadinarayana, C.; Wilson, O. M.; Yan, Z.; Goodman, D. W.;

Crooks, R. M. Journal of the American Chemical Society 2005, 127, 1380-

1381.

Scott, R. W. J.; Wilson, O. M.; Crooks, R. M. Journal of Physical Chemistry B 2005,

109, 692-704.

Scott, R. W. J.; Wilson, O. M.; Crooks, R. M. Chemistry of Materials 2004, 16, 5682-

5688.

Segalman, R. A. Materials Science & Engineering, R: Reports 2005, R48, 191-226.

276

Sellin, M. F.; Webb, P. B.; Cole-Hamilton, T. J. Chem. Commum. 2001, 781-782.

Serp, P.; Hierso, J. C.; Feurer, R.; Kihn, Y.; Kalck, P.; Faria, J. L.; Aksoylu, A. E.;

Pacheco, A. M. T.; Figueiredo, J. L. Carbon 1999, 37, 527-530.

Shah, P. S.; Holmes, J. D.; Doty, R. C.; Johnston, K. P.; Korgel, B. A. J. Am. Chem.

Soc. 2000, 122, 4245-4246.

Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. J. Phys. Chem. B 2001, 105,

9433-9440.

Shah, P. S.; Husain, S.; Johnston, K. P.; Korgel, B. A. J. Phys. Chem. B 2002, 106,

12178-12185.

Shah, P. S.; Novick, B. J.; Hwang, H. S.; Lim, K. T.; Carbonell, R. G.; Johnston, K.

P.; Korgel, B. A. Nano Letters 2003, 3, 1671-1675.

Shao, Y.; Yin, G.; Gao, Y. Journal of Power Sources 2007, 171, 558-566.

Shao, Y.; Yin, G.; Gao, Y.; Shi, P. Journal of the Electrochemical Society 2006, 153,

A1093-A1097.

Shevchenko, E. V.; Talapin, D. V.; Kotov, N. A.; O'Brien, S.; Murray, C. B. Nature

(London, United Kingdom) 2006, 439, 55-59.

Shibata, G.; Deguchi, M.; Taomoto, A.; Ozaki, T. 2004, 8 pp.

Shukla, A. K.; Ravikumar, M. K.; Roy, A.; Barman, S. R.; Sarma, D. D.; Arico, A.

S.; Antonucci, V.; Pino, L.; Giordano, N. Journal of the Electrochemical

Society 1994, 141, 517-22.

Shyue, J.-J.; De Guire, M. R. Journal of the American Chemical Society 2005, 127,

12736-12742.

Sinfelt, J. H.; Taylor, W. F.; Yates, D. J. C. Journal of Physical Chemistry 1965, 69,

95-101.

Singh, U. K.; Albert Vannice, M. Journal of Catalysis 2000, 191, 165-180.

Sinha, A. K.; Seelan, S.; Tsubota, S.; Haruta, M. Topics in Catalysis 2004, 29, 95-

102.

277

Skaff, H.; Lin, Y.; Dinsmore, A.; Russell, T.; Emrick, T. Abstracts of Papers, 224th

ACS National Meeting, Boston, MA, United States, August 18-22, 2002 2002,

COLL-026.

Somorjai, G. A. Catalysis Letters 2005, 101, 1-3.

Somorjai, G. A.; Contreras, A. M.; Montano, M.; Rioux, R. M. Proceedings of the

National Academy of Sciences of the United States of America 2006, 103,

10577-10583.

Somorjai, G. A.; Rioux, R. M.; Grunes, J. Clusters and Nano-Assemblies: Physical

and Biological Systems, [International Symposium], Richmond, VA, United

States, Nov. 10-13, 2004 2005, 97-125.

Song, H.; Rioux, R. M.; Hoefelmeyer, J. D.; Komor, R.; Niesz, K.; Grass, M.; Yang,

P.; Somorjai, G. A. Journal of the American Chemical Society 2006, 128,

3027-3037.

Song, H.; Rioux, R. M.; Kim, F.; Somorjai, G. A.; Yang, P. Abstracts of Papers,

227th ACS National Meeting, Anaheim, CA, United States, March 28-April 1,

2004 2004, INOR-877.

Song, Z.; Cai, T.; Hanson, J. C.; Rodriguez, J. A.; Hrbek, J. Journal of the American


Spange, S.; Simon, F.; Heublein, G.; Jacobasch, H. J.; Boerner, M. Colloid and

Polymer Science 1991, 269, 173-8.

Sra, A. K.; Ewers, T. D.; Schaak, R. E. Chemistry of Materials 2005, 17, 758-766.

Srivastana, R.; Mani, P.; Hahn, N.; Strasser, P. Angewandte Chemie, International

Edition 2007, 46, 8988-8991.

Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.; Lucas, C. A.;

Markovic, N. M. Science (Washington, DC, United States) 2007, 315, 493-

497.

Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.; Wang,

G.; Ross, P. N.; Markovic, N. M. Nature Materials 2007, 6, 241-247.

Steele, B. C. H.; Heinzel, A. Nature (London, United Kingdom) 2001, 414, 345-352.

278

Steigerwalt, E. S.; Deluga, G. A.; Lukehart, C. M. Journal of Physical Chemistry B

2002, 106, 760-766.

Stevens, D. A.; Dahn, J. R. Carbon 2004, 43, 179-188.

Stevens, D. A.; Hicks, M. T.; Haugen, G. M.; Dahn, J. R. Journal of the


Stowell, C. A.; Korgel, B. A. Nano Letters 2005, 5, 1203-1207.

Strasser, P. Journal of Combinatorial Chemistry 2008, 10, 216-224.

Strasser, P.; Koh, S.; Yu, C. ECS Transactions 2007, 11, 167-180.

Strongin, D. R.; Carrazza, J.; Bare, S. R.; Somorjai, G. A. Journal of Catalysis 1987,

103, 213-15.

Subramanian, B.; Rajewski, R. A.; Snavely, K. J. Pharm. Sci. 1997, 86, 885-890.

Subramoney, S. Advanced Materials (Weinheim, Germany) 1998, 10, 1157-1171.

Sun, Q.; Sun, R.-A. Chemical Research in Chinese Universities 2002, 18, 307-310.

Sun, S. Advanced Materials (Weinheim, Germany) 2006, 18, 393-403.

Sun, S.; Fullerton, E. E.; Weller, D.; Murray, C. B. IEEE Transactions on Magnetics

2001, 37, 1239-1243.

Sun, S.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Science (Washington, D. C.)

2000, 287, 1989-1992.

Sun, X.; Kang, S.; Harrell, J. W.; Nikles, D. E.; Dai, Z. R.; Li, J.; Wang, Z. L. Journal

of Applied Physics 2003, 93, 7337-7339.

Tanev, P. T.; Pinnavaia, T. J. Chemistry of Materials 1996, 8, 2068-2079.

Tanev, P. T.; Pinnavaia, T. J. Science (Washington, D. C.) 1995, 267, 865-7.

Tang, Z.; Geng, D.; Lu, G. J Colloid Interface Sci 2005, 287, 159-66. .

Taylor, W. F.; Sinfelt, J. H.; Yates, D. J. C. Journal of Physical Chemistry 1965, 69,

3857-63.

Thomas, J. M.; Johnson, B. F. G.; Raja, R.; Sankar, G.; Midgley, P. A. Accounts of

Chemical Research 2003, 36, 20-30.

Thomas, L.; Lionti, F.; Ballou, R.; Gatteschi, D.; Sessoli, R.; Barbara, B. Nature

(London) 1996, 383, 145-147.

279

Tian, J. M.; Wang, F. B.; Shan, Z. H. Q.; Wang, R. J.; Zhang, J. Y. Journal of Applied


Tian, Z. Q.; Jiang, S. P.; Liu, Z.; Li, L. Electrochemistry Communications 2007, 9,

1613-1618.

Tibiletti, D.; Fonseca, A. A.; Burch, R.; Chen, Y.; Fisher, J. M.; Goguet, A.;

Hardacre, C.; Hu, P.; Thompsett, D. J. Phys. Chem. B 2005, 109, 22553-22559.

Tsang, S. C.; Chen, Y. K.; Harris, P. J. F.; Green, M. L. H. Nature (London) 1994,

372, 159-62.

Tsubota, S.; Cunningham, D. A. H.; Bando, Y.; Haruta, M. Studies in Surface Science

and Catalysis 1995, 91, 227-35.

Tsubota, S.; Haruta, M.; Kobayashi, T.; Ueda, A.; Nakahara, Y. Studies in Surface

Science and Catalysis 1991, 63, 695-704.

Tsung, C.-K.; Hong, W.; Shi, Q.; Kou, X.; Yeung, M. H.; Wang, J.; Stucky, G. D.

Advanced Functional Materials 2006, 16, 2225-2230.

Tuinstra, F.; Koenig, J. L. Journal of Chemical Physics 1970, 53, 1126-30.

Valden, M.; Lai, X.; Goodman, D. W. Science (Washington, D. C.) 1998, 281, 1647-

1650.

Van Bergen, F.; Gale, J.; Damen, K. J.; Wildenborg, A. F. B. Energy (Amsterdam,

Netherlands) 2004, 29, 1611-1621.

Vasiliev, M. A. Journal of Physics D: Applied Physics 1997, 30, 3037-3070.

Vegard, L. Zeitschrift fuer Physik 1921, 5, 17-26.

Vegard, L. Zeitschrift fuer Physik 1921, 5, 393-5.

Venezia, A. M.; Liotta, L. F.; Deganello, G.; Schay, Z.; Guczi, L. Journal of

Catalysis 1999, 182, 449-455.

Vigne-Maeder, F.; Amrani, S. E.; Gelin, P. J. Catal. 1992, 134, 536.

Vijayaraghavan, G.; Stevenson, K. J. Langmuir 2007, 23, 5279-5282.

Vondrova, M.; Klimczuk, T.; Miller, V. L.; Kirby, B. W.; Yao, N.; Cava, R. J.;

Bocarsly, A. B. Chemistry of Materials 2005, 17, 6216-6218.

280

Wagner, F. T.; Gasteiger, H. A.; Makharia, R.; Neyerlin, K. C.; Thompson, E. L.;

Yan, S. G. ECS Transactions 2006, 3, 19-29.

Waje, M. M.; Li, W.; Chen, Z.; Larsen, P.; Yan, Y. ECS Transactions 2007, 11,

1227-1233.

Wakabayashi, N.; Takeichi, M.; Uchida, H.; Watanabe, M. Journal of Physical

Chemistry B 2005, 109, 5836-5841.

Wakayama, H.; Fukushima, Y. Ind. Eng. Chem. Res. 2000, 39, 4641-4645.

Wakayama, H.; Fukushima, Y. Industrial & Engineering Chemistry Research 1344,

ACS ASAP.

Wang, A.-Q.; Chang, C.-M.; Mou, C.-Y. Journal of Physical Chemistry B 2005, 109,

18860-18867.

Wang, C.; Daimon, H.; Onodera, T.; Koda, T.; Sun, S. Angewandte Chemie,

International Edition 2008, 47, 3588-3591.

Wang, C.; Zhu, G.; Li, J.; Cai, X.; Wei, Y.; Zhang, D.; Qiu, S. Chemistry--A

European Journal 2005, 11, 4975-4982.

Wang, J.; Yin, G.; Shao, Y.; Wang, Z.; Gao, Y. Journal of Physical Chemistry C

2008, 112, 5784-5789.

Wang, K.; Morris, M. A.; Holmes, J. D. Chemistry of Materials 2005, 17, 1269-1271.

Wang, X. Q.; Liang, C.; Dai, S. Langmuir 2008, in press.

Wang, Z. L.; Petroski, J. M.; Green, T. C.; El-Sayed, M. A. Journal of Physical

Chemistry B 1998, 102, 6145.

Wark, M.; Wellmann, H.; Rathousky, J. Thin Solid Films 2004, 458, 20-25.

Watkins, J. J.; Blackburn, J. M.; McCarthy, T. J. Chemistry of Materials 1999, 11,

213-215.

Wei, M.; Musie, G. T.; Busch, D. H.; Subramanian, B. Green Chem. 2004, 6, 387-

393.

Wei, M.; Musie, G. T.; Busch, D. H.; Subramanian, B. J. Am. Chem. Soc. 2002, 124,

2513-2517.

281

Weimann, T.; Geyer, W.; Hinze, P.; Stadler, V.; Eck, W.; Golzhauser, A.

Microelectronic Engineering 2001, 57-58, 903-907.

Wikander, K.; Ekstroem, H.; Palmqvist, A. E. C.; Lundblad, A.; Holmberg, K.;

Lindbergh, G. Fuel Cells (Weinheim, Germany) 2006, 6, 21-25.

Wikander, K.; Petit, C.; Holmberg, K.; Pileni, M.-P. Langmuir 2006, 22, 4863-4868.

Williams, D. B.; Carter, C. B. Transmission Electon Microscopy, Springer: New

York, 1996.

Williams, J. W.; Krchma, I. J. J. Am. Chem. Soc. 1926, 48, 1888-96.

Wynne, K. J.; Shenoy, S.; Fujiwara, T.; Irie, S.; Woerdeman, D.; Sebra, R.; Garach,

A.; McHugh, M. Polymer Preprints (American Chemical Society, Division of

Polymer Chemistry) 2002, 43, 888.

Xie, X.; Liotta, C. L.; Eckert, C. A. Ind. Eng. Chem. Res. 2004, 43, 7907-7911.

Xiong, L.; Manthiram, A. Journal of the Electrochemical Society 2005, 152, A697-

A703.

Xiong, L.; Manthiram, A. Electrochimica Acta 2005, 50, 2323-2329.

Xu, Y.; Ruban, A. V.; Mavrikakis, M. Journal of the American Chemical Society

2004, 126, 4717-4725.

Xu, Z.; Xiao, F. S.; Purnell, S. K.; Alexeev, O.; Kawi, S.; Deutsch, S. E.; Gates, B. C.

Nature (London) 1994, 372, 346-8.

Yan, X. M.; Kwon, S.; Contreras, A. M.; Koebel, M. M.; Bokor, J.; Somorjai, G. A.

Catalysis Letters 2005, 105, 127-132.

Yang, C.-m.; Kalwei, M.; Schuth, F.; Chao, K.-j. Applied Catalysis, A: General 2003,

254, 289-296.

Yang, D. Q.; Sacher, E. Surface Science 2002, 516, 43-55.

Yang, D. Q.; Zhang, G. X.; Sacher, E.; Jose-Yacaman, M.; Elizondo, N. Journal of


Yano, H.; Kataoka, M.; Yamashita, H.; Uchida, H.; Watanabe, M. Langmuir 2007,

23, 6438-6445.

282

Ye, H.; Crooks, R. M. Journal of the American Chemical Society 2007, 129, 3627-

3633.

Ye, H.; Scott Robert, W. J.; Crooks Richard, M. Langmuir 2004, 20, 2915-20. .

Yeo, S.-D.; Debenedetti, P. G.; Radosz, M.; Schmidt, H.-W. Macromolecules 1993,

26, 6207-6210.

Yin, Y.; Alivisatos, A. P. Nature (London, United Kingdom) 2005, 437, 664-670.

Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angewandte Chemie, International Edition

1999, 38, 56-77.

Yoo, J. W.; Hathcock, D. J.; El-Sayed, M. A. Journal of Catalysis 2003, 214, 1-7.

Yu, R.; Song, H.; Zhang, X.-F.; Yang, P. Journal of Physical Chemistry B 2005, 109,

6940-6943.

Yu, X.; Ye, S. Journal of Power Sources 2007, 172, 145-154.

Zanella, R.; Giorgio, S.; Henry, C. R.; Louis, C. Journal of Physical Chemistry B

2002, 106, 7634-7642.

Zhang, G.; Yang, D.; Sacher, E. Journal of Physical Chemistry C 2007, 111, 565-570.

Zhang, J.-M.; Ma, F.; Xu, K.-W. Applied Surface Science 2004, 229, 34-42.

Zhang, J.; Han, B.; Hou, Z.; Liu, Z.; He, J.; Jiang, T. Langmuir 2003, 19, 7616-7620.

Zhang, J.; Lima, F. H. B.; Shao, M. H.; Sasaki, K.; Wang, J. X.; Hanson, J.; Adzic, R.

R. Journal of Physical Chemistry B 2005, 109, 22701-22704.

Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science (Washington, DC, United

States) 2007, 315, 220-222.

Zhang, L.; He, R.; Gu, H.-C. Applied Surface Science 2006, 253, 2611-2617.

Zhang, W.-H.; Daly, B.; O'Callaghan, J.; Zhang, L.; Shi, J.-L.; Li, C.; Morris, M. A.;

Holmes, J. D. Chemistry of Materials 2005, 17, 6407-6415.

Zhang, Z.; Lagally, M. G. Science 1997, 276, 377-383.

Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Frederickson, G. H.; Chmelka, B. F.;

Stucky, G. D. Science (Washington, D. C.) 1998, 279, 548-552.

Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. Journal of the American


283

Zhao, D.; Yang, P.; Margolese, D. I.; Stucky, G. D. Chemical Communications

(Cambridge) 1998, 2499-2500.

Zhao, L.; Zhu, G.; Zhang, D.; Chen, Y.; Qiu, S. Journal of Solid State Chemistry

2005, 178, 2980-2986.

Zheng, N.; Stucky, G. D. Journal of the American Chemical Society 2006, 128,

14278-14280.

284

Vita

Gaurav Gupta was born in Gwalior, India to Ramesh Chandra and Archana

Gupta. Upon graduating from Jawahar Lal Nehru High School, Bhopal, India in 1999,

he enrolled at Indian Institute of Technlogy (IIT) in Mumbai, India. While at IIT

Mumbai, he conducted undergraduate research and a senior research project under the

direction of Dr. Jayesh Bellare. He received his B.Tech. in Chemical Engineering in

May of 2003. He enrolled at The University of Texas at Austin in September of 2003

for graduate studies. As a graduate student he conducted research under the

supervision of Dr. Keith P. Johnston. He will receive his Ph.D. in Chemical

Engineering in August 2008. He will join ExxonMobil Upstream research center as a

senior research scientist in Nov. 2008.

Permanent Address: 50, Amaltas Phase-2,

Chunna Bhatti, Kolar Road

Bhopal, Madhya Pradesh, Pin Code- 462016

India

This dissertation was typed by the author.

copyright by gaurav gupta 2008

Documents