Platinum Alloy Nanoparticles: Composition, Shape, Structure. and Electrocatalytic Property

Transcription

1 Platinum Alloy Nanoparticles: Composition, Shape, Structure and Electrocatalytic Property By Zhenmeng Peng Submitted in Partial Fulfillment of the Requirement for the Degree Doctor of Philosophy Supervised by Professor Hong Yang Department of Chemical Engineering Arts, Science and Engineering Edmund A. Hajim School of Engineering and Applied Sciences University of Rochester Rochester, NY 2010

2 ii Curriculum Vitae The author was born in Linyi, Shandong Province, People s Republic of China, in He received his B.S. degree in Materials Physics in 2002, and his M.S. degree in Materials Physics and Chemistry in 2005, both from University of Science and Technology of China (USTC). The author came to the University of Rochester in fall 2005 to pursue his Ph.D. degree. He began his graduate studies in the Department of Mechanical Engineering and received a 2 nd M.S. degree in Materials Science in June Thereafter he transferred to the Department of Chemical Engineering and continued his studies under the direction of Professor Hong Yang. The author has received a Leon Huntington Hooker Fellowship in His doctoral research is focused on preparation and characterization of platinum alloy nanoparticles and their applications as proton exchange membrane fuel cell (PEMFC) electrocatalysts.

3 iii Acknowledgements I would like to express my gratitude to my advisor, Prof. Hong Yang, and thank him for the gracious help during my graduate study. It has been a great experience for me conducting research under his direction. He has been always teaching and encouraging me to think creatively and independently. I will benefit from his deep understanding of research and generous transfer of his knowledge in my future career. I would like to thank Prof. Jacob Jorne, Prof. Mitchell Anthamatten and Prof. Paul D. Funkenbusch for being on my committee, and for their great suggestions to my research. I would also like to thank Prof. Younan Xia, Dr. Eric Lee, Dr. Eric Formo and Dr. Pedro H. C. Camargo at Washington University in St. Louis for our intimate collaborations. I would like to thank Dr. Frederick T. Wagner and Dr. Junliang Zhang at General Motors Corporation for their kind suggestions on electrochemical measurements, and appreciate the generous help from my colleagues, especially from Dr. Xiaowei Teng, Dr. Yong Wang, Sean Maksimuk, Dr. Shengchun Yang, Dr. Rui Shen, Hongjun You, Jianbo Wu, Miao Shi, Qing Du, and Isthier Chaudhury. I would like to thank Mr. Brian McIntyre for use of the electron microscopes, Ms. Christine Pratt for her help with the X-ray diffractometer, and Prof. Matthew Z. Yates and his research group for use of their UV-vis spectrometer. I would

4 iv like to acknowledge the kind assistances from our department staff, especially from Ms. Sandra Willison, Ms. Gina Eagan, Ms. Andrea Costales, Ms. Tiffany Markham, Ms. Rosario E. Malaver and Mr. Larry Kuntz. Finally, I would like to thank all my family members. Without their kind support I could not study abroad and pursue my research. I especially thank my wife, Jing Li, for her understanding and great sacrifice during these several years. Only with her self-giving support and encouragement I can continue my research. I would also thank my son, Yiyang, whose arrival brings me tremendous happiness.

5 v Abstract With the increasing environmental concern and accelerated depletion of fossil fuel, there are renewed research interests in the development of new technology using alternative energy sources, most noticeably those fuels that can be utilized by proton exchange membrane fuel cells (PEMFCs). Platinum has been widely used as electrocatalyst in PEMFCs because of its outstanding catalytic property over others metals. While platinum is often essential to ensure outstanding catalytic properties, cost, activity and durability are still some of the key issues that have hindered its real world applications. Exploration of highly efficient and durable electrocatalysts with low Pt content is pivotal in advancing the fuel cell technology. Studies have demonstrated that both d-band electrons and surface geometric structure can greatly affect the activity of catalysts and can be optimized by tailoring their composition and shape and overall structure. The cost issue can be addressed in part by improving the specific activity and durability of the catalysts and by reducing the amount of platinum used. In this thesis I present my studies on the synthesis, characterization and electrochemical study of platinum alloy nanoparticles. A series of Pt-on-Metal (M=Ag, Au, Cu, Pd) heterogeneous nanostructures have been prepared and their electrocatalytic properties have been tested. The Pt-on-Pd catalyst exhibited both enhanced activity and much improved stability in oxygen reduction reaction

6 vi (ORR). We have been able to make other platinum alloy nanostructures from these Pt-on-M nanoparticles. Pt-Au bimetallic nanoparticles produced by thermal treatment were much more active than Pt in catalyzing formic acid oxidation reaction (FAOR). Platinum hollow nanospheres and cubic nanoboxes were obtained by an electrochemical approach and exhibited significant improvement in ORR and methanol oxidation reaction (MOR) activities. Study on novel platinum nanoalloys with composition in their bulk miscibility gap and their electrocatalytic property has resulted in the synthesis of PtAg alloy nanoparticles with a wide range of composition. An electrochemical method has been developed for the preparation of heterogeneous PtAg alloy catalysts with low-pt content cores and Pt-rich surfaces. The optimal PtAg catalyst not only can have much improved activity but also show limited degradation in FAOR. A generic chemical dealloying method has also been developed for making tiny metal nanoparticles with their size down to 1 nm. Their catalytic activity has been demonstrated using p-nitrophenol reduction as the model reaction.

7 vii Table of Contents Chapter 1 Introduction 1.1 Composition, Shape, and Structure Controls of Platinum 1 1 Alloy Nanoparticles Composition Control Composition Control of Alloys in Macroscopic 1 7 Miscibility Gap Shape Control Structure Control 1.2 Platinum Alloy Nanoparticles as PEMFCs Electrocatalysts Reaction Mechanism Current Status 1.3 Development of Next Generation PEMFC Catalysts Deterministic Factors in Electrocatalysis Design of Advanced Platinum Alloy Nanoparticles as Fuel Cell Catalyst 1.4 Objective and Overview of Thesis 1.5 References Chapter 2 PtAg Alloy Nanoparticles and their Formation of 42 Pt-Surface Rich Electrocatalyst 2.1 PtAg Alloy Nanoparticles with the Compositions in 42 Miscibility Gap Introduction Experimental Section Results and Discussion Conclusions

8 viii 2.2 Understanding the Composition-Dependent Formation of 57 Platinum-Silver Nanowires Introduction Experimental Section Results and Discussion Conclusions 2.3 An Electrochemical Approach to Pt-Surface Rich PtAg Alloy Nanostructure Introduction Experimental Section Results and Discussion Conclusions 2.4 References Chapter 3 Preparation and Electrocatalytic Property of Platinum 113 Hollow Nanoparticles and Cubic Nanoboxes 3.1 Preparation and Electrocatalytic Property of Platinum 113 Hollow Nanoparticles Made from Pt-on-Ag Heteronanostructures Introduction Experimental Section Results and Discussion Conclusions 3.2 An Electrochemical Restructuring Method for Making Ultrafine Platinum Cubic Nanoboxes Introduction Experimental Section

9 ix Results and Discussion Conclusions 3.3 References Chapter 4 Synthesis and Electrocatalytic Property of Pt-on-Pd 163 Heteronanostructures 4.1 Introduction 4.2 Experimental Section 4.3 Results and Discussion 4.4 Conclusions 4.5 References Chapter 5 Post-Synthesis Modification of Pt-on-Au 180 Heteronanostructures and Their Electrocatalytic Property 5.1 Introduction 5.2 Experimental Section 5.3 Results and Discussion 5.4 Conclusions 5.5 References Chapter 6 Ultrafine Metal and Metal Alloy Nanoparticles for 205 Catalytic Applications 6.1 Introduction 6.2 Experimental Section 6.3 Results and Discussion 6.4 Conclusions 6.5 References

10 x Chapter 7 Conclusions and Research Outlook 7.1 Summary and Concluding Remarks 7.2 Research Outlook New Alloy Compositions for Catalysis Platinum Alloy Nanowires via Oriented Attachment and their Application as Supportless Catalysts Finely Controlling the Surface Composition of 226 Alloy Catalyst Based on Electrochemistry Preparation of Supported Metal Alloy Tiny 227 Nanoparticle Catalyst via Chemical Dealloying 7.3 References 228 Appendix: Publication List of the Work Conducted at University of 230 Rochester

11 xi List of Figures Figure Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 1.5 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Title Schematic representation of three types of platinum alloys: (a) random, (b) clustered, and (c) ordered forms. Phase diagram of bulk Ag-Pt binary system, redrawn after ref. 10. Schematic illustration of (a) overall excess free energy G r and (b) average number of clusters N r as a function of cluster size r for heterogeneous and homogeneous nucleation. Reaction pathways of (a) oxygen reduction (ORR), (b) hydrogen oxidation (HOR), (c) methanol oxidation (MOR) and (d) formic acid oxidation reactions (FAOR). Schematic illustration of various advanced Pt-based nanostructures. TEM images of PtAg nanoparticles with different molar ratios of Pt/Ag precursors: (a) 1:4; (b) 1:2; (c) 2:1; (d) 4:1. XRD patterns of as-prepared PtAg nanoparticles with different molar ratios of Pt/Ag precursors. (a) Representative EDX spectrum of PtAg nanoparticles made from a mixture with molar ratio of Pt/Ag precursors equal to 1:4, and (b) relationship between Pt atomic ratios in final PtAg alloys versus those in the metal precursors. Experimental and theoretical data of lattice parameter and composition relationship for as-synthesized and thermally-annealed Ag-Pt nanostructures. The theoretical data were calculated based on the Vegard s law. Page

12 xii Figure 2.5 Figure 2.6 Figure 2.7 Figure 2.8 Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 Figure 2.13 Figure 2.14 PXRD patterns of thermally annealed Ag-Pt alloy nanostructures in (a) θ range and (b) the enlarged (111) and (200) diffractions. TEM images of carbon-supported Ag 74 Pt 26 nanoparticles at (a, b) 3, (c, d) 10, and (e, f) 70 wt.% metal loadings (a, c, e) before and (b, d, f) after the thermal annealing. PXRD patterns in (a) θ range and (b) the enlarged (111) and (200) diffractions for carbon-supported Ag 74 Pt 26 NPs after annealing at 700 C. The loading is based on the total weight of metals. TEM images of (a) Pt 26 Ag 74, (b) Pt 39 Ag 61, (c, d) Pt 53 Ag 47, (e) Pt 73 Ag 27, and (f) Pt 86 Ag 14 nanostructures, respectively. XRD pattern of worm-like Pt 53 Ag 47 nanostructures. TEM images of Pt 53 Ag 47 alloy nanostructures made after the reaction taking place for: (a) 2, (b) 10, (c) 30, and (d) 180 min, respectively. TEM images and the corresponding schematic illustrations showing the early growing stages from (a) primary particle to two-particle system through (b) MA and (c) TA growthes, and (d-i) three-particle systems of Pt 53 Ag 47 nanowires through either MA or TA growth, respectively. Potential energy (E p ) as a function of distance between two 3-nm Pt 50 Ag 50, Pt and Ag nanoparticles. Side (top images) and top (bottom images) views of the two possible configurations of amine functional group on Pt 50 Ag 50 {111} surfaces. Color code: orange, Ag; dark blue, Pt; yellow, H; gray, C; and blue, N. Adsorption energy (E ad ) of functional groups for (a) OAm

13 xiii Figure 2.15 Figure 2.16 Figure 2.17 Figure 2.18 Figure 2.19 Figure 2.20 Figure 2.21 and (b) OA molecules on the three low-index surfaces of Pt 50 Ag 50, Ag and Pt, respectively. TEM images of (a) Ag and (b) Pt nanoparticles prepared under the same conditions as those for making Pt 53 Ag 47 nanowires. Simulation of time-dependent (a) mean square displacement (MSD) of surface atoms of 3-nm nanoparticles of PtAg alloys, Ag and Pt metals; (b) changes in particle-particle distance (D) and (c) total energy (E); and (d-f) pair correlation function, g(r), as a function of r at the interfacial regions between two colliding Pt 50 Ag 50 particles. (a) Low and (b) high magnification TEM images and (c) PXRD pattern of PtAg alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratio of 1/6. Representative TEM images of PtAg mixed alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratios of (a) 1/8, (b) 1/4, (c) 1/3, and (d) 1/2, respectively. (a) Representative EDX spectrum of PtAg alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratio of 1/6, and (b) relationship of Pt atomic percentage (at.%) in the products and their corresponding precursors. The Pt at.% was calculated based on the ratio of mole number of Pt over the total mole number of Pt and Ag metals. (a) Low and (b) high magnification TEM images and (c) PXRD pattern of carbon-supported Pt 18 Ag 82 mixed alloy nanoparticles after the thermal treatment. CV curves for Ag dissolution from carbon-supported (a) Ag and (b-c) Pt x Ag y alloy nanoparticles at the different upper

14 xiv Figure 2.22 Figure 2.23 Figure 2.24 Figure 2.25 Figure 2.26 potentials ranging from 0.6 to 1.2 V at 0.1-V increment. The composition of the original alloy nanoparticles was Pt 18 Ag 82. For clarity, only the initial one and a half cycles are shown in (b) and the enlarged hydrogen adsorption regions are shown in (c). CV curves of ten continuous cycles for Ag dissolution from carbon-supported Ag nanoparticles with a scan range between 0 and 0.8 (top) and 1.0 V (bottom), respectively. See experimental sections for details. CV curves of ten continuous cycles for silver dissolution from carbon-supported Pt 18 Ag 82 nanoparticles with a scan range between 0 and a given upper potential limit of 0.6, 0.8, 1.0, or 1.2 V, respectively. See experimental sections for details. Representative TEM images and size distribution analyses of carbon-supported nanoparticles made from Pt 18 Ag 82 alloy nanoparticles after the dissolution of Ag at the upper potentials of (a, b) 0.6, (c, d) 1.0, and (e, f) 1.2 V, respectively. See text for changes of compositions and nanostructures of the resulting alloy particles. Representative HR-TEM, STEM images and corresponding single-point EDX analyses of three different types of heterogeneous alloy nanostructures that had overall compositions of (a-d) Pt 18 Ag 82, (e-h) Pt 34 Ag 66, and (i-l) Pt 58 Ag 42, respectively. Representative STEM images (left) and their corresponding EDX line scans (right) of the three types of heterogeneous alloy nanoparticles that had overall compositions of (a,b)

15 xv Figure 2.27 Figure 2.28 Figure 2.29 Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Pt 18 Ag 82, (c,d) Pt 34 Ag 66, and (e, f) Pt 58 Ag 42, respectively. Schematic illustration of compositional and structural changes of the heterogeneous PtAg alloy nanoparticles made by controlled dissolution of Ag metal. CV curves showing the (a) mass current density (i mass ) and (b) area-specific current density (i area ) in the forward (solid line) and backward (dash line) scans, (c) bar graph illustrating both i mass and i area, and (d) changes in i mass at 0.6 V over multiple formic acid oxidation cycles catalyzed by the resultant heterogeneous Pt 34 Ag 66 and pure Pt electrocatalysts. CV curves showing the (a) area-specific current density (i area ) in the forward (solid line) and backward (dash line) scans, and (b) bar graph illustrating i area in formic acid oxidation catalyzed by the resultant heterogeneous PtAg alloy nanoparticles with overall compositions of Pt 18 Ag 82, Pt 34 Ag 66, and Pt 58 Ag 42. Representative TEM and HR-TEM images of as-prepared (a, b) Ag and (c, d) Pt-on-Ag nanoparticles, respectively. Representative HAADF-STEM images and their corresponding elemental maps of Pt-on-Ag nanoparticles at (a-d) low and (e-h) high magnifications. (a) PXRD patterns of as-prepared Ag (bottom) and Pt-on-Ag (top) nanoparticles, and (b) the corresponding EDX spectrum for Pt-on-Ag nanoparticles. UV-vis spectra of Ag (0 min) and Pt-on-Ag nanoparticles made after the reaction for a time period between 2 and 120 min, respectively

16 xvi Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 Figure 3.9 Figure 3.10 Figure 3.11 Figure 3.12 Representative TEM and HR-TEM images of Pt-on-Ag nanoparticles made after the reaction for (a, b) 2, (c, d) 30, and (e, f) 120 min., respectively. Representative TEM images of Pt-on-Ag nanoparticles made at (a) 170 and (b) 190 C, respectively. The reaction time was 1 h and Pt(acac) 2 /Ag metal molar ratio was kept at one. Representative TEM images of Pt-on-Ag nanoparticles made at the Pt(acac) 2 /Ag metal molar ratio of (a) 1/2 and (b) 2/1, respectively. These reactions took placed at 180 C for 1 h. Representative TEM and HR-TEM images of carbon-supported (a, b) Pt-on-Ag nanoparticles, and the resulting products after (c, d) acid and (e, f) thermal treatments. PXRD patterns of acid-treated Pt-on-Ag nanostructures (bottom) and thermally-treated PtAg nanoparticles (top). Inset: enlarged region for the (111) diffractions. Representative (a) TEM, (b) HR-TEM, (c) STEM image, and (d, e) elemental maps of carbon-supported Pt hollow structures made from carbon-supported Pt-on-Ag nanoparticles. Representative (a) TEM, (b) HR-TEM, (c) STEM image and the corresponding (d) Ag and (e) Pt elemental maps of carbon-supported PtAg alloy nanoparticles after the same acid treatment as that for making Pt hollow nanostructures from Pt-on-Ag nanoparticles. CV curves of carbon-supported Pt nanoparticle (E-TEK,

17 xvii Figure 3.13 Figure 3.14 Figure 3.15 Figure 3.16 Figure 3.17 Figure 3.18 wt%), hollow, and PtAg alloy catalysts. All tests were performed in a 0.1-M HClO 4 aqueous solution and the CV curves were recorded at a scan rate of 50 mv/s. (a) ORR polarization curves, (b) intrinsic mass current density, (c) CV, and (d) hydroxyl surface coverage (Θ OHad ) of carbon-supported Pt-on-Ag nanoparticle, PtAg alloy, Pt hollow and particle reference catalysts. Area-specific current density of oxygen reduction reaction using carbon-supported catalysts of platinum nanoparticles (E-TEK, 20 wt%), hollows and PtAg alloy nanostructures. Representative (a) TEM and (b) HR-TEM images of carbon-supported platinum hollow structures after an accelerated electrochemical stability test. Oxidation-reduction cycles were carried out through applying 30,000 linear potential sweeps between 0.6 and 1.0 V. TEM images of as-prepared (a) Ag and (b) Pt-on-Ag nanoparticles. Representative TEM images of Pt hollow nanocubes at (a) low and (b) high magnifications, and (c) individual cubes imaged under various tilting angles with respective to the direction of imaging beam. (a) HR-TEM image, (b) EDX spectrum, (c) STEM images with a Pt-M line scan and (d) TEM images taken at different rotating angles of the stage. The characterizations were done with the same Pt hollow cube, except EDX spectrum shown in (b), which was obtained using large amount of Pt nanoboxes with the SEM technique

18 xviii Figure 3.19 Figure 3.20 Figure 3.21 Figure 3.22 Figure 3.23 Figure 4.1 Figure 4.2 Figure 4.3 Representative TEM images of various morphologies of Pt nanostructures obtained with different potential scanning (a, b) range, (c) profile, and (d) rate. TEM images of Pt nanostructures after linear potential cycles in the range of (a) V and (b-c) V, respectively. The total numbers of cycles were 3000 for those samples shown in (a) and (c), and 800 for that shown in (c). (a) TEM, (b) HR-TEM, (c) STEM images superimposed with Pt-M line scan, and (d) EDX spectrum of Pt hollow nanostructures after the removal of Ag cores. TEM images of Pt hollow nanostructures after linearly cycling the potential in the range of V for (a) 1000 and (b) 6000 times, respectively. (a) Methanol oxidation reaction catalyzed by Pt cubic nanoboxes, hollow nanoshperes and the commercial Pt catalyst (TKK, 46.7 wt.% Pt), and (b) turnover frequencies (TOF) at the peak potentials. The CV curves of these two prepared catalysts in a sulfuric acid aqueous solution are shown in the inset of (a). Representative TEM image of as-synthesized Pd nanoparticles. PXRD patterns of as-prepared Pd and Pt-on-Pd nanoparticles. The intensity and position for Pt (blue) and Pd (purple) references were taken from the JCPDS database. Representative (a) TEM, (b) HR-TEM, (c) HAADF-STEM images and (d, e) elemental maps for Pd and Pt metals of

19 xix Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.8 Figure 4.9 Figure 5.1 Pt-on-Pd bimetallic nanoparticles; and (f) TEM and (g) HR-TEM images of carbon-supported Pt-on-Pd bimetallic catalysts after the thermal treatments. Energy dispersive X-ray (EDX) spectrum of as-prepared Pt-on-Pd nanoparticles collected on a field emission scanning electron microscope (FE-SEM, Zeiss-Leo DSM982). (a) HAADF-STEM image and (b, c) elemental maps of Pd and Pt metals, and (d) PXRD pattern of carbon-supported Pt-on-Pd bimetallic nanoparticles after thermal treatments. The intensity and position for Pt (blue) and Pd (purple) references were taken from the JCPDS database. (a) CV, (b) hydroxyl surface coverage (Θ OH ), (c) ORR polarization curves and (d) specific kinetic current densities (i k ) for carbon-supported Pt-on-Pd and Pt catalysts. CV and ORR polarization curves for carbon-supported (a, b) Pt-on-Pd and (c, d) Pt catalysts before and after 30,000 cycles. Representative (a) TEM, (b) HR-TEM image, and (c) EDX spectrum of carbon-supported Pt-on-Pd bimetallic nanoparticles after accelerated stability test (30,000 cycles). TEM images and particle size distribution analyses of Pt catalysts (E-TEK, 20wt.% Pt): (a, b) as-received and (c, d) after 30,000 CV cycles of accelerated stability test. Representative TEM images of (a) Au and (b) Pt-on-Au nanoparticles (NPs), (c) HAADF-STEM image and (d, e) the elemental maps for Au and Pt metals, and (f) HR-TEM image of a single Pt-on-Au nanoparticle; and (g) the

20 xx Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 corresponding XRD patterns of Au and Pt-on-Au nanoparticles. All Pt-on-Au nanoparticles were made at 180 C for 1 h. EDX spectrum of as-prepared Pt-on-Au nanoparticles from Au nanoparticles and Pt(acac) 2. The Au nanoparticle/pt(acac) 2 ratio was 1:4 based on the mole numbers of the metals. UV-vis spectra and TEM images (insets) of Au (0 min) and Pt-on-Au nanoparticles made after predetermined periods of reaction times. The Au/Pt precursor ratio was 1:4 and the reaction temperature was 180 C. TEM images of Pt-on-Au nanoparticles formed at four different reaction temperatures: (a) 170, (b) 180, (c) 190, and (d) 200 C. The Au nanoparticle/pt(acac) 2 feeding ratio was 1:4 and the reaction time was 1 h. Representative TEM images of Pt-on-Au nanoparticles made from precursors with Au nanoparticle/pt(acac) 2 molar ratios of (a) 1:1, (b) 1:2, (c) 1:3 and (d) 1:6, respectively. The reactions were carried out at 180 C for 1 h. Representative TEM images of Pt-on-Au nanoparticles on carbon supports (a) before and (b) after the thermal treatment; (c) HR-TEM image, (d) STEM image, and (e, f) the corresponding the elemental maps for Au and Pt metals, and (g) PXRD pattern of PtAu alloy heteronanostructures obtained through the thermal treatment. Cyclic voltammetry of formic acid oxidations by carbon-supported PtAu heteronanostructures, Pt and Au nanoparticle references. All experiments were conducted

21 xxi Figure 5.8 Figure 5.9 Figure 5.10 in a 0.5-M formic acid aqueous solution with 0.5-M H 2 SO 4 as the supporting electrolyte. Representative TEM images of carbon-supported (a) Pt (60 wt%), and (b) Au (20 wt.%) catalysts from E-TEK. Cyclic voltammetry curves of carbon-supported Pt (E-TEK, 60 wt%), Au (E-TEK, 20 wt%) catalysts and PtAu bimetallic nanostructures. The CV was conducted in a 0.5-M H 2 SO 4 aqueous solution under the protection of Ar. All measurements were run at ambient room temperatures and with a scan rate of 50 mv/s. CV curves of CO stripping on surfaces of Pt (60 wt.% Pt) and PtAu bimetallic nanostructures. All tests were Figure 5.11 Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4 conducted in 0.1-M HClO 4 aqueous solutions under argon protection. Change of current density over CV cycles at 0.3 and 0.5 V for carbon-supported PtAu heteronanostructures and Pt nanoparticles. All experiments were conducted in a 0.5-M formic acid aqueous solution using 0.5-M H 2 SO 4 as the supporting electrolyte. (a) TEM image, (b) size distribution, and (c) EDX spectrum of as-synthesized trace-pt content PtAg nanoparticles. TEM images of trace-pt content PtAg nanoparticles on (a) carbon, (b) silicon carbide, and (c) titanium oxide supports, and (d-f) their corresponding STEM images. (a) TEM image and (b) size distribution of the prepared carbon-supported tiny Pt nanoparticles. STEM and HR-TEM images of the prepared tiny Pt nanoparticles on (a, b) carbon, (c, d) silicon carbide, and (e,

22 xxii Figure 6.5 Figure 6.6 Figure 6.7 Figure 6.8 Figure 7.1 f) titanium oxide supports. EDX spectra of the prepared tiny Pt nanoparticles on (a) carbon, (b) silicon carbide, and (c) titanium oxide supports. (a) TEM and HR-TEM (Inset) images and (b) EDX spectrum of carbon-supported trace-au content AuAg nanoparticles, and (c) TEM and HR-TEM (Inset) images and (d) EDX spectrum of carbon-supported trace-pd content PdAg nanoparticles. (a) STEM and HR-TEM (Inset) images and (b) EDX spectrum of the prepared carbon-supported tiny Au nanoparticles, and (c) STEM and HR-TEM (Inset) images and (d) EDX spectrum of the prepared carbon-supported tiny Pd nanoparticles. (a) The extinction spectra at different reaction time, with the disappearance of the peak for p-nitrophenol due to its reduction into p-aminophenol, and (b) the Arrhenius plot for this reaction catalyzed by carbon-supported platinum tiny nanoparticles. Schematic illustration of energy changes upon mixing two immiscible metals in bulk

23 xxiii List of Tables Table Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 3.1 Table 7.1 Title Compositions of Pt-Ag Alloy Nanostructures Prepared at Different Pt(acac) 2 /Ag(St) Feeding Ratios. Simulation Results on Adsorption Energy of Functional Groups of Oleylamine (OAm) and Oleic Acid (OA) for the Strongest Adsorptions on Three Low-Index Ag, Pt and Pt 50 Ag 50 Surfaces. Particle size, size distribution, and composition of PtAg nanoalloys prepared at different Pt(acac) 2 /Ag(St) feeding ratios. Calculation of silver loss, overall and surface compositions, and diameter of the resultant Pt x Ag y heterogeneous nanostructures made from Pt 18 Ag 82 nanoparticles. Summary of Pt-on-Ag heteronanostructures prepared under various reaction conditions. Summary of immiscible bimetals. Page

24 Chapter 1 1 Chapter 1 Introduction 1.1 Composition, Shape, and Structure Controls of Platinum Alloy Nanoparticles Composition Control Alloys are solid solutions of two or more metals, among which bimetallic systems have been widely studied. 1 A major advantage of alloys over their corresponding pure metals is their broad ranges of compositions and properties. Alloys can be classified as random, clustered and ordered forms according to their atomic arrangements. An alloy that has long-range atomic orders is also called as an intermetallic compound, or intermetallic. 1, 2 The need to control the composition and structure of platinum-based materials in order to have certain desired property has been the major driving force for researchers to study Pt alloys and intermetallics. For an ideal binary system, the excess Gibbs free energy upon mixing ( ) Gmix where atoms are arranged randomly can be expressed by the changes in enthalpy ( H mix ) and entropy ( S mix ): 1, 3 ( X A GV, A + X B GV B ) = H mix T S mix Gmix = GV AB,, (1.1) H mix = N AzX A X Bε (1.2) 1 ε = ε AB AA + 2 mix ( ε ε ) BB ( X lnx X lnx ) A A B B (1.3) S = R + (1.4)

25 Chapter 1 2 where G V, AB G, G,, V A and V B are molar free energy terms for alloy AB, pure metals A and B, respectively; X A and X B mole fractions of A and B in alloy, z number of bonds per atom, and ε energy difference between A-B bond energy ( ε AB ) and average energy terms of A-A ( ε AA ) and B-B ( ε BB ) bonds. In a bulk system, bimetallics can lower its overall free energy through the formation of alloys if Gmix becomes negative. Based on Equation (1.4), the change in entropy is positive upon mixing two metals, thus the formation of alloys is always entropically favored. If the bond formation between metal A and B is exothermic (ε<0), the change in enthalpy should be negative. Under such conditions the excess free energy upon mixing ( spontaneously. Gmix ) is negative and an alloy should form The situation becomes complicated if the mixing is endothermic (ε >0, H mix >0). The formation of alloy should then be temperature dependent. 1, 4 At high temperatures, contribution from entropy becomes larger than that from enthalpy and the excess free energy turns negative, alloys should form. However, the excess Gibbs free energy can be positive at low temperatures according to Equation (1.1). In the latter case, two metals become immiscible, and as a consequence, there is a miscibility gap, within which alloys do not form. Real bimetallic systems can behave differently from those situations predicted by the idealized model, which do not consider stain energy. 4, 5 The difference in size between metal atoms is the main cause for the strain energy. The atomic

26 Chapter 1 3 arrangement in real bimetallic systems is usually different from the idealized case, where random arrangement of atoms has the lowest Gmix. Either ordered or clustered atomic arrangements can be preferred enthalpically. In another word, the enthalpy for mixing two metals can be lowered by reducing the number of A-B bonds for those cases when ε is negative or by increasing this number when ε is positive. On the other hand, random arrangement of atoms is always favored entropically. The formation of random, clustered or, ordered alloy is obtained when Gmix value reaches the minimum under given conditions (Figure 1.1). Figure 1.1 Schematic representation of three types of platinum alloys: (a) random, (b) clustered, and (c) ordered forms. The formation of colloidal alloy nanoparticles can be qualitatively understood using the classical nucleation theory. 3, 6 To form alloys, a system has to have large negative excess Gibbs free energy upon mixing ( the volume free energy of the alloy ( G V, AB Gmix ). In another word, ) should be much more negative than those of two metals themselves, G V, A G, and V B. The essential parameters * r, * G, * r N and dn r * dt for nucleation and growth of alloy nanoparticles 3, 7-9 can be described in the following equations:

27 Chapter 1 4 r γ G * 2 AB = AB V, AB (1.5) G * AB 16πγ = 3 G 3 AB 2 V, AB (1.6) ( [ ] ) * G AB N * = N A A eq S A + [ B] eq SB exp (1.7) rab RT dn dt * AB r = f AB N A [ A] [ A] eq ( ) * AB S + B S exp RT eq A [ ] eq B G (1.8) Usually the surface energy terms (γ ) for metal and alloy are comparable, which can result in r < r r and * AB * * A, B G < G G, suggesting alloy * AB *, * A B nanoparticles should form stable nuclei at smaller diameters than those for pure metal counterparts. Based on the relationship between the number of clusters, N r, and the size, r, most of the nuclei formed should be made of alloys and the nucleation for pure metals is negligible. There exist a range of platinum binary alloys as bulk solids, such as CoPt, CuPt, FePt, NiPt, PdPt, PtRh and PtRu One important consideration for colloidal synthesis of these alloy nanostructures has to do with the availability of zero-valence solutes of corresponding metals in solution during the reaction. The solute atoms of platinum and other metals need to be present in solution simultaneously at both nucleation and growth stages and at designed concentrations in order to produce nanoalloys with predetermined compositions. Most of metal precursors used in colloidal synthesis are inorganic or organometallic salts where metal ions can be reduced to zero-valence atoms. In

28 Chapter 1 5 some cases, thermally decomposable metal precursors are suitable for use in the synthesis. While platinum ions can be easily reduced to zero valence metal in solutions because of its high standard electrode potentials, reduction for most of other transition metals are not that as straight forwards. 5 Strong reducing agents are often needed to insure all metal precursors can be reduced simultaneously at 5, proper rates. Sodium borohydride, hydrazine and hydrogen, are among the 11, 19, 20 commonly used reducing agents for making platinum alloys nanoparticles. By using the co-reduction method, Pt-Co, 11 Pt-Ni, 19 Pt-Cu, 15 Pt-Pd 20, 21 and Pt-Ru nanoparticles have been made, though these binary nanoalloys do not have uniform size, shape and composition. As counter ions and solvents can change the redox potentials of given metal ions, choosing proper metal precursors and reaction conditions can sometimes effectively adjust the difference in reducing rates of those metal ions involved. 22 The changes in reaction kinetics facilitate the formation of alloy nanostructures with designed compositions. Nanoparticles of Ni x Pt 1-x alloys (x= 0 to 1) were synthesized from nickel(ii) acetate and platinum(ii) acetylacetonate at elevated temperatures using 1,2-hexadecanediol as the reducing agent. 23 Ultrafine FePt nanoparticles were prepared using a similar method, where iron and platinum acetylacetonates were reduced by long carbon-chain diols in organic solvents. 24 Co-reduction of Pt(acac) 2 and Ru(acac) 3 by 1,2-hexadecanediol in organic solvents led to the 25, 26 formation of fairly uniform PtRu nanoparticles.

29 Chapter 1 6 Thermal decomposition is an alternative way to produce zero valence atom solutes for metals other than platinum which is reduced from the salt precursors. Metal carbonyl is such a class of precursors that are often used in nonhydrolytic synthetic media, as the decomposition products are generally quite clean and easy to control kinetically. Iron pentacarbonyl has been widely used in the synthesis of FePt alloy nanoparticles, where Fe solute atoms could be produced by controlling reaction temperatures Iron platinum nanostructure is one of the few alloys where composition, size, and even shape can be controlled well. 30 Nanoparticles of CoPt alloys have also been synthesized using cobalt carbonyl as 31, 32 the thermally decomposable precursor. Random alloys do not necessarily have the minimum free energy. When H mix dominates, metal atoms bond in ordered fashion and form intermetallic compounds which have certain stoichiometry and long-range order. The ordered arrangement in intermetallics often breaks the symmetry of original crystal structures of the pure metals in order to minimize the system total free energy that can be affected by factors, such as relative atomic size and electronegativity of the metal atoms. 1, 5 FePt nanoparticles with a face-centered tetragonal (fct) phase (L1 0 ) are among 27, 33 the best known intermetallic alloys in recent years. However, colloidal FePt nanoparticles can hardly be prepared under relatively mild synthetic conditions. 27, 34 An indirect approach is to thermally treat the disordered fcc FePt

30 Chapter 1 7 nanoparticles, which are synthesized by thermal decomposition of iron pentacarbonyl and reduction of platinum acetylacetonate in the presence of long carbon-chain diols, , 30 to induce the phase transition. Yang and coworkers were able to anneal Pt@Fe 2 O 3 core-shell nanoparticles prepared via a sequential 35, 36 procedure in a reductive atmosphere to produce fct FePt alloys. Similarly intermetallic CoPt nanoparticles have been prepared via phase transition from 37, 38 disordered phases. While it is difficult to control the shape of intermetallic nanostructures other than spheres, uniform nanorods of intermetallic PtPb have been made recently through co-reduction of acetylacetonates of platinum and lead by tert-butylamine borane (TBAB) in diphenyl ether and adamantanecarboxylic acid, 17, 18 hexadecanethiol and hexadecylamine as the capping agents. The different kinetic stability of the growing low-index surfaces of a NiAs-type hexagonal phase is thought to be the driving force for the anisotropic growth. 18 Further studies revealed the growth of PtPb nanorods experienced a phase evolution, from Pt-rich cubic phase to thermodynamically stable hexagonal phase PtPb intermetallic Composition Control of Alloys in Macroscopic Miscibility Gap Narrowing or complete disappearance of miscibility gaps for nanostructured platinum alloys is fundamentally interesting, and potentially very useful for

31 Chapter 1 8 controlling the properties that cannot be achieved through bulk alloys. 39, 40 Figure 1.2 shows a representative Ag-Pt binary phase diagram, which shows a large miscibility gap between Ag 2 Pt 98 and Ag 95 Pt 5 at < ~400 C. Silver and platinum do not mix well and form bulk alloys in this range of composition. As being discussed, when the size of particle reduces to nanometer regime, contribution of free energy from surfaces can no longer be ignored. This surface energy can result in the change of phase behaviour, and the phase diagram based on bulk bimetallic systems cannot be simply applied for nanoalloys. Theoretical as well as experimental studies show that miscibility between different metals can be greatly increased with the decrease in particle size Figure 1.2 Phase diagram of bulk Ag-Pt binary system, redrawn after ref. 10. Study focusing on synthesis and property of nanostructured platinum alloys with compositions that do not exist in bulk materials is still limited. Besides PtRu, AgPt and AuPt are two of the platinum binary systems that have been 39, 45, 46 characterized recently. In the case of PtRu alloys, both structural and

32 Chapter 1 9 compositional properties can be size dependent, as evidenced by the observation of those compositions that are not expected in the bulk phase diagram. 47 In the bulk phase diagrams, Au-Pt alloys have a miscibility gap in the composition range between around Au 1 Pt 99 to Au 85 Pt 15 at 400 C or below. 10 Nanoparticles of AuPt alloys with composition in this gap have been prepared by co-reduction of HAuCl 4 and H 2 PtCl 6 with sodium borohydride in a water-toluene two-phase 45, 46 system Shape Control The shapes of platinum and its alloy nanoparticles can be controlled by both thermodynamic and kinetic factors, which are dictated by both the intrinsic structural properties of platinum and reaction systems such as, solvent, capping agent, and reducing agent. Metal nanoparticles form facets to minimize surface energy and total excess free energy. Platinum, which has a face-centered cubic (fcc) symmetry, is usually bound by three low-index planes, namely {100}, {110} and {111} surfaces. Among these three planes, the (111) planes have the lowest 48, 49 surface energy while the (110) planes have the highest one. Other reaction conditions including concentration, time and temperature are also critical. Platinum precursors can be chosen from hexachloroplatinic acid (H 2 PtCl 6 ), potassium hexachloroplatinate (K 2 PtCl 6 ), potassium tetrachloroplatinate (K 2 PtCl 4 ), and platinum acetylacetonate (Pt(acac) 2 ),

33 Chapter 1 10 depending on the choice of solvents (either water or organic liquids), reductants, surfactants, and other additives. 50 A range of reducing agents including borohydride, hydrazine, hydrogen, citrate, and ascorbic acid, can be used in aqueous systems. Meanwhile, polyols, diols and amines are commonly used reducing agents in organic systems. These reducing agents offer a fairly board range of reducing ability. Besides, many other chemicals including both organic and inorganic molecules as well as ions can be used to either passivate or activate particular surfaces and affect the natural growth habits of platinum and its alloys. In principle, one may expect to obtain platinum nanoparticles with predetermined shapes having certain low indexed planes through judicious selections of these reaction parameters. Organic capping agents play pivotal roles in several different aspects in shape control of colloidal platinum nanostructures. 49 Long carbon chains of organic capping agents are hydrophobic and have the stereo hindrance effect to prevent direct contacts among relatively high-energy surfaces of platinum, thus stabilize Pt nanoparticles from aggregation. The decrease in total excess free energy, due to the adsorption of capping agents, effectively results in preventing Pt nanoparticles from further growth and Ostwald ripening. When capping agents absorb selectively onto given platinum surfaces, the morphology of the nanocrystals can 49, 51, 52 therefore be controlled. Since platinum nanocrystals have different electronic structures and atomic arrangements for various facets, one can expect

34 Chapter 1 11 that given capping agents should absorb onto these surfaces differently. The preferred adsorption onto one set of surfaces over the others should result in different growth rates along various given crystallographic directions. The solute atoms would more likely attach to those less protected platinum surfaces, leading to an anisotropic growth. One key criterion in selecting proper capping agents for shape control is the proper interaction between the guest molecules and various platinum facets, which should have the proper equilibrium constant and facet-selected adsorption. El-Sayed and co-workers successfully prepared Pt nanocrystals with cube- and tetrahedron-like shapes by using acrylic acid and polyacrylate in a colloidal system, although in this early work both size and shape are not uniform for the particles reported in this work Even today, the synthesis of uniform and high-yield tetrahedral platinum nanoparticles remains to be a challenging issue. Tetradecyltrimethylammonium (C 14 TABr) is an effective capping agent that was thought to preferentially adsorb to {100} surfaces of platinum in aqueous 57, 58 solutions, as observed by Yang and co-workers. By adjusting the reduction rate, cubic, cubooctahedral and porous Pt nanoparticles have been obtained. When NaBH 4 was used, the reducing ability could be adjusted by changing ph values of the reaction system. 57 Furthermore, slow reduction could facilitate the formation of cubes in alkaline solutions. 57 The reduction process can be accelerated in acid solutions, leading to formation of cubooctahedral

35 Chapter 1 12 nanoparticles. 57 By using ascorbic acid as reducing agent, porous structures 57, 58 were obtained. In organic systems, oleic acid and oleylamine have routinely been used to control the growth direction of Pt nanoparticles Cubes, tripods, octapods, prisms, rods and various other shapes have been obtained in presence of these two surfactants. 59 Yang and co-workers prepared Pt cubes, tripods and multipods by applying adamantanecarboxylic acid (ACA) and hexadecylamine (HDA) as capping agents Noticeably, in many colloidal systems, the interactions between nitrogen and platinum atoms are important and amine has been a preferred functional group. Besides direct interactions with platinum surfaces, the competing binding between amine groups with other organic ligands such as adamantanecarboxylic acid results in exchange kinetics that favour the anisotropic 32, 62, 64 growth of platinum nanostructures. Inorganic ions and other small molecules, which have been overlooked for quite sometimes on their functions in shape control, play increasingly important roles in the design of shape-controlled nanoparticles of various metals including platinum Similar to the role of adsorptive organic molecules, these inorganic species show preferred adsorption to specific facets of platinum. Unlike their organic counterparts though, such interactions can either promote or 66, 68, 69 inhibit the further growth along given directions. Silver species have been found to affect the shape of Pt nanocrystals in a big

36 Chapter 1 13 way. 63, 65, 66 Silver is thought to adsorb preferentially on (100) over (111) surfaces in the form of either Ag 0 or Ag 2+ 4 and alter the growth rates along these directions. Cubic, cubooctahedral and octahedral shapes of Pt nanoparticles have been prepared by adding different amount of silver ions in the reaction mixtures. 66 Xia and co-workers have prepared Pt nanowires and other nanostructures by introducing trace amount of Fe(II) and Fe(III) ions, which mediate the reduction rates of Pt(IV) species on various surfaces of platinum differently They also applied nitric ions (NO 2 - ) in the synthesis of platinum tetrahedra and octahedra with well-defined facets. 70 In this case, nitric ions form complexes with Pt(II) and Pt(IV) species and induce the anisotropic growths preferentially along <111> directions. Copper ions have also been found to interact with Pt surfaces strongly and used to prepare platinum nanocubes. 71 Concentration of Cu (II) ion is showed to dramatic influence the size and shape of Pt. In absence of this ion spheres are formed instead of cubes. In some cases even water can be critical in the shape of platinum nanocrystals. 72 When Pt(IV) species form the hydrated complexes such as PtCl 5 (H 2 O) - and PtCl 4 (H 2 O) 2, they selectively adsorb onto Pt (111) surfaces, resulting in the formation of platinum nanocubes. In this system, the selective absorption accelerates the growth along the interacting surfaces, rather than hindrance the growth which is more commonly seen for organic capping agents. Recently, Tian et. al. have obtained Pt tetrahexahedral nanocrystals electrochemically. 73 Tetrahexahedral nanocrystal is bound by 24

37 Chapter 1 14 facets with {730}, {210}, and {520} surfaces and one of the few well-defined shapes with high-index planes. The adsorbed oxygen and hydroxyl species are thought to play the essential role for the formation of these unusual platinum nanocrystals. The total excess free energy, which is critical for the growth and shape control of particles, can be further reduced by introducing crystal defects, stacking faults in (111) plane in particular. The formation of twin planes is thermodynamically favored if the decrease in surface energy can compensate the increase in other energy terms induced by defects, which can be described by the following 49, 74 expression: γ γ + γ < γ ' (1.9) + twin strain where γ ' and γ are crystal-solution interfacial energy with and without twin planes, respectively; γ twin surface energy for the formation of twin plane, and γ strain strain energy induced by the formation of twin plane due to lattice distortion or geometric mismatch. The formation of twin plane in Pt nanoparticles can be largely attributed to the reduction of surface energy. The introduction of twin planes lowers the symmetry in crystal structures and changes the nature growth habits of Pt nanoparticles, leading to various new morphologies. Yang and his coworkers have successfully obtained monopods, bipods, planar 62, 64 tripods, and multipods of platinum based on this concept. Formation of different shapes can attribute to the number of twin defects in the seed crystals and

38 Chapter 1 15 subsequent preferred growth. Growth along three equivalent <211> directions of a seed crystal with a single (111) twin plane results in planar tripods, while the growth from a single spot of a multiply twinned crystal leads to monopods. Interestingly, while various low-dimensional rod-like platinum nanostructures form due to the twinning in (111) plane, the growth mechanisms can be quite different and careful lattice and micro-structural analyses are necessary Structure Control Multicomponent platinum nanostructures can be made using colloidal synthesis. Platinum can be deposited in either monolayer or multilayer fashion on a core particle, or served as core for the subsequent depositions of other metals or metal oxides. Platinum can also grow on selective regions of the surface of core particles in an island growth mode. Controllable epitaxial growth is essential for the growth of monolayer on metal nanoparticles. Under-potential deposition (UPD) is one of the best pathways to the fabrication of platinum monolayer on metal nanoparticles. 75 UPD has traditionally been used in the deposition of metal atoms electrochemically onto another metal substrate at a potential slightly less negative than the equilibrium 76, 77 potential for its reduction to metal. It usually occurs when the interaction between electrodepositing and substrate metal atoms is stronger than those in pure metal. In many cases, atoms can only deposit up to a single monolayer, after

39 Chapter 1 16 which a continuing deposition is no longer energetically favourable. When platinum nanostructures are used as the support metal, several metals including Ru, Pd and Ag can be directly deposited as monolayer While UPD method cannot be directly applied to prepare platinum monolayer on designed metals directly due to the mismatch of redox potentials. 81 In this approach, a sacrificial monolayer such as copper is first deposited on a predetermined metal using the UPD method, followed by a redox replacement reaction by platinum-containing ions. This approach successfully broadens the possible metal elements that can be directly or indirectly deposited by applying the UPD technique. Platinum monolayer has been grown on various metal nanoparticles using this two-step approach Small nanoparticles can release surface energy through interaction with other species, which can alter the subsequent nucleation and growth of clusters. The heterogeneous nucleation generally has a smaller critical energy ( G * hetero ) barrier than that for homogeneous one while they have similar critical nucleus radius (r * ) (Figure 1.3a). If sufficient sites are available for heterogeneous nucleation, both number of critical clusters (N r* ) and nucleation rate (dn r* /dt) should be larger for heterogeneous nucleation than those for homogeneous one (Figure 1.3b). In another word, the solute atoms nucleate and grow heterogeneously. One strategy is to introduce large amount of core particles

40 Chapter 1 17 Figure 1.3 Schematic illustration of (a) overall excess free energy G r and (b) average number of clusters N r as a function of cluster size r for heterogeneous and homogeneous nucleation. into reaction mixtures that use mild reducing agents or small amount of precursors. Platinum core-shell nanostructures of Au@Pt, Pd@Pt, 93 Cu@Pt, 94 Pt@Cu, 94 and Pt@Fe 2 O 3 95 have been prepared using this method. In these systems, platinum serves as either core or shell metal. For instance, to make Au@Pt core-shell nanoparticles, gold nanoparticles were first generated in an aqueous solution, followed by the addition of platinum precursors using hydrogen as the reducing agent Ru@Pt core-shell nanoparticles were synthesized by a two-step method, 96 where ruthenium acetylacetonate and platinum chloride were added sequentially and reduced in polyols. Recently, Yang and co-workers have successfully controlled the shape of Pt@Pd core-shell nanoparticles using

41 Chapter 1 18 tetradecyltrimethylammonium bromide (TTAB) and NO 2 following an epitaxial growth of Pd on Pt nanocubes. 97 In this system, while the seeding crystals of platinum had cubic shape, growth of palladium resulted in not only cubes bound by {100} planes, but also octahedra bound by {111} planes, and cuboctahedra, a truncated shape. Several other solution phase methods have also been reported for making platinum core-shell nanoparticles. Pt@Pd core-shell nanoparticles were produced electrochemically and stabilized by tetrabutylammonium bromide. 98 These nanoparticles were prompt to oxidation and one metal was enriched by the other. Colloidal Pt@Ru core-shell nanoparticles were obtained sonochemically from their corresponding metal precursors in aqueous solutions. 99 Both particle size and reaction rate were sensitive to the stabilizers used. The diameters of particles produced were between 5 and 10 nm when sodium dodecyl sulphate (SDS) was used, and around 5 nm when polyvinyl-2-pyrrolidone (PVP) was used. The radiolytic reduction method was applied in the synthesis of Au@Pt and Pt@Au core-shell nanoparticles. 100 Solvated electron and radical produced by γ-ray irradiation of the solvents reduced the metal precursors. In some cases, platinum core-shell nanostructure can be obtained even if two 99, 101 metal precursors are mixed from the very beginning. This phenomenon usually occurs in reaction mixtures with mild reducing agents and is reasonable, since reduction of metal precursors happens sequentially depending on the

42 Chapter 1 19 difference in the standard reduction potentials of the metal ions. An example is the Pt-Ru system where co-reduction of Pt and Ru actually leads to Pt@Ru clusters due to the high reduction potential of Pt. 99 Growth of epitaxial layers can sometimes follow either Volmer-Weber (VW, or island growth) or Stranski-Krastanov (SK, or island-on-wetting-layer growth) mode to form heteronanostructures, depending on the interplay among surface, interfacial and strain energies. Pt-on-Au heteronanostructures have also been reported, where the growth of Pt seemed to follow a SK mode Palladium was found to overgrow on cubic Pt nanocrystals upon reduction of its precursors by ascorbic acid, while Pt@Pd core-shell nanoparticles formed under different conditions. 97 The Pt nanowires on metal gauze is another type of heteronanostructures. 69 Platinum atoms are thought to nucleate at the defect sites of the gauze. By slowing down the reducing rate with Fe(II) or Fe(III), Platinum nuclei grow into nanowires instead of particles in the presence of PVP. Some Pt-based metal-oxide heteronanostructures have also been made. FePt-iron oxide heterodimers are prepared using a sequential method, where FePt nanoparticles are synthesized in advance separately. 105 Pt nanoparticles and nanowires could also grow directly on TiO 2 nanofibers Platinum Alloy Nanoparticles as PEMFC Electrocatalysts In recent years, with the increasing environmental concerns and accelerated

43 Chapter 1 20 depletion of fossil fuels, there are resurgent research interests for the development of new technology using alternative energy sources other than fossil fuel, most noticeably hydrogen in polymer electrolyte membrane (or proton exchange membrane) fuel cells (PEMFCs) for automobiles PEMFCs are devices that directly convert chemical energy stored in fuel molecules into electric energy. During operations, oxygen gas is fed from cathode and electrochemically reduced while fuel molecules with low standard redox potential are electrochemically oxidized at anode. Based on the type of fuels used, PEMFCs can be further categorized into direct hydrogen fuel cells (DHFC), direct methanol fuel cell (DMFC), direct formic acid fuel cell (DFAFC) and so on. While PEMFCs can have much higher efficiency than the combustion of fossil fuels and are not limited by the Carnot cycle in terms of energy conversion, in practice overpotentials existing at both electrodes can dramatically reduce the overall efficiency. Slow reaction kinetics contributes greatly to the overall overpotential in a cell and development of high efficient electrocatalysts is pivotal. Platinum is the most widely used PEMFC electrocatalyst among pure metals, 110 and its alloys show much improved catalytic activities because of the favourable electronic and geometric structures. Besides size, both composition and shape can play critical roles for the observed enhancement in activity. Depending on the type of fuel cells, the critical reactions can be at either cathode or anode half-cell, and the emphasis of catalyst developments varies accordingly.

44 Chapter Reaction Mechanism In a PEMFC, oxygen is electrochemically reduced on catalytic surfaces through oxygen reduction reaction (ORR) at the cathode. The reaction typically follows either four-electron or two-electron reduction pathway, both of which undergo 111, 112 several elementary steps and involve multiple electrons (Figure 1.4a). Water is produced directly when oxygen is reduced through the four-electron mechanism. The two-electron reduction, on the other hand, occurs at low potentials and produces hydrogen peroxide as the intermediate. In general, the four-electron route is more efficient than the two-electron process for ORR. Figure 1.4 Reaction pathways of (a) oxygen reduction (ORR), (b) hydrogen oxidation (HOR), (c) methanol oxidation (MOR) and (d) formic acid oxidation reactions (FAOR).

45 Chapter 1 22 Various chemicals including hydrogen, methanol, formic acid and even ethanol, can be fed as fuels at anode of PEMFCs. When hydrogen is used as the choice of fuel, the anodic reaction has a small overpotential, because H-H bond is relatively easy to break catalytically and the corresponding HOR involves only two electrons (Figure 1.4b). 113 For hydrogen fuel cells, the sluggish kinetic is mostly attributed to ORR, which involves simultaneously absorbed oxygen, 107, 114 hydrated proton, and electron on active surface sites. Complete oxidation of methanol, however, involve six electrons and requires careful selection of electrocatalysts to decrease its large overpotential at the anodic half cell for MOR 115, 116 (Figure 1.4c). The activation of C-H bond is a key step and thought to happen ahead of O-H bond due to the relatively low bonding energy. A dual pathway, analogous to that for ORR, has been proposed for MOR, in which methanol can be oxidized either directly to CO 2 or through the adsorbed CO ads intermediate. 114 Similarly, electrochemical oxidation of formic acid can also have two major possible pathways, namely the dehydrogenation process to form 117, 118 CO 2 directly, or via a dehydration step (Figure 1.4d) Current Status As expensive platinum group metals (PGM) are essential elements to ensure outstanding catalytic properties, reducing the high cost of catalysts is one of the 107, 119 major challenges for wide-spread applications of PEMFCs. The cost issue

46 Chapter 1 23 can be addressed in part by improving the specific activity and durability of electrocatalysts and decreasing the usage of PGM in fuel cells. A range of design parameters, such as composition of alloy, heteronanostructure and support may all be worthwhile to be explored or further examined. A few recent examples will be discussed to show some of the new developments of advanced electrocatalysts. Improving the sluggish kinetics of ORR using new catalysts is critical in the 107, 120, 121 development of PEM fuel cells. Although carbon-supported platinum nanoparticles are still used predominantly as cathodic catalysts for ORR, increasing amount of research efforts have been put on new electrocatalysts of structurally well-defined platinum alloys. 107 One thrust is on Pt 3 M (M=Ni, Co, Fe, Ti, V) alloys which show substantial improvement in specific activity for ORR 119, 122 compared with platinum. Specifically, a recent study indicates that Pt 3 Ni (111) surface is about ninety-time more active than the state-of-art Pt/C catalysts for ORR. 119 The difference in catalytic activity among the three low-indexed planes of Pt 3 Ni surfaces is substantial, and can be attributed to the changes of d-band center ( ε d ) and surface atom arrangement of platinum and decreases in coverage of non-reactive oxygenated species (OH ad ) due to the incorporation of nickel atoms. 119 Platinum monolayer or skin layer on various metal supports has also shown enhanced activities for ORR Dramatic increase of ORR activity is

47 Chapter 1 24 observed when Pd (111) surface is used as the substrate and attributed to the change in ε d of platinum monolayer induced by its interactions with underlining Pd metal. 124 The activity can be further increased by using mixed monolayers of platinum and another metal, where atomic geometry on surface plays important roles as well. 126 In general, Pt content can be reduced to a very low level by using monolayer, and still possesses five to eight times or even higher mass-specific activity than Pt/C. 125 Recently, Zhang et al. demonstrated that electrocatalysts of Pt nanoparticles on carbon supports could be stabilized by small gold clusters. 127 These Au-modified Pt catalysts showed insignificant changes in activity and in surface area after over 30,000 cycles. This observation is in sharp contrast to heavy losses observed 108, 109 with carbon supported Pt catalysts tested under the same conditions. This work suggests that advanced electrocatalysts with excellent durability and low degradation rate could be achieved by using heterogeneous nanostructures. Unlike hydrogen fuel cell, the anodic half-cell reactions have relatively large overpotential and slow kinetics in DMFCs. 128 The CO ads -like intermediates generated during MOR interact strongly with platinum atoms and can only be 115, 116 oxidized at potentials much higher than the working range for DMFCs. The CO ads -like molecules can accumulate on and block the active sites, and subsequently reduce the activity of platinum catalysts. Removal of adsorbed poisoning intermediates from active sites is necessary to achieve reasonable

48 Chapter 1 25 activity for MOR. Carbon-supported PtRu alloy nanoparticles are the most widely used electrocatalysts, as they have much lower on-set potentials and higher activity than platinum. 129 The enhanced performance by incorporation of Ru can be explained using the dual function model. First, Ru atoms strongly adsorb hydroxyl species (OH ads ), which can oxidize the CO ads species adsorbed on Pt 115, 129 sites, leading to a recovery of active surfaces. Second, both surface atomic arrangement and electronic structure of Pt atoms are modified by the neighbouring Ru atoms, resulting in an improvement in surface catalytic reaction. Recently, Abruna, DiSalvo and co-workers reported exceptionally high MOR electrocatalytic activity over PtPb intermetallic surfaces The current density increased by 1-2 orders of magnitude and the on-set potential was also reduced by 100 mv in comparison with pure Pt. 132 No detectable adsorbed CO intermediates were observed in the in-situ electrochemical study and Fourier transform infrared (FTIR) measurement. 133 They concluded that electro-oxidation of methanol on PtPb did not undergo the pathway that produce adsorbed CO ads. Yang and co-workers have prepared uniform hexagonal phase intermetallic PtPb nanorods in diphenyl ether. 18 The colloidal PtPb nanorods in disordered face-center cubic phase and with different compositions could be made. 17 Much improved activities in catalysing MOR were observed for sintered PtPb nanoparticles. Similar to MOR, poisoning caused by CO ads intermediate species during

49 Chapter 1 26 electro-oxidation of formic acid is a major issue related to observed slow reaction 117, 118 kinetics in some catalytic systems. Accumulation of CO ads intermediates on surfaces of pure platinum metal is commonly observed for formic acid oxidation reactions. Both PtPd alloys and palladium metal however, show superior performance due to the depressed dehydration pathway because of the 118, 134, 135 electronic effect. Nanoparticles of palladium overgrown on Pt also 97, 103 exhibit much less poisoning effect than pure platinum. Both PtAu alloy nanoparticles 136 and Pt-modified Au nanoparticles 137 have very good performance toward the oxidation of formic acid as well. 1.3 Development of Next Generation PEMFC Catalysts Deterministic Factors in Electrocatalysis The design of fuel cell catalysts aims at creating surfaces that balance the adsorption of reactants and desorption of products, and at the same time lower the activation energy of reactions. Both theoretical and experimental studies have shown that electronic structure and surface atom arrangement are two key factors to determine the performance of electrocatalysts. 107 Nøskov and co-workers have calculated d-band centers of transition metals ( ε d, relative to the Fermi level) and correlated them with their ability of chemisorptions and catalytic activity It is found that d-band center ε d and chemisorption energy E chem has a linear relationship. This relation partially explains the

50 Chapter 1 27 volcano-shaped curve for the relationship between activity in ORR and d-band center of electrocatalysts. 124 When the d-band center is narrow, the chemisorption of reactant molecules to catalytic surface becomes easy, but the product molecules can be hard to remove from the surface. The accumulation of product molecules on catalyst surface blocks the active sites and slows down the overall reaction. One the other hand, few reactant molecules can be absorbed onto the catalyst surfaces if d-band center is deep. The optimal d-band center should be located in between these two extreme conditions. Besides electronic consideration, the surface structure is another important factor, since multiple active sites are usually involved in catalytic reaction. 141 Proper surface atomic arrangement is essential to achieve optimal activities. Such geometric effect can be sensitive to relatively small changes and configurations other than the optimal often results in sharply decreased activity. Surface structures are generally different from those in bulk solids because of the reconstruction. While computational simulation has improved greatly in recent years in understanding the possible surface configurations and electronic structures of transition metals and their alloys, design of superior electrocatalysts can be best achieved by combining the theoretical tools with experimental exploration Design of Advanced Platinum Alloy Nanoparticles as Fuel Cell

51 Chapter 1 28 Catalysts Based on the above discussions, optimizations in electronic structure and surface atom arrangement, both of which vary with facets of nanocrystals, are required for good performance of catalysts. By controlling the shape of pure metal nanoparticles, we could tune their catalytic properties. Because electronic structure and surface geometry are usually correlated, the best performance cannot be achieved sometimes for pure individual metals. Alloys bimetallic compounds in particular, have been widely studied. 1 Good catalysts may be realized by finely controlling the composition and shape of alloys. Another interesting group of catalysts is Pt-based multifunctional nanostructures, which include 90-92, 97, , 142 core-shell, dumbbell, and particle-on-particle structures. The electronic structures could be finely tuned by coupling with selected nanostructures while the surface atom geometry is maintained. Figure 1.5 lists some possible Pt-based structures. The shape of Pt nanoparticles can be controlled by using selected capping agents. The concept of homogeneous nucleation can be instructive for the formation of alloyed nanoparticles. For complex structures, Pt can be deposited in either monolayer or multiplayer fashion on a core particle, or served as core for the subsequent depositions of other metals or metal oxides. Pt can also grow on selective regions of the surface of core particles to form islands. Three different growth modes, Frank-van der Merwe, Volmer-Weber, and Stranski-Krastanov, can help to understand these structures.

52 Chapter 1 29 Figure 1.5 Schematic illustration of various advanced Pt-based nanostructures. 1.4 Objective and Overview of Thesis This PhD thesis is focusing on the preparation of novel Pt-based alloys and nanostructures, and the study of their electrocatalytic properties. The main objective is to explore efficient and cheap fuel cell catalysts, which can be addressed in part by improving their specific activity and durability and by reducing the amount of platinum used in fuel cells. In Chapter 2, preparation of PtAg alloy nanostructures in a broad composition range will be presented. Although silver platinum binary alloys with compositions between about Ag 2 Pt 98 and Ag 95 Pt 5 at <~ 400 C have largely not been observed in bulk due to the large immiscibility, they can form solid solution at the nanometer-scale region. The results indicate that lattice parameter changes almost linearly with composition in these Ag-Pt nanomaterials. In another word, lattice

53 Chapter 1 30 parameter and composition relationship follows the Vegard s law, which is a strong indication for the formation of metal alloys. The synthesis and formation of PtAg alloy nanowires in the presence of oleylamine and oleic acid are discussed based on the concept of oriented attachment, which is shown to be composition dependent. Worm-like nanowires can occur largely at the composition around Pt 50 Ag 50 while only sphere-like or faceted nanoparticles form in both Pt and Ag rich alloy regions under the same reaction conditions. Density functional theory (DFT) calculation is used to understand the interactions between the functional groups of capping agents and low index planes of PtAg alloys. The structural order of interfaces after collision between primary particles is obtained by molecular dynamic (MD) simulation. The experimental data and simulation results indicate that the formation of alloy nanowires is mostly driven by the interplay between the binding energy of capping agents on alloy surfaces and the diffusion of atoms at the interface upon the collision of primary nanoparticles. I will describe a new approach to the synthesis of Pt-surface rich heterogeneous alloy nanoparticles through selective electrochemical dissolution of silver from PtAg alloy nanoparticles. The surface and bulk compositions, size and architecture of these heterogeneous nanostructures are controlled synergistically by changing the upper limits of cycling potentials in perchloric acid (HClO 4 ) aqueous solution to dissolve silver deterministically. These Pt-surface rich PtAg nanostructures exhibit much higher activity than Pt

54 Chapter 1 31 nanoparticles and good durability as electrocatalyst in formic acid oxidation reaction (FAOR). In Chapter 3-5, the electrocatalytic study of Pt-on-M (M=Ag, Au, Cu, Pd) and their derived nanostructures is presented. A series of such Pt-on-M heterogeneous nanostructures have been made using a sequential synthetic method and by intentionally introducing abundant metal nanoparticles to promote the heterogeneous nucleation of Pt on their surface. The growth of Pt-on-M nanoparticles has been controlled by choosing mild reducing agent and proper reaction temperature ranges. The Pt-on-Pd catalysts exhibit both enhanced activity and much improved stability in oxygen reduction reaction. Other platinum alloy nanostructures have been made from these Pt-on-M nanoparticles. For instance, Pt-Au bimetallic nanoparticles have been produced by thermal treatment and shown higher activity than Pt in catalyzing formic acid oxidation reaction. Other nanostructures such as Pt hollow nanospheres and cubic nanoboxes can also be obtained by an electrochemical approach. These new forms of Pt nanostructures exhibit improvement in ORR and methanol oxidation reaction activities. In Chapter 6, I will describe a generic chemical dealloying method for making ultra-small metal and metal alloy nanoparticles on various supports including C, SiC and TiO 2. A series of 3-4 nm alloy nanoparticles with nominal composition of Pt 1 Ag 30, Au 1 Ag 30, Pd 1 Ag 30 have been prepared using colloidal synthetic methods and subsequently loaded onto the supports. After the dealloying

55 Chapter 1 32 process in solution silver can be removed, resulting in the formation of 1-nm pure noble metal nanoparticles of Pt, Au or Pd. The supported tiny platinum metal nanoparticles are highly active in catalytic reduction of p-nitrophenol. Author Contributions: This chapter is in part based on the following review article (Z. M. Peng, H. Yang, Designer Platinum Nanoparticles: Shape, Composition in Alloys, Nanostructure and Electrocatalytic Property, Nano Today, 2009, 4, ). Professor Hong Yang and Zhenmeng Peng conceived and co-wrote the paper. 1.5 References (1) Porter, D. A.; Easterling, K. E., Phase Transformtions in Metals and Alloys. 2nd ed.; Chapman & Hall: London, (2) Sauthoff, G., Intermetallics. VCH-Wiley: Weinheim, (3) Mullin, J. W., Crystallization. 3rd ed.; Oxford University Press: Oxford, (4) Pfeiler, W., Alloy Physics: a Comprehensive Reference. Wiley-VCH: Weinheim, (5) Ferrando, R.; Jellinek, J.; Johnston, R. L. Chem. Rev. 2008, 108, 845. (6) Park, J.; Joo, J.; Kwon, S. G.; Jang, Y.; Hyeon, T. Angew. Chem.-Int. Edit. 2007, 46, (7) Volmer, M.; Weber, A. Z. Phys. Chem. 1926, 119, 227. (8) Volmer, M., Kinetic der Phasenbildung. Steinfopff, Leipzig: (9) Becker, R.; Doring, W. Ann. Phys. 1935, 24, 719.

56 Chapter 1 33 (10) Massalski, T. B.; Murray, J. L.; Bennett, L. H.; Baker, H., Binary Alloy Phase Diagrams. American Society for Metals: Metals Park, Ohio, (11) Kumbhar, A.; Spinu, L.; Agnoli, F.; Wang, K. Y.; Zhou, W. L.; O'Connor, C. J. IEEE Trans. Magn. 2001, 37, (12) Hyun, M. S.; Kim, S. K.; Lee, B.; Peck, D.; Shul, Y.; Jung, D. Catal. Today 2008, 132, 138. (13) Shimazaki, Y.; Kobayashi, Y.; Yamada, S.; Miwa, T.; Konno, M. J. Colloid Interface Sci. 2005, 292, 122. (14) Deivaraj, T. C.; Lee, J. Y. J. Power Sources 2005, 142, 43. (15) Wang, W. H.; Tian, X. L.; Chen, K.; Cao, G. Y. Colloid Surf. A-Physicochem. Eng. Asp. 2006, 273, 35. (16) Bonnemann, H.; Richards, R. M. Eur. J. Inorg. Chem. 2001, (17) Yang, S. C.; Peng, Z. M.; Yang, H. Adv. Funct. Mater 2008, 18, 1. (18) Maksimuk, S.; Yang, S. C.; Peng, Z. M.; Yang, H. J. Am. Chem. Soc. 2007, 129, (19) Travitsky, N.; Ripenbein, T.; Golodnitsky, D.; Rosenberg, Y.; Burshtein, L.; Peled, E. J. Power Sources 2006, 161, 782. (20) Solla-Gullon, J.; Rodes, A.; Montiel, V.; Aldaz, A.; Clavilier, J. J. Electroanal. Chem. 2003, 554, 273. (21) Solla-Gullon, J.; Montiel, V.; Aldaz, A.; Clavilier, J. Electrochem. Commun. 2002, 4, 716. (22) Lide, D. R., CRC Handbook of Chemistry and Physics, 88th Edition. CRC Press/Taylor and Francis: Boca Raton, (23) Ahrenstorf, K.; Albrecht, O.; Heller, H.; Kornowski, A.; Gorlitz, D.; Weller, H. Small 2007, 3, 271. (24) Elkins, K. E.; Vedantam, T. S.; Liu, J. P.; Zeng, H.; Sun, S. H.; Ding, Y.; Wang, Z. L. Nano Lett. 2003, 3, (25) Liu, Z. F.; Ada, E. T.; Shamsuzzoha, M.; Thompson, G. B.; Nikles, D. E.

57 Chapter 1 34 Chem. Mat. 2006, 18, (26) Lee, Y. H.; Lee, G.; Shim, J. H.; Hwang, S.; Kwak, J.; Lee, K.; Song, H.; Park, J. T. Chem. Mat. 2006, 18, (27) Sun, S. H. Adv. Mater. 2006, 18, 393. (28) Chen, M.; Kim, J.; Liu, J. P.; Fan, H. Y.; Sun, S. H. J. Am. Chem. Soc. 2006, 128, (29) Chen, M.; Liu, J. P.; Sun, S. H. J. Am. Chem. Soc. 2004, 126, (30) Sun, S. H.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Science 2000, 287, (31) Shevchenko, E. V.; Talapin, D. V.; Rogach, A. L.; Kornowski, A.; Haase, M.; Weller, H. J. Am. Chem. Soc. 2002, 124, (32) Shevchenko, E. V.; Talapin, D. V.; Schnablegger, H.; Kornowski, A.; Festin, O.; Svedlindh, P.; Haase, M.; Weller, H. J. Am. Chem. Soc. 2003, 125, (33) Guo, Q. J.; Teng, X. W.; Yang, H. Adv. Mater. 2004, 16, (34) Jeong, U.; Teng, X. W.; Wang, Y.; Yang, H.; Xia, Y. N. Adv. Mater. 2007, 19, 33. (35) Teng, X. W.; Yang, H. J. Am. Chem. Soc. 2003, 125, (36) Teng, X. W.; Yang, H. Nanotechnology 2005, 16, S554. (37) Rossi, G.; Ferrando, R.; Mottet, C. Faraday Discuss. 2008, 138, 193. (38) An, B. H.; Wu, J. H.; Liu, H. L.; Ko, S. P.; Ju, J. S.; Kim, Y. K. Colloid Surf. A-Physicochem. Eng. Asp. 2008, 313, 250. (39) Peng, Z. M.; Yang, H. J. Solid State Chem. 2008, 181, (40) Faraday Discuss.-Special Issue on Nanoalloys: From Theory to Application 2008, 138, 1. (41) Shirinyan, A. S.; Gusak, A. M.; Wautelet, M. Acta Mater. 2005, 53, (42) Wautelet, M.; Dauchot, J. P.; Hecq, M. Nanotechnology 2000, 11, 6. (43) Ouyang, G.; Tan, X.; Wang, C. X.; Yang, G. W. Nanotechnology 2006, 17,

58 Chapter (44) Mori, H.; Yasuda, H. Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 2001, 312, 99. (45) Luo, J.; Maye, M. M.; Petkov, V.; Kariuki, N. N.; Wang, L. Y.; Njoki, P.; Mott, D.; Lln, Y.; Zhong, C. J. Chem. Mat. 2005, 17, (46) Bond, G. C. Platinum Metals Review 2007, 51, 63. (47) Hills, C. W.; Mack, N. H.; Nuzzo, R. G. J. Phys. Chem. B 2003, 107, (48) Yang, H.; Teng, X. W.; Maksimuk, S., Multipods and Dendritic Nanoparticles of Platinum: Colloidal Synthesis and Electrocatalytic Property. In Metal Nanoclusters in Catalysis and Materials Science: the Issue of Size Control. Part II: Methodologies, Corain, B.; Schmid, C. G.; Toshima, N., Eds. Elsevier: Amsterdam, 2008; pp 307. (49) Tao, A. R.; Habas, S.; Yang, P. D. Small 2008, 4, 310. (50) Peng, Z. M.; Yang, S. C.; Yang, H., Approaches to the Synthesis and Characterization of Spherical and Anisotropic Platinum Nanomaterials. In Metallic and Metal Oxides Nanomaterials, Kumar, C., Ed. VCH-Wiley Verlag: Weinheim, 2008; Vol. 1. (51) Milliron, D. J.; Hughes, S. M.; Cui, Y.; Manna, L.; Li, J. B.; Wang, L. W.; Alivisatos, A. P. Nature 2004, 430, 190. (52) Peng, X. G.; Manna, L.; Yang, W. D.; Wickham, J.; Scher, E.; Kadavanich, A.; Alivisatos, A. P. Nature 2000, 404, 59. (53) Ahmadi, T. S.; Wang, Z. L.; Green, T. C.; Henglein, A.; ElSayed, M. A. Science 1996, 272, (54) Ahmadi, T. S.; Wang, Z. L.; Henglein, A.; ElSayed, M. A. Chem. Mat. 1996, 8, (55) Petroski, J. M.; Wang, Z. L.; Green, T. C.; El-Sayed, M. A. J. Phys. Chem. B 1998, 102, (56) Petroski, J. M.; Green, T. C.; El-Sayed, M. A. J. Phys. Chem. A 2001, 105,

59 Chapter (57) Lee, H.; Habas, S. E.; Kweskin, S.; Butcher, D.; Somorjai, G. A.; Yang, P. D. Angew. Chem.-Int. Edit. 2006, 45, (58) Bratlie, K. M.; Lee, H.; Komvopoulos, K.; Yang, P. D.; Somorjai, G. A. Nano Lett. 2007, 7, (59) Ren, J.; Tilley, R. D. Small 2007, 3, (60) Ren, J. T.; Tilley, R. D. J. Am. Chem. Soc. 2007, 129, (61) Wang, C.; Daimon, H.; Lee, Y.; Kim, J.; Sun, S. J. Am. Chem. Soc. 2007, 129, (62) Maksimuk, S.; Teng, X. W.; Yang, H. Phys. Chem. Chem. Phys. 2006, 8, (63) Teng, X. W.; Yang, H. Nano Lett. 2005, 5, 885. (64) Maksimuk, S.; Teng, X.; Yang, H. J. Phys. Chem. C 2007, 111, (65) Rioux, R. M.; Song, H.; Grass, M.; Habas, S.; Niesz, K.; Hoefelmeyer, J. D.; Yang, P.; Somorjai, G. A. Top. Catal. 2006, 39, 167. (66) Song, H.; Kim, F.; Connor, S.; Somorjai, G. A.; Yang, P. D. J. Phys. Chem. B 2005, 109, 188. (67) Chen, J. Y.; Herricks, T.; Xia, Y. N. Angew. Chem.-Int. Edit. 2005, 44, (68) Lee, E. P.; Chen, J. Y.; Yin, Y. D.; Campbell, C. T.; Xia, Y. N. Adv. Mater. 2006, 18, (69) Lee, E. P.; Peng, Z. M.; Cate, D. M.; Yang, H.; Campbell, C. T.; Xia, Y. J. Am. Chem. Soc. 2007, 129, (70) Herricks, T.; Chen, J. Y.; Xia, Y. N. Nano Lett. 2004, 4, (71) Qu, L. T.; Dai, L. M.; Osawa, E. J. Am. Chem. Soc. 2006, 128, (72) Yu, Y. T.; Wang, J.; Zhang, J. H.; Yang, H. J.; Xu, B. Q.; Sun, J. C. J. Phys. Chem. C 2007, 111, (73) Tian, N.; Zhou, Z. Y.; Sun, S. G.; Ding, Y.; Wang, Z. L. Science 2007, 316,

60 Chapter (74) Hofmeister, H. Cryst. Res. Technol. 1998, 33, 3. (75) Leiva, E. Electrochim. Acta 1996, 41, (76) Kolb, D. M.; Przasnys.M; Gerische.H. J. Electroanal. Chem. 1974, 54, 25. (77) Gerische.H; Kolb, D. M.; Przasnys.M. Surf. Sci. 1974, 43, 662. (78) Chrzanowski, W.; Wieckowski, A. Langmuir 1997, 13, (79) Attard, G. A.; Bannister, A. J. Electroanal. Chem. 1991, 300, 467. (80) Domke, K. F.; Xiao, X. Y.; Baltruschat, H. Phys. Chem. Chem. Phys. 2008, 10, (81) Brankovic, S. R.; Wang, J. X.; Adzic, R. R. Surf. Sci. 2001, 474, L173. (82) Park, S.; Yang, P. X.; Corredor, P.; Weaver, M. J. J. Am. Chem. Soc. 2002, 124, (83) Kim, Y. G.; Kim, J. Y.; Vairavapandian, D.; Stickney, J. L. J. Phys. Chem. B 2006, 110, (84) Huang, M. H.; Jin, Y. D.; Jiang, H. Q.; Sun, X. P.; Chen, H. J.; Liu, B. F.; Wang, E. K.; Dong, S. J. J. Phys. Chem. B 2005, 109, (85) Tang, H.; Chen, J. H.; Wang, M. Y.; Nie, L. H.; Kuang, Y. F.; Yao, S. Z. Appl. Catal. A-Gen. 2004, 275, 43. (86) Jin, Y. D.; Shen, Y.; Dong, S. J. J. Phys. Chem. B 2004, 108, (87) Sasaki, K.; Mo, Y.; Wang, J. X.; Balasubramanian, M.; Uribe, F.; McBreen, J.; Adzic, R. R. Electrochim. Acta 2003, 48, (88) Vukmirovic, M. B.; Zhang, J.; Sasaki, K.; Nilekar, A. U.; Uribe, F.; Mavrikakis, M.; Adzic, R. R. Electrochim. Acta 2007, 52, (89) Adzic, R. R.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.; Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Top. Catal. 2007, 46, 249. (90) Zhao, D.; Xu, B. Q. Phys. Chem. Chem. Phys. 2006, 8, (91) Zhao, D.; Xu, B. Q. Angew. Chem.-Int. Edit. 2006, 45, 4955.

61 Chapter 1 38 (92) Park, I. S.; Lee, K. S.; Jung, D. S.; Park, H. Y.; Sung, Y. E. Electrochim. Acta 2007, 52, (93) Wang, Y.; Toshima, N. J. Phys. Chem. B 1997, 101, (94) Zhou, S. H.; Varughese, B.; Eichhorn, B.; Jackson, G.; McIlwrath, K. Angew. Chem. -Int. Edit. 2005, 44, (95) Teng, X. W.; Black, D.; Watkins, N. J.; Gao, Y. L.; Yang, H. Nano Lett. 2003, 3, 261. (96) Alayoglu, S.; Nilekar, A. U.; Mavrikakis, M.; Eichhorn, B. Nat. Mater. 2008, 7, 333. (97) Habas, S. E.; Lee, H.; Radmilovic, V.; Somorjai, G. A.; Yang, P. Nat. Mater. 2007, 6, 692. (98) Kolb, U.; Quaiser, S. A.; Winter, M.; Reetz, M. T. Chem. Mat. 1996, 8, (99) Vinodgopal, K.; He, Y. H.; Ashokkumar, M.; Grieser, F. J. Phys. Chem. B 2006, 110, (100) Henglein, A. J. Phys. Chem. B 2000, 104, (101) Belloni, J.; Mostafavi, M.; Remita, H.; Marignier, J. L.; Delcourt, M. O. New J. Chem. 1998, 22, (102) Zhou, S. G.; McIlwrath, K.; Jackson, G.; Eichhorn, B. J. Am. Chem. Soc. 2006, 128, (103) Lee, H. J.; Habas, S. E.; Somorjai, G. A.; Yang, P. D. J. Am. Chem. Soc. 2008, 130, (104) Pellegrino, T.; Fiore, A.; Carlino, E.; Giannini, C.; Cozzoli, P. D.; Ciccarella, G.; Respaud, M.; Palmirotta, L.; Cingolani, R.; Manna, L. J. Am. Chem. Soc. 2006, 128, (105) Figuerola, A.; Fiore, A.; Di Corato, R.; Falqui, A.; Giannini, C.; Micotti, E.; Lascialfari, A.; Corti, M.; Cingolani, R.; Pellegrino, T.; Cozzoli, P. D.; Manna, L. J. Am. Chem. Soc. 2008, 130, 1477.

62 Chapter 1 39 (106) Formo, E.; Peng, Z. M.; Lee, E.; Lu, X. M.; Yang, H.; Xia, Y. N. J. Phys. Chem. C 2008, 112, (107) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, 9. (108) Yu, X. W.; Ye, S. Y. J. Power Sources 2007, 172, 133. (109) Yu, X. W.; Ye, S. Y. J. Power Sources 2007, 172, 145. (110) Norskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard, T.; Jonsson, H. J. Phys. Chem. B 2004, 108, (111) Markovic, N.; Gasteiger, H.; Ross, P. N. J. Electrochem. Soc. 1997, 144, (112) Gattrell, M.; MacDougall, B., Reaction mechanism of the O 2 reduction/evolution reaction. In Handbook of Fuel Cells - Fundamentals Technology and Applications, Vielstich, W.; Lamm, A.; Gasteiger, H. A., Eds. John Wiley & Sons. Ltd.: West Sussex, England, 2003; Vol. 2. (113) Breiter, M. W., Reaction mechanism of the H 2 oxidation/evolution reaction. In Handbook of Fuel Cells - Fundamentals Technology and Applications, Vielstich, W.; Lamm, A.; Gasteiger, H. A., Eds. John Wiley & Sons. Ltd.: West Sussex, England, 2003; Vol. 2. (114) Lipkowski, J.; Ross, P. N., Electrocatalysis. Wiley-VCH: New York, (115) Hamnett, A. Catal. Today 1997, 38, 445. (116) Wasmus, S.; Kuver, A. J. Electroanal. Chem. 1999, 461, 14. (117) Park, S.; Xie, Y.; Weaver, M. J. Langmuir 2002, 18, (118) Rice, C.; Ha, R. I.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.; Barnard, T. J. Power Sources 2002, 111, 83. (119) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (120) Koh, S.; Strasser, P. J. Am. Chem. Soc. 2007, 129, (121) Srivastava, R.; Mani, P.; Hahn, N.; Strasser, P. Angew. Chem.-Int. Edit.

63 Chapter , 46, (122) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.; Wang, G. F.; Ross, P. N.; Markovic, N. M. Nat. Mater. 2007, 6, 241. (123) Zhang, J.; Lima, F. H. B.; Shao, M. H.; Sasaki, K.; Wang, J. X.; Hanson, J.; Adzic, R. R. J. Phys. Chem. B 2005, 109, (124) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Angew. Chem.-Int. Edit. 2005, 44, (125) Zhang, J.; Mo, Y.; Vukmirovic, M. B.; Klie, R.; Sasaki, K.; Adzic, R. R. J. Phys. Chem. B 2004, 108, (126) Zhang, J. L.; Vukmirovic, M. B.; Sasaki, K.; Nilekar, A. U.; Mavrikakis, M.; Adzic, R. R. J. Am. Chem. Soc. 2005, 127, (127) Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, 220. (128) Olah, G. A.; Goeppert, A.; Prakash, G. K. S., Beyond Oil and Gas: the Methanol Economy. Wiley-VCH: Weinheim, (129) Liu, H. S.; Song, C. J.; Zhang, L.; Zhang, J. J.; Wang, H. J.; Wilkinson, D. P. J. Power Sources 2006, 155, 95. (130) Alden, L. R.; Roychowdhury, C.; Matsumoto, F.; Han, D. K.; Zeldovich, V. B.; Abruna, H. D.; DiSalvo, F. J. Langmuir 2006, 22, (131) Alden, L. R.; Han, D. K.; Matsumoto, F.; Abruna, H. D.; DiSalvo, F. J. Chem. Mat. 2006, 18, (132) Roychowdhury, C.; Matsumoto, F.; Zeldovich, V. B.; Warren, S. C.; Mutolo, P. F.; Ballesteros, M.; Wiesner, U.; Abruna, H. D.; DiSalvo, F. J. Chem. Mat. 2006, 18, (133) Wang, H. S.; Alden, L.; DiSalvo, F. J.; Abruna, H. D. Phys. Chem. Chem. Phys. 2008, 10, (134) Zhou, W. P.; Lewera, A.; Larsen, R.; Masel, R. I.; Bagus, P. S.; Wieckowski, A. J. Phys. Chem. B 2006, 110, (135) Rice, C.; Ha, S.; Masel, R. I.; Wieckowski, A. J. Power Sources 2003, 115,

64 Chapter (136) Choi, J. H.; Jeong, K. J.; Dong, Y.; Han, J.; Lim, T. H.; Lee, J. S.; Sung, Y. E. J. Power Sources 2006, 163, 71. (137) Park, I. S.; Lee, K. S.; Choi, J. H.; Park, H. Y.; Sung, Y. E. J. Phys. Chem. C 2007, 111, (138) Greeley, J.; Norskov, J. K.; Mavrikakis, M. Annu. Rev. Phys. Chem. 2002, 53, 319. (139) Hammer, B.; Norskov, J. K. Adv. Catal. 2000, 45, 71. (140) Ruban, A.; Hammer, B.; Stoltze, P.; Skriver, H. L.; Norskov, J. K. J. Mol. Catal. A-Chem. 1997, 115, 421. (141) Demirci, U. B. J. Power Sources 2007, 173, 11. (142) Yu, H.; Chen, M.; Rice, P. M.; Wang, S. X.; White, R. L.; Sun, S. H. Nano Lett. 2005, 5, 379.

65 Chapter Chapter 2 PtAg Alloy Nanoparticles and their Formation of Pt-Surface Rich Electrocatalysts 2.1 PtAg Alloy Nanoparticles with the Compositions in Miscibility Gap Introduction Scaling-related thermodynamic property of bimetallic nanoparticles is of great current research interest because of the catalytic, magnetic and electronic applications of alloys and intermetallics. 1-6 With the decrease of particle size, especially to the point where the critical dimensions are in the nanometer regime, many properties can be significantly different from those of their bulk counterparts. 4 For example, surface plasmon resonance becomes significant for bimetallic Ag-Au nanoparticles and tunable with the change of compositions Although bulk phase diagrams for many bimetallic materials have been well studied and widely used for decades, 12 it is unclear whether they can be simply extended to the regime of nanometer-sized materials as a large fraction of atoms reside on the surfaces when sizes of particles become small. Surface atoms can contribute significantly to the increase of Gibbs free energy and result in a large difference in property between nanoparticles and bulk materials, such as melting point 13 and crystal order. 14 The large fraction of atoms on the surface and at interfaces of bimetallic nanoparticles should affect their phase behaviors, as being

66 Chapter predicted by theoretical simulation In the case of Au-Ge alloys, both the composition and temperature of eutectic are shown to scale with the size of nanostructures. 19 Surface segregation of bimetallic nanoparticles is another phenomenon in which chemical composition at surface differs from that in bulk It becomes obvious when the difference in atomic size, surface energy, and strain energy between given two metal elements is large. The surface atoms can reconstruct to form core-shell, sandwich, 21, 29, 30 and onion-like 31, 32 structures. In this context, the miscibility of individual element in bimetallic nanoparticles is important in understanding the structure and property. Change in miscibility may help to design novel alloy nanomaterials that possess unique properties that may not be obtained as bulk materials. Both theoretical studies and some limited experimental data have shown that miscibility between the metal elements can be increased with the decrease in 15, 16, 18, particle size. While nanoparticles of different alloys have been made, 1, 2, 36, 37 most of these materials have compositions existed in bulk forms. Only a 38, 39 few nanoalloys with miscibility gaps have been observed, such as Au-Pt, Pt-Ru 40 and Pb-Sn. 41 The nanometer scale phenomena have hardly been explored. It is unclear if the enhancement of miscibility in these selective nanomaterials can also be the case for Ag-Pt or other bimetallic systems. It is also interesting to study the meta-stability of these nanoalloys by examining the

67 Chapter sintering behaviors of those bimetallic materials, if the miscibility is scalable with size. In this section, the phase and composition behaviors of Ag-Pt nanoalloys are presented. Silver platinum bimetallics have been studied in recent years as 42, 43 lead-free solder materials. In the bulk, Ag-Pt bimetallics have a large miscibility gap at the temperature below about 1190 C 12 and form alloys only at very high atomic ratio of either Ag or Pt. 44 When the atomic composition falls outside the range between Ag 2 Pt 98 and Ag 95 Pt 5 at <~ 400 C, the bulk materials exist in more than one phase: one Pt enriched and the other Ag enriched alloys. We show in this work that Ag-Pt alloy nanostructures can be synthesized in solution from the molecular precursors of their corresponding salts. Furthermore, both miscibility and crystal phase of these nanoparticles can change upon annealing at 700 C Experimental Section Materials. Platinum acetylacetonate (Pt(acac) 2, Strem Chemicals, 98%) and silver stearate (Alfa Aesar) were purchased from VWR. Oleic acid (90%, technical grade), oleylamine (70%, technical grade), 1, 2-hexadecanediol (HDD, 90%, technical grade), and diphenyl ether (DPE, 99%, ReagentPlus ) were purchased from Aldrich. All chemicals and reagents were used as received. Synthesis. In a standard procedure, Pt(acac) 2 (0.03 g or 75 µmol), silver

68 Chapter stearate (0.12 g or 0.30 mmol), HDD (0.49 g or 1.9 mmol) were mixed with oleic acid (0.3 ml or 0.9 mmol), oleylamine (0.3 ml or 0.9 mmol), and DPE (5mL or 31.5 mmol) in a 25-mL three neck round-bottom flask. The experiments were carried out under an argon atmosphere using the standard Schlenk line technique. A heating mantle was used in these reactions and the temperature was controlled by a thermal couple connected to a controller (J-KEM Scientific Inc.). The reaction mixture was heated to the reflux temperature (around 260 C) at 5 C/min and held for 1 h. To study the compositions, we systematically adjusted the relative amount between Pt(acac) 2 and silver stearate while kept the total mole number of these two precursors at mmol. After the reaction, nanoparticles were separated by dispersing the obtained reaction mixtures with 2 ml of hexane and 6 ml of ethanol, followed by centrifugation at 5000 rpm for 5 min. The particles obtained were redispersed in 10 ml of hexane and 5 ml of ethanol, followed by centrifugation at 5000 rpm for 5 min. After this size selection procedure, the particles were collected for further study. Thermal treatment. To study the phase behaviors during the annealing, both as-made and carbon-supported Ag-Pt nanoparticles were used. Several different loadings were examined for carbon-supported Ag-Pt nanoparticles. The heat treatment of these samples was conducted in a ceramic boat using a programmable tube furnace. The tube was heated to 300 C in air at 5 C /min and kept for 1 h,

69 Chapter followed by ramping the temperature to 700 C at 3 C /min in forming gas (5 vol.% H 2 in Ar). The annealing was conducted at 700 C for 1 h. Characterization. Powder X-ray diffraction (PXRD) patterns were recorded using a Philips MPD diffractometer with a Cu K α X-ray source (λ= Å). Transmission electron microscopy (TEM) images were taken on a JEOL JEM 2000EX microscope at an accelerating voltage of 200 kv. For TEM study, the specimens were prepared by drop-casting hexane-dispersed particles onto carbon-coated copper grids. Energy dispersive X-ray (EDX) analysis was carried out on a field emission scanning electron microscope (FE-SEM, Zeiss-LEO DSM 982) operating at 20 kv. The SEM specimens, typically tens of micrometers in thickness, were prepared by transferring concentrated particle dispersions in hexane onto sample holders using pipettes. The thermal gravimetric analysis (TGA) was conducted on an SDT-Q600 TA instrument. In a standard procedure, 5-10 mg powder sample was transferred into an alumina pan, which was subsequently placed onto the dual beam sample holder. The flow rate of air was set at 50 ml/min. The typical temperature range was from room temperature to 800 C and the heating rate was 10 C/min Results and Discussion Silver platinum nanoparticles could be produced after a reaction period of 1 h for the reaction mixture at Ag stearate/pt(acac) 2 molar ratio of 4:1 (Figure 2.1a).

70 Chapter The average diameter of these nanoparticles was about 8.2 nm with a standard deviation of 1.1 nm. When Ag stearate/pt(acac) 2 molar ratio changed to 2:1, the particles obtained had a slightly larger average diameter of 9.2 ± 1.9 nm (Figure 2.1b). The morphology of final particles became non-spherical, when Ag stearate/pt(acac) 2 molar ratio decreased below above unity. Network of cross-linked nanowires was observed for the nanostructures formed at Ag stearate/pt(acac) 2 molar ratio of 1:2 and 1:4 (Figure 2.1c, d). The diameter of these nanowires was found to be about 6 nm. The large difference in particle morphology could be due to the surface segregation for the different compositions or a change in the growth kinetics with Ag stearate/pt(acac) 2 ratio. In the later case, growth kinetic can be affected by the interactions between capping agents and surface atoms. Such interaction can be sensitive to the concentration of precursors and reaction condition. Figure 2.1 TEM images of PtAg nanoparticles with different molar ratios of Pt/Ag precursors: (a) 1:4; (b) 1:2; (c) 2:1; (d) 4:1.

71 Chapter Figure 2.2 shows the PXRD patterns of these Ag-Pt nanostructures, pure silver ( , JCPDS-ICDD) and platinum metals ( , JCPDS-ICDD), respectively. All four Ag-Pt nanomaterials show diffractions that could be indexed to (111), (200), (220), (311) and (222) planes of a face-centered cubic (fcc) lattice. No other diffraction peak was observed, indicating these products were phase pure. The observed diffraction angles for Ag-Pt nanostructures fell in between those of the two pure metal elements, suggesting the formation of alloy nanoparticles. The diffraction peaks of alloy samples had higher 2θ angles than those for pure silver metal and moved to high angles with the decrease of Ag stearate/pt(acac) 2 molar ratio. Figure 2.2 XRD patterns of as-prepared PtAg nanoparticles with different molar ratios of Pt/Ag precursors.

72 Chapter EDX analysis was used to determine the elemental compositions for the final products. Platinum, silver, carbon and oxygen were the elements that could be detected (Figure 2.3a). The carbon and oxygen signals were most likely due to the capping agents on the surfaces of nanoparticles. The atomic percentage of platinum in these bimetallic alloy products increased with the decrease in Ag stearate/pt(acac) 2 molar ratio (Figure 2.3b). The amount of silver in the nanoalloys seemed to be slightly lower than that in silver stearate used, suggesting that the silver precursors were not completely converted into alloy products after the reaction for 1 h. Figure 2.3 (a) Representative EDX spectrum of PtAg nanoparticles made from a mixture with molar ratio of Pt/Ag precursors equal to 1:4, and (b) relationship between Pt atomic ratios in final PtAg alloys versus those in the metal precursors.

73 Chapter A linear relationship between lattice parameter and composition described by the Vegard s law can be a good indication of miscible binary alloys that form solid solution. This relation can be described by the following empirical equation: 45 a1 a2 a = a x1 (2.1) a2 where a, a 1, and a 2 are lattice parameters for the binary solid solution and the two corresponding pure metals, respectively, and χ 1 the molar fraction of Component 1. The lattice parameter for Ag-Pt alloys can be determined experimentally according to the following equation: 46 λ 2sinθ 2 2 a = h + k + l 2 (2.2) where λ is the wavelength of X-ray which equals to Å for Cu K α source; θ the diffraction angle, and ( hkl ) the Miller indices. Figure 2.4 shows the relationship between composition and lattice parameter of these Ag-Pt nanoalloys. The compositions were based on the EDX analysis and the lattice parameters were from the PXRD data. The average atomic compositions were Ag 74 Pt 26, Ag 58 Pt 42, Ag 26 Pt 74 and Ag 16 Pt 84, respectively, for alloy nanostructures made at the Ag stearate/pt(acac) 2 molar ratio of 4:1, 2:1, 1:2 and 1:4. These experimental data agreed well with those theoretical values obtained based on the Vegard s law, indicating that the Ag-Pt particles were indeed alloys. The deviation between the experimental data and theoretical prediction might be largely due to the measurement errors although scaling effect could play a role in the calculation

74 Chapter where bulk lattice parameters for pure metals were used. Recent studies suggest that 6 nm platinum and silver nanoparticles could have roughly 0.5% deviation in 47, 48 lattice parameter when compared with their bulks. Figure 2.4 Experimental and theoretical data of lattice parameter and composition relationship for as-synthesized and thermally-annealed Ag-Pt nanostructures. The theoretical data were calculated based on the Vegard s law. Figure 2.5 PXRD patterns of thermally annealed Ag-Pt alloy nanostructures in (a) θ range and (b) the enlarged (111) and (200) diffractions.

75 Chapter The effect of temperature on the phase stability of Ag-Pt nanoalloys were studied by treating the as-made particles at 300 C for 1 h in air and then 700 C for 1 h in the forming gas. The rationale for choosing 700 C for the annealing experiments is that the phase transition and segregation for Ag-Pt nanoalloys occurs at around this temperature. In comparison, the phase transition temperature for the bulk Ag-Pt bimetallics is around 1200 C. Our TEM study indicates that the Ag-Pt nanoparticles sintered dramatically after the heat treatment. Figure 2.5 shows the PXRD patterns of the sintered product from the four Ag-Pt alloy nanoparticles. The enlarged (111) diffraction regions were presented to show the change after this thermal treatment. For a given Ag-Pt alloy, at least one new set of diffraction peaks other than those for the alloy were detected, indicating that phase segregation happened. The relative peak intensities also changed with the atomic compositions of Ag-Pt alloy nanostructures. We noticed that the peak positions of these annealed products were not centered between those for pure silver and platinum metals, suggesting that Ag- and Pt-enriched alloys were most likely the main final products. Their lattice parameters were calculated based on both sets of diffraction peaks (Figure 2.4), which moved far away from that for Ag-Pt alloy nanostructures but between those for pure silver and platinum metals. The composition of these two enriched alloys changed little with different Ag-Pt alloy nanostructures and was near the phase boundary regions in the bulk phase diagram (Figure 2.4).

76 Chapter The Ag 74 Pt 26 alloy nanoparticles were used to study the effect of particle size on the phase behavior. Carbon-supported alloy nanoparticles were also used to control the growth of particle size during the annealing process. The low, medium and high metal loadings were tested at 3, 10 and 70 wt.%, respectively. Figure 2.6 shows the TEM images of carbon-supported Ag-Pt nanoparticles at these three different loadings before and after the thermal treatment. At the low metal loading (3 wt.%), the size of supported alloy particles did not change dramatically, as they were far apart. When the particle loading was increased to around 10 wt.%, sintering became obvious. When the loading reached to around 70 wt.%, the nanoparticles sintered in large numbers and formed crystals with the diameter or edge length up to several hundreds of nanometers. Figure 2.6 TEM images of carbon-supported Ag 74 Pt 26 nanoparticles at (a, b) 3, (c,

77 Chapter d) 10, and (e, f) 70 wt.% metal loadings (a, c, e) before and (b, d, f) after the thermal annealing. Figure 2.7 PXRD patterns in (a) θ range, and (b) the enlarged (111) and (200) diffractions for carbon-supported Ag 74 Pt 26 nanoparticles after annealing at 700 C. The loading is based on the total weight of metals. Figure 2.7 shows the PXRD patterns of these carbon-supported Ag-Pt nanoparticles after the thermal treatment. There were hardly any changes for the diffraction patterns at the low loading level of alloy nanoparticles. A careful examination shows that the diffraction peaks of these annealed Ag-enriched alloy nanoparticles were at higher angles than those after thermal treatment without carbon supports. This observation indicates that the carbon-supported Ag-Pt particles should have a less degree of phase segregation for samples with carbon supports than those without. In another word, the Ag-Pt alloy nanoparticles on carbon-support were thermodynamically stable at this enhanced temperature and

78 Chapter did not phase segregate as dramatically as the large particles during the annealing process. With the increase of metal loading, a shoulder peak around 40 2θ and next to the (111) diffraction of Ag-Pt alloy appeared. Its position was close to the diffraction for pure platinum. The intensity of this shoulder peak increased with the metal loading on carbon support, indicating strong phase segregation after the annealing process. The diffraction peaks also became sharper than before due to largely the growth of crystal domain size, which agreed with the TEM observation Conclusions Silver platinum alloy nanoparticles with composition through the entire miscibility gap of the bulk were successfully prepared. The lattice parameter and composition of these nanoparticles were experimentally determined and their relationship followed the Vegard s law, a strong indication of the alloy formation between platinum and silver at nanometer scale. These Ag-Pt alloy nanoparticles can grow in size and turn into two-phase solid-state materials after annealing at high temperature. These multiphase materials should be Pt- and Ag-rich alloys along the boundary regions in bulk phase diagram. Our PXRD and TEM data further suggest that the phase segregation of Ag-Pt alloy nanostructures should be size-dependent. Understanding of this phenomenon can be important for designing novel alloy nanomaterials which may have unique electronic, magnetic

79 Chapter and catalytic properties. Author Contributions: This section is based on a published work (Z. M. Peng, H. Yang, Ag-Pt Alloy Nanoparticles with the Compositions in Miscibility Gap, Journal of Solid State Chemistry, 2008, 181, ). Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, analysed the data and co-wrote the paper.

80 Chapter Understanding the Composition-Dependent Formation of Platinum-Silver Nanowires Introduction The colloidal nanocrystals form through nucleation, followed by a growth process in which nutrients are consumed Recently, in-situ transmission electron microscopy (TEM) studies show that final resulting nanoparticles can be generated from the primary colloidal Pt nanoclusters In some cases, primary particles can grow into nanowires through oriented attachment For semiconducting quantum dots and oxides, dipolar interaction seems to the key driving force for the formation of such nanowires. The oriented attachment however can lead to the formation of low dimensional morphologies in materials with no permanent dipole as well, suggesting factors other than dipolar interaction 59, may also be important. Nanostructured platinum alloys are selected for this study, because they are important in a range of industrial applications, from electronics to catalysts and 49, electrocatalysts. The needs for green technology and sustainable industrial processes put new challenges in the design and synthesis of highly active and selective nanocatalysts, 74,75 which requires better understanding on how to precisely control crystal morphology, as shape and surface composition of nanocrystals are typically governed by the underlying crystal symmetry. For the

81 Chapter synthesis of Pt alloy nanowires, PtAg, PtCo, PtCu, and PtFe have been 66, synthesized in both molecular solvent and ionic liquid media. These alloy nanowires have typically been obtained through the reduction of metal salts and thermal decomposition of metal carbonyls in presence of capping agents It has also been found recently that PtAg can not only form alloys at the nanometer-sized scale, but also possess different nanostructures when they are produced in a non-hydrolytic system. 66 In this section, we present the formation of Pt 53 Ag 47 nanowires in the presence of oleic acid and oleylamine while Pt or Ag rich alloys form faceted or sphere-like nanoparticles. This composition-dependent growth is used to understand the key factors that govern the formation of metal alloy nanowires. In this context, DFT is used to compare the absorption energy between the functional groups of capping agents and low index surfaces of Pt-Ag alloys, while MD simulation is used to examine the collisions between individual particles and the surface atomic diffusion upon collision. The experimentally observed formation of PtAg nanowire can be explained well by the simulation data Experimental Section Materials. Platinum acetylacetonate (Pt(acac) 2, Strem Chemicals, 98%) and silver stearate (Ag(St), Alfa Aesar) were purchased from VWR. Oleic acid (OA, 90%, technical grade), oleylamine (OAm, 70%, technical grade), 1,

82 Chapter hexadecanediol (HDD, 90%, technical grade), and diphenyl ether (DPE, 99%, ReagentPlus ) were purchased from Aldrich. All chemicals and reagents were used as received. Preparation of PtAg alloy nanostructures. All experiments were carried out under argon atmosphere using a standard Schlenk technique. In a typical procedure, HDD (0.49 g or 1.9 mmol), OA (0.3 ml or 0.9 mmol), and OAm (0.3 ml or 0.9 mmol) were mixed with DPE (4 ml or 25.2 mmol) in a 25-mL three neck round-bottom flask and preheated to 200 C. In a separate flask, Pt(acac) 2 (0.025 g or mmol) and Ag(St) (0.025 g or mmol) were mixed with DPE (1 ml or 6.3 mmol) and heated to 80 C until all solids were dissolved. The latter solution, which had a light yellow color, was injected into the flask at 200 C and held for 1 h unless stated otherwise. PtAg alloy nanostructures with different Pt/Ag molar ratios were made by using predetermined amounts of Pt(acac) 2 and Ag(St), while keeping the sum of these two precursors constant at mmol. For comparison, pure Pt and Ag nanoparticles were prepared using the same procedure. After the reactions, the resulting mixtures were washed twice with 2 ml of hexane and 6 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. The precipitate was re-dispersed in 2 ml of hexane for further characterizations. Characterization. Transmission electron microscopy (TEM) images were taken on a Hitachi 7100 microscope at the accelerating voltage of 80 kv.

83 Chapter High-resolution transmission electron microscopy (HR-TEM) images were obtained using a FEI TECNAI F-20 field emission microscope operated at 200 kv, which has an optimal resolution of 1 Å in TEM mode. Energy dispersive X-ray (EDX) analysis on ensembles of nanoparticles was carried out on a field emission scanning electron microscope (FE-SEM, Zeiss-Leo DSM982) installed with an EDAX detector. Powder X-ray diffraction (PXRD) patterns were recorded using a Philips MPD diffractometer with a Cu K α X-ray source (λ= Å). DFT Calculation. The Dmol3 code developed by Delley was used to obtain the adsorption energy (E ad ) of the two functional groups on metal and alloy surfaces. 83, 84 The structures were optimized using Perdew-Wang exchange-correlation function (PW91) based on the generalized gradient approximation (GGA). No symmetry and spin restrictions were applied in the calculation. A double numerical basis set plus a polarization p-function (DNP), DFT semi-core pseudopots (DSPP), and an octupole scheme were selected to describe the multipolar expansion of the charge density and Coulomb potential. A thermal smearing of hartree (0.136 ev) was set for the energy level of occupied orbits in order for them to converge. The following criteria were used to obtain the optimized final structure. First, convergence tolerance of self-consistent field (SCF) energy was less than 10 6 hartree ( ev) in the conjugate gradient algorithm. Second, the maximum displacement of an atom was less than Å, and the force due to the displacement was less than 0.002

84 Chapter hartree/å (0.054 ev/ Å). For Pt 50 Ag 50 alloy, ordered surfaces were used in the simulation. The structures of both capping agents and metal surfaces were fully optimized before they were brought together. The entire structure containing both the adsorbed molecule and metal surface was then optimized to get the most stable form. The E ad was obtained using the following equation: E ad = E E E (2.3) total s m where E total, E s and E m are the bond energy of the whole system, metal surface and free molecule, respectively. MD Simulation. The MD simulations were performed in an atomic number-volume-temperature (NVT) canonical ensemble and the simulation time step was set to be one femtosecond (fs) for all calculations. The atomic interactions were described with the PCFF30 force field in the Forcite package, which is obtained by using the ab initio principle and empirical parameters, and can be expressed by the following equation: E = E + b E + a E + elec + E VDW E o + E t + E bb + E ab + E aa + E at + E bt (2.4) where E b, E a, E o, E t, E bb, E aa, E ab, E at and E bt are valence energy terms consisting of those from the distortions of bond length (E b ), bond angle (E a ), out-of-plane bending angle (E o ), torsion angle (E t ), two bands with one common atom (E bb ), two angles with a common band (E aa ), coupling between a bond and an angle (E ab ), interactions between a bond and a dihedral angle of rotation about the bond (E bt ) and an angle and a torsion (E at ). E elec and E VDW are non-bonded energy terms

85 Chapter and separately derived from electrostatic interactions and the van der Waals forces. The potential energy (E p ) between metal nanoparticles was calculated by using two 3-nm particles at an initial distance of 3 nm. Randomly mixed Pt 50 Ag 50 nanoparticles were optimized using PCFF force field and then used in all the MD simulations. E p value was obtained by calculating the difference in total energy of the two particles before (E i ) and after (E f ) the interaction, which can be depicted as: E p = E E (2.5) f i The atom diffusions on the surface of 3-nm metal nanoparticles were simulated at different temperatures ranging from 2 to 1000 K with a step size of 50 K. A simulation time of 200 ps was used at each temperature. The formations of nanowires due to the attachment of primary nanoparticles were modeled using two fully optimized 3-nm nanoparticles. The initial distance between these two particles was set at 1 nm and the temperature was set at 480 K, which was used in the experiments. Changes in particle-particle distance (D), total energy (E) and pair correlation function (g(r)) as a function of simulation time were calculated Results and Discussion

86 Chapter Figure 2.8 TEM images of (a) Pt 26 Ag 74, (b) Pt 39 Ag 61, (c, d) Pt 53 Ag 47, (e) Pt 73 Ag 27, and (f) Pt 86 Ag 14 nanostructures, respectively. Figure 2.8 shows the TEM images of Pt-Ag nanostructures made at different platinum acetylacetonate/silver stearate (Pt(acac) 2 /Ag(St)) molar ratios. Energy dispersive X-ray (EDX) analysis was used in obtaining the composition of these alloy nanostructures and the results are summarized in Table 2.1. There was a clear correlation between the shape and composition of Pt-Ag nanostructures.

87 Chapter Sphere-like nanoparticles, obtained at Pt(acac) 2 /Ag(St) molar ratio of 1/4, had an average composition of Pt 26 Ag 74 and a diameter of 3.2 ± 0.4 nm (Figure 2.8a and Table 2.1). Similarly, 3.4 ± 0.4 nm Pt 39 Ag 61 nanoparticles formed at Pt(acac) 2 /Ag(St) molar ratio of 1/2 (Figure 2.8b). When this ratio increased to 1/1, the shape of nanoparticles was dominated by worm-like nanowires (Figure 2.8c). The average diameter of the nanowires was 2.8 ± 0.4 nm based on the TEM measurement. High-resolution TEM (HR-TEM) images reveal that these nanowires were composed of multiple crystalline domains, and the observed fringes were from {111} planes (Figure 2.8d). The domains of primary crystals were about 3 nm, comparable to diameters of nanoparticles prepared at low Pt(acac) 2 /Ag(St) molar ratios. These observations suggest the growth of nanowires undergo an oriented attachment process where the primary nanoparticles form first 59, 60 and grow into nanowires. EDX analysis indicates that the average composition of these nanowires was Pt 53 Ag 47, which was similar to the Pt(acac) 2 /Ag(St) feeding ratio. The length of the nanowires decreased dramatically when Pt(acac) 2 /Ag(St) molar ratio was changed to 2/1 (Figure 2.8e). EDX data indicate that the nanoparticles had an average composition of Pt 73 Ag 27. Nanowires could hardly be observed, when Pt(acac) 2 /Ag(St) molar ratio changed to 4/1, and composition in the resulting nanostructures reached Pt 86 Ag 14 (Figure 2.8f). Figure 2.9 shows the PXRD pattern of Pt 53 Ag 47 nanostructures made at Pt(acac) 2 /Ag(St) molar ratio of 1/1. The diffraction pattern could be readily indexed to (111), (200),

88 Chapter Table 2.1 Compositions of Pt-Ag Alloy Nanostructures Prepared at Different Pt(acac) 2 /Ag(St) Feeding Ratios Sample No. Pt(acac) 2 /Ag(St) molar ratio PtAg nanoalloys a Pt (at%) Ag (at%) Composition 1 1/ ± ± 0.5 Pt 26 Ag / ± ± 0.2 Pt 39 Ag / ± ± 0.6 Pt 53 Ag / ± ± 0.4 Pt 73 Ag / ± ± 0.8 Pt 86 Ag 14 a : The atomic percentage and composition of the alloys were obtained based on EDX analysis. Figure 2.9 XRD pattern of worm-like Pt 53 Ag 47 nanostructures. (220), (311) and (222) planes of a face-centered cubic (fcc) lattice and in between those for pure Ag ( , JCPDS-ICDD) and Pt metals ( , JCPDS-ICDD). This observation indicates that Pt and Ag formed alloy in the form

89 Chapter of a solid solution. 66 Similar diffraction patterns at different diffraction positions can be obtained for the nanoparticles made at other Pt(acac) 2 /Ag(St) molar ratios. The growth of Pt 53 Ag 47 nanostructures was further studied by TEM (Figure 2.10). Both small primary particles and some short rod-like nanostructures formed 2 min after the injection of metal precursors (Figure 2.10a). With the increase of reaction time, worm-like morphology appeared (Figure 2.10b, c). The nanowires reached their maximum length after 1 h, and no major changed in morphology even after 3 h (Figure 2.10d). The width of nanowires changed very little throughout the process, suggesting that these Pt 53 Ag 47 nanowires grow from the 59, 65 primary nanoparticles through the oriented attachment. Figure 2.10 TEM images of Pt 53 Ag 47 alloy nanostructures made after the reaction taking place for: (a) 2, (b) 10, (c) 30, and (d) 180 min, respectively.

90 Chapter Figure 2.11 TEM images and the corresponding schematic illustrations showing the early growing stages from (a) primary particle to two-particle system through (b) MA and (c) TA growthes, and (d-i) three-particle systems of Pt 53 Ag 47 nanowires through either MA or TA growth, respectively. HR-TEM is used to study the shape evolution of Pt 53 Ag 47 nanowires and the data show the attachment occurred mostly on {111} facets (Figure 2.11). Two growth modes namely lattice matched attachment (MA) and twinning attachment (TA) were observed. The grain boundary and lattice defect were not obvious if

91 Chapter two primary particles attached and grew together via the MA mode (Figure 2.11a,b). A twin plane with the mirror image-like orientation could be observed if two particles attached through the {111} surfaces (Figure 2.11c). Both linear and bending morphologies could be obtained when additional particles attached to the existing nanostructures through either MA or TA growth (Figure 2.11d-i). Figure 2.12 Potential energy (E p ) as a function of distance between two 3-nm Pt 50 Ag 50, Pt and Ag nanoparticles. Several factors may contribute to the observed composition-dependent shape evolution of nanowires, among which effectiveness of the collisions between primary nanoparticles should be an important factor for the attachment growth. In a colloidal system, the particles move because of Brownian motions, and collide with a frequency that can be estimated according to the following equation f = ktρ 2 3, where ρ is the density of nanoparticles with a radius of r in a 3πη r solution with viscosity of η, and T is temperature. 89 Van der Waals interactions become dominated and increase the probability of collision when two particles are in close proximity. The potential energy (E p ) profiles due to atomic interaction

92 Chapter between two 3-nm nanoparticles of Pt 50 Ag 50 (to mimic composition of Pt 53 Ag 47 ), pure Pt and Ag were calculated using MD simulation (Figure 2.12, see experimental section for details) The minimal E p was obtained when the interparticle distance was around 2.6 Ǻ, roughly equal to the lattice spacing of (111) planes. Further decrease in interparticle distance led to a sharp increase in potential energy, associated with the repulsive forces from the overlapping electrons. The E p -d profile between two Pt 50 Ag 50 nanoparticles resembled those of atomic interactions in some aspects. 90 The values of E p depended strongly on the composition of nanoparticles, with platinum having the largest decrease in minimal E p, followed by Pt 50 Ag 50. Silver nanoparticles had the smallest decrease in minimal E p value. Figure 2.13 Side (top images) and top (bottom images) views of the two possible configurations of amine functional group on Pt 50 Ag 50 {111} surfaces. Color code: orange, Ag; dark blue, Pt; yellow, H; gray, C; and blue, N.

93 Chapter As these Pt-Ag nanostructures were produced in solution with capping agents, the effect of oleylamine (OAm) and oleic acid (OA) needs to be considered carefully. In the DFT calculation, carboxylic acid and amine were considered to be the key functional groups in obtaining the adsorption energy (E ad ) on metal and metal alloy surfaces. To simplify the simulations, we used short carbon-chain molecules, propylamine and propanoic acid, to model the functional groups of oleylamine and oleic acid respectively. This approach was validated by comparing the values of adsorption energy between oleylamine and propylamine on Ag (100) surface. It was found that the E ad was ev for oleylamine and ev for propylamine. The difference in E ad between these two values was about 1%, which is insignificant. For simplicity, an ordered Pt 50 Ag 50 alloy surface was employed in the DFT calculation. For a given family of Pt 50 Ag 50 alloy surface, there can be more than one configuration of surface atomic arrangement and for the adsorption of the head group of capping agent. Figure 2.13 show the side and top views of the two most likely configurations of amine functional group on Pt 50 Ag 50 {111} surfaces. In this case, the structure of adsorbing molecule was 83, 84 fully optimized before the DFT calculation. Adsorption energy of the capping agents was obtained by calculating the energy difference before and after the adsorption of functional groups on the Pt 50 Ag 50 surfaces. The E ad was ev when nitrogen atom is adsorbed on surface Pt atoms, and ev for the situation when the adsorption is on Ag atoms (Figure 2.13). The more negative the

94 Chapter E ad value is, the stronger the adsorption is. Thus, the more negative E ad value of the two was used to represent the adsorption strength of the functional group on a given surface, when multiple possible configurations for the adsorption on the alloy surface existed. Similarly, there were two possible adsorption configurations of functional group for oleylamine adsorbed on Pt 50 Ag 50 {100} surfaces, and two on Pt 50 Ag 50 {110} surfaces, respectively. For the adsorption of oleic acid (OA) on Pt 50 Ag 50 surfaces, there should be multiple configurations for the adsorptions on the above three low index Pt 50 Ag 50 surfaces because both oxygen atoms of carboxylic acid group can interact with either Pt or Ag atom. The calculated adsorption energy ranges from ev for the situation where carbonyl oxygen of the COOH group adsorbed on Ag atom and the hydroxyl oxygen on Pt atom to ev for both oxygen atoms on Ag. There are six most likely configurations for carboxylic group on Pt 50 Ag 50 {100} surfaces and four on Pt 50 Ag 50 {110} surfaces. The E ad values were calculated for all these configurations of Pt 50 Ag 50 low-index surfaces, together with those of pure Pt and Ag metal surfaces. The E ad values for the strongest adsorptions of amine and carboxyl acid groups on these surfaces are summarized in Table 2.2. These values can be used to compare to the strength of molecular adsorption on surfaces and the ability to prevent the alloy surfaces of primary particles from directly contacting with each other because of the steric hindrance effect of the 61, 81 capping agents.

95 Chapter Figure 2.14 Adsorption energy (E ad ) of functional groups for (a) OAm and (b) OA molecules on the three low-index surfaces of Pt 50 Ag 50, Ag and Pt, respectively. Table 2.2 Simulation Results on Adsorption Energy of Functional Groups of Oleylamine (OAm) and Oleic Acid (OA) for the Strongest Adsorptions on Three Low-Index Ag, Pt and Pt 50 Ag 50 Surfaces a surface (100) (110) (111) OAm OA OAm OA OAm OA Ag Pt Pt 50 Ag a : All values are in ev. Figure 2.14 show the bar diagram of these calculated E ad of functional groups for OAm and OA on the three low-index surfaces of Pt 50 Ag 50 alloys. In general, amine functional group bounded more strongly than carboxylic acid group.

96 Chapter Among the configurations of amine functional group on the three low-index surfaces, (111) plane had the smallest E ad, suggesting OAm should preferably reside on the surface of nanoparticles and the {111} facets be the least protected. These calculations explain the experimental observation that the oriented growth happened preferably on the {111} surfaces. The relative large E ad and small E p for pure Ag, both of which can result in reducing the direct contact of metal alloy surface between two particles, or the so-called effective collision, are likely the key reasons for the observation that Ag nanoparticles did not form worm-like structures though the oriented attachment under the similar reaction conditions (Figure 2.15a). Figure 2.15 TEM images of (a) Ag and (b) Pt nanoparticles prepared under the same conditions as those for making Pt 53 Ag 47 nanowires.

97 Chapter Figure 2.16 Simulation of time-dependent (a) mean square displacement (MSD) of surface atoms of 3-nm nanoparticles of PtAg alloys, Ag and Pt metals; (b) changes in particle-particle distance (D) and (c) total energy (E); and (d-f) pair correlation function, g(r), as a function of r at the interfacial regions between two colliding Pt 50 Ag 50 particles. An effective collision between two primary nanoparticles is only the precondition for the formation of nanowires through oriented attachment. The metal atoms at the colliding interfaces need to form strong metallic bonds in order to grow into nanowires. To under the bond formation and interfacial structures upon collisions between two nanoparticles, we used MD simulation to obtain the mean square displacements (MSD) as a function of temperature for Pt 50 Ag 50 alloy,

98 Chapter Pt and Ag metals (Figure 2.16a). The results show that Ag atoms could diffuse easily at Pt 50 Ag 50 interfacial regions under the reaction temperature of 200 C (or 473 K). Interestingly, Pt nanoparticles show almost no change in MSD at this temperature. The lack of atomic diffusion upon collision suggests that Pt-Pt bond cannot be readily broken and reconstruct at low reaction temperatures. Experimentally, we observed the formation of highly faceted Pt nanoparticles under the same reaction conditions as those for making the worm-like alloy nanowires (Figure 2.15b). While Pt atoms in Pt 50 Ag 50 alloy had a larger change in MSD value than Pt at 200 C, Ag atoms in Pt 50 Ag 50 alloys had a even more dramatic increase in MSD value than Pt, suggesting the diffusion should be driven preferably by Ag atoms (Figure 2.16a). Changes in both particle-particle center distance (D) and total energy (E) of the colliding Pt 50 Ag 50 alloy particles were calculated as a function of simulation time (Figure 2.16b, c). The results indicate that collision between two Pt 50 Ag 50 nanoparticles was inelastic as the changes in both D and E were gradual. The particle-particle center distance decreased rapidly after the collisions, and the process was dominated by the plastic deformation (Figure 2.16b). In this inelastic deformation, the nanostructure went through a reconstruction due to the slide of atomic planes and surface atom diffusion. The primary particles could rotate to match the crystal planes. Our simulation data further indicate that the surface atoms of Pt-Ag alloy particles reorganized and moved fairly rapidly to the neck

99 Chapter region between the colliding particles. This fast movement was likely driven by 89, 92 the difference in chemical potential between the convex and concave regions. The time-dependent pair correlation function (g(r)) calculation is used to evaluate the structure order of the interfaces after two Pt 50 Ag 50 particles collided (Figure 2.16d-f). The narrower the peak is, the higher the crystalline order is. The peaks for Pt 50 Ag 50 interface became narrow with the increase of simulation time, indicating the atoms in this region changed from disordered to ordered structures. In comparison, two colliding Pt nanoparticles reached stable states in a much shorter simulation time than Pt 50 Ag 50 particles, and with little relaxations in both D and E. The corresponding g(r) functions for Pt nanoparticles show no large lattice deformation. This result suggests that even if two Pt nanoparticles collide, reconstruction at the interface and the formation of metallic bonds most likely cannot occur readily Conclusions In summary, composition-dependent formation of Pt 53 Ag 47 worm-like nanowires provide an excellent platform to understand the key factors for the shape control of Pt alloys. The oriented attachment, which is contributed to the formation of metal nanoparticles in solution, 56 is the dominant mode of formation of Pt-Ag nanowires. DFT calculation and MD simulation data indicate the synergistic effect of several critical factors including composition, surface capping

100 Chapter agent and reaction temperature is essential for the shape control of colloidal nanocrystals. Author Contributions: This section is adapted from a paper to appear in ACS Nano (Z. M. Peng, H. J. You, H. Yang, Understanding the Composition-Dependent Formation of Platinum Silver Nanowires, 2010, DOI: /nn ). Professor Hong Yang, Zhenmeng Peng, and Hongjun You conceived and designed the experiments, analysed the data and co-wrote the paper. Zhenmeng Peng conducted the synthesis and performed the electron miscopy characterization. Hongjun You carried out the DFT calculation and MD simulation.

101 Chapter An Electrochemical Approach to Pt-Surface Rich PtAg Alloy Nanostructure Introduction Development of highly active and low cost catalysts is a major challenge that needs to be overcome to practical applications of proton exchange membrane fuel cells (PEMFCs) As platinum and its alloys have been the major choices of electrocatalysts for years in most of the low-temperature fuel cell devices, the synthesis of low Pt-content nanostructures is essential. An intriguing recent development is to create Pt-rich surface on particle supports made of less-noble metals The synthesis of such heterogeneous nanostructures, however, is not trivial, although limited success on either discontinuous particles or continuous thin films of platinum or a mixture of the both have been made. 99 The architecture with discontinuous coverage of platinum has been demonstrated based on the particle-on-particle or dendritic nanostructures Two major methods have been developed on producing continuous noble metal surfaces. Adzic and co-workers have used an under potential deposition (UPD) and redox replacement 107, 112, 113 method to produce platinum monolayer on selective metal surfaces. Copper is used as a sacrificial layer in the procedure because of the mismatch of redox potentials between platinum and the desirable supporting metals. The platinum skin layers obtained via UPD are highly active because of the electronic

102 Chapter and surface geometric contributions of the supporting metals. Tao et al. have recently developed a chemical reaction-driven restructuring method to produce core-shell bimetallic nanostructures. 114 Platinum or rhenium core and palladium shell nanoparticles are generated from their corresponding alloys by controlling the reaction atmosphere at elevated temperatures. The formation of such heterogeneous alloy nanostructure is thought to be the result of minimization of global surface energy. These studies show that engineering of colloidal alloy nanoparticles can potentially make a paradigm shift in the design of multifunctional catalysts. With the recent success on making mixed alloy nanoparticles including those 115, 116 metals with large bulk miscibility gaps, complex heteronanostructures should be obtainable through controlled chemical reactions. While gas phase reactions offer the advantage of in-situ tuning of surface compositions, the method is limited by the choice of possible materials. 114 Platinum shell and metal core nanoparticles have not been produced so far based on this chemical reaction-driven method. Electrochemical approach, on the other hand, can in principle offer a great deal of flexibility on the selective removal of non-pt metals from nanostructured alloys by applying predetermined potentials. In this section, I will describe a new approach to heterogeneous alloy nanostructures with Pt-surface rich PtAg alloy nanoparticles through a selective electrochemical removal of Ag atoms. The rationales for choosing PtAg

103 Chapter nanoalloys are multifold. First, the dealloying process should be favored, because these two metals have a large difference between their standard reduction potentials. Furthermore, PtAg mixed alloys form largely through chemical synthesis at nanometer scale, and are immiscible as bulk materials when temperature is below about 1190 o 115, 117 C. This phenomenon suggests that platinum should favor the homogeneous nucleation and growth and form stable metal form during the process of continuous removal of silver metal. Second, the d-band center (ε d ) for PtAg shift upwards in comparison with pure platinum. 118, 119 The changes in electronic structure alter the chemisorption toward a given 112, 118, 119 reactant species. Segregation of silver atoms can be favored on the PtAg alloy surface, which results in a large decrease in the number of neighboring 120, 121 platinum atoms. They help reduce the number of adsorption sites for poisoning species that slow down the electrocatalytic reactions in PEMFCs Thus, the resultant nanostructures can be a model system to understand the composition and electrocatalytic property relationship. Third, studies on catalytic performance of PtAg alloys are still quite limited, mostly because of the 115, 122, 126 availability of this class of metal alloys. This work shows that the removal of non-pt metals does not result in the dissolution of alloy nanoparticles, but the formation of Pt-surface rich core-shell like nanostructures. Using FAOR as the model reaction, it is further demonstrated that the optimal PtAg heteronanostructures can not only have much improved activity but also show

104 Chapter limited degradation over reaction cycles. While dealloying process has been shown to improve the electrocatalytic properties, little structural details are available for such heterogeneous alloy nanostructures Experimental Section Materials. Platinum acetylacetonate (Pt(acac) 2, Strem Chemicals, 98%) and silver stearate (Ag(St), Alfa Aesar) were purchased from VWR. Oleic acid (OA, 90%, technical grade), oleylamine (OAm, 70%, technical grade), 1, 2-hexadecanediol (HDD, 90%, technical grade), and diphenyl ether (DPE, 99%, ReagentPlus ) were from Aldrich. All chemicals and reagents were used as received without further purification. The reference catalysts of carbon-supported platinum (20 wt.% Pt) and silver (20 wt.% Ag) were from E-TEK, Inc. Preparation of low-pt content PtAg alloy nanoparticles. All experiments were carried out under argon using the standard Schlenk line technique. In a typical procedure, HDD (0.49 g or 1.9 mmol), OA (0.3 ml or 0.9 mmol), and OAm (0.3 ml or 0.9 mmol) were mixed with DPE (4 ml or 25.2 mmol) in a 25-mL three neck round-bottom flask and preheated to 200 C. In a separate flask, Pt(acac) 2 (7.0 mg or mmol) and Ag(St) (43.0 mg or mmol) were mixed with DPE (1 ml or 6.3 mmol) and heated to 80 C until the solid was completely dissolved. The latter solution, which had a light yellow color, was

105 Chapter injected into the flask preheated at 200 C and held for 1 h. PtAg alloy nanostructures with different molar ratios were synthesized from Pt(acac) 2 and Ag(St) under the condition that the total amount of these two precursors was mmol. After the reactions, the resultant mixtures were washed twice with 2 ml of hexane and 6 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. Preparation of Pt-surface rich PtAg heteronanostructures. Carbon black (Vulcan XC-72) was used to support the as-made PtAg nanoparticles. In general, the suspension of carbon black in 15 ml of toluene was sonicated for 1 h before the addition of PtAg nanoparticles dispersed in 5 ml of toluene. The amounts of PtAg nanoparticles were predetermined and varied from 15 wt.% to 20 wt.% of the final products. The resultant mixture was further stirred overnight before the solids were collected via centrifugation at 6000 rpm for 5 min. These carbon-supported PtAg nanoparticles (PtAg/C) were thermally treated at 300 C for 1 h in air at a rate of 2 C/min and further annealed at the same temperature for 1 h in a forming gas of 5 vol.% hydrogen in argon. The selective removal of Ag atoms was conducted electrochemically using a CHI 760 dual channel electrochemical workstation from CH Instruments, Inc. The three-electrode system consisted of a glassy carbon working electrode (5 mm in diameter), a platinum wire counter electrode, and a hydrogen reference electrode (HydroFlex, Gaskatel). The process began by dispersing PtAg/C in a mixture of deionized water, isopropanol and 5 wt.% Nafion solution

106 Chapter (V water /V 2-propanol /V 5% Nafion = 0.8/0.2/0.005), followed by sonication for 10 min. This mixture was drop-cast onto a glassy carbon electrode and dried under a stream of air. A typical amount of PtAg metals used in the sample was 0.5 µg, which was determined by thermal gravimetric analysis (TGA) recorded in air up to 800 C. A 0.1-M perchloric acid (HClO 4 ) aqueous solution was used as the supporting electrolyte. The set potentials for removing silver atoms were monitored using cyclic voltammetry (CV). Before each experiment, the solution was bubbled with argon for 30 min to remove dissolved oxygen. The silver dissolution experiment was performed by scanning predetermined potential ranges for 20 cycles at a scan rate of 50 mv/s and the RDE rotating rate of 1600 rpm unless stated otherwise. Characterization. Transmission electron microscopy (TEM) images were taken on a JEOL JEM 2000EX microscope at an accelerating voltage of 200 kv or a Hitachi 7100 microscope at 80 kv. High-resolution transmission electron microscopy (HR-TEM), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and energy dispersive X-ray (EDX) analysis were taken on an FEI TECNAI F-20 field emission microscope operated at 200 kv. The optimal resolutions of this microscope are 1 Å in TEM mode and 1.4 Å in STEM mode. EDX analysis on ensembles of nanoparticles was carried out on a field emission scanning electron microscope (FE-SEM, Zeiss-Leo DSM982) installed with an EDAX detector. Powder X-ray diffraction (PXRD)

107 Chapter patterns were recorded using a Philips MPD diffractometer with a Cu K α X-ray source (λ= Å). TGA was conducted on an SDT-Q600 simultaneous TGA/DSC system from TA Instruments, Inc. Electrochemical properties were measured on the CHI 760 dual channel electrochemical workstation using the three-electrode system described above. For comparison, carbon-supported platinum (20 wt.% Pt) and silver (20 wt.% Ag) catalysts were loaded onto the electrodes according to the same protocols described above, and the total amount of the metals used was fixed at 1 µg. An aqueous solution of HClO 4 (0.1 M) was used for the electrochemical active surface area (ECSA) measurements. A mixed aqueous solution of 0.1-M HClO 4 and 0.5-M formic acid was used as the electrolyte for the electrochemical oxidation of formic acid. Before each measurement, the solution was bubbled with argon for 30 min to remove dissolved oxygen. The CV curves were recorded after the initial 20 cycles at a scan rate of 50 mv/s and the RDE rotating rate of 1600 rpm unless stated otherwise Results and Discussion The mixed PtAg alloy nanoparticles were synthesized using the hot-injection 130, 131 method. Although platinum and silver are largely immiscible as bulk alloys at low temperatures, they could form stable mixed alloys from molecular 115, 116, 122 precursors using a colloidal synthetic approach. Unlike the heating-up

108 Chapter method, 115 nanoparticles produced at Pt(acac) 2 /Ag(St) molar ratio of 1/6 had a small average diameter of 3.9 ± 0.7 nm when the hot-injection method was used (Figure 2.17a). The size of these nanoparticles are in the optimal range for fuel cell catalyst. 94, 132 HR-TEM image shows the fringes of (111) plane in individual nanoparticles, indicating the existence of good crystalline order (Figure 2.17b). Figure 2.17c shows representative PXRD pattern of as-prepared PtAg nanoparticles and the positions of diffraction peaks for pure Ag ( , JCPDS-ICDD) and Pt metals ( , JCPDS-ICDD). The diffraction pattern of these nanoparticles could be readily indexed to (111), (200), (220), (311) and (222) planes of face-centered cubic (fcc) lattice and fell in between those of the two pure 115, 133 metals, suggesting the formation of mixed alloys. Figure 2.17 (a) Low and (b) high magnification TEM images and (c) PXRD pattern of PtAg alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratio of 1/6.

109 Chapter Figure 2.18 Representative TEM images of PtAg mixed alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratios of (a) 1/8, (b) 1/4, (c) 1/3, and (d) 1/2, respectively. The PtAg alloy nanoparticles with different compositions could be prepared by changing the molar ratio between Pt(acac) 2 and Ag(St) in the reaction mixtures. TEM images show that the nanoparticles formed were in a narrow size range when the total amount of Pt(acac) 2 and Ag(St) were kept constant (Figure 2.18). The nanoparticles formed had an average diameter of 3.3 ± 0.5 nm at Pt(acac) 2 /Ag(St) molar ratio of 1/2, and 3.8 ± 0.6 nm at 1/8. The size and size distribution of nanoparticles made at different precursor molar ratios are summarized in Table 2.3. EDX spectrum shows both platinum and silver were present in these nanoparticles (Figure 2.19a). Figure 2.19b shows the atomic percentage of platinum based on the total amount of metals in the precursors and

110 Chapter the resultant PtAg alloy nanoparticles as determined by EDX analysis. A linear relationship could be established between the amount of Pt in the precursors and final products. The atomic percentage of Pt in the resultant Pt x Ag y nanoparticles was only slightly higher than that in the precursors fed into the reaction mixtures, suggesting the amount of unreacted precursors was fairly low (Figure 2.19b). The compositions of the mixed PtAg alloy nanoparticles made are also summarized in Table 2.3. Figure 2.19 (a) Representative EDX spectrum of PtAg alloy nanoparticles made at Pt(acac) 2 /Ag(St) molar ratio of 1/6, and (b) relationship of Pt atomic percentage (at.%) in the products and their corresponding precursors. The Pt at.% was calculated based on the ratio of mole number of Pt over the total mole number of Pt and Ag metals.

111 Chapter Table 2.3 Particle size, size distribution, and composition of PtAg nanoalloys prepared at different Pt(acac) 2 /Ag(St) feeding ratios. Sample No. Pt(acac) 2 /Ag(St) As-prepared nanoparticles molar ratio Size (nm) Composition 1 1/2 3.3 ± 0.5 Pt 38 Ag /3 3.5 ± 0.7 Pt 31 Ag /4 3.4 ± 0.5 Pt 24 Ag /6 3.9 ± 0.7 Pt 18 Ag /8 3.8 ± 0.6 Pt 13 Ag 87 Figure 2.20 (a) Low and (b) high magnification TEM images and (c) PXRD pattern of carbon-supported Pt 18 Ag 82 mixed alloy nanoparticles after the thermal treatment. The controlled removal of silver metal was carried out using primarily Pt 18 Ag 82 alloy nanoparticles, which was prepared from the mixture at Pt(acac) 2 /Ag(St) molar ratio of 1/6. Carbon support was used in order to minimize the sintering of

112 Chapter resulting nanoparticles. These carbon-supported Pt 18 Ag 82 nanoparticles were thermally treated first in air and then in forming gas (5% H 2 in Ar) at 300 C. Figure 2.20a shows the Pt 18 Ag 82 nanoparticles were evenly distributed on carbon supports after the thermal treatment. The average diameter of the alloy nanoparticles grew slightly to 5.4 ± 1.6 nm and the size distribution became slightly broad as well. HR-TEM image shows that the particles had good crystallinity after the thermal treatment and lattice fringes could be observed readily throughout the whole particles (Figure 2.20b). The PXRD patterns of these thermally treated alloy nanoparticles contained all five diffraction peaks for the fcc lattice as observed in as-made samples (Figure 2.20c). The broad peak observed at about 25 2θ belonged to the graphitic carbon used as the support. The diffraction peaks for Pt 18 Ag 82 nanoparticles again fell in between those of Pt 115, 133 and Ag metals, and could be fit with Vegard s law. While the positions remained to be the same after this thermal treatment, the corresponding peaks of these alloys became sharp indicating the crystallinity and domain size of the nanoparticles was improved. No new crystal phases were formed and nanoparticles remained to be mixed alloys upon this thermal treatment. This observation could be understood since PtAg mixed alloys only began to phase separate above ~700 C and as ensembles of nanoparticles. 115 Silver metal can dissolve selectively from the PtAg mixed alloys using an electrochemical method because it has a much lower standard electrode potential

113 Chapter than platinum For carbon-supported Ag nanoparticles (Ag/C), the anodic reaction had an onset potential slightly above 0.6 V in the forward sweep and peaked at 0.88 V (Figure 2.21a). Silver was largely removed from the electrode after the Figure 2.21 CV curves for Ag dissolution from carbon-supported (a) Ag and (b-c) Pt x Ag y alloy nanoparticles at the different upper potentials ranging from 0.6 to 1.2 V at 0.1-V increment. The composition of the original alloy nanoparticles was Pt 18 Ag 82. For clarity, only the initial one and a half cycles are shown in (b) and the enlarged hydrogen adsorption regions are shown in (c). first cycle when the upper-limit potential was larger than 1 V, judging by the disappearance of the anodic reaction peak. The complete removal of Ag could also be accomplished through multiple cycles with an upper-limit potential equal

114 Chapter or even smaller than 0.8 V (Figure 2.22). The amount of Ag dissolved can be calculated from the CV curves based on the following equation: Ag E 2 n IdE Ag E1 Loss% = = (2.5) n n Fv Ag,0 Ag,0 where n Ag,0 is the loading amount of silver on electrode, n Ag is the amount of silver dissolved, I is the current with the capacitive signal deducted, E is the potential, E 1 and E 2 are the low and high end potential limits respectively, F is the Faraday constant, and v is the potential sweeping rate. Figure 2.22 CV curves of ten continuous cycles for Ag dissolution from carbon-supported Ag nanoparticles with a scan range between 0 and 0.8 (top) and 1.0 V (bottom), respectively. See experimental sections for details.

115 Chapter Figure 2.23 CV curves of ten continuous cycles for silver dissolution from carbon-supported Pt 18 Ag 82 nanoparticles with a scan range between 0 and a given upper potential limit of 0.6, 0.8, 1.0, or 1.2 V, respectively. See experimental sections for details. Based on this calculation, almost all the atoms in pure Ag nanoparticles could be dissolved through the electrochemical oxidation at the proper potential (Figure 2.21a). For carbon-supported Pt 18 Ag 82 alloy nanoparticles, the anodic currents associated with the dissolution of Ag were detected when the potential was higher than 0.6 V (Figure 2.21b). Unlike that in pure silver nanoparticles, for which currents were observed continually over each subsequent cycle until all the silver atoms were removed (Figure 2.22), the number of electrons transferred due to Ag dissolution from Pt 18 Ag 82 nanoparticles was dependent on the scan range. Current associated with Ag electro-oxidation was not detected in the subsequent cycles. This behavior was independent of the potential range for the scan (Figure 2.23).

116 Chapter This observation indicates that the dissolution of silver from Pt 18 Ag 82 /C should be completed mostly during the first sweep. Two anodic peaks centered at 0.73 and 0.98 V were detected for Pt 18 Ag 82 nanoparticles when the upper limit potential was above 1 V. The currents were much smaller than that for Ag metal under the same conditions, indicating partial dissolution of Ag from the mixed alloy. The formation of mixed alloys appeared to help preventing Ag from rapid dissolution. 134 The gradual dissolution of Ag from Pt 18 Ag 82 nanoparticles could result in the accumulation of Pt atoms on particle surface. The increase of hydrogen adsorption at potentials below 0.4 V suggests the formation of Pt-rich surfaces (Figure 2.21c). There was a strong dependence of Ag loss on the upper limit potentials, indicating this electrochemical method is a controllable approach to the control of compositions and fine nanostructures of PtAg mixed alloys. Figure 2.24 shows the TEM images and size distribution analysis of the resulting carbon-supported nanoparticles made from Pt 18 Ag 82 /C after potential cycles to 0.6, 1.0 and 1.2 V, respectively. Based on the TEM study, the particle size and size distribution did not seem to change much when the upper-limit potential was 0.6 V (Figure 2.24a-b). No anodic signal or electron transfer was detected in the corresponding CV curve (Figure 2.21b) and the composition of the resulting alloy nanoparticles remained to be the same based the EDX analysis.

117 Chapter Figure 2.24 Representative TEM images and size distribution analyses of carbon-supported nanoparticles made from Pt 18 Ag 82 alloy nanoparticles after the dissolution of Ag at the upper potentials of (a, b) 0.6, (c, d) 1.0, and (e, f) 1.2 V, respectively. See text for changes of compositions and nanostructures of the resulting alloy particles. When the upper potential changed to 1.0 V, the average diameter of the resulting particles decreased to 4.2 nm (Figure 2.24c-d). The average composition of the alloy nanoparticles changed from Pt 18 Ag 82 to Pt 41 Ag 59. The amount of silver loss and the reduction in particle size were estimated from the

118 Chapter CV curve based on the electron transfer according to Equation 2.5. In this calculation, the shape of nanoparticles was approximated as sphere, and electrons transferred due to the dissolution of Ag followed the Faradaic process. In addition, the lattice constant of the alloy remained to be around 4.05 Å and Pt atoms deposited evenly on the surface of particles during the Ag dissolution. With these assumptions, the PtAg alloy nanostructures should have different compositions and architectures when upper-limit potential was raised from 0.6 to 1.2 V (Table 2.4). When this potential was 1.0 V, about 57% of Ag atoms in the alloy dissolved. The resultant nanostructures should have an average particle size of 4.4 nm and overall average composition of Pt 34 Ag 66. The surface composition was calculated to be Pt 77 Ag 23. When the upper potential was raised to 1.2 V, the size and composition of the particles changed dramatically. The average diameter decreased further to about 3.2 nm and the overall composition became Pt 57 Ag 43 based on EDX analysis (Figure 2.24e-f). These values agreed very well with those from the calculations based on the CV curves, in which over 80% of silver atoms were removed, resulting in an overall composition of Pt 58 Ag 42 and an average size of about 3.6 nm (Table 2.4). Platinum atoms should cover essentially the whole surface of the nanoparticles after such treatment.

119 Chapter Table 2.4 Calculation of silver loss, overall and surface compositions, and diameter of the resultant Pt x Ag y heterogeneous nanostructures made from Pt 18 Ag 82 nanoparticles * Upper-limi Electron Ag Composition Diameter* t potential ** (µc) loss Overall Surface ** (nm) (V) (at.%) Pt 18 Ag 82 Pt 18 Ag Pt 18 Ag 82 Pt 19 Ag Pt 19 Ag 81 Pt 23 Ag Pt 24 Ag 76 Pt 40 Ag Pt 34 Ag 66 Pt 77 Ag Pt 49 Ag 51 Pt Pt 58 Ag 42 Pt 3.6 *: Calculations were based on the CV curves. **: Electron transfer was calculated from the first CV cycle. ***: The resultant nanoparticles were assumed to be spherical and the thickness of a single atomic layer was 0.23 nm, given the lattice spacing for (111) surface is 2.26 Å for Pt and 2.35 Å for Ag. Figure 2.25 Representative HR-TEM, STEM images and corresponding single-point EDX analyses of three different types of heterogeneous alloy nanostructures that had overall compositions of (a-d) Pt 18 Ag 82, (e-h) Pt 34 Ag 66, and (i-l) Pt 58 Ag 42, respectively.

120 Chapter The heterogeneity in composition of these PtAg alloy nanoparticles was studied by HR-TEM and HAADF-STEM (Figure 2.25). Lattice fringes could be observed from all three types of PtAg nanostructures. The compositions in the center and at the edge of individual nanoparticles were characterized using spot EDX analysis under HAADF-STEM mode. No major difference in Pt/Ag atomic ratio was found at the center and edge regions of the alloy nanoparticle treated in the potential range between 0 and 0.6 V (Figure 2.25b-d), indicating an even distribution of these two elements in these nanoparticles. EDX spot analysis showed a large increase in Pt/Ag atomic ratios at both the center (Pt 40 Ag 60 ) and edge (Pt 50 Ag 50 ) areas after the potential was cycled to 1.0 V (Figure 2.25f-h). A significant increase of relative intensity for Pt could be observed, especially near the edges of the nanoparticle, indicating a substantial increase of Pt at the near surface region. The silver signal became weak at the edge of nanoparticle after the potential increased further to 1.2 V (Figure 2.25j-l). The overall trend on the changes of metal compositions agreed quite well with the results from CV study. The heterogeneity in metal distributions across these three types of nanoparticles was further confirmed by EDX line scan (Figure 2.26). When the potential was below 0.6 V, both Ag L line and Pt M line had their maximum intensities at the center. The shapes of the line scans for Pt and Ag elements were similar, indicating a homogeneous distribution of these two metals across the alloy

121 Chapter nanoparticle (Figure 2.26a-b). With the cycle potential increased to 1 V, Ag L line still had the maximum intensity in the center, while the Pt M line became flat (Figure 2.26c-d). This observation indicates that electrochemical removal of silver was substantial and platinum at the surface region was enriched dramatically. Finally, with the potential up to 1.2 V, the intensity for Ag L line was low across the entire alloy nanoparticle though the weakest region was at the edges (Figure 2.26e-f). Figure 2.26 Representative STEM images (left) and their corresponding EDX line scans (right) of the three types of heterogeneous alloy nanoparticles that had overall compositions of (a,b) Pt 18 Ag 82, (c,d) Pt 34 Ag 66, and (e, f) Pt 58 Ag 42, respectively.

122 Chapter The convergent evidence from the HR-TEM, HAADF-STEM EDX and CV studies indicates that different types of alloy nanostructures exist after the electrochemical treatments (Figure 2.27). When the Pt 18 Ag 82 alloy nanoparticles were treated at 0.6 V or lower, i.e., below the onset potential for electrochemical oxidation of Ag, little dissolution of Ag occurred and the alloy composition remained largely intact (Figure 2.27). The mixed alloy nanoparticles changed to core-shell like heterogeneous nanostructures with a reduced diameter and an overall composition of Pt 34 Ag 66 upon the partial removal of silver metal at 1.0 V. The exposed Pt atoms reconstructed upon the dissolution of Ag and formed Pt-surface rich heterogeneous nanostructures (Figure 2.27). Thus, composition, size and architecture of PtAg alloy nanostructures could be synergistically controlled electrochemically. Figure 2.27 Schematic illustration of compositional and structural changes of the heterogeneous PtAg alloy nanoparticles made by controlled dissolution of Ag metal.

123 Chapter The electrocatalytic property of these heterogeneous PtAg nanoparticles produced from the Pt 18 Ag 82 alloys was characterized using FAOR as the model reaction. In FAOR, the specific current density depends heavily on the surface electronic and geometric structures of PtAg nanoalloys and both dehydrogenation and dehydration pathways are involved. 138 Among these catalysts, the one obtained after 20 CV cycles to an upper potential of 1 V, which had an overall composition of Pt 34 Ag 66 and surface composition of about Pt 77 Ag 23, exhibited the highest activity (Figure 2.28). The mass-specific current density (i mass ) was 3380 ma/mg-pt, more than six times of that for the commercial Pt catalyst reference (E-tek, 20 wt.% Pt), which had a current density of 502 ma/mg-pt at 0.6 V. The relative increase of i mass was even more pronounced at the even lower potential range, which is important for their applications as anode catalysts. The i mass values for these PtAg nanoparticles were 417 ma/mg-pt at 0.2 V and 2204 ma/mg-pt at 0.4 V. In comparison, mass-specific current density for pure Pt catalyst was only 31 ma/mg-pt at 0.2 V and 231 ma/mg Pt at 0.4 V. The increase in i mass could largely be attributed to its enhanced intrinsic activity, which can be described using the area-specific current density (i area ) based on the electrochemical active surface area (ECSA) (Figure 2.28c). The area-specific current density was 8.6 ma/cm 2 -Pt at 0.6 V for Pt 34 Ag 66 and nearly 13 times of that for the commercial Pt catalyst reference (0.67 ma/cm 2 -Pt). Similarly, the i area

124 Chapter of Pt 34 Ag 66 catalyst was about 25 times higher at 0.2 V or 18 times higher at 0.4 V than that for the Pt/C reference catalyst. Figure 2.28 CV curves showing the (a) mass current density (i mass ) and (b) area-specific current density (i area ) in the forward (solid line) and backward (dash line) scans, (c) bar graph illustrating both i mass and i area, and (d) changes in i mass at 0.6 V over multiple formic acid oxidation cycles catalyzed by the resultant heterogeneous Pt 34 Ag 66 and pure Pt electrocatalysts. The enhancement of FAOR activity for these heterogeneous PtAg nanoparticles and their large difference in performances after different treatments could be explained in terms of surface structures. The small fraction of surface silver could suppress the dehydration pathway by reducing the number of ensembles of adjacent platinum atoms, which are required for both the generation and oxidation

125 Chapter , 121, 123, 138 of absorbed CO species through a so-called third-body mechanism. The shift in the d-band electron center for platinum due to the incorporation of silver also helped Noticeably, only a slight decrease in specific mass current density was observed over multiple cycles, indicating very good stability (Figure 2.28d). Figure 2.29 CV curves showing the (a) area-specific current density (i area ) in the forward (solid line) and backward (dash line) scans, and (b) bar graph illustrating i area in formic acid oxidation catalyzed by the resultant heterogeneous PtAg alloy nanoparticles with overall compositions of Pt 18 Ag 82, Pt 34 Ag 66, and Pt 58 Ag 42. Besides the Pt 34 Ag 66 heterogeneous catalyst, two other types of Pt-Ag heterogeneous nanostructures with overall compositions of Pt 18 Ag 82 and Pt 58 Ag 42 were also characterized. The bulk and surface compositions for the Pt 18 Ag 82

126 Chapter catalyst were similar, while the Pt 58 Ag 42 catalyst most likely had a Pt surface (Figure 2.27). Not surprisingly, Pt 34 Ag 66 alloy catalyst had the highest i area among these three types of Pt-Ag heterogeneous catalysts (Figure 2.29). The i area for Pt 18 Ag 82 was 0.66 ma/cm 2 -Pt at 0.2 V and lower than that for the Pt 34 Ag 66 heterogeneous catalyst. The Pt 58 Ag 42 catalyst was largely inactive at this potential and had a negligible i area of ma/cm 2 -Pt (Figure 2.29b). Similar trends were observed for the current density measured at 0.4 and 0.6 V, where Pt 34 Ag 66 also exhibited much higher intrinsic activities than the other two Pt-Ag heterogeneous alloy and Pt catalysts. The large difference in activity among these Pt-Ag heterogeneous alloy catalysts could be associated with their different surface structures. These data also suggest Pt 77 Ag 23 (or Pt 3 Ag) represents the optimal surface composition of Pt-Ag alloy catalysts for FAOR Conclusions An electrochemical method has been developed to prepare heterogeneous PtAg alloy nanostructures with Pt-surface rich alloy nanoparticles. The architecture and composition of these heterogeneous alloy nanostructures can be obtained from the same type of PtAg nanoparticles. In comparison with UPD and the chemical 107, reaction-driven methods, this selective electrochemical dissolution approach is highly versatile and controllable in generating core-shell like alloy

127 Chapter nanostructures. The Pt-surface rich nanoparticles can potentially have broad ramifications in the development of novel catalysts and electrocatalysts. Author Contributions: This section is adapted from a paper to be published (Z. M. Peng, H. J. You, H. Yang, An Electrochemical Approach to Pt-Surface Rich PtAg Alloy Nanostructures, Advanced Functional Materials, 2010, accepted). Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, analysed the data and co-wrote the paper. Zhenmeng Peng conducted the synthesis and performed the electrochemical study. 2.4 References (1) Jeong, U.; Teng, X. W.; Wang, Y.; Yang, H.; Xia, Y. N. Adv. Mater. 2007, 19, 33. (2) Sun, S. H. Adv. Mater. 2006, 18, 393. (3) Wieckoweski, A.; Savinova, E. R.; Vayenas, E. G., Catalysis and Electrocatalysis at Nanoparticle Surface. Marcel Dekker, Inc.: New York, (4) Ozin, G. A.; Arsenault, A. C., Nanochemistry: A Chemical Approach to Nanomaterials. RSC Publishing Cambridge, 2005; p 628. (5) Koper, O.; Winecki, S., Specific Heats and Melting Points of Nanocrystalline Materials In Nanoscale Materials in Chemistry, Klabunde, K. J., Ed. John Wiley & Sons, Inc.: New York, 2001; pp 263. (6) Faraday Discuss.-Special Issue on Nanoalloys: From Theory to Application 2008, 138, 1.

128 Chapter (7) Sun, Y. G.; Xia, Y. N. Science 2002, 298, (8) Xia, Y. N.; Halas, N. J. MRS Bull. 2005, 30, 338. (9) Rodriguez-Gonzalez, B.; Burrows, A.; Watanabe, M.; Kiely, C. J.; Marzan, L. M. L. J. Mater. Chem. 2005, 15, (10) Alvarez-Puebla, R. A.; Ross, D. J.; Nazri, G. A.; Aroca, R. F. Langmuir 2005, 21, (11) Kariuki, N. N.; Luo, J.; Maye, M. M.; Hassan, S. A.; Menard, T.; Naslund, H. R.; Lin, Y. H.; Wang, C. M.; Engelhard, M. H.; Zhong, C. J. Langmuir 2004, 20, (12) Massalski, T. B.; Murray, J. L.; Bennett, L. H.; Baker, H., Binary Alloy Phase Diagrams. American Society for Metals: Metals Park, Ohio, (13) Xiao, S. F.; Hu, W. Y.; Yang, J. Y. J. Phys. Chem. B 2005, 109, (14) Sun, Y.; Zhuang, L.; Lu, J.; Hong, X.; Liu, P. J. Am. Chem. Soc. 2007, 129, (15) Shirinyan, A. S.; Wautelet, M. Mater. Sci. Eng. C-Biomimetic Supramol. Syst. 2006, 26, 735. (16) Shirinyan, A. S.; Gusak, A. M.; Wautelet, M. Acta Mater. 2005, 53, (17) Shirinyan, A. S.; Gusak, A. M. Philos. Mag. 2004, 84, 579. (18) Wautelet, M.; Dauchot, J. P.; Hecq, M. Nanotechnology 2000, 11, 6. (19) Adhikari, H.; Marshall, A. F.; Goldthorpe, I. A.; Chidsey, C. E. D.; McIntyre, P. C. ACS Nano 2007, 1, 415. (20) Tersoff, J. Appl. Phys. Lett. 2003, 83, 353. (21) Wang, G. F.; Van Hove, M. A.; Ross, P. N.; Baskes, M. I. Prog. Surf. Sci. 2005, 79, 28. (22) Mainardi, D. S.; Balbuena, P. B. Int. J. Quantum Chem. 2001, 85, 580. (23) Lequien, F.; Creuze, J.; Berthier, F.; Legrand, B. Faraday Discuss. 2008, 138, 105. (24) Mejia-Rosales, S. J.; Fernandez-Navarro, C.; Perez-Tijerina, E.;

129 Chapter Montejano-Carrizales, J. M.; Jose-Yacaman, M. J. Phys. Chem. B 2006, 110, (25) Liu, H. B.; Pal, U.; Perez, R.; Ascencio, J. A. J. Phys. Chem. B 2006, 110, (26) Cheng, D. J.; Huang, S. P.; Wang, W. C. Phys. Rev. B 2006, 74, (27) Van Hoof, T.; Hou, M. Phys. Rev. B 2005, 72, (28) Mainardi, D. S.; Balbuena, P. B. Langmuir 2001, 17, (29) Qin, F.; Jiang, C.; Anderegg, J. W.; Jenks, C. J.; Gleeson, B.; Sordelet, D. J.; Thiel, P. A. Surf. Sci. 2007, 601, 376. (30) Wang, G. F.; Van Hove, M. A.; Ross, P. N.; Baskes, M. I. J. Chem. Phys. 2005, 122, (31) Cheng, D. J.; Wang, W. C.; Huang, S. P. J. Phys. Chem. B 2006, 110, (32) Baletto, F.; Mottet, C.; Ferrando, R. Phys. Rev. Lett. 2003, 90, (33) Ouyang, G.; Tan, X.; Wang, C. X.; Yang, G. W. Nanotechnology 2006, 17, (34) Shirinyan, A.; Wautelet, M.; Belogorodsky, Y. J. Phys.-Condes. Matter 2006, 18, (35) Mori, H.; Yasuda, H. Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 2001, 312, 99. (36) Cable, R. E.; Schaak, R. E. Chem. Mat. 2005, 17, (37) Schaak, R. E.; Sra, A. K.; Leonard, B. M.; Cable, R. E.; Bauer, J. C.; Han, Y. F.; Means, J.; Teizer, W.; Vasquez, Y.; Funck, E. S. J. Am. Chem. Soc. 2005, 127, (38) Bond, G. C. Platinum Metals Review 2007, 51, 63. (39) Luo, J.; Maye, M. M.; Petkov, V.; Kariuki, N. N.; Wang, L. Y.; Njoki, P.; Mott, D.; Lln, Y.; Zhong, C. J. Chem. Mat. 2005, 17, (40) Hills, C. W.; Mack, N. H.; Nuzzo, R. G. J. Phys. Chem. B 2003, 107, 2626.

130 Chapter (41) Lee, J. G.; Mori, H. Phys. Rev. Lett. 2004, 93, (42) Nousiainen, O.; Kangasvieri, T.; Ronka, K.; Rautioaho, R.; Vahakangas, J. Solder Surf. Mt. Tech. 2007, 19, 15. (43) Nousialanen, O.; Rautioaho, R.; Kautio, K.; Jaaskelainen, J.; Leppavuori, S. Solder Surf. Mt. Tech. 2005, 17, 32. (44) Rhines, F. N., Phase Diagrams in Metallurgy: Their Development and Application. New York, McGraw-Hill: (45) Vegard, L.; Dale, H. Zeits. Krist. 1928, 67, 148. (46) Nuffield, E. W., X-Ray Diffraction Methods. Wiley: New York, (47) Medasani, B.; Park, Y. H.; Vasiliev, I. Phys. Rev. B 2007, 75. (48) Jiang, Q.; Liang, L. H.; Zhao, D. S. J. Phys. Chem. B 2001, 105, (49) Peng, Z. M.; Yang, H. Nano Today 2009, 4, 143. (50) Xia, Y.; Xiong, Y. J.; Lim, B.; Skrabalak, S. E. Angew. Chem.-Int. Edit. 2009, 48, 60. (51) Yin, Y.; Alivisatos, A. P. Nature 2005, 437, 664. (52) Tao, A. R.; Habas, S.; Yang, P. D. Small 2008, 4, 310. (53) Park, J.; Joo, J.; Kwon, S. G.; Jang, Y.; Hyeon, T. Angew. Chem.-Int. Edit. 2007, 46, (54) Sugimoto, T.; Shiba, F.; Sekiguchi, T.; Itoh, H. Colloid Surf. A-Physicochem. Eng. Asp. 2000, 164, 183. (55) Mullin, J. W., Crystallization. 3rd ed.; Oxford University Press: Oxford, (56) Zheng, H. M.; Smith, R. K.; Jun, Y. W.; Kisielowski, C.; Dahmen, U.; Alivisatos, A. P. Science 2009, 324, (57) Watzky, M. A.; Finney, E. E.; Finke, R. G. J. Am. Chem. Soc. 2008, 130, (58) Penn, R. L.; Banfield, J. F. Science 1998, 281, 969. (59) Cademartiri, L.; Ozin, G. A. Adv. Mater. 2009, 21, 1013.

131 Chapter (60) Tang, Z. Y.; Kotov, N. A. Adv. Mater. 2005, 17, 951. (61) Zhang, Q.; Liu, S. J.; Yu, S. H. J. Mater. Chem. 2009, 19, 191. (62) Pradhan, N.; Xu, H. F.; Peng, X. G. Nano Lett. 2006, 6, 720. (63) Banfield, J. F.; Welch, S. A.; Zhang, H. Z.; Ebert, T. T.; Penn, R. L. Science 2000, 289, 751. (64) Cademartiri, L.; Malakooti, R.; O'Brien, P. G.; Migliori, A.; Petrov, S.; Kherani, N. P.; Ozin, G. A. Angew. Chem.-Int. Edit. 2008, 47, (65) Teng, X. W.; Han, W. Q.; Ku, W.; Hucker, M. Angew. Chem.-Int. Edit. 2008, 47, (66) Peng, Z. M.; Yang, H. J. Solid State Chem. 2008, 181, (67) Fenske, D.; Borchert, H.; Kehres, J.; Kroger, R.; Parisi, J.; Kolny-Olesiak, J. Langmuir 2008, 24, (68) Fang, J. X.; You, H. J.; Kong, P.; Yi, Y.; Song, X. P.; Ding, B. J. Cryst. Growth Des. 2007, 7, 864. (69) Burda, C.; Chen, X. B.; Narayanan, R.; El-Sayed, M. A. Chem. Rev. 2005, 105, (70) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, 9. (71) Burch, R.; Breen, J. P.; Meunier, F. C. Appl. Catal. B - Environ. 2002, 39, 283. (72) Chen, J. Y.; Lim, B.; Lee, E. P.; Xia, Y. N. Nano Today 2009, 4, 81. (73) Sun, S. H. Adv. Mater. 2006, 18, 393. (74) Somorjai, G. A.; Rioux, R. M. Catal. Today 2005, 100, 201. (75) Lee, E. P.; Peng, Z. M.; Cate, D. M.; Yang, H.; Campbell, C. T.; Xia, Y. J. Am. Chem. Soc. 2007, 129, (76) Wang, Y.; Yang, H. J. Am. Chem. Soc. 2005, 127, (77) Zhang, Z. T.; Blom, D. A.; Gai, Z.; Thompson, J. R.; Shen, J.; Dai, S. J. Am. Chem. Soc. 2003, 125, 7528.

132 Chapter (78) Maksimuk, S.; Yang, S. C.; Peng, Z. M.; Yang, H. J. Am. Chem. Soc. 2007, 129, (79) Yang, S. C.; Peng, Z. M.; Yang, H. Adv. Funct. Mater. 2008, 18, (80) Chen, M.; Pica, T.; Jiang, Y. B.; Li, P.; Yano, K.; Liu, J. P.; Datye, A. K.; Fan, H. Y. J. Am. Chem. Soc. 2007, 129, (81) Wang, C.; Hou, Y. L.; Kim, J. M.; Sun, S. H. Angew. Chem.-Int. Edit. 2007, 46, (82) Liu, Q. S.; Yan, Z.; Henderson, N. L.; Bauer, J. C.; Goodman, D. W.; Batteas, J. D.; Schaak, R. E. J. Am. Chem. Soc. 2009, 131, (83) Delley, B. J. Chem. Phys. 1990, 92, 508. (84) Delley, B. J. Chem. Phys. 2000, 113, (85) Sun, H.; Mumby, S. J.; Maple, J. R.; Hagler, A. T. J. Am. Chem. Soc. 1994, 116, (86) Sun, H. J. Comput. Chem. 1994, 15, 752. (87) Sun, H. Macromolecules 1995, 28, 701. (88) Hill, J. R.; Sauer, J. J. Phys. Chem. 1994, 98, (89) Halder, A.; Ravishankar, N. Adv. Mater. 2007, 19, (90) Kittel, C., Introduction to Solid State Physics. Wiley: New York, (91) Kyle J. M. Bishop; Christopher E. Wilmer; Siowling Soh; Bartosz A. Grzybowski. Small 2009, 5, (92) Koparde, V. N.; Cummings, P. T. J. Phys. Chem. B 2005, 109, (93) Wieckowski, A.; Savinova, E. R.; Vayenas, C. G. Catalysis and Electrocatalysis at Nanoparticle Surfaces, Marcel Dekker, New York (94) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, 9. (95) Antolini, E. Mater. Chem. Phys. 2003, 78, 563. (96) Shao, Y. Y.; Yin, G. P.; Gao, Y. Z. J. Power Sources 2007, 171, 558.

133 Chapter (97) Yu, X. W.; Ye, S. Y. J. Power Sources 2007, 172, 133. (98) Chen, J. Y.; Lim, B.; Lee, E. P.; Xia, Y. N. Nano Today 2009, 4, 81. (99) Peng, Z. M.; Yang, H. Nano Today 2009, 4, 143. (100) Zhong, C. J.; Luo, J.; Njoki, P. N.; Mott, D.; Wanjala, B.; Loukrakpam, R.; Lim, S.; Wang, L.; Fang, B.; Xu, Z. C. Energy Environ. Sci. 2008, 1, 454. (101) Chen, Z. W.; Waje, M.; Li, W. Z.; Yan, Y. S. Angew. Chem.-Int. Edit. 2007, 46, (102) Tian, N.; Zhou, Z. Y.; Sun, S. G.; Ding, Y.; Wang, Z. L. Science 2007, 316, 732. (103) Liang, H. P.; Zhang, H. M.; Hu, J. S.; Guo, Y. G.; Wan, L. J.; Bai, C. L. Angew. Chem. -Int. Ed. 2004, 43, (104) Susut, C.; Nguyen, T. D.; Chapman, G. B.; Tong, Y. Electrochim. Acta 2008, 53, (105) Greeley, J.; Mavrikakis, M. Nat. Mater. 2004, 3, 810. (106) Chen, J. G.; Menning, C. A.; Zellner, M. B. Surf. Sci. Rep. 2008, 63, 201. (107) Adzic, R. R.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.; Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Top. Catal. 2007, 46, 249. (108) Peng, Z. M.; Yang, H. J. Am. Chem. Soc. 2009, 131, (109) Lim, B.; Jiang, M. J.; Camargo, P. H. C.; Cho, E. C.; Tao, J.; Lu, X. M.; Xia, Y. N. Science 2009, 324, (110) Peng, Z. M.; Yang, H. Nano Res. 2009, 2, 406. (111) Zhou, S. G.; McIlwrath, K.; Jackson, G.; Eichhorn, B. J. Am. Chem. Soc. 2006, 128, (112) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Angew. Chem.-Int. Edit. 2005, 44, (113) Zhang, J. L.; Vukmirovic, M. B.; Sasaki, K.; Nilekar, A. U.; Mavrikakis, M.; Adzic, R. R. J. Am. Chem. Soc. 2005, 127,

134 Chapter (114) Tao, F.; Grass, M. E.; Zhang, Y.; Butcher, D. R.; Renzas, J. R.; Liu, Z.; Chung, J. Y.; Mun, B. S.; Salmeron, M.; Somorjai, G. A. Science 2008, 322, 932. (115) Peng, Z. M.; Yang, H. J. Solid State Chem. 2008, 181, (116) Hwang, B. J.; Kumar, S. M. S.; Chen, C. H.; Chang, R. W.; Liu, D. G.; Lee, J. F. J. Phys. Chem. C 2008, 112, (117) Massalski, T. B.; Murray, J. L.; Bennett, L. H.; Baker, H. Binary Alloy Phase Diagrams, American Society for Metals, Metals Park, Ohio (118) Ruban, A.; Hammer, B.; Stoltze, P.; Skriver, H. L.; Norskov, J. K. J. Mol. Catal. A-Chem. 1997, 115, 421. (119) Hammer, B.; Norskov, J. K. Adv. Catal. 2000, 45, 71. (120) Greeley, J.; Norskov, J. K.; Mavrikakis, M. Annu. Rev. Phys. Chem. 2002, 53, 319. (121) Demirci, U. B. J. Power Sources 2007, 173, 11. (122) Xu, J. B.; Zhao, T. S.; Liang, Z. X. J. Phys. Chem. C 2008, 112, (123) Choi, J. H.; Jeong, K. J.; Dong, Y.; Han, J.; Lim, T. H.; Lee, J. S.; Sung, Y. E. J. Power Sources 2006, 163, 71. (124) Yu, X. W.; Pickup, P. G. J. Power Sources 2008, 182, 124. (125) Rice, C.; Ha, R. I.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.; Barnard, T. J. Power Sources 2002, 111, 83. (126) Schaal, M. T.; Hyman, M. P.; Rangan, M.; Ma, S.; Williams, C. T.; Monnier, J. R.; Medlin, J. W. Surf. Sci. 2009, 603, 690. (127) Koh, S.; Strasser, P. J. Am. Chem. Soc. 2007, 129, (128) Srivastava, R.; Mani, P.; Hahn, N.; Strasser, P. Angew. Chem.-Int. Edit. 2007, 46, (129) Strasser, P.; Koha, S.; Greeley, J. Phys. Chem. Chem. Phys. 2008, 10, 3670.

135 Chapter (130) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc. 1993, 115, (131) Jana, N. R.; Peng, X. G. J. Am. Chem. Soc. 2003, 125, (132) Kinoshita, K. J. Electrochem. Soc. 1990, 137, 845. (133) Vegard, L.; Dale, H. Zeits. Krist. 1928, 67, 148. (134) Snyder, J.; Asanithi, P.; Dalton, A. B.; Erlebacher, J. Adv. Mater. 2008, 20, (135) Lide, D. R. Ed. CRC Handbook of Chemistry and Physics, 88th Edition (Internet Version 2008), CRC Press/Taylor and Francis, Boca Raton, FL. (136) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410, 450. (137) Qian, L. H.; Chen, M. W. Appl. Phys. Lett. 2007, 91, (138) Neurock, M.; Michael, J.; Wieckowski, A. Faraday Discuss. 2008, 140, 363.

136 Chapter Chapter 3 Preparation and Electrocatalytic Property of Platinum Hollow Nanoparticles and Cubic Nanoboxes 3.1 Preparation and Electrocatalytic Property of Platinum Hollow Nanoparticles Made from Pt-on-Ag Heteronanostructures Introduction A challenge that needs to be overcome in PEMFCs is the slow kinetics of ORR catalyzed by nanometer (nm)-sized Pt metal particles in the cathode membranes. 1-4 While the mass current density of Pt catalysts for hydrogen fuel cells has improved dramatically during the past several years, further development is still urgently needed. 5, 6 In this context, design of novel platinum nanostructures has been the research focus of many scientists in search of next generation 1, 2, 7-14 high-performance electrocatalysts. Two areas have attracted particular attentions recently in the development of platinum-based cathode catalysts. Enhanced activity and improved stability in catalyzing ORR by Pt-on-Pd heteronanostructures or nanodendrites have been reported. 12, 15 The synthesis of such catalysts begins with the production of core Pd nanoparticles through the colloidal synthesis. Platinum nanoparticles grow subsequently onto the Pd core nanoparticles from molecular species Besides Pt-Pd bimetallics, a few other Pt-on-M (M=metal) heterogeneous nanoparticles

137 Chapter have also been synthesized and some have been tested for their electrocatalytic performance Another intriguing new class of Pt-based electrocatalysts is porous nanostructures which include tubes, hollows, porous dendrites and networks. Both Pt nanotubes and hollow spheres have been made and 22, 23 demonstrated to have higher ORR activities than pure Pt particles. Besides ORR reduction, some of these Pt nanostructures possess very good catalytic property in MOR Hollow and porous nanostructures of platinum bimetals and alloys have also been studied, and in some systems excellent catalytic properties 25, have been found. In this chapter, I will present the design and synthesis of Pt-on-Ag heterogeneous nanostructures and their applications in the preparation of a new kind of Pt hollow spheres consisting of continuous interconnected fine primary 28, 29, nanoparticles. Unlike the commonly used galvanic replacement method, these hollow nanostructures were prepared through a selective removal of silver cores. The platinum hollow structures can be controlled very well and show superior activity in ORR when compared with the commercial Pt catalysts. The PtAg mixed nanoalloys made from Pt-on-Ag nanoparticles through thermal treatment, on the other hand, had a much worse ORR activity than the same reference Pt catalyst Experimental Section

138 Chapter Synthesis. The Pt-on-Ag bimetallic heterogeneous nanoparticles were 12, 36, 37 prepared using a sequential synthetic method, which involved the generation of silver nanoparticles in organic mixtures and then the deposition of platinum metal. Air and water dissolved in the mixtures were removed by a vacuum pump and a flow of argon gas was used during the synthesis. Preparation of silver nanoparticles. The 9.5-nm silver nanoparticles were prepared using a modified procedure published elsewhere. 38 Specifically, silver trifluoroacetate (AgTFA, 98%, Aldrich, 0.22 g or 1 mmol) was mixed with oleylamine (OAm, 70%, Aldrich, 0.99 ml or 3 mmol) and isoamyl ether (99%, Aldrich, 5 ml) in a 25-mL three-neck flask and slowly heated to 160 o C in an oil bath in 80 min. The reaction was kept at this temperature for 90 min. The product was washed with 10 ml of ethanol, followed by centrifugation at 4000 rpm for 5 min. The precipitate was washed one additional time with 5 ml of ethanol and then redispersed in 10 ml of diphenyl ether (DPE, 99%, Aldrich). Preparation of Pt-on-Ag bimetallic heteronanostructures. Platinum was deposited onto surfaces of as-prepared Ag nanoparticles to produce Pt-on-Ag bimetallic heteronanostructures. In a typical procedure, a dispersion of Ag nanoparticles (50 µmol Ag metal) in DPE was added into a solution of platinum acetylacetonate (Pt(acac) 2, 98%, Strem, 0.02 g or 50 µmol), OAm (0.3 ml or 0.9 mmol) and DPE in a 25-mL three-neck flask. The total volume of DPE was fixed at 5 ml for all experiments. The mole number of metals added was

139 Chapter determined by thermal gravimetric analysis (TGA) conducted on an SDT-Q600 system from TA Instruments, Inc. The mixture was heated to 180 C using an oil bath and kept at this temperature for 1 h unless stated otherwise. The product was washed and precipitated out by adding 10 ml of ethanol as the anti-solvent and centrifugation at 6000 rpm for 5 min. The precipitate was collected and dispersed in 2 ml of toluene. To study the effects of key reaction conditions on the formation of Pt-on-Ag heteronanostructures, we varied the amount of Ag nanoparticles used between 25 to 100 mol, reaction temperature from 170 to 190 C, and reaction time were from 2 min to 2 h. Only one variable was adjusted each time while all other conditions were kept the same. Preparation of carbon-supported Pt-on-Ag nanoparticles. Carbon black (Vulcan XC-72) was used to support as-prepared Pt-on-Ag nanoparticles before any further treatment. In a typical procedure, 50 mg of carbon black was dispersed in 5 ml of toluene and sonicated for 1 h before the addition of Pt-on-Ag nanoparticles in toluene. The amount of Pt-on-Ag nanoparticles was predetermined using TGA analysis and fixed at 15 wt% of the final product. The mixed dispersions were stirred overnight before being collected via centrifugation at 6000 rpm for 5 min. The precipitate was allowed to dry at room temperature under ambient conditions. Thermal treatment of carbon-supported Pt-on-Ag bimetallic nanoparticles. Two different procedures were used to remove surface capping agents on the

140 Chapter Pt-on-Ag nanoparticles for the preparation of catalyst. One was an acid-based treatment used for the removal of surfactants on palladium nanoparticles. 39 To be specific, 30 mg of carbon-supported Pt-on-Ag nanoparticles were added into 10 ml of acetic acid and heated for 10 h at 70 C, where acetic acid reacted with the capped surfactants. The reaction mixture was cooled down to room temperature and washed for three times with 20 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. This procedure was repeated once with ethanol instead of acetone. The resulting product was dried at room temperature under ambient conditions. Alternatively, 30 mg of carbon-supported Pt-on-Ag nanoparticles was heated to 300 C in air at a ramping rate of 2 C/min and held at that temperature 12, 19 for 1 h in a tube furnace (Lindberg/Blue, Mini-Lite). The product was further treated at 300 C for 1 h under a reductive atmosphere of 5 vol% hydrogen in argon. With this thermal treatment, the Pt-on-Ag bimetallic nanoparticles turned into PtAg mixed nanoalloys. Preparation of Pt hollow nanostructures. Pt-hollow nanostructures were made via a selective removal of Ag metal from the Pt-on-Ag nanoparticles electrochemically. The experiment was conducted in a 125-mL five-neck flask using a CHI 760 dual channel electrochemical workstation (CH Instruments, Inc.). The three-electrode system consisted of a rotating disk working electrode made of glassy carbon (RDE, 5 mm in diameter), a platinum wire counter electrode, and a hydrogen reference electrode (HydroFlex, Gaskatel). The HydroFlex electrode

141 Chapter was calibrated by performing hydrogen evolution reaction (HER) with two Pt electrodes. All the potentials were recorded with respect to a reversible hydrogen electrode (RHE). 5 mg of acetic acid-treated carbon-supported Pt-on-Ag nanoparticles were dispersed in a mixture of deionized water, isopropanol and 5 wt% Nafion solution (V water /V 2-propanol /V 5% Nafion = 0.8/0.2/0.005), followed by sonication for 10 min. This mixture was deposited onto a RDE and dried under a stream of air. The amount of metals used in the experiment was 0.5 µg unless stated otherwise, which was determined by thermal gravimetric analysis (TGA). A 0.1-M perchloric acid (HClO 4 ) aqueous solution was used as the supporting electrolyte. The potentials for removing silver atoms were monitored using cyclic voltammetry (CV). Before each experiment, the solution was bubbled with argon for 30 min to remove dissolved oxygen. The potential was cycled between 0 and 1.3 V for 20 times at a scan rate of 50 mv/s and a RDE rotating rate of 1600 rpm to complete the dissolution. Characterization. A detailed description regarding characterization of the samples has been presented in Chapter 2.3. Briefly, TEM and HR-TEM images were taken on a FEI TECNAI F-20 microscope. STEM and elemental maps were carried out under a HAADF mode on the same microscope. EDX spectra were taken on a FE-SEM Zeiss-Leo DSM982. PXRD patterns were recorded using a Philips MPD diffractometer. UV-vis spectra were collected with a UV/VIS/NIR

142 Chapter spectrometer. The loading amount of metals on carbon was determined using an SDT-Q600 system. Electrochemical properties were measured on the same electrochemical workstation described above. Carbon-supported platinum (20 wt% Pt, E-tek) was used as a reference catalyst. The protocols for loading this catalyst onto the electrodes followed those described above, and the total amount of the metals used was fixed at 0.5 µg unless stated otherwise. An oxygen-free aqueous solution of HClO 4 (0.1 M), which was obtained through degassing with argon for 30 min, was used for the electrochemical active surface area (ECSA) measurement. After 20 cycles, The CV curves were recorded at a scan rate of 50 mv/s and the RDE rotating rate of 1600 rpm unless stated otherwise. All ORR tests were conducted at room temperature in an oxygen-saturated aqueous solution of HClO 4 (0.1 M). In all these ORR studies, the loading amount of catalysts was kept at 2 µg based on the amount of metals. The polarization curves were obtained by sweeping the potential from 0 to 1 V at the scan rate of 10 mv/s and rotating speed of 1600 rpm. To evaluate the long-term stability of Pt hollow nanoparticles, an accelerated test was conducted by running 30,000 linear potential sweeps between 0.6 and 1.0 V at a scan rate of 100 mv/s in an argon-protected 0.1-M HClO 4 aqueous solution Results and Discussion

143 Chapter Figure 3.1a shows a representative TEM image of as-synthesized silver nanoparticles made by using a procedure published elsewhere. 38 Silver trifluoroacetate was used as the precursor because it has good solubility in various organic solvents and can be thermally reduced to silver metal. 40, 41 The silver nanoparticles were fairly uniform and had an average size of 9.5 ± 1.3 nm. HR-TEM studies show that most of the silver nanoparticles had truncated octahedral shape (Figure 3.1b). These Pt-on-Ag nanoparticles were obtained by reducing Pt(acac) 2 at 180 C at Pt(acac) 2 /Ag-nanoparticle molar ratio of one. The reaction took 1 h to complete and oleylamine acted as both reducing and capping agent. 1, 42, 43 The Ag nanoparticles served as seed crystals for the growth of Pt-on-Ag heteronanostructures and could also assist the platinum reduction process. 44, 45 The Ag cores were mostly covered by multiple 3-nm Pt nanoparticles (Figure 3.1c). Figure 3.1d shows the HR-TEM image of an individual Pt-on-Ag nanoparticle. The Pt nanoparticles seemed to grow directly on the surfaces of Ag nanoparticles, since no clear discontinuity or obvious boundary was observed between their lattices. The distribution of Ag and Pt in the nanoparticles was studied by EDX analysis using the HAADF-STEM. Figure 3.2 shows representative STEM image of Pt-on-Ag nanoparticles and the corresponding Ag and Pt elemental maps. The Ag signal was readily detectable and concentrated in the cores of Pt-on-Ag nanostructures, judging by its distribution with respect to the particles as shown in

144 Chapter Figure 3.1 Representative TEM and HR-TEM images of as-prepared (a, b) Ag and (c, d) Pt-on-Ag nanoparticles, respectively. Figure 3.2 Representative HAADF-STEM images and their corresponding elemental maps of Pt-on-Ag nanoparticles at (a-d) low and (e-h) high magnifications. the STEM image (Figure 3.2b). The distribution of Pt metal was quite different from that of Ag in both location and intensity of the signal (Figure 3.2c). This difference was more pronounced when their elemental maps were compared using

145 Chapter overlays based on the signals from Pt-M and Ag-L lines (Figure 3.2d). The signals from Ag were completely surrounded by those from Pt. A careful STEM and EDX study on the individual Pt-on-Ag nanoparticle shows that the particle-on-particle morphology could be observed under the HAADF-STEM mode (Figure 3.2e). The Ag distributed mostly in central region, suggesting that core was consisted of Ag nanoparticle (Figure 3.2f). The distribution of Pt however looked quite different, with the EDX signals appearing both at the edge and in the center (Figure 3.2g). These converging data revealed an overgrowth of platinum on surface of silver nanoparticles as shown in the overlay of the nanoparticle (Figure 3.2h). The PXRD patterns were recorded for both as-prepared Ag and Pt-on-Ag nanoparticles (Figure 3.3a). The diffraction peaks can be indexed to (111), (200), (220), (311) and (222) planes of a face-centered cubic (fcc) structure. The well-defined shape and fairly good intensity of these peaks suggested reasonable crystallinity for both types of nanoparticles. The diffraction peaks became asymmetric, especially in the high-angle region, after Pt was deposited on the Ag nanoparticles. The diffractions from Pt metal were readily detectable and the corresponding XRD patterns can be deconvoluted into two sets of peaks for pure Ag and Pt, respectively. EDX analyses on the ensembles of Pt-on-Ag nanoparticles indicates that these particles had an empirical composition of Pt 43 Ag 57, close to the feeding ratio of one between Pt(acac) 2 and Ag particle precursors (Figure 3.3b and Table 3.1).

146 Chapter Figure 3.3 (a) PXRD patterns of as-prepared Ag (bottom) and Pt-on-Ag (top) nanoparticles, and (b) the corresponding EDX spectrum for Pt-on-Ag nanoparticles. Table 3.1 Summary of Pt-on-Ag heteronanostructures prepared under various reaction conditions Sample No. T ( C) Pt(acac) 2 /Ag molar ratio Morphology Atomic composition :1 Pt-on-Ag Pt 14 Ag :2 Pt-on-Ag Pt 29 Ag 71 3* 180 1:1 Pt-on-Ag Pt 43 Ag : :1 Pt-on-Ag and free Pt NPs Pt-on-Ag and free Pt NPs N/A N/A *: The optimal condition for making Pt-on-Ag nanoparticles, which were used subsequently for preparing Pt hollow structures, and Pt-on-Ag and PtAg alloy catalysts.

147 Chapter Figure 3.4 UV-vis spectra of Ag (0 min) and Pt-on-Ag nanoparticles made after the reaction for a time period between 2 and 120 min, respectively. UV-vis spectroscopy was used to monitor the growth process of Pt-on-Ag nanoparticles by taking out aliquots of the mixtures periodically. Ag nanoparticles are known to have a surface plasmon resonance (SPR) band in the visible range. The intensity of this SPR peak is sensitive to the surface coverage and was used to study the deposition of Pt. Figure 3.4 shows a series of spectra collected at different reaction times. The pure Ag nanoparticles exhibited a characteristic SPR peak at around 390 nm, close to the values reported. 38, 41 The intensity of Ag plasmon band gradually decreased over the reaction time, indicating a gradual increase in the surface coverage by Pt nanoparticles. Weak but detectable peak could still be observed after the reaction for 120 min. This

148 Chapter signal was associated with either a small fraction of exposed Ag surface or a damping effect on its SPR due to thin coating layer of Pt. Figure 3.5 Representative TEM and HR-TEM images of Pt-on-Ag nanoparticles made after the reaction for (a, b) 2, (c, d) 30, and (e, f) 120 min., respectively. TEM was used to further study the growth of Pt-on-Ag nanoparticles. No obvious large deposition of Pt on the surface of Ag nanoparticles after the reaction for 2 min (Figure 3.5a). HR-TEM study, however, shows that there were subtle changes in the morphology and sub nm-sized tiny particles were observed (Figure 3.5b). It seems that Pt nucleated and grew directly on the Ag surfaces. The size of the Pt nanoparticles became sufficiently large for imaging after reaction for 30 min (Figure 3.5c-d). This growth process continued until the precursors were exhausted after the reaction for about 120 min (Figure 3.5e-f). The observed particle-on-particle morphology resembled the product formed through either SK or VW mode. 1, 46, 47 These growth modes, governed by the interplay between

149 Chapter global surface energy and interfacial energy, led to the formation of either island-on-wetting-layer or isolated islands on substrate. Both the reaction temperature and Pt(acac) 2 /Ag metal molar ratio were varied systematically to optimize the conditions for the synthesis of pure Pt-on-Ag nanostructures. Both the population and size of Pt nanoparticles decreased dramatically if the reaction took place at 170 C for 1 h while keeping the Pt(acac) 2 /Ag particle precursor ratio at unity (Figure 3.6a). The resulting products had an appearance overall composition of Pt 14 Ag 86, indicating an incomplete reduction of Pt(acac) 2. When the reaction temperature was raised to 190 C while keeping all other conditions the same, large amount of free Pt nanoparticles formed besides the Pt-on-Ag nanoparticles, indicating the homogenous nucleation and growth became possible upon the reduction of Pt(acac) 2 under this temperature (Figure 3.6b). Figure 3.6 Representative TEM images of Pt-on-Ag nanoparticles made at (a) 170 and (b) 190 C, respectively. The reaction time was 1 h and Pt(acac) 2 /Ag metal molar ratio was kept at one.

150 Chapter If the Pt(acac) 2 /Ag nanoparticle molar ratio was lowered to 1/2 while keeping the reaction at 180 C for 1 h, the surface coverage of Ag nanoparticles by Pt was reduced, resulting in a composition of Pt 29 Ag 71 (Figure 3.7a). When Pt(acac) 2 /Ag nanoparticle molar ratio raised to 2, both Pt-on-Ag and individual free Pt nanoparticles were found (Figure 3.7b). These results are summarized in Table 3.1. The correlation between reaction parameters and final product can be understood in terms of reaction kinetics and a competing process between heterogeneous and homogeneous nucleation and growth. Low reaction temperature and Pt(acac) 2 concentration favor heterogeneous nucleation and growth since the energy barriers should be lower than that for its competitive process. 1 A low reaction temperature also leads to slow reaction kinetics according to the Arrhenius law. The optimal temperature for preparing Pt-on-Ag heteronanostructures with minimum free Pt nanoparticles were experimentally determined to be about 180 C at Pt(acac) 2 /Ag nanoparticle molar ratio of 1/1. Figure 3.7 Representative TEM images of Pt-on-Ag nanoparticles made at the Pt(acac) 2 /Ag metal molar ratio of (a) 1/2 and (b) 2/1, respectively. These reactions took placed at 180 C for 1 h.

151 Chapter We hypothesized that if the Pt-on-Ag nanoparticles resembled the products formed through the SK or VW growth mode, selective removal of Ag core might be possible to create Pt hollows if the Pt nanocrystals packed sufficiently dense. Such hollow nanostructures could have high specific surface areas and Pt primary particles in the optimal size range for catalyzing ORR. Pt-on-Ag and PtAg nanoparticles should not be the choice of catalysts, as incorporation of Ag often lowers the Pt catalytic activity towards ORR due to largely the electronic effect. 23, 48, 49 Figure 3.8 Representative TEM and HR-TEM images of carbon-supported (a, b) Pt-on-Ag nanoparticles, and the resulting products after (c, d) acid and (e, f) thermal treatments. Vulcan carbon was used as the support in this study in order to minimize the possible aggregation and sintering among the nanoparticles during the treatment.

152 Chapter Figure 3.8a-b show representative TEM images of carbon-supported Pt-on-Ag nanoparticles. The type of Pt-on-Ag nanoparticles used was made at 180 C with a Pt(acac) 2 /Ag metal molar ratio of 1/1 as shown in Figure 3.1 and had a composition of Pt 43 Ag 57. The loading amount was 16.3 wt% of metals on carbon. These Pt-on-Ag nanoparticles were uniformly dispersed on the carbon particles. Treating the particles with acetic acid or at 300 C in air and then a forming gas of diluted hydrogen in argon were the two methods used to remove the surface capping agents, mostly composed of oleylamine. The morphology of Pt-on-Ag nanoparticles was largely intact after the treatment with acetic acid (Figure 3.8c-d). The thermal treatment on the other hand, made the particle surfaces smooth, resulting in the loss of their dendritic structures (Figure 3.8e-f). Such morphological change is often the results of formation of alloy or other rearrangement processes. Figure 3.9 shows the PXRD patterns of these two types of products. The XRD pattern of acid-treated sample looked similar to the as-prepared Pt-on-Ag nanoparticles and the dominant diffraction was from the (111) plane of the relatively large Ag cores. The diffraction peak at around 25 2θ was from the graphitic carbon fragments of the carbon support. The (111) diffraction shifted dramatically for the sample with the treatment at 300 C and fell in between those for pure Pt and Ag metals. Such observation is a strong indication of the formation of PtAg mixed alloys. 50

153 Chapter Figure 3.9 PXRD patterns of acid-treated Pt-on-Ag nanostructures (bottom) and thermally-treated PtAg nanoparticles (top). Inset: enlarged region for the (111) diffractions. Carbon-supported platinum hollow nanostructures were produced from those acetic acid-treated Pt-on-Ag nanoparticles through the electrochemical dissolution of silver cores (Figure 3.10). The potential range used in this procedure was from 0 to 1.3 V and the supporting electrolyte was perchloric acid. The upper limit of the scan potential was well above the standard potential of 0.80 V for electrochemical oxidation of Ag metal, 51 and it should help in removing the organic residues on the particle surfaces after the acetic acid treatment. Figure 3.10a shows representative TEM images of the hollow particles that had the characteristic contrast difference between the center and edge regions. HR-TEM image show the hollow particles had good crystallinity with clearly visible (111) lattices of fcc phase (Figure 3.10b). STEM image and the corresponding

154 Chapter elemental maps based on Ag-L and Pt-M lines reveal that the shell of hollow structure was primarily made of Pt and negligible amount of Ag was detected (Figure 3.10c-e). It is worthwhile to point out the fact that continuous shell of hollow structures can be produced provides additional proof for a likely Stranski-Krastanov growth mode of Pt on Ag cores. 47 Unlike the products obtained after exposure to acid acid, PtAg alloy formed after heat treatment. Silver metal could not be removed electrochemically from these nanoparticles. Both Pt and Ag elements distributed uniformly in such nanoparticles, as shown by the TEM and EDX studies (Figure 3.11). Figure 3.10 Representative (a) TEM, (b) HR-TEM, (c) STEM image, and (d, e) elemental maps of carbon-supported Pt hollow structures made from carbon-supported Pt-on-Ag nanoparticles.

155 Chapter Figure 3.11 Representative (a) TEM, (b) HR-TEM, (c) STEM image and the corresponding (d) Ag and (e) Pt elemental maps of carbon-supported PtAg alloy nanoparticles after the same acid treatment as that for making Pt hollow nanostructures from Pt-on-Ag nanoparticles. Carbon-supported Pt hollow nanoparticles (Pt-hollow), together with Pt-on-Ag, PtAg and Pt (Pt/C, 20 wt%, E-TEK) catalysts, were studied for their ORR activities. The carbon-supported Pt-on-Ag nanoparticles show little hydrogen adsorption-desorption (HAD, E< 0.4 V) indicating these particles were largely electrochemically inactive. Although the electronic effect due to the incorporation of Ag into Pt could not be ruled out, 48, 49 incomplete removal of surface capping agents using acetic acid should be the key factor contributing to the low hydrogen adsorption on Pt-on-Ag nanoparticles, as HAD signals were observed for both PtAg nanoalloys and Pt-hollow structures (Figure 3.12). The ECSA values were calculated to be 34.6 m 2 /g Pt for PtAg nanoalloys and 60.9

156 Chapter m 2 /g Pt for Pt-hollow nanostructures. In comparison, the ECSA for the Pt reference from E-tek was 83.5 m 2 /g Pt. The relatively high surface area for the commercial Pt catalysts is because of the small particle size. 12 Figure 3.12 CV curves of carbon-supported Pt nanoparticle (E-TEK, 20 wt%), hollow, and PtAg alloy catalysts. All tests were performed in a 0.1-M HClO 4 aqueous solution and the CV curves were recorded at a scan rate of 50 mv/s. While it had lower electrochemical surface area than the Pt reference, Pt-hollow nanostructures exhibited higher activity than both Pt and PtAg alloy catalysts in ORR. Figure 3.13a shows the ORR polarization curves for these four different types of catalysts. The Pt-on-Ag nanostructures were least active, having a detectable onset current below about 0.5 V. The large negative on-set potential shift in comparison to the Pt reference catalyst revealed a poor ORR activity of PtAg catalyst. The ORR polarization curve for Pt hollows almost overlapped with the one for the commercial Pt nanocatalysts, even though these hollow

157 Chapter particles had a smaller ECSA and lower Pt loading on the RDE than the Pt reference catalyst. Figure 3.13 (a) ORR polarization curves, (b) intrinsic mass current density, (c) CV, and (d) hydroxyl surface coverage (Θ OHad ) of carbon-supported Pt-on-Ag nanoparticle, PtAg alloy, Pt hollow and particle reference catalysts. The intrinsic current density of these catalysts were calculated by applying the Koutecky-Levich equation, 52, 53 which can be depicted as follows: = + = + 1/ 2 (3.1) i i i i Bω k d k where, i is the measured current density, i k is the kinetic current density, i d is the diffusion (mass-transfer) limited current density, B is a constant which is a function of concentration (C O2 ) and diffusion coefficient (D O2 ) of O 2 in the electrolyte and the viscosity of electrolyte (ν), and ω is the rotation rate of the

158 Chapter electrode in unit of rpm. Based on this equation, the Pt hollow catalyst had an intrinsic mass current density of 322 µa/µg Pt at 0.9 V, almost twice the value of 187 µa/µg Pt for Pt nanoparticles (Figure 3.13b). The improvement in area-specified current density was even more pronounced, which was 529 µa/cm 2 Pt at 0.9 V and over two times of the value of 224 µa/cm 2 Pt for Pt nanoparticles (Figure 3.14). 12 Both were higher than the PtAg alloy made from Pt-on-Ag nanoparticles. As indicated above, the low ORR activity of PtAg alloy catalysts can be partially explained in terms of the electronic structures, i.e., an up-shift in 48, 49 the d-band of Pt due to the interaction with Ag metal. The result of such up-shift in d-band structure is an increased coverage of OH group on the catalyst 8, 10 surfaces. The OH coverage on catalysts is detrimental to the ORR activity. A visible difference in OH coverage was also observed between the Pt hollow and Figure 3.14 Area-specific current density of oxygen reduction reaction by using carbon-supported catalysts of Pt nanoparticles (E-TEK, 20 wt%), hollows and PtAg alloy nanostructures.

159 Chapter reference catalysts. As shown in Figure 3.13c, the value of OH coverage is the quotient of OH-adsorption area (shaded area on forward sweep between 0.6 and 1.2 V) divided by the overall active surface area (shaded area of backward sweep between ~ 0.05 and 0.4 V) in the CV curves. The Pt hollow catalyst clearly had a lower OH ad coverage than that of Pt reference over the entire potential range (Figure 3.13d). Small size of the Pt primary nanocrystals of a continuous shell could be an important structural factor contributing to this enhancement of ORR 2, 10, 54 activity. These Pt hollow nanostructures were physically stable after multiple electrochemical cycles. In this accelerated stability test, potential was continuously swept between 0.6 and 1.0 V linearly for 30,000 times in a 0.1-M HClO 4 aqueous solution. TEM images show the overall structure of the hollows retained largely intact, although the surface topology changed substantially (Figure 3.15a). HR-TEM study indicates that the surface of these Pt hollows became smooth and the population of dendrite-like features diminished after the multiple cycles (Figure 3.15b). Figure 3.15 Representative (a) TEM and (b) HR-TEM images of carbon-supported platinum hollow structures after an accelerated electrochemical stability test. Oxidation-reduction cycles were carried out through applying

160 Chapter ,000 linear potential sweeps between 0.6 and 1.0 V Conclusions Pt-on-Ag bimetallic heteronanostructures can be obtained by reducing platinum precursors in presence of Ag nanoparticles. The formation of Pt-on-Ag nanoparticles is shown to follow a heterogeneous nucleation and growth process resembling the products formed through the Stranski-Krastanov growth mode. Platinum hollow nanostructures can be obtained from Pt-on-Ag nanoparticles using an electrochemical method. The Pt hollow nanostructures are more active than a commercial reference Pt catalyst in oxygen reduction reaction, suggesting the hollow nanostructures contribute to the enhanced catalytic activity. Author Contributions: This section is adapted from a paper published in Chemistry of Materials (Z. M. Peng, J. B. Wu, H. Yang, Synthesis and Oxygen Reduction Electrocatalytic Property of Platinum Hollow and Platinum-on-Silver Nanoparticles, 2010, 22, ). Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, analysed the data and co-wrote the paper. Zhenmeng Peng conducted the synthesis and performed the electrochemical study. Jianbo Wu helped with the synthesis of platinum-on-silver nanoparticles.

161 Chapter An Electrochemical Restructuring Method for Making Ultrafine Platinum Cubic Nanoboxes Introduction Metal nanostructures with hollow interiors have attracted a great deal of attention in recent years because they have both unique structural features such as high specific surface, low density and large void ratio, and important electronic, optical, and catalytic properties. 55, 56 For example, hollow palladium nanoshperes have been reported as good catalyst in Suzuki cross coupling reactions. 57 Both platinum nanotubes and hollow spheres can exhibit higher activity than pure platinum particles in ORR and MOR Gold nanocages have been conjugated with monoclonal antibodies for biological imaging and cancer therapy, and can also be modified with polymers to enable a controlled release of pre-loaded anticancer drugs and enzymes using near-infrared light. 65 Metal hollow nanostructures have been made in several different morphologies including box, cage, sphere and tube. Templating method has long been used as the main approach to the synthesis of metal hollow nanostructures. 55 In general, a metal nanoparticle is used as the template for depositing the designed metals using the so-called galvanic replacement reaction. The metal shells are generated from metal salt precursors through the oxidation of metal core templates. The shapes of the resulting hollow structures are largely determined by those of

162 Chapter the templates. Currently, sizes of those shape-controlled hollow nanostructures are usually in the range of tens of nanometers although smaller hollow structures would be preferred in catalysis and many other applications since large specific surface area is often required for high mass specific activity. Beside the size of hollow nanoparticles, shape is another important factor that needs to be considered in order to produce highly activity catalysts. Making shape-defined metal hollow structures at sub 10-nm small size range is still quite challenging and there is no reliable method available yet. In this section I will present a new method for making size and shape fairly uniform Pt cubic nanoboxes with an edge length of about 6 nm and a wall thickness of 1.5 nm or about 4-unit cell thick. These cubic nanoboxes are produced through electrochemical reconstruction of Pt hollow nanospheres which are made using truncated octahedral and other faceted Ag nanoparticles as the templates in a solution phase synthesis. We further demonstrate shape-dependent catalytic property of these nanoboxes in MOR. The enhanced catalytic activity can be attributed to the cubic morphology in which (100) surface is dominated and preferred for direct oxidation of methanol Experimental Section In general, Pt cubic nanoboxes were obtained by electrochemically removing Ag from Pt-on-Ag bimetallic heterogeneous nanoparticles. The bimetallic

163 Chapter heteronanostructures were prepared using a sequential deposition method. 66 The synthetic mixtures were degassed for 5 min before the reaction which was performed under a flow of argon gas. Preparation of Pt-on-Ag nanoparticles. The 9.5-nm silver nanoparticles were synthesized using a modified procedure published elsewhere. 67 Specifically, silver trifluoroacetate (AgTFA, 98%, Aldrich, 0.22 g or 1 mmol) was mixed with oleylamine (OAm, 70%, Aldrich, 0.99 ml or 3 mmol) and isoamyl ether (99%, Aldrich, 5 ml) in a 25-mL three-neck flask and slowly heated to 160 C in an oil bath at a ramping rate of 2 C/min. The reaction was kept at 160 C for 90 min. After the reaction, the product was washed with 10 ml of ethanol, followed by centrifugation at 4000 rpm for 5 min. The precipitate was washed again with 5 ml of ethanol and then redispersed in 10 ml of diphenyl ether (DPE, 99%, Aldrich). Platinum was deposited on these Ag nanoparticles to produce Pt-on-Ag bimetallic heteronanostructures. Specifically, a dispersion of Ag nanoparticles (100 µmol based on the amount of Ag) in DPE (1 ml) was added into a mixture of platinum acetylacetonate (Pt(acac) 2, 98%, Strem, 0.02 g or 50 µmol), OAm (0.3 ml or 0.9 mmol) and DPE (4 ml) in a 25-mL three-neck flask. The mixture was heated to 180 C in 15 min using an oil bath and kept at this temperature for 1 h. The final product was mixed with 10 ml of ethanol and centrifuged at 6000 rpm for 5 min. The precipitate was collected and dispersed in 2 ml of toluene.

164 Chapter Preparation of carbon-supported Pt-on-Ag nanoparticles. Carbon black (Vulcan XC-72) was used to support as-prepared Pt-on-Ag nanoparticles before further treatment. In a typical procedure, 50 mg of carbon black was dispersed in 5 ml of toluene and sonicated for 1 h before the addition of Pt-on-Ag nanoparticles. The amount of Pt-on-Ag nanoparticles was determined using thermogravimetric analysis (TGA) and fixed at 10 wt% of the final product. The mixtures were stirred overnight before the final products were collected via centrifugation at 6000 rpm for 5 min. The precipitate was dried at room temperature under ambient conditions. These Pt-on-Ag nanoparticles were treated with acetic acid first before the electrochemical dealloying process. 68 To be specific, 30 mg of carbon-supported Pt-on-Ag nanoparticles were added into 10 ml of acetic acid and heated for 10 h at 70 C. This mixture was cooled down to room temperature and washed for three times with 20 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. This procedure was repeated once with acetone instead of ethanol. The resulting product was dried at room temperature under ambient conditions. Preparation of Pt hollow nanostructures. Pt hollow nanostructures were made after the removal of Ag from the Pt-on-Ag nanoparticles electrochemically. The synthesis was conducted in a 125-mL five-neck flask using a CHI 760 dual channel electrochemical workstation (CH Instruments, Inc.). The three-electrode system is consisted of a rotating disk working electrode that is made of glassy

165 Chapter carbon (RDE, 5 mm in diameter), a platinum wire counter electrode, and a hydrogen reference electrode (HydroFlex, Gaskatel). The HydroFlex electrode was calibrated by performing hydrogen evolution reaction (HER) with two Pt electrodes. All the potentials were recorded against a reversible hydrogen electrode (RHE). Five milligrams of acetic acid-treated carbon-supported Pt-on-Ag nanoparticles were dispersed in a mixture of deionized water, isopropanol and 5 wt% Nafion at V water /V 2-propanol /V 5% Nafion ratio of 0.8/0.2/0.005, followed by sonication for 10 min. This mixture was deposited onto a RDE and dried under a stream of air. The amount of metals used in the experiment was 0.5 µg unless stated otherwise. The mass used in the preparation was determined by TGA (SDT-Q600, TA Instruments, Inc). A 0.1-M perchloric acid (HClO 4 ) aqueous solution was used as the supporting electrolyte in the removal of Ag. The potential for removing silver was monitored based on cyclic voltammetry (CV). Before each experiment, the solution was purged with argon for 30 min to remove dissolved oxygen gas. The potential was cycled between 0 and 1.3 V for 20 times at a scan rate of 50 mv/s and a RDE rotating rate of 1600 rpm to complete the dissolution. The electrode was then transferred into a 0.5-M sulfuric acid (H 2 SO 4 ) aqueous solution for further electrochemical treatment. A linear potential was cycled for 3,000 times in order to prepare Pt cubic nanoboxes. The optimal potential range was between 0.6 and 1.0 V at the scan rate of 100 mv/s. The effects of scan range,

166 Chapter rate, and number and profile of cycles on the shape of these Pt hollow nanostructures were systematically examined by adjusting one parameter at a time while keeping all other variables unchanged. Characterization. A detailed description regarding characterization of the samples has been presented in Chapter 2.3. Briefly, TEM and HR-TEM images were taken on a FEI TECNAI F-20 microscope. STEM and elemental maps were carried out under a HAADF mode on the same microscope. EDX spectra were taken on a FE-SEM Zeiss-Leo DSM982. PXRD patterns were recorded using a Philips MPD diffractometer. The loading amount of metals on carbon was determined using an SDT-Q600 system. Electrochemical properties were measured on the same electrochemical workstation described above. The total amount of the metals used in each test was fixed at 0.5 µg unless stated otherwise. An 0.5-M H 2 SO 4 aqueous solution was degassed with argon for 30 min before the electrochemical active surface area (ECSA) measurement. The CV curves were recorded at a scan rate of 20 mv/s after 20 cycles, unless stated otherwise. Methanol oxidation reaction (MOR) was conducted at room temperature in an oxygen-free aqueous mixture of H 2 SO 4 (0.5 M) and MeOH (0.5 M), and the CV curves were recorded after 50 potential cycles. The potential was cycled between 0.05 and 1.2 V at a scan rate of 50 mv/s. The loading of catalysts was kept at 0.5 µg based on the amount of metals in all the MOR studies.

167 Chapter Results and Discussion The Pt nanoboxes were prepared by treating Pt-on-Ag nanoparticles electrochemically using different potential cycling profiles. The Ag nanoparticle cores had an average diameter of 9.5 nm and were faceted (Figure 3.16a). They were synthesized by reducing silver trifluoroacetate (AgTFA) in isoamyl ether in the present of oleylamine (OAm) at 160 C, 67, and used as the supports for the growth of Pt nanoparticles. Platinum acetylacetonate (Pt(acac) 2 ) deposited on the surface of Ag nanoparticles at 180 C to form Pt-on-Ag heteronanostructures (Figure 3.16b). These Pt-on-Ag nanoparticles were then supported on carbon before being further treated with acetic acid to remove the surface capping 66, 68 agents. Twenty potential cycles were carried out between 0.0 and 1.3 V in a 0.1-M perchloric acid (HClO 4 ) aqueous solution to remove Ag metal cores. The sample was then transferred into a 0.5-M sulfuric acid (H 2 SO 4 ) aqueous solution for further treatments under different potential cycling conditions. Figure 3.17 shows representative TEM images of the nanoboxes collected after 3,000 linear sweeps with a scan range between 0.6 and 1.0 V, and at a scan rate of 100 mv/s. Most of these Pt nanostructures were cube-like nanoboxes and had fairly uniform size. The average edge length of these cubes was about 6 nm. The contrast difference between the core and shell regions suggests that these nanocubes were hollow. These Pt cubic nanoboxes had an average wall

168 Chapter thickness of around 1.5 nm, which is equivalent to ~4 unit cell length. HR-TEM image of an individual nanobox reveals that the d-spacing was 2.01 Å for those lattices normal to the surfaces and 2.26 Å for those in 45 angle with the surfaces (Figure 3.17b). These fringes can be readily indexed to {100} and {111} planes of pure Pt metal, further indicating that the hollow structures were Pt nanoboxes bound by {100} surfaces. As TEM images are two-dimensional (2D) projections of 3D objects, cubic nanoboxes can be projected as a range of different hollow 2D structures, ranging rectangle to hexagon depending on the angle between imaging electronic beam and the position of boxes on the grid (Figure 3.17c). Under certain angles, such as β=-30, the difference in contrast could disappear between core and shell regions for the cubic nanoboxes. Figure 3.16 TEM images of as-prepared (a) Ag and (b) Pt-on-Ag nanoparticles.

169 Chapter Figure 3.17 Representative TEM images of Pt hollow nanocubes at (a) low and (b) high magnifications, and (c) individual cubes imaged under various tilting angles with respective to the direction of imaging beam. To further elucidate the structure and chemical composition of nanoboxes, we conducted the HR-TEM and energy dispersive X-ray (EDX) analysis with individual nanoboxes (Figure 3.18). EDX analysis shows the cubic nanoboxes were made of pure Pt, and no Ag can be detected in the final product (Figure 3.18a-c). The observed Cu signals were from the grid used in our EDX study. EDX line-scan in combination with the HAADF-STEM image shows the Pt

170 Chapter Figure 3.18 (a) HR-TEM image, (b) EDX spectrum, (c) HAADF-STEM images with a Pt-M line scan and (d) TEM images taken at different rotating angles of the stage. The characterizations were done with the same Pt hollow cube, except EDX spectrum shown in (b), which was obtained using large amount of Pt nanoboxes with the SEM technique. elemental distribution in the nanobox (Figure 3.18c). The sharp contrast between the center and edge of the nanobox in the HAADF-STEM image is evidence strong indication of the hollow structure. The Pt-M line cross-section scans show a saddle-type shape, with the weakest signal in the core region and intensive peaks at both ends. By rotating along either x (α) or y (β) axis at a step angle of 10, we could unambiguously confirm that the Pt particle was hollow

171 Chapter despite there existed low contrast difference across the cubes under some imaging angles, such as β= 30 (Figure 3.18d). Both square and rhombic shapes were observed for the same hollow cube when the tilting angles of the TEM grid relative to imaging beam direction changed. The cube was projected as a rhombic shape in the 2D micrographs, when the tilting angles along the y-axis (β) were -10, -20 and -30, respectively. The nearly perfect cubic nanobox could be projected as a truncated shape with round corners. For instance, at the tilting angles of α=30 or β=20, parts or all of the sharp external corners of the square shown in Figure 3.18a became round. Figure 3.19 Representative TEM images of various morphologies of Pt nanostructures obtained with different potential scanning (a, b) range, (c) profile, and (d) rate.

172 Chapter As the Ag nanoparticle templates had truncated octahedral or other non-cubic shapes, the formation of Pt cubic nanoboxes should be quite different from the Au nanoboxes made through galvanic replacement reaction. Experimentally, we observed that several parameters can play important roles in determining the final morphology of the Pt nanostructures. The rate of scan, number of cycle, range and profile of the applied potentials are some of the critical ones. Change in one or more of these parameters could result in the formation of ill-defined morphologies other than nanoboxes (Figure 3.19). For instance, if the potential scan range was changed to between 0.6 V and 0.8 V in the CV cycles, hollow nanospheres instead of cubic nanoboxes were obtained (Figure 3.19a). It appears that under this condition, the surface Pt atoms show limited mobility and the hollow shells largely preserved the truncated shapes of the Ag templates. The surface of these hollow nanoparticles was relatively rough. On the other hand, if the scan range increased to 1.1 V, the Pt shells could collapse to form hollows with reduced void space and overall particle size (Figure 3.19b). Dense nanoparticles were also observed, indicating extended reconstruction happened. The hollow particles formed after the treatment with upper-limit pontential (E u ) of 1.1 V typically were largely spherical and had smooth surfaces (Figure 3.20a). This size reduction process was accelerated with a further increase of E u to 1.2 V. Hollow nanospheres with reduced size could be observed after 800 linear scans (Figure 3.20b). These hollow spheres became solid and had even smaller sizes

173 Chapter after 1500 scans. They eventually turned into 2-nm Pt dense solid nanoparticles after 3000 scans (Figure 3.20c). This observation can be readily understood as the electrochemical oxidation of Pt metal occurs at about 1.2 V, which can result 73, 74 in rapid dissolution of Pt species. Thus electrochemical reconstruction at E u of 1.2 V became too harsh for conversion of Pt nanostructures into cubic nanoboxes. Figure 3.20 TEM images of Pt nanostructures after linear potential cycles in the range of (a) V and (b-c) V, respectively. The total numbers of cycles were 3000 for those samples shown in (a) and (c), and 800 for that shown in (c). The scan profile can also affect the final shape of Pt nanoparticles dramatically. Previously, square wave potential has been used to control the shape of Pt crystal electrochemically, 75 so we tested the shape conversion of Pt hollow spheres using this kind of scan profile in the potential range between 0.6 and 1.0 V. Many of the Pt shells collapsed into solid nanoparticles after 3000 cycles (Figure 3.19c), indicating the square wave was not the optimal cycling profile for the conversion

174 Chapter to nanoboxes. Finally, the potential scanning rate was examined systematically to determine if the reaction kinetics could play a role in the formation of nanoboxes. We observed that if the scan rate was changed from ~100 mv/s to 20 mv/s, the population of the boxes became smaller and size distribution became broader (Figure 3.19d). If the scan rate became too fast, well-defined cubic nanoboxes could not be generated either. Figure 3.21 (a) TEM, (b) HR-TEM, (c) STEM images superimposed with Pt-M line scan, and (d) EDX spectrum of Pt hollow nanostructures after the removal of Ag cores. The above observations indicate that careful control of the reaction kinetics is required for obtaining the well-defined cubic nanoboxes. As the shape of

175 Chapter original Ag nanoparticle templates was not cubic, extensive reconstruction of Pt shells had to occur during the electrochemical treatments. Figure 3.21 shows the TEM images and EDX analysis of Pt nanostructures produced right after 20 cycles of CV treatment in 0.1-M HClO 4 aqueous solution. The nanostructures had pseudo-spherical shapes resembling those of Ag nanoparticles (Figure 3.21a). Small nanoparticles could readily be found on the surface of these hollow nanoparticles (Figure 3.21b). The overall sizes of these hollow structures were comparable with those of as-made Pt-on-Ag nanoparticles, and no obvious shrinkage was observed at this stage. There was no evidence for the large degree of reconstruction and the formation of nanobox. The elemental analysis of representative hollow particle was carried out using HAADF-STEM technique (Figure 3.21c). The EDX scan of Pt-M line shows a saddle-like shape across the particle indicating the structure was hollow. No silver was detected by EDX analysis, suggesting the complete removal of Ag metal (Figure 3.21d). Mixed cubic nanoboxes and hollow spheres were observed after the hollow nanoparticles were treated electrochemically with linear potential scan for 1000 cycles in a 0.5-M H 2 SO 4 solution (Figure 3.22). The resulting hollow particles shrank in size after the treatment. This observation suggests that the restructuring process was gradual and the initial hollows were not the most stable structures. However, once cubic nanoboxes were formed, they seemed to be stable and showed little difference in shape between those obtained after 3000 and 6000 cycles (Figure

176 Chapter b). The formation of cubic nanoboxes is most likely related to the relative stability among the low indexed surfaces of platinum metals. It has been reported that the adsorption of hydroxyl molecules varies according to the low-index surfaces of platinum. 76 A repetitive adsorbing-desorbing process associated with the cycling can lead to a reconstruction to the platinum structures 75, 77, 78 with most stable surfaces. The Pt (100) surface is more difficult to restructure than other low-index surfaces with this linear potential cycling. 79 Chemisorption of (bi)sulfate anions on the three low-index Pt surfaces was also quite different These two factors might contribute to the formation of Pt cubic nanoboxes. Figure 3.22 TEM images of Pt hollow nanostructures after linearly cycling the potential in the range of V for (a) 1000 and (b) 6000 times, respectively. As cubic crystals are bound by {100} surfaces, we use these nanoboxes to study their catalytic property in electrochemical oxidation of methanol, a reaction that 83, 84 depends highly on the surface electronic structure and atomic arrangement.

177 Chapter The electrochemical active surface areas (ECSAs) of both the Pt nanoboxes and hollow nanospheres as shown in Figure 3.21a were obtained by integrating the shaded areas of CV curves shown in the inset of Figure 3.23a. The shaded areas were between 0.05 and 0.4 V (vs. RHE) in the cyclic voltammetry (CV) curves and associated with hydrogen adsorption on Pt. The Pt hollow nanoshperes Figure 3.23 (a) Methanol oxidation reaction catalyzed by Pt cubic nanoboxes, hollow nanoshperes and the commercial Pt catalyst (TKK, 46.7 wt.% Pt), and (b) turnover frequencies (TOF) at the peak potentials. The CV curves of these two hollow catalysts in a sulfuric acid aqueous solution are shown in the inset of (a). had an ECSA of 82.4 m 2 /g Pt, close to that of the state-of-the-art commercial Pt/C catalysts that have much particle sizes in the range of 2-3 nm. The CV curve of Pt nanoboxes was different from that of Pt hollow spheres and could be associated with their morphology. The ECSA of Pt cubic nanoboxes was calculated to be 58.0 m 2 /g Pt, indicating some loss of surface area during the surface reconstruction. The electrocatalytic property of both Pt nanoboxes and hollow nanospheres toward the MOR was tested and compared with that of the commercial Pt catalyst.

178 Chapter Figure 3.23a shows the CV curves recorded at a scan rate of 50 mv/s after 50 cycles, which has two characteristic peaks for pure Pt in their forward and backward scans. There is little difference in shape and peak potential among these three catalysts, suggesting the reaction mechanism should be similar. The specific activity for Pt nanoboxes at its peak potential was 1.10 ma/cm 2 Pt, higher than 0.74 ma/cm 2 Pt for the hollow nanospheres and 0.63 ma/cm 2 Pt for TKK. This result indicates that nanoboxes can have a better specific activity than the other Pt nanostructures. We further analyzed the turnover frequency (TOF) of these catalysts at their peak potentials. The TOF is defined as the number of oxidized methanol molecules per Pt surface site per second and calculated based on the oxidation current in the CV, the ECSA value, and the number of electrons transferred in the reaction (Figure 3.23b). The Pt hollow nanospheres showed a small enhancement in TOF when they were compared to the TKK catalyst (0.59 vs s -1 ). Pt cubic nanoboxes, on the other hand, had a higher TOF of 0.89 s -1 at 0.85 V than other two catalysts, indicating a much improved catalytic activity. As Pt (100) surface is more active in general than (111) surface towards MOR, 83, the observed difference in TOF can be the outcome of such dependency on shape Conclusions

179 Chapter A new electrochemical method for making Pt cubic nanoboxes is developed. It is intriguing that the final cubic shape is the result of the requirement for most stable surfaces under the reaction conditions rather than the shapes of the original templates. Thus, both the size and shape uniformity of the obtained hollow nanostructures should be controllable. Since the generation of cubic nanoboxes is based on the difference in standard reduction potentials between two metals and the restructuring to most stable form, it should be generic for making metal cubes and other shapes from metal-on-metal or core-shell nanostructures. As both the size (6 nm in edge length) and wall thickness (1.5 nm) are among the smallest nanoboxes attainable so far and approaching to the ultimate limits of few unit cell length, these Pt hollow nanostructures should be of great interests in the development of low-cost catalysts and for those biological applications pertinent to metal nanoboxes. Author contributions: Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, and co-wrote the paper. Hongjun You conducted the synthesis and Zhenmeng Peng performed the electron microscopy characterization. All authors contributed to the discussion of the manuscript. 3.3 References (1) Peng, Z. M.; Yang, H. Nano Today 2009, 4, 143.

180 Chapter (2) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, 9. (3) Lovic, J. D.; Tripkovic, A. V.; Gojkovic, S. L. J.; Popovic, K. D.; Tripkovic, D. V.; Olszewski, P.; Kowal, A. J. Electroanal. Chem. 2005, 581, 294. (4) Chen, J. Y.; Lim, B.; Lee, E. P.; Xia, Y. N. Nano Today 2009, 4, 81. (5) de Bruijn, F. A.; Dam, V. A. T.; Janssen, G. J. M. Fuel Cells 2008, 8, 3. (6) Ferreira, P. J.; la O, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.; Gasteiger, H. A. J. Electrochem. Soc. 2005, 152, A2256. (7) Zhong, C. J.; Luo, J.; Njoki, P. N.; Mott, D.; Wanjala, B.; Loukrakpam, R.; Lim, S.; Wang, L.; Fang, B.; Xu, Z. C. Energy Environ. Sci. 2008, 1, 454. (8) Zhang, J. L.; Mo, Y.; Vukmirovic, M. B.; Klie, R.; Sasaki, K.; Adzic, R. R. J. Phys. Chem. B 2004, 108, (9) Zhang, J. L.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, 220. (10) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (11) Koh, S.; Strasser, P. J. Am. Chem. Soc. 2007, 129, (12) Peng, Z. M.; Yang, H. J. Am. Chem. Soc. 2009, 131, (13) Lee, E. P.; Peng, Z. M.; Cate, D. M.; Yang, H.; Campbell, C. T.; Xia, Y. J. Am. Chem. Soc. 2007, 129, (14) Lee, E. P.; Peng, Z. M.; Chen, W.; Chen, S. W.; Yang, H.; Xia, Y. N. ACS Nano 2008, 2, (15) Lim, B.; Jiang, M. J.; Camargo, P. H. C.; Cho, E. C.; Tao, J.; Lu, X. M.; Xia, Y. N. Science 2009, 324, (16) Tao, F.; Grass, M. E.; Zhang, Y. W.; Butcher, D. R.; Renzas, J. R.; Liu, Z.; Chung, J. Y.; Mun, B. S.; Salmeron, M.; Somorjai, G. A. Science 2008, 322, 932. (17) Guo, Q. J.; Teng, X. W.; Yang, H. Adv. Mater. 2004, 16, (18) Teng, X. W.; Yang, H. J. Am. Chem. Soc. 2003, 125,

181 Chapter (19) Peng, Z. M.; Yang, H. Nano Res. 2009, 2, 406. (20) Zhou, S. G.; McIlwrath, K.; Jackson, G.; Eichhorn, B. J. Am. Chem. Soc. 2006, 128, (21) Beard, K. D.; Borrelli, D.; Cramer, A. M.; Blom, D.; Zee, J. W. V.; Monnier, J. R. ACS Nano 2009, DOI: /nn900214g. (22) Chen, Z. W.; Waje, M.; Li, W. Z.; Yan, Y. S. Angew. Chem.-Int. Edit. 2007, 46, (23) Chen, H. M.; Liu, R. S.; Lo, M. Y.; Chang, S. C.; Tsai, L. D.; Peng, Y. M.; Lee, J. F. J. Phys. Chem. C 2008, 112, (24) Liang, H. P.; Zhang, H. M.; Hu, J. S.; Guo, Y. G.; Wan, L. J.; Bai, C. L. Angew. Chem. -Int. Ed. 2004, 43, (25) Guo, S. J.; Dong, S. J.; Wang, E. Chem.-Eur. J. 2008, 14, (26) Teng, X. W.; Liang, X. Y.; Maksimuk, S.; Yang, H. Small 2006, 2, 249. (27) Liang, H. P.; Guo, Y. G.; Zhang, H. M.; Hu, J. S.; Wan, L. J.; Bai, C. L. Chem. Commun. 2004, (28) Chen, G.; Xia, D. G.; Nie, Z. R.; Wang, Z. Y.; Wang, L.; Zhang, L.; Zhang, J. J. Chem. Mat. 2007, 19, (29) Vasquez, Y.; Sra, A. K.; Schaak, R. E. J. Am. Chem. Soc. 2005, 127, (30) Teng, X. W.; Maksimuk, S.; Frommer, S.; Yang, H. Chem. Mat. 2007, 19, 36. (31) Ding, Y.; Chen, M. W.; Erlebacher, J. J. Am. Chem. Soc. 2004, 126, (32) Xu, C. X.; Wang, L. Q.; Wang, R. Y.; Wang, K.; Zhang, Y.; Tian, F.; Ding, Y. Adv. Mater. 2009, 21, (33) Mayers, B.; Jiang, X. C.; Sunderland, D.; Cattle, B.; Xia, Y. N. J. Am. Chem. Soc. 2003, 125, (34) Sun, Y. G.; Mayers, B.; Xia, Y. N. Adv. Mater. 2003, 15, 641. (35) Sun, Y. G.; Mayers, B. T.; Xia, Y. N. Nano Lett. 2002, 2, 481. (36) Teng, X. W.; Black, D.; Watkins, N. J.; Gao, Y. L.; Yang, H. Nano Lett.

182 Chapter , 3, 261. (37) Lee, H. J.; Habas, S. E.; Somorjai, G. A.; Yang, P. D. J. Am. Chem. Soc. 2008, 130, (38) Lu, X. M.; Tuan, H. Y.; Chen, J. Y.; Li, Z. Y.; Korgel, B. A.; Xia, Y. N. J. Am. Chem. Soc. 2007, 129, (39) Mazumder, V.; Sun, S. H. J. Am. Chem. Soc. 2009, 131, (40) Fields, E. K.; Meyerson, S. J. Org. Chem. 1976, 41, 916. (41) Lin, X. Z.; Teng, X. W.; Yang, H. Langmuir 2003, 19, (42) Capdevielle, P.; Lavigne, A.; Sparfel, D.; Barannelafont, J.; Cuong, N. K.; Maumy, M. Tetrahedron Lett. 1990, 31, (43) Zhong, X.; Feng, Y.; Lieberwirth, I.; Knoll, W. Chem. Mater. 2006, 18, (44) Song, H.; Kim, F.; Connor, S.; Somorjai, G. A.; Yang, P. D. J. Phys. Chem. B 2005, 109, 188. (45) Teng, X. W.; Yang, H. Nano Lett. 2005, 5, 885. (46) Bauer, E.; Vandermerwe, J. H. Phys. Rev. B 1986, 33, (47) Chambers, S. A. Surf. Sci. Rep. 2000, 39, 105. (48) Greeley, J.; Norskov, J. K.; Mavrikakis, M. Annu. Rev. Phys. Chem. 2002, 53, 319. (49) Hammer, B.; Norskov, J. K., Theoretical surface science and catalysis - Calculations and concepts. In Advances in Catalysis, Academic Press Inc.: San Diego, USA, 2000; Vol. 45, pp 71. (50) Peng, Z. M.; Yang, H. J. Solid State Chem. 2008, 181, (51) Lide, D. R., CRC Handbook of Chemistry and Physics, 88th Edition (Internet Version 2008). CRC Press/Taylor and Francis: Boca Raton, FL. (52) Bard, A. J.; Faulkner, L. R., Electrochemical Methods - Fundamentals and Application. 2nd ed.; John Wiley & Sons: New York, (53) Vielstich, W.; Lamm, A.; Gasteiger, H. A., Handbook of Fuel Cells,

183 Chapter Fundametnals Technology and Applications. John Wiley & Sons Ltd.: 2003; Vol. 2. (54) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Radmilovic, V.; Markovic, N. M.; Ross, P. N. J. Phys. Chem. B 2002, 106, (55) Sun, Y. G.; Mayers, B.; Xia, Y. N. Adv. Mater. 2003, 15, 641. (56) Zhang, Q.; Wang, W. S.; Goebl, J.; Yin, Y. D. Nano Today 2009, 4, 494. (57) Kim, S. W.; Kim, M.; Lee, W. Y.; Hyeon, T. J. Am. Chem. Soc. 2002, 124, (58) Chen, Z. W.; Waje, M.; Li, W. Z.; Yan, Y. S. Angew. Chem.-Int. Edit. 2007, 46, (59) Chen, H. M.; Liu, R. S.; Lo, M. Y.; Chang, S. C.; Tsai, L. D.; Peng, Y. M.; Lee, J. F. J. Phys. Chem. C 2008, 112, (60) Liang, H. P.; Zhang, H. M.; Hu, J. S.; Guo, Y. G.; Wan, L. J.; Bai, C. L. Angew. Chem. -Int. Ed. 2004, 43, (61) Guo, S. J.; Dong, S. J.; Wang, E. Chem.-Eur. J. 2008, 14, (62) Skrabalak, S. E.; Chen, J. Y.; Sun, Y. G.; Lu, X. M.; Au, L.; Cobley, C. M.; Xia, Y. N. Accounts Chem. Res. 2008, 41, (63) Song, K. H.; Kim, C. H.; Cobley, C. M.; Xia, Y. N.; Wang, L. V. Nano Lett. 2009, 9, 183. (64) Au, L.; Zheng, D. S.; Zhou, F.; Li, Z. Y.; Li, X. D.; Xia, Y. N. ACS Nano 2008, 2, (65) Yavuz, M. S.; Cheng, Y. Y.; Chen, J. Y.; Cobley, C. M.; Zhang, Q.; Rycenga, M.; Xie, J. W.; Kim, C.; Song, K. H.; Schwartz, A. G.; Wang, L. V.; Xia, Y. N. Nature Mater. 2009, 8, 935. (66) Peng, Z. M.; Wu, J. B.; Yang, H. Chem. Mater. 2010, 22, (67) Lu, X. M.; Tuan, H. Y.; Chen, J. Y.; Li, Z. Y.; Korgel, B. A.; Xia, Y. N. J. Am. Chem. Soc. 2007, 129, 1733.

184 Chapter (68) Mazumder, V.; Sun, S. H. J. Am. Chem. Soc. 2009, 131, (69) Wang, Y.; Wong, J. F.; Teng, X. W.; Lin, X. Z.; Yang, H. Nano Lett. 2003, 3, (70) Lin, X. Z.; Teng, X. W.; Yang, H. Langmuir 2003, 19, (71) Wang, C.; Hu, Y. J.; Lieber, C. M.; Sun, S. H. J. Am. Chem. Soc. 2008, 130, (72) Wang, C.; Yin, H. G.; Chan, R.; Peng, S.; Dai, S.; Sun, S. H. Chem. Mat. 2009, 21, 433. (73) de Bruijn, F. A.; Dam, V. A. T.; Janssen, G. J. M. Fuel Cells 2008, 8, 3. (74) Ferreira, P. J.; la O, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.; Gasteiger, H. A. J. Electrochem. Soc. 2005, 152, A2256. (75) Tian, N.; Zhou, Z. Y.; Sun, S. G.; Ding, Y.; Wang, Z. L. Science 2007, 316, 732. (76) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (77) Furuya, N.; Ichinose, M.; Shibata, M. J. Electroanal. Chem. 1999, 460, 251. (78) Furuya, N.; Shibata, M. J. Electroanal. Chem. 1999, 467, 85. (79) Clavilier, J.; Orts, J. M.; Feliu, J. M. J. Phys. IV 1994, 4, 303. (80) Markovic, N. M.; Gasteiger, H. A.; Ross, P. N. J. Phys. Chem. 1995, 99, (81) Mostany, J.; Herrero, E.; Feliu, J. M.; Lipkowski, J. J. Phys. Chem. B 2002, 106, (82) Hoshi, N.; Sakurada, A.; Nakamura, S.; Teruya, S.; Koga, O.; Hori, Y. J. Phys. Chem. B 2002, 106, (83) Hamnett, A. Catal. Today 1997, 38, 445. (84) Lee, S. W.; Chen, S. O.; Sheng, W. C.; Yabuuchi, N.; Kim, Y. T.; Mitani, T.; Vescovo, E.; Shao-Horn, Y. J. Am. Chem. Soc. 2009, 131,

185 Chapter (85) Kita, H.; Gao, Y. Z.; Nakato, T.; Hattori, H. J. Electroanal. Chem. 1994, 373, 177. (86) Sun, S. G.; Clavilier, J. J. Electroanal. Chem. 1987, 236, 95. (87) Herrero, E.; Franaszczuk, K.; Wieckowski, A. J. Phys. Chem. 1994, 98, 5074.

186 Chapter Chapter 4 Synthesis and Oxygen Reduction Electrocatalytic Property of Pt-on-Pd Bimetallic Heteronanostructures 4.1 Introduction One key area in advancing PEMFCs is to improve both the sluggish kinetics and long-term stability of ORR catalysts. 1-3 A major strategy for improving the activity is to use platinum alloys instead of pure platinum as the electrocatalysts, 4 though leaching of non-platinum metal over time is a major issue. The loss of surface area due to Ostwald ripening or grain growth is another major factor that often results in the degradation of catalytic performance. 5 One solution to improve the durability is to deposit Au nanoclusters on Pt catalysts. 6 In this chapter, I will describe a new approach to address both the activity and stability issues synergistically by using palladium metal nanoparticles as support for Pt catalysts based on a particle-on-particle structure. Such heterogeneous bimetallic nanocrystals are expected to integrate several different functionalities in one structure, which is difficult to accomplish in a single-component material. The rationales for choosing palladium as metallic support for platinum nanoparticles are based on the following two major factors: First, both metals have a face-centered cubic (fcc) phase with a unit length of 3.92 Å for Pt and 3.89 Å for Pd. The small lattice mismatch means that the epitaxial growth should be

187 Chapter favored. 7 Second, Pt monolayer on Pd surface shows higher catalytic activity than pure Pt in ORR. 8 Selective growth of metals on semiconductors and metal oxides and semiconductors on semiconductors has been investigated in recent years, though examples on well-defined metal-on-metal heteronanostructures are still quite limited. 7,9-11 Heterogeneous nucleation and growth however, should be favored thermodynamically over the homogenous nucleation for the corresponding individual elements, when global energy minimum is dominated by interfacial energy. 3,12 Additional reduction in free energy can further be achieved through epitaxial growth. 4.2 Experimental Section Synthesis. Pt-on-Pd bimetallic nanoparticles were prepared by a sequential synthetic method. All the experiments were conducted under an argon atmosphere using a standard Schlenk line. In a typical procedure, palladium nanoparticles were produced first. To make Pd nanoparticles, palladium acetylacetonate (Pd(acac) 2, 99%, Strem, 0.1 g or 0.33 mmol) was dissolved in trioctylphosphine (TOP, 90%, Aldrich, 0.5 ml or 1.1 mmol) to form Pd-TOP complex in a 25-mL three-neck flask. Oleylamine (OAm, 70%, Aldrich, 10 ml) was then added into this flask. The resulting solution was slowly heated to 250 C at a rate of 2 C/min and maintained at this temperature for 30 min before cooling down to

188 Chapter ambient room temperatures. The product was precipitated out by washing twice with a mixture of 6 ml of hexane and 18 ml of ethanol, followed by centrifugation at 6000 rpm for 10 min. The precipitate was dispersed in 4 ml of DPE (Aldrich, 99%) for the subsequent use in the preparation of Pt-on-Pd nanoparticles, which began by mixing Pt(acac) 2 (Strem, 98%, 0.02 g or 50 µmol) and Pd nanoparticles (12.5 µmol in metal) in 5 ml of DPE and 0.3 ml of OAm. The concentration of Pd nanoparticles in DPE was determined by TGA. This mixture was heated to 180 C and kept at this temperature for 1 h. The product was precipitated out by washing with 10 ml of ethanol, followed by centrifugation at 6000 rpm for 10 min. The precipitate was dispersed in 3 ml of hexane. Preparation of carbon-supported Pt-on-Pd nanostructure catalysts. A suspension of 40 mg of carbon black (Vulcan XC72R) in 2 ml of hexane was sonicated for 1 h, followed by the addition of Pt-on-Pd nanoparticles that contained 10 mg of metals. The amount of Pt-on-Pd nanoparticles in hexane was determined by TGA analysis. This mixture was stirred overnight and the resulting solid was precipitated out via centrifugation. The carbon-supported Pt-on-Pd nanoparticles were thermally treated at 300 C for 1 h in air to remove the surface capping agents and at 80 C for 2 h in a forming gas of H 2 /Ar to ensure the surface composing of zero valence metals. Characterization. TEM images were taken on a JEOL JEM 2000EX

189 Chapter microscope at an accelerating voltage of 200 kv. HR-TEM images were taken on a FEI TECNAI F-20 microscope operated at 200 kv. STEM and elemental maps were carried out under the HAADF mode. EDX analysis of particle ensembles was carried out on a field emission SEM (Zeiss-Leo DSM982). PXRD patterns were recorded using a Philips MPD diffractometer with a Cu K α X-ray source (λ= Å). The TGA was conducted on an SDT-Q600 system from TA Instruments, Inc. Electrochemical properties of the catalysts were measured on a CHI 760 dual channel electrochemical workstation from CH Instruments, Inc. using a three-electrode system that consists of a glassy carbon working electrode (5 mm in diameter), a platinum leaf electrode, and a hydrogen reference electrode (HydroFlex, Gaskatel). The hydrogen reference electrode was placed in a separate cell containing 0.05-M H 2 SO 4 aqueous solution and connected with the main cell using a salt bridge. The HydroFlex electrode was calibrated by performing hydrogen evolution reaction (HER) with two Pt electrodes. All the potentials were recorded with respect to a reversible hydrogen electrode (RHE). The catalyst dispersions were prepared by mixing 5 mg of designed catalyst in 5 ml of aqueous solution containing 1 ml of iso-propanol and 25 µl of a 5 wt.% Nafion solution (V water : V 2-propanol : V 5% Nafion = 0.8: 0.2: 0.005), followed by ultrasonication for 10 min. A designed amount of dispersion was drop-cast onto the glassy carbon electrode (5 mm in diameter) and dried for 30 min before the

190 Chapter measurements. For ECSA study, the cyclic voltammetry (CV) was conducted at catalyst loading of 2 µg of metals in an Ar-protected 0.1 M perchloric acid (HClO 4 ) aqueous solution. The scan rate was 20 mv/s. The ECSA values were calculated by integrating the area under curve in the hydrogen adsorption range between 0.05 and 0.4 V for the backward sweep in the CV. The adsorption of hydroxyl species was calculated based on the OH ad peak in the CV curves at the potential larger than 0.6 V. Dividing the hydroxyl adsorption area by the overall active surface area resulted in the surface coverage of OH ad species (Θ OHad ). All ORR tests were conducted at ambient room temperature in O 2 -saturated 0.1-M HClO 4 aqueous solutions. The loading was based on the ECSA of the catalysts, which were set at 2 cm 2 for all the ORR studies. The polarization curves were obtained by sweeping the potential from 0 to 1 V at the scan rate of 10 mv/s and rotation disk speed of 1600 rpm. Before recording the data, the current densities for the prepared catalysts were checked to make sure that they reached the diffusion limited of about 6 ma/cm 2 and mass-transport was not an issue in these measurements. The Koutecky-Levich equation was applied to calculate kinetic current density based on ORR polarization curves, which can be described as follow: = + = + (4.1) 1/ 2 i i i i Bω k d k where, i is the measured current density, i k is the kinetic current density, i d is the diffusion (mass-transfer) limited current density, B is a constant and a function of

191 Chapter concentration (C O2 ), diffusion coefficient (D O2 ) of O 2 in the electrolyte and viscosity of the electrolyte (ν), and ω is the rotation rate of the electrode in unit of rpm. The value obtained for kinetic current was independent of diffusion and could be used to evaluate the intrinsic activity of the catalysts. The area-specified current density (i k ) was obtained by normalizing the current based on the active surface areas of the catalysts. The accelerated stability tests were conducted at ambient room temperature in Ar-protected 0.1-M HClO 4 aqueous solutions. The loading was based on the ECSA of the catalysts and set at 2 cm 2. The corresponding mass loadings of the catalysts were 5.4 µg of metal for Pt-on-Pd nanostructures and 2.4 µg of metal for Pt nanoparticles (E-TEK, 20 wt.% Pt, diameter: 2.5 nm). The potentials were cycled continuously for 30,000 times between 0.6 and 1.0 V (vs. RHE) at a scan rate of 100 mv/s. The ECSA after stability test was calculated based on the CV curve by integrating the area for hydrogen adsorption. 4.3 Results and Discussion The Pt-on-Pd nanoparticles were prepared using a sequential synthetic method. The palladium nanoparticles were made from palladium acetylacetonate in oleylamine based on a modified procedure. 13,14 The as-made palladium nanoparticles were fairly monodisperse and had an average diameter of 5.3±0.6

192 Chapter nm (Figure 4.1). PXRD pattern shows the particles were made of Pd metal (Figure 4.2). Figure 4.3a shows the representative TEM image for the as-synthesized Figure 4.1 Representative TEM image of as-synthesized Pd nanoparticles. Figure 4.2 PXRD patterns of as-prepared Pd and Pt-on-Pd nanoparticles. The intensity and position for Pt (blue) and Pd (purple) references were taken from the JCPDS database. Pt-on-Pd nanoparticles. The Pt nanoparticles had an average diameter of about 3 nm and distributed evenly on the surface of palladium nanoparticles.

193 Chapter High-resolution TEM image shows that these nanoparticles have good crystallinity and with well-defined fringes (Figure 4.3b). The Pt nanoparticles grew along the (111) crystal planes on Pd supports. No obvious grain boundaries or defects could be observed, as only a very small lattice mismatch of 0.77% existed between Pt and Pd metals. 7 Similar architectures have been reported in other bimetallic systems, such as Au-Pt. 11 The elemental distributions of these two metals were studied by a HAADF-STEM. Figure 4.3c-e shows representative STEM image and the corresponding EDX maps for Pd and Pt of a Pt-on-Pd nanoparticle. While Pd could only be detected in the core region, Pt was found throughout the entire particle including the branch regions, indicating the formation of Pt-on-Pd nanoparticles. PXRD pattern shows multiple diffraction peaks could be observed and indexed to an fcc lattice (Figure 4.2). The atomic ratio between Pd and Pt in the bimetallic nanostructures was about 1/3 based on the EDX analysis using a FE-SEM, close to the Pd metal/pt(acac) 2 molar ratio in the reaction mixture (Figure 4.4). The as-made nanoparticles could be loaded onto carbon support (Vulcan XC-72R) and thermally treated to make Pt-on-Pd bimetallic catalysts. Figure 4.3f shows representative TEM images of carbon-supported Pt-on-Pd nanoparticles at 20 wt.% loading of metals. The Pt-on-Pd nanoparticles were still evenly distributed on carbon supports without obvious sintering or growth of particles after the thermal treatment. HR-TEM study and EDX maps show the particle-on-particle morphology and elemental

194 Chapter distributions remained largely intact (Figure 4.3g and Figure 4.5a-c). The PXRD diffraction peaks became slightly sharper than before (Figure 4.5d), indicating an improvement in crystallinity. Figure 4.3 Representative (a) TEM, (b) HR-TEM, (c) HAADF-STEM images and (d, e) elemental maps for Pd and Pt metals of Pt-on-Pd bimetallic nanoparticles; and (f) TEM and (g) HR-TEM images of carbon-supported Pt-on-Pd bimetallic catalysts after the thermal treatments.

195 Chapter Figure 4.4 EDX spectrum of as-prepared Pt-on-Pd nanoparticles collected on a field emission scanning electron microscope (FE-SEM, Zeiss-Leo DSM982). Figure 4.5 Representative (a) HAADF-STEM image and (b, c) elemental maps of Pd and Pt metals, and (d) PXRD pattern of carbon-supported Pt-on-Pd bimetallic nanoparticles after thermal treatments. The intensity and position for Pt (blue) and Pd (purple) references were taken from the JCPDS database.

196 Chapter Figure 4.6 shows the electrochemical properties of carbon-supported Pt-on-Pd heteronanostructure and Pt reference catalysts (E-TEK, 20 wt.% Pt, diameter: 2.5 nm) (See Supporting Information for details). The electrochemical surface area (ECSA) for the Pt-on-Pd catalyst was found to be 37.3 m 2 /g metal (or 44.5 m 2 /g Pt) based on the CV data (Figure 4.6a). Well-defined chemical adsorption peaks of hydrogen on different Pt low-index surfaces became less definable for the Pt-on-Pd catalyst. The incorporation of Pd also greatly altered the ability to absorb hydroxyl species (OH ad, E >0.6 V) (Figure 4.6b). Both the onset and peak potentials for the Pt-on-Pd catalyst had positive shifts in comparison with pure Pt on the backward sweep, suggesting the fast hydroxyl desorption from the Pt-on-Pd surfaces. The ORR tests were conducted in O 2 -saturated HClO 4 aqueous solutions at ECSA of 2.0 cm 2 for all catalysts on a glass carbon electrode (GCE). Figure 4.6c shows the polarization curves for both carbon-supported Pt-on-Pd and pure Pt catalysts. The Pt-on-Pd catalysts exhibited more positive on-set potential and higher activity than the pure Pt nanoparticles. The area-specific current density (i k ), which represents the intrinsic activity of the catalysts and calculated using Koutecky-Levich equation, 15 was 307 µa/cm 2 Pt at 0.9 V for the Pt-on-Pd nanostructures and nearly doubled that for the Pt catalyst (166 µa/cm 2 metal) (Figure 4.6d). These results agreed well with those reported for Pt skin layers on Pd surfaces. 8 As the adsorbed OH ad species has a negative impact on the ORR and

197 Chapter low OH ad coverage on the surface of Pt-on-Pd catalysts helps improve the kinetics and thus enhance the activities. 3,8 Figure 4.6 (a) CV, (b) hydroxyl surface coverage (Θ OH ), (c) ORR polarization curves and (d) specific kinetic current densities (i k ) for carbon-supported Pt-on-Pd and Pt catalysts. The long-term stability of Pt-on-Pd catalyst was evaluated by applying linear potential sweeps between 0.6 and 1.0 V based on an established procedure. 5a,6 After 30,000 cycles, Pt-on-Pd catalysts loss about 12% of the initial ECSA (Figure 4.7a) and showed a small degradation of 9 mv in half-wave potential (Figure 4.7b). The particle-on-particle morphology, size and even composition remained after the accelerated tests (Figure 4.8). In sharp contrast, the degradation of Pt catalyst was quite serious, with a loss of 39% of the initial ECSA and a large

198 Chapter decrease of 35 mv in half-wave potential after the test (Figure 4.7c-d). The Pt nanoparticles also experienced a dramatic growth in diameter after the cycles, changing from 2.5 to 3.9 nm (Figure 4.9). The much improved stability thus could be due to the favored interfacial structures between Pt and Pd supports, as well as the larger than usual overall particle size of Pt-on-Pd nanostructures, which prevented the small Pt from dissolution in the ORR. This platinum-on-metal architecture provides a new design strategy for making hydrogen fuel cell cathode catalysts with both excellent activity and stability. Figure 4.7 CV and ORR polarization curves for carbon-supported (a, b) Pt-on-Pd and (c, d) Pt catalysts before and after 30,000 cycles.

199 Chapter Figure 4.8 Representative (a) TEM, (b) HR-TEM image, and (c) EDX spectrum of carbon-supported Pt-on-Pd bimetallic nanoparticles after accelerated stability test (30,000 cycles). Figure 4.9 TEM images and particle size distribution analyses of Pt catalysts (E-TEK, 20wt.% Pt): (a, b) as-received and (c, d) after 30,000 CV cycles of accelerated stability test.

200 Chapter Conclusions In summary, heterogeneous Pt-on-Pd bimetallic nanoparticles were prepared. The growth of platinum nanoparticles on the palladium surface is driven by the gaining of free energy at the initial nucleation stage and the loss of strain energy at the growth stage. The catalyst exhibits a 2-fold activity increase for ORR in comparison with a reference platinum catalyst. The carbon-supported Pt-on-Pd bimetallic nanoparticles also show a much better stability than the reference platinum catalyst, indicating palladium nanoparticles can be used to prevent platinum catalysts from dissolution in the ORR. This synthetic approach provides a new design strategy for making advanced fuel cell cathode catalysts with both great activity and excellent stability. Author Contributions: This section is adapted from a paper published in Journal of the American Chemical Society (Z. M. Peng, H. Yang, Synthesis and Oxygen Reduction Electrocatalytic Property of Pt-on-Pd Bimetallic Heteronanostructures, 2009, 131, ). Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, analysed the data and co-wrote the paper. Zhenmeng Peng conducted the synthesis and performed the electrochemical study. 4.5 References

201 Chapter (1) (a) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, (b) de Bruijn, F. A.; Dam, V. A. T.; Janssen, G. J. M. Fuel Cells 2008, 8, (2) Chen, G. Y.; Delafuente, D. A.; Sarangapani, S.; Mallouk, T. E. Catal. Today 2001, 67, (3) Peng, Z. M.; Yang, H. Nano Today 2009, 4, (4) (a) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, (b) Koh, S.; Strasser, P. J. Am. Chem. Soc. 2007, 129, (c) Chen, S.; Ferreira, P. J.; Sheng, W. C.; Yabuuchi, N.; Allard, L. F.; Shao-Horn, Y. J. Am. Chem. Soc. 2008, 130, (d) Stamenkovic, V.; Schmidt, T. J.; Ross, P. N.; Markovic, N. M. J. Phys. Chem. B 2002, 106, (5) (a) Ferreira, P. J.; la O, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.; Gasteiger, H. A. J. Electrochem. Soc. 2005, 152, A2256-A2271. (b) Shao, Y. Y.; Yin, G. P.; Gao, Y. Z. J. Power Sources 2007, 171, (c) Yu, X. W.; Ye, S. Y. J. Power Sources 2007, 172, (6) Zhang, J. L.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, (7) (a) Habas, S. E.; Lee, H.; Radmilovic, V.; Somorjai, G. A.; Yang, P. Nat. Mater. 2007, 6, (b) Lee, H. J.; Habas, S. E.; Somorjai, G. A.; Yang, P. D. J. Am. Chem. Soc. 2008, 130, (8) (a) Zhang, J. L.; Mo, Y.; Vukmirovic, M. B.; Klie, R.; Sasaki, K.; Adzic, R. R. J. Phys. Chem. B 2004, 108, (b) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Angew. Chem.-Int. Edit. 2005, 44, (9) (a) Mokari, T.; Rothenberg, E.; Popov, I.; Costi, R.; Banin, U. Science 2004, 304, (b) Mokari, T.; Sztrum, C. G.; Salant, A.; Rabani,

202 Chapter E.; Banin, U. Nat. Mater. 2005, 4, (c) Yu, H.; Chen, M.; Rice, P. M.; Wang, S. X.; White, R. L.; Sun, S. H. Nano Lett. 2005, 5, (d) Cozzoli, P. D.; Manna, L. Nat. Mater. 2005, 4, (e) Gu, H. W.; Zheng, R. K.; Zhang, X. X.; Xu, B. J. Am. Chem. Soc. 2004, 126, (f) Yang, J.; Elim, H. I.; Zhang, Q. B.; Lee, J. Y.; Ji, W. J. Am. Chem. Soc. 2006, 128, (10) Milliron, D. J.; Hughes, S. M.; Cui, Y.; Manna, L.; Li, J. B.; Wang, L. W.; Alivisatos, A. P. Nature 2004, 430, (11) (a) Peng, Z. M.; Yang, H. Nano Res. 2009, 2, 406. (b) Zhou, S. G.; McIlwrath, K.; Jackson, G.; Eichhorn, B. J. Am. Chem. Soc. 2006, 128, (12) (a) Roder, H.; Schuster, R.; Brune, H.; Kern, K. Phys. Rev. Lett. 1993, 71, (b) Wilcoxon, J. P.; Provencio, P. P. J. Am. Chem. Soc. 2004, 126, (c) Chambers, S. A. Surf. Sci. Rep. 2000, 39, (13) Kim, S. W.; Park, J.; Jang, Y.; Chung, Y.; Hwang, S.; Hyeon, T.; Kim, Y. W. Nano Lett. 2003, 3, (14) Liu, Q. S.; Bauer, J. C.; Schaak, R. E.; Lunsford, J. H. Angew. Chem.-Int. Edit. 2008, 47, (15) (a) Bard, A. J.; Faulkner, L. R. Electrochemical Methods - Fundamentals and Application; 2nd ed.; John Wiley & Sons: New York, (b) Vielstich, W.; Lamm, A.; Gasteiger, H. A. Handbook of Fuel Cells, Fundametnals Technology and Applications; John Wiley & Sons Ltd., 2003; Vol. 2.

203 Chapter Chapter 5 Post-Synthesis Modification of Pt-on-Au Heteronanostructures and Their Electrocatalytic Property 5.1 Introduction Metal alloys, those containing platinum in particular, have attracted lots of attention in recent years because of the increasing demand for high performance catalysts in fuel cells and other applications. 1-9 It has been shown that Pt 3 Ni (111) surface is much more active than the corresponding Pt (111) surface in oxygen reduction reaction, and over 90-fold more active than the state-of-the-art Pt/C catalysts. 4 The enhancement is associated with the surface electronic and 3, 4, 10 geometric structures of platinum atoms. Thus in order to make highly active catalysts, the surface compositions of nanoalloys need to be engineered 1, 2, 11, 12 synergistically. Post-synthesis modification of alloys emerges as a viable approach to address the challenges of controlling the location of different metal elements in the nanostructures. Under designed reactive environments, surface structures and compositions of Rh 50 Pd 50 can be tuned by using NO, CO and other gases. 11 This study demonstrates that high level of controls can be achieved for metal alloys. In this chapter I will present the design and synthesis of Pt-surface rich PtAu bimetallic nanostructures from Pt-on-Au nanoparticles. It has demonstrated the

204 Chapter sequential synthesis of core-shell nanoparticles from molecular precursors are readily achievable, 13, 14 as heterogeneous nucleation and growth are often favored thermodynamically. 15 The advantage of sequential method is its ability to deposit one component at a time. The as-made core-shell nanoparticles can be converted to new compositions on solid supports under relatively mild temperatures. 14 Heterogeneity in composition should be attainable by carefully controlling the reaction conditions, because solid state diffusion is typically slow near room temperature. Such multifunctional heterogeneous alloy nanostructures may potentially be used as highly active and low platinum content electrocatalysts, 1, 16, 17 a critical area for the development of PEMFCs which is a major class of low-temperature energy conversion devices using small molecules 18, 19 as fuels. Effects of the formation of alloys on electrocatalytic properties of platinum are studied using direct oxidation of formic acid as a model reaction. This fuel has several advantages over other commonly used small molecules for PEMFCs A direct formic acid fuel cell (DFAFC) has a fairly high theoretical open circuit voltage (1.45 V) and experiences fewer problems in fuel crossover than direct a 24, 25 methanol fuel cell (DMFC). The kinetic of chemical oxidation of formic acid is sensitive to the surface structure and composition changes of the electrocatalysts. It is known that this oxidation can undergo both

205 Chapter dehydrogenation and dehydration, which are described by the following two 18, 21, 26 reactions: dehydrogenation HCOOH CO H + 2e (5.1) dehydration HCOOH CO ads + H + 2 O CO2 + 2H + 2e (5.2) In the dehydrogenation pathway, formic acid is completely oxidized to CO 2 in a single step on a catalyst, typically made of carbon-supported platinum nanoparticles. In contrast, formic acid first generates CO ads on the catalytic surfaces in the dehydration pathway. These intermediates are then oxidized electrochemically to produce CO 2. In the latter reaction pathway, the absorbed CO intermediates can block the active sites and slow down the overall reaction. Dehydrogenation is the main pathway in the low potential range between 0.2 and 18, 20, V, although adsorbed CO ads can still be generated via dehydration. The coverage of Pt active sites by CO ads species increases with the potential, resulting in the loss of available active sites. The CO poisoning can by greatly suppressed by using platinum bimetallic electrocatalysts. It has been shown that both PtPb and PtBi intermetallics can have higher activities than pure platinum metal in catalyzing this oxidation reaction The improved adsorption of formic acid is attributed to the electronic effect, while the reduced poisoning by CO ads is because of the geometric effect High performance toward the 25, 34 oxidation of formic acid has been observed for both PtAu alloy nanoparticles 5, 35, 36 and Pt-modified Au nanoparticles.

206 Chapter Experimental Section Synthesis of gold nanoparticles. Oleylamine-stabilized gold nanoparticles 37, 38 were prepared using a modified Brust-Schiffrin method. A solution of tetraoctylammonium bromide (TOAB, 98%, Aldrich, 0.73 g or 1.3 mmol) in toluene (99.8%, Aldrich, 26.7 ml) was mixed with 10-mL aqueous solution of gold (III) chloride trihydrate (HAuCl 4 3H 2 O, >99.9%, Aldrich, 0.06 g or 0.15 mmol) in a 100-mL flask. This two-phase mixture was vigorously stirred using a mechanic stirrer (VWR) until all tetrachloroaurate ions were transferred into the toluene phase. An 8.3-mL freshly-made aqueous solution of sodium borohydride (99%, Aldrich, 0.13g or 3.3 mmol) was slowly added into the toluene phase under vigorous stir for 10 min. After reaction for additional 1 h, oleylamine (70%, Aldrich, 0.88 ml) was added to the toluene phase. The particles dispersed in toluene were separated out from the reaction mixtures and mixed with 50 ml of ethanol. The products of nanoparticles were precipitated out by centrifugation at 6000 rpm for 5 min (Beckman Coulter, Allegra TM 21 Centrifuge). This precipitate was washed with 7.5 ml of chloroform and 22.5 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. The as-made gold nanoparticles were dispersed in 5 ml of diphenyl ether (DPE, 99%, Aldrich) for further use. Synthesis of Pt-on-Au nanoparticles. In a typical procedure, oleylamine (0.3 ml or 0.9 mmol) was added to a solution of platinum acetylacetonate (Pt(acac) 2,

207 Chapter %, Strem, 0.02 g or 50 µmol) in DPE (4 ml), followed by the addition of a dispersion of gold nanoparticles (12.5 µmol) in DPE (1 ml). The molar number of gold was calculated based on the amount of metal determined by the TGA technique using air as carrier gas. The total volume of DPE was fixed at 5 ml for all experiments. The reaction was conducted at 180 C under argon for 1 h, unless stated otherwise. The product was precipitated out from the mixture by adding 10 ml of ethanol, which was used as the antisolvent, followed by centrifugation at 6000 rpm for 10 min. The precipitate was collected and dispersed in 2 ml of hexane. Feeding ratio between these two metal precursors, reaction temperature and time were varied systematically to study the effect of these conditions on the formation of Pt-on-Au nanostructures. To study the effect of feeding ratio, we adjusted the amount of gold nanoparticles while all other variables kept the same. The Pt/Au atomic ratio in the reaction mixture was changed from one to six by adding different amounts of Pt(acac) 2 and Au nanoparticles. Preparation of PtAu bimetallic heteronanostructures. A suspension of 50 mg of carbon black (Vulcan XC-72) in 15 ml of hexane was sonicated at ambient room temperature for 1 h using a sonicator (Brandson, Model 2510), followed by the addition of 10 mg of Pt-on-Au nanoparticles. The total amount of metals of nanoparticles added was determined by TGA using air as carrier gas to a maximum temperature set at 800 C. This mixture was stirred at room

208 Chapter temperature overnight using a magnetic stirrer and precipitated out via centrifugation at 5000 rpm for 5 min. The resultant powder was thermally treated at 300 C for 1 h in air at a ramping rate of 2 C/min and then at 400 C for 2 h in a forming gas of 5 vol.% hydrogen in argon in a tube furnace (Lindberg/Blue, Mini-Lite tube furnace). The ramping rate was 1 C/min for the segment between 300 and 400 C. Characterization. A detailed description regarding characterization of the samples has been presented in Chapter 2.3. Briefly, TEM and HR-TEM images were taken on a FEI TECNAI F-20 microscope. STEM and elemental maps were carried out under a HAADF mode on the same microscope. EDX spectra were taken on a FE-SEM Zeiss-Leo DSM982. PXRD patterns were recorded using a Philips MPD diffractometer. UV-vis spectra were collected with a UV/VIS/NIR spectrometer. The loading amount of metals on carbon was determined using an SDT-Q600 system. Electrochemical property was studied on a CHI 760 dual channel electrochemical workstation (CH Instruments, Inc.) using a three-electrode system that consisted of a glassy carbon working electrode (5 mm in diameter), a platinum wire counter electrode, and a hydrogen reference electrode (HydroFlex, Gaskatel). The ink of carbon-supported catalysts was prepared by dispersing in a mixture of deionized water, iso-propanol and 5 wt% Nafion solution (V water :V 2-propanol :V 5% Nafion = 0.8:0.2:0.005), followed by sonication at an ambient

209 Chapter room temperature for 10 min. The catalyst ink which contained 1 µg of PtAu metals was drop-cast onto the glassy carbon electrode and dried under a gentle stream of air. The amount of metals in a catalyst was quantified by TGA as well. For comparison, carbon-supported platinum (60 wt% Pt, E-TEK) and gold (20 wt%, E-TEK) catalysts were prepared according to the same protocol. The total amount of the catalyst was kept at 1 µg of metals for all the tests. For ECSA measurement, a 0.5-M H 2 SO 4 was used as the supporting electrolyte. For electrocatalytic oxidation of formic acid, the electrolyte was 0.5-M H 2 SO 4 and 0.5-M formic acid in an aqueous solution. Before each measurement, the solution was bubbled with argon for 30 min to remove dissolved oxygen gas. For CO stripping experiments, the electrodes were pretreated by immersing in a CO-saturated 0.1-M HClO 4 aqueous solution and held at 0.25 V for 20 min before the measurement. The CVs were typically run at the ambient room temperatures using a scan rate of 50 mv/s, unless stated otherwise. 5.3 Results and Discussion Figure 5.1 shows TEM images, elemental maps, and PXRD patterns of Pt-on-Au nanoparticles, and a representative TEM image of gold nanoparticles used in the synthesis. The gold nanoparticles, which were made using the Brust-Schiffrin method, 37, 38 had an average diameter of 7.4 ± 1.0 nm (Figure 5.1a). The PXRD pattern of these nanoparticles shows five diffraction peaks that can be

210 Chapter indexed to (111), (200), (220), (311) and (222) planes of face-centered cubic (fcc) gold metal, respectively. Upon the addition of Pt(acac) 2, secondary particles formed on the surface of Au cores could be clearly observed under TEM after reaction for 1 h at 180 C (Figure 5.1b). The formation of platinum on gold nanoparticles was most likely followed either island or island-on-wetting layer growth. The EDX analysis indicates these nanoparticles contained both Pt and Au elements and had an average Pt/Au atomic ratio of 0.63 (Figure 5.2). The distributions of these two metal elements in nanoparticle were investigated by EDX using STEM mode. Figure 5.1c-e shows the STEM image of an individual Pt-on-Au nanoparticle and its corresponding elemental maps for Pt and Au, respectively. Both Au and Pt metals were readily detectable, though their distributions within the particle were quite different. While Au could only be found in the core region, with a size much smaller than that of the nanoparticle (Figure 5.1c), platinum distributed throughout the entire particle including the branches that were outside the core. The HR-TEM study shows that these nanoparticles were crystalline, judging by the well-defined fringes (Figure 5.1f). The labeled lattices could be indexed to (200) crystal planes of fcc platinum metal. These observations suggest that nanoparticles were made of gold cores and platinum branches. No obvious grain boundaries between the branches and cores could be observed, and the heterogeneous nucleation and growth most likely followed either SK or VW mode. 39 The growth of heterogeneous nanostructures

211 Chapter should be favored thermodynamically, 40 as the lattice mismatch between gold and platinum metals is only about 4% and the strain energy for heterogeneous nucleation is small in such case because of a gradual relaxation of platinum lattices at the interfaces between Au and Pt. 41 Our PXRD pattern shows that the diffraction patterns for these nanoparticles could be deconvoluted very well into two sets of diffractions for pure gold and platinum, respectively (Figure 5.1g). Figure 5.1 Representative TEM images of (a) Au and (b) Pt-on-Au nanoparticles (NPs), (c) STEM image and (d, e) the elemental maps for Au and Pt metals, and (f) HR-TEM image of a single Pt-on-Au nanoparticle; and (g) the corresponding XRD patterns of Au and Pt-on-Au nanoparticles. All Pt-on-Au nanoparticles were made at 180 C for 1 h.

212 Chapter Figure 5.2 EDX spectrum of as-prepared Pt-on-Au nanoparticles from Au nanoparticles and Pt(acac) 2. The Au nanoparticle/pt(acac) 2 ratio was 1:4 based on the mole numbers of the metals. Figure 5.3 UV-vis spectra and TEM images (insets) of Au (0 min) and Pt-on-Au nanoparticles made after predetermined periods of reaction times. The Au/Pt precursor ratio was 1:4 and the reaction temperature was 180 C.

213 Chapter The growth of platinum on gold nanoparticles was followed by UV-vis spectroscopy and TEM studies (Figure 5.3). The plasmon resonance band from gold nanoparticles in the visible range is sensitive to the change of surface and has 37, 42 been used to study the coating by inorganic materials. Figure 5.3 shows the representative UV-vis spectra and their corresponding TEM images of Au and Pt-on-Au nanoparticles obtained at a series of reaction points. The gold particles 37, 43, 44 show the characteristic plasmon resonance band at around 510 nm. This absorbance became smaller for the formed Pt-on-Au nanoparticles over times, and disappeared almost completely for the particles made after the reaction for 60 min. No absorbance peak was observed for those Pt-on-Au nanoparticles made after 120 min. The UV-vis spectroscopic data agreed well with the TEM observation (Figure 5.3). The evolution of the branches on the cores could clearly be observed. Fairly large and multilayered particle-on-particle morphology was observed for those made at 120 min. Besides reaction time, the overall morphology and size of platinum on gold nanoparticles could also be controlled by reaction temperature (Figure 5.4). At the Au nanoparticle/pt(acac) 2 feeding ratio of 1:4 and with a reaction time of 1 h, increase in reaction temperature resulted in growth of large branches. The surface coverage of nanoparticles also increased and the optimal temperature range was found to be between 180 and 190 C (Figure 5.4). For those Pt-on-Au nanoparticles made at 170 C, the platinum particles on Au surfaces were small,

214 Chapter while those made at 200 C tended to overly grow. Similarly, Au nanoparticle/pt(acac) 2 feeding ratio could also affect the morphology of the branches and surface coverage of Pt-on-Au nanoparticles (Figure 5.5). Figure 5.4 TEM images of Pt-on-Au nanoparticles formed at four different reaction temperatures: (a) 170, (b) 180, (c) 190, and (d) 200 C. The Au nanoparticle/pt(acac) 2 feeding ratio was 1:4 and the reaction time was 1 h. Figure 5.5 Representative TEM images of Pt-on-Au nanoparticles made from precursors with Au nanoparticle/pt(acac) 2 molar ratios of (a) 1:1, (b) 1:2, (c) 1:3 and (d) 1:6, respectively. The reactions were carried out at 180 C for 1 h.

215 Chapter These Pt-on-Au nanoparticles provide an excellent platform to study the controlled synthesis of heterogeneous PtAu alloy nanostructures, since thermal properties of metal nanoparticles are scaled with size. In this study, carbon was used as the support to minimize the sintering among alloy nanoparticles. The carbon-supported Pt-on-Au nanoparticles were thermally treated first in air at 300 C and then in forming gas at 400 C. The treatment at 300 C in air was to facilitate the removal of surface capping agents through calcination. Figure 5.6 shows the representative (S)TEM images, elemental maps, and PXRD pattern of carbon-supported Pt-on-Au nanoparticles and the corresponding PtAu alloy nanostructures obtained after the thermal treatment. The TEM image shows that Figure 5.6 Representative TEM images of Pt-on-Au nanoparticles on carbon supports (a) before and (b) after the thermal treatment; (c) HR-TEM image, (d) STEM image, and (e, f) the corresponding the elemental maps for Au and Pt

216 Chapter metals, and (g) PXRD pattern of PtAu alloy heteronanostructures obtained through the thermal treatment. the Pt-on-Au nanoparticles dispersed well on carbon supports (Figure 5.6a). After the thermal treatment, the particles remained to be evenly distributed. While the overall size of the bimetallic particles did not seem to change dramatically, their surfaces became much smoother than those before the treatment (Figure 5.6b). HR-TEM study shows that the PtAu nanostructure was highly crystalline and possessed multiple domains (Figure 5.6c). STEM image and elemental maps for Au and Pt indicate that Au still distributed mainly in the core area, while Pt element could be detected throughout the entire PtAu nanoparticle with relatively strong signals at the outer regions (Figure 5.6d-f). Some Au signals could be detected in localized regions outside of the core, suggesting Au atoms diffused into the surface layers of Pt particles upon annealing at 400 C. These findings suggest that alloy between Pt and Au formed during the thermal treatment resulting in PtAu alloy nanostructures with surfaces rich in Pt. PXRD pattern provides additional evidence that heterogeneity in PtAu alloy phases existed in these annealed nanoparticles (Figure 5.6g). Noticeably, after the treatment, the diffraction pattern split into two distinctive sets, which were in between those for pure Au and Pt metals. The separation became obvious for those peaks at high 2θ angles. These two sets of diffraction peaks could be assigned to Au- and Pt-rich PtAu alloys, respectively

217 Chapter (Figure 5.6g). The compositions of these two alloys were estimated to be Pt 21 Au 79 and Pt 81 Au 19, respectively, based on the calculation on the shift of peak positions using Vegard s law. 45, 46 As the overall Pt/Au atomic ratio was 0.63, the bimetallic nanostructures should therefore consist of about 70 % of Pt 21 Au 79 and 30% of Pt 81 Au 19 alloys. While they do not form good solid solution and have a large miscibility gap in their bulk phase diagram, platinum and gold metals can 25, 34 form alloys at the nanometer scale at relatively low temperature. Similar phase behavior has also been observed in PtAg alloy nanostructures made from 46, 47 molecular precursors. Surface energy could be the main driving force for the formation of these nanoalloys. 48 PXRD pattern also shows crystallinity of these bimetallic nanostructures (PtAu/C) improved substantially as the diffraction peaks became sharper than those for Pt-on-Au nanoparticles. Figure 5.7 Cyclic voltammetry of formic acid oxidations by carbon-supported PtAu heteronanostructures, Pt and Au nanoparticle references. All experiments were conducted in a 0.5-M formic acid aqueous solution with 0.5-M H 2 SO 4 as the supporting electrolyte.

218 Chapter Figure 5.8 Representative TEM images of carbon-supported (a) Pt (60 wt%), and (b) Au (20 wt.%) catalysts from E-TEK. Figure 5.7 shows the CV curves for the oxidation of formic acid using PtAu/C as the electrocatalysts. The carbon-supported Pt (Pt/C, 60 wt% Pt E-TEK) and Au (Au/C, 20 wt.%, E-TEK) catalysts commercial were tested for comparison. The average size and size distribution of these two commercial catalysts determined by TEM were 3.5±0.7 nm in diameter for Pt/C and 7.6±3.1 nm for Au/C (Figure 5.8). The current density was normalized according to the unit mass of metals rather than ECSA, because the conventional method for determining 49, 50 ECSA is based on hydrogen adsorption/desorption. Since gold metal does not show measurable hydrogen adsorption, such electrochemical method cannot be used to measure the true surface areas of PtAu alloys (Figure 5.9). The mass current density is also the preferred measure for practical applications.

219 Chapter Figure 5.9 Cyclic voltammetry curves of carbon-supported Pt (E-TEK, 60 wt%), Au (E-TEK, 20 wt%) catalysts and PtAu bimetallic nanostructures. The CV was conducted in a 0.5-M H 2 SO 4 aqueous solution under the protection of argon. All measurements were run at ambient room temperatures and with a scan rate of 50 mv/s. For bimetallic PtAu/C catalysts the peak current density reached 359 ma/mg PtAu at around 0.55 V, which was substantially higher than 43 ma/mg Pt for Pt/C catalysts (Figure 5.7). The specific activity based on surface area should be even higher for PtAu/C catalysts, considering the average size of PtAu alloy nanoparticles was larger than that of Pt nanoparticles used as the reference. It seems that when platinum was used as the catalyst, both dehydrogenation and dehydration occurred in the oxidation of formic acid. The dehydrogenation happened at low potential range, while the dehydration became dominant at the high potential range resulting in increase in the surface coverage by adsorbed CO

220 Chapter intermediates. When potential was about 0.9 V, the adsorbed CO ads began to be oxidized. The observed high current density in the reverse sweep was associated 18, 20, 21 with direct formic acid oxidation after the removal of CO ads species. As a comparison, carbon-supported pure Au nanoparticles show no active toward the oxidation of formic acid (Figure 5.7). The mass current density associated with the oxidation of CO ads species, which is centered around 0.9 V, was much smaller for bimetallic PtAu/C catalysts than that for the Pt/C catalyst. The ratio between these two peak current densities (I 0.9V /I 0.55V ) in the positive scan decreased from 4.2 for Pt/C to 0.28 to PtAu/C. This result indicates dehydrogenation reaction was preferred over the Pt-rich surfaces of PtAu alloy heteronanostructures. The change of surface electronic structures could be a factor for the much improved activity in this bimetallic system, as the d-band center can shift from ev for platinum to around ev for either PtAu alloys or Pt overlayers on Au based on the DFT calculation Interestingly, the d-band center for PtAu alloy is similar to that for pure Pd metal (-1.83 ev), which catalyzes the oxidation of 51, 52 formic acid predominantly via the dehydrogenation pathway. CO stripping was used to study the shift of d-band center, as the affinity of CO molecule on Pt sites can be directly correlated to the change of peak positions due to the oxidation when the potential was above 0.60 V (Figure 5.10). For pure Pt catalysts, the current density for this oxidation peaked at 0.73 V. The absorbed CO could be completely removed from the surface during the first cycle, judging

221 Chapter by the disappearance of this signature peak and the subsequent appearance of hydrogen adsorption/desorption peaks. The on-set potential for PtAu/C catalysts however shifted positively by more than 100 mv during the first sweep. A peak centered at 0.84 V was readily detectable but diminished over cycles. These observations indicate the interactions between CO and the active site on PtAu surface were stronger than that for pure Pt, which could be attributed to an up shift in Pt d-band center due to the incorporation of Au. The accumulation of CO ads 26, 53 species on surface generally leads to reduced activities over times. Figure 5.11 shows the activities of carbon-supported PtAu and pure Pt catalysts after multiple CV cycles. The PtAu catalyst has a much higher mass current density than Pt at both 0.3 and 0.5 V during both the initial and subsequent cycles, demonstrating the superior performance of alloy catalyst. Overall, the mass current density for PtAu catalysts decreased only slightly and stabilized over cycles, suggesting the majority of CO ads species can be removed via potential cycling. Besides the electronic factor, incorporation of Au atoms on the surface helps in suppressing the dehydration pathway, which requires the presence of neighboring Pt sites for both generation and oxidation of CO-like intermediates. 54, 55 The modification in electronic and geometric structures of surface platinum due to the formation of PtAu bimetallic should be the main reasons for the observed enhancement in catalytic activities.

222 Chapter Figure 5.10 CV curves of CO stripping on surfaces of Pt (60 wt.% Pt) and PtAu bimetallic nanostructures. All tests were conducted in 0.1-M HClO 4 aqueous solutions under argon protection. Figure 5.11 Change of current density over CV cycles at 0.3 and 0.5 V for carbon-supported PtAu heteronanostructures and Pt nanoparticles. All experiments were conducted in a 0.5-M formic acid aqueous solution using 0.5-M H 2 SO 4 as the supporting electrolyte. 5.4 Conclusions Despite the lack of miscibility between Pt and Au metals in bulk materials at low temperature, Pt-surface rich PtAu bimetallic heteronanostructures can be produced from Pt-on-Au nanoparticles, which are made from molecular

223 Chapter precursors. Carbon-supported PtAu bimetallic catalysts exhibit much higher electrocatalytic activity than Pt in the oxidation of formic acid, partially because Pt-rich PtAu alloy surfaces have the appropriate electronic configurations that favor the dehydrogenation pathway and the geometric structures that help to reduce the CO poisoning. The conversion from particle-on-particle to Pt-surface rich bimetallic heteronanostructures can be very useful in the design and preparation of multifunctional nanoalloys. Author Contributions: This section is adapted from a paper published in Nano Research (Z. M. Peng, H. Yang, PtAu Bimetallic Heteronanostructures Made by Post-Synthesis Modifications of Pt-on-Au Nanoparticles, 2009, 2, ). Professor Hong Yang and Zhenmeng Peng conceived and designed the experiments, analysed the data and co-wrote the paper. Zhenmeng Peng conducted the synthesis and performed the electrochemical study. 5.5 References (1) Peng, Z. M.; Yang, H. Nano Today 2009, 4, 143. (2) Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R. R. Science 2007, 315, 220. (3) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J. J.; Lucas, C. A.; Wang, G. F.; Ross, P. N.; Markovic, N. M. Nat. Mater. 2007, 6, 241. (4) Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G. F.; Ross, P. N.; Lucas, C. A.; Markovic, N. M. Science 2007, 315, 493. (5) Park, I. S.; Lee, K. S.; Choi, J. H.; Park, H. Y.; Sung, Y. E. J. Phys. Chem.

224 Chapter C 2007, 111, (6) Habas, S. E.; Lee, H.; Radmilovic, V.; Somorjai, G. A.; Yang, P. Nat. Mater. 2007, 6, 692. (7) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal. B-Environ. 2005, 56, 9. (8) Burda, C.; Chen, X. B.; Narayanan, R.; El-Sayed, M. A. Chem. Rev. 2005, 105, (9) Rodriguez, J. A.; Goodman, D. W. Science 1992, 257, 897. (10) Stamenkovic, V. R.; Mun, B. S.; Mayrhofer, K. J. J.; Ross, P. N.; Markovic, N. M. J. Am. Chem. Soc. 2006, 128, (11) Tao, F.; Grass, M. E.; Zhang, Y. W.; Butcher, D. R.; Renzas, J. R.; Liu, Z.; Chung, J. Y.; Mun, B. S.; Salmeron, M.; Somorjai, G. A. Science 2008, 322, 932. (12) Koh, S.; Strasser, P. J. Am. Chem. Soc. 2007, 129, (13) Teng, X. W.; Black, D.; Watkins, N. J.; Gao, Y. L.; Yang, H. Nano Lett. 2003, 3, 261. (14) Teng, X. W.; Yang, H. Nanotechnology 2005, 16, S554. (15) Kashchiev, D., Nucleation: Basic Theory with Applications. Butterworth Heinemann: Oxford, (16) Zhang, J. L.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adzic, R. R. Angew. Chem.-Int. Edit. 2005, 44, (17) Adzic, R. R.; Zhang, J.; Sasaki, K.; Vukmirovic, M. B.; Shao, M.; Wang, J. X.; Nilekar, A. U.; Mavrikakis, M.; Valerio, J. A.; Uribe, F. Top. Catal. 2007, 46, 249. (18) Markovic, N. M.; Gasteiger, H. A.; Ross, P. N.; Jiang, X. D.; Villegas, I.; Weaver, M. J. Electrochim. Acta 1995, 40, 91. (19) Rice, C.; Ha, R. I.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.; Barnard, T. J. Power Sources 2002, 111, 83.

225 Chapter (20) Chen, Y. X.; Heinen, M.; Jusys, Z.; Behm, R. B. Angew. Chem.-Int. Edit. 2006, 45, 981. (21) Lovic, J. D.; Tripkovic, A. V.; Gojkovic, S. L. J.; Popovic, K. D.; Tripkovic, D. V.; Olszewski, P.; Kowal, A. J. Electroanal. Chem. 2005, 581, 294. (22) Zhu, Y. M.; Ha, S. Y.; Masel, R. I. J. Power Sources 2004, 130, 8. (23) Rice, C.; Ha, S.; Masel, R. I.; Wieckowski, A. J. Power Sources 2003, 115, 229. (24) Jeong, K. J.; Miesse, C. A.; Choi, J. H.; Lee, J.; Han, J.; Yoon, S. P.; Nam, S. W.; Lim, T. H.; Lee, T. G. J. Power Sources 2007, 168, 119. (25) Choi, J. H.; Jeong, K. J.; Dong, Y.; Han, J.; Lim, T. H.; Lee, J. S.; Sung, Y. E. J. Power Sources 2006, 163, 71. (26) Capon, A.; Parsons, R. J. Electroanal. Chem. 1973, 45, 205. (27) Matsumoto, F.; Roychowdhury, C.; DiSalvo, F. J.; Abruna, H. D. J. Electrochem. Soc. 2008, 155, B148. (28) Alden, L. R.; Han, D. K.; Matsumoto, F.; Abruna, H. D.; DiSalvo, F. J. Chem. Mat. 2006, 18, (29) Roychowdhury, C.; Matsumoto, F.; Zeldovich, V. B.; Warren, S. C.; Mutolo, P. F.; Ballesteros, M.; Wiesner, U.; Abruna, H. D.; DiSalvo, F. J. Chem. Mat. 2006, 18, (30) 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. J. Am. Chem. Soc. 2004, 126, (31) Greeley, J.; Norskov, J. K.; Mavrikakis, M. Annu. Rev. Phys. Chem. 2002, 53, 319. (32) Hammer, B.; Norskov, J. K., Theoretical surface science and catalysis - Calculations and concepts. In Advances in Catalysis, Academic Press Inc.: San Diego, USA, 2000; Vol. 45, pp 71. (33) Ruban, A.; Hammer, B.; Stoltze, P.; Skriver, H. L.; Norskov, J. K. J. Mol.

226 Chapter Catal. A-Chem. 1997, 115, 421. (34) Lee, J. K.; Lee, J.; Han, J.; Lim, T. H.; Sung, Y. E.; Tak, Y. Electrochim. Acta 2008, 53, (35) Kristian, N.; Yan, Y.; Wang, X. Chem. Commun. 2008, 353. (36) Kim, J.; Jung, C.; Rhee, C. K.; Lim, T. H. Langmuir 2007, 23, (37) Duan, H. W.; Nie, S. M. J. Am. Chem. Soc. 2007, 129, (38) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc.-Chem. Commun. 1994, 801. (39) Bauer, E.; Vandermerwe, J. H. Phys. Rev. B 1986, 33, (40) Porter, D. A.; Easterling, K. E., Phase Transformation in Metals and Alloys. Chapman & Hall: London, (41) Chambers, S. A. Surf. Sci. Rep. 2000, 39, 105. (42) Ji, X. H.; Song, X. N.; Li, J.; Bai, Y. B.; Yang, W. S.; Peng, X. G. J. Am. Chem. Soc. 2007, 129, (43) Hu, M.; Chen, J. Y.; Li, Z. Y.; Au, L.; Hartland, G. V.; Li, X. D.; Marquez, M.; Xia, Y. N. Chem. Soc. Rev. 2006, 35, (44) Xia, Y. N.; Halas, N. J. MRS Bull. 2005, 30, 338. (45) Vegard, L.; Dale, H. Zeits. Krist. 1928, 67, 148. (46) Peng, Z. M.; Yang, H. J. Solid State Chem. 2008, 181, (47) Xu, J. B.; Zhao, T. S.; Liang, Z. X. J. Phys. Chem. C 2008, 112, (48) Adhikari, H.; Marshall, A. F.; Goldthorpe, I. A.; Chidsey, C. E. D.; McIntyre, P. C. ACS Nano 2007, 1, 415. (49) Luo, J.; Wang, L. Y.; Mott, D.; Njoki, P. N.; Lin, Y.; He, T.; Xu, Z. C.; Wanjana, B. N.; Lim, I. S.; Zhong, C. J. Adv. Mater. 2008, 20, (50) Zhao, D.; Xu, B. Q. Phys. Chem. Chem. Phys. 2006, 8, (51) Zhou, W. P.; Lewera, A.; Larsen, R.; Masel, R. I.; Bagus, P. S.; Wieckowski, A. J. Phys. Chem. B 2006, 110, (52) Hoshi, N.; Kida, K.; Nakamura, M.; Nakada, M.; Osada, K. J. Phys. Chem.

227 Chapter B 2006, 110, (53) Jiang, J.; Kucernak, A. J. Electroanal. Chem. 2002, 520, 64. (54) Kristian, N.; Yan, Y. S.; Wang, X. Chem. Commun. 2008, 353. (55) Neurock, M.; Michael, J.; Wieckowski, A. Faraday Discuss. 2008, 140, 363.

228 Chapter Chapter 6 Ultrafine Metal and Metal Alloy Nanoparticles for Catalytic Applications 6.1 Introduction Ultrafine metal nanoparticles with their size as small as sub-nanometer have attracted significant amount of research attentions for their novel reactivity that has notbeen observed as advanced catalysts in the large particles. 1-5 One example is the size-dependent catalytic activity of gold nanoparticles in low-temperature CO oxidation. Gold is almost inactive at 300 K with particle size larger than about 6 nm but can exhibit dramatically high activity once it reaches below certain critical size Current understandings of their distinct catalytic property have been associated with the large fraction of under-coordinated surface atoms and their possible effects on the electronic structures, both of which can result in dramatically different interactions with the reacting molecules and even alter the reaction pathway. Another benefit from using ultrafine metal nanoparticles as catalysts is their extra large specific surface area, which can maximize the utilization of metals. It is especially meaningful for noble metal 15, 16 catalysts that the cost is often an important issue. Despite of all the advantages that ca be harvested from ultrafine metal nanoparticle catalysts, there have only been very limited successes in the 4, 17 preparation of such catalysts using traditional synthetic methods. Preventing the growth of particles after the nucleation remain to be a main challenge in

229 Chapter making ultrafine metal nanoparticles. The size of metal nanoparticles synthesized using colloidal methods is typically well above 2 nm, even though burst nucleation can help to maximize the generation of nuclei and the use of capping agents can prohibit the overgrowth of the formed nanoparticles. 18 In this chapter, I present an alternate method to prepare supported ultrafine metal nanoparticles. Rather than a direct synthesis of ultrafine metal nanoparticles, the procudure starts from the synthesis of alloy nanoparticles which contain only a small amount of the targeted metal and a second sacrificial metal element. Dealloying based on the difference in their redox potentials can lead to a selective dissolution of the sacrificial element Uniform 1-nm platinum, gold and palladium nanoparticles have been obtained upon a complete dissolution of silver metal in an acid solution from 4-nm Pt 1 Ag 27, Au 1 Ag 25 and Pd 1 Ag 28 nanoparticles, respectively. These ultrafine metal nanoparticles have been loaded on various supports to study their catalytic property in nitrophenol reduction reaction. Carbon-supported ultrafine platinum nanoparticles are highly active compared with that of reference platinum catalyst with a large size. 6.2 Experimental Section Materials. Platinum acetylacetonate (Pt(acac) 2, Strem Chemicals, 98%), gold acetate (Au(ac) 3, Alfa Aesar, 99.9%), palladium acetylacetonate (Pd(acac) 2, Strem Chemicals, 99%) and silver stearate (Ag(St), Alfa Aesar) were purchased from

230 Chapter VWR. Oleic acid (OA, 90%, technical grade), oleylamine (OAm, 70%, technical grade), 1, 2-hexadecanediol (HDD, 90%, technical grade), diphenyl ether (DPE, 99%, ReagentPlus ), and titanium oxide nanoparticle dispersions in xylene (TiO 2, 99.9% trace metal basis) were from Aldrich. All chemicals and reagents were used as received without further purification. The commercial carbon-supported platinum (20 wt.% Pt) catalyst was from E-TEK, Inc. Synthesis of trace-m content MAg (M= Pt, Au, Pd) alloy nanoparticles. All experiments were carried out under argon using the standard Schlenk line technique. In a typical procedure, HDD (0.49 g or 1.9 mmol), OA (0.3 ml or 0.9 mmol), and OAm (0.3 ml or 0.9 mmol) were mixed with DPE (4 ml or 25.2 mmol) in a 25-mL three neck round-bottom flask and preheated to 200 C. In a separate flask, 4 µmol of metal precursors (1.6 mg of Pt(acac) 2, 1.5 mg of Au(ac) 3, or 1.2 mg of Pd(acac) 2 ) and Ag(St) (50.0 mg or mmol) were mixed with DPE (1 ml or 6.3 mmol) and heated to 80 C until the solid was completely dissolved. The latter solution was injected into the flask preheated at 200 C and held for 1 h. The resultant mixtures were washed twice with 2 ml of toluene and 6 ml of ethanol, followed by centrifugation at 6000 rpm for 5 min. Preparation of carbon-supported MAg (M = Pt, Au, Pd) nanoparticles. Carbon black (Vulcan XC-72R) was used to support the as-made MAg nanoparticles. In general, a suspension of 60 mg of carbon black in 5 ml of toluene was sonicated for 60 min before the addition of MAg nanoparticles

231 Chapter dispersed in 2 ml of toluene. The resultant mixture was stirred overnight before the solids were collected via centrifugation at 6000 rpm for 5 min, which were dried at ambient room temperature. Preparation of SiC-supported PtAg nanoparticles. In a typical procedure, 100 mg of SiC support was dispersed in 5 ml of chloroform and sonicated for 60 min, followed by mixing with the MAg nanoparticles in 2 ml of chloroform. The resultant mixture was stirred overnight before the solids were collected via centrifugation at 6000 rpm for 5 min and dried at room temperature. Preparation of TiO 2 -supported PtAg nanoparticles. Generally a suspension of 100 mg of TiO 2 nanoparticles in 0.2 ml of xylene was added into 5 ml of toluene and sonicated for 60 min. The MAg nanoparticles dispersed in 2 ml of toluene was mixed thereafter. The resultant mixture was stirred overnight before the solids were collected via centrifugation at 6000 rpm for 5 min and dried at room temperature. Preparation of supported ultrafine Pt, Au and Pd nanoparticles. The ultrafine metal nanoparticles were obtained by an acid treatment of the supported MAg nanoparticles, which involved the dissolution of Ag metal. Specifically, 50 mg of MAg nanoparticles on various supports were mixed with 30 ml of nitric acid aqueous solution (3 M) and sonicated for 10 min to make a uniform dispersion. The resultant mixture was stirred overnight before the solids were collected via centrifugation at 6000 rpm for 5 min, and dried at ambient room

232 Chapter temperature. An AgPd sample of 20 mg was dispersed in 30 ml of 1-M nitric acid aqueous solution and mixed for 3 h to prepare ultrafine palladium nanoparticles since palladium metal can dissolve in concentrated nitric acid solution. Characterization. A detailed description regarding characterization of the samples has been presented in Chapter 2.3. Briefly, TEM and HR-TEM images were taken on a FEI TECNAI F-20 microscope. STEM images were carried out under a HAADF mode on the same microscope. EDX spectra were taken on a FE-SEM Zeiss-Leo DSM982. Catalytic property. The reduction of p-nitrophenol by sodium borohydride (NaBH 4 ) to p-aminophenol was used as a model reaction to evaluate the catalytic activity of carbon-supported ultrafine platinum nanoparticles. In a typical procedure, 5 mg of carbon-supported platinum catalyst was dispersed in 5 ml of distilled water and sonicated for 10 min to make the uniform ink. 50 µl of the catalyst ink was mixed with 30 ml of p-nitrophenol aqueous solution (0.1 mm), which was purged ahead of the test by a flow of argon for 30 min to remove the dissolved oxygen. After this step, 0.5 ml of sodium borohydride solution (2.5 M) was added to initiate the reaction. The reaction was followed by measuring the extinction of the solution at 400 nm in situ using a fiber optic spectrometer (Ocean Optics, USB2000). Commercial Pt catalyst (TKK, 46.7 wt% Pt) was

233 Chapter also tested using a similar procedure for comparison. The reduction is a first-order reaction: 22 dc = kappc (6.1) dt or ln C C t 0 = k app t (6.2) where C t and C 0 are the concentrations of p-nitrophenol at time t and right before the reaction, k app is the apparent rate constant. The activation energy for this reaction, E a, can be obtained by using the Arrhenius equation: k app Ea / RT = Ae (6.3) where A is the pre-exponential factor and R is the gas constant. 6.3 Results and Discussion Figure 6.1a shows the TEM image of as-synthesized PtAg alloy nanoparticles using a non-hydrolytic method reported elsewhere. 23 Both platinum and silver precursors can be reduced simultaneously at 200 C in the presence of 1,2-hexadecanediol and form alloy nanoparticles. The particles appeared to be spherical in shape and monodisperse in size distribution, with an average diameter of 4.0 ± 0.5 nm (Figure 6.1b). Both platinum and silver metals were detected using EDX) spectroscopy, suggesting existence of both elements in the prepared nanoparticles (Figure 6.1c). The copper and carbon signals were from the TEM

234 Chapter grid and capping agent on the surface of particles, respectively. The atomic ratio between Pt and Ag in the PtAg alloy nanoparticles was 1:27 based on EDX analysis. This value was close to the Pt(acac) 2 /Ag(St) molar ratio of 1:30, indicating most metal precursors were reduced during the reaction. These Pt 1 Ag 27 nanoparticles were loaded onto different supports. Figure 6.2a-c show the TEM images of carbon-, silicon carbide-, and titanium oxide-supported Pt 1 Ag 27 nanoparticles, which were observed distributing uniformly on these supports. Lattice fringes could be clearly observed from these metal nanoparticles, suggesting good crystallinity. These supported Pt 1 Ag 27 particles could also be observed using HAADF-STEM. Figure 6.1 (a) TEM image, (b) size distribution, and (c) EDX spectrum of as-synthesized PtAg nanoparticles.

235 Chapter Figure 6.2 (a-c) TEM and (d-f) STEM images of PtAg nanoparticles on (a,d) carbon, (b,e) silicon carbide, and (c,f) titanium oxide supports. Figure 6.3 (a) TEM image and (b) size distribution of the carbon-supported ultrafine Pt nanoparticles. The supported Pt 1 Ag 27 nanoparticles were mixed with a 3-M HNO 3 aqueous solution to removeag metal selectively. Only ultrafine metal nanoparticles around 1 nm in size could be observed on the carbon support under TEM after acid-treatment, indicating a dramatic shrinkage of the size of Pt 1 Ag 27 nanoparticles (Figure 6.3a). Statistic analysis on the obtained metal

236 Chapter nanoparticles gave an average particle size of 1.3 ± 0.2 nm, one third of that for the Pt 1 Ag 26.8 particles, and is a good indication that silver species can be remove without collapsing the nanoparticles. The different atomic number between the platinum and carbon elements resulted in a good contrast between these two elements under a HAADF-STEM mode (Figure 6.4a), where the platinum metal Figure 6.4 STEM and HR-TEM images of the prepared ultrafine Pt nanoparticles on (a, b) carbon, (c, d) silicon carbide, and (e, f) titanium oxide supports. has larger Z number and thus significantly brighter contrast than carbon. HR-TEM study shows most of the Pt nanoparticles contain only dozens of

237 Chapter atoms with distorted lattice fringes most likely due to their ultra small size (Figure 6.4b). 12 No silver signal could be detected using EDX spectroscopy, suggesting the acid treatment was efficient in removing silver metal and the resultant particles were made of pure platinum (Figure 6.5a). Ultrafine platinum nanoparticles on other supports, such as silicon carbide and titanium oxide, could also be prepared using a similar approach and characterized by using HAADF-STEM and HR-TEM (Figure 6.4c-f). Only those singles from platinum and elements of the supporting materials were observed, suggesting pure platinum on these two supports (Figure 6.5b,c). Figure 6.5 EDX spectra of the prepared ultrafine Pt nanoparticles on (a) carbon, (b) silicon carbide, and (c) titanium oxide supports.

238 Chapter Figure 6.6 (a) TEM and HR-TEM (Inset) images and (b) EDX spectrum of carbon-supported AuAg nanoparticles, and (c) TEM and HR-TEM (Inset) images and (d) EDX spectrum of carbon-supported PdAg nanoparticles. This dealloying method can be used for preparation of ultrafine nanoparticles of metals other than Pt. We have synthesized low-au content AuAg nanoparticles and low-pd content PdAg nanoparticles using a similar procedure. Both AuAg and PdAg had an average particle size of about 4 nm (Figure 6.6). Au(ac) 3 was used as the gold precursor because the commonly used HAuCl 4 reacts with Ag(St) and produces AgCl as the by-product that interfere with the reaction. Figure 6.6a shows the TEM image of AuAg nanoparticles on carbon support. HR-TEM studies on individual AuAg nanoparticles show their well-defined lattice fringes and indicate a good crystallinity. The EDX spectroscopy detected the existence of both gold and silver metals and the chemcial composition was Au 1 Ag 25 (Figure 6.6b). The low-pd content PdAg nanoparticles were also observed to be

239 Chapter distributed evenly on the carbon surface and had a good crystallinity (Figure 6.6c). The overall composition of the nanoparticles was calculated to be Pd 1 Ag 28, indicating the successful preparation of low-pd content PdAg nanoparticles on carbon support. Figure 6.7 shows the obtained ultrafine gold and palladium nanoparticles by dissolution of silver in nitric acid solution. Both metal nanoparticles were around 1 nm in size and had clear lattice fringes (Figure 6.7a, c). EDX spectroscopy suggested a complete removal of silver species, with pure ultrafine gold and palladium nanopaticles as the final product (Figure 6.7b, d). Figure 6.7 (a) STEM and HR-TEM (inset) images and (b) EDX spectrum of the prepared carbon-supported ultrafine Au nanoparticles, and (c) STEM and HR-TEM (Inset) images and (d) EDX spectrum of the prepared carbon-supported ultrafine Pd nanoparticles.