Energy & Environmental Science PAPER

Energy & Environmental Science

View Online / Journal Homepage / Table of Contents for this issue

C

Dynamic Article Links
0.9 V), further confirming the formation of a complete Pt shell on an IrNi alloy core. On the other hand, the XANES spectra of the Pt L3 edge from PtML/IrNi/C (Fig. 2(a)) indicate a very small potential dependence. This effect is more evident in the relative change of the X-ray absorption peak intensity of the Pt L3 edge spectra for PtML/IrNi/C and Pt/C as a function of potential (Fig. 2(d)). The increase in the intensity of the absorption edge peak for the PtML/IrNi/C electrocatalyst commences at considerably higher potentials than does that for the Pt/C catalyst. The difference in absorption properties between PtML/IrNi/C and pure Pt is due to modification of the electronic properties of the surface atoms by the underlying metal via geometric strain and ligand interactions.15,28 The high Energy Environ. Sci., 2012, 5, 5297–5304 | 5299

Downloaded on 16 February 2012 Published on 28 October 2011 on http://pubs.rsc.org | doi:10.1039/C1EE02067F

View Online

Fig. 2 In situ XANES for the Pt monolayer IrNi electrocatalyst at various potentials, (a) Pt L3 edge, (b) Ir L3 edge and (c) Ni K edge. (d) A comparison of the change of the adsorption peak as a function of potential for PtML/IrNi/C and Pt/C.

Pt oxidation of the PtML/IrNi/C electrocatalyst (significant delay in PtOH formation) clearly indicates the better stability of the electrocatalyst, which is also discussed further in this paper. The Pt and Ir mole ratio was 3 : 7 which was determined by the ex situ XAS analysis (ESI, Fig. S1†). So, from combined XAS and EDX/TEM elemental analysis we get the final product molar ratio as Pt Ir2.29 Ni1.29. 3.3 ORR activity of Pt monolayer on IrNi nanoparticle electrocatalyst The cyclic voltammetry curves for the thin-film electrodes of the PtML/IrNi/C and the IrNi/C nanoparticles in 0.1 M HClO4 solution show no anodic currents that can be ascribed to the oxidation/dissolution of Ni, demonstrating that Ni is protected

by the Ir shell (ESI, Fig. S2†). Furthermore, the apparent peaks of upd of Cu imply that the surface mainly consists of Ir, since upd does not take place on a Ni surface. The Pt loading is 8.65 nmol (8.61 mg cm2), calculated from the Cu upd. Fig. 3(a) compares the currents obtained during the ORR in the anodic sweep cycle on the PtML/IrNi/C, PtML/Ir/C and Pt/C (E-tek 20%) electrocatalyst at 1600 rpm in 0.1 M HClO4 solution. Though the kinetic currents for Pt/C (10 mg cm2 (10 nmol) Pt loading) are slightly higher than PtML/Ir/C (3.6 mg cm2 Pt), there is not much change in the ORR activity. Moreover, alloying Ni with Ir (PtML/IrNi/C) one can clearly observe the increases in diffusion current and activity for the ORR. This enhancement in activity may be due to electronic modifications induced by strain, ligand and segregation effects from bimetallic IrNi cores. It has always been difficult to separate the role of strain and ligand effects in

Fig. 3 (a) Polarization curves for the ORR of the PtML/IrNi/C and PtML/Ir/C and Pt/C (E-tek 20%) nanoparticle electrocatalysts at 1600 rpm in oxygensaturated 0.1 M HClO4. (b) A comparison of Pt mass activities for Pt/C (E-tek 20%), PtML/Ir/C and PtML/IrNi/C synthesized using RDE and scale-up synthesis method.

5300 | Energy Environ. Sci., 2012, 5, 5297–5304

This journal is ª The Royal Society of Chemistry 2012


View Online

ORR activity enhancements as both play a cumulative effect on the modification of chemical properties of these surfaces.30 Using the polarization curves of the PtML/IrNi/C electrocatalyst in 0.1 M HClO4 solution purged by O2 at different rotation speeds at a scan rate of 10 mV s1, the inverse current density (1/j) was plotted as a function of the inverse of the square root of the rotation rate (u1/2), also known as a Koutecky–Levich plot (ESI, Figs S3 and S4, respectively†). The linearity and parallelism of the Koutecky- Levich plots suggest first-order kinetics with respect to molecular oxygen.29 Fig. 3(b) compares the Pt mass activities at 0.9 V for Pt/C (20%E-tek), PtML/Ir/C and PtML/IrNi/C nanoparticles. The Pt mass activities for our scaledup PtML/IrNi/C electrocatalyst and RDE synthesized PtML/IrNi/ C electrocatalyst are, respectively, three to six times that of the Pt/C electrocatalyst. The Pt specific activities of the PtML/IrNi/C, PtML/Ir/C, and Pt/C electrocatalysts at 0.9 V/RHE are 0.60, 0.13, and 0.24 mA cm2, respectively. We undertook DFT calculations to elucidate the observed enhancement of activity of PtML/IrNi/C compared to that of Pt/C and PtML/Ir/C electrocatalysts. To simulate the nanoparticles properly, and to save computational time, we constructed sphere-like nanoparticle models of 1.7 nm based on a truncated octahedron (Fig. 4). Since our model nanoparticle is smaller than the average size of the experimental ones (5.0 nm),18 we could not use the experimental mole ratio of Ni/ Ir ¼ 0.56 with the two Ir monolayers for the PtMLIrNi model.18

Thus, we used only the mole ratio of the IrNi subcore (Ni/Ir ¼ 0.37). As Fig. 4(d) shows, while the IrNi core has one Ir submonolayer (60 Ir atoms), it retains the Ir/Ni ratio of 0.36 by replacing 14 of the 33 Ir atoms in the subcore with Ni atoms. As Fig. 4 illustrates, we calculated the BE–O that serves as a descriptor for scaling the ORR activity.26 Fig. 4(e) shows that partially replacing Ni in the core weakens the BE–O, and introduces more contraction therein (3.45 eV and 4.12%, respectively) compared to that of PtMLIr (3.52 eV and 3.81%, correspondingly). Fig. 4(f) plots the BE–O variations against the d-band center, in which the trend from Pt to PtMLIr nanoparticles agrees with that reported for the extended surfaces (Pt(111) and PtMLIr(111)).31 In addition, this figure also shows that using Ir or IrNi as a core shifts the d-band center of Pt in the shell away from the Fermi level, thereby lowering the BE–O.10 Interestingly, the partial replacement of Ir with Ni entails a slightly weaker BE–O than that of PtMLIr (0.1 eV) (Fig. 4(f)). Therefore, it is likely that extending the partial replacement of Ir by Ni in PtMLIrNi should augment surface strain, and accordingly, lower the BE–O; expectedly, there should be a maximum for this due to the low stability of the Ni core under the ORR conditions.32,33 To verify the trend of weakening of the BE–O upon mixing Ni with Ir, we calculated the extreme case with a pure Ni core (PtMLNi); this led to the most strained surface and the weakest BE–O (5.60% and

Fig. 4 (a) A schematic of a sphere-like nanoparticle considered in DFT calculations representing adsorption of atomic oxygen at the fcc site. Cross sectional views of nanoparticle models of (b) Pt, (c) PtMLIr, and (d) PtMLIrNi with 1.7 nm. The dotted circles represent Ir in sub-core. (e) The DFT calculated binding energy of oxygen (BE–O) as a function of strain on PtMLIrNi, PtMLIr and Pt using the nanoparticle models. (f) The DFT calculated binding energy of oxygen (BE–O) as a function of the average d-band center of metals interacting with O on PtIr1,MLIrNi, PtMLIrNi, PtMLIr, and Pt. (g) The Pt specific activity against BE–O on PtIr1,MLIrNi, PtMLIr and Pt. ‘‘ML’’ is the monolayer.


Energy Environ. Sci., 2012, 5, 5297–5304 | 5301


View Online

3.12 eV, respectively). Careful structural examination showed that PtMLNi is slightly collapsed inward, causing more surface contraction than the other nanoparticles do. Based on the calculated strain, d-band center, and BE–O, a descriptor for the ORR activity,26 it is anticipated that PtMLIr and PtMLIrNi may have similar ORR activities, which cannot fully support experimental findings. Thus to gain a better understanding of the experimentally measured ORR activity, we also considered the possibility of anti-segregation under the ORR condition. Possibly, the Ir in the core might be exchanged with the Pt in the shell, driven by a stronger Ir–O interaction than that of Pt–O. To examine anti-segregation effects, we interchanged one Ir atom in the core with its interconnected Pt atom in the shell (in our notation, PtIr1,MLIrNi for PtMLIrNi, and PtIr1,MLIr for PtMLIr. In agreement with a previous study by others on extended surfaces,10 our calculations clearly show that for bare clusters, Ir preferentially remains in the core due to the energy cost of such transitions, i.e., 0.64 eV for PtMLIr / PtIr1,MLIr, and 0.53 eV for PtMLIrNi / PtIr1,MLIrNi. In addition, we noted that it is energetically more costly to pull out Ir from the PtMLIr core and move it to the shell than it is when Ir is in a PtMLIrNi core. Therefore, we considered only the anti-segregation of Ir of the latter, rather than the former. On exposure to oxygen, as shown in Fig. 4(f), O predominantly moves to the Pt–Pt–Ir hybrid hollow site, and the corresponding BE–O is negatively shifted, 3.83 eV. Consequently, the cost of pulling one Ir atom out to the shell declines from 0.53 eV to 0.14 eV. That is, on interacting with oxygen, the anti-segregation of Ir is energetically more likely than that on the bare cluster, so strengthening the O–PtMLIrNi interaction. Fig. 4(g) illustrates the calculated BE–Os versus the measured specific activities for Pt/C, PtMLIr/C, and PtMLIrNi/C. We note that the BE–O for PtMLIrNi corresponds to the strength of O binding after the exchange of one Ir atom in the core with one Pt atom in the shell (PtIr1,MLIrNi). Pt nanoparticles bind O too strongly to be removed from the catalyst, while PtMLIr does so too weakly to dissociate oxygen; hence, it lowers the ORR activity more than Pt does. In contrast, the partial replacement of Ir in the core with Ni favors the anti-segregation of Ir to the shell, and PtMLIrNi supports a moderate binding of oxygen between Pt and PtMLIr, so facilitating the efficient breakage of the O–O bond, and allowing the facile removal of O under fuel cell conditions. Overall, it is evident that the geometric, electronic, and segregation effects under fuel cell conditions have a key role in rationalizing the increase in the ORR activity of PtMLIrNi/C compared to that of PtMLIr/C and Pt/C.

3.4 Stability of Pt monolayer on IrNi nanoparticle electrocatalyst The long-term stability of ORR electrocatalysts is another critical requirement for applying them in fuel cells. We evaluated the stability of the Pt monolayer electrocatalyst by an accelerated test involving potential cycling between 0.6 V and 1.0 V/RHE at a scan rate of 50 mV s1 in an air-saturated HClO4 solution. Table 1 summarizes the loss in Pt and noble metal mass activity for the PtML/IrNi/C electrocatalyst prepared on the glassy carbon electrode and in the scale-up synthesis method. Voltammetry was used to determine the Pt surface area of the electrodes by measuring H adsorption before and after potential cycling. Integrating the charge between 0 and 0.33 V associated with H adsorption for PtML/IrNi/C shows a small loss of Pt surface area (ESI, Fig. S5 inset†). The table also shows the percentage loss in the initial surface area and the half wave potential after the accelerated cycling test. A decrease in H adsorption and desorption charges was observed after the potential cycling test compared with those before cycling. The catalyst made on a glassy carbon electrode showed a 45% loss in the surface area and 18 mV loss in half-wave potential after 30 000 cycles. The catalyst activity was much more stable for the sample obtained using the scale-up synthesis method. The Pt mass activity of the PtML/IrNi/C electrocatalyst after 50 000 cycles is 0.57 A mg1Pt at 0.9 V/RHE. The polarization curves at 1600 rpm on the PtML/IrNi/C sample before and after 50 000 potential cycles in 0.1 M HClO4 show no change in activity (ESI, Fig. S5†). The catalyst made by the scale-up synthesis method showed only a 23% loss in the Pt surface area and no loss in half-wave potential was observed even after 50 000 cycles. This demonstrates the advantages of applying the galvanic displacement of a Cu upd layer to the large-scale synthesis of catalysts with welldefined core–shell structures. There have always been questions about Ni leaching out upon potential cycling in an acidic environment, so to check the hypothesis we carried out STEM/EELS analysis on numerous individual particles obtained after potential cycling. Fig. 5(b) shows the distribution of Pt and IrNi components in a single representative nanoparticle after 50 000 cycles accelerated potential cycling. Fig. 5(a) shows a HAADF image of a single nanoparticle after 50 000 potential cycles. The intensity of the spectra for the Pt M-edge, Ir M-edge and Ni L-edge are integrated along the line for a given energy region and the integrated values along the scanned line (as indicated in Fig. 5(a) and shown in Fig. 5(b)). It is clearly shown that the core–shell structure of

Table 1 Mass activity, half-wave potentials and % loss in Pt surface area for the PtML/IrNi/C catalyst before and after potential cycling Synthesis on glassy carbon electrode

Pt PGM

Scale-up synthesis method

Mass activity (jk/A mg1)

Half-wave potential (mV)

Initial (0.9 V)

After 30000 cycles (0.9 V)

Initial

After 30 000 cycles

% loss in surface area After 30 000 cycles

1.4 0.78

0.48 0.27

890

872

45


Mass activity (jk/A mg1)

Half-wave potential (mV)

Initial (0.9 V)

After 50 000 cycles (0.9 V)

Initial

After 50 000 cycles

% loss in surface area After 50 000 cycles

0.71 0.22

0.57 0.17

851

851

23



View Online

Fig. 5 (a) A representative HAADF-STEM image of a PtML/IrNi/C nanoparticle after potential cycling test. (b) A comparison of the corresponding HAADF Pt (black), Ir (red) and Ni (blue) EELS intensity profiles in a line scan indicated.

the IrNi nanoparticle is maintained even after 50 000 cycles of potential cycling. The Pt and Ir shell protects Ni from leaching out. The Pt shell on IrNi nanoparticles is also intact which clearly justifies that the IrNi core is a very stable substrate for the Pt monolayer.

4. Conclusions We demonstrated the formation of a new class of core–shell electrocatalysts for the ORR, consisting of a Pt monolayer shell and bimetallic IrNi core. Thus, we coupled highly stable, inexpensive core–shell nanoparticles with a Pt monolayer to produce electrocatalysts with high activity and very high stability. The results show that the activity of Pt monolayer electrocatalysts for the ORR is strongly substrate metal-dependent and our IrNi core–shell structured substrate is a new very suitable core for Pt monolayer electrocatalysts. The electronic structure of substrates is an important factor, because it can alter the catalytic activity of deposited Pt. Our DFT calculations using a sphere-like model clearly demonstrate that mixing Ni with Ir (PtML/IrNi/C) induces geometric, electronic, and segregation effects and thus weakens the binding energy of oxygen, resulting in higher activity than pure Pt/C and PtML/Ir/C electrocatalysts. The galvanic displacement of a Cu upd layer demonstrated through gram-scale synthesis of Ptshell on IrNi core catalysts seems very promising in scale-up synthesis of highly active, cost-effective nanoparticles with a core– shell structure suitable for specific catalytic reaction requirements. We consider that these data confirm that the Pt monolayer approach holds excellent potential for creating efficient fuel cell electrocatalysts ready for commercialization.

Acknowledgements This work is supported by U.S. Department of Energy, Divisions of Chemical and Material Sciences under the Contract No. DE-AC02-98CH10886. AIF acknowledges support by DOE BES Grant DE-FG02-03ER15476. Beamlines X19A at the NSLS are supported in part by the Synchrotron Catalysis Consortium, U. S. Department of Energy Grant No DE-FG0205ER15688.We thank the National Energy Research Scientific Computing (NERSC) Center, BNL’s Center for Functional Nanomaterials (CFN), and Prof. M. C. Lin for CPU time. This journal is ª The Royal Society of Chemistry 2012

References 1 W. Vielstich, A. Lamm, H. A. Gasteiger, Handbook of Fuel Cells – Fundamentals, Technology and Applications, John Wiley & Sons, Chichester, 2003. 2 H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal., B, 2005, B 56, 9. 3 R. R. Adzic, in Frontiers in Electrochemistry, Vol. 5, Electrocatalysis, ed. J. Lipkowski and P. N. Ross, VCH Publishers, New York, 1998; p 197. 4 S. Mukerjee, S. Srinivasan, M. P. Soriaga and J. Mcbreen, J. Phys. Chem., 1995, B 99, 4577. 5 S. Chen, P. J. Ferrira, W. Sheng, N. Yabuuchi, L. Allard and Y. ShaoHorn, J. Am. Chem. Soc., 2008, 130, 13818. 6 V. R. Stamenkovic, B. Fowler, B. S. Mun, G. F. Wang, P. N. Ross, C. A. Lucas and N. M. Markovic, Science, 2007, 315, 493. 7 R. R. Adzic, J. Zhang, K. Sasaki, M. B. Vukmirovic, M. Shao, J. X. Wang, A. U. Nilekar, M. Mavrikakis and F. Uribe, Top. Catal., 2007, 46, 249. 8 M. H. Shao, K. Sasaki, N. S. Marinkovic, L. Zhang and R. R. Adzic, Electrochem. Commun., 2007, 9, 2848. 9 J. Zhang, F. H. B. Lima, M. H. Shao, K. Sasaki, J. X. Wang, J. Hanson and R. R. Adzic, J. Phys. Chem. B, 2005, 109(48), 22701–22704. 10 B. Hammer and J. K. Nørskov, Adv. Catal., 2000, 45, 71. 11 J. A. Rodriguez, Surf. Sci. Rep., 1996, 24, 225. 12 W. Zhou, K. Sasaki, D. Su, Y. Zhu, J. Wang and R. R. Adzic, J. Phys. Chem. C, 2010, 114, 8950–8957. 13 K. Lee, L. Zhang and J. Zhang, J. Power Sources, 2007, 170, 291. 14 J. Qiao, R. Lin, B. Li, J. Ma and J. Liu, Electrochim. Acta, 2010, 55, 8490–8497. 15 F. H. B. Lima, J. L. Zhang, M. H. Shao, K. Sasaki, M. B. Vukmirovic, E. A. Ticianelli and R. R. Adzic, J. Phys. Chem. C, 2007, 111, 404–410. 16 J. R. Kitchin, J. K. Norskov, M. A. Barteau and J. G. Chen, Phys. Rev. Lett., 2004, 93, 156801. 17 M. Mavrikakis, B. Hammer and J. K. Norskov, Phys. Rev. Lett., 1998, 81, 2819. 18 K. Sasaki, K. Kuttiyiel, L. Barrio, D. Su, A. I. Frenkel, N. Marinkovic, D. Mahajan and R. Adzic, J. Phys. Chem. C, 2011, 115, 9894–9902. 19 K. Sasaki, J. X. Wang, H. Naohara, N. Marinkovic, K. More, H. Inada and R. R. Adzic, Electrochim. Acta, 2010, 55, 2645– 2652. 20 J. McBreen, W. E. O’Grady, K. I. Pandya, R. W. Hoffman and D. E. Sayers, EXAFS Study of the Nickel-Oxide Electrode, Langmuir, 1987, 3, 428–433. 21 B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537. 22 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter, 1993, 47, 558–561. 23 G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter, 1996, 54, 11169–11186. 24 P. E. Bl€ ochl, Phys. Rev. B: Condens. Matter, 1994, 50, 17953– 17979.

Energy Environ. Sci., 2012, 5, 5297–5304 | 5303

View Online 29 N. A. Anastasijevic, V. Vesovic and R. R. Adzic, J. Electroanal. Chem., 1987, 229, 317. 30 J. R. Kitchin, J. K. Nørskov, M. A. Barteau and J. G. Chen, Phys. Rev. Lett., 2004, 93, 15. 31 J. L. Zhang, M. B. Vukmirovic, Y. Xu, M. Mavrikakis and R. R. Adzic, Angew. Chem., Int. Ed., 2005, 44, 2132. 32 E. Antolini, J. R. C. Salgado and E. R. Gonzalez, J. Power Sources, 2006, 160, 957. 33 M. Watanabe, K. Tsurumi, T. Mizukami, T. Nakamura and P. Stonehart, J. Electrochem. Soc., 1994, 141, 2659.


25 B. Hammer, L. B. Hansen and J. K. Nørskov, Phys. Rev. B: Condens. Matter, 1999, 59, 7413–7421. 26 J. X. Wang, H. Inada, L. Wu, Y. Zhu, Y. Choi, P. Liu, W.-P. Zhou and R. R. Adzic, J. Am. Chem. Soc., 2009, 131, 17298– 17302. 27 K. Sasaki, H. Naohara, Y. Cai, Y. Choi, P. Liu, M. B. Vukmirovic, J. Wang and R. R. Adzic, Angew. Chem., Int. Ed., 2010, 49, 8602– 8607. 28 Ma Yuguang and Perla B. Balbuena, J. Electrochem. Soc., 2010, 157, B959–B963.

