New Opportunity for in Situ Exsolution of Metallic Nanoparticles on

Jul 25, 2016 - One of the main challenges for advanced metallic nanoparticles (NPs) supported functional perovskite catalysts is the simultaneous achi...
2 downloads 8 Views 6MB Size
Download PDF
Letter pubs.acs.org/NanoLett

New Opportunity for in Situ Exsolution of Metallic Nanoparticles on Perovskite Parent Yi-Fei Sun,† Ya-Qian Zhang,† Jian Chen,‡ Jian-Hui Li,*,§,∥ Ying-Tao Zhu,∥ Yi-Min Zeng,⊥ Babak Shalchi Amirkhiz,⊥ Jian Li,# Bin Hua,*,† and Jing-Li Luo*,† †

Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta T6G 1H9, Canada National Institution for Nanotechnology, National Research Council, Edmonton, Alberta T6G 2M9, Canada § National Engineering Laboratory for Green Chemical Productions of Alcohols-Ethers-Esters, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen, Fujian 361005, P. R. China ∥ Department of Physics, University of Changji, Changji, Xinjiang 831100, P. R. China ⊥ Canmet MATERIALS, Natural Resources Canada, Hamilton, Ontario L8P 0A5, Canada # School of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan, Hubei 430074, P. R. China ‡

S Supporting Information *

ABSTRACT: One of the main challenges for advanced metallic nanoparticles (NPs) supported functional perovskite catalysts is the simultaneous achievement of a high population of NPs with uniform distribution as well as long-lasting high performance. These are also the essential requirements for optimal electrode catalysts used in solid oxide fuel cells and electrolysis cells (SOFCs and SOECs). Herein, we report a facile operando manufacture way that the crystal reconstruction of double perovskite under reducing atmosphere can spontaneously lead to the formation of ordered layered oxygen deficiency and yield segregation of massively and finely dispersed NPs. The real-time observation of this emergent process was performed via an environmental transmission electron microscope. Density functional theory calculations prove that the crystal reconstruction induces the loss of coordinated oxygen surrounding B-site cations, serving as the driving force for steering fast NP growth. The prepared material shows promising capability as an active and stable electrode for SOFCs in various fuels and SOECs for CO2 reduction. The conception exemplified here could conceivably be extended to fabricate a series of supported NPs perovskite catalysts with diverse functionalities. KEYWORDS: Crystal reconstruction, oxygen deficiency, double perovskite, in situ exsolution, in situ characterizations

T

alloy NPs, promoting electrochemical performance, sulfur tolerance, and coking resistance.11,12 The “pinned” unique feature of these NPs contribute to their exceptional stabilities.13 With the exception of the ABO3-type perovskite, the concept of exsolution was also extended to the layered K2NiF4-type oxides,14−16 which had been widely used as electrodes for SOFCs and SOECs with excellent performance. Nevertheless, to date, the detailed characterization data for exsolution phenomena shown are almost under ex situ condition. The evidence obtained by high temperature atomic force microscope (AFM) indicates that the formation of surface pits is the initial step prior to the emergence of NPs from the ABO3 structure.14 However, atomic-scale investigations of the exsolution mechanism, particularly for the double perovskite group, are scarce. In fact, for the whole perovskite family,

remendous effort has been devoted to the design and manufacture of highly active nanoparticle (NP)-supported catalysts for energy conversion and storage devices. Planting NPs has been identified as an effective approach simultaneously achieving catalytic activity and durability of state-of-the-art perovskite catalysts for applications in solid oxide fuel cells1 and electrolysis cells2 (SOFCs and SOECs). Unlike the conventional impregnation method, in situ exsolution delivers thermally stable and evenly dispersed NPs. LaFeO3 perovskite with exsolved Pd, Pt, and Rh has been successfully created and applied as the well-known automotive emission catalysts during the last few decades.3,4 The comprehensive studies of SOFCs also demonstrate that the exsolution of Ag,5 Ru,6 Ni,7,8 and Cu9 on perovskites could boost electrochemical activity. Particularly, the alternation of the stoichiometry of perovskite from A/B = 1 to A/B < 1 breaks the bottleneck of exsolution level and shed light on the facilitation of in situ segregation of metallic NPs with a larger population.10 For instance, A-site deficient lanthanum chromites could achieve exsolution of metallic and © XXXX American Chemical Society

Received: July 4, 2016

A

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. (a) Thermogravimetric analysis of PBMCo in 10% H2−N2. (b) HAADF image of the PBMCo-850-4h materials overlaid with the EDX elemental map of Pr (b-1), Ba (b-2), Co (b-3), and highlighted Co and Ba nanoparticles (b-4). (c) High-resolution transmission electron microscopy (HRTEM) images of fresh PBMCo (c-1), PBMCo-500 (c-2), and PBMCo-850-4h (c-3). The diffractograms and their simulations are shown as insets in the images. The regions where the diffractograms were obtained are marked by yellow squares.

information is still unavailable in terms of either in situ visual observations or theoretical calculations to determine the natural triggering force of exsolution. This letter aims to respond to this lack of information by a detailed investigation of the exsolution driving force for double perovskite. There is more than one possible evolution of the supported NP catalysts when switching from an operando condition to an ex situ condition, including reoxidation of metallic NPs or structure evolution. Therefore, the exsolution of NPs on double perovskites requires further investigation such as determination of the driving force, the initial exsolution temperature, and the accomplishment of exsolution. To clarify these influencing factors, in situ characterizations must be carried out. In this work, we found that in situ crystal reconstruction was confirmed as the triggering force steering the fast growth of finely dispersed Co NPs on Co-doped Pr 0.5 Ba 0.5 MnO x (PBMCo) double perovskite with exceptional performance in SOFC and SOEC models. The emergent crystal reconstruction under operando condition induced the formation of a layered structure and the creation of massive ordered oxygen vacancies, further decreasing the oxygen coordinating number surrounding Co and lowering the energy required to drag out the B-site cations. In addition, for the very first time, we utilized atomicscale in situ characterization technologies to directly visualize this exsolution process, which provided a deeper and more thorough understanding of this phenomena. Due to the fact that Mn2+ has a similar radius to Co2+, the Mn in B-site of PBMCo should be partially substituted by Co.

A previous study showed the crystal reconstruction of Pr0.5Ba0.5MnO3 (PBMO) combined with the loss of oxygen ions in a thermogravimetric analysis (TGA) test in reducing atmosphere.15 Similarly, the multistep weight loss of PBMCo associated with the high temperature reduction treatment was initially deduced by TGA shown in Figure 1a. Stage 1 of the weight loss completed below 800 °C, which was similar to the previous results of PBMO.15,16 From a structural standpoint, the formula of the material was changed to a layered PrBaMn1.8Co0.2O5 during this period. The A-site in one single unit cell was doubled, and B-sites were sandwiched between the PrOx and BaO layers with the formula: ([BaO]-[Mn/CoOx][PrOx]-[Mn/CoOx]-[BaO]). Simultaneously, oxygen in the PrOx plane was removed to create two-dimensional oxygen diffusion channels, which could be responsible for the initial weight loss of 5.5 wt %. After reaching 850 °C, the system was held at this temperature for another 260 min during which there were three distinct stages. The partial removal of the lattice oxygen strongly bonding with perovskite may occur in stage 2. Afterward, the weight of the sample decreased at a higher rate. Previous investigation on the exsolution of Co from La 0.3Sr0.7TiO 3−δ (LST) at high temperature had been reported.17 Moreover, by comparing the reducibility of each element, it can be speculated that the majority of the weight loss should be due to the loss of oxygen bonding with Co and reduction of ionic Co. In stage 4, any further increase of the B

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. (a−c) In situ TEM images of PBMCo with 0.5 Pa of H2 supplied from 770 to 850 °C with one image captured every 40 °C with a ramping rate of 20 °C min−1. (d−f) TEM images of PBMCo in 0.5 Pa of H2 supplied at 850 °C over a time period of 500 s.

reduction time did not lead to obvious weight losses, indicating the completion of reduction. So far no direct evidence of the exsolution from double perovskites has been shown. Therefore, more reliable evidence is required to verify the accuracy of our hypothesis. First, in situ X-ray diffraction (XRD) was used to detect the crystal reconstruction process during the thermal and reducing atmosphere treatment, following the protocol shown in Figure S-1(a). The assembly of in situ XRD can be found in Figure S1(b). As demonstrated by XRD patterns in Figure S-1(c,d), a shift of the (200) peak to a lower angle and a gradual crystal reconstruction from hexagonal to cubic were observed. More importantly, the formation of metallic cobalt located at 44° for temperatures higher than 700 °C was detected from the XRD patterns. The SEM image in Figure S-2(d,e) for the sample after the TGA test proved the existence of exsolved Co particles. A large amount of uniformly dispersed particles with average size of 20−30 nm could be found being pinned on the surface of the

perovskite. The high angle annular dark-field (HAADF) images displayed in Figure 1(b-1) to (b-3) and in Figure S-3 present the distribution of each element. The metallic Co composition of NPs, presented as bright orange spots (Figure 1(b-3)), is also determined. In comparison, only a few visible bright spots can be detected on PBMCo-850 [Figure S-2(c)], and no exsolved NPs can be seen on the fresh PBMCo and PBMCo500 [Figure S-2(a,b)]. In addition, it was also found that the 4 h TGA reduction treatment led to the annealing of bulk particles but still facilitated the increase of BET surface area from 9.48 to 15.37 m2 g−1, which may be due to the formation of defects in the lattice and extra metallic Co islands. High-resolution TEM (HRTEM) was employed to analyze the crystal evolution of the materials at the atomic-scale. Figure 1(c-1) shows a typical HRTEM image of a region in fresh PBMCo. The simulation of the diffractogram displays that selected area of PBMCo possesses a Pm3m structure [see the insets in Figure 1(c-1) and Figure S-4(a)]. When PBMCo was subjected to reduction under H2 atmosphere at 500 °C, the C

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. Schematic representation of double perovskites (a) without and (b) with crystal reconstruction, showing the ordering of each element, the formation of oxygen vacancies, and the exsolved metallic Co.

temperature further increased to 850 °C, large particles attached to the edge of the bulk were found. These particles with a spherical shape became clearer when the temperature reached the target temperature. These in situ images perfectly matched the results of in situ XRD. It can be inferred from the results of TGA, SEM, and in situ TEM images that the exsolution of Co mainly happened at a temperature of 800 °C and higher at a relatively high rate. To investigate the exsolution process at this temperature point, we also used environmental TEM to analyze a selected clean area. Here, gas was not introduced into the vacuum chamber until the temperature reached Ttargeted. Figure 2(d−f) illustrates a series of TEM images of PBMCo taken at 850 °C in 0.5 Pa of H2 supplied over a period of 500 s. At 0 s, no apparent exsolved particles can be found [Figure 2(d)]. Afterward, some particles with diameter of approximately 5 nm were detected by TEM after reduction for 250 s [Figure 2(e)]. As time went on, more particles with even distribution over a single bulk could be detected. Moreover, the Co NPs anchored on the edge of the bulk were observed. When the time reached 500 s, the welldispersed particles with average diameter of 15 nm became clearer [Figure 2(f)]. Meanwhile, some smaller particles with diameters less than 5 nm appeared, indicating the continuous exsolution of new particles. The phase of metallic Co was identified by selected area electron diffraction (SAED) (Figure S-7). The cubic Co pattern simulation matches the SAED pattern obtained. Before, in situ exsolution has been regarded as an effective way for preparing supported NPs perovskite catalysts. This process usually happens at high temperature, meaning that the interaction between NPs and perovskite should be strong enough to maintain thermal stability. However, limited exsolution population and slow exsolution rate have restricted its wider development. For example, more than 100 h was required for La0.8Sr0.2Cr0.85Ni0.15O3−δ to slowly generate a small number of Ni particles, and stability tests showed a decreased performance after 190 h in H2.6 The cobalt-doped LST prepared by our group was able to produce high-density precipitated Co particles within 10 h. However, its reduction temperature was as high as 1400 °C.17 The slow exsolution rate and critical reaction conditions were attributed to inactive lattice oxygen. Although the A-site deficiency in perovskite

basic cubic Pm3m structure remained, but an orthorhombic phase with an Imma structure was also found, as shown in Figure 1(c-2) and Figure S-4(b). There was an orientational relationship (OR) between the cubic PBMCo structure and the orthorhombic phase, e.g., [002]o∥[100]c, [100]o∥[1̅10]c (o denotes the orthorhombic phase, while c denotes cubic PBMCo). When the reduction temperature was further increased to 850 °C, extra reflections emerged in the diffractogram [see an inset in Figure 1(c-3) and Figure S4(c)]. Based on the ⟨110⟩ zone axis pattern of cubic PBMCo, these circled reflections in the diffractogram indicate that the space between the cation layers of the original PBMCo was doubled along the c direction, which in turn confirmed the layered sandwiched structure of the reduced sample. The TGA results above show that the crystal reconstruction of PBMCo (AxA′1−xByB′1−yO3 to A2xA′2−2xB2yB′2−2yO5) led to the formation of a large amount of oxygen vacancies in PrOx plane. As the oxygen atoms escaped, the electroneutrality of the materials should be maintained via the reduction of Mn4+ to Mn3+ (even to Mn2+) and of Co3+ to Co2+ (even to Co0). X-ray photoelectron spectroscopy (XPS) was used to investigate the surface composition of the samples treated under different conditions (Figures S-5 and S-6). Different reduction treatments exhibited limited effect on the valence of Mn, and the ratios of Mn3+ to Mn2+ were roughly equal to 1 for PBMCo500, PBMCo-850, and PBMCo-850-4h. On the contrary, either a higher reduction temperature or a longer reduction time obviously favored the formation of metallic Co. The percentage of Co0 on PBMCo-850-4h occupied is higher than 50%, which is consistent with our inference of surface exsolution. Considering that XRD and XPS can only provide indirect evidence to confirm the exsolution of Co from the perovskite scaffold, the results would be more reliable if this process could be visually observed under in situ conditions. Therefore, environmental TEM was applied to observe the exsolution behavior in real time. Figures 2(a−c) display a series of TEM images of PBMCo in 0.5 Pa of H2 supplied from 770 to 850 °C with one image captured every 40 °C. Specifically, no obvious exsolution could be detected for the sample at 770 °C [Figure 2(a)]. When the temperature risen to 810 °C, several bright spots appeared [Figure 2(b)], suggesting the occurrence of exsolution. As the D

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 4. Electrochemical performances of PBMCo anode in SOFC and SOEC models: (a) current−voltage (I−V) curves and the corresponding power densities of the PBMCo anode in C2H6 and syngas fuels. (b) Temperature-dependent I−V curves for CO2 electrolysis with a feed composition of V(CO2)/V(CO) = 7/3 at increasing temperatures from 800 to 900 °C at the ramping rate of 2 °C/min.

could lower the reduction temperature and facilitate the occurrence of exsolution by forcing the perovskite toward a stoichiometric structure for La0.52Sr0.28Ni0.06Ti0.94O3−δ,10 it was still time-consuming to have a relative full reduction of Ni cations. In our case, the in situ observation and TGA analysis have illustrated the fast exsolution of PBMCo after crystal reconstruction. As the atomic-scale structure model in Figure 3 shows, the theoretical oxygen coordination number of Co for a perfect stoichiometric perovskite is 6. Ideally, after the removal of oxygen on [PrOx] plane, the coordination number of a layered perovskite (AA′BB′O5) decreases to 5. According to a pioneering study, the migration and diffusion of B-site cations were particularly favorable on the (110) plane of perovskite.18 Consequently, in our Co-doped Pr0.5Ba0.5MnO3 system, the removal of oxygen ions on the [PrOx] plane should also facilitate the exsolution process of Co. However, the decrease of bonding energy stabilizing Co should also contribute to the exsolution of metallic Co. To verify our prediction, the energy consumed for dragging out a Co atom was calculated based on the density functional theory (DFT) using the Vienna ab initio simulations package (VASP) code. The schematic representations of the simulation of the exsolution processes are shown in Figure 3(a,b). The Mn (3p, 3d, 4s), O (2s, 2p), Co (3d, 4s), Ba (5s, 5p, 6s), and Pr (5s, 5p, 4f, 6s) electrons were doped as valence states, with the remaining electrons being frozen as core states. To investigate the formation energy of Co vacancy defects in PrBaMn2−xCoxO 5 (after crystal reconstruction) and in PrBaMn2−xCoxO6 (before crystal reconstruction), we applied a 2 × 2 × 1 supercell with a = b = 7.96 Å and c = 7.78 Å (or a = b = 7.88 Å and c = 7.65 Å) as the initial structure (the unit parameters were obtained from XRD patterns after Rietveld refinement). Brillouin zone integrations were performed using the Monkhorst−Pack grid scheme, and the 5 × 5 × 5 k point meshes were used with the Γ point included. To calculate the defect formation, we used the following formulations:

Ed ‐ O6 = E(PrBaMn2 − x□O5) + E(Co) − E(PrBaMn2 − xCox O6 )

where E(PrBaMn1−xCoxO5) and E(PrBaMn1−xCoxO6) are the energies of PrBaMn2‑xCoxO5 and PrBaMn2‑xCoxO6, and □ labels the Co vacancy site. Note that E(Co) as a reference is arbitrary. Here, we adopted the energy of face-central cubic Co metal unit cell as E(Co). It can be seen from the calculated defect formation energies that the formation energy of PrBaMn1−x□O5 (AA′BB′O5) with the appearance of one Co atom was calculated to be −3.62 eV. In comparison, that of PrBaMn2−xCoxO6 (AA′BB′O6) was estimated to be −4.82 eV. Consequently, it can be concluded that the formation of a layered structure is favorable for the exsolution of B-site Co. In comparison, the exsolution of B-site cations from perovskite without crystal reconstruction is relatively kinetically unfavorable, owing to its full coordination properties and absence of consecutive diffusion paths. Moreover, Figure S-8 shows the relation between oxygen nonstoichiometry and different treatments on the material determined by TGA analyses. PBMCo had the stoichiometry of Pr0.5Ba0.5Mn0.9Co0.1O2.89 and PrBaMn1.8Co0.2O4.99 before and after TGA analysis, indicating that it had higher levels of vacancies than PBMO without the Co dopant. The results are also relatively consistent with the XPS composition analysis in Figures S-4 and S-5. These oxygen vacancies generated in a reducing atmosphere resulted in an enhanced ionic conductivity. As shown by Figure S-9, the PBMCo perovskite structure retained high electrical conductivities of 125.25 and 2.73 S cm−1 in air and in 10% H2−N2 at 800 °C, respectively, which is mainly due to the created fast two-dimensional diffusion channels for the transportation of other lattice oxygen ions. To characterize the performance of this type of functional catalyst as anode, the performance of a single SOFC was tested in various fuels. The I−V curves of the SOFC in H2 at various temperatures are presented in Figure S-10 after ohmic compensation. The maximum power densities of the cell were as high as 520, 830, and 1100 mW cm−2 at 800, 850, and 900 °C, respectively. The cell also exhibited a high catalytic activity toward the oxidation of syngas and ethane. As shown in

Ed ‐ O5 = E(PrBaMn2 − x□O5) + E(Co) − E(PrBaMn2 − xCox O5) E

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

and B.H. synthesized the sample and did the structural and chemical analyses. Y.F.S. and B.H. performed the electrochemical measurements. Y.F.S. and J.C. performed the environmental TEM characterizations. Y.T.Z., Y.F.S., and J.H.L. did the DFT calculations. J.H.L. and Y.F.S. analyzed the data with assistance from J.L. and B.H. All authors discussed the results and commented on the manuscript. J.L.L., B.H., and J.H.L. coordinated and supervised the overall project.

Figure 4a, the maximum power density of the cell fed with syngas reached 900 mW cm−2 at 900 °C. The exsolved Co particles had high thermal stability and carbon deposition resistance. The image shown in Figure S-11 indicates that around 50 vol % of the exsolved metallic particles was submerged into the oxide parent, suggesting its “pinned” situation. Thus, the interdiffusion between these two phases should be so significant that the adhesion can be sharply enhanced.19 In comparison, the thermal stability of deposited Co particles prepared by impregnation reveals the obvious agglomeration of particles over a short-time in situ treatment possibly because of weak metal−support interaction (seen SEM-EDX results in Figure S-12). Previous report on Ni vapordeposited La0.4Sr0.4TiO3−δ also showed similar results with a much quicker increase in particle size over time compared to Ni exsolved La0.4Sr0.4TiO3−δ.13 Due to the fact that PBMCo shows promising performance for SOFC, the application of PBMCo as a potential cathode in CO2 electrolysis is worth further investigation. Figure 4(b) shows the temperature-dependent I−V curve for CO 2 electrolysis between 800 and 900 °C under the gas conditions of 70 vol % CO2/30 vol % CO to the cathode and ambient air to the anode. It was found that at 1.5 V, the PBMCo was more effective for CO2 electrolysis with high cathodic current densities of 2.5, 1.6, and 0.85 A cm−2 at 900, 850, and 800 °C, respectively. The attractive electrochemical performance can be attributed to fast oxygen ion mobility20,21 and high adsorption amount of CO2 caused by the large amount of oxygen vacancies.22−26 Furthermore, the existence of stable metallic Co NPs with a uniform distribution facilitated the transportation of electrons in our reaction, thus promoting catalytic activity kinetically. In summary, the crystal reconstruction of PBMCo to a layered double perovskite structure induced phase reconstruction leading to the formation of an oxygen deficient layer and numerous in situ exsolved Co NPs. This fast exsolution process was, for the first time, observed and monitored by environmental TEM. The DFT results indicate a much lower energy barrier for dragging out Co from the layered perovskite structure. The fabricated material demonstrates outstanding performance in both SOFC and SOEC modes. Thus, this preparation method represents a brand-new opportunity to further design NP supported functional perovskite catalysts in a reasonable and convenient way.

■

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was supported by the Natural Sciences and Engineering Research Council (NSERC) of the Canada Strategic Project Grant, China Scholarship Council (CSC) and the National Natural Science Foundation of China under grant 21303141.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b02757. Additional information, experimental section, and figures (PDF)

■

REFERENCES

(1) Atkinson, A.; Barnett, S.; Gorte, R. J.; Irvine, J. T. S.; McEvoy, A. J.; Mogensen, M.; Singhal, S. C.; Vohs, J. Nat. Mater. 2004, 3, 17−27. (2) Yang, C.; Yang, Z.; Jin, C.; Liu, M.; Chen, F. Int. J. Hydrogen Energy 2013, 38, 11202−11208. (3) Nishihata, Y.; Mizuki, J.; Akao, T.; Tanaka, H.; Uenishi, M.; Kimura, M.; Okamoto, T.; Hamada, N. Nature 2002, 418, 164−167. (4) Tanaka, H.; Taniguchi, M.; Uenishi, M.; Kajita, N.; Tan, I.; Nishihata, Y.; Mizuki, J. i.; Narita, K.; Kimura, M.; Kaneko, K. Angew. Chem. 2006, 118, 6144−6148. (5) Zhu, Y.; Zhou, W.; Ran, R.; Chen, Y.; Shao, Z.; Liu, M. Nano Lett. 2016, 16, 512−518. (6) Kobsiriphat, W.; Madsen, B. D.; Wang, Y.; Shah, M.; Marks, L. D.; Barnett, S. A. J. Electrochem. Soc. 2010, 157, B279−B284. (7) Park, B. H.; Choi, G. M. Solid State Ionics 2014, 262, 345−348. (8) Arrivé, C.; Delahaye, T.; Joubert, O.; Gauthier, G. J. Power Sources 2013, 223, 341−348. (9) Adijanto, L.; Balaji Padmanabhan, V.; Kungas, R.; Gorte, R. J.; Vohs, J. M. J. Mater. Chem. 2012, 22, 11396−11402. (10) Neagu, D.; Tsekouras, G.; Miller, D. N.; Ménard, H.; Irvine, J. T. S. Nat. Chem. 2013, 5, 916−923. (11) Sun, Y.-F.; Li, J.-H.; Cui, L.; Hua, B.; Cui, S.-H.; Li, J.; Luo, J.-L. Nanoscale 2015, 7, 11173−11181. (12) Sun, Y.-F.; Li, J.-H.; Wang, M.-N.; Hua, B.; Li, J.; Luo, J.-L. J. Mater. Chem. A 2015, 3, 14625−14630. (13) Neagu, D.; Oh, T.-S.; Miller, D. N.; Menard, H.; Bukhari, S. M.; Gamble, S. R.; Gorte, R. J.; Vohs, J. M.; Irvine, J. T. S. Nat. Commun. 2015, 6, 8120−8127. (14) Oh, T.-S.; Rahani, E. K.; Neagu, D.; Irvine, J. T. S.; Shenoy, V. B.; Gorte, R. J.; Vohs, J. M. J. Phys. Chem. Lett. 2015, 6, 5106−5110. (15) Sengodan, S.; Choi, S.; Jun, A.; Shin, T. H.; Ju, Y.-W.; Jeong, H. Y.; Shin, J.; Irvine, J. T. S.; Kim, G. Nat. Mater. 2015, 14, 205−209. (16) Trukhanov, S. V.; Lobanovski, L. S.; Bushinsky, M. V.; Fedotova, V. V.; Troyanchuk, I. O.; Trukhanov, A. V.; Ryzhov, V. A.; Szymczak, H.; Szymczak, R.; Baran, M. J. Phys.: Condens. Matter 2005, 17, 6495. (17) Cui, S.-H.; Li, J.-H.; Zhou, X.-W.; Wang, G.-Y.; Luo, J.-L.; Chuang, K. T.; Bai, Y.; Qiao, L.-J. J. Mater. Chem. A 2013, 1, 9689− 9696. (18) De Souza, R. A.; Islam, M. S.; Ivers-Tiffee, E. J. Mater. Chem. 1999, 9, 1621−1627. (19) Chambers, S. A.; Gu, M.; Sushko, P. V.; Yang, H.; Wang, C.; Browning, N. D. Adv. Mater. 2013, 25, 4001−4005. (20) Bi, L.; Boulfrad, S.; Traversa, E. Chem. Soc. Rev. 2014, 43, 8255− 8270. (21) Anderson, M. T.; Vaughey, J. T.; Poeppelmeier, K. R. Chem. Mater. 1993, 5, 151−165. (22) Zhang, J.; Xie, K.; Wei, H.; Qin, Q.; Qi, W.; Yang, L.; Ruan, C.; Wu, Y. Sci. Rep. 2014, 4, 7082.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (J.H.L.). *E-mail: [email protected] (B.H.). *E-mail: [email protected] (J.L.L.). Author Contributions

J.L.L., B.H., and J.H.L. conceived and designed the experiments. Y.F.S. performed the experiments, analyzed the data, and wrote the manuscript with help from all the authors. Y.F.S., Y.Q.Z., F

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (23) Li, H.; Sun, G.; Xie, K.; Qi, W.; Qin, Q.; Wei, H.; Chen, S.; Wang, Y.; Zhang, Y.; Wu, Y. Int. J. Hydrogen Energy 2014, 39, 20888− 20897. (24) Qi, W.; Gan, Y.; Yin, D.; Li, Z.; Wu, G.; Xie, K.; Wu, Y. J. Mater. Chem. A 2014, 2, 6904−6915. (25) Li, S.; Li, Y.; Gan, Y.; Xie, K.; Meng, G. J. Power Sources 2012, 218, 244−249. (26) Ebbesen, S. D.; Mogensen, M. J. Power Sources 2009, 193, 349− 358.

G

DOI: 10.1021/acs.nanolett.6b02757 Nano Lett. XXXX, XXX, XXX−XXX