English

• Platinum (Pt) is the most widely used electrode catalyst in eco-friendly energy devices, especially the oxygen reduction reaction (ORR) catalysts in the cathode of fuel cells [1,2]. To control the overall cost of the devices, optimizing the utilization of the precious Pt metal in the electrode is highly important [3]. High-index facets (HIFs) contain abundant low-coordinated noble metal atoms which significantly reduce the oxygen adsorption, and then enhance the ORR activity [4–6]. Thus, it is reasonable to create HIFs on the surface Pt and Pt-transitional metal alloy (Pt₃M) nanocrystals to boost their ORR performance with low Pt usage. However, the size of HIFs-contained Pt and Pt₃M nanocrystals can be relatively large due to the high surface energy of those active facets, which may decrease their ORR activity because of low surface area [7–9].

  Generally, the methods for creating HIFs can be concluded to wet-chemical method [10–12], electrochemical method [13–16], and high-temperature alloying/dealloying method [17–19]. Among them, wet-chemical method is the most widely used route because of its simplicity. To tackle the issue of the overgrowth of the HIFs-contained nanocrystals synthesized by traditional wet-chemical method, new approaches have been developed such as selectively etching surface atoms and capping specific facets by choosing appropriate surfactant [20,21]. Several studies have verified that halide ions/oxygen (e.g., Cl⁻/O₂ and Br⁻/O₂) and metal ions (e.g., Cu⁺/Cu²⁺ and Fe²⁺/Fe³⁺) are proper candidates for the corrosion of surface atoms and forming HIFs [22–24]. Meanwhile, halide ions in etching agents as well as surfactants also play an important role on creating HIFs since they can be selectively adsorbed onto the {100}, {110} and {111} facets of Pt or Pd nanocrystals [25]. Those halide ions directly hinder the deposition of newly-reduced metal atoms from precursors, thereby controlling the growth rate and mode of crystal surface and tuning its morphology [26,27]. In our previous work, we have developed a facile chemical etching route to successfully create {311} facets on Pt₃Ni nanocrystals within several nanometers by introducing hydrochloric acid (HCl) into the reaction [28]. The primary results reveal that the etching effect of Cl⁻/O₂ to the metal nanocrystals and the capping effects of Cl⁻ on specific facets (such as [111]) together determine the formation of those HIFs [29,30]. Nonetheless, a deep insight into the formation mechanism of this special structure needs to be further studied.

  Besides Pt₃Ni, other fcc-type Pt₃M nanocrystals are also widely used on advanced electrocatalysts. For example, Pt₃Co is considered as the most stable and active ORR catalysts in proton-exchange membrane fuel cells [31,32]. Inspired by the successful applications of HCl chemical-etching to create HIFs on Pt₃Ni nanocrystals, it is worth investigating the effectiveness of this method towards other Pt₃M alloys. Nevertheless, the etching rate of surface atoms by Cl⁻/O_2, and the adsorption of Cl⁻ on the surface can be diverse towards different metal nanocrystals, therefore producing unique morphology and composition. Based on our previous study about HCl-etched Pt₃Ni nanocrystals (HCl-Pt-Ni NCs), we further expanded this system to Pt (HCl-Pt NCs) and Pt₃Co (HCl-Pt-Co NCs) following a similar wet chemical method (Fig. 1a). Theoretical calculation, combining with a series of elaborate characterizations, was employed to explore the different structure and composition evolution process of these nanocrystals under HCl etching. Owing to the highly active sites and the proper amount of transitional metal, HCl-Pt-Ni NCs exhibit the highest ORR activity among all the samples. Meanwhile, HCl-Pt-Ni NCs also possess extraordinary ORR durability.

  Figure 1

  

  Figure 1.  (a) Schematic illustration of the preparation process of samples. Structural characterizations of HCl-Pt and HCl-Pt-M NCs. (b) TEM and (c) HAADF-STEM image of HCl-Pt NCs. (d) TEM and (e) HAADF-STEM images of HCl-Pt-Ni NCs. (f) TEM and (g) HRTEM images of HCl-Pt-Co NCs. The inset of (e) shows the formation of unsaturated atoms on HCl-Pt-Ni NCs. EDS-mapping profile of (h) HCl-Pt-Co and (i) HCl-Pt-Ni NCs.

  DownLoad: Full-Size Img PowerPoint

  Typically, the HCl-Pt and HCl-Pt₃M nanocrystals were obtained by heating the N,N-dimethylformamide (DMF) and metal precursor solution containing polyvinylpyrrolidone (PVP) and HCl at 160 ℃ for 6 h. The representative transmission electron microscopy (TEM) images of these samples (Figs. 1b, d and f) illustrate the uniform formation of nanocrystal structure. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of HCl-Pt NCs (Fig. 1c) shows that a complete octahedron shape is almost well preserved. However, the octahedron morphology of HCl-Pt-Ni NCs is somewhat irregular (Fig. 1e). Meanwhile, various unsaturated atoms (edges, steps and corners) can be observed on the surface of HCl-Pt-Ni NCs (inset of 1e), demonstrating the successful formation of HIFs. As for HCl-Pt-Co NCs, the high-resolution TEM (HRTEM) image in Fig. 1g reveals that the shape of most of the nanocrystals is close to oval-shaped nanorods rather than nanoparticles, and no obvious unsaturated atoms can be found. Obviously, the interaction between HCl and different Pt-based nanocrystals can be dissimilar, resulting in the distinctive morphology among each sample. The average sizes of HCl-Pt NCs and HCl-Pt-Ni NCs were measured as 2.74 and 5.77 nm, respectively (Fig. S1 in Supporting information). The enlargement of the size of HCl-Pt-Ni NCs can be attributed to the increase of surface energy brought by HIFs. For HCl-Pt-Co NCs, their sizes were ~7.69 nm in length and ~3.68 nm in diameter (Fig. S2 in Supporting information). The elemental mapping profiles (Figs. 1h and i) display that both Pt and the transitional metal can be uniformly distributed throughout HCl-Pt-Ni NCs and HCl-Pt-Co NCs. Besides, the energy-dispersive spectrometer (EDS, Fig. S3 in Supporting information) profile indicates that the composition of HCl-Pt-Ni NCs is Pt/Ni = 98.5/1.5, while that of HCl-Pt-Co NCs is Pt/Co = 99.3/0.7, which is close to the results of ICP-OES measurement (Pt/Ni = 98.5/1.5 for HCl-Pt-Ni, and Pt/Co = 99.4/0.6 for HCl-Pt-Co). In order to highlight the effect of HCl, the Pt, Pt₃Ni and Pt₃Co nanocrystals were also synthesized without adding HCl into DMF. Fig. S4 (Supporting information) depicts that all of those nanocrystals have a homogeneous octahedron-like structure, with an average edge length of 6.76 nm for Pt NCs, 4.97 nm for Pt₃Ni NCs, and 4.95 nm for Pt₃Co NCs (Fig. S5 in Supporting information). The EDS profile in Fig. S6 (Supporting information) further suggests that the composition of Pt₃Ni and Pt₃Co NCs is Pt/Ni = 76.3/23.7 and Pt/Co = 77.2/22.8, respectively, nearly consistent with ICP-OES data (Pt/Ni = 76.1/23.9 for Pt₃Ni NCs, and Pt/Co = 77.0/23.0 for Pt₃Co NCs). Compared with their HCl-etched counterparts, the transitional element contents are apparently decreased due to the leaching effect of acid.

  The phase structure of the aforementioned samples was studied by X-ray diffraction (XRD) technique. As shown in Fig. 2a, all of the HCl-etched nanocrystals possess three main diffraction peaks corresponding to highly crystalline fcc structure. Compared with the standard pure Pt pattern (JCPDS No. 04–0802), the right shift of the diffraction peaks of HCl-Pt-Co and HCl-Pt-Ni indicates the incorporation of transitional metals into the lattice (Fig. S7 in Supporting information). XRD patterns of Pt, Pt₃Ni and Pt₃Co NCs also depict the successful formation of fcc-phase nanostructure (Fig. S8a in Supporting information). The leaching of transitional metals in HCl-Pt-Ni and HCl-Pt-Co NCs can be confirmed by the left-shift of the main diffraction peaks compared with their pure structure (Fig. S8b in Supporting information). In addition, such left-shift effect may suggest the construction of a Pt-rich surface on the nanocrystals [33,34]. To investigate the possible change of electronic state, the X-ray photoelectron spectroscopy data were recorded. The Pt 4f spectra of both pure (Fig. S9 in Supporting information) and HCl-etched (Fig. 2b) nanocrystals reveal the dominant presence of zerovalent Pt in these samples. The shift of main Pt peaks to the low electron energy range discovered from Pt₃Ni NCs, Pt₃Co NCs, HCl-Pt-Ni NCs, and HCl-Pt-Co NCs is due to the strong interaction between Pt and other transitional metals. On the other hand, the Co 2p XPS spectra (Fig. 2c) of Pt₃Co NCs mainly consist of oxidized Co^δ+ peak resulted from strong Co-Pt or Co-O interaction [35–37]. The signal of Co 2p can be hardly detected in HCl-Pt-Co NCs possibly because most of Co on the surface is leached by HCl. In Fig. 2d, the Ni 2p spectra of Pt₃Ni contain both metallic Ni⁰ and oxidized Ni^δ+ signs, while the signal of Ni⁰ is dramatically suppressed in HCl-Pt-Ni NCs due to the dealloying of Ni [38].

  Figure 2

  

  Figure 2.  Structural characterizations of the nanocrystals. (a) XRD patterns and (b) Pt 4f XPS spectra of HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs. (c) Co 2p spectra of Pt₃Co NCs and HCl-Pt-Co NCs. (d) Ni 2p XPS spectra of Pt₃Ni NCs and HCl-Pt-Ni NCs.

  DownLoad: Full-Size Img PowerPoint

  To understand the mechanism of the morphology and composition evolution, we have performed density functional theory (DFT) calculation using Pt(111) and various Pt₃M(111) surfaces. Based on the structural characterizations, Pt₃M(111) surfaces with a layer of Pt-skin (Pt₃M_Ptskin) were also constructed (for details, see DFT calculations part in Supporting information). The adsorption of chlorine was firstly investigated to understand its hindering effect on the formation of HIFs. It is widely known that on clean Pt surface, four possible adsorption sites are available for heteroatoms, namely top site, bridge site, fcc site and hcp site. Nevertheless, Pt₃M alloys, along with their segregated structures, can provide more possible nonequivalent adsorption sites due to their different local atom arrangement [39]. Totally, 9 nonequivalent sites (two top sites, three bridge sites, two fcc sites and two hcp sites) can be provided by Pt₃M surface, while Pt₃M_Ptskin surface contains 14 distinct adsorption sites (three top sites, five bridge sites, three fcc sites and three hcp sites), as illustrated in Fig. S10 (Supporting information). The adsorption energy of chloride ion (ΔE_Cl) on those possible sites are calculated, and the results are summarized in Tables S1 and S2 (Supporting information). The theoretical models of the most stable adsorption structure on different surfaces are illustrated in Fig. 3a. For Pt(111) surface, the most favorable adsorption site is fcc. In the case of non-segregated Pt₃Ni, the Cl⁻ tends to be adsorbed on bridge-1 site, i.e., right above the Pt-Ni bond on the surface. Similar results are discovered in the Pt₃Ni(111) surface with Pt-skin (bridge-5 site). As for non-segregated Pt₃Co, the Cl⁻ is favored to be tightly attracted by the hcp-Pt₂M site, i.e., coordinated with two Pt atoms and one transitional metal atom, while chlorine adsorbs preferentially at fcc-1 site on Pt₃Co_Ptskin(111), in which Cl⁻ is coordinated with three Pt atoms. Among different surfaces, Pt₃Ni(111) shows the highest ability to adsorb chlorine on its surface. It should be noted that the surface with ideal stoichiometric proportion tends to bind Cl⁻ stronger than skin-type surfaces, which is in agreement with previous reports [39,40]. The ΔE_Cl variation among different sites on the surface is also compared (Fig. 3b). Significant differences can be found among adsorption sites on Pt(111) and Pt₃Ni(111), along with skin-type surface, indicating that the mobility of chlorine is hindered on them. Compared with Pt(111), Pt₃Ni(111) has a wider ΔE_Cl distribution, which suggests that Cl⁻ can be attached more tightly. On the other hand, the adsorption energy of chlorine on Pt₃Co is rather smooth, especially for the non-segregated Pt₃Co(111) whose ΔE_Cl is almost the same among different sites. The results demonstrate that chlorine has a good mobility on Pt₃Co surface. In addition, the analysis of projected density of states (PDOS) further reveals the interaction between chlorine and metal substrates. Specifically, the p-states electrons in Cl⁻ can almost fully overlap with the d-states electrons in Pt(111) and Pt₃Ni(111), especially between −7.5 eV and −3.8 eV (relevant to the Fermi level, E_f), indicating that chlorine can be strongly bound by these two metal nanocrystals (Fig. 3c). On the other hand, the overlapped area of Cl-p states electrons with the d-states electrons of Pt₃Co(111) is relatively small, suggesting a weak interaction between them. For the case of metal nanocrystals with a Pt-skin (Fig. 3d), most of the Cl-p states electrons can interact with metal-d states electrons on Pt₃Ni(111)_Ptskin, while only partial Cl-p states electrons have a similar behavior on Pt₃Co(111)_Ptskin. Based on the calculation results, we can conclude that Pt₃Ni(111) has the strongest ability to adsorb chlorine, and the chlorine is hardly to move on its surface, thereby producing the largest obstruction for the deposition of new atoms. The smallest average Pt-Cl bond length of Pt₃Ni(111)-Cl measured from the stable adsorption models (Table S3 in Supporting information) also reflects this overwhelming adsorption effect.

  Figure 3

  

  Figure 3.  DFT calculations. (a) Theoretical structures of the Cl⁻-adsorbed nanocrystal surface. (b) The range of Cl⁻ adsorption energies on different nanocrystal surfaces. PDOS analysis for nanocrystals (c) without and (d) with a Pt-skin on the surface. (e) Vacancy formation energies (Pt, Ni and Co) for Pt NCs, Pt₃Ni NCs, and Pt₃Co NCs.

  DownLoad: Full-Size Img PowerPoint

  Another key point for the construction of HIFs is the etching of surface atoms. In addition, it also influences the final composition of nanocrystals. Therefore, we have calculated the vacancy formation energy (ΔE_V) of surface Pt and M atoms on different structures. As summarized in Fig. 3e, the vacancy formation energy of Pt (ΔE_V-Pt) is obviously higher than Ni and Co atoms, suggesting that the transitional metal is mainly etched rather than Pt when HCl is added. Besides, the Co vacancy formation energy on Pt₃Co(111) is extremely lower than that of Ni on Pt₃Ni(111). This result can explain the low transitional metal content of HCl-Pt-Co compared with HCl-Pt-Ni.

  From the above analysis, we can deduce the difference of the structure and composition evolution in Pt, Pt₃Ni and Pt₃Co nanocrystals under HCl etching: (1) Pt₃Ni has the strongest chlorine ion adsorption on its surface, which prevent the deposition and the reduction of newly-generated metal atoms, thus regulating the morphology of surface. Besides, Ni is effectively etched by HCl, which further facilitates the growth of HIFs. (2) Although Cl⁻ can also be attached by Pt(111) surface, the high ΔE_V-Pt hinders the remove of surface atoms. Therefore, it is difficult to effectively create HIFs on pure Pt surface by HCl etching. (3) The existence of HCl can effectively remove Co atoms in Pt₃Co, but the chlorine ions are relatively mobile on the surface. In this case, the hindering effect of new metal atoms deposition cannot be well developed. Hence, the construction of HIFs is also hard for HCl-Pt-Co NCs.

  The promotion of the electrocatalytic performance by HIFs was further verified by a series of electrochemical ORR tests. Before the electrochemical test, the as-prepared nanocatalysts were homogeneously deposited onto Vulcan XC-72 carbon and pre-activated. Their electrocatalytic performance was compared with commercial Pt/C catalysts. The loading amount of Pt was 2.6, 2.5, 2.7, and 4.0 µg for HCl-Pt-Ni NCs, HCl-Pt-Co NCs, HCl-Pt NCs, and commercial Pt/C catalysts, respectively (based on ICP-OES). As shown in the cyclic voltammetry (CV) curves (Fig. 4a), the electrochemical active surface area (ECSA) of Pt/C, HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs is 60.0, 61.6, 59.5 and 63.8 m²/g_Pt. Compared with pure nanocrystals (Fig. S11a and Table S4 in Supporting information), the ECSA value of HCl-etched samples is improved. To further evaluate the ORR performance, the linear sweep voltammetry (LSV) ORR polarization curves were recorded in Fig. 4b. It can be seen that the half-wave potential (E_1/2) of different catalysts follows a trend as: HCl-Pt-Ni (0.91 V) > HCl-Pt-Co (0.87 V) > HCl-Pt (0.87 V) > Pt/C (0.84 V). The kinetic currents of the catalysts were calculated based on the Koutecky-Levich equation. Then, the kinetic currents were further normalized with ECSAs and the loading amount of Pt to obtain the mass activity and specific activity of each catalyst. As illustrated in Fig. 4c, HCl-Pt-Ni NCs exhibits a high mass activity of 1.95 A/mg_Pt at 0.9 V, which is significantly higher than that of Pt/C (0.18 A/mg_Pt), HCl-Pt NCs (0.43 A/mg_Pt) and HCl-Pt-Co NCs (0.49 A/mg_Pt). As for the specific activity (at 0.9 V, Fig. 4d), HCl-Pt-Ni NCs also shows a better enhancement (3.07 mA/cm_pt²) than Pt/C (0.27 mA/cm_pt²), HCl-Pt NCs (0.71 mA/cm_pt²) and HCl-Pt-Co NCs (0.82 mA/cm_pt²). The ORR activities of pure Pt and Pt₃M NCs were also investigated (Figs. S11b-d and Table S4 in Supporting information). The mass and specific activities of HCl-Pt-Ni NCs are significantly improved compared with pure Pt₃Ni NCs, while HCl-Pt NCs and HCl-Pt-Co NCs exhibit no apparent increase in ORR activity. The remarkable ORR activity of HCl-Pt-Ni NCs is resulted from the HIFs and the unique electronic structure [28]. Meanwhile, the ORR durability of HCl-Pt-Ni were also studied by employing an accelerated durability test (ADT) at room temperature. Fig. 4e demonstrates that only an 8 mV negative shifts can be found from the LSV curves of HCl-Pt-Ni NCs, while that for Pt/C, HCl-Pt and HCl-Pt-Co is 90 mV, 30 mV and 29 mV, respectively (Fig. S12 in Supporting information). The mass activity of HCl-Pt-Ni NCs after ADTs is 1.62 A/mg_Pt (Fig. 4f), with a 16.9% loss compared with its original value. However, its durability still exceeds Pt/C, HCl-Pt NCs and HCl-Pt-Co NCs, with a mass activity loss of 44.7%, 32.5% and 38.1% after ADT. Since ΔE_V-Pt is an effective descriptor of the durability of Pt-based nanocrystals, the excellent stability of HCl-Pt-Ni NCs can be reflected by the high ΔE_V-Pt of Pt₃Ni, especially when the Pt skin is formed on the surface (Fig. 3e) [41].

  Figure 4

  

  Figure 4.  Electrocatalytic ORR performance. (a) CV curves, (b) LSV curves, (c) mass activity (@0.9 V), and (d) specific activity (@0.9 V) of Pt/C, HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs. (e) LSV curve evolution of HCl-Pt-Ni after ADTs of 10,000 cycles. (f) mass activity (@0.9 V) variations of Pt/C, HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs after ADTs of 10,000 cycles.

  DownLoad: Full-Size Img PowerPoint

  In conclusion, we have demonstrated the effect of HCl etching on different Pt and Pt₃M nanocrystals. Under HCl chemical etching, only Pt₃Ni NCs can successfully form surface steps (HIFs) compared with Pt and Pt₃Co NCs. The difference of structure and composition evolution among those nanocrystals can be concluded to the variation of Cl⁻ adsorption energy as well as the etching rate of surface atoms. Owing to its highly active sites, HCl-Pt-Ni NCs exhibits superior ORR activity with a ~10.8 and ~11.3 times higher mass and specific activities than commercial Pt/C catalyst. Besides, HCl-Pt-Ni also possesses good ORR durability due to the high Pt vacancy formation energy. This study provides a general guidance for the structural and compositional design of high-performance ORR catalysts.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was financially supported by Nature Science Foundation of Hunan Province (No. 51502343) and the National Key R & D Program of China (No. 2018YFB1900603).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.109188.

  
  
• 1. [1]

     G. Fisseha, Y. Hu, Y. Yu, et al., Chin. Chem. Lett. (2023) 108445.

  2. [2]

     L. Gao, T. Sun, X. Tan, et al., Appl. Catal. B 303 (2022) 120918. doi: 10.1016/j.apcatb.2021.120918

  3. [3]

     L. Gao, X. Li, Z. Yao, et al., J. Am. Chem. Soc. 141 (2019) 18083–18090. doi: 10.1021/jacs.9b07238

  4. [4]

     C. Li, N. Clament Sagaya Selvam, J. Fang, Nano-Micro Lett. 15 (2023) 83. doi: 10.2307/jj.2430613.11

  5. [5]

     L. Tan, N. Yang, X. Huang, et al., Chem. Commun. 55 (2019) 14482–14485. doi: 10.1039/c9cc06132k

  6. [6]

     W. Long, J. Wang, F. Xu, et al., Chin. Chem. Lett. 31 (2020) 269–274. doi: 10.1016/j.cclet.2019.03.044

  7. [7]

     L. Bu, J. Ding, S. Guo, et al., Adv. Mater. 27 (2015) 7204–7212. doi: 10.1002/adma.201502725

  8. [8]

     C. Xiao, N. Tian, W. Li, et al., CrystEngComm 23 (2021) 6655–6660. doi: 10.1039/d1ce00949d

  9. [9]

     C. Xiao, B. Lu, P. Xue, et al., Joule 4 (2020) 2562–2598. doi: 10.1016/j.joule.2020.10.002

  10. [10]

      Z. Wang, B. Zhang, S. Liu, et al., Adv. Funct. Mater. 32 (2022) 2200893. doi: 10.1002/adfm.202200893

  11. [11]

      B. Jiang, D. Tian, Y. Qiu, et al., Nano-Micro Lett. 14 (2021) 40.

  12. [12]

      M. Liu, B. Lu, G. Yang, et al., Adv. Sci. 9 (2022) 2200147. doi: 10.1002/advs.202200147

  13. [13]

      N. Tian, Z. Zhou, S. Sun, et al., Science 316 (2007) 732–735. doi: 10.1126/science.1140484

  14. [14]

      N. Tian, Z. Zhou, N. Yu, et al., J. Am. Chem. Soc. 132 (2010) 7580–7581. doi: 10.1021/ja102177r

  15. [15]

      S. Hu, N. Tian, M. Li, et al., Chem. Synth. 3 (2023) 4. doi: 10.20517/cs.2022.32

  16. [16]

      L. Liang, Q. Feng, X. Wang, et al., Angew. Chem. Int. Ed. 62 (2023) e202218039. doi: 10.1002/anie.202218039

  17. [17]

      P. Chen, J.S. Du, B. Meckes, et al., J. Am. Chem. Soc. 139 (2017) 9876–9884. doi: 10.1021/jacs.7b03163

  18. [18]

      L. Huang, M. Liu, H. Lin, et al., Science 365 (2019) 1159–1163. doi: 10.1126/science.aax5843

  19. [19]

      L. Huang, H. Lin, C.Y. Zheng, et al., J. Am. Chem. Soc. 142 (2020) 4570–4575. doi: 10.1021/jacs.0c00045

  20. [20]

      C. Wang, L. Zhang, H. Yang, et al., Nano Lett. 17 (2017) 2204–2210. doi: 10.1021/acs.nanolett.6b04731

  21. [21]

      L. Zhang, J. Shi, M. Liu, et al., Chem. Commun. 50 (2014) 192–194. doi: 10.1039/C3CC46423G

  22. [22]

      H. Zhang, X. Xia, W. Li, et al., Angew. Chem. Int. Ed. 49 (2010) 5296–5300. doi: 10.1002/anie.201002546

  23. [23]

      L. Zhang, L. Roling, X. Wang, et al., Science 349 (2015) 412–416. doi: 10.1126/science.aab0801

  24. [24]

      H. Du, S. Luo, K. Wang, et al., Chem. Mater. 29 (2017) 9613–9617. doi: 10.1021/acs.chemmater.7b03406

  25. [25]

      Q. Zhang, W. Li, L. Wen, et al., Chem. Eur. J. 16 (2010) 10234–10239. doi: 10.1002/chem.201000341

  26. [26]

      X. Han, Q. Kuang, M. Jin, et al., J. Am. Chem. Soc. 131 (2009) 3152–3153. doi: 10.1021/ja8092373

  27. [27]

      W. Gong, Z. Jiang, R. Wu, et al., Appl. Catal. B 246 (2019) 277–283. doi: 10.1016/j.apcatb.2019.01.061

  28. [28]

      L. Sun, Q. Wang, M. Ma, et al., Nano Energy 103 (2022) 107800. doi: 10.1016/j.nanoen.2022.107800

  29. [29]

      L. Ma, C. Wang, M. Gong, et al., ACS Nano 6 (2012) 9797–9806. doi: 10.1021/nn304237u

  30. [30]

      Q. Li, M. Rellán-Piñeiro, N. Almora-Barrios, et al., Nanoscale 9 (2017) 13089–13094. doi: 10.1039/C7NR03889E

  31. [31]

      X.X. Wang, S. Hwang, Y. Pan, et al., Nano Lett. 18 (2018) 4163–4171. doi: 10.1021/acs.nanolett.8b00978

  32. [32]

      Z. Zhao, Z. Liu, A. Zhang, et al., Nat. Nanotechnol. 17 (2022) 968–975. doi: 10.1038/s41565-022-01170-9

  33. [33]

      M. Xie, Z. Lyu, R. Chen, et al., J. Am. Chem. Soc. 143 (2021) 8509–8518. doi: 10.1021/jacs.1c04160

  34. [34]

      C. Li, S. Kwon, X. Chen, et al., Nano Lett. 23 (2023) 3476–3483. doi: 10.1021/acs.nanolett.3c00567

  35. [35]

      Z. Zhang, X. Tian, B. Zhang, et al., Nano Energy 34 (2017) 224–232. doi: 10.1016/j.nanoen.2017.02.023

  36. [36]

      Y. Li, H. Li, G. Li, et al., Nanoscale 14 (2022) 14199–14211. doi: 10.1039/d2nr04316e

  37. [37]

      F. Yang, J. Ye, L. Gao, et al., Adv. Energy Mater. 13 (2023) 2301408. doi: 10.1002/aenm.202301408

  38. [38]

      J. Cui, P. Ma, W. Li, et al., Nano Res. 14 (2021) 4714–4718. doi: 10.1007/s12274-021-3410-y

  39. [39]

      I. Pašti, N. Gavrilov, S. Mentus, Electrochim. Acta 130 (2014) 453–463. doi: 10.1016/j.electacta.2014.03.041

  40. [40]

      H. Ke, C. Taylor, J. Electrochem. Soc. 167 (2020) 111502. doi: 10.1149/1945-7111/aba44e

  41. [41]

      K. Li, X. Li, H. Huang, et al., J. Am. Chem. Soc. 140 (2018) 16159–16167. doi: 10.1021/jacs.8b08836

• 

  Figure 1  (a) Schematic illustration of the preparation process of samples. Structural characterizations of HCl-Pt and HCl-Pt-M NCs. (b) TEM and (c) HAADF-STEM image of HCl-Pt NCs. (d) TEM and (e) HAADF-STEM images of HCl-Pt-Ni NCs. (f) TEM and (g) HRTEM images of HCl-Pt-Co NCs. The inset of (e) shows the formation of unsaturated atoms on HCl-Pt-Ni NCs. EDS-mapping profile of (h) HCl-Pt-Co and (i) HCl-Pt-Ni NCs.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Structural characterizations of the nanocrystals. (a) XRD patterns and (b) Pt 4f XPS spectra of HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs. (c) Co 2p spectra of Pt₃Co NCs and HCl-Pt-Co NCs. (d) Ni 2p XPS spectra of Pt₃Ni NCs and HCl-Pt-Ni NCs.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  DFT calculations. (a) Theoretical structures of the Cl⁻-adsorbed nanocrystal surface. (b) The range of Cl⁻ adsorption energies on different nanocrystal surfaces. PDOS analysis for nanocrystals (c) without and (d) with a Pt-skin on the surface. (e) Vacancy formation energies (Pt, Ni and Co) for Pt NCs, Pt₃Ni NCs, and Pt₃Co NCs.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Electrocatalytic ORR performance. (a) CV curves, (b) LSV curves, (c) mass activity (@0.9 V), and (d) specific activity (@0.9 V) of Pt/C, HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs. (e) LSV curve evolution of HCl-Pt-Ni after ADTs of 10,000 cycles. (f) mass activity (@0.9 V) variations of Pt/C, HCl-Pt NCs, HCl-Pt-Co NCs and HCl-Pt-Ni NCs after ADTs of 10,000 cycles.

  下载: 全尺寸图片 幻灯片
•