A fundamental viewpoint on the hydrogen spillover phenomenon of electrocatalytic hydrogen evolution

Hydrogen spillover phenomenon of metal-supported electrocatalysts can significantly impact their activity in hydrogen evolution reaction (HER). However, design of active electrocatalysts faces grand challenges due to the insufficient understandings on how to overcome this thermodynamically and kinetically adverse process. Here we theoretically profile that the interfacial charge accumulation induces by the large work function difference between metal and support (∆Φ) and sequentially strong interfacial proton adsorption construct a high energy barrier for hydrogen transfer. Theoretical simulations and control experiments rationalize that small ∆Φ induces interfacial charge dilution and relocation, thereby weakening interfacial proton adsorption and enabling efficient hydrogen spillover for HER. Experimentally, a series of Pt alloys-CoP catalysts with tailorable ∆Φ show a strong ∆Φ-dependent HER activity, in which PtIr/CoP with the smallest ∆Φ = 0.02 eV delivers the best HER performance. These findings have conclusively identified ∆Φ as the criterion in guiding the design of hydrogen spillover-based binary HER electrocatalysts.

1. Numerous studies manifest that the hydrogen spillover is highly relevant to the properties of supports, such as reducibility and facets (Nature 541, 68-71 (2017), Nat Nanotech 15, 848-853 (2020)). They used in-suit XAFS or control experiments to confirm the phenomenon of hydrogen spillover. In this work, how can the authors confirm the hydrogen spillover on CoP supports without any convincing experiment characterizations? We do suspect the existence of hydrogen spillover in this work. If the authors can't solve this question, the viewpoint of this work is unconvincing and may be misleading to other researchers.
2. In this work, the authors claimed that the introduction of the second metal can tune the work functions of Pt alloy nanoparticles, which then correlate the work function difference with hydrogen evolution activity. However, the authors can't rule out the activity contribution of second metal such as Ir, which is also reported as a superior HER catalyst (Adv Mater 30, 1805606 (2018)), as well as the electronic structure change of platinum alloying induced activity enhancement (J Am Chem Soc 141, 19964-19968 (2019)). The authors just discussed it in DFT calculations without any experiment results. The authors should at least design some control experiments to try to clarify it.

Reviewer #2 (Remarks to the Author):
The manuscript submitted by Ma et al. reported the synthesis and characterization of PtM/CoP catalysts (M=Ir, Rh, Pd, Ag, and Au) to fundamentally understand what are the key factors behind hydrogen spillover phenomenon for HER. The measured activity of Pt2Ir1/CoP catalyst is higher than even ~ other reported catalysts. Theoretical calculations are employed to obtain insight toward key factors of hydrogen spillover. The proposed concept of work-function difference is a novel and new approach to unveil hydrogen spillover in electrochemical system, and the authors conducted characterization accordingly. Considering these point, I will recommended this paper to be published on Nat. Commun after carrying out a minor revision.
1. Measured currents during catalytic evaluation are the sum of the anodic (HOR) and cathodic (HER) currents. Because Pt-based materials are highly active catalyst for both of HER and HOR, it is important to maintain H2 atmosphere in electrolyte during the experiment. According to important for us to improve the quality of our studies. We would like to address his/her comments. We do agree that the reviewer gave very constructive comments on our studies, which could help us to give a clear presentation and convincing evidences on these issues. Generally, a catalytic system is complicated, in which many parameters could affect the catalytic performance of a catalyst. In this revised manuscript, we would like to supply and reorganize more experimental results as well as presentation and to provide the solid experimental evidences and demonstrate the hydrogen spillover process in our study. Considering that these three questions are causally related, the response to the Comment #1 and #2 are integrated as below.

What we presented in this manuscript is a fundamental understanding on hydrogen spillover phenomenon of metal-supported HER electrocatalysts and the importance of work function difference between metal and support (Δφ
As mentioned, except for the contributions of hydrogen spillover, the improvements on HER performance for the Pt2Ir1/CoP catalysts may be also related to the contributions of Pt2Ir1 or CoP itself. This view is widely accepted while revisiting the previous efforts on the metal-supported HER electrocatalysts. For instance, Baek et.al (Adv. Mater. 2018, 30, 1805606) reported the superior HER activity of catalysts of encapsulating Ir nanoparticles inside a cage-like organic network (Ir@CON), which took benefits from the intrinsically decent HER activity of Ir species and further optimization for its spatial structures by CON. Wu et. al (J. Am. Chem. Soc. 2019, 141, 19964-19968) reported the catalysts of depositing Pt submonolayer on an intermetallic Pd3Pb nanoplate (AL-Pt/Pd3Pb) for efficient HER electrocatalysis, which could be ascribed to the largely optimized atomic efficiency and electronic structure of the catalytically active Pt layer. Yang et. al (Energy Environ. Sci.,2020, 13, 3110) reported that the catalysts of introducing single-atom Pt dopant into the Co2P catalysts (Pt-Co2P) significantly optimized the electronic structures and thereby HER process of Co2P, affording the superior HER performance. Clearly, the catalysts mentioned above have either high metal loading (Ir@CON,22 wt.%) or small metal size (Pt-Co2P and AL-Pt/Pd3Pb, atomic scale). These characters were beneficial to achieve abundant HER catalytic sites on metal as well as strong electronic metal-support interaction, which were the preconditions of achieving the profound contributions of metal or support itself for the overall HER activity. Considering the low total Pt2Ir1 loading (1.0 wt.%) and the large Pt2Ir1 size (~ 1.6 nm) in Pt2Ir1/CoP catalysts, it was expected that Pt2Ir1 or CoP itself in Pt2Ir1/CoP catalysts would not contribute profoundly to the overall HER activity. To experimentally examine the roles of Pt2Ir1 or CoP and figure out the nature behind the HER activity improvements, various control experiments have been designed and carried out.   In Figure R3a, the rGO was catalytically inert for HER, consistent with the previous reports (J. Am. Chem. Soc. 2011, 133, 7296-7299;Adv. Funct. Mater. 2016, 26, 6785-6796). Thus, the apparent HER activity of Pt/rGO and Pt2Ir1/rGO could effectively embody the catalytic contributions of the loaded metals themselves.
Experimentally, the Pt2Ir1/rGO catalysts showed far higher overpotential of 193 mV to reach 20 mA/cm 2 (20) and much larger Tafel slope of 86.2 mV/dec, in comparison with those of Pt2Ir1/CoP (20 = 7 mV and Tafel slope = 25.2 mV/dec) and even CoP (20 = 156 mV and Tafel slope = 103 mV/dec), suggesting the non-dominant contributions of Pt2Ir1 in Pt2Ir1/CoP. In addition, the 20 and Tafel slope of Pt2Ir1/rGO were slightly less than that of Pt/rGO (20 = 166 mV and Tafel slope = 86.2 mV/dec), further excluding the case that Pt2Ir1 itself in Pt2Ir1/CoP catalysts enhanced and dominated the HER activity improvements in our catalytic system. These results could also correspond to the DFT calculations in the revised manuscript ( Figure 7a).
As the reviewer noticed, compared with the Pt/CoP model, the Pt2Ir1/CoP showed the changes in the ΔGH on Pt2Ir1 itself (site 1': -0.20 eV → -0.39 eV; site 2': -0.06 eV → -0.15 eV), indicating the stronger hydrogen adsorption at the Pt2Ir1 sites. When considering the separate HER on Pt2Ir1 itself, the stronger hydrogen adsorption enables faster proton supply for the reaction, however, on the other side, this leads to weaker hydrogen desorption and slower release of active sites, still limiting the overall HER rate (Science, 2017, 355, eaad4998). From this view, the less HER activity of Pt2Ir1/rGO relative to Pt/rGO is predictable.
(b) To investigate the catalytic contributions of the CoP itself, we have incorporated thiocyanate ions (SCN − ), which are known to block and deactivate the Pt and Ir sites under acidic conditions (Adv. Mater. 2018, 30, 1805606;Energy Environ. Sci., 2020, 13, 4921-4929). In Figure R3b, the CoP-like HER activity for SCN-Pt2Ir1/CoP (20 = 149 mV and Tafel slope = 107.1 mV/dec) provided the convincing experimental evidences that the catalytic contribution of CoP itself was non-dominating. In addition, the HER activity of SCN-Pt2Ir1/CoP was very close to that of SCN-Pt/CoP (20 = 151 mV and Tafel slope = 100.1 mV/dec), excluding the dominated contribution of CoP itself in Pt2Ir1/CoP catalysts for the HER activity improvements.
Above control experiments provided strong evidences to prove the significantly enhanced HER activity on Pt2Ir1/CoP catalysts was not resulted from Pt2Ir1 or CoP itself. Reviewer also mentioned Ir as good HER catalysts (Adv. Mater. 2018, 30, 1805606) and Pt alloying with regulated electronic structures as the promoted HER catalysts (J. Am. Chem. Soc. 2019, 141, 19964-19968). The mentioned Ir@CON catalysts (Adv. Mater. 2018, 30, 1805606) carried out a large Ir usage and a unique strategy for optimizing the spatial structures of Ir by CON, affording the abundant catalytically Ir sites for HER. In this way, the intrinsically good HER activity of Ir could be completely utilized, thus creating superior overall HER performance.
Especially, the Ir loading in this study was over 500 μg/cm 2 , which was much higher than our case with a very low Ir loading of ~0.2 μg/cm 2 (Table R1). In addition, the above experimental observations showed that the HER activity of Pt2Ir1/rGO and Pt2Ir1/CoP catalysts were totally different even under the same Ir loading of ~0.2 μg/cm 2 . Thus, the roles of Ir should be different, compared to this previous study.  Mater. 2018, 30, 1805606 On the other hand, the mentioned AL-Pt/Pd3Pb catalysts (J. Am. Chem. Soc. 2019, 141, 19964-19968) have an atomic-scale Pt submonolayer for HER, affording the high atomic efficiency. Upon this, the contributions of electronic structure regulation for Pt could be effectively magnified, eventually resulting in superior overall HER performance. However, the Pt utilization activity in this study was 115.6 A/mgPt at -50 mV vs RHE, which was much lower than our case with a Pt utilization activity of 165.2 A/mgPt but with a much larger particle size (~1.6 nm) of Pt2Ir1 (Table R2). In addition, it was shown that the Pt utilization activity of Pt2Ir1/rGO and Pt2Ir1/CoP catalysts were totally different even under the similar size (~1.6 nm) of loaded Pt2Ir1.
All results suggested the different catalytic mechanism for such high catalytic performance of Pt2Ir1/CoP, compared to those previous studies. For a better understanding, the updated discussions can be found in our revised manuscript (Page 11-14 and Page 19).  Chem. Soc. 2019, 141, 19964-19968 (2) Evidences of hydrogen spillover phenomenon: the operando electrochemical

impedance spectra (EIS) and cyclic voltammetry (CV) investigations
The reviewer proposed that hydrogen spillover was highly relevant to the properties of supports in heterogeneous catalysis and the utilization of the operando spectroscopy (such as in-situ XAFS) on supports could confirm the hydrogen spillover phenomenon (Nature, 2017, 541, 68-71;Nat. Nanotech. 2020, 15, 848-853 Figure R4 and Table R3).
Following a previous recognition (Energy Environ. Sci., 2019, 12, 2298-2304 Table R3. In this way, the corresponding hydrogen adsorption kinetics should also change. Considering the potential-dependent R2 for all catalysts, it is rational to quantify their hydrogen adsorption kinetics via plotting log R2 vs. overpotential and calculating the EIS-derived Tafel slopes by virtue of the Ohm's law (J. Power Sources 2006, 158, 464-476). As shown in Figure R6, the similar EIS-derived Tafel Int. Edit. 2019, 58, 16038-16042;J. Am. Chem. Soc. 2009, 131, 14756-14760;J. Mater. Chem. A 2014, 2, 3954-3960;Int. J. Hydrogen Energy 2018, 43, 1251-1260. As shown in Figure R7, the CV curves shows that the intensities of the hydrogen desorption peaks for Pt/rGO and Pt2Ir1/rGO are equally weak. Similar characteristics were also found on the hydrogen desorption peaks of bare CoP,  C 2011, 115, 11880-11886). Thus, it is rational to quantify their hydrogen desorption kinetics via plotting hydrogen desorption peak position vs. scan rate and comparing the fitted slopes (Figure R8a-8c). As shown in Figure R8d, the similar fitted slope of Pt/CoP compared with that of bare CoP suggests its unaltered hydrogen desorption kinetics. Comparatively, the significantly reduced slope for Pt2Ir1/CoP suggests its drastically accelerated hydrogen desorption kinetics. It was reported that the hydrogen desorption kinetics for metal-support electrocatalysts could be effectively accelerated by hydrogen spillover effect (Angew. Chem. Int. Edit. 2019, 58, 16038-16042).
Therefore, the unaltered kinetics of the hydrogen desorption for Pt/CoP should be related to the limited hydrogen spillover from Pt to CoP, while the faster kinetics of the hydrogen desorption for Pt2Ir1/CoP should be originated from the efficient hydrogen spillover from Pt2Ir1 to CoP. (c) Consistence of theoretically calculated and experimentally measured Tafel slopes. Referring to the previous reports (Nat. Commun. 2016, 7, 12272;Adv. Funct. Mater. 2017, 27, 1700359), the HER pathway of Pt2Ir1/CoP was described by the following equations: The reaction velocity of hydrogen evolution could be written as r = k3θCoP-H*CH + , where r is the reaction rate; k is the rate constant; θ is the hydrogen coverage of on active sites; and CH + is the concentration of hydrogen ion.
In the steady state, Therefore, the calculated Tafel slope for Pt2Ir1/CoP catalysts is: V/dec (assuming α = 0.5, F is the Faraday constant, R the Rydberg gas constant and T the absolute temperature). The coincidence between this theoretical value and experimental observation ( Figure R3a, 25.2 mV/dec) for Pt2Ir1/CoP clearly confirmed this hydrogen-spillover-based HER pathway.
(d) pH-and temperature-dependent HER performance. Such HER pathway could be further verified by investigating the pH-dependent relation of HER ( Figure   R9a and R9b). In this way, the reaction order of Pt2Ir1/CoP was experimentally determined to be 1.98, which was also in accord with the theoretical value of 2 (Equation 4). Moreover, the temperature-dependent relation of HER for Pt2Ir1/CoP catalysts was also investigated ( Figure R9c and R9d). A linear relationship between log j0 with 1/T is exhibited in a semi-logarithmic plot and then electrochemical activation energies could be calculated according to the Arrhenius equation (log j0 = log(FKc) -ΔG0/2.303RT, where R is the gas constant, ΔG0 is the apparent activation energy, F is the Faraday constant). The calculated value of ΔG0 for Pt2Ir1/CoP is 20.9 kJ/mol. Such low activation energy is beneficial to HER process. All these characters of Pt2Ir1/CoP corresponded well to those of the previously reported hydrogen-spillover-based HER electrocatalysts in acid media (Nat. Commun. 2016, 7, 12272;Adv. Funct. Mater. 2017, 27, 1700359) and strongly supported its successful hydrogen spillover process and profound contributions to the HER activity.

Depending on the well-organized evidences, (1) the case that the enhanced Pt2Ir1 or CoP itself in Pt2Ir1/CoP catalysts dominated the HER activity improvements was excluded and (2) the contributions of hydrogen spillover was confirmed. Hence, it can be convincing that small Δφ for Pt2Ir1/CoP model catalysts in our case results in their efficient hydrogen spillover phenomenon along with superior HER activity and the Δφ has the predictive power for the HER activities of the PtM/CoP
electrocatalysts. For a better understanding, we have updated the relative discussion in Page 14-18.

(3) Embodiment of the predictive power of the Δφ
To better show the predictive ability of the Δφ for hydrogen spillover in our studies, we also attempted to verify their efficacy with a previously reported work on other HSBB electrocatalysts. Shao et. al (Adv. Funct. Mater., 2017, 27, 1700359) reported a series of HSBB catalysts of loading various metal (Ag, Au, Re, Ru, Rh and Pd) on molybdenum disulfide (MoS2) support for HER, in which Rh/MoS2 delivered the best HER activity. In this study, authors did not give a deep understanding on the best performance of Rh among all metals. Thus, we analysed their HER activity parameters (15) as a function of Δφ. As shown in Figure    The proposed concept of work-function difference is a novel and new approach to unveil hydrogen spillover in electrochemical system, and the authors conducted characterization accordingly. Considering these points, I will recommend this paper to be published on Nat. Commun after carrying out a minor revision.
We thank the reviewer's constructive comments on our manuscript. We would like to address those comments as below.  Soc., 2015Adv. Mater. 2019, 31, 1900813). As shown in Figure R1, Pt2Ir1/CoP catalysts deliver an overpotential of 9 mV to reach 20 mA/cm 2 and a Tafel slope of 25.0 mV/dec. The detailed comparison of HER activity in H2-saturated 0.5 M H2SO4 electrolyte for Pt2Ir1/CoP with other reported catalysts is shown in Table R1. Clearly, Pt2Ir1/CoP presents the superior HER activity among the state-of-the-art HER electrocatalysts, especially Pt-based catalysts. We have added the relevant discussions in the revised manuscript (Page 10).  Response: Thanks for pointing this out. We have already revised this error as shown in Figure R2. The updated results can be found in our revised manuscript ( Figure   6b).  Figure R3 and R4) to the catalysts in the initial submission, therefore their influences can be excluded.
In Figure R3a, the rGO is catalytically inert for HER as previously reported.       To prepare CoP support, the as-prepared Co(OH)2 nanosheets (100 mg) and NaH2PO2·H2O (2 g) were put at two separate positions in a quartz boat with NaH2PO2 at the upstream side of the furnace. Subsequently, the temperature of the tube furnace was raised to 300 °C with a ramping rate of 5 °C/min and maintained at 300 °C for 60 min, and then naturally cooled to room temperature under the protection of Ar gas with a flow rate of 100 mL/min. Powder XRD pattern of the product is shown in Figure R7a and consistent with that of CoP standard (JCPDS # 29-0497), suggesting the successful synthesis of CoP support. In the high-resolution XPS of the product ( Figure R7b), the typical signals of Co and P species in CoP were identified, further confirming the formation of CoP support (J. Am. Chem. Soc. 2016, 138, 14686-14693). TEM image ( Figure R7c)  and Ir 4f region for commercial Pt/C (20 wt.%) and Ir/C (wt.%) catalysts were added into the Figure S14 of the initial manuscript. The updated results were shown in Figure R8. Compared with Pt/CoP, the XPS peaks of Pt 4f7/2 and Pt 4f5/2 for Pt2Ir1/CoP shift to the low binding energy, while its XPS peaks of Co 2p3/2 and Co 2p2/1 shift to the high binding energy. In addition, the XPS peaks of Pt 4f7/2 (71. Electroanal. Chem., 1997, 424, 141-151;Electroch. Acta, 1997, 42, 323-330;J. Power Sources, 2006, 158, 464-476 , 2010, 46, 8401-8403;Chem. Mater., 2009, 21, 3649-3654;J. Am. Chem. Soc., 2014, 136, 1280-1283Langmuir, 2010Langmuir, , 26, 2339Langmuir, -2345 In this case, to experimentally support this recognition, we have used a sensitive and reliable electrochemical method, the stripping of adsorbed carbon monoxide (CO) measurements ( Figure R9), in which the CO stripping peak characteristics were validated as an indicator of the composition of Pt-based metal (Chem. Commun., 2014, 50, 11558-11561;J. Mater. Chem. A, 2016, 4. 15400-15410). In the case of the Pt/CoP and Ir/CoP, the CO stripping peaks are seen to be centered at 0.83 and 0.93 V vs. RHE respectively, corresponding well to the previous reports (J. Mater. Chem. A, 2016, 4. 15400-15410;Langmuir, 1997, 13, 6713-6721). The much more positive CO stripping peak and much larger full width at half maximum (FWHM) for Ir/CoP over those for Pt/CoP could be explained as arising from the higher desorption activation energy of CO from Ir (22 kcal/mol) compared to that from Pt (13 kcal/mol) (J. Phys. Chem., 1988, 92, 5213-5221 I think the main problem of this paper is that they do not really show there is H spillover in their system. They seem to assume that this is the case, and then study how to optimize the component materials so that the H would not be trapped at the interface. There could be other reasons (without involving the H spillover) for the observed improved performance: for example, the nanoparticle itself may get better.
In fact, as stated in page 13, "the changes in the ΔGH on site 1' and site 2' range from -0.20 eV to -0.39 eV and from -0.06 eV to -0.15 eV…"this means that by alloying the particle itself gets better. There seems to be no need to involve the interface and further H spill over. The authors must try best to clarify strengthen this part, factual evidence of logical support for spillover action, in a good revision.

Response:
We thank the reviewer for raising the useful comments on our manuscript.

What we presented in this manuscript is a fundamental understanding on hydrogen spillover phenomenon of metal-supported HER electrocatalysts and the importance of work function difference between metal and support (Δφ) on the hydrogen spillover as well as HER activity for the hydrogen-spillover-based binary (HSBB)
catalyst of Pt2Ir1/CoP. The nature of the ∆Φ and its contribution to HER performance are mainly embodied in two aspects. The reviewer mentioned the previous theoretically study on the hydrogen spillover phenomenon (ACS Nano, 2009, 3, 1657 of metal-support catalysts, which proposed that the metal should have a large enough hydrogen chemical potential relative to the support to enable the thermodynamically favorable hydrogen spillover. Inspired by this, it was realized that the Pt2Ir1/CoP catalysts in our case should be essentially similar. These alloyed Pt2Ir1 sites with enriched electron density were endowed with enhanced proton adsorption. Hence, the hydrogen chemical potential on Pt2Ir1 should be significantly increased, thermodynamically facilitating the interfacial hydrogen spillover from Pt2Ir1 to CoP. This previous theoretically work provides a strong support on our study and helps us better explain the energetically favorable hydrogen spillover in Pt2Ir1/CoP catalysts. We have added the relevant discussions in Page

20-21 and cited this paper as Ref [73] in our revised manuscript.
Most importantly, reviewer raised the concern on the existence of hydrogen spillover phenomenon in our Pt2Ir1/CoP catalysts. Indeed, the catalytic system in our work is complicated, in which many parameters could affect the catalytic activity.
In this revised manuscript, we would like to supply and reorganize more experimental results as well as presentation and to provide the solid experimental evidences and demonstrate the hydrogen spillover process in our study.

(1) Explicitation of catalytic contributions: the catalytic performance of various control catalysts.
Except for the contributions of hydrogen spillover, the HER activity improvements for the Pt2Ir1/CoP catalysts may be also related to the contributions of     In Figure R3a, the rGO was catalytically inert for HER, consistent with the previous reports (J. Am. Chem. Soc. 2011, 133, 7296-7299;Adv. Funct. Mater. 2016, 26, 6785-6796). Thus, the apparent HER activity of Pt/rGO and Pt2Ir1/rGO could effectively embody the catalytic contributions of the loaded metals themselves.
Experimentally, the Pt2Ir1/rGO catalysts showed far higher overpotential of 193 mV to reach 20 mA/cm 2 (20) and much larger Tafel  considering the solo HER on Pt2Ir1 itself, the stronger hydrogen adsorption enables faster proton supply for the reaction, however, on the other side, this leads to weaker hydrogen desorption and slower release of active sites, still limiting the overall HER rate (Science, 2017, 355, eaad4998). From this view, the less HER activity of Pt2Ir1/rGO relative to Pt/rGO is predictable.

(b)
To investigate the catalytic contributions of the CoP itself, we have incorporated thiocyanate ions (SCN − ), which are known to block and deactivate the Pt and Ir sites under acidic conditions (Adv. Mater. 2018, 30, 1805606;Energy Environ. Sci., 2020, 13, 4921-4929). In Figure R3b, the CoP-like HER activity for

(2). Evidences of hydrogen spillover phenomenon: the operando electrochemical impedance spectra (EIS) and cyclic voltammetry (CV) investigations
The previous efforts proposed that hydrogen spillover was highly relevant to the properties of supports in heterogeneous catalysis and the utilization of the operando spectroscopy (such as in-situ XAFS) on supports could confirm the hydrogen spillover phenomenon (Nature, 2017, 541, 68-71;Nat. Nanotech. 2020, 15, 848-853 (Energy Environ. Sci., 2019, 12, 2298-2304. The recorded Nyquist plots were simulated by a double-parallel equivalent circuit model ( Figure R4 and Table R3). Following a previous recognition (Energy Environ. Sci., 2019, 12, 2298-2304, the second parallel components (Cφ and R2) reflect the hydrogen adsorption behavior on catalyst surface, where Cφ and R2 represent the hydrogen adsorption pseudo-capacitance and resistance, respectively.  Table R1. As shown in Figure R5, the integration of Cφ vs. η profiles gives the hydrogen adsorption charge (QH) on catalyst surfaces during HER. In this way, the corresponding hydrogen adsorption kinetics should also change.
Considering the potential-dependent R2 for all catalysts, it is rational to quantify their hydrogen adsorption kinetics via plotting log R2 vs. overpotential and calculating the EIS-derived Tafel slopes by virtue of the Ohm's law (J. Power Sources 2006, 158, 464-476). As shown in Figure R6, the similar EIS-derived Tafel slope of Pt/CoP compared with that of bare CoP suggests its unaltered hydrogen adsorption kinetics.
Hence, Pt/CoP showed the individual hydrogen adsorption on respective Pt and CoP, supporting the limited hydrogen spillover from Pt to CoP due to the sluggish spillover kinetics. Comparatively, the significantly declined EIS-derived Tafel slope for  Int. Edit. 2019, 58, 16038-16042;J. Am. Chem. Soc. 2009, 131, 14756-14760;J. Mater. Chem. A 2014, 2, 3954-3960;Int. J. Hydrogen Energy 2018, 43, 1251-1260. As shown in Figure R7, the CV curves shows that the intensities of the hydrogen desorption peaks for Pt/rGO and Pt2Ir1/rGO are equally weak. Similar characteristics were also found on the hydrogen desorption peaks of bare CoP, SCN-Pt/CoP and SCN-Pt2Ir1/CoP. Above facts further supported that the enhancement of Pt2Ir1 or CoP itself in Pt2Ir1/CoP catalysts was too slight to dominate its hydrogen desorption behavior and thereby HER activity improvements. Naturally, the similar hydrogen desorption peak of Pt/CoP compared with that of bare CoP suggests its almost non-increased amount of desorbed hydrogen, corresponding to the limited hydrogen spillover from Pt to CoP and thus the lack of abundant spillovered hydrogen for desorption. In contrast, the significantly stronger hydrogen desorption peak of Pt2Ir1/CoP compared to that of Pt/CoP strongly indicated the existence of hydrogen spillover from Pt2Ir1 to CoP and thus the excess spillovered hydrogen on CoP for efficient desorption.  C 2011, 115, 11880-11886). Thus, it is rational to quantify their hydrogen desorption kinetics via plotting hydrogen desorption peak position vs. scan rate and comparing the fitted slopes ( Figure R8a-8c). As shown in Figure R8d, the similar fitted slope of Pt/CoP compared with that of bare CoP suggests its unaltered hydrogen desorption kinetics. Comparatively, the significantly reduced slope for Pt2Ir1/CoP suggests its drastically accelerated hydrogen desorption kinetics. It was reported that the hydrogen desorption kinetics for metal-support electrocatalysts could be effectively accelerated by hydrogen spillover effect (Angew. Chem. Int. Edit. 2019, 58, 16038-16042). (c) Consistence of theoretically calculated and experimentally measured Tafel slopes. Referring to the previous reports (Nat. Commun. 2016, 7, 12272;Adv. Funct. Mater. 2017, 27, 1700359), the HER pathway of Pt2Ir1/CoP was described by the following equations: The reaction velocity of hydrogen evolution could be written as r = k3θCoP-H*CH + , where r is the reaction rate; k is the rate constant; θ is the hydrogen coverage of on active sites; and CH + is the concentration of hydrogen ion.
In the steady state, Therefore, the calculated Tafel slope for Pt2Ir1/CoP catalysts is: kJ/mol. Such low activation energy is beneficial to HER process. All these characters of Pt2Ir1/CoP corresponded well to those of the previously reported hydrogen-spillover-based HER electrocatalysts in acid media (Nat. Commun. 2016, 7, 12272;Adv. Funct. Mater. 2017, 27, 1700359) and strongly supported its successful hydrogen spillover process and profound contributions to the HER activity.  Large difference in the Fermi energies of metal and support drives the interfacial charge flow until the system reaches an equilibrium, followed by the Schottky barrier formation and charge accumulation at the interface, thus strongly tapping proton at interface. In this case, the interfacial hydrogen spillover has to overcome a large energy barrier, leading to the unsatisfactory HER activity. Thus, the ∆Φ is the key determinant of hydrogen spillover barrier and used to reflect the overall HER activity.  (Nature 2017, 541, 68-71;Nat. Nanotech. 2020, 15, 848-853;Chem. Rev., 2012, 112, 2714-2738.

Figure R9 | (a)
Unlike those cases, in electrocatalytic hydrogen evolution reaction (HER), hydrogen could also be adsorbed on support (like CoP in our case) and thus bring the ambiguity whether the hydrogen intermediates on support originate from the hydrogen adsorption on itself or hydrogen spillover. In addition, the electrolyte environment (such as H2O, H3O + , SO4 2-, H + , OHand K + ) during HER goes against the monitoring of the spillovered hydrogen on support. Faced with these common issues, the use of the state-of-the-art operando techniques (e.g. XAS, FTIR) to demonstrate the electrocatalytic hydrogen spillover faces great challenges. Thus, it is still lack of solid experimental evidences on the presence of hydrogen spillover in electrocatalytic HER. Generally, theoretical simulation and indirect evidences are adopted as the state-of-the-art approach to understand the hydrogen spillover in electrocatalytic HER, such as the calculations of the corresponding thermodynamic and kinetic parameters (e.g. activation energy, energy difference, Tafel slope and reaction order) or investigations on the hydrogen spillover induced phase change of support (Energy & Environmental Science, 2019, 12, 2298-2304Nat. Commun., 2016, 7, 1-7;Angew. Chem. Int. Ed., 2020, 59, 20423-20427;Nano Energy, 2020, 71, 104653).
To minimize the impact of the above common issues, what we enabled in this work is a strategy of utilizing operando electrochemical investigations on the hydrogen adsorption/desorption behavior on support, which is equivalent to the concept of monitoring the spillovered hydrogen on support in heterogeneous catalysis and experimentally support the existence of hydrogen spillover in our case.
As also stated by the Reviewer #3, this strategy is an opportunity to provide solid experimental evidences on electrocatalytic hydrogen spillover phenomenon but leaving debate. To be specific, such electrochemical investigations were performed in a standard three-electrode system, containing centimeter-sized reference,  Figure S12, S19 and Figure 6) b). Hydrogen adsorption behavior analysis (Figure 6a and 6b); c). Hydrogen desorption behavior analysis (Figure 6c and 6d); d). Comparison of theoretically calculated and experimentally measured Tafel slopes (Figure 4b and Equation S5); e). pH-and temperature-dependent HER performance (Figure 6e and 6f); f). Interfacial charge dilution and hydrogen spillover channel formation revealed by DFT calculations (Figure 7).
Although these evidences are indirect, all above results demonstrate the hydrogen spillover phenomenon as the most likely reason for the significantly improved HER performance herein. We have done what we can do to demonstrate the electrocatalytic hydrogen spillover phenomenon in our case, which is indeed a significant advance compared with other studies on the HER electrocatalysts based on hydrogen spillover.
Also, we provide a critical descriptor and a deep understanding on the design of highly performed HER electrocatalysts through hydrogen spillover.
We do agree with the importance of the direct evidences on hydrogen spillover proposed by the Reviewer #1. To obtain this goal, it could completely solve the above common issues on utilizing the state-of-the-art operando techniques for tracking the spillovered hydrogen. Unfortunately, with the current methodologies and technologies, this is almost impossible due to the inevitability of the existence of the adsorbed hydrogen on support as well as the H2O in electrolyte. To address the Reviewer #1's concerns, the following aspects might be considered in the future:

a) Developing the transient imaging technology:
If existing the efficient hydrogen spillover, large amounts of spillovered hydrogen may accumulate on support to form nanobubbles (Proc. Nati. Acad. Sci., 2018, 115, 5878-5883). Thus, developing the transient imaging technology (like super-resolution fluorescence microscopy with suitable fluorescence dye molecules) to label the hydrogen nanobubbles and observe their nucleation, growth and migration will provide the solid evidences on hydrogen spillover. However, it requires a very high spatial resolution for this technique to identify the metals and supports.
b) Developing the nano-sized three-electrode electrochemical system: The limitations of our operando electrochemical investigations are derived from the much larger size of the three-electrode device compared to that of the catalysts. Developing the nano-sized three-electrode device might provide the possibility to investigate the local electrochemical response on catalyst surface. In this way, a transient electrochemical response at the interface will be detected during hydrogen spillover.

c) Developing the special techniques to avoid the interference of solvents and electrolytes:
The common issues of utilizing the state-of-the-art operando spectroscopy (EXFAS and FT-IR) to trace the spillovered hydrogen in current HER electrocatalysts are the interference of adsorbed hydrogen of support and various ion/molecule in ambience (H2O, H3O + , SO4 2-, H + , OHand K + ) when using the common catalytic system (H2SO4 or KOH aqueous electrolyte). Thus, developing the special system may solve the above common issues without influencing the HER process of the catalysts, which lays the foundation for re-enabling the operando spectroscopy to trace the spillover hydrogen.
We will work in this regard and further promote the development of the novel HER electrocatalysts based on hydrogen spillover.
The Reviewer #1 also mentioned that the Ir/CoP catalysts should be also prepared for comparison. We thank the reviewer for this constructive suggestion. Herein, we have supplemented the control catalysts by loading Ir nanoparticles on CoP (Ir/CoP) through the similar approach to examine the contributions of Ir for such high catalytic performance of Pt2Ir1/CoP. It was found that the chemical and morphological characters of the loaded Ir in Ir/CoP ( Figure R1), especially the loading (~ 1.0 wt.%) and size (~ 1.63 nm) were similar to those of Pt/CoP and Pt2Ir1/CoP (Figure 3 and S11), therefore excluding their size influences on the catalytic performance. The Ir/CoP catalysts showed an overpotential of 144 mV to reach 20 mA/cm 2 (20) and Tafel slope of 106.2 mV/dec ( Figure R2), which were similar to these of bare CoP (20 = 156 mV and Tafel slope = 108.1 mV/dec) as well as Pt/CoP (20 = 120 mV and Tafel slope = 103.1 mV/dec) and much higher than those of Pt2Ir1/CoP (20 = 7 mV and Tafel slope = 25.2 mV/dec). The results indicated the non-dominant contributions of Ir itself in Pt2Ir1/CoP and the significance of the alloyed Pt2Ir1 for the improved catalytic activity. To support these results, the hydrogen adsorption and desorption behavior of Ir/CoP were further evaluated and compared with other control catalysts by the operando EIS and CV investigations. The recorded Nyquist plots were simulated by a double-parallel equivalent circuit model ( Figure R3 and Table R1). Figure R3 | Nyquist plots for Ir/CoP catalysts at various HER overpotentials. Zoom-in parts were correspondingly presented as inset. The scattered symbols represent the experimental results, and the solid lines are simulated fitting results. The inset also shows the equivalent circuit for the simulation. The fitted parameters are summarized in Table R1. To investigate the hydrogen desorption behavior, operando CV investigations were also carried out on Ir/CoP and other control catalysts and their hydrogen desorption peak was monitored during CV scanning in the double layer region. As shown in Figure R4b, the CV curves showed that the intensity of the hydrogen desorption peaks for bare CoP, Pt/CoP and Ir/CoP was equally weak, supported that the enhancement from Ir itself for Pt2Ir1/CoP catalysts was too weak to dominate their hydrogen desorption behavior.
Overall, the above facts further confirm that Ir itself in Pt2Ir1/CoP catalysts should not dominate their HER activity improvement. For a better understanding, the updated discussions can be found in our revised manuscript (Page 7-16).

Reviewer 3
The revision is OK but the Response is so excessively long (43 pages, not to forget also 39 pages of the paper) that it reminds the "filibuster method" in some parliamentary proceedings. Anyways, the authors tried hard to address reviewers' concerns. I still do not see clear convincing evidence of the spillover process but I can accept it here as an overall plausible scenario of the process, OK. I do see and sympathize with the stronger reservations of the Reviewer 1. Yet from my perspective the paper overall is thorough and its publication can be useful, even though leaving some moot points debatable in the future. I can endorse its acceptance to Nat. Comm.
at this point.

Response:
We thank the reviewer's constructive comments on our manuscript.
Through the revision, we feel a significant improvement in the quality of this manuscript. In the future, we expect to directly observe the electrocatalytic hydrogen spillover phenomenon based on the development of methodology and technology and further provides the insights on the hydrogen-spillover-based HER electrocatalysts.