Sustainable production of hydrogen with high purity from methanol and water at low temperatures

Carbon neutrality initiative has stimulated the development of the sustainable methodologies for hydrogen generation and safe storage. Aqueous-phase reforming methanol and H2O (APRM) has attracted the particular interests for their high gravimetric density and easy availability. Thus, to efficiently release hydrogen and significantly suppress CO generation at low temperatures without any additives is the sustainable pursuit of APRM. Herein, we demonstrate that the dual-active sites of Pt single-atoms and frustrated Lewis pairs (FLPs) on porous nanorods of CeO2 enable the efficient additive-free H2 generation with a low CO (0.027%) through APRM at 120 °C. Mechanism investigations illustrate that the Pt single-atoms and Lewis acidic sites cooperatively promote the activation of methanol. With the help of a spontaneous water dissociation on FLPs, Pt single-atoms exhibit a significantly improved reforming of *CO to promote H2 production and suppress CO generation. This finding provides a promising path towards the flexible hydrogen utilizations.

4. At least CO stripping experiment should be performed by cyclic voltammetry measurement for all the synthesized catalysts. 5. The surface area and porosity of CeO2 nanorods and Pt single atom nanorods CeO2 catalyst should be disclosed by BET analysis. 6. The wt % of Pt atoms should be ascertained by ICP-OES measurements. 7. The frustrated Lewis pairs in the PN-CeO2, NR-CeO2, and NP-CeO2 catalyst after loading of Pt is not studied. So, it is hard to understand whether Pt influences the frustrated Lewis pairs formation or not. 8. Why does a 20 % loss of catalytic performance after the stability test ( Figure 3C)? Is Pt single atom leaching out from CeO2 nanorods? 9. Besides, after the catalytic reactions, the existence of frustrated Lewis pairs sites should be studied.
Reviewer #1 (Remarks to the Author): The paper by Zhang et al. tackles an interesting subject, namely low temperature methanol water reforming over a catalyst of dual dual-active sites of Pt single-atoms and frustrated Lewis pairs (FLPs) on porous nanorods of CeO2.
This reviewer ranks the experimental catalysis and the DFT calculation in this paper as important for the understanding of the activation of H2O and methanol at low temperature. However, the role of the frustrated Lewis pairs (FLPs), the synergy of single Pt and FLPs in this study were not demonstrated and discussed sufficiently. This reviewer supports the publication in Nature Comm., but only after addressing the open points in the results analysis section in an appropriate manner: Response: Thank for the positive and constructive comments raised by the reviewer. We would like to address those comments in details.
Firstly, there are some confusions, this reviewer would like to discuss with the authors. It is about the key works of "Sustainable production of hydrogen, high purity, methanol and water at low temperatures" in the title.
• Page 10, line 187-189 "To the best of our knowledge, it is the lowest temperature for all previously reported heterogeneous catalysts to achieve the efficient H2 generation of reforming of methanol and water in the absence of any additives." This conclusion is inaccurate: please check the study of ACS Appl. Mater. Interfaces 2021, 13,

24702−24709
Response: Thank you for your kind reminder. The H2 production of reforming of methanol and water has been successfully achieved by the N-doped carbon dots/g-C3N4 catalysts in the study of ACS Appl. Mater. Interfaces 2021, 13, 24702-24709. The carbon dots provide a catalytic hydrogen generation from methanol and H2O without transition metal catalysts, which is of great significant in this field. In that study, the H2 generation rate was 19.5 μmol g -1 h -1 , which is much lower than the rate catalyzed by Pt1/PN-CeO2 (6400 μmol g -1 h -1 ) in this work. Nonetheless, our judgment is inaccurate in the initial submission. We have summarized the relevant data in the Table S3 to facilitate readers to grasp the research progress more comprehensively.
Meanwhile, we have modified the sentence to "To the best of our knowledge, it is among one of the lowest temperatures for all previously reported heterogeneous metal catalysts to achieve the efficient H2 generation of reforming of methanol and water in the absence of any additives." (Page 11, • In abstract section: (page two, line 19-22) "Herein, we demonstrate that the dual-active sites of Pt single-atoms and frustrated Lewis pairs (FLPs) on porous nanorods of CeO2 enable the efficient additive-free H2 generation with a low CO generation (0.027 %) through APRM at 120 °C." As we know, CO is a poison for fuel cell, and normally, even miniscule amounts of carbon monoxide (50-100 ppm) in the hydrogen are sufficient to bind to the platinum catalysts and prevent them working. Here, the 0.027 % is 270 ppm of CO. The author should discuss how to handle such high CO concentration in real application case.
Response: This is a practical question for the commercial application of liquid methanol as hydrogen storage. Methanol and H2O system offers a promising methodology for transportation and storage. In this study, the selectivity of CO is among one of the lowest levels in comparison with those of the previous studies. However, the generated H2 from this system has to be purified before it supplies the power of fuel cell at this stage.
We have added the relevant discussion in the revised manuscript, as following: "Due to hydrogen with the low CO concentration required of fuel cells as well as other applications, herein, the catalytically generated hydrogen with a CO concentration of 270 ppm from methanol and H2O by Pt1/PN-CeO2 has to be handled carefully and purified as the power supplies of fuel cell at this stage. Nevertheless, this new catalyst featured with the facile synthesis and high activity as well as the suppressed CO generation at low temperatures still paves a possible way towards a commercially achievable liquid sunshine roadmap." (Line 440-446, Page 21) • In table S1, with the same single Pt catalyst, H2 generation rate (molH2 molPt -1 h -1 ) at 135 o C is 199, while, at 120 o C it is decreased to 33. What are the advantages of carrying out this reaction with 15 o C difference, but losing 83.4% activity [(1-33/199)%]? In page 10, line 192-193 "When the reaction temperature was further increased to 165 °C, the H2 generation rate was significantly enhanced to 1103 molH2 molPt -1 h -1 ." The activity losing with 165 o C as a standard is about 97 % [(1-33/1103)%]. It seems that a lower temperature (100-135 °C) is not in the optimal working range, why the authors emphasize the advantages of low temperature? This reviewer would suggest the authors, to compare your catalyst and the catalyst in Nature, 2017, 544, 80-83 (Reference 11 in manuscript) at a wider reaction temperature range.
Response: Thank you for your comments. Up to now, the reforming of methanol and water by those catalysts still faces two big obstacles: (1) the high temperatures (＞250 °C) to boost catalytic reaction and (2) the low purity of H2 accompanied with the generation of CO at a high level. Recent advances in developing new heterogeneous catalysts have greatly decreased the operation temperatures as low as 150 °C for the aqueous phase reforming of methanol by using the atomically dispersed Pt on α-MoC. Afterwards, the further decrease of the reaction temperatures to yield a satisfactory hydrogen generation rate is extremely difficult and rarely realized on heterogeneous catalysts yet up to now. Therefore, developing high-efficient catalysts capable of in-situ releasing of H2 at even lower temperatures and the suppressed CO generation is highly desirable for the large-scale production of hydrogen, bringing us a step closer to methanol economy.
Herein, we demonstrate that the dual-active site catalysts composed of the single-atom Pt and frustrated Lewis pairs (FLPs) on the atomically dispersed Pt anchored on porous nanorods of CeO2 (Pt1/PN-CeO2) enables a stabilized H2 generation at a low-temperature of 120 °C through a base-free aqueous phase reforming of methanol. However, the hydrogen generation rate (1592 molH2 molPt -1 h -1 ) catalyzed by Pt1/PN-CeO2 at 180 °C is lower than the rate catalyzed by 2% Pt/α-MoC at 170 °C (1755 molH2 molPt -1 h -1 ) in the reference (Nature, 2017, 544, 80-83). We did not perform the hydrogen generation at a higher temperature owing to the pressure limitation of our autoclave. The main contribution of our manuscript is the satisfactory hydrogen generation with a low CO generation from methanol and H2O at low temperature (120 °C and even 100 °C) via the novel dual-active sites of single-atom Pt and FLPs sites for methanol and H2O activation, respectively. Therefore, this new catalyst featured with facile synthesis, high activity and suppressed CO generation at low temperature paves the way towards a commercially achievable liquid sunshine roadmap.
• The last one, the authors should further clarify the concept of "Sustainable production of hydrogen" with methanol as H2 carrier. Industry methanol is produced from synthesis gas (carbon monoxide and hydrogen). And one product of methanol water reforming is CO2.
Response: Undoubtedly, industry methanol is produced from synthesis gas. However, from the perspective of technological development, the sustainable production of hydrogen is achieved through the following steps: (I) As a promising methodology for hydrogen storage via methanol, the H2 source must be green hydrogen, which is transformed from solar, wind and/or other renewable energy. (II) After that, green methanol can be generated by the catalytic reaction between CO2 and H2. Also, the green methanol can be obtained from transformation of biomass.
(III) Finally, hydrogen production can be achieved via reforming of green methanol and H2O. (IV) After reforming, generated CO2 can be further transformed into methanol again via hydrogenation of CO2 by green H2. Therefore, sustainable production of hydrogen is achieved through green methanol and CO2 recycling. This concept has been reported and summarized in a previous review by Prof. Choon FongShih, Prof. Tao Zhang, Prof. Jinghai Li, and Prof.
Secondly, this study provides a systematic experiment and demonstrates a unique catalytic performance of loading single Pt atom on the surface of CeOx. There are some comments this reviewer would like to share with the authors: • Single Pt/CeOx catalyst has been widely studied. The single metal Pt-CeOx support interaction is quite important for its catalytic performance. The authors introduce a concept of "the dual-active sites of Pt single-atoms and frustrated Lewis pairs (FLPs)" to understand the surface chemistry. This reviewer consider of this concept is the major innovation in this study. This reviewer suggests the authors to provide more evidence to further clarify: the 1) coordination configuration of single Pt; 2) geometry model of FLPs, ( Figure 1D is hard to understand); 3) the geometry model of single Pt with FLPs; 4) a detailed reaction pathway on this Single Pt-FLPs (reedit Figure 6). In the reaction mechanism model, is there any synergy between the single Pt with FLPs. And the distance between single Pt and FLPs matter?
Response: Thank you for your constructive comments. We would like to answer these questions in the following order:

1) Geometry model of FLPs:
The geometry model and construction process have been systematically discussed in our previous reports (Nat. . 2017, 8, 15266;Chem. Soc. Rev. 2018, 47, 5541-5553;J. Am. Chem. Soc. 2019, 141, 11353-11357). To make it easier for the reader to understand, the geometry model of FLPs was added in our revised manuscript. As shown in Figure S1, the FLPs site is constructed by the two adjacent surface Ce 3+ as Lewis acidic site and the neighboring surface lattice oxygen as Lewis basic site on CeO2(110) surface. The constructive process of FLP site has been also described in the revised manuscript. As shown in Figure 2a Figure   2b). However, the electronic interaction between CeI and OIa/OIc will still hinder the activation of small molecules on the CeI-OIIc active sites ( Figure S2e). Therefore, the weak FLPs-like activation is obtained on the traditional CeO2 materials. When the second adjacent surface oxygen (OIa) is removed, the reduced Ce cations (CeI and CeII) and surface lattice oxygen (OIIc) are independent Lewis acidic and basis sites ( Figure S2c and S2f), respectively. Notably, two adjacent reduced surface Ce sites (CeI and CeII) and lattice OIIc is constructed the FLPs site of (CeI,CeII)-OIIc with a shorter distance (3.99 Å).

Commun
However, the construction of surface FLPs site cannot be realized by removing surface oxygen atoms on the CeO2(111) facet. Meanwhile, due to the low formation of oxygen vacancy on CeO2(100) surface, the relatively unstable oxygen defect leads to the spatial configuration for formation of FLPs sites. Therefore, CeO2(110) surface instead of CeO2(100) and CeO2 (111)

2) Coordination configuration of single Pt.
According to the aberration-corrected high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image (Figure 2a), the single-atom Pt in catalysts were experimentally demonstrated, which was further verified from the uniform Pt distribution on PN-CeO2 by the energy dispersive spectroscopy (EDS) mapping ( Figure 2b). X-ray absorption near edge structures (XANES) of Pt K-edge revealed that the white line peak of the Pt1/PN-CeO2 catalysts located at 11568.7 eV (Figure 2c), which was very close to that of PtO2. More improtantly, the k 3 -weight Fourier transforms of extended X-ray absorption fine structure (EXAFS) spectra of Pt1/PN-CeO2 delivered one prominent peak at ~1.63 Å, which was labeled as Pt-O bond ( Figure 2d). Therefore, the single-atom Pt was undoubtedly coordinated with O atom on CeO2 surface.
Meanwhile, DFT simulation was used to further explore the coordination configuration of single Pt. As shown in Figure S3, the most stable configuration is the Pt atom located in oxygen vacancy of CeO2(110) surface with the lowest adsorption energy. Meanwhile, single Pt atom can be stabilized on oxygen vacancy adjacent to the FLPs sites of CeO2(110). On this configuration, the oxygen vacancies adjacent to the FLPs sites is occupied by a single Pt atom.
Consistent with the experimental results, the single Pt atom coordinates with different number of O atoms in these two typical configurations. Therefore, the single atom Pt is located on the oxygen vacancy and coordinated with O atom on CeO2 surface. Figure S3. Summary of the adsorption energy of Pt single-atom at various sites on CeO2(110) surface.

3) The geometry model of single Pt with FLPs:
Response: Considering that the single atom Pt active sites exhibit higher capability for CH3OH activation than Pt nanoparticles, the constructed dual-active sites of single-atom Pt and FLPs (Pt1-FLP) can effectively activate both H2O and CH3OH, potentially reducing the reaction temperatures of APRM to generate H2. DFT calculations were also used to explore the possible spatial structure of Pt atom on CeO2(110) surface. As shown in Figure S3, the singleatom Pt is easier to form on the oxygen defect of CeO2(110) surface owing to the lowest formation energy (1.35 eV).
In this spatial configuration, the dual-active sites of single-atom Pt and FLPs sites in the distance is successfully designed, as type I in Figure 1d. where Pt atom is located on the oxygen vacancy of CeO2(110) surface and in the distance with FLPs sites. In addition, the single-atom Pt could occupy one of the oxygen vacancies adjacent to the FLPs sites owing to the sightly high formation energy (1.70 eV). For this spatial configuration, the Pt1-FLP dualactive sites is spatially adjacent with each other, as type II in Figure 1d. Fortunately, the spatial configuration of FLPs site, which is constructed by two adjacent Ce 3+ as Lewis acidic site and lattice O atom as Lewis basic site, is not affected by the nearby single Pt atom as well as single Pt atom in the distance of the two configurations.  We have modified and supplemented relevant description in the revised manuscript, as following: "Then, DFT calculations were further used to explore the possible spatial structures of Pt atom on CeO2 (110) surface, which delivered two configurations. As shown in Figure S3, the Pt atom prefers to occupy the oxygen defect of CeO2 (110)  Response: The CO-FTIR analysis was added in the revised manuscript. As shown in Figure S8, peaks at approximately 2090 cm -1 were associated with the linearly adsorbed CO on isolated ionic Pt 2+ , revealing that the Pt active sites existed as single-atom on PN-CeO2 and coordinated with O atoms. 1-3 Specifically, the two apparent peaks at 2099 cm -1 and 2076 cm -1 could be attributed to the single-atom Pt located in spatial configuration of the type I and type II (Figure 1d), respectively, owing to the lower valence state of Pt on oxygen vacancy adjacent to the FLPs sites than it on traditional oxygen vacancy ( Figure S3). • H/D isotope experiments and DFT calculations are well done. However, to further understand the reaction pathway of methanol-water reforming at low temperature, the authors should provide in-situ spectroscopic evidence, to detect/monitor the reaction intermediates.
Response: Thank you for your constructive suggestion. The mechanism of methanol decomposition over Pt/CeO2 catalysts has been carefully investigated in the previous reports. 4 Based on their conclusion, the CeO2 supports exhibited no catalytic ability for methanol dissociation. While, the methoxy species adsorbed on the Pt active sites were the curtail to dehydrogenate of methanol to carbon monoxide and H2. Herein, the Lewis acidic Ce 3+ as a co-active site interacts with CH3OH molecule, further promoting the CH3OH dissociation afterwards on the singleatom Pt active sites. To experimental detect the possible assistant role of Lewis acidic Ce 3+ , the adsorption behavior of methanol on catalyst surface was explored by FTIR. As shown in Figure S21, methanol molecule only adsorbed on the oxygen vacancy of PN-CeO2 via bridged methoxy species. After introduced single-atom Pt on the surface of PN-CeO2, another characteristic peak of bridged methoxy species at 1058 cm -1 was observed from the FTIR spectra.
Due to the absence of bridged methoxy species on the two adjacent Ce atoms of PN-CeO2, it could be confirmed that the methoxy species was bridged on the single-atom Pt and adjacent Ce. Therefore, the methoxy species were detected as the reaction intermediates for methanol dissociation, which was similar with previous reports. 4 Figure S21. The FTIR spectrograms of PN-CeO2 and Pt1/PN-CeO2 after adsorption of methanol.
We have added the relevant description in our revised manuscript, as following: "To experimentally detect the roles of Lewis acidic Ce 3+ , the adsorption behavior of methanol on the surface of Pt1/PN-CeO2 catalysts were investigated. As shown in Figure S21    Response: Thank you for your comments.
The concept of frustrated Lewis pairs (FLPs) on CeO2 surface has been developed experimentally and theoretically in our previous reports (Nat. Commun. 2017, 8, 15266). Due to that the detailed constructive process of FLPs has been discussed, we lacked of some necessary description in this manuscript. To make it easier for the reader to understand, we have added the relevant illustration in our revised manuscript, as following: As shown in Figure S1, the FLPs site is constructed by the two adjacent surface Ce 3+ as Lewis acidic site and the neighboring surface lattice oxygen as Lewis basic site on CeO2(110) surface. The constructive process of FLPs site has been also described in the revised manuscript. As shown in Figure 2a However, the electronic interaction between CeI and OIa/OIc will still hinder the activation of small molecules on the CeI-OIIc active sites ( Figure S2e). Therefore, the weak FLPs-like activation is obtained on the traditional CeO2 materials. When the second adjacent surface oxygen (OIa) is removed, the reduced Ce cations (CeI and CeII) and surface lattice oxygen (OIIc) are independent Lewis acidic and basis sites ( Figure S2c and S2f), respectively. Notably, two adjacent reduced surface Ce sites (CeI and CeII) and lattice OIIc is constructed the FLPs site of (CeI,CeII)-OIIc with a shorter distance (3.99 Å).
However, the construction of surface FLPs site cannot be realized by removing surface oxygen atoms on the CeO2(111) facet. Meanwhile, due to the low formation of oxygen vacancy on CeO2(100) surface, the relatively unstable oxygen defect leads to the spatial configuration for formation of FLPs sites. Therefore, CeO2 (110)   Unfortunately, it is hard to directly observe the FLP sites on the surface of CeO2 or other reported catalysts with FLPs performance. We are constantly trying various characterization techniques to directly observe or quantify the FLPs sites. Meanwhile, we are also actively seeking cooperation to overcome this challenge. However, so far, there is no satisfactory results. We will continue to focus on this research.
To expose the influence of FLPs sites on the catalytic performance, we have developed indirect evaluation parameter to evaluate the number of FLPs sites on various CeO2 surface. Based on the constructive process, the adjacent oxygen vacancy is the critical to form FLPs sites on CeO2 surface. Therefore, higher oxygen vacancy concentration inevitably results in a higher probability to form adjacent oxygen vacancy. Due to that the number of adjacent oxygen vacancy corresponds to the number of FLPs sites, the concentration of oxygen vacancy of CeO2 is used as a parameter to indirectly describe the number of FLPs sites.
We have added this relevant description in the revised manuscript. (Supporting Information, Figure S1 and Therefore, the Pt1/PN-CeO2 catalysts also exhibited the weakest bonding with CO molecule. However, due to the abundance of surface *OH species from FLPs sites, the rate of *CO reforming on single-atom Pt is still satisfactory compared with this transformation on Pt nanoparticles ( Figure S4). In contrast, the low d-band center of single-atom Pt on PN-CeO2 will exhibit the weak interaction with CO2 molecule, facilitating its desorption form single-atom Pt active sites.  To further confirm the adsorption of CO on Pt1/PN-CeO2, the CO-FTIR analysis was added in the revised manuscript. As shown in Figure S8, peaks at approximately 2090 cm -1 were associated with the linearly adsorbed CO on isolated ionic Pt 2+ , 1-3 revealing that the Pt active sites existed as single-atom on PN-CeO2 and coordinated with O atoms. Specifically, the two apparent peaks at 2099 cm -1 and 2076 cm -1 could be attributed to the single-atom Pt located in spatial configuration of the type I and type II (Figure 1d), respectively, owing to the lower valence state of Pt on oxygen vacancy adjacent to the FLPs sites than it on traditional oxygen vacancy ( Figure S3).
More details have been added in the revised manuscript, as following: "Then, the dispersion and chemical environments of Pt on PN-CeO2 were studied by diffuse-reflectance infrared 5. The surface area and porosity of CeO2 nanorods and Pt single atom nanorods CeO2 catalyst should be disclosed by BET analysis.

Response:
The surface area and porosity of PN-CeO2 and Pt1/PN-CeO2 were tested by BET analysis. The surface area of PN-CeO2 was 109 m 2 g -1 . As shown in Figure S6b, the main pore in PN-CeO2 was observed with pore size ~10 nm. Meanwhile, the pores with size of < 5nm was also observed. Importantly, the porous structural feature of PN-CeO2 was strongly confirmed combing with the dark-field TEM image. After loading of single-atom Pt, the Pt1/PN-CeO2 catalysts exhibited the similar nitrogen adsorption-desorption isotherm plot. The surface area of Pt1/PN-CeO2 increased slightly to 118 m 2 g -1 compared with the PN-CeO2 supports, which could be attributed to the new appeared pores with size of <2 nm. Therefore, the porous structural feature was not destroyed during loading of single-atom Pt.
The relevant description has been updated in our revised manuscript, as following: "The specific surface area of PN-CeO2 was 109 m 2 g -1 , as derived from the N2 adsorption/desorption isotherm plot (Figure 6a). The porous structure with a size of 1.5~3.0 nm was revealed from TEM images (Figure 5b) as well as the Brunauer-Emmett-Teller (BET) measurements (Figure 6b and 6c Figure S7 and Table S1). Therefore, the FLP sites could be formed on the PN-CeO2 supports owing to the high concentration of oxygen defect on the CeO2(110) surface, as described in our previous reports. 9,10 Then, the single-atom Pt anchored on PN-CeO2 (Pt1/PN-CeO2) with 0.36 wt.% loading was successfully synthesized through a photo-assisted deposition process due to the strong trapping of metal species on the defective sites of PN-CeO2 from the DFT calculations ( Figure S3). Both the specific surface area/pore structure ( Figure S6) and levels of surface oxygen defects ( Figure S7 Table S2. 7. The frustrated Lewis pairs in the PN-CeO2, NR-CeO2, and NP-CeO2 catalyst after loading of Pt is not studied.
So, it is hard to understand whether Pt influences the frustrated Lewis pairs formation or not.

Response:
The surface properties of PN-CeO2, NR-CeO2 and NP-CeO2 supports after loading of Pt was studies by XPS analysis, as shown in Figure S7, S11 and S12. Meanwhile, the surface fractions of Ce 3+ and Ce 3+ -O were also summarized in Table S1. Due to the reduction of H2, the surface defects of various CeO2 supports exhibited a slight increase. Take Pt1/PN-CeO2 catalysts as example, the surface Ce 3+ fraction increased from 30.8% to 34.7% compared with the PN-CeO2 supports, along with the increased surface Ce 4+ -O fraction. Therefore, the FLPs sites on Pt1/PN-CeO2 supports still existed and might be increased owing to the increased surface defects concentration.
Although the level of surface defects of Pt/NR-CeO2 also increased after the Pt-loading, the surface Ce 3+ fraction was still obviously lower than that of Pt1/PN-CeO2. Thus, the Pt1/PN-CeO2 catalysts also exhibited more amount of FLPs sites than the Pt/NR-CeO2 catalysts. In addition, there is no FLPs sites on the surface of Pt/NP-CeO2 due to the mismatch of crystal face.
More details could be found in Page 9, Page 13 and Page 14 (highlighted by yellow).    8. Why does a 20 % loss of catalytic performance after the stability test ( Figure 3C)? Is Pt single atom leaching out from CeO2 nanorods?
Response: After the hydrogen generation, the reaction solution was analyzed by ICP-OES. However, there was no Pt ions in the reaction solution. The surface Ce 3+ fraction along with the Ce 3+ -O fraction of the used Pt1/PN-CeO2 catalysts was similar as the as-synthesized Pt1/PN-CeO2 ( Figure S11), revealing the FLPs sites on PN-CeO2 was stability during the H2 generation. After careful analysis of the HAADF-STEM images ( Figure S10c), the Pt nanoclusters were observed on the surface of used Pt1/PN-CeO2 at 165 °C. Previous report has proved that the perimeter Pt active sites in the Pt-CeO2 catalytic system are the critical active sites for the reforming of *CO and remain dynamically mobile. 6 Therefore, the decrease in H2 generation rate of the Pt1/PN-CeO2 catalysts would be attributed to the increased size of Pt active sites owing to the possible mobility on the surface of catalysts.
We have modified the relevant description in our revised manuscript, as following: "  9. Besides, after the catalytic reactions, the existence of frustrated Lewis pairs sites should be studied.

Response:
After the hydrogenation generation, the Pt1/PN-CeO2 catalysts were separated from the reaction solution. Then, the used Pt1/PN-CeO2 catalysts were also analyzed by XPS. As shown in Figure S11, the surface Ce 3+ fraction was 32.7% along with the 47.3% of Ce 3+ -O fraction, similar with the values of as-synthesized Pt1/PN-CeO2 catalysts. Therefore, the absence of reduced surface oxygen defects suggested that the FLPs sites were still abundant on the used Pt1/PN-CeO2 surface.
The relevant description has been modified in our revised manuscript, as following: