Lead halide perovskites for photocatalytic organic synthesis

Nature is capable of storing solar energy in chemical bonds via photosynthesis through a series of C–C, C–O and C–N bond-forming reactions starting from CO2 and light. Direct capture of solar energy for organic synthesis is a promising approach. Lead (Pb)-halide perovskite solar cells reach 24.2% power conversion efficiency, rendering perovskite a unique type material for solar energy capture. We argue that photophysical properties of perovskites already proved for photovoltaics, also should be of interest in photoredox organic synthesis. Because the key aspects of these two applications are both relying on charge separation and transfer. Here we demonstrated that perovskites nanocrystals are exceptional candidates as photocatalysts for fundamental organic reactions, for example C–C, C–N and C–O bond-formations. Stability of CsPbBr3 in organic solvents and ease-of-tuning their bandedges garner perovskite a wider scope of organic substrate activations. Our low-cost, easy-to-process, highly-efficient, air-tolerant and bandedge-tunable perovskites may bring new breakthrough in organic chemistry.

. (a) The photograph for CsPbBr 3 P1 disperse in less polar solvents: Hexane, Toluene, 1,4-dioxane, ethyl acetate; polar solvents: Acetone, acetonitrile, DMF, and DMSO at ambient light (top) and 365 nm UV light (bottom) 3. Is there any possibility that the enhancement of PL intensity of perovskite NCs with acids is caused by binding the defects on perovskites?
Reply: We indeed appreciate this reviewer's comment on the binding of defects on perovskite. We noticed that bind of acid on perovskite may result in a better PL as well as elevated solar cell performance, according to reference Energy Environ. Sci., 2018,11, 3480-3490. In this reference, Lewis acid is added into perovskite film to significantly suppress defects through a synergistic effect, hence a better PL performance is indeed observed in this work. And finally such process leads to a highly efficient and stable perovskite solar cell. We therefore agree with this reviewer that there is a possibility that that the enhancement of PL intensity of perovskite NCs with acids is caused by binding the defects on perovskites. Related discussion is added in the manuscript.
In these regards, we really appreciate this reviewer's excellent comments to indeed enhance the quality of this work.
Regarding our recent published JACS, it is our initial exploration of perovskite towards photocatalytic α-alkylation of aldehydes. It is a first attempt that a very simple C-C bond formation reaction can be achieved using perovskite in high yield. However, as we discussed in the introduction, more general acceptance of perovskite toward photoredox reactions that are of broader interests is still unknown. Particularly, the challenging synthesis towards much more useful organic synthesis, i.e. drug molecules: such as ring-closure C-C bond forming, N-hetereocyclization, aromatic amination or aryl-esterification; or materials synthesis: polymerizations. We discuss the key parameters that are of interests from research fields both in perovskite materials and in organic chemistry. Such key parameters, include but not limited to, size effect, stability and tolerance, capping ligand effects, comparisons with known photocatalysts etc. More importantly, we have conducted experimental evidence to show that the bandgap-tuning via halides exchanging can be applied to activate different reactions and catalyze previously unachievable reactions. At the end of our recently published JACS paper, we briefly mentioned about the theoretical possibility that the band-tuning's effect on catalytic reaction may be achievable. However, so far as we know, band-tuning's real effects on photocatalytic reactions have not been experimentally discussed, nor explored. In this work, we are also devoted to unveiling the mystery of band-tuning of perovskite with regards to their photocatalytic ability. Therefore, we believe "the novelty and systematic study in this manuscript is at another level".
The comments mentioned from previous work on perovskite/NiO x /TiO 2 solar cell photocatalysts published on ACS Energy Lett regarding C-H activations (C-H oxidation to form aldehyde or alcohol) and benzylic alcohol oxidation are very interesting and related to our work. And we are very sorry to miss these references and are happy to add the related discussions. We also notice that these authors are very careful in terms of terminology. For example, such C-H activation or benzylic alcohol oxidation is not described as "organic synthesis" by these authors in these two ACS Energy Letter papers. In fact, these authors are very careful to describe the capability of such heterogeneous TiO 2 /perovskite system for organic chemical activation. One critical point is that the product yield they provide for C-H activation, for example, is much less than 1%, some examples are less than 0.1%, NOT practical for organic synthesis. Another important point needs to mention is that oxidation of organics (or organic pollutants) using metal oxide mixed/doped TiO 2 (i.e. NiO x in this paper) or visible light absorption dye-modified TiO 2 (i.e. perovskite in these papers) has been extremely common in the field of environmental science research area, but such oxidation can't be described or categorized as organic synthesis.
But we definitely agree with this reviewer that this type of work is relevant, and we would love to discuss and add these missing references. This type of discussion now is added in the revised manuscript. We do not think these papers will influence the fundamental insight of our work. But we greatly appreciate this reviewer's comments and would certainly love to address his/her comments one by one as demonstrated below.
1. This paper gives the reader a good view on the organic reactions that perovskite can perform, however no details about the mechanism is provided. I strongly suggest that the author should give us more insight in the reaction cycle, intermediates, active species… Reply: Thank you very much for this suggestion. We have the proposed mechanism discussion in our first version when submitted to Nat. Chem. We removed it because we think the proposed mechanisms for many of the discussed reactions may demonstrate no obvious difference comparing to Iridium and Ruthenium noble-metal photocatalysts. However, thanks to this reviewer's comments, we therefore re-investigate each and every type of the reactions on their respected mechanism. And we do find interesting difference and we also discovered, observed or isolated some important intermediates to prove the proposed mechanism. During the mechanism exploration, we employed a radical scavenger TEMPO to trap the related intermediate to prove our proposed mechanism. Here in this revised manuscript, we have added the mechanism discussion/comparison of C-C bond formation via C-H activations in the absence of air or in the presence of air in Scheme 1a. And we also added the proposed mechanism of N-heterocyclization in Scheme 1c with experimentally observed intermediate. In scheme 1b, we have added the verified key radical intermediate via TEMPO trapped experiment. The rest of the proposed mechanisms, i.e. C-O, C-N bonds formations, or polymerizations etc. in supporting information as shown in Scheme S1-S2. The revision is listed below: More importantly, oxygen may be of an essential component in certain photoredox organic reactions, in which perovskite may show a superior performance than oxygen sensitive photocatalysis as shown in Scheme 1. For instance, in C-C bond formations, 1d is achieved in nitrogen atmosphere while in a similar setup, air or oxygen atmosphere produces a ring-closure 1e (crystal structure provided). Oxygen is found to be the key reagent to be reduced as the hydrogen atom acceptor that further induced the C-H activation on phenyl rings. 40,41 As shown in Scheme 1, the reaction mechanisms are proposed in which the key radical intermediates have been investigated. For instance, perovskite PL quenching by 1e-A was observed and resulted in 1e-B radical in the presence of oxygen. Intermediate 1e-B and 1e-C have both been verified via radical trapping experiment employing 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO), a radical scavenger, through LC-MS (Scheme 1b). In the absence of oxygen, radical 1d-B has also been confirmed by TEMPO-trapped product and further verified with an isolated self-coupling compound 1d-C via 1 H NMR (SI). It is worth mentioning that the presence of air leads to more 1d-C formation and ultimately diminishes the yield of 1d. In Scheme 1c, under air condition for 2a and 2f formations, substrate quenching of perovskite also leads to the formation of 2a-A or 2f-A, corroborating with 2a-B molecule via TEMPO trapped experiment. Furthermore, the TEMPO trapped product for the radical intermediate 2f-B has also been observed in LC-MS (Scheme 2b). Other proposed mechanisms for C-O, C-N formation and ATRP in our perovskite photocatalysis under air are shown in Scheme S1-S2. However, the competition between oxygen as a quencher for molecular photocatalyst and oxygen as a reactant, is essentially problematic and leading to low yield or no reaction of such oxygen-involving reactions (Table 1)

Radical Trapping Experiments
The experimental procedure for trapping radicals with TEMPO 1) The trapping experiment for the synthesis of 1d. To a 4 mL vial, CsPbBr 3 NCs (1.0 mg), 2-phenyl-1,2,3,4-tetrahydroisoquinoline (0.25 mmol, 1.0 equiv.), 3-buten-2-one (0.5 mmol, 2.0 equiv.), TEMPO (0.5 mmol, 2.0 equiv.), trifluoroacetic acid (0.05 mmol, 20 mol%), and 2 mL DCM, the mixture was bubbled with N 2 for 10 min and then stirred. After irradiation with Blue LED for 8h, trace amount of 1d and 1d-C was isolated, while 1d-B-TEMPO was detected by LC-MS. The crude self-coupling product 1d-C was purified by column chromatography to afford a mixture of diastereoisomers with 1     2. In the second part of this paper, the author try to reveal the size effect of perovskite on the reaction. However, only two sizes, 6nm and 9nm, were tested. From the principle of statistics, it is difficult to obtain any information from this limited comparison.
Reply: Thanks for the comment. It is true that two sizes are not enough and we have synthesized two more new batches of QDs materials (4 nm, λ PL = 467 nm and 14 nm, λ PL = 525 nm) according to the reference (J. Phys. Chem. Lett. 2018, 9, 3093-3097) to further prove our conclusion that "small size NCs in general promote a faster reaction rate, but not necessarily a higher yield unless presenting in a perovskite friendly reaction environment". The UV-vis and PL spectra with these new batches are shown and updated in Fig. 2d, and the comparison for the reactions of 1b and 2a are shown in Fig. S4 in supporting information. And the related discussions are updated in the revised manuscript as shown below: In contrast, using a high temperature synthetic method, we also synthesized size-controlled CsPbBr3 NCs (P2, 4 nm, λ PL = 467 nm; P3, 6 nm, λ PL = 508 nm; P4, 9 nm, λ PL = 515 nm; P5, 14 nm, λ PL = 521 nm, Fig. 2b-d and Fig. S3). As shown in Fig. 2d, these NCs show a blue shift probably due to quantum confinements. Their photocatalytic ability has also been explored in the same reaction condition (catalyst loading 1.0 mg). For example, in C-H activation, at the early stage of the reaction, we find that smaller size NCs, i.e. P3-P5 show much higher initial reaction rate compare to the original synthesized P1 NCs. (Fig. S4-5). However, small size NCs' catalytic reactivity diminished quickly. When breaking a C-Br bond to form 1b, the reaction yield is recorded as 54-64% using P3-P5 in less than 40 min, and longer reaction time leads to a marginal increase of the yield of 1b. Much lower yield, ~8% was observed within P2 not only because of a faster deactivation using a smaller size, but also a significant blue shift leading to less visible absorption. Whereas using P1, the reaction rate is slower, however, the yield was observed to continuously increase and reach to 85% in ca. 5 hours.

Figure 2. (d) UV-vis and PL spectra of CsPbBr 3 P2-P5;
3. The perovskite is instability in CH 2 Cl 2 , as the author said it can be attributed to the ion exchange, XRD or XPS is a better method to prove it. "whereas no obvious PL changes are observed in non-halide solvents (Fig. S6)." but I do not see the relevant PL spectrum in Figure S6. In Fig 2f and 2g, the author try to use XRD to prove the ion change during the photocatalytic reaction, but the logic is too complex. It would be more intuitive if the author make a comparison on CsPbBr3 between before and after reaction.

Reply:
We are very sorry for this confusion and thank you for pointing this confused points out. a) As we mentioned in the manuscript, the perovskite is quite stable in CH 2 Cl 2 without light; however, with the irradiation of Blue LED, Br/Cl halide exchange happens. This observation is also corroborated with previous work (J. Am. Chem. Soc. 2017, 139, 4358-4361;Catal. Sci. Technol., 2018, 8, 4257-4263). Regarding the statement "whereas no obvious PL changes are observed in non-halide solvents ( Fig. S6)", we are very sorry, we only give the contrast images under UV lamp in original Figure S6a to indicate there is no change of the PL in hexane under the LED illumination. However, we did not show the PL spectra before and after the irradiation of blue LED. In the revised manuscript we have added the relevant PL spectra in hexane and 1,4-dioxane in Fig. S7c-d. The updated version included the missing information for the non-halogenated solvent as shown in Figure S7.  S7. (a) The photograph for CsPbBr 3 P3 before and after the irradiation of Blue LED for 2h in Hexane, CH 2 Cl 2 and CH 2 Br 2 at ambient light (top) and 365 nm UV light (bottom); (b) The UV-vis and PL spectra for P3 before the irradiation of LED and after irradiation in CH 2 Cl 2 for 1h; (c) The PL spectra for P3 before the irradiation of LED and after irradiation in Hexane for 5h; (d) The PL spectra for P3 before the irradiation of LED and after irradiation in 1,4-dioxane for 5h.
In addition, regarding the XRD, we are sorry for the confusion and complexity. We demonstrate the comparison before and after reaction in Fig. 2f and 2g. We are sorry for the complexity here. As shown in the updated Figure 2f and 2g, we record the XRD in three different time periods, the top one is the synthesized CsPbBr 3 catalysts before mixing with anything; the middle XRD is the isolated solid after mixing with substrate and co-catalyst but before illumination; the bottom one is the solid isolated after reaction. The middle XRDs indicate a fast anion-change with co-catalyst (ClCH 2 CH 2 ) 2 NH 2 Cl, forming a CsPbBr x Cl 3-x within both reaction 1b and 1c. However, the bottom XRDs are very different, 1b ending with CsPbBr 3 due to the chemical-equivalent byproduct Branions, and 1c ending with CsPbCl 3 due to the chemical-equivalent byproduct Clanions. Related discussions have been added in the manuscript to further clear the confusion. 4. The author use acid to enhance the PL intensity and obtain a conclusion that "non-halide organic acid may not only stabilize the perovskite NCs, but also may increase the overall catalytic efficiency for respective reactions (1a, 1d and 1e)". Can the author use photocatalytic reactions (1a, 1d and 1e) to prove this conclusion?
Reply: Thanks for this comment and we are sorry to miss the discussion of the correlation between amount of non-halide acid and the final catalytic efficiency. In fact, in our Table S1, S4 and S5. We demonstrate that no product or trace amount of the product were afforded without adding TFA. This indicated, in the catalytic mechanism that, the protons are the key component for the catalytic cycle to accomplish. Without acid, many reactions cannot proceed. This is also corroborated with our discussion at the beginning of this paragraph: "Acidity or free protons in perovskite reaction mixture may play a role in organic synthesis. For instance, carboxylic acids such as propionic acid, benzoic acid or trifluoroacetic acid (TFA), were used either as the co-catalyst (1a, 1d and 1e) or as a substrate (3a-3f)…" In addition, we also specifically explore the TFA amount with regards to our final catalytic efficiency and yield. Different amount of TFA in photocatalytic reaction (1a, 1d, and 1e) has been explored to prove our conclusion. Such data has been added and updated in the Table S1, S4 and S5. We find that suitable amount of TFA (catalytic amount for 1a, and ~1.0 equiv. for 1d and 1e) is essential for these reactions to proceed. Zero amount or too much amount of TFA in the reaction mixture will result in a significantly decreased reaction yield. This observation is also consistent with the decreasing of PL Intensity in too high concentration of TFA (Fig. 2h).
In this regard, we really appreciate this reviewer's comment to help to improve the quality of such discussion and the manuscript has also been revised below accordingly regarding this comment. "The maximum PL enhancement was observed using TFA at a concentration of ca. 6.5-13 mM, more acid leads to a diminishing PL probably because large number of protons may start to initiate a deactivation process. Interestingly, such optimized TFA concentration also leads to a maximum product yield of 1a, 1d and 1e as shown in Table S1, S4 and S5 respectively, indicating a high PL of the photocatalyst may increase the catalytic conversion. Therefore, non-halide organic acid may not only stabilize the perovskite NCs, but also may increase the overall catalytic efficiency for respective reactions." 5. From the stability test in SI, it seems like the stability of perovskite during photo reaction is too poor. P2 seems like loss the activity suddenly. what is happening at that point?
Reply: Thanks for the comment. It is true that the small perovskites particles in the quantum dots level that decompose very quickly and are too poor in reactivity. That is one of the differences that our photocatalytic (P1) system comparing to some other reported perovskite results. We noticed that P3 (previous version noted as P2) seems like a sudden loss of reactivity according to Figure S4a. One of the assumptions is that the time intervals are too wide in the current exploration, but the deactivation is too quick. We extracted a data point in every 20 minutes use NMR studies. And we basically find that it lose activity after 2 data points. We would like to know more information, particularly in between 20 min and 40 min. Therefore, we have re-test the 1 H NMR yield with regards to the time using P3, and we can extract one more data point in between at 10 min and 30 min. As a result, we find the yield is ~20% at 10 min, and ~50 % at 30 min which is close to saturated line. We re-plot the data points in Figure S4a and no sudden loss like previous figure was observed. Therefore, we do not think there is a sudden catalytic activity lose here, but because our data points are too few, and not enough to demonstrate such change. Hence, we have updated the Figure S4 with more experimental data points, 10, 20, 30, 40 min and now it has been showing that a graduate deactivation process and not a sudden loss, as shown below. We are sorry for this mis-leading information in the first version. And we indeed thank this reviewer very much for this helpful comment. 6. I am confused by the author's statement about the ligands. CdSe QD is stabilized by oleic acid, the author's sample still need ligand to stabilize, n-octylammonium, I do not see any advantage about this point.
Reply: we are really sorry for the confusion. This reviewer is right, both CdSe QDs and perovskite need to stabilize by capping ligands during the synthesis. After isolation, the capping ligands are still there. However, when using as a photocatalyst for photoredox reaction, as shown in the reference (J. Am. Chem. Soc. 2017, 139, 4250-4253), CdSe QDs need to be further stabilized with adding extra oleic acid and trioctylphosphine as ligand in ODE solution. The catalyst has to be loaded in ODE or in hexanes solution, otherwise reaction is not working. Such stabilization strategy for CdSe is very important and has been discussed exclusively in the paper with various stabilization ligands including A, stearate and trioctylphosphine; B, oleate and trioctylphosphine; C, oleate and diphenylphosphine. And they find the B give the best yield. In order to make this comparison more reasonable, we also conduct experiments. Simple stabilized with oleic acid using commercially available CdSe in 3.0 nm is not working or resulting in trace amount of product (Table S2, S6-S11). However, in this reference, the reason on the ligand's role, particularly phosphine's role has not been discussed/illustrated in this paper. In our perovskite system, the ligand role is more straightforward, it is used to for the colloidal solution to form in the DCM or other organic solvents. And such capping ligand (ammonium as A site) is part of the APbX 3 structure, hence more stable and therefore, we can directly load the catalyst in solid form without any pre-suspension. Here, we added experiments in Table S3-S5 to evaluate the capping ligand's role for our photocatalytic reaction and we find that capping ligand is not altering any catalytic results at all in our system. we revised our manuscript as shown below: For example, addition of trioctylphosphine and extra oleic acid was found to be essential to stabilize CdSe QDs during the reported catalytic reactions, otherwise very low or no yield of products can be obtained. 39 Discrepancy on various phosphine stabilizing ligands or carboxylic capping ligands leads to significant change on the product yield (12%-70%) using CdSe QDs. 39 In addition, introducing extra oleic acid may result in complexity of organic reactions, for example, competition in esterification is observed for reaction 3a. While changing capping ligand on perovskite plays little role in the final product yield as shown in Table S3-S5. The is probably because the capping ligands (e.g., n-octylammonium) that stabilize perovskite colloids are reported to function as A site to the perovskite APbX 3 structure, 31 hence no extra stabilization protocol is required using perovskite nanocrystal for photocatalysis. 7. In the last part, replacing Br with Cl and I to tune the bandgap has been described in many publications. Some contents are similar to the author's JACS paper.
Reply: Thanks for this comment. Indeed, the bandgap tuning via halides exchanging has been described in many publications, including our recent published JACS paper (briefly mentioned at the end regarding the theoretical possibility or envision regarding the band-tuning's effect on catalytic reaction). However, so far as we know, band tuning's experimental effects on photocatalytic reaction have not been discussed, nor explored. Thanks to this reviewer's comments. In this revised version, we have accordingly removed the similar contents on the theoretical envision, but focused on the new experimental discovery in this work. To our knowledge, here it is the first time to illustrate that band-tuning of perovskite that can result in significant difference in photocatalytic reaction outcomes. We demonstrate that intentional or in situ band-tuning experiments of CsPbBr 3 NCs successfully drive previously unachievable reactions, i.e. reactions showing in Figure 3d, 2i and 3f via halide exchange as discussed in our manuscript. Specifically, we first hypothesize that band-tuning via in-situ ion-exchange (PL blue-shift) of perovskite is similar to increase the triplet energy E T in Ir molecules, thus leading to a higher yield of product. And then we have conducted a systematic band-tuning experiment (via adding known amount of TMSCl to tune the bandgap) to confirm such correlation between the bandedges and the product yield.
More importantly, our hypothesis is transformative and also proved that such band-tuning is very powerful. It may also let the people to re-think the general assumption in organic chemistry that C-Cl bond are stronger than C-Br bond and hence harder to activate. For instance, such banding-tuning with halide exchange can activate C-Cl bond towards C-C bond formation reactions. Without band-tuning, a weaker C-Br bond, (α-bromoketone) is actually non-reactive in the same system. We therefore proved that band-tuning may result in an absolute discrepancy in photo-activation reactions.
Overall, we really appreciate the valuable suggestions and comments from the reviewers who helped us to improve the quality of this work. And we are also grateful for the opportunity to re-submit our manuscript. We hope this revised manuscript is acceptable for publication in "Nature Communications".
The points raised in the previous round of review have been satisfactorily addressed by the authors. I would therefore recommend its publication in Nature Communications at its current format.
Reviewer #3 (Remarks to the Author): I will start by saying that I agree with the premise that perovskites (bulk and nanostructured) should have a large role in photocatalysis for organic synthesis, given their relatively high stability under water-free conditions. However, this manuscript, while ambitious, reads like a condensed review, not a communication. The problem starts early on, where the questions posed are so broad that one can't possibly be expected to evaluate their answers in a single mansucript. Manuscripts, especially communications, should contain one main thesis --this manuscript contains 10 or more. I appreciate the scope of work here, but the condensation of so many reactions PLUS characterization of the perovskite nanocrystals under certain conditions (which should clearly be a separate paper) dilutes the discussion so dramatically that I cannot discern the major claims of the paper, other than that perovskites can perform some chemical reactions in high yield. What is missing here is (i) comments about reproducibility, (ii) thorough comparison of critical catalytic parameters (TON, TOF, driving force, etc) with the state-ofthe-art, (iii) discussion of catalyst loading, (iv)action spectra proving photocatalytic activity of the material and identifying the contributing populations within clearly heterogeneous samples, (iv) ANY precise discussion of mechanism.
Because of the condensed nature of the discussion, many conclusions are speculative and comparisons to, say QD or molecular photocatalysts, include statements that just aren't true in general. There's also an abundance of imprecise language --talking about, for example, "more catalytic sites" and "larger surface areas" of smaller NCs rather than the more correct surface area-to-volume ratio. Also the discussion of "deactivation" of the nanocrystals really contains no physical insight.
In summary, the authors make an definite impression that perovskites are promising materials for this application, but the data provided and the discussion do not, in my opinion, constitute a complete scientific study of any one reaction or any one catalytic material, so I cannot recommend this paper for publication in its current form.

Reviewer #1
Comments: The points raised in the previous round of review have been satisfactorily addressed by the authors. I would therefore recommend its publication in Nature Communications at its current format.

Reply:
We indeed thank this reviewer very much for the valuable comments and suggestions again.

Response to Reviewer 3
Reviewer #3: 1. I will start by saying that I agree with the premise that perovskites (bulk and nanostructured) should have a large role in photocatalysis for organic synthesis, given their relatively high stability under water-free conditions.

Reply:
We indeed thank this reviewer very much for the valuable comments and suggestions and we also would like to revise according to these suggestions and comments.
2. However, this manuscript, while ambitious, reads like a condensed review, not a communication. The problem starts early on, where the questions posed are so broad that one can't possibly be expected to evaluate their answers in a single manuscript. Manuscripts, especially communications, should contain one main thesis --this manuscript contains 10 or more. I appreciate the scope of work here, but the condensation of so many reactions PLUS characterization of the perovskite nanocrystals under certain conditions (which should clearly be a separate paper) dilutes the discussion so dramatically that I cannot discern the major claims of the paper, other than that perovskites can perform some chemical reactions in high yield.
Reply: Thank you for your comments on this point. We agree that too many reactions indeed distract the main thesis of this manuscript. And many certain condition discussions can be a separate paper. Accordingly, we focused on three types general reactions, C-C, C-N and C-O and also remove the distracted reactions: halide reduction, polymerization reactions. We note that these two types of reactions have not been actually discussed too much in the previous version at all. Inspiring by this reviewer's comment, we also note that the reaction 1a and 3g-3i in the old version also have NOT been thoroughly discussed. Although adding these reactions into the discussion seems likely expanding the perovskite's general catalytic ability or scope of the work towards photocatalysis, yet they do not provide essential points toward understanding the key chemistry here but introduce distractions. We take the advice and remove these distracted reactions, but still keep the three fundamental types of reactions in the category of C-C, C-N and C-O bond formation reactions. In this way, we may not only demonstrate the general acceptance of our perovskite towards photocatalytic activation of three fundamental types of organic reactions, but also present in a neat way to highlight the key catalytic parameters, mechanisms and unique properties of perovskite. In following discussion after Figure 1, we also categorized such reactions into three types and focus on discussion of mechanisms using perovskite comparing to other photocatalysts. We hope our revised manuscript is in an acceptable manner to demonstrate the key point that the Lead Halide Perovskites is truly a potential photocatalyst candidate for general acceptance in fundamental organic synthesis.
3. What is missing here is (i) comments about reproducibility, (ii) thorough comparison of critical catalytic parameters (TON, TOF, driving force, etc) with the state-of-the-art, (iii) discussion of catalyst loading, (iv) action spectra proving photocatalytic activity of the material and identifying the contributing populations within clearly heterogeneous samples, (v) ANY precise discussion of mechanism.
Reply: We thank this reviewer very much for these insightful comments on the key parameters. These comments indeed helped us to improve the quality of this work. Accordingly, here in the revised manuscript we have added the discussion of the following: To test the reproducibility, we conducted each type of reactions, 1a, 1c, 1d, 2a, 2g, 3f for three times in their corresponding optimized conditions using the solvent and materials treatment in SI. The respective data has been updated in Fig. 1 and Table S8 was also listed in the new version. The results show that our photocatalyst leads to a reasonable reproducibility for C-C, C-N, C-O bond formation reactions under such optimized conditions.   (Yields of 1a, 1c, 1d, 2a, 2g, 3a are the average yields of three times reactions details in Table S8; Inset: perspective view of 1d's single crystal structure with the thermal ellipsoids drawn at 50% probability level and the H atoms omitted for clarity.) ii) Critical catalytic parameters of TON and driving force have been explored and their comparison with reported photocatalysts have also been illustrated as followed. Accordingly, we have explored the TON as suggested and the following part has been updated in the manuscript as well as in SI: "Catalytic turnover number (TON) is compared and listed in Table 1. Heterogeneous catalyst, i.e. 3.0 nm CdSe QDs were reported to optimally render a TON of 79,100 (based on QD's molecular weight Mw, 88,000 g/mol) in glove box. 39 However, in our condition under air, no yield (nor TON) of 1, 2 and 3 can be obtained using CdSe QDs. In addition to air-sensitivity, CdSe's performance was also dependent on size and capping ligands. 39 While changing capping ligand on perovskite plays little role in the yield as shown in Table S2-S4. This is probably because the capping ligands (e.g., n-octylammonium) that stabilize perovskite colloids are reported to function as A site to the perovskite APbX 3 structure, 31 hence no extra stabilization protocol is required using perovskite nanocrystal for photocatalysis. Using the method in CdSe QDs 39 to calculate TON, P2 NCs (14 nm, based on Mw, 8,015,000 g/mol, P1-P5 TON details see Table S9) renders 2,565,000. Perovskites' heterogeneous catalytic ability is validated via regaining strong PL after recovering the catalyst via centrifuge after reaction (Fig. S7). To compare TON with molecular catalysts, TON calculation based on mole of metal (independent of size, CsPbBr 3 , 579.8 g/mol) was carried out instead. For instance, four cycles of the reactions render a TON of at least 9,100 for 1a (Table  1, details see SI). Overall, one or two orders of higher TONs under our condition are observed using perovskite than others, except reaction 3, in which TON may rely on both perovskite and Ni co-catalyst."

Driving force:
"To further elucidate the reaction mechanism, electrochemical studies were conducted. (Fig. S24-31) According to the comparison between redox potentials of the key substrates and the band energy of perovskite, the respective driving force is listed in Figure 3. Driving force for HT in reaction 1c, 1d and 2a is observed among ~ 0.1 to 0.3 eV, consistent with the Stern-Volmer quenching results (Fig S11-17) as well as the mechanistically verified intermediates in Scheme 1. However, 2g-B disfavors HT due to a more positive oxidation potential (E ox , 1.42V), corroborating with the previous observation that direct radical forming from 2g-B is difficult, unlike reaction 2a pathway. Moreover, driving force for ET is listed from ~0.2 to 0.5 eV, confirming our discussion on ET in Scheme 1. However, noticeable exception, 2,4'dichloroacetophenone, though presenting a more negative reduction potential (E red , -1.47V), reacts to form respective pyrrole. We postulate that in-situ band-tuning of perovskite may play a role here and is discussed below." (iii) Catalyst loading exploration has also been explored and added in the revised manuscript as shown in Figure 1 as well as in the discussion context of the manuscript.
The loading exploration and their comparison has been also showing in Table S1-S7. "Catalyst loading has also been explored (Table S1-S7) and respective minimum loading for typical reactions of ~ 0.1-0.5 mmol has been listed in Fig. 1. These reactions result in respective products in moderate to high yields without need for anaerobic sparging." (iv) We have conducted further spectroscopy investigation including PL and their related Stern-Volmer quenching studies to illustrate the heterogeneous nature of NCs catalyst. As shown in Fig. S7, we found that the recycled photocatalyst CsPbBr 3 (after a completely reaction cycle) in EtOAc is still emissive. Recycling of these NCs via centrifuge (inset) indicated the strong PL of the perovskite. Reloading such catalyst leads to the second cycle of the catalytic reactions. Moreover, the Stern-Volmer quenching studies also support our proposed mechanism as shown below.     CsPbBr 3 NCs Emission quenching by (E)-1-benzylidene-2phenylhydrazine. k q = 8.8 ×10 9 M -1 s -1 .
We have spent most of time in this period to explore the key mechanism of these fundamental reactions. In fact, the exploration of the first four points have substantially helped us to address the last key point: mechanism. For example, the driving force is indeed helping us to understand the charge transfer process as the key step for electrontransfer (ET) or hole-transfer (HT). We have thoroughly conducted the electrochemical experiments to illustrate such driving force. First, for the mechanism exploration, we have distinguished oxygen's role via its presence or absence in C-C bond formations.
More importantly, during the revision, we have also distinguished the mechanism in C-N bond formations, via comparison of the initial radical formation pathway. The driving force is corroborating with most of the key steps that we proposed in terms of radical formation in scheme 1. Such study demonstrated the differences of possible HT between the formation of pyrrole and pyrazole due to the different oxidation level of substrates. Such driving force result (a completely different observation on HT for reaction 2a and 2g) encouraged us to further explore the key intermediate because there must be something different between pyrrole and pyrazole formations. Interestingly, but not surprisingly, we do successfully identify another key intermediate, directly HT transfer forming 2a-C (instead of 2g-C) via TEMPO-trapping experiment. Such result has been added in manuscript and in SI as well shown as below. Here we really appreciate reviewer 3's comments that help illustrate the different pathway of pyrrole and pyrazole formations.  Overall, according to TEMPO trapping experiment on the key intermediates, the Stern-Volmer quenching results, and the determined redox potentials of the key substrates and possible driving force for the charge transfer, we rationalized our mechanism as followed. "Mechanism. Oxygen may be of an essential component in certain photoredox reactions. For instance in Scheme 1a, radical addition product 1c is achieved in nitrogen atmosphere while in a similar setup, air or oxygen atmosphere produces a ring-closure 1d (crystal structure provided in Fig. 1). Oxygen is found to be the key reagent as the hydrogen atom acceptor that further induced the C-H activation on phenyl rings. 40,41 As shown in Scheme 1, the reaction mechanisms are proposed in which the key radical intermediates have been investigated. Upon Stern-Volmer PL quenching studies (Fig. S11-17), perovskite PL quenching by 1d-A was observed (k q = 3.6 ×10 8 M -1 s -1 , Fig.  S12) and resulted in 1d-B radical in the presence of oxygen. Intermediate 1d-B and 1d-C have been verified via radical trapping experiment employing 2,2,6,6-tetramethyl-1piperidinyloxy (TEMPO) as a radical scavenger, through LC-MS (Fig. S19-20). In the absence of oxygen, radical 1c-B is also confirmed by TEMPO-trapped product (Fig. S18) and further verified by the self-coupling diastereoisomers 1c-C via 1 H NMR (see SI). It is worth mentioning that the presence of air leads to more 1c-C formation and ultimately diminishes the yield of 1c.

4) The trapping experiment of 2a-C.
Scheme 1b shows the proposed mechanism of C-N formations, in which both oxidative (ET, 2a-A) and reductive quenching product (HT, 2a-B) in reaction 2a have been trapped by TEMPO (either observed via 1 H NMR or LCMS), indicating a strong charge separation and transfer ability induced by perovskite. This pathway is similar to our previous mechanism exploration in α-alkylation of aldehydes. 17 Radical coupling between 2a-A and 2a-C leads to the intermediate of 2a-D. Then C-N formation via intramolecular cyclization and a final dehydration leads to the pyrazole product 2a. In contrast, the radical formation from 2g-B via direct HT has not been observed, instead 2g-C was verified via radical-trapping, likely demonstrating a different mechanism of pyrrole formation as shown in Scheme 1b. Mechanism of reaction 3a is also proposed and shown in Scheme S1 similar to previous reported mechansim. 45 To further elucidate the reaction mechanism, electrochemical studies were conducted. (Fig. S24-31) According to the comparison between redox potentials of the key substrates and the band energy of perovskite, the respective driving force is listed in Figure 3. Driving force for HT in reaction 1c, 1d and 2a is observed among ~ 0.1 to 0.3 eV, consistent with the Stern-Volmer quenching results (Fig S11-17) as well as the mechanistically verified intermediates in Scheme 1. However, 2g-B disfavors HT due to a more positive oxidation potential (E ox , 1.42V), corroborating with the previous observation that direct radical forming from 2g-B is difficult, unlike reaction 2a pathway. Moreover, driving force for ET is also listed from ~0.2 to 0.5 eV, confirming our discussion on ET in Scheme 1. However, noticeable exception, 2,4'dichloroacetophenone, though presenting a more negative reduction potential (E red , -1.47V), still reacts to form respective pyrrole. We postulate that in-situ band-tuning of perovskite may play a role here and is discussed below." 4. Because of the condensed nature of the discussion, many conclusions are speculative and comparisons to, say QD or molecular photocatalysts, include statements that just aren't true in general. There's also an abundance of imprecise language -talking about, for example, "more catalytic sites" and "larger surface areas" of smaller NCs rather than the more correct surface area-to-volume ratio. Also the discussion of "deactivation" of the nanocrystals really contains no physical insight. Reply: We really appreciate this reviewer's comments in this regards that would certainly help us to improve the quality of this work. Imprecise language regarding "the catalytic site" and "surface area" has also been corrected with surface area-to-volume ratio. And the discussion of "deactivation" has been revised accordingly. Such details are listed in SI and as showed below.

Calculation of NC cubic edge length d for P2-P5:
According to the well-established size-dependent absorbance spectrum of CsPbBr 3 NC from ref S3.

Calculation of Surface area-to-volume ratio:
Surface area of one NC = 6d 2 ; volume of a NC = d 3 (nm 3 ) Thus, surface area-to-volume ratio: 6/d (nm -1 ) We have also carefully revised speculative comment and comparisons. And in the new version, particularly regarding discussion of QDs, we have revised as below: "Catalytic turnover number (TON) is compared and listed in Table 1. Heterogeneous catalyst, i.e. 3.0 nm CdSe QDs were reported to optimally render a TON of 79,100 (based on QD's molecular weight Mw, 88,000 g/mol) in glove box. 39 However, in our condition under air, no yield (nor TON) of 1, 2 and 3 can be obtained using CdSe QDs. In addition to air-sensitivity, CdSe's performance was also dependent on size and capping ligands. 39 While changing capping ligand on perovskite plays little role in the yield as shown in Table S2-S4. This is probably because the capping ligands (e.g., n-octylammonium) that stabilize perovskite colloids are reported to function as A site to the perovskite APbX 3 structure, 31 hence no extra stabilization protocol is required using perovskite nanocrystal for photocatalysis. Using the method in CdSe QDs 39 to calculate TON, P2 NCs (14 nm, based on Mw, 8,015,000 g/mol, P1-P5 TON details see Table S9) renders 2,565,000. Perovskites' heterogeneous catalytic ability is validated via regaining strong PL after recovering the catalyst via centrifuge after reaction (Fig. S7). …" 5. In summary, the authors make a definite impression that perovskites are promising materials for this application, but the data provided and the discussion do not, in my opinion, constitute a complete scientific study of any one reaction or any one catalytic material, so I cannot recommend this paper for publication in its current form.

Reply:
We really appreciate this reviewer's valuable comments that indeed helped us to understand the mechanism and help us to improve the quality of manuscript. We hope that the revised work is at a stage to support publication in Nature Communications.