Templated growth of oriented layered hybrid perovskites on 3D-like perovskites

The manipulation of crystal orientation from the thermodynamic equilibrium states is desired in layered hybrid perovskite films to direct charge transport and enhance the perovskite devices performance. Here we report a templated growth mechanism of layered perovskites from 3D-like perovskites which can be a general design rule to align layered perovskites along the out-of-plane direction in films made by both spin-coating and scalable blading process. The method involves suppressing the nucleation of both layered and 3D perovskites inside the perovskite solution using additional ammonium halide salts, which forces the film formation starts from solution surface. The fast drying of solvent at liquid surface leaves 3D-like perovskites which surprisingly templates the growth of layered perovskites, enabled by the periodic corner-sharing octahedra networks on the surface of 3D-like perovskites. This discovery provides deep insights into the nucleation behavior of octahedra-array-based perovskite materials, representing a general strategy to manipulate the orientation of layered perovskites.

occur during solution thinning, after spin-coating or during thermal annealing? 2. The PDS (PbI2-DMF solvated phase) plays a critical role in the growth of oriented layered perovskites. Could the DMF in PDS be other solvent molecules, e.g. DMSO?
3. Ammounium halide salts such as NH4Cl improve the solubility of PDS, which has been proved by the authors. Could the authors further explain why these salts can improve the solubility of PDS? 4. As reported in literature, e.g. Adv. Energy Mater. 2018, 8 (21), 1800185 and J. Am. Chem. Soc. 2017, 139, 1432-1435, RP perovskite films prepared from nominal n values are actually mixtures with various RP phases. These RP perovskite films show graded phase distribution even with out-of-plane orientation. Supplementary Figure 17 in this work also shows a similar observation. Could the authors comment on the growth mechanism of graded RP phases with out-of-plane orientation?

Reviewer #3 (Remarks to the Author):
The manuscript by Wang et al. presents a thorough investigation of the growth mechanism for oriented layered perovskites. The main finding is that the firstly formed solvated perovskites in solution can template the sequential growth of layered perovskites, the orientation of layered perovskites is determined by the lattice matching between solvated perovskite and layered perovskites. In my opinion, since H. Tsai et al. reported hot-cast RP-type layered perovskites with out-of-plane orientation (Nature 536, 312, (2016)), the detailed nucleation process of layered perovskites is still unclear. The finding in this manuscript is very exciting to the 2D perovskite field, because the authors clarified several unsolved questions with this templated growth mechanism, especially how the out-of-plane orientation was formed. Moreover, it was well demonstrated that the formation of solvated perovskites can be engineered, which could be a useful method to achieve layered perovskite films with out-of-plane orientation and hence improve device performance. The manuscript is well written and organized with ample relevant references. I consider this work to be suitable for the broad readership in Nature Communications and I would recommend publication of this manuscript after revision.
The following are some specific concerns: 1)According to some previous results (e.g. K. Yan et al. J. Am. Chem. Soc. 137, 4460 (2015)), the perovskite precursor solutions are generally colloidal dispersions, so I am wondering whether the excessive NH4Cl also impacted on the colloidal size and hence tuned the grain size and final morphology of the layered perovskite films?
2)It is impressive that the out-of-plane orientation and high crystallinity of the RP perovskites can be equally achieved in a group of films with excessive AX salts (A = NH4+ or MA+; X = Cl-, Br-or I-) as additives. So how about the device performance for solar cells resulted from precursor solutions with other AX salts than NH4Cl?
3)In Fig. S14a, the XRD pattern of RP films with different amounts of NH4Cl addition was shown, the absence of diffraction peaks below 10° when the molar ratio of NH4Cl: PbI2>0.3 was believed to be due to the dominating OP orientation induced by NH4Cl additive. However, there is an exception for 1.0 NH4Cl, in which some extra diffraction peaks show up. Could the authors provide some discussion on this?

Response to Referees Letter
Referee #1: Comment 1: The work titled "Templated Growth of Oriented Layered Hybrid Perovskites on Solvated Perovskites" aims to study the mechanism of 2D perovskites growth and to understand how the orientation of layered perovskites is controlled with different treatments. The authors first study the effect of NH 4 Cl on the orientation of 2D perovskite thin films. Then the authors investigate the growth of single crystals and the effect of a PbI 2 -DMF solvate phase on the growth and orientation of the crystals, and strive to connect the observations to the mechanisms that control 2D perovskite orientation in perovskite films. Finally, the authors fabricate solar cells with the goal of corroborating their findings.
This reviewer found that there were many interesting observations in this work. However, there are also oversights, unwarranted assumptions, and a lack of flow throughout the manuscript. In its present state the manuscript cannot be accepted for publication in Nature Communications. With the appropriate major revisions, however, it might become suitable.
Here are major items requiring attention: The authors give no motivation for studying NH 4 Cl treatment. Has this been studied before? If so, why are you studying it again? The NH 4 Cl work seems very disjointed relative to the rest of the manuscript and there is no convincing argument for the connection between this and the PbI 2 -DMF solvate work. If the authors insist on including the NH 4 Cl work it must be introduced, motivated, and connected with the rest of the work.
Response 1: We appreciate the referee for the insightful comments which helped us to improve the quality of this manuscript tremendously. We also thank the referee's recognition of the novelty of this work, and think this manuscript can be accepted after revisions. We followed the referee's suggestion to enhance our proving as will be shown in the following point-by-point responses.
The major contribution of this work is to understand the nucleation and alignment of layer perovskites. We revealed that, in order to make layered perovskite grow directionally, a surface initialized crystallization is needed, which only become dominating when the nucleation inside the solution is suppressed and so that nucleation starts at liquid surface. Due to this reason, NH 4 Cl additive is necessary in our work. As we mentioned in the manuscript and will discuss below in this response, NH 4 Cl additive played an important role in the suppression of the precipitation of PbI 2 -DMF and MAI-PbI 2 -DMF solvated phase phases, which guarantee the downward growth of RP perovskite from liquid surface as dominating mechanism. Without employing NH 4 Cl as additive, the templated growth of RP perovskite from quasi-3D perovskite can still happen, but sole templated growth mechanism does not result in RP perovskite films with high crystallinity and well aligned crystal orientations at RT. In another word, we agree that employing NH 4 Cl additives has been reported previously mainly in MAPbI 3 -based 3D perovskites (e.g. Ding, L. and et al., Nanoscale 6, 9935-9938 (2014);Chen, T. and et al., J. Mater. Chem. A 3, 18514-18520 (2015); Liang, Z. and et al., Chem. Mater. 27, 1448-1451(2015). However, to the best of our knowledge, the influence of the NH4Cl additives on the nucleation of PbI2-based solvated phase in solution has never been noticed and discussed before. We firstly discovered the role of NH4Cl additive in this study. Since the nucleation engineering is crucial for the control of the final morphology of RP perovskites, we further broadened the idea that using additives like NH 4 Cl, MACl, MAI, MABr or NH 4 I can be general ways to engineer the nucleation of solvated phase and hence the RP perovskites. It's worthy to mention that our study also demonstrated the good working stability (Fig. 5c) of the RP perovskite solar cells fabricated with NH 4 Cl additive method, which has never been discussed before, to the best of our knowledge. On the other hand, using NH 4 Cl as additives is a very promising method for fabricating solar cells with comparable high PCE (see Supplementary Figure  S18 of the revised SI) with that of hot casting method (Tsai, H. et al. Nature 536, 312-316 (2016); Zhang, X. et al. Energ. Environ. Sci. 10, 2095-2102(2017), which makes our conclusion representative.
So logically, introducing NH4Cl treatment is necessary, which is closely connected with other analysis in our manuscript. We prefer to remain it in our manuscript. As the referee pointed out, we missed the motivation of using NH 4 Cl as additive in our previous manuscript. We accepted this suggestion and made some revises as follows: In page 4, last paragraph, we add a sentence: "As inspired by the success of NH 4 Cl additive in promoting the crystallinity of 3D perovskites, recently NH 4 Cl additive was also employed in RP perovskite for grains with OP orientation". Response 2: We thank referee for the comments and suggestions.
By following the referee's suggestion, we compared the absorption spectra of BA-based layered perovskite films fabricated from precursor solution with and without NH 4 Cl, as it is shown in Supplementary Figure 1 of revised SI file. The absorption spectra of sample with NH 4 Cl as additives show exciton absorption peaks located at 605, 640 and 668 nm (Supplementary Figure 1) which can be assigned to layered perovskite with n = 3, 4 and 5, respectively. The absorption peaks agree well with other papers (Stoumpos, C. C. et al. Chem. Mater. 28, 2852-2867(2016; Quintero-Bermudez, R. et al. Nat. Mater., 17, 900-907 (2018)). We can see sample with NH 4 Cl additive show relatively higher content of layered perovskites phase with n=3-5. On the contrast, sample without NH 4 Cl additive show higher content of layered perovskites phase with high n value (with absorption around 700~770 nm) and low n value phase (e.g. n=2 at 567 nm), but less content in n=3-5. We did not assuming that both the control and NH 4 Cl treated film has the same phase of 2D perovskite in our manuscript. The result is that using NH 4 Cl additive can suppress the nonuniform distribution of layer number (n), other than increase the average layer number (because there were no NH 4 + and Clions remained after annealing as we will prove it below).
Accordingly, in our revised Supplementary  For a better demonstration of that the photocurrent in layered perovskite (w NH 4 Cl additive) is mainly contributed by low-n value (e.g. n<5) phase (other than high-n value phase), we further compared the EQE spectra of perovskite solar cells based on layered perovskite (BA <n>=4) and 3D perovskite with the same solar cells structure (i.e. ITO/PEDOT:PSS/perovskite/PCBM/C60/BCP/Cu) in Figure R1. The EQE drop dramatically to be less than 40% in the region of wavelength > 700 nm, which is significantly lower than that of MAPbI 3 based 3D perovskites.  contributions in photocurrent are layered perovskites with low n value. We believe that the absorption and EQE spectra had provide enough evidence for the existence of low-n layered perovskite as dominating component, since the NH 4 Cl additive does not remain in final layered perovskite films. BA <n>=4 w NH 4 Cl glass identified when light was incident from glass side. The PL is dominated by the peak at 750 nm, mainly contributed by materials phase with low band gap. So it's not straightforward for us to estimate the amount of layered perovskite with low-n value from photoluminescence spectra. We agree that transient absorption spectroscopy can provide more information in the phase distribution of 2D perovskite (Quintero-Bermudez, R. et al. Nat. Mater., 17, 900-907 (2018)), but such facility is not accessible for us at this moment.
On the other hand, we want to mention that NH 4 Cl do not increase the amount of 3D perovskite or high-n layered perovskite phases because it does not remained in the final perovskite film. From the tolerance factor point of view, the incorporation of NH 4 + and Clions into the layered perovskites or 3D perovskite phase is thermal dynamically unstable due to their small ion radius ( (2015)). For further proof, we carried out XPS spectra analysis (see Figure Figure 1a) together with the absence of diffraction peaks along q z axis in the range of 0~10 nm-1 (Fig. 1b) suggests a dominating OP orientations…" As for the unequivocal and unconventional the XRD labelling, the 4 numbers labels (e.g. 2102 or 0102) were actually 3 number (i.e. (2 10 2) or (0 10 2)). In order to avoid misunderstanding, we added space in all the labels in our revised manuscript.

Comment 4:
The language throughout the manuscript is more definitive than is justified by the evidence/arguments. Some of the observations made may support the hypothesis of the authors, but it is not absolute evidence for it. The authors should point to any evidence that could support their claims instead of making unsubstantiated claims.
I list some examples: The authors claim that the 750 nm absorption is evidence for a solvated perovskite? What is the evidence for this? This could also be bulk perovskite. DMF is known to evaporate fairly easily at room temperature.
In our revised manuscript, for a better accuracy, we changed the term of "solvated perovskite" to "3D-like perovskite" since the black color phase is similar to 3D perovskite. We use the term of "3D-like" because the corner-sharing PbI 6 octahedra networks (i.e. the black colored phase) is not complete as compared to ideal bulk perovskite due to the absorption of DMF molecules and less continuous in its octahedra network. Here we thank referee for the constructive comments. Response 5: We thank the referee for the insightful comment. We do not exclude the formation of MAI-PbI 2 -DMF phase during the precipitation process. We collected the precipitation from precursor solution (formed by injecting CB anti-solvent as shown in Figure 2h and Supplementary  Figure 9) and carried out XRD studies. We confirm the presence of (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) together with PbI 2 -DMF as shown in the Supplementary Figure 4c. So we agree that the precipitate can be a mixture of PbI 2 -DMF and MAI-PbI 2 -DMF phases. This is reasonable because the (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) is the intermediate phase for PbI 2 -DMF phase converted to bulk perovskites (Petrov, A. A. et al. J. Phys. Chem. C, 121, 20739-20743 (2017);Jung, M. et al. Chem. Soc. Rev., 48, 2011(2019). On the other hand, in the perovskite film spin coated from precursor solution without NH 4 Cl as additives, i.e. the PbI 2 -based precipitation will become severe, excessive PbI 2 -DMF and (MA) 2 (DMF) 2 Pb 3 I 8 was found remaining in the final film, which lead to much increased XRD peak intensity around 2θ=9.63° and 2θ=6.63°, respectively (Supplementary Figure 10). In the revised manuscript, we define "PbI 2 -DMF-contained solvated phase" as "PDS", which involves PbI 2 -DMF and MAI-PbI 2 -DMF phases. The formation of MAI-PbI 2 -DMF phase didn't change our conclusion on the templated growth of 2D perovskite. MAI-PbI 2 -DMF phase like (MA) 2 (DMF) 2 Pb 2 I 6 and (MA) 2 (DMF) 2 Pb 3 I 8 are also needle-like crystal (Petrov, A. A. et al. J. Phys. Chem. C, 121, 20739-20743 (2017)). As shown in Figure  Supplementary Figure 8, we intentionally soak (MA) 2 (DMF) 2 Pb 3 I 8 fibers into the oversaturated BA based RP perovskite (<n>=2) solution (similar to the experiment shown in Figure 2 of main text), same directional growth of layered perovskite from the (MA) 2 (DMF) 2 Pb 3 I 8 fibers can be observed. This is because MAI-PbI 2 -DMF phase also lead to the formation of 3D-like perovskite on its surface (black colored in Supplementary Figure 8) which is the key issue to trigger directional growth of layered perovskites.

Supplementary
For a better accuracy, we follow referee's suggestion and modified our description as follows: In page 6, paragraph 2 of our revised manuscript, we modified the sentence of "by forming one-dimensional PbI 2 -DMF solvate phase (PDS)" to be "by forming one-dimensional PbI 2 -DMF-contained solvate phases (PDS)" In page 6, paragraph 2 of our revised manuscript, we add a sentence of "The PDS formed in MAI-rich solution can be a mixture of PbI 2 -DMF and (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) phases (see Supplementary Figure 4 and Supplementary Note 2), the latter of which has been reported to be the intermediate phase for the formation of perovskites" In page 8, paragraph 1 of our revised manuscript, we add a sentence of "Replacing the PbI 2 -DMF phase in Fig   Comment 6: The authors state that lattice-matching defines orientation. The authors do not have the evidence to make such strong claims.

Response 6:
We thank the referee for the comment. Generally, observing the quasi-3D perovskite/layered perovskite interface with high resolution transmission electron microscope (HRTEM) would be a straightforward method to demonstrate the lattice matching assumption. However, unfortunately, the facility is not accessible for us at this moment. So we conducted the templated growth of 2D perovskite (T-RP) on 3D MAPbI 3 single crystal (3D-SC) to observe how the facets of 3D-SC (i.e. with different lattice constant) impact on the orientation of macro RP perovskite crystals and verify the lattice matching: Similar to the templated growth experiment shown in Figure 2d,g, templated growth of BA-based RP perovskite crystals on different facets of 3D-SC has been investigated by replacing PDS fibers with MAPbI 3 3D-SC. [redacted] Accordingly, we found the T-RP also shows two typical kinds of orientation: i), On the facets with (1 1 2), (2 0 0) or (0 2 0) planes are exposed (i.e. the edges of PbI 6 octahedral pointing out), the T-RP sheets stand on the surface with a titling angle of ~45° to the terminal planes of 3D-SC (either be left or right titled). [redacted] ii), On the facets with (1 1 0) or (0 0 2) planes are exposed (i.e. the corners of PbI 6 octahedral pointing out), the T-RP sheets stand vertically on the surface. [redacted] [ Figure redacted] [redacted] As a conclusion, the templated growth of RP perovskite on 3D-SC experiment clearly demonstrated that the orientation of the [PbI 6 ] 4octahedral on 3D perovskite surface greatly determined the orientation of the [PbI 6 ] 4octahedral on RP perovskite, the lattice matching between 3D perovskite and layered perovskite play the key role in the final orientation of RP perovskites. We hope these additional studies can convince the referee. [redacted]

Comment 7. The authors state that the lowest energy configuration between the PbI 2 -phase and liquid is along the long side. Again, what evidence is there for this?
Response 7: Thanks for the comment. To confirm the lowest energy configuration, we intentionally dropped some needle-like PbI 2 -DMF crystals in micrometer scale on the top of the saturated RP perovskite precursor solution for the optical observation. As shown in Figure R8, the needle-like PbI 2 -DMF crystals lying on the liquid top with its length direction parallel to liquid/air interface. This is reasonable if taking the surface tension into consideration, because generally liquid tends to minimize its surface area due to its surface tension, while the lying of PbI 2 -DMF crystals on top of liquid helps to minimize the surface area of the liquid phase than other possible configurations as illustrated in Figure R8c.  In page 8, paragraph 1 of our revised manuscript, we changed the sentence of "The high speed warranties the formation of RP perovskites within one second" to be "The high speed enables the formation of RP perovskites within one second".
The language in our revised manuscript was double checked now.

Reviewer #2:
Comment 1: In this manuscript, Wang et al. report the templated growth of layered perovskites with out-of-plane orientation on solvated perovskites. The growth mechanism demonstrated in this work is reasonably proved by experimental data, which provides insightful understanding on the growth of vertically oriented layered perovskite films using perovskite precursor solutions with additional ammonium halide salts. In addition, the presented RP solar cells show device performance comparable to the best of the reported layered solar cells. The manuscript is well written and would be of great interest to the perovskite community. Therefore, I strongly recommend the publication of this work in Nature Communications. The following comments are provided to the authors to further strengthen the manuscript.
When soaked in oversaturated RP precursor solution, the surface of PDS powders turn black in a few seconds (Supplementary Figure 5). Then the surface of the PDS induces the growth of oriented RP perovskites. However, it is unclear when the PDS totally turns into perovskite phase. Does it occur during solution thinning, after spin-coating or during thermal annealing?
Response 1: We appreciate the referee's very positive comments and suggestions, which helped us to improve the quality of this manuscript tremendously.
The "PDS" in our previous manuscript represent PbI 2 -DMF phase. In the revised manuscript, the (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) phases are also involved in our analysis. So here the PDS include the PbI 2 -DMF phase and (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) phases in our revised manuscript. It's still difficult for us to carry out in-situ study on the conversion of PDS to perovskite phase. We found the PbI 2 -DMF phase is stable at RT (e.g. see Figure 2 of our manuscript), and the (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) phases are also reported stable at RT (Petrov, A. A. et al. J. Phys. Chem. C, 121, 20739-20743 (2017)). So it's not necessarily all the PDS will be converted into perovskite during solution thinning, depending on the dynamic of DMF evaporation. For the precursor solution with NH 4 Cl additives, the amount of precipitated PDS before the growth of RP perovskite is largely reduced, and most of the PDS phase should be converted to perovskite during solution thinning. Due to this reason, we do not found the presence of PbI 2 -DMF phase or (MA) 2 (DMF) 2 Pb m I 2m+2 (m=2,3) phases by XRD measurement after spin coating process (Supplementary Figure 10a). Nevertheless, we could like to mention that remaining of PbI 2 -DMF or MAI-PbI 2 -DMF related phase is thermodynamically possible. For the precursor solution without NH 4 Cl additives, much more PDS phases would be formed before the growth of RP perovskites (see the illustration in Figure 3a). In this case, there is a possibility that some of the PDS phases are not converted to perovskite during solution thinning. As shown in Supplementary Figure 10b, the peak around 2θ=9.5° indicates the presence of residual PbI 2 -DMF phase in sample spun from precursor without NH 4 Cl additives, and the XRD diffraction peaks around 2θ=6.6° and 2θ=8.1° also indicate the presence of (MA) 2 (DMF) 2 Pb 3 I 8 as the intermediate phase for the formation of bulk perovskites (see Ref. Petrov, A. A. et al. J. Phys. Chem. C 121, 20739-20743 (2017)). These remained PbI 2 -DMF and (MA) 2 (DMF) 2 Pb 3 I 8 phases will totally converted to perovskite phase after a thermal annealing at 100 , so that no PbI 2 -DMF and (MA) 2 (DMF) 2 Pb 3 I 8 phases can be detected in the final RP perovskite films as shown in Supplementary Figure 11a and Figure 4f in our revised manuscript.
In page 9, paragraph 2 of our revised manuscript, we add the following sentence "It's not necessarily all the PDS will be converted into perovskite phase during solution thinning, depending on the dynamic of DMF evaporation. So that the PbI 2 -DMF and (MA) 2 (DMF) 2 Pb m I 2m+2 phases can be detected in spin coated film (Supplementary Figure 10) until heated at elevated temperature of 70~100 ." In page 10, paragraph 1 of our revised manuscript, we add the following sentence "Since the preformed PDS are much suppressed by NH 4 Cl additives, these residual solvated phase may not be necessarily detectable by XRD (Supplementary Figure 10)." In our Supplementary (Yan, K. et al., J. Am. Chem. Soc. 137, 4460-4468 (2015)). According to P. Kamat group's results (K. Stamplecoskie et al., Energy Environ. Sci., 8, 208 (2015)), the Pb 2+ ions in precursor solution can form plumbate complexes with halide ions (X -) such as PbI 3 and PbI 4 2- . With the presence of excessive AX salt, there are increasing in the concentration of halide ions, which would lead to higher degree halide complex with more negative charges on the Pb-I based colloidal particles. When the Pb-I based colloidal particles become more negatively charge, due to the electrostatic repelling force, the aggregation of the Pb-I based colloidal particles in solution become difficult. The stabilization of Pb-I based colloidal particles in DMF by excessive AX salts can be one of the possible reason for the increased solubility (Tong, G. et al. Materials Today Energy 5, 173-180 (2017)).

Comment 4:
As reported in literature, e.g. Adv. Energy Mater. 2018, 8 (21), 1800185 and J. Am. Chem. Soc. 2017, 139, 1432-1435, RP perovskite films prepared from nominal n values are actually mixtures with various RP phases. These RP perovskite films show graded phase distribution even with out-of-plane orientation. Supplementary Figure 17 in this work also shows a similar observation. Could the authors comment on the growth mechanism of graded RP phases with out-of-plane orientation?
Responses 4: Thanks for raising this issue which are also very general phenomenon. We believe this frequently observed vertical phase separation can be understood with our results. In our explanation, the downward growth of the RP perovskite was confirmed as the major crystal growth mechanism. So the vertical phase separation is closely related to the nonuniform materials supply (i.e. the precipitated raw materials from dissolved state) during solution thinning. We confirmed the solubility of BAI, MAI and PbI 2 in DMF follow an order of PbI 2 <MAI<BAI, as shown in Supplementary  Figure 2b. This lead to PbI 2 -DMF and MAI-PbI 2 -DMF solvated phase preferably formed on top of the resulted film, which turns to perovskite with large n value or 3D perovskite phases in the lateral thermal annealing process. And the much higher solubility of BAI make the layered perovskite with low n value preferably formed on bottom of the resulted film. Although thermal annealing can promote the interdiffusion of these species, a detectable graded vertical phase separation is still expected.
In page 15, paragraph one of our revised manuscript, we changed the last sentence "At the last stage of spin coating or doctor blading, the solvated phase on top converted to a thin layer consisted of larger-n RP perovskites due to the excessive MA+ cations and lacked BA+ cation in the preformed solvated phase, contributing to the frequently observed vertical phase separation with large-n RP perovskites on top" to be "Due to the solubility difference shown in Figure 2b, there should be a preferably precipitation of PbI 2 -DMF and MAI-PbI 2 -DMF solvated phase on top and BA-rich phase on bottom during spin coating or doctor blading, which leads to relative more larger-n RP perovskites formed on top of the resulted film, contributing to the frequently observed vertical phase separation"; and we added the reference of "J. Am. Chem. Soc., 139, 1432-1435(2017  Comment 1: The manuscript by Wang et al. presents a thorough investigation of the growth mechanism for oriented layered perovskites. The main finding is that the firstly formed solvated perovskites in solution can template the sequential growth of layered perovskites, the orientation of layered perovskites is determined by the lattice matching between solvated perovskite and layered perovskites. In my opinion, since H. Tsai et al. reported hot-cast RP-type layered perovskites with out-of-plane orientation (Nature 536, 312, (2016)), the detailed nucleation process of layered perovskites is still unclear. The finding in this manuscript is very exciting to the 2D perovskite field, because the authors clarified several unsolved questions with this templated growth mechanism, especially how the out-of-plane orientation was formed. Moreover, it was well demonstrated that the formation of solvated perovskites can be engineered, which could be a useful method to achieve layered perovskite films with out-of-plane orientation and hence improve device performance. The manuscript is well written and organized with ample relevant references. I consider this work to be suitable for the broad readership in Nature Communications and I would recommend publication of this manuscript after revision.
The following are some specific concerns: According to some previous results (e.g. K. Yan et al. J. Am. Chem. Soc. 137, 4460 (2015)), the perovskite precursor solutions are generally colloidal dispersions, so I am wondering whether the excessive NH4Cl also impacted on the colloidal size and hence tuned the grain size and final morphology of the layered perovskite films?
Response 1: We appreciate the referee's recognition and very positive comments. Figure R9 | Particle size distribution of RP perovskite precursor solution with and without NH 4 Cl, where adding NH 4 Cl donot impact a lot on the particle size of colloidal in our perovskite solution.
We check the colloidal size in our precursor solution with and without additive by dynamic laser scattering method with 780 nm laser as light source (Nanotrac Wave II, Microtrac), as shown in Figure R9. We find that the particle size of colloidal in our perovskite solution with and without additive are all less than 3 nm, which is different with the cased reported in by Tong, G. et al. (Materials Today Energy 5, 173-180 (2017)). The absence of colloid particles with large diameter might be due to the different purity of PbI 2 raw materials (99.999%) and supplier (Sigma-Aldrich) in our case. Since the distributions of the particle size are very close for both cases, we can confirm that the increased crystallinity and grain orientation by the NH 4 Cl additives is not directly related to the changes in colloidal sizes.
Comment 2: It is impressive that the out-of-plane orientation and high crystallinity of the RP perovskites can be equally achieved in a group of films with excessive AX salts (A = NH 4 + or MA + ; X = Cl -, Bror I -) as additives. So how about the device performance for solar cells resulted from precursor solutions with other AX salts than NH 4 Cl?
Response 2: We thank referee for raising this issue. Although many other AX salts can lead to similar OP orientation of RP perovskites, in our study, we decided to choose NH 4 Cl for the systematic device optimizing. This is because NH 4 Cl additives delivered the best PCE, as show in Figure R10. We note that both NH 4 + ions and Clions are too small to be stable in perovskite phase, so that this kind of additive is readily to be removed during thermal annealing, which has been confirmed by our XPS analysis ( Figure R2) and agree with some published works (Si, H. N. et al. Adv. Funct. Mater., 27, 1701804 (2017); Xie, F. X. et al. ACS Nano, 9, 639-646 (2015)). The easy evaporation of NH 4 Cl helps to minimize the influence of additives on the stoichiometric ratio of RP perovskites and hence its average n-values, which is superior than using excessive MA + ions, or excessive Bror Iions. In 3d BA <n>=4 w NH 4 Cl C 1s Figure R2 | X-ray photoelectron spectroscopy (XPS) of BA based layered perovskite (<n>=4) fabricated from precursor solution with 0.5 molar ratio of NH 4 Cl additive. The absence of Cl 2p core level peak around 200 eV suggests undetectable Clions remained in the perovskite film. Fig. S14a, the XRD pattern of RP films with different amounts of NH4Cl addition was shown, the absence of diffraction peaks below 10° when the molar ratio of NH 4 Cl: PbI 2 >0.3 was believed to be due to the dominating OP orientation induced by NH 4 Cl additive. However, there is an exception for 1.0 NH 4 Cl, in which some extra diffraction peaks show up. Could the authors provide some discussion on this?

Comment 3: In
Response 3: Thanks for pointing this phenomenon out. We agree that the peaks below 10° in Supplementary Figure 17a (i.e. the Fig. S14a in our previous version) indicating IP orientation become detectable when too much NH 4 Cl was used (e.g. NH 4 Cl:PbI 2 =1.0). From the cross-section SEM image shown in Figure R11, crystal grain with random orientation can be observed when too much NH 4 Cl additive was used. Nevertheless, we can still explain it in the frame of solvated phased triggered templated growth of RP perovskites. In our story, the nucleation of RP perovskite is strongly determined by the location and amounts of preformed PDS phase, which dominated the finally crystal orientation and morphologies. However, the amount of PDS precipitation can be much reduced by NH 4 Cl additives. There should be an optimized range for the loading of NH 4 Cl, i.e. if there are too much NH 4 Cl additive in precursor solution, the preformed PDS precipitation will be insufficient to cover the whole liquid-air interface, and then the templated growth of RP perovskite will become nonuniform, or alternatively, become noncompetitive with homogenous nucleation inside the liquid phase. As a consequence, grains with IP or random orientation will show up. We believe th NH 4 Cl add
In th orien the e grow hom his unfavora ditive.