Bottom-up growth of homogeneous Moiré superlattices in bismuth oxychloride spiral nanosheets

Moiré superlattices (MSLs) are modulated structures produced from homogeneous or heterogeneous 2D layers stacked with a twist angle and/or lattice mismatch. Expanding the range of available materials, methods for fabricating MSL, and realization of unique emergent properties are key challenges. Here we report a facile bottom-up synthesis of homogeneous MSL based on a wide-gap 2D semiconductor, BiOCl, using a one-pot solvothermal approach with robust reproducibility. Unlike previous MSLs usually prepared by directly stacking two monolayers, our BiOCl MSLs are realized in a scalable, direct way through chemical growth of spiral-type nanosheets driven by screw-dislocations. We find emergent properties including large band gap reduction (∼0.6 eV), two-fold increase in carrier lifetime, and strongly enhanced photocatalytic activity. First-principles calculations reveal that such unusual properties can be ascribed to the locally enhanced inter-layer coupling associated with the Moiré potential modulation. Our results demonstrate the promise of MSL materials for chemical and physical functions.

1. There is a lack of introduction to BiOCl and other 2D materials having screw-dislocation driven growth. What is the current research status of BiOCl nanosheets? (i.e., synthesis methods, properties and applications) This would make the manuscript much more comprehensive to the wider field.
2.Are all the synthesized BiOCl spiral nanosheets formed with a small degree of interlayer twist angle? Why is it only unique to BiOCl and not observed for pervious 2D materials or nanoplates with spiral structures (or this should be present for other 2D materials as well)? Since the formation of interlayer twist is important here, its mechanism should be explained. Could this be associated to the strain generated by the screw-dislocation similar to "Eshelby twist" for 1D nanowires?
3.In Figure 2b, there are some peaks missing in their XRD spectra for 40 min and above. What does the missing peaks indicate? 4.The term "Real-time" in the heading "Real-time characterization of growth…" is misleading. TEM was done after synthesis on transferred samples, not in "real-time". 5.For multiple spiraling layers besides two layers, is there a constant interlayer twist between the successive layers? For example, does the periodicity of the Moiré pattern decreases with increasing layers? What effect does this have towards its emergent properties?
Reviewer #3 (Remarks to the Author): In this work, 2D Moiré superlattices (MSLs) BiOCl was synthesized via solvothermal reaction, which was considered to follow a screw-dislocation driven mechanism. The fabrication of BiOCl MSLs leads to strong modulation of optoelectronic properties with 0.6 eV band gap reduction and substantial carrier lifetime increase, which may cause the enhanced visible-light photocatalytic activity. First-principles calculations further revealed that the locally enhanced inter-layer coupling effect associated with the Moiré potential modulation may play an important role in the modulation of optoelectronic properties. Only after addressing the following questions can this work be published.
1. I am wondering if surfaces of BiOCl are clean enough, without capping agents residuals. If PDDA caps on the surfaces of BiOCl, how to remove it? Whether PDDA participate in catalytic process?
2. There are many methods to modify BiOCl, such as doping and heterostructures regulation. Compared to the traditional methods, what is the unique advantage of screw dislocation?
3. For the growth mechanism, more quantified information is need. For example, the authors claim the low supersaturation condition responsible for the screw dislocation, then how low, or what is the benchmark? Why low supersaturation condition is important? 4. TEM images in the manuscript just exhibited one single BiOCl nanosheet. TEM images that show more nanosheets are needed to further prove the monodispersity. Moreover, some cracks and irregular edges appeared in TEM images. What caused the existence of this kind of structure? 5. Cs-corrected HADDF STEM was widely used to characterize nanoscale structure down to atomic level. In this work, Cs-corrected HADDF-STEM should be used to characterize the differences among HH-stack, AH-stack, and AA-stack. What was the atomic arrangement in the uniaxial-oriented of BiOCl MSLs? Would strain exist in the uniaxial-oriented of BiOCl MSLs? 6. I am wondering if high-indexed facets could exist on step sites in the spiral nanosheets. If so, please index the specific facets. 7. The authors mentioned bidirectional growth with AFM images. Are the growth speeds in two directions the same? 8. The manuscript discussed HH-stack and AA-HH-stack a lot, but did not pay too much attention on the HA-stack structure. Could the authors explain the unique properties of HA-stack and how it worked in this system? 9. Whether the L-edge of XANES and EXAFS for Bi in spiral nanosheets and bulk BiOCl are the same? 10. What are the reaction sites on BiOCl MSLs during catalytic degradation, the oriented axis, highindexed facets, edge sites, or others?
Note: The Reviewers' Comments are with italic, our Responses are in blue, the new changes made in the manuscript are in red.

Reviewer #1
(1) Referee: This Fig. 1k. We have also revised Fig. 1j to clearly indicate the actual structure of our experimentally synthesized BiOCl nanosheet. They are shown also as Fig. R1 below for the convenience of the referee. The specific atomic structures within the unit cell of the MSL are shown in Fig. 3a and 3c. The squared Moiré patterns with the period L from 7.7 nm to 14.3 nm and the corresponding twist angle from 3.0° to 1.6°, constructed using the bilayer MSL model, are now provided in the new Fig. S16 (shown also as Fig. R2 below). Four cases with the twist angles of 1.6°, 2.0°, 2.5°, and 3.0° are shown in this figure.  The reasonableness of our constructed model can be seen from the fact that the MSL and the resulted functional properties that we observe are due to the incommensurate structure formed between two outmost layers with thicker inert sections away from this. Experimentally, the twisting is observed between the adjacent nanosheets, each composed of seven BiOCl monolayers. The incommensurate structure occurs between the bottom monolayer of the upper nanosheet and the top monolayer of the lower nanosheet. The key of the MSL is spatial in-plane modulation of atomic-scale structures, which leads to a spatial modulation of electronic structure and related properties. Such spatial modulation is directly related to the two outmost BiOCl monolayers stacked with a twist angle. The inner BiOCl layers of the nanosheets do not (or barely) contribute to such spatial modulation. Hence, our constructed model, which is tailored for DFT calculations, captures the key structural features that are responsible for the spatial modulation of physical properties of interest. To make this point clearer, we have added the above discussion into p. 6, 2 nd paragraph of the revised manuscript.   Fig.1i would not be the as-claimed distance of AA-stack but the length of c axis for BiOCl lattice. Furthermore, the influence of AA-stack to bandgap reduction would be reduced, then how to understand the huge bandgap reduction of ~0.6 eV for BiOCl, while that for graphene MSL was merely 0.05 eV.
Reply: Our synthesized BiOCl MSLs are formed by the twisted stacking of two 7-monolayers nanosheets. The distance (~0.736 nm) between the two adjacent nanosheets is determined from the AFM step height image (Fig. S1) and TEM measurement (Fig. 1i). This distance is equal to the length of c axis for BiOCl lattice as noticed by the referee. The inset is now added to Fig.S1b (shown also as Fig. R4 below for the convenience of the referee) to illustrate this point more explicitly. In our DFT calculations, the bilayer MSL model adopts the experimentally determined distance between the two adjacent nanosheets (~0.736 nm). The predicted 0.55 band gap reduction agrees with the experimentally observed value of ~0.6 eV.
Such large band gap reduction is traced down to the AA-stack region of the MSL, which sustains the shorter interlayer distance than its equilibrium condition. Based on our calculation, the equilibrium interlayer distance for the AA-stack is as large as 0.877 nm. This originates from the AA-stack having crystal structure with the head-to-head arrangement of the Cl atoms from adjacent layers (Fig. 3c). Such shorter interlayer distance substantially enhances the interlayer coupling, upshifts the valence band edge, and reduces the band gap. This has been discussed in detail in the section of Origin of band-gap reduction in the manuscript.        Reply: This has been now clarified more explicitly in our revised text, Fig.1j, and newly added Fig.1k (see also our replies to the referee's comments #2 and #3). We appreciate the referee for offering us the opportunity to improve the clearness of our presentation in the manuscript. Reply: We thank the referee for the above positive comments, and for the below constructive suggestions that help us improve the manuscript quality.
(2) Referee: There is a lack of introduction to BiOCl and other 2D materials having screw-dislocation driven growth. What is the current research status of BiOCl nanosheets? (i.e., synthesis methods, properties and applications) This would make the manuscript much more comprehensive to the wider field.
Reply: By following the referee's good suggestion, we have now added the following texts on the introduction to BiOCl in p. 4 line 5 of the revised manuscript: "As a stable layered wide-gap semiconductor, bismuth oxychloride (BiOCl) has received great attention owing to its optical and electrical properties feasible for photocatalysis applications, such as water splitting 1 , dye degradation 2 , N2 reduction 3   As noted by the referee, the formation of interlayer twist angle between two nanosheets is essential for emergence of Moiré pattern, but not a sufficient condition.
Among the synthesized 2D spiral nanosheets with screw-dislocation driven growth, the interlayer twist angle between two nanosheets was indeed reported, for instance in

Reply:
We thank the referee for the above positive comments on our work.       R14 below for the convenience of the referee), in which we also show the structures from our theoretical simulation (Fig. R14d,g,h) for direct comparison. We can see that the Moiré patterns involving three distinct regions of HH-stack, AH-stack, AA-stack structures are observed in Fig. R14c. The results are in agreement with the structures from our theoretical simulation (Fig. R14d). From the high-resolution enlarged arrangements of the largest Bi atoms, which also agrees with the atomic structures in HH-stack and AH-stack region from the simulation results ( Fig. R14g-h). We have now summarized this newly added results briefly in the main text (on p. 10 line 3).
Regarding the referee's question on the possibly existing strain in the uniaxial-oriented of BiOCl MSLs, we think the center of screw-dislocation and the edge of spiral nanosheets provide channels to mechanically relax residual strains.
Therefore, the BiOCl MSLs should be free of strain distribution.      nanosheets. There are also screw dislocation axis and edge sites as mentioned by the referee. These are potential reaction sites during catalytic degradation. We now explicitly state these possibilities, with the comment that future work to distinguish them will be of interest (in p. 14 line 23 of the main text).
changes made in the manuscript are in red.

Reviewer #1
(1) Referee: The authors have done a substantial revision and the quality of the paper was improved. However, there were still critical issues remaining.

Reply:
We thank the referee for making further constructive and specific comments to help us improve the quality of our work.
(2) Referee: Previously, the authors mistook the distance of BiOCl c axis (0.736 nm) for the distance between the two adjacent MSL nanosheets. In the last revision, the authors revised this point in Figure 1 but did not correct it in Figure 3d, i.e., there was no experimental evidence to demonstrate the distance of AA-stack. Since the lack of such experimental result would make the calculation results in Figure 3 meaningless in terms of bandgap determination, the authors are strongly recommended to provide a side-view HRTEM image of BiOCl MSL to show the distance between the two adjacent MSL nanosheets.

Reply:
We thank the referee for this good suggestion. We note that the actual distance between the two adjacent nanosheets was determined from the joint analysis of the AFM step height (Supplementary Fig. 1) and TEM data (Fig. 1i). It is equal to the length of c axis for BiOCl lattice (0.736 nm) as noticed by the referee. Following the referee's suggestion, we have now performed the side-view HRTEM measurement of BiOCl MSL. The cross section of the MSL was obtained by using the focused ion beam, in which procedure the measured micro area is protected by deposited platinum before being etched by gallium ions. The results are shown in the newly added Supplementary Fig. 9 (shown also as Fig. R1 below for the convenience of the referee). We observe straightly and clearly that the distance between the two adjacent nanosheets is equal to the length of c axis for BiOCl lattice.  R1 (Supplementary Fig. 9): The side-view HRTEM image of BiOCl MSL nanosheet. The cross section of the MSL was obtained by using the focused ion beam, in which procedure the measured micro area is protected by deposited platinum before being etched by gallium ions.
Accordingly, in p. 7 lines 174-176 of the revised manuscript we have now added the following discussion: "The side-view HRTEM image (where the cross section of the MSL was obtained by using the focused ion beam, Supplementary Fig. 9) further indicates that the distance between the two adjacent nanosheets is equal to the length of c axis for BiOCl lattice."

Reviewer #2
(1) Referee: The authors had clarified all our questions and had made the appropriate revisions. Therefore, I would recommend publication of this paper.

Reply:
We appreciate the referee for the valuable efforts in reviewing our manuscript.

Reviewer #3
(1) Referee: The revision has been made accordingly to my comments and improved significantly. I recommend acceptance of this paper in this form.

Reply:
We thank the referee for the constructive comments that help us greatly improve the manuscript quality.
The novelty of this work is the Moiré superlattice of BiOCl for band gap reduction, increase in carrier lifetime, and thus the enhanced photocatalytic activity. The first report of BiOCl Moiré superlattice is of great interest, however, the assignment of band gap reduction merely to the Moiré superlattice is risky owing to the presence of defects, as evidenced by Bi LIII-edge EXAFS spectra and DRS spectra of BiOCl MS. The authors have commented on the influence of defect to DRS spectra in the last revision, I think the manuscript could be accepted for publication after the authors comment on the possible influence of defect to band gap reduction, which would help the readers to get an unbiased knowledge of bandgap modulation for BiOCl.