Dynamic multiphase semi-crystalline polymers based on thermally reversible pyrazole-urea bonds

Constructing responsive and adaptive materials by dynamic covalent bonds is an attractive strategy in material design. Here, we present a kind of dynamic covalent polyureas which can be prepared from the highly efficient polyaddition reaction of pyrazoles and diisocyanates at ambient temperature in the absence of a catalyst. Owing to multiphase structural design, poly(pyrazole-ureas) (PPzUs) show excellent mechanical properties and unique crystallization behavior. Besides, the crosslinked PPzUs can be successfully recycled upon heating (~130 °C) and the molecular-level blending of polyurea and polyurethane is realized. Theoretical studies prove that the reversibility of pyrazole-urea bonds (PzUBs) arises from the unique aromatic nature of pyrazole and the N-assisting intramolecular hydrogen transfer process. The PzUBs could further broaden the scope of dynamic covalent bonds and are very promising in the fields of dynamic materials.

• As far as the comparison between the experimental and computed parameters presented in Table S3 is concerned, a qualitative agreement is achieved, but the results are far from being quantitative. Is it due to the role of a different alkylation on the isocyanate (alkyl chain or methyl, for experiment and computation, respectively)? Does neglecting possible local and specific solventsolute effects (in calculation the solvent DCM is accounted only for through a continuum by PCM) might play some role of the different behavior registered?
• Why where the B3LYP/6-31+G* results shown in the main text and the B3LYP/6-31+G** shifted in the SI? If the authors think the B3LYP is more accurate then M06-2X, the results obtained with the largest basis set should be shown. In any case, they should motivate their choices, otherwise I don't see the point of presenting results with these three particular functionals. Incidentally, the title within Figure 5 should be corrected, since an asterisk seems to be missing in the basis set specification (6-31+G instead of 6-31+G*). On the contrary, if no polarization function was employed (6-31+G), it should be added in the text.
• In Table S3, it is not clear from the b and d notes, how the free energy is computed for each row. I guess the b and d subscripts in the last 4 rows should be exchanged. Please clarify.
Reviewer #2 (Remarks to the Author): This manuscript describes synthesis and self-healability of polyurea based on pyrazole-urea bonds. Authors studied dynamic chemistry of pyrazole-urea bonds using small molecule compound and theoretical studies.
The use of pyrazole moieties is not new. In fact, pyrazole has been used in the development of blocked isocynate materials in coating industries. Also, similar research using hindered urea has been published recently. Furthermore, the temperature required for dynamic chemistry of pyzazole-urea bonds is higher than 110C which may present some limitation of the chemistry to materials applications.
Although the manuscript is well written with experimental data required for self-healability, it would be published in a specific journal.
Reviewer #3 (Remarks to the Author): The authors prepared dynamic, multiphase, and semi-crystalline polymers based on thermally reversible pyrazole-urea bonds. For the first time, dynamic chemistry of pyrazole-urea bonds was well-studied by using model compounds. The authors also proposed a molecular-level welding concept between dynamic polyurea and polyurethane. The work of this manuscript is practical and logical and should be published after a minor revision. Here are some questions and comments: 1. It is interesting that the uncross-linked PPzU 6 shows a better mechanical strength than crosslinked PPzU 7 (Supplementary Table S2 and Figure 3). This is not usual. Why crosslink? Can PPzU 6 maintain the mechanical strength effectively after the same thermal recycling treatment? Moreover, the yield strength and modulus of PPzU 7a and PPzU 7b had a significant decline after thermal recycling (Supplementary Table S2). 2. The author should not use the "gel fraction0.97" to demonstrate the densely covalently crosslinking of the materials (Line 118, Page 7). What's more, the PPzU materials in this work can be hardly defined as densely covalently crosslinking, especially for PPzU 7a, whose crosslink density is 233 mol•m-3 and swelling ratio is up to 576%. 3. The authors put forward the concept of molecular-level welding. It is not so justified as the experiment was conducted at an elevated temperature and solvent. Welding should be accomplished in solid state. In my view, macromolecular exchange reaction or "polymer scrambling" reaction (J. Am. Chem. Soc. 2003, 125, 4064) is more correct. Also what are the values for the proposed molecular-level welding? 4. There is an error in Line 135, Page 8, the "PPzU 5" should be "PPzU 6". 5. Supplementary Figure 10 (D) was not clear and the soaking time in solvent was not given. It is better if PPzU 8, which has a lower crystalline degree, can be tested as a contrast sample. 6. In the introduction part, about the review on dynamic urea bond: Except for introducing bulky substituents to make urea bond dynamic (Nat. Commun. 2014, 5, 3218), the addition of metal ion is also effective to reduce the dissociation activation energy of urea bond (J. Mater. Chem. A, 2019, 7, 15933-15943).

Dear reviewers,
On behalf of my co-authors, we thank all the reviewers for the helpful comments, which have helped us to greatly improve our manuscript. We carefully considered all reviewers' comments and did our best to address these concerns. In the text below we provide a point-by-point response to all comments made by the three reviewers. The changes were marked in the revised Manuscript and Supplementary Information. Thank you and best regards. In this paper the authors report on new kind of dynamic covalent polyureas, poly(pyrazole-ureas) (PPzUs), which are shown to exhibit excellent mechanical properties and peculiar crystallization behavior, in contrast to reported dynamic polymers, whose crystallization is inhibited because of a different structural design. The first part of the manuscript is devoted to illustrate the reversible character of the pyrazole-urea bonds (PzUBs) through a wide variety of methods, ranging from experimental spectroscopic and thermodynamics technique to theoretical calculation, carried out at DFT level. As detailed by the authors in the second part of the manuscript, the PzUBs reversibility is thereafter exploited to recycle, upon heating, the PPzUs which can be easily prepared, at ambient temperature in the absence of a catalyst, by a highly efficient polyaddition reaction of pyrazoles and diisocyanates.
Given the potential impact of PzUBs in the field of dynamic materials and the growing attention to efficient strategies to improve the structural design of responsive polymers, the topic is in my opinion appropriate for Nature Communications. The manuscript is well written and the results, well supported by both experimental and theoretical evidences, discussed accurately. However, I have few concerns regarding essentially the computational part, and the following remarks should in my opinion be addressed, prior to publication.
Response: We appreciate the reviewer for the time reviewing our manuscript and the positive comments. The manuscript has been carefully revised according to the suggestions from the reviewer.
• The paragraphs devoted to the discussion of the computational results are presented only at page 10, after the discussion on experimental findings on PPZUs. Since the calculations concern with the PzUBs of the polymer building blocks, whose experimental findings are discussed in the first part of the manuscript, rather than with the polymer itself, it would seem more appropriate to shift the computational results section at the end of such first part, before the section "Synthesis and characterization of poly(pyrazole-ureas)". Otherwise, the computational section seems a bit "appended" at the end of the paper.
Response: Many thanks for this suggestion. We have made this shift.
• From a computational point of view, my major concern deals with the inclusion of dispersion within the calculations, which is tested by comparing the B3LYP (which does not account for dispersion) and M06-2X (which includes dispersion to a certain extent) functional with the same basis set (6-31+G**). Given the importance, as discussed by the authors, of the aromaticity of the pyrazole unit, do the differences found between the two functionals, and reported in Table S3, can be attributed to such interactions?
Response: Thanks. We do think the aromaticity of the pyrazole plays an important role in the easy dissociation of the C-N bond. However, it looks like that the difference between the two functionals is mainly caused by the C-N bond formation process. The following table lists the free energies of these intermediates and transition states of the two functionals (Table R1). As we can see, the differences between the free energies of the two functionals (ΔG 1 ) significantly decrease during and after the C-N bond formation, indicating that the computed bond energy difference should be the main contribution to the divergence between the two functionals.
We also investigated the influence of the aromaticity and internal hydrogen bond on the stability of the final product by comparing the energy of the following three structures (Table R2). In R1, the influence of the aromaticity effect on carbonyl moiety is ruled out while the internal hydrogen bond is retained. In R2, the internal hydrogen bond is removed while the aromaticity effect is retained. In R3, both the aromaticity effect and internal hydrogen bond are removed. The enthalpy and free energy of R1 are 0.77 and 1.07 kcal·mol -1 lower than P1, respectively, indicating the aromaticity of the pyrazole indeed diminishes the conjugation effect between the pyrazole nitrogen and the carbonyl group. The energy of R2 is much larger than P1. On the one hand, the removal of the internal hydrogen bond can increase the energy; on the other hand, the repulsion between the lone electron pairs on Nitrogen and Oxygen also causes the destabilization. As we can see in R3, no lone electron pair repulsion exists, and the energy of R3 is comparable with P1.
These results indicate that the aromaticity and internal hydrogen bonding only show little influence on the stability of the final product. Therefore, we don't think aromaticity is the main contribution to the divergence between the two functionals.  Response: As we stated above, although the main difference between M06-2X and B3LYP is the treatment of dispersion, we don't think it is the reason for the large free energy difference with M06-2X. The involvement of dispersion is indeed very important when dealing with weak interaction, but in this reaction system, no significant weak interaction exists. Table R1 shows that the free energy differences between M06-2X and B3LYP (ΔG 1 ) significantly decrease during and after the C-N bond formation (P1 << TS2, Int2 < TS1 << Int1 ≈ MeNCO+pyrazole). The similar free energy differences (ΔG 1 ) for hydrogen-bonded Int1 and starting materials (MeNCO+pyrazole) may exclude the possibility that dispersion has a big influence on this system. Instead, the strong C-N bond formation accounts for the large difference of the thermodynamic free energy (ΔG b ).
We have added the comments in the Supplementary Information (Supplementary Table 2).
• As far as the comparison between the experimental and computed parameters presented in Table   S3 is concerned, a qualitative agreement is achieved, but the results are far from being quantitative.
Is it due to the role of a different alkylation on the isocyanate (alkyl chain or methyl, for experiment and computation, respectively)? Does neglecting possible local and specific solvent-solute effects (in calculation the solvent DCM is accounted only for through a continuum by PCM) might play some role of the different behavior registered?
Response: Thanks for the suggestion. We think the different alkyl subsistents may cause a certain difference between experiment and computation results. In addition, the small basis set used in the manuscript can also cause some inaccuracy. We further calculated the energy using a bigger basis set (6-311++G(2df,2pd)) with SMD solvation model ( Fig. 2a and Supplementary Fig. 10). The activation energy of the reversed reaction significantly decreases (Supplementary Table 1). We apologize for missing a statement in the Computational Methods, that the solvation model we used is CPCM with SMD-coulomb atomic radii (page 4 in the Supplementary Information). We have also examined the SMD model with B3LYP and M06-2X methods, and no significant difference was found.
• Why where the B3LYP/6-31+G* results shown in the main text and the B3LYP/6-31+G** shifted in the SI? If the authors think the B3LYP is more accurate then M06-2X, the results obtained with the largest basis set should be shown. In any case, they should motivate their choices, otherwise I don't see the point of presenting results with these three particular functionals.
Incidentally, the title within Figure 5 should be corrected, since an asterisk seems to be missing in the basis set specification (6-31+G instead of 6-31+G*). On the contrary, if no polarization function was employed (6-31+G), it should be added in the text.
Response: We are sorry for the mistake. The results shown in the main text were B3LYP/6-31+G(d), and now we have updated the data to the level of B3LYP/6-311++G(2df,2pd)//B3LYP/6-31+G(d,p) (Fig. 2a). As mentioned in the Computational Methods (page 4 in the Supplementary Information), we employed B3LYP/6-31+G(d) for all calculations including the favored pathway, disfavored pathway, and resonance energy analysis.
Besides, for the favored pathway we also used M06-2X functional for comparison and bigger basis sets to improve accuracy ( Fig. 2a and Supplementary Fig. 10).
• In Table S3, it is not clear from the b and d notes, how the free energy is computed for each row.
I guess the b and d subscripts in the last 4 rows should be exchanged. Please clarify.
Response: The subscript of b refers to the binding reaction (forward reaction) and d refers to the dissociation reaction (backward process). The activation free energies ΔG ≠ b and ΔG ≠ d are calculated from Supplementary Figs 3 and 9, respectively. The reaction free energy ΔG b is calculated from Supplementary Fig 5. We have clarified these in Supplementary Table 1  Authors studied dynamic chemistry of pyrazole-urea bonds using small molecule compound and theoretical studies.
The use of pyrazole moieties is not new. In fact, pyrazole has been used in the development of blocked isocyanate materials in coating industries. Also, similar research using hindered urea has been published recently. Furthermore, the temperature required for dynamic chemistry of pyzazole-urea bonds is higher than 110C which may present some limitation of the chemistry to materials applications.
Response: We thank the reviewer very much for spending precious time on our manuscript. We would like to further explain the novelty and importance of the manuscript by the following responses to the comments of the reviewer.
It is of great scientific and technological importance to discover the dynamic nature of old chemistry and then utilize it for dynamic polymer materials. Although pyrazoles have been studied as blocked molecules for isocyanate, the pyrazole-urea bonds have not been applied for the design of dynamic polymers with reprocessing/recyclable capacities. What's more, the kinetics, thermodynamics and reversible mechanism of pyrazole-urea bonds were not intensively investigated before. The dynamic nature of pyrazole-urea bonds is a significant contribution to the fields of dynamic polymer materials. The detailed reaction parameters of pyrazole-urea bonds from both experimental and theoretical aspects may also be valuable to other authors that use dynamic covalent bonds, such as dynamic combinatorial chemistry, self-assembly, shape-memory materials, 3D printing, etc.
Cheng et al. reported interesting hindered urea bonds which are significantly destabilized by bulky substituents and reversible at room temperature. The hindered urea bonds are used for the design of self-healing/recyclable amorphous polyureas, which are mechanically weak organogels (Nat. Commun. 2014, 5, 3218)  They have better thermal resistance and can withstand higher temperatures without loss of strength or change of structure, which allows for the applications requiring a higher temperature.
Although the manuscript is well written with experimental data required for self-healability, it would be published in a specific journal.
Response: We appreciate that the referee recognizes the overall quality of the manuscript. As described above, our work is of interest for chemists and materials scientists. The quality of the manuscript has been improved after addressing the comments from the referees. We believe that the revised manuscript accomplishes the requisite of generality and broad interest for the community of Nature Communications.
Reviewer #3 (Remarks to the Author): The authors prepared dynamic, multiphase, and semi-crystalline polymers based on thermally reversible pyrazole-urea bonds. For the first time, dynamic chemistry of pyrazole-urea bonds was well-studied by using model compounds. The authors also proposed a molecular-level welding concept between dynamic polyurea and polyurethane. The work of this manuscript is practical and logical and should be published after a minor revision. Here are some questions and comments: Response: We thank the reviewer for the positive comments on our work. We are glad that the reviewer finds this work important and practical.
1. It is interesting that the uncross-linked PPzU 6 shows a better mechanical strength than cross-linked PPzU 7 (Supplementary Table S2 and Figure 3). This is not usual. Why crosslink?
Can PPzU 6 maintain the mechanical strength effectively after the same thermal recycling treatment? Moreover, the yield strength and modulus of PPzU 7a and PPzU 7b had a significant decline after thermal recycling (Supplementary Table S2).
Response: We agree with this reviewer that usually crosslinking could enhance mechanical strength. In our polymers, besides the chemical crosslinking, crystallization which serves as physical crosslinking also has a significant contribution to the mechanical properties of PPzUs.
For highly covalently crosslinked PPzUs 7, crystallization is strongly inhibited, which leads to a decrease in tensile strength. The similar phenomenon that covalently crosslinked polymers shows lower mechanical strength than the linear ones can also be found in crystalline polyethylene (Polymer 1982(Polymer , 23, 1944(Polymer -1952. Compared with thermoplastic polymers, covalently crosslinked thermosets usually have excellent solvent resistance and thermal stability. In our work, PPzUs 7 are not soluble in solvents including CH 2 Cl 2 and CHCl 3 , in which PPzU 6 dissolve. The TGA and DMA experiments also prove that the thermostability of PPzUs 7 is better than that of linear 6 ( Supplementary Figs 16 and 19).

Strain (%)
Recycle 1 Origin crosslinking of the materials (Line 118, Page 7). What's more, the PPzU materials in this work can be hardly defined as densely covalently crosslinking, especially for PPzU 7a, whose crosslink density is 233 mol•m-3 and swelling ratio is up to 576%.
Response: We thank the reviewer for this valuable suggestion. We have clarified this in the revised version of our paper with the following sentence: "Of particular note is that PPzU 7c with the highest crosslinking degree (average molecular weight between crosslinks (M c ) = 1.7 kg·mol -1 , Supplementary Response: Many thanks for this insightful suggestion. More widely used term, "blending", is now used instead of "welding" throughout the manuscript. We think this molecular-level blending is a new technique to make polymer blends or hybrid materials. As we know, different species of polymers are immiscible in most cases. Owing to the introduction of reversible covalent bonds, polymer chains from different species of polymers can exchange. Therefore, this technique can be applied to synthesize polymer mixture with better compatibility. For example, recently enhanced adhesion of poly(methyl methacrylate) and polyethylene was achieved through dynamic dioxaborolane metathesis reaction (Science 2017, 356, 62-65). In addition, the macromolecular exchange reaction can control topological transformations of polymers (Nat. Chem. 2017, 9, 817-823). We have modified the corresponding statement to make it clear in our revised manuscript as follows: "Different from traditional polymer blending methods, this artful methodology using dynamic covalent bonds provides a new approach to polymer blends." 4. There is an error in Line 135, Page 8, the "PPzU 5" should be "PPzU 6".