Frontal polymerization-triggered simultaneous ring-opening metathesis polymerization and cross metathesis affords anisotropic macroporous dicyclopentadiene cellulose nanocrystal foam

Multifunctionality and effectiveness of macroporous solid foams in extreme environments have captivated the attention of both academia and industries. The most recent rapid, energy-efficient strategy to manufacture solid foams with directionality is the frontal polymerization (FP) of dicyclopentadiene (DCPD). However, there still remains the need for a time efficient one-pot approach to induce anisotropic macroporosity in DCPD foams. Here we show a rapid production of cellular solids by frontally polymerizing a mixture of DCPD monomer and allyl-functionalized cellulose nanocrystals (ACs). Our results demonstrate a clear correlation between increasing % allylation and AC wt%, and the formed pore architectures. Especially, we show enhanced front velocity (vf) and reduced reaction initiation time (tinit) by introducing an optimal amount of 2 wt% AC. Conclusively, the small- and wide-angle X-ray scattering (SAXS, WAXS) analyses reveal that the incorporation of 2 wt% AC affects the crystal structure of FP-mediated DCPD/AC foams and enhances their oxidation resistance.

This is a nice paper about the use of FROMP to obtain polyDCPD foams by a novel approach. The manuscript is well written, data are well presented and consistent with the discussion and conclusions. The authors investigated many parameters, including composition and reactor geometry. I do reccomend it for publication after some minor modifications. Actually, the authors should clarify these points: l. 147 -the chemical formula would help l. 148 -FP cannot exhibit a catalytic activity (IUPAC definition) Table 1 and text -the use of so many decimal digits in non-quantitative data is a non-sense l. 187 -was the more used descending mode (ignition at the top) tried? Why using the ascending one? ls. 264,265 -the number of significant digits is not consistent (errors cannot have two decimal digits while the averages have just one) Fig. 5a -pDCPD solid is not a correct way to distinguish this sample from the others. Actually, all of them are solid. l. 281 -while Vf and Tmax are parameters well known in FP, they should need a better definition for readers who are not familiar with the technique. l. 282 -how reliable long distance IR measurement are? I suspect that they could be affected by the glass surface; in addition, temperature taken inside the sample instead of the external wall should be more significant. l. 303 -Are all the digits significant? Fig. 6b -perfectly straight lines are displayed, thus indicating that no deviation from linearity was observed, no experimental error affected the measurement, hundreds of measurements, each one taken a very short time, was performed. This sound strange. Are the displayed lines "real" or are they the interpolations of several points that are not shown?
Reviewer #2 (Remarks to the Author): In the article, Anisotropic Macroporous Dicyclopentadiene / Cellulose Nanocrystal Foam via Frontal Polymerization-triggered Simultaneous Ring-opening Metathesis Polymerization and Cross Metathesis, Park and Kwak explored the use of cellulose nanocrystals in the synthesis of poly-dicyclopentadiene anisotropic foams using frontal polymerization. The authors claimed that an allyl-functionalized cellulose nanocrystal will participate in a cross-metathesis reaction during the polymerization of dicyclopentadiene. This parallel reaction will improve the phase separation between the two components and as a result the control of pore morphology. The authors synthesized and characterized two allyl-functionalized cellulose nanocrystal batches. Then, they frontally polymerized dicyclopentadiene in the presence of the nanocrystals. Morphological characterization of the foams showed a variation of the pore size as a function of the loading of nanocrystals in the formulation. The authors also studied the influence of the nanocrystals on frontal velocity and maximum temperature of the system. SAXS analyses were used to study the alignment of the cellulose nanocrystals in the foam. Finally, the authors proposed a mechanism for the formation of the foam that involves the chirality of the nanocrystals and their use as nucleating agents.
Major concerns 1. The main claim of the manuscript about the synthesis of anisotropic macroporous dicyclopentadiene/cellulose nanocrystal foams is mainly supported by literature precedent instead of performed experiments. 2. The relation between the foaming process and phase separation caused by the cross-metathesis reaction is not well-explained and supported by experiments. 3. Although the authors claim the formation of anisotropic foams, the alignment of the voids was not determined. 4. Authors need more control experiments to support the multiple claims in their manuscript.
Minor concerns 1. Line 32-35: The authors claim that understanding the relation between stereochemistry and crystal structures during FP will be useful to enhance the oxidative stability of pDCPD. It is hard to follow the logic of this claim and is beyond the scope of the current publication, which is about the use of cellulose nanocrystals to make pDCPD foams. 2. Line 51-52: Authors should describe what characteristics are they referring to. 3. Line 107-110: The authors claim that an effective internal phase separation would occur caused by the difference in the rate of DCPD crosslinking and nanocrystals cross-metathesis. However, there are no experiments in the paper related to the ROMP and CM rates. 4. Line 139-141: Is 1 wt% enough to induce the formation of porous structures? In the following sections the authors used up to 4 wt% of allylic-modified cellulose nanocrystals, it would be important to show the results with 4 wt% of the CTD-modified cellulose nanocrystals. 5. Line 154: Authors need to explain what an IG-decoupling sequence is. 6. Line 154-166: The authors correctly stated that CP-MAS NMR is not quantitative. However, it is hard to follow their use of this technique to quantify the percentage of allylation in the LAC sample. First, in order to compare the relative intensities of 13C-NMR signals across different samples is better to use an internal standard instead of the other cellulose signals. Second, did the authors take the IG NMR spectrum for LAC? What was the result? Third, significant figures of the ratios and percentages need to be double-checked. Fourth, there is a typo in line 166, instead of Table S1 it should say Table  1. 7. Line 178: The claim needs a reference. 8. Line 181: It is unclear why the authors studied the agglomeration in aqueous media instead of the monomer. 9. Line 193-197: If the authors started with the same mass in all the experiments, why did they observe a decrease in the mass after polymerization? It is not clear why by using a higher percentage of AC the authors obtained foams with lower mass. 10. Line 200: The authors claim the directional anisotropy is maintained. However, there are no calculations that support this claim. 11. Line 202-208: The authors suggest LAC performs worse than AC in forming polymeric solid foams. However, this claim is based solely on the comparison of images 4a and 4b. Authors need to calculate alignment, average diameter, and distribution to support this claim.  Table 2 and reconsider the importance of the inclusion of R2. 14. Line 231-232: Reconsider rephrasing for more clarity. 15. Line 231-233: The claim about the no crosslinking between DCPD and AC is hard to support by 13C-NMR because at these concentrations the AC signals are not observed in Figure S5. Therefore, it is highly unlikely to observe a signal corresponding to cross-metathesis in the pDCPD signal around 130 ppm. 16. Line 237-239: The authors suggest that bubbles are forming from a phase separation phenomenon. It would be important that the authors expand this concept. Where are the bubbles coming from? Are they formed from dissolved gases? Why phase separation produces bubbles in this system? 17. Line 239-250: The authors claim that the wrinkles observed in the foams are caused by the polymerization rate difference between ROMP and CM. However, as the authors stated, this behavior had been observed in another system where there is only one type of reaction (ROMP). Therefore, the claim can't be supported by the experiment. 18. Line 248-267: The authors discussed the observation of spin mode in formulations with low content of AC. However, no connection of the spin mode with the hypothesis of the manuscript was provided. It is hard to understand the reason why this observation was included in the manuscript. Moreover, Figures 4a, 4b, and S3a-b are not related to the study of the spin mode. Why was spin mode not observed in the other formulations? Why were two different imaging techniques used to compare Figure S6? 19. Line 274: The authors should include the images of other formulations together with Figure S6c to support the increase in the multichannel size. 20. Line 282: The use of an IR camera to measure maximum temperature is not accurate because glass absorbs IR radiation. 21. Line 290-298: Authors suggest the increase in frontal velocity is due to an increase in the amount of monomers containing C=C moieties. To support this claim authors need to demonstrate that the amount of C=C coming from the AC is higher than the amount of replaced C=C from DCPD. Additionally, this suggestion does not explain why the 4 wt% has a lower frontal velocity. Claim about allyl-moieties being more reactive than cyclopentene is misleading because dicyclopentadiene contains both high strain and low strain cyclic alkenes. Therefore, a discussion about the low reactivity of one particular moiety cannot be done separately. 27. Line 339-347: Authors suggest that polydispersity is problematic in the determination of crystallinity via SAXS. However, they claim the samples are more crystalline which indicates a higher anisotropy. If SAXS generated unclear results about the alignment of the AC, the authors need to include other experiments that corroborate their assumption. 28. Line 346-349: Authors need to expand on the relation between DCPD oxidation and the foaming mechanism. 29. Line 353: It is hard to see the difference in oxidation with the images provided in Figure 7b. FTIR characterization will be helpful to corroborate their differences. 30. Line 369-372: How do the authors know that an organization of the polymerizing DCPD is happening in their system? It is hard to support this claim just by referencing literature on other systems, which are dramatically different from the current one. 31. Line 372-375: Authors need to provide quantitative oxidation and crystallinity values for all the samples to support that they are correlated with each other. 32. Line 383: How do the authors know that the 2wt% mixture has a stoichiometric amount of functional C=C and chiral-inducting hydroxyl groups? 33. Line 385-387: The authors use the word "synergistically" to explain their claims about the organization of the CN. However, a better discussion is needed to support this hypothesis. 34. Line 400-418: Conclusions need to be rewritten based on the suggestion made along with the manuscript revision.  Table 1 and text -the use of so many decimal digits in non-quantitative data is a nonsense The reviewer advised to provide the chemical formulae to better help readers understand the synthesized cellulose structure. We most definitely agree with the reviewer's advice and edited it accordingly.
The reviewer has kindly pointed out the significance of the non-quantitative data obtained from solid state 13 C CP-MAS NMR analyses of allyl-functionalized cellulose l. 147 -the chemical formula would help nanocrystal (AC) and low allyl-AC (LAC). Previously, these values were used to approximate the % allylation of AC and LAC. However, we entirely agree that a more precise quantitative data should be provided to calculate % allylation. Therefore, we have conducted 13 C Inversegated solid NMR (IG-NMR) for LAC sample and obtained its quantitative peak integration values.
In the current IG-NMR analyses, carbon peak at C1 position in the AC AGU unit was integrated as the standard peak (1.00). The integrated carbons peak values that were assigned to allyl-groups in AC were 0.1363 (C7), 0.9210 (C8), and 1.3199 (C9). Those in LAC were 0.1325 (C7), 0.2159 (C8), and 0.0441 (C9). Using the typical method of calculating degree of substitution of cellulose 42,43 , a complete allylation or the sum of peak integration across C7 -9 positions was assumed to be 3. Therefore, % allylation was obtained by % allylation = (sum of the obtained peak integrated values of C7, C8, C9) 3 In this regard, % allylation of AC and LAC are approximately 79.24% and 13.08% respectively.
Unmodified NMR spectra of AC and LAC are provided in Fig. S1 and S2.
We appreciate the reviewer's concern over the direction of ignition. We have previously tried initiation of frontal polymerization (FP) from the top of the cylinder. However, there were a couple of aspects relating to the reaction that led us to make the decision of initiating the reaction from the bottom. First, when igniting from the top, our AC-DCPD liquid mixture had to be in direct contact with the soldering iron. In this case, foams started to form from the tip of the soldering iron, and we had to hold the iron until the foams were large enough for the iron to be removed from the unreacted, liquid mixture. We believed that this method did not qualify for a 'localized initiation' in which the thermal stimulus would be removed upon observation of FP initiation. Additionally, our earlier experiments measuring frontal velocity raised the possibility that the reactions initiated from the top and bottom may be different. We provide a front displacement vs time plot that compares different initiation methods in Fig. R1.
Although the reaction is unclear, our group concluded that for this study, FP reaction should be initiated from the bottom of the glass tube without making a direct contact between the soldering iron and the AC-or LAC-DCPD liquid mixture. l. 187 -was the more used descending mode (ignition at the top) tried? Why using the ascending one?
Fig. 5a -pDCPD solid is not a correct way to distinguish this sample from the others. Actually, all of them are solid.
As the reviewer kindly advised, we have revised the relevant section in our newly edited manuscript. We greatly appreciate the reviewer's attention to the details in our manuscript.
We definitely agree with the reviewer's comment that the description for Fig. 5a needed a clearer mode of distinguishing each DCPD solids or foams. Grateful for this advice, we have made an appropriate revision on our new manuscript.
We are grateful that the reviewer has kindly raised the necessity to include the definitions of frontal velocity (vf) and maximum temperature (Tmax) and their importance in the reaction of FP. Therefore, we have included further elaboration of the two terms in our new manuscript.
The reviewer has kindly pointed out to the details of reporting the obtained maximum temperature (TMax) values. With the newly measure TMax values, we addressed our observation in line 291 which was shown in our previous edit above.
Once again, we are incredibly grateful for your time and consideration, and we would love to express our immense gratitude for the important questions the reviewer had asked that eventually led us to have discovered more in this study.
Thank you so much. Sincerely,

Prof. Seung-Yeop Kwak
We greatly appreciate the reviewer's concern over the scope of the current manuscript.
With the help of the reviewer's followed comments across our manuscript, the newly found relationship between the formed DCPD/AC foam crystal structures and the unique oxidation resistance of DCPD/2wt% AC foam was highlighted in the later section of our newly edited manuscript. Briefly, by reviewer's suggestion, we have first quantified the levels of oxidized states of the DCPD/AC foams through the ATR-FTIR analysis. Then from the SAXS and WAXS analyses, it was revealed that 2wt% AC induced a formation of a long-range order of its crystalline domains independently from those of DCPD, and that, within the DCPD/2wt% AC foam, a homogeneous distribution of tightly packed crystallite structures relatively suppressed the oxygen diffusion. Based on these findings, the abstract portion of our new manuscript was edited.
Here we show a rapid production of cellular solids by frontally polymerizing a mixture of DCPD monomer and allyl-functionalized cellulose nanocrystals (AC). Our results demonstrated the correlation between increasing % allylation and AC wt%, and the formed Line 32-35: The authors claim that understanding the relation between stereochemistry and crystal structures during FP will be useful to enhance the oxidative stability of pDCPD. It is hard to follow the logic of this claim and is beyond the scope of the current publication, which is about the use of cellulose nanocrystals to make pDCPD foams.
Line 51-52: Authors should describe what characteristics are they referring to.
pore architectures. Especially, with the optimal amount of 2wt% AC, more rapid front velocity and reaction initiation time were observed. From the small-and wide-angle X-ray diffraction analyses, it was revealed that the introduction of 2wt% AC affected the crystal structure of frontally polymerized DCPD/AC foams and thus induced a relative resistance to oxidation.
We certainly agree with the reviewer that the properties rising from the materials microstructures needed much clarity to efficiently convey our contextual messages to the readers.

Line 107-110: The authors claim that an effective internal phase separation would occur caused by the difference in the rate of DCPD crosslinking and nanocrystals cross-metathesis. However, there are no experiments in the paper related to the ROMP and CM rates.
Relevant Section in the New Manuscript: Figure S12 ******************************************************************* Figure  Relevant Section in the New Manuscript: p. 18 Lines 334 -356 *************************************************************************** The x vs. t plot showed stable, linearly increasing graphs of the prepared DCPD/AC specimens, indicating pure FPs ( Figure 5b). As shown in Table 3, vf steadily increased from 0.89 ± 0.2 mm s -1 to 1.7 ± 0.1 mm s -1 as AC wt% increased up to 2wt%, then slightly decreased to 1.1 ± 0.1 mm s -1 when it reached 4wt%. The relative decrease in vf in the case of 4wt% AC may have resulted from the mild interruption in the chain reaction during CM due to the increased amount of non-allylated hydroxyl groups on CNC surfaces, since % allylation in AC did not yield 100% from our result. 55 Nevertheless, similar trend of increasing vf with increasing C=C moieties was reported in a number of previously studied FP systems whose feature was attributed to the increased crosslinking degree. 56,57 Experimentally, this phenomenon may be better understood by comparing the reaction velocities of DCPD/AC, -LAC, and -CTD. Figure  displayed increased vf while DCPD/CTD samples exhibited significant decrease in vf with increasing CTD wt%., presumably due to the hindrance of an uninterrupted reaction by the non-allylic moieties. The impact of % allylation pertaining to the reaction dynamics can also be realized. While FP containing ACs generally showed increasing trend of vf, those incorporating LACs revealed negligible differences between the measured vf values. This phenomenon appears clearer when comparing the smallest wt% of AC, LAC, and CTD in FP of DCPD ( Figure S12b). In this sense, the effect of incorporating highly reactive olefin metathesis group like allyl-moiety into FP of DCPD was translated into an overall increase in vf that exhibited further rise upon the compositional increment of reactive C=C bonds.
We appreciate the reviewer's concern over the feasibility of tannic acid-decylamine (CTD) modified CNCs (CTD-CNCs) to inducing porous structures during FP of DCPD.
Perhaps our group lacked clarity in describing this portion of our manuscript as we tried to deliver the fact that CTD-CNCs, even at 1 wt%, were not able to induce successful phase separation and produce porous structures. However, based on the reviewer's kind advice, our group performed FP of DCPD/CTD-CNC 4 wt% on two separate occasions. Unfortunately, we were not able to initiate FP reaction of the DCPD monomer-4wt% CTD-CNC mixtures. The first problem our group witnessed was in the dispersion of CTD-CNCs. As seen in Fig. R2a, even after treating the monomer-CTD-CNC mixture at the same amplitudes and over the same duration as we did to the monomer-AC/LAC mixture in the current study, our group observed a difficulty in dispersing CTD-CNC NPs. When initiated from the bottom of the glass tube, the reaction did not occur. On the other hand, when initiated from above, in which the phase mostly consisted of DCPD monomer, FP was initiated but did not propagate when the reaction front reached the DCPD-CTD layer that was accumulated to the bottom of the glass tube.
To be assured that the size of the added NPs was not the leading factor of this phenomenon, we freeze-milled the CTD-CNC powder produce even-finer powders. With these, our group seemed to have overcome the aforementioned challenge regarding the dispersion of CTD-CNC powders. However, FP was still not initiated when the soldering iron was indirectly in contact with the bottom of the tube. When we dipped the tip of the soldering iron into the liquid monomer-CTD mixture, FP reaction was initiated from the surface of the tip; however, the reaction did not propagate leaving only a partial amount of the monomer-CTD mixture polymerized ( Fig. R2c) From these results, our group has reconfirmed that without C=C moiety on CNCs, the production of DCPD foams leveraging its chemistry was not possible.
Furthermore, the more the amount of CTD-CNCs was present in a DCPD FP system, the more difficult the reaction initiation. We postulate that the non-allylic moieties decorated on CNC hindered a reaction of uninterrupted frontal ring-opening metathesis polymerization (FROMP) of DCPD monomer.
We agree with the reviewer's comment that raised the necessity for an elaboration of the IG-decoupling sequence in 13 C NMR analysis. Therefore, we have made an appropriate adjustment in our new manuscript.

References
In the case of AC (Fig. 2a) Since an AGU unit of CNC contains three OH groups that could be modified, a complete allylation or the sum of peak integration across C7, C8, and C9 positions was considered to be equal to 3. Therefore, % allylation can be obtained by % allylation = (sum of the obtained peak integrated values of C7, C8, C9) 3 Using this method, % allylation of AC and LAC are approximately 79 % and 13 %, respectively (Table 1).
We certainly agree with the reviewer's comment that the general claim about the effects of nanoparticle dispersity to the reaction product. Therefore, the relevant section was edited appropriately.
Relevant Section in the New Manuscript: Lines 189 -190 *************************************************************************** Both AC and LAC were suspended in DCPD liquid monomer at five different wt%: 0.5, 1, 2, 4 wt%, prior to being thermally initiated for FP. It is widely understood that the NP dispersity greatly contributes to the characteristics of reaction products. We appreciate the reviewer's concern as for the reason why the agglomeration of AC was done in aqueous media (i.e., water), not in liquid monomer. In our manuscript, we mentioned that DCPD was solid at room temperature and becomes liquidous at an elevated temperature (50℃). At a depreciated temperature, the monomer turned back into solid. These conditions made difficult for a TEM analysis confirming the dispersity or agglomeration of AC in DCPD monomer to be conducted. Since the objective of the TEM analysis was to observe the dispersity of AC due to its surface modification and size, we had dispersed our material in water which has a relatively similar low-viscosity (~1 cP) compared to that of the monomer previous work 4 had reported (~1.5 cP). We incredibly appreciate the reviewer's critical comment pointing out the mass reduction that was observed from our result. From the current state of study, the reason for mass reduction could not be fully elucidated. However, from Tables S1 and S2, it may be observed that the density of the produced DCPD/AC and -LAC foams generally decreased for increasing the wt% of AC and LAC, respectively. According to a previous work 5 , the more extensive foaming process (i.e., higher interconnectivity and increased porosity) by the addition of reactive reactants may be related to the decrease in density. Since with the increasing % allylation (LAC, 13%, to AC, 79%) and AC wt% seemed to have steadily decreased the density of the reaction product, we presumed this effect resulted from the more aggressive foaming process which may be observed by the increasing pore sizes and % porosity from the μ-CT analysis.  Table 2, it can be seen that with increasing AC wt%, pores became larger and the interconnectivity between the formed pores became increasingly apparent. On the other hand, as shown in Fig. 3c, DCPD/LAC foams exhibited poor formation of pores thus displayed significantly lower interconnectivity between the pores.

References
In this sense, our group postulated that differentiation of individual but connected pores were not made by the software. Therefore, like it is shown in Fig. S7, we speculated that the slight decrease in anisotropy was due to the pore interconnectivity by the increasing pore sizes.
Therefore, we sought to measure the % porosity along the coronal planes to support our claim.
As shown in Fig. 3e, a clear trend of increasing % porosity and mildly decreasing DA was observed with increasing AC wt%. Unlike DCPD/AC foams, DCPD/LAC foams exhibited significantly lower porosity from which DA stayed relatively constant across the examined samples. By these findings, it was understood that with increasing AC wt%, pore diameters of the formed millichannel increased, while maintaining their directional anisotropy. These results were discussed in the relevant section: page 10, lines 187 -227; of the newly edited manuscript.
Relevant Section in the New Manuscript: Figure 3, Table 2 We appreciate the reviewer's concern over the overall readability of the data presented in Figure 4. Following the reviewer's advice, we have made an appropriate correction in the newly prepared Figure 3.
Relevant Section in the New Manuscript: Figure 3, Table 2 Line 231-232: Reconsider rephrasing for more clarity.
We certainly agree with the reviewer's comment regarding the necessity for correct significant figures in reporting the average diameters of the pores formed in the DCPD/AC foams. Moreover, with the appropriate addition made in Figure 3d, we have also included average diameters of the pores formed in DCPD/LAC foams in Table 2.
Relevant Section in the New Manuscript: Figure 3, Table 2 We greatly appreciate the reviewer's suggestion to provide more clarity in elaborating our claim about the phase separation-mediated foaming mechanism presented in this study.
Based on the reviewer's suggestion, we have made appropriate changes to the relevant section in our newly written manuscript.

Relevant Section in the New Manuscript: Lines 235 -258
In the conventional methods of producing porous DCPD solids, the monomer  Table 2 and reconsider the importance of the inclusion of R2.
96 -99.5wt%) and the minority (AC, 0.5 -4wt%) components were simultaneously polymerized into continuous phases that also created a macroporous architecture in a one-pot FP system.
To support the claim that phase separation drove the production of DCPD/AC foams, our group first hypothesized that no new C=C bonds between the two reactants would have formed in discrete metathesis polymerizations of DCPD and AC. Therefore, solid state 13 C CP-MAS NMR analysis of the produced DCPD/AC foams was conducted to examine any changes (e.g., shifts in δ or peak intensities) to the C=C bond peaks at approximately 130 ppm (C8, C10) in parallel with increasing AC wt% ( Figure S8). Despite the increase in AC wt%, no significant δ shifts or peak intensity changes were observed from the 13 C CP-MAS NMR spectra of the prepared foams (0 -4 wt% AC). On the other hand, the X-ray diffraction (XRD) analysis of the same specimens displayed an increasing trend of intensity pertaining to the cellulose (004) plane characteristic peak 37 at approximately 2θ ≈ 31.7° with the increment of AC wt% ( Figure   S9). Collectively, these results may suggest that the C=C moieties of ACs did not participate in the ROMP of DCPD while being present in the one-pot production of FP-mediated DCPD/AC foams.  Figure S5. Therefore, it is highly unlikely to observe a signal corresponding to crossmetathesis in the pDCPD signal around 130 ppm. Figure S9. (ab) The X-ray diffraction (XRD) patterns of unmodified CNC, AC, neat DCPD, and DCPD/0.5 -4wt% AC foams are shown. With increasing AC wt%, peak intensity pertaining to the cellulose (004) plane became increasingly apparent, suggesting the unlikeliness of ACs being interacted with DCPD undergoing ROMP while being present in the same reaction pot.

Line 231-233: The claim about the no crosslinking between DCPD and AC is hard to support by 13C-NMR because at these concentrations the AC signals are not observed in
We greatly appreciate the reviewer's critical point made over the originally presented solid state 13 C CP-MAS NMR spectra of CNC, AC, and DCPD/AC foams in the explanation of the phase-separation mediate foaming mechanism.
In the case of the CP-MAS NMR analyses of neat DCPD and DCPD/AC foams, we were looking not only looking for potential carbon shifts (δ) but also the changes in the peak intensity at 130 ppm as it corresponded to C=C bonds from the peak assigning of AC, LAC, and DCPD. We hypothesized that if there were chemical interactions between the allyl groups of AC and the crosslinking C=C groups of DCPD, any chemical signs induced by this effect would be observed through the solid 13 C CP-MAS NMR analysis. Without significant changes to those parameters being observed, we conducted an X-ray diffraction (XRD) analysis using the same specimens and observed a characteristic peak of the cellulose (004) plane at around 2θ ≈ 31.7°. Collectively, we presumed that ACs may not have participated in the ROMP Line 237-239: The authors suggest that bubbles are forming from a phase separation phenomenon. It would be important that the authors expand this concept. Where are the bubbles coming from? Are they formed from dissolved gases? Why phase separation produces bubbles in this system? reaction of DCPD while being present in the one-pot production of FP-mediated DCPD/AC foams.
These new explanations were included in the newly edited version of our manuscript at their relevant sections.

Relevant Section in the New Manuscript: Lines 235 -258
Relevant Section in the New Manuscript: Figures S8, S9 We greatly appreciate the reviewer's suggestion to include the elaboration of phase separation and the subsequent bubble formation.
In the edited section of our manuscript, we have compared the already known methods of producing porous DCPD: high internal phase emulsion (HIPE) and chemically induced phase separation (CIPS), to our method. We have also investigated the feasibility of other reaction parameters such as the initiator/reactant degradation, solvent boiling, and dissolution of gas into the liquid monomer/AC mixtures that might have played a fundamental role in the bubble formation. After deducing the factors that were unlikely to have induced foaming of DCPD/AC or -LAC mixtures, our group explored the more traditional foaming process that was widely accepted as the underlying mechanism in the foam production via the HIPE method: bubbles were formed by the evacuation of the individually dispersed droplets during the formation of a new continuous internal phase. Under the presumption that the C=C moiety on ACs were undergoing CM, we sought to trace the remarks of crosslinking by AC allyl groups from the produced foams, the reaction kinetics (e.g., front velocity, initiation time, and TMax).
The more detailed explanation related to the bubble formation by phase separation may be found in the following articles.
Relevant Section in the New Manuscript: Lines 235 -387, Figures 4, 5, Table 3 In the conventional methods of producing porous DCPD solids, the monomer To support the claim that phase separation drove the production of DCPD/AC foams, our group first hypothesized that no new C=C bonds between the two reactants would have formed in discrete metathesis polymerizations of DCPD and AC. Therefore, solid state 13 C CP-MAS NMR analysis of the produced DCPD/AC foams was conducted to examine any changes (e.g., shifts in δ or peak intensities) to the C=C bond peaks at approximately 130 ppm (C8, C10) in parallel with increasing AC wt% ( Figure S8). Despite the increase in AC wt%, no significant δ shifts or peak intensity changes were observed from the 13 C CP-MAS NMR spectra of the prepared foams (0 -4 wt% AC). On the other hand, the X-ray diffraction (XRD) analysis of the same specimens displayed an increasing trend of intensity pertaining to the cellulose (004) plane characteristic peak 37 at approximately 2θ ≈ 31.7° with the increment of AC wt% ( Figure   S9). Collectively, these results may suggest that the C=C moieties of ACs did not participate in the ROMP of DCPD while being present in the one-pot production of FP-mediated DCPD/AC foams.
The voids or bubbles formed in the produced foams may also corroborate our hypothesis that FP-driven phase separation generated the cellular DCPD/AC solids. According to previous studies, these bubbles were generated by the evacuation of the dispersed individual "droplets" during the formation of a new continuous internal phase 47 , the elimination of the dissolved gas or water in monomer 48 , and during solvent boiling or thermal decomposition of initiators 49 . In the case of the current study, the N2-purged reaction environment, and the degradation of either DCPD, AC, or LAC seemed less likely to have induced the formation of bubbles. First, the reaction of DCPD/0.5-1 wt% CTD, which were performed in the same reaction environment as those of DCPD/AC and -LAC, did not exhibit any pore formation.
Moreover, the thermogravimetric analysis (TGA) of AC and LAC ( Figure S10) revealed their robust thermal stability at temperatures (~260 ℃) that were significantly higher than the measured maximum reaction temperatures of this study (~150 ℃, shown in Fig. 6c), deviating from the likelihood of decomposition of reactants to induce bubble formation during FP.
Based on these results, we hypothesized that the bubbles observed in DCPD/AC and -LAC foams were the outcomes of more than one chemical reaction (ROMP of DCPD) that was present in a FP reaction. Specifically, we presumed that the foaming mechanism in this study  Interestingly, however, millichannels became increasingly wider by being merged with one another when AC wt% increased ( Figure S11b t plot showed stable, linearly increasing graphs of the prepared DCPD/AC specimens, indicating pure FPs ( Figure 5b). As shown in Table 3, vf steadily increased from 0.89 ± 0.2 mm s -1 to 1.7 ± 0.1 mm s -1 as AC wt% increased up to 2wt%, then slightly decreased to 1.1 ± 0.1 mm s -1 when it reached 4wt%. The relative decrease in vf in the case of 4wt% AC may have resulted from the mild interruption in the chain reaction during CM due to the increased amount of non-allylated hydroxyl groups on CNC surfaces, since % allylation in AC did not yield 100% from our result. 55 Nevertheless, similar trend of increasing vf with increasing C=C moieties was reported in a number of previously studied FP systems whose feature was attributed to the increased crosslinking degree. 56,57 Experimentally, this phenomenon may be better understood by comparing the reaction velocities of DCPD/AC, -LAC, and -CTD. Figure S12a Figure S12b). In this sense, the effect of incorporating highly reactive olefin metathesis group like allyl-moiety into FP of DCPD was translated into an overall increase in vf that exhibited further rise upon the compositional increment of reactive C=C bonds. Figure 6c shows a reaction temperature profile of neat DCPD and DCPD/0.5 -4 wt% AC for which temperature values were recorded at every two seconds for the duration of the FP reaction. The graphs were characterized by sharp temperature spikes (i.e., TMax) that were observed shortly after the reaction was thermally initiated. Without the application of local stimulus, the reaction temperature remained rather constant at around 21 ℃.
The flatter regions of the graphs indicate the reaction initiation time (tinitiation) or the amount of time the initiation zone was heated with a soldering iron before front propagations were observed. With increasing AC wt%, the tinitiation increased with the exception of 2wt% AC from which the quickest reaction initiation was observed. Nonetheless, the general trend of increasing tinitiation was also observed from the previously reported FP systems with secondary reactive components that slightly lowered the reactivity of reactions. 24,58 Especially, when micro-/nanoparticles were added into these systems, as heat sinks, they absorbed the released heat and also decreased TMax values of the reactions. 21,56,59 On the other hand, our system exhibited small differences between TMax values of neat DCPD and the prepared DCPD/AC formulations at around 151 ℃ (Table 3). We first investigated the size of ACs as a possible factor to have induced such effect to TMax. However, the dimension (w x l) of AC was approximately 14 ± 3 x 88 ± 13 nm (aspect ratio ~ 6.6 ± 2) from the TEM analysis ( Figure S6

Figure 5. FP of DCPD/AC was recorded using an IR-and a digital camera. (a) For increasing AC wt%, the reaction lasted for shorter amounts of time. (b) Front velocities (vf) of each reaction were obtained by plotting front displacement (x) against time (t). (c) The temperature vs. time plots exhibited negligible changes in TMax for increasing AC wt%. On the other hand, a general trend of increasing reaction initiation time (tinitiation) was observed with increasing AC wt%. (d) Reaction
kinetics of 2wt% AC was highlighted as it exhibited the most rapid vf and tinitiation amongst the tested specimens.  Line 239-250: The authors claim that the wrinkles observed in the foams are caused by the polymerization rate difference between ROMP and CM. However, as the authors stated, this behavior had been observed in another system where there is only one type of reaction (ROMP). Therefore, the claim can't be supported by the experiment. We greatly appreciate the reviewer's concern over the discontinuity in the proposed hypothesis and the observation of spin mode from our original manuscript. Based on the suggestion that the reviewer had given us in the previous comments, we have included more corroborated explanation as to how the discovery of spin mode from the lower AC wt% foams intrigued us to further explore the possibility of varying crosslinking degree and determine the heart of the foaming mechanism.
Moreover, in Fig. S11, we have shown the formed voids in the DCPD/AC foams that were produced in 1.5 mm borosilicate capillary tubes. Here, with increasing AC wt%, the walls spiral-like voids started to merge with each other that consequently produced larger pores.
From this, it was understood that the apparent sign of spin mode was not observed from higher AC wt% due to the increased pore sizes by the merging of voids that were produced by the spin Figure S6c to support the increase in the multichannel size.

Line 274: The authors should include the images of other formulations together with
Line 282: The use of an IR camera to measure maximum temperature is not accurate because glass absorbs IR radiation.
mode of FP. Furthermore, as reasoned in our newly prepared manuscript, FP of DCPD/0.5 wt% AC was performed in two different reactors during the investigation of the leading cause of the spin mode. While the inner void structure was visible through the digital OM for thinner (1.5 mm in thickness) foam specimens, those of the larger (25 mm in thickness) specimens required for additional preparation of the sample to visualize its void microstructure.
Relevant Section in the New Manuscript: Lines 235 -387, Figures 4, 5, Table 3 We appreciate the reviewer's suggestion to include images of other AC formulations that were frontally polymerized in borosilicate capillary tubes with diameters of 1.5 mm. We have provided the relevant images in Fig. S11.
Relevant Section in the New Manuscript: Figure S11 We most definitely agree with the reviewer that using an IR camera from a distance for Tmax measurement was not the most accurate method. Thus, we obtained temperature values for every 2 seconds intervals from the inside of the sample using commercially available Ttype thermocouples from Omega Engineering. With these new set of data, we provide a more Line 290-298: Authors suggest the increase in frontal velocity is due to an increase in the amount of monomers containing C=C moieties. To support this claim authors need to demonstrate that the amount of C=C coming from the AC is higher than the amount of replaced C=C from DCPD. Additionally, this suggestion does not explain why the 4 wt% has a lower frontal velocity. Claim about allyl-moieties being more reactive than cyclopentene is misleading because dicyclopentadiene contains both high strain and low strain cyclic alkenes. Therefore, a discussion about the low reactivity of one particular moiety cannot be done separately.
Relevant Section in the New Manuscript: Figure 5 We greatly appreciate the reviewer's suggestion to retest the FP reaction of DCPD/4wt% AC. First, in We greatly appreciate the reviewer's thoughtful suggestion to provide more validity to the reported reaction dynamics values. The appropriate changes were made in the newly prepared manuscript ( Table 3).
Relevant Section in the New Manuscript: Lines 447 -450 The obtained 2D scattering images scans were transformed into one-dimensional (1D) scattering patterns using the azimuthal integration method in which the scattered light signals were averaged along the concentric circle around the incident beam by the wavevector (q) or the azimuth angle (ψ = 360°). 67,68 We greatly appreciate the critical points raised by the reviewer. Based on the reviewer's suggestion, we have made an appropriate change to our new manuscript that described the transformation of 2D SAXS/WAXS patterns into 1D azimuthal integrated patterns. 173° ( Figure 6c). Unlike the scattering patterns of these formulations, that of 2wt% AC exhibited rather periodic pattern of several structural peaks between ψ ~ 29° and 132°, which may suggest a relative homogeneous distribution of crystalline domains along its vertical axis.

Relevant Section in the New
This phenomenon could be supported by the scattering intensity vs. scattering vector (ln(I(q)) vs. q) plot of 2wt% AC. While no intensity peaks were observed from other AC wt%, sharp intensity peaks at qmax = 0.02846 Å -1 and qmin = 0.03070 Å -1 were observed for 2wt% AC ( Figure 6d). By the equation dL = 2π/qmax 68,69 , the crystalline domain spacing (dL) of 2wt% AC was calculated as dL = 3.514 nm. Based on these results, it could be understood that a longrange order of crystalline domains was formed by 2wt% AC that were independent of those generated by DCPD. Then, the effects of introducing ACs to the crystal and lattice structures of DCPD were investigated through the WAXS analysis.
We are grateful for the reviewer's comment on the initially proposed mechanism. After conducting additional experiments based on the reviewer's suggestion, we present further evidence that may support the proposed mechanism of relative oxidation resistance observed from the DCPD/2wt% AC foam.
As shown in Figure 6a, the relative oxidation resistance of DCPD/2wt% AC was confirmed from the ATR-FTIR analysis of neat DCPD and the produced DCPD/AC foams. It was previously studied that the proneness of the DCPD alkene backbone towards oxidative damage was the cause of DCPD foam oxidation. 62 Therefore, hydrogenation of the polymerized DCPD material was mainly highlighted as a method to prevent oxidation for its ability to affect the C=C chain conformation. 63,64 Since our experimental procedure was devoid of hydrogenation posttreatment procedure, we sought to investigate the crystal structure of the formed DCPD/2wt% AC foam.
First, the 2D small-angle X-ray scattering (SAXS) analysis was performed for neat DCPD solid and DCPD/0.5 -4wt% AC foams by frontally polymerizing the specimens in 1.5 mm borosilicate capillary tubes. The SAXS analysis was conducted in a transmission mode where the primary beam was emitted in a perpendicular direction to the vertical axes of the samples (Figure 6b). Here, the collected scattering pattern of neat DCPD was considered the background. Therefore, the analyses of background subtracted DCPD/AC patterns provided the information about the crystalline domains that were formed by the varying amounts of AC.
From the azimuthal integrated SAXS patterns of ACs (Figure 6c), a more homogenous distribution of crystalline domains was observed across ψ ~ 29° and 132° for 2wt% AC, while other AC wt% displayed rather heterogenous distribution of structure peaks. In Fig. R1, we show how our group interpreted the obtained 1D SAXS patterns of the varying AC wt%.
Experimentally, oxidation of DCPD/AC foams was clearly observed from the normalized ATR-FTIR spectra of the prepared samples that displayed pronounced C=O and O-H absorption peaks at around 1700 cm -1 and 3390 cm -1 , respectively, which agreed with the previously reported results (Figure 6a). 62 Unlike those of the other DCPD/AC formulations, however, the IR spectrum of DCPD/2wt% AC exhibited significantly lower intensity of the C=O and O-H peaks, that confirmed the relative oxidation-resisting ability of DCPD/2wt% AC foam.
Line 372-375: Authors need to provide quantitative oxidation and crystallinity values for all the samples to support that they are correlated with each other.
We greatly appreciate the reviewer's concern over the reliability of the previous claim we made in the original manuscript. To support the newly proposed relationship between oxidation of DCPD/AC foams and the formed crystal structures, we subtracted the obtained SAXS intensity of neat DCPD from those obtained from the DCPD/AC foams. Therefore, by presenting background subtracted SAXS patterns of ACs, we have shown the formation of crystalline domains that was purely induced by ACs.

Relevant Section in the New Manuscript: Lines 401 -508
Relevant Section in the New Manuscript: Figure 6c, 6d We greatly appreciate the reviewer's advise to present quantitative evidence that our observation for oxidation may be explained. With the newly presented finds that were elaborated in pages 22 -26, we highlight the calculated crystalline domain spacing (dL = 3.514 nm), lattice spacing (d-spacing) and crystallite size (D) values that were obtained from the SAXS and WAXS analyses.
Line 383: How do the authors know that the 2wt% mixture has a stoichiometric amount of functional C=C and chiral-inducting hydroxyl groups?
Relevant Section in the New Manuscript: Table 4   Table 4. The shifts in 2θ values, lattice spacing (d-spacing), and crystallite sizes (D) were determined by the WAXS analyses of neat DCPD and DCPD/0.5 -4wt% AC foams.

(d-spacing) (nm) (nm)
We appreciate the reviewer's insightful comment pointing out the claim that lacked sufficient corroboration from our original manuscript. Based on the new findings we have made through the reviewer's thoughtful comments and suggestions, the newly edited manuscript described the effect of 2wt% AC to have induced a long-range order of crystalline domains that was independent of the crystalline domains that may have formed by DCPD. Moreover, from the WAXS patterns, we showed that the introduction of 2wt% AC to the crystal structure of DCPD resulted the narrowest lattice spacing and thus the smallest crystallite sizes. Collectively, these results supported the relative oxidation resistance that was observed from the DCPD/2wt% foam. Line 400-418: Conclusions need to be rewritten based on the suggestion made along with the manuscript revision.

Relevant Section in the New
Our group appreciate the reviewer's concern over the validity in the claims we had made in the original manuscript. Based on the new findings, we have made appropriate changes to the relevant section of the new manuscript.
Relevant Section in the New Manuscript: Lines 401 -508 We are grateful for the reviewer's insightful suggestions that we are ascertained of their importance to significantly raising the quality of the current study. Based on all the corrections made in the new manuscript, we have made changes in our conclusions.

Relevant Section in the New Manuscript: Conclusions
In summary, we have demonstrated an efficient method of producing solid polymeric foams by frontally polymerizing DCPD/AC mixture. Two allyl-functionalized CNC samples were prepared by the varying % allylation: 79 % (AC) and 13 % (LAC). Both AC and LAC were dispersed in liquid monomer at different wt%: 0.5, 1, 2, 4. FP of DCPD/AC and -LAC mixtures were initiated by a localized thermal stimulus and cellular solids with varying pore microstructures were produced by controlling the AC and LAC wt%. μ-CT scan analyses clearly demonstrated that for increasing % allylation and wt% of allylated CNCs, the pore size and distribution became larger and increasingly homogeneous, respectively. Moreover, with increasing AC wt%, % porosity of the foams increased while maintaining the relative DA. The FE-SEM analysis and the obtained vf, TMax, and tinitiation values displayed the effects of varying crosslinking degree to the construct of the formed DCPD/AC foams by the amount of C=C Line 385-387: The authors use the word "synergistically" to explain their claims about the organization of the CN. However, a better discussion is needed to support this hypothesis.