Large-scale assembly of isotropic nanofiber aerogels based on columnar-equiaxed crystal transition

Ice-templating technology holds great potential to construct industrial porous materials from nanometers to the macroscopic scale for tailoring thermal, electronic, or acoustic transport. Herein, we describe a general ice-templating technology through freezing the material on a rotating cryogenic drum surface, crushing it, and then re-casting the nanofiber slurry. Through decoupling the ice nucleation and growth processes, we achieved the columnar-equiaxed crystal transition in the freezing procedure. The highly random stacking and integrating of equiaxed ice crystals can organize nanofibers into thousands of repeating microscale units with a tortuous channel topology. Owing to the spatially well-defined isotropic structure, the obtained Al2O3·SiO2 nanofiber aerogels exhibit ultralow thermal conductivity, superelasticity, good damage tolerance, and fatigue resistance. These features, together with their natural stability up to 1200 °C, make them highly robust for thermal insulation under extreme thermomechanical environments. Cascading thermal runaway propagation in a high-capacity lithium-ion battery module consisting of LiNi0.8Co0.1Mn0.1O2 cathode, with ultrahigh thermal shock power of 215 kW, can be completely prevented by a thin nanofiber aerogel layer. These findings not only establish a general production route for nanomaterial assemblies that is conventionally challenging, but also demonstrate a high-energy-density battery module configuration with a high safety standard that is critical for practical applications.

The paper is also filled with exaggerated and hype statements which I found annoying (« excellent », « outstanding », etc.) and not very scientific. Great science does not need hype.
Specific comments: -the term « snow » is fancy but misleading. Snow crystals are grown in very different conditions with a different physics (solid/vapor equilibrium). I think the authors should use a different term.
-line 59-60: « it is challenging to scale up the thick bulk material fabrication while maintaining consistent properties ». This challenge is supposidely solved here but actually not demonstrated. No mention is made of the maximum thickness obtained, for example. From the figure 1b and f, it seems that the thickness is of ~7-8cm (?), which is about twice larger than what has been demonstrated by conventional freeze casting (pieces several cm thick have already been reported). What is the maximum thickness achieved here? Scaling up the lateral dimensions is not so challenging (several papers have already been published on this), the challenge in scale-up so far has always been the increase in thickness, which is not clearly reported and discussed here. From what is exposed here, I suspect it is actually not so simple, since the « snow » is mixed with additional suspension into a block that needs to be frozen completely, as in conventional freezecasting. These limits are not discussed in the manuscript. -l72: « these building blocks often results in interconnected macro-scale open pores, whose dimensions are greater than the standard free path of air ». The authors imply here (and later in the paper) that their approach is better in this regards. However: -no pore characterization and pore size distribution is reported here -the SEM pictures reveal that the porous structure is clearly macroporous (micro-or mesopores could also be present but not visible on the SEM pics) and interconnected. I thus don't see anything unique or specific in these structures. Such strong claims should be supported by experimental data, which is not the case here. -line 87: the « snow » is mixed with additional suspension and cast into blocks that need to be frozen completely. Is the additional suspension cooled before being mixed with the snow ? Is the mix homogeneized? Are there any precautions to take to avoid thawing? How much pressure is applied to packed the mix and ensure no pores are left? No mention is made of how long it takes to freeze such large blocks. No mention is made either about the required freeze-drying time. Too little informations about this important step are provided in the manuscript. -line 89: « the SC process was was easily scaled in a linear fashion to over 1.2m2 area without alteration ». Where are the data supporting this statement? I only see a (nice) picture of a very large piece, but I don't see a comparison of the microstructure/properties for pieces of increasingly large dimensions.
-line 100 (and other places): many claims are made about the supposed advantages of the process, but these advantages are not really discussed, and no mention is made of the limits. For example, the SC process is more complex (more steps) than the standard freeze-casting process. The first freezing step may be faster, but a second freezing stage is required. Overall, it is very likely that the overal complexity and thus cost of the process is at least similar if not greater than conventional freeze-casting.
-line 104 « This was noteworthy as SC technology was incompatible with slurry-forming technology ». It don't understand this statement, could you please elaborate? -line 122: « in a snow slurry system, the ice crystal tend to nucleate simultaneously ». If I understood correctly the process, the « snow » scrapped from the rotating drum is collecting and packed. The ice crystals are thus already present in the snow and not thawed at this stage. The ice crystals in the snow thus continue growing, there is no need to nucleate novel crystals. -line 127: « we sucessfully realize complex architectures in snow-casted materials ». It's not particularly complex, it's either random (and somewhat isotropic), or already reported before (e.g. Fig S6).
-line 136 « this structure exhibited a special micro-orientation and macro-isotropic nature ». I don't see what's special (what does « macro-isotropic » means?) here, more explanations should be provided. The structure shown in Fig S6 is typical of freeze-casted fibrous structures (e.g. freeze-casted carbon nanotubes structures), which has been reported in numerous papers before. It's nice but I don't see anyting special or novel here. Overall, the paper is not focussed. Too little details are provided about the novel process and its control and limits, and too many informations are provided about a myriad of very different materials. This variation of freeze-casting appears thus promising but too little informations are provided about the process at this stage to make it a convincing, real advancement.
Reviewer #3 (Remarks to the Author): The authors describe the methodology and successful synthesis of an ultralight, large-scale, thermal super insulating, and flexible nanofiber aerogel using a continuous rotating ice crystallization. The team also demonstrated the capability of a 5 mm thick film to decrease the risk of thermal propagation, which is a major challenge with lithium-ion batteries and an area of the field with a continuous need of advancements.
I recommend the authors to discuss the challenges and trade-offs of mitigating the risk of propagation and the added weight. What increase in parasitic mass and volume ratios do the nanofiber aerogel film introduce to the cell stack? Parasitic mass ratio is defined as mass of cell stack divided by mass of cells only. This may show that adding thermal capacitance to the cell stack is not an effective mitigation strategy for propagation resistance.
I recommend the paper for publication.
According to these comments and suggestions from three referees, we have carefully revised the manuscript with all the changes highlighted. The comments are reproduced and our responses are given directly in different color (blue). The pointby-point responses to Reviewer #1, Reviewer #2, and Reviewer #3 are listed in the following pages. All page numbers refer to the revised manuscript file with tracked changes.

Responses to Referee #1
This manuscript presents a new concept of snow-casting (SC) technique, which is combined with a continuous rotating ice crystallization (RIC) technology, to solve the problem of non-uniform temperature gradient occurring in the conventional freeze casting method. This work also demonstrates the feasibility of SC method to produce ultralight, large-scale, thermal super-insulating (0.020 W/m·K) and flexible Al2O3·SiO2 nanofibers aerogels (ASNAs), which can find potential applications in aircraft, unmanned aerial vehicles and flying cars. In addition, the ASNAs are claimed to have potential to enhance the battery safety by blocking the thermal runaway in modules when used as separator between cells. The concept of SC also offers high flexibility of the shape of final products, expanding the range of applications. The manuscript has sufficient overall novelty suitable for publication in the Journal. However, there are several critical issues that need to be addressed before it can be reconsidered for publication.

Comment 3:
The time taken for ice crystal formation using SC is compared with that using unidirectional freeze-casting (FC). Because both the freezing temperatures used and the microstructures produced from these two methods are different, the authors should justify the choice of unidirectional FC for comparison. In addition, the unidirectional FC technique produces frozen samples in the final shape in one-step from the solution, whereas the SC involves a RIC process to prepare snowflakes and second freeze-casting in a cryogenic environment where the nanofiber suspensions are mixed with snowflakes to eliminate the gap between snowflakes. The entire fabrication times should also be compared rather than the time for ice nucleation alone. Essentially, it is difficult to make a direct comparison between the two techniques as each of these techniques stands alone with different approaches.
Response: Thank you very much for your insightful comments and suggestions. In general, nondirectional freeze-casting is a common method for preparing huge ice block. However, the nondirectional freezing process usually encumbered by the great nucleation resistance and low heat transfer capacity owing to the relatively low thermal conductivity of ice. We often observe phenomena of supercooled water (liquid water below 0 ℃) and hollow ice block (center is liquid). This process is timeconsuming and inefficient (Fig. 1e), and has been replaced by the induction nucleation of metal materials, which can also be called unidirectional freezing.
In unidirectional freezing process, the heat conduction is in the axial direction and the heat is barely not conductive in the vertical direction. Specifically, when the slurry was frozen on a cryogenic plate, ice nucleation and growth can expel nanofibers to the surface of crystal column to form the aligned walls. The nanofibers were then formed a long-range lamellar structure and cast a solid anisotropic network. This method combined with formation mechanism and regulation strategies has been widely reported in various literatures [1][2][3][4][5][6][7][8][9][10][11][12] . At the experimental level, liquid nitrogen is usually used to fabricated the aerogels. However, once expanded to largescale samples, the controllability and cost of this method will face huge challenges.
As the reviewer noted, we compared the ice block production rate between the conventional undirectional (one step) and directional freezing (one step) wi crushed ice casting (two step). The same mold size (18×18 cm 2 in-plane) and freezing environment (−20°C) were adopted. The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of undirectional freezing process,when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5).

Revise details:
 Page 5, Line 20, added "The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5)."  Page 17, Line 9, added in Methods, "When comparing the ice block production rate of the conventional undirectional and directional freezing to crushed ice casting, the same mold size (18×18 cm 2 in-plane) and freezing environment (−20 °C) were adopted."  Fig. 7).
However, when the fiber concentration is too low (less than 0.1%), the cell wall consisted of the nanofibers will be loose during the subsequent freezing process, and it is even difficult to connect each other. When the fiber concentration is too high (higher than 1.8%), the nanofibers will be agglomerated and difficult to evenly disperse in the solution, and the obtained cell wall will be highly compact. Moreover, the thermal conductivity and elastic modulus of the obtained nanofiber aerogel increase as the density increases (Supplementary Table 1 Revise details:  Page 12, Line 26, we added, "We also compared the thermal conductivity of our ASNF aerogels with those of the lamllar sponges and anisotropic aerogels along the axial and transverse direction (Fig. 4g) Revise details:  Page 12, Line 26, the "~20 mg cm −1 " has revised as "~20 mg cm −3 ".

Responses to Referee #2
The authors expose in this manuscript a variation of ice-templating enabling the production of large (m 2 range) pieces, with isotropic porous microstructures. The applicability of the process is then demonstrated with different materials (proof of concept).
The novelty of the process is its division into two freezing steps: -A first step where only small bits of suspension (or solution) are ice-templated and scrapped from the cold surface repeatedly. This ensures the relatively rapid production of a large volume of frozen bits (which the authors refer to as "snow").
-A second step where the "snow" is mixed with additional native suspension (to fill the gaps between the snow bits), and frozen, before being freeze-dried.
Since the novelty is in the process itself, I would have expected a deep, parametric investigation of the process and the influence of the various parameters, that would have provided a good overview and understanding of the process, its possibilities, control, and limitations. Instead, only a small part of the paper (approx. 20%) is devoted to the process, and almost no information and data are provided about its control and properties. The majority of the paper is about the proof-of-concepts materials obtained with this process. Although it's nice to demonstrate what the process can do, this poor balance makes the paper confusing and a bit shallow.
Examples of these points are discussed below.
Response summary: Thank you for your constructive, detailed and accurate comments and suggestions. We also appreciate for your interest in the whole article including reading the details, understanding our ideas, and suggesting to the manuscript. In the past three months, we have carefully gone through and discussed these thoughtful and meaningful comments with our colleagues. We have gained a lot including the preciseness of writing, scientific depth, material performances, and experimental details. We have revised the manuscript according to the reviewer's suggestions. Moreover, we have carefully addressed, point by point, all the comments the reviewer has raised. We hope that the reviewer will find our responses satisfactory and convincing.
As suggested by the reviewer, the structure of the paper has been readjusted and we have increased the content proportion of crushed ice casting method and formation mechanism. We have revised the manuscript according to the reviewer's suggestions and added the relevant experimental results mentioned by the reviewer. For example, we added the effect of roller speed on output and the ratio of ice to integrating agent on freezing time and the final structure. We also added the section headings to enhance the readability. A logical thinking chain of "preparation-mechanismperformance-application" has been formed.

Revise details:
 In Page 4, Line 1, we added, "The rotating speed was set as 50 rad min −1 , owing to a higher speed will cause the incomplete freezing of the nanofiber dispersion." According to experimental research, a 5:1 mass ratio for ice crystal and dispersion is appropriate for the following blending and freezing."  Page 5, Line 22, we added, "The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5). The whole process time from raw materials to products is also much lower than the traditional fiber felt/silica aerogel composites (Supplementary Table 1  Page 12, Line 4, we added section headings, "Thermal Insulation performance."  Page 13, Line 26, we added section headings, "Proof-of-concept experiments."

Comment 1:
The paper is also filled with exaggerated and hype statements which I found annoying ("excellent", "outstanding", etc.) and not very scientific. Great science does not need hype.

Response:
We are thankful for your constructive comments and suggestions. We have double-checked the manuscript and deleted all the exaggerated and hype statements (such as "excellent", "outstanding", "unique", and "remarkably") in the revised manuscript to ensure that all the clams revised on properties and performance are scientific and objective.
 Page 16, Line 24, "excellent" has been revised as "high". Response: Thank you for your important comments and suggestions. We agree that the term "snow" is inappropriate to describe the product in the RIC process. During the RIC process, the nanofiber dispersion was rapidly frozen on the surface of a rotating cryogenic drum to form ice with large area and scraped into crushed ice with small sized by the blade on the machine. We have changed the word into "crushed ice" in the revised manuscript.

Revise details:
 All "snow" or "snowflakes" has been revised as "crushed ice". From what is exposed here, I suspect it is actually not so simple, since the "snow" is mixed with additional suspension into a block that needs to be frozen completely, as in conventional freeze-casting. These limits are not discussed in the manuscript.

Response:
Thank you for your meaningful comments and suggestions. We are very sorry for bad expression on the progress in the improvement of ice crystal growth mechanism in thickness direction using crushed ice casting. We agreed that the challenge in scale-up so far has been the increase in thickness. Most surface growth processes would gradually slow down or even stop, owing to the gradual vanish of concentration or temperature gradient and reduction of mass or heat transfer. We must emphasize that increasing thickness is not the core of difficulty, but maintaining freezing efficiency. For conventional freezing methods, the greater the thickness, the lower the heat transfer efficiency, and the higher the energy and time costs involved in manufacturing materials.
(i) Firstly, we compared the ice block production rate between the conventional nondirectional and directional freezing with crushed ice casting. Compared with the traditional ice-template method, our crushed ice casting method The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block.
(ii) We achieved the decoupling of the ice nucleation and growth processes. The 3D expansion of the growth sites will greatly accelerate the crystallization process, outperformance the 2D nucleation from directional freezing, especially in large-scale manufacturing. Moreover, we confirmed the high energy transfer efficiency of rotating freezing in the preparation of crushed ice through computational fluid dynamics (CFD) simulations 16 . The principle of heat and mass transfer enhancement in this thin-film rotating reactor has been widely reported [17][18][19][20] .
(iii) To probe the mechanism differences between directional freezing and crushed ice casting, we in-situ observed the cooling process under an optical-fluorescence microscope. We recorded the temperature profile during cooling, the slurry temperature decreased linearly before ice nucleation and a sudden temperature rise can be observed when the slurry temperature reached 1.3 °C, which is attributed to the heat release upon nucleation (Fig. 2b). After 24 seconds, the cooling slope is almost identical with initial state, indicating the crystallization process has been completed.
However, in the case of crushed ice casting strategy, all pre-existing and 3D random distributed ice crystals can act as the primitive growth sites. The adjacent crystals will unconstrainedly grow with no favorable location and orientations, and ultimately form a multidomain bulk materials. The ultrafast crystallization process (1.2 seconds) obtained from sequential optical images further confirmed this phenomenon.
(iv) According to classical crystal theory, the transformation of liquid into solid at low temperature is divided into two parts: Nucleation requires a certain degree of undercooling and solidification environment; and the growth rate is proportional to the number of nucleation. On the basis of the above theory and observations, we propose here a possible mechanism for enhancing the cooling efficiency and controlling the 3D structures by crushed ice casting.
We successfully combine these two mechanisms to achieve isotropic architectures in ice-templating materials. (i) Ice crystals rapid nucleation on rotating cryogenic drum can effectively increase heat transfer efficiency. (ii) Multi-point growth from three-dimensional distribution of crystal nucleus can effectively accelerate the formation of bulk crystals, resulting in a macroscopically cellular distribution of nanofibers.
(v) Through blending the crushed ice and the dispersion evenly, the ice crystals are distributed uniformly and randomly after the mixture is poured into the pre-fabricated mold, and grow almost simultaneously in the freezing environment, thus ensuring the consistency of the structure as the size increases.

Revise details:
 Abstract, in Page 1, Line 29, we added, "Herein, we describe a general icetemplating technology through rotating freezing, crushing, and re-casting nanofiber slurry based on decoupling the ice nucleation and growth processes."  In Page 2, Line 26, we deleted, "It is challenging to scale up the thick bulk material fabrication while maintaining consistent properties matching the variable application scenario."  In Page 5, Line 19, we added, "Compared with the traditional ice-template method, our crushed ice casting method decouple the ice nucleation and growth processes. The 3D expansion of the growth sites will greatly accelerate the crystallization process, outperformance the 2D nucleation from directional freezing, especially in thick sample manufacturing. The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block." pores, whose dimensions are greater than the standard free path of air". The authors imply here (and later in the paper) that their approach is better in this regard. However: -no pore characterization and pore size distribution were reported here -the SEM pictures reveal that the porous structure is clearly macroporous (micro-or mesopores could also be present but not visible on the SEM pics) and interconnected.
I thus don't see anything unique or specific in these structures. Such strong claims should be supported by experimental data, which is not the case here.
Response: Thank you for your constructive comments and suggestions. I am sorry the demonstration of our pore structure is unreasonable for reading. As the reviewer noted, we have carried out more detailed characterization of the pore structure of the isotropic ASNF aerogels. The size of pores in isotropic ASNF aerogels are contributed to the ice crystals and the compositions. According to IUPAC classification, the pores are classified in to the following groups according to their diameter: those with size less than 2 nm are called micropores; the size larger than 50 nm is called macropore; As suggested, we tested the area pore size distribution of isotropic ASNF aerogels by an automated mercury porosimeter. This is a standard test method for pore size distribution of porous materials. The obtained average pore size is 39.86 μm, and the pore size is mainly distributed in the range of 6 to 106 μm, which is corresponded well with the ice block size. In additional, the nanofiber cell walls of ASNF aerogels consist of numerous mesopores due to the adding of commercial silica aerogel powders, which exhibited a small pore size of 2-4 nm, as shown in Supplementary Fig.   18, characterized by the nitrogen sorption isotherms test. These results indicate the ASNF aerogels are mainly consisted of macropores, while some mesoporous materials are embedded in the structure.

Revise details:
 In Page 1, Line 29, "These highly porous skeletons could effectively reduce solid heat conduction. However, these building blocks often result in interconnected macro-scale open pores, whose dimensions are greater than the standard mean free path of air (69 nm); and fail to restrict the free movement of molecules and mitigate the gaseous heat conduction." have been replaced by "However, because of the channel dimensions are much greater than the standard mean free path of air (69 nm), which fail to restrict the free movement of molecules and mitigate the gaseous heat conduction 21 ."  In Page 9, Line 21, we added, "The superelastic characteristics and negative Poisson's ratios are mainly derived from the bonded nanofiber networks and interconnected separate cellular structures ( Supplementary Fig. 11)."  In Page 12, Line 18, we added, "The macropore size distribution of ASNF aerogels was characterized by an automated mercury porosimeter (Fig. 4f). The average size of macropore in the ASNF aerogels was 39.86 μm and the porosity was estimated as 90.73%."  In Page 12, Line 23, we added, "Contrasting the nitrogen sorption isotherms of these materials to our aerogels showed that our aerogels had the highest Brunauer-Emmet-Teller (BET) specific surface area (21.89 m 2 g −1 ) and Barrett-Joyner-Halenda (BJH) average pore sizes (3.86 nm), owing to the existence of commercial silica aerogel powders (Supplementary Fig. 18 and Table 3)."  We added Supplementary Fig. 11.  (1) In the fabrication process, the additional suspension was not cooled before being blending with the crushed ice. Normal temperature dispersion is sufficient because the mass ratio of dispersion to crushed ice is 1:5.
(2) For typical experiments, a 200 g salad-like ice-dispersion mixture can be obtained after half-minute blending using a thermally insulated mixing tank and customized agitator paddles. The time spent in this process can be ignored for the whole process.
(3) Supplementary Fig. 24 show the optical images of crushed-ice-dispersion mixture. In the blending process, the ice will be further refined and melting under the action of external force, ensuring the uniform distribution of the ice and water. Partially melting of ice crystallization are inevitable during the mixing of ice and solution, although we use thermally insulated containers. The homogeneity of the mixture is confirmed by the stability of the mechanical properties of different parts at a large-scale sample.
(4) Owing to the salad-like ice-dispersion mixture has a certain fluidity, it can fill the mold without leaving pores. times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5).
(6) We added the freeze-drying time in Supplementary Table 1. We compared the whole process time from raw materials to products between isotropic ASNF aerogels and the traditional fiber felt/silica aerogel composites (Use 300 × 300× 10 mm 3 sample as standard). The whole process time from raw materials to products is also much lower than the traditional fiber felt/silica aerogel composites (Supplementary Table 1) 13 .

Revise details:
 In Page 5, Line 21, We added, "The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5). The whole process time from raw materials to products is also much lower than the traditional fiber felt/silica aerogel composites (Supplementary Table 1    Response: Thank the reviewer for the valuable comments. With respect to the production of large dimensions, the fabrication process was consistent with the process for pieces of small dimensions, including spinning, homogenizer, freezing into crushed ice, freezing dryer and calcination. Therefore, our preparing method of the nanofiber aerogel samples is highly scalable (Fig. R1, these images have been included throughout the article).
In the case of fabrication process for large-scale aerogels, there are nothing changed except the enlargement capacity of the freeze-dryer. Thus, the process still followed the forming mechanism of crushed-ice-template process. As shown in Supplementary   Fig. 8, we have produced ASNAs with size of 35.0×34.0×2.5 cm 3 . Then five positions were randomly selected on the large size samples to characterize the morphology, mechanical properties and thermal properties, thus confirming that the method is applicable to the preparation of large size samples and will not affect the structure and properties.

Comment 7:
Many claims are made about the supposed advantages of the process, but these advantages are not really discussed, and no mention is made of the limits. For example, the SC process is more complex (more steps) than the standard freezecasting process. The first freezing step may be faster, but a second freezing stage is required. Overall, it is very likely that the overall complexity and thus cost of the process is at least similar if not greater than conventional freeze-casting.
Response: Thank you for your constructive comments and suggestions. Indeed, we claim too many advantages, but insufficient evidence and disadvantages in the previous version. We have added discussion the advantages and disadvantages of the crushed ice casting method to fabricate the aerogels in the revised manuscript.
(1) In contrast with the traditional lamellar aerogels formed by directional freezing, our method decouples the nucleation and growth processes, thus, the preparation process is relatively complex. However, the directional freezing process usually encumbered by the great nucleation resistance and low heat transfer capacity owing to the relatively low thermal conductivity of ice. In the crushed ice casting process, the 3D expansion of the growth sites will greatly accelerate the crystallization process, outperformance the 2D nucleation from directional freezing, especially in large-scale manufacturing.
Moreover, we confirmed the high energy transfer efficiency of rotating freezing in the preparation of crushed ice through computational fluid dynamics (CFD) simulations 16 . The principle of heat and mass transfer enhancement in this thin-film rotating reactor has been widely reported [17][18][19][20] .
Therefore, the total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5).
(2) Past attempts to modulate thermal insulation properties focused on developing a family of aerogels using directional freezing method based on  Revise details:  Page 5, Line 20, added "The total required freezing time is 1.38 times shorter than that of conventional directional freezing process, and 1.92 times shorter than that of nondirectional freezing process, when producing 1.2 kg ice block ( Fig. 1e and Supplementary Fig. 5)."  Page 17, Line 9, added in Methods, "When comparing the ice block production rate of the conventional undirectional and directional freezing to crushed ice casting, the same mold size (18×18 cm 2 in-plane) and freezing environment (−20 °C) were adopted." We are sorry for the careless mistake. We agree that the ice crystals continue growing rather than nucleating. Therefore, the freeze process will be faster due to the omission of nucleation time. As noted, we have revised the relevant statements. As mentioned in the manuscript, when freezing slurry with the same mass, freezing the mixture of crushed ice and slurry will significantly increase the freezing rate, thus improving the efficiency of sample growth and preparation.

Revise details:
 In Page 7, Line 14, we added, "However, in the case of crushed ice casting strategy, all pre-existing and 3D random distributed ice crystals can act as the primitive growth sites."  In Page 7, Line 16, we added, "Owing to no nucleation process exist in the system, the setting temperature and slurry temperature are basically consistent (Fig. 2g). Moreover, the ultrafast crystallization process (1.2 seconds) obtained from sequential optical images further confirmed this phenomenon ( Fig. 2h, i, and Supplementary Video 5)." Response: Thank you very much for the insightful comments. We agreed that the structure is not particularly complex, and similar structure has been reported using foaming method (bubble templated) 22-24 , cellulose-based nondirectional freezing method 25 , and freeze-casted carbon nanotubes structures 26, 27 , but not common on ceramic nanofiber aerogels. Indeed, this expression is inappropriate in the manuscript.
We agreed that the structure is not particularly special. Similar claim in the manuscript has been revised.
The "macro-isotropic" means that the macroscopic materials obtained by the  Revise details:  All the "fatigue resistance" and been replaced by "damage tolerance". Response: Thank you for your valuable comment and suggestions. As mentioned in the manuscript, the densities of the ASNF aerogels can be regulated by changing the concentrations of the precursor dispersions. This process is highly reproducible, which has also been confirmed in many other literature 5, 10, 28-34 . We have measured the correlation between density and concentration, as shown in Supplementary Fig. 7.
The concentrations of the precursor dispersion can range from 0.05% to 1.8%. For concentrations overtop 1.8%, the high concentration of nanofiber makes the agitation resistance great, which leads to the fiber is difficult to disperse evenly. For concentrations below 0.05%, too few fibers make it difficult to construct an effective cell wall, leaving the structure loose or barely formed. Therefore, the sample will lose its elastic properties and be prone to result collapse.

Revise details:
 Page 7, Line 25, we added, "To investigate mechanical property of the obtained ASNF aerogels, we first systematically investigated the effect of density regulation and size enlargement on the mechanical properties of the resultant materials. The densities of the ASNF aerogels can be readily regulated by changing the concentrations of the precursor dispersions, and the minimum density achieved is 0.59 mg·cm −3 (Supplementary Fig. 7).
However, when the fiber concentration is too low (less than 0.1%), the cell wall consisted of the nanofibers will be loose during the subsequent freezing process, and it is even difficult to connect each other. When the fiber concentration is too high (higher than 1.8%), the nanofiber swill be agglomerated and difficult to evenly disperse in the solution, and the obtained cell wall will be highly compact. Moreover, the thermal conductivity and elastic modulus of the obtained nanofiber aerogel increase as the density increases (Supplementary Table 1). Eventually, to achieve both low thermal conductivity and high elastic modulus, a density of 5 mg cm -3 was selected for test samples.".  and then select the appropriate process conditions. Therefore, the number of experiments can be greatly reduced, thus saving human, material and financial resources. Therefore, it is necessary to use the model to simulate the process.

Comment 15:
Overall, the paper is not focused. Too little details are provided about the novel process and its control and limits, and too many information is provided about a myriad of very different materials. This variation of freeze-casting appears thus promising but too little information is provided about the process at this stage to make it a convincing, real advancement.
Response: Thank you for your constructive, detailed and accurate comments and suggestions. We also appreciate for your interest in the whole article including reading the details, understanding our ideas, and suggesting to the manuscript. In the past three months, we have carefully gone through and discussed these thoughtful and meaningful comments with our colleagues. We have gained a lot including the preciseness of writing, scientific depth, material performances, and experimental details. We have addressed these comments point by point by performing additional theoretical and experimental studies. We believe that with your help, the quality and impact of the revised paper have been evaluated to a higher level. Although we have made a lot of efforts, there were still some errors and omissions in the manuscript. We hope you can continue to point out them.

Responses to Referee #3
The authors describe the methodology and successful synthesis of an ultralight, largescale, thermal super insulating, and flexible nanofiber aerogel using a continuous rotating ice crystallization. The team also demonstrated the capability of a 5 mm thick film to decrease the risk of thermal propagation, which is a major challenge with lithium-ion batteries and an area of the field with a continuous need of advancements. Traditional fire walls will add weigh and volume significantly, thus diminishing gravimetric and volumetric energy densities of the battery. While for phase change materials, one of the main disadvantages is the flash point of the customarily employed kinds of paraffin, which contribute significantly to the fire load. In general, air cooling system is limited by the low thermal conductivity and small specific heat capacity of air. Although the use of electric fans will increase the airflow rate, it will lead to additional energy consumption costs. While for the liquid cooling system, the flow channels and cooling medium could indeed result in large space occupation and weight increasing. And the hoses, pipes, or mini channels will increase the complexity, which is costly and will occupy a large volume in the battery module and pack.
For ceramic fiber aerogels prepared by the crushed ice template method, we calculated the increase of parasitic mass and volume ratio after adding aerogels according to the comments of reviewers. For the sample with a thickness of 0.5 cm, the density of the sample is 5 mg cm -3 , which can be ignored compared with the mass of the battery pack. The ASNF aerogels could completely block the TR propagation without additional assistants and specific energy loss. The volumetric energy density of the module is determined as ∼677.6 Wh L −1 in the initial state and ∼589.3 Wh kg −1 with ASNF aerogels.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): The revised manuscript has addressed most of our comments and the quality has been improved significantly. However, there remain a few comments/suggestions as follows.
1. On Comment 3: Although the authors provided the total freezing times for different freezing methods, it is still a concern of making direct comparison between different approaches although these samples have distinct microstructures. 2. In view of the capability for large-scale production of aerogels, the most critical limitation of conventional freeze-casting technique is the maximum freezing distance with uniform properties. As such, the freezing distance should be considered the parameter for comparison rather than weight of ice block (Fig R1e). In addition, there is not much time difference in ice block weight less than 800g between the directional freezing and crushed ice casting. Explain why. 3. Fig. 4g: The thermal conductivity of lamellar sponges in the axial direction (circle) is found lower than that in the transverse direction (triangle) at all temperatures studied. Explain why and provide more convincing discussion. In addition, anisotropic aerogels have been proven to have better insulation performance than their isotropic counterparts due to the improved thermal management by dissipating the heat in the axial direction and inhibiting the heat conduction in the transverse direction [J. I went through the revised version of the manuscript. Although the authors replies to the points raised, I found many of these arguments still unconvincing and/or based on unfair comparisons. Regarding the freezing process itself, for example, which is supposedly the central novelty of the paper, the key argument proposed by the author is that the overall freezing time is faster, and this should enable commercialization of these materials. I think the authors do not have sufficient understanding of the costs and hurdles of commercializing novel materials. Freezing is (energywise) relatively cheap, the expensive part is the drying, which is the same in both cases (previous work and current work), and the cost of the process is just a small part of the equation to decide whether a novel material can be commercialized succesfully or not. Besides, the authors already mentioned in the introduction that freeze-casting is low cost. The main benefit of the novel approach (claimed by the authors) is that pieces of large dimensions can be processed, but the same dimensions can be obtained with standard freeze-casting. The scientific advances are thus incremental at best, and there are still unfortunatly still little exploration of the science behind the proposed approach. The proposed explanations (decoupling of the nucleation and growth phases) are not very convincing either, as the control of nucleation in « standard » freeze-casting have been achieved with many different strategies (including epitaxial growth of ice from a pre-existing ice template).
The key idea upon which the paper is based (preparing ice crystals and then mixing them with the solution) has already been reported in many papers: see The only novelty (in terms of process) that I see here is thus that it's made at a larger scale than before, which again is nice but very incremental in my opinion.
Overall, although the claimed novelty is on the process (bearing in mind my concerns above), it still represents only 20% of the paper or so, the rest is dedicated to different materials obtained by the process, I thus still find the paper poorly balanced.
Other comments: -non-scientific expressions are still present throughout the manuscript (e.g. « salad-like », which does not mean anything, « ultralight feathers » ?) -my concerns about reproducibility have not really been adressed, I still don't see error bars in Fig  S7a, for example.
-non-standard tests (Fig 3a) are still present. These are uninformative since we cannot compare the behavior with that of previously reported materials. -Fig 4g and 4i proves that the thermal conductivity properties are essentially the same than previously reported, which is not surprising since macropores dominate the thermal behaviors in both cases. There are thus no improvements here.
-the discussion section is actually the conclusion -the comparison of the freezing time (Fig 2b and 2g) are misleading, since the starting temperatures are different in both cases ! If the standard freeze-casting was performed by starting with a suspension at 0°C, the total freezing time would be much shorter.

Comment 1:
The revised manuscript has addressed most of our comments and the quality has been improved significantly. However, there remain a few comments/suggestions as follows. Although the authors provided the total freezing times for different freezing methods, it is still a concern of making direct comparison between different approaches although these samples have distinct microstructures. In view of the capability for large-scale production of aerogels, the most critical limitation of conventional freeze-casting technique is the maximum freezing distance with uniform properties. As such, the freezing distance should be considered the parameter for comparison rather than weight of ice block (Fig R1e). In addition, there is not much time difference in ice block weight less than 800g between the directional freezing and crushed ice casting. Explain why?

Response:
We are thankful for the constructive comments. As the reviewer suggested, we have deleted the comparison of the freezing time between two different approaches.
We are more concerned about the time required to form ice blocks with the same thickness. We have provided a comparison of the freezing distance to replace the weight of the ice block to prove the capability for large-scale production of aerogels ( Supplementary Fig. 10).
When comparing time consumption between different methods, the set freezing area and final freezing environment are constant. When the ice block weights less than 800g, the thickness of the liquid film is 2 cm. In this case, the heat transfer resistance is relatively high. The freezing rate gradually decreases from bottom to top due to the gradual vanishing of temperature gradient and reduction of heat transfer. This thickness is much greater than the liquid film thickness on the surface of the rotating cryogenic drum (~3 mm). Although decoupling the ice nucleation and growth processes can enhance the freezing process, the crushed ice casting will not exhibit particularly Time consumption of crushed ice manufacturing (first step) and crushed ice re-casting steps (second step). c Freezing rate of crushed ice re-casting and unidirectional freezing.

Revise details:
 Page 9, Line 3, added, "Next, we compared the overall freezing efficiency of different freezing methods. When the freezing distance was low, the CIC method had no significant advantage over directional freezing because two separate steps were required. However, when producing ice block with a high thickness of 3 cm, the total freezing time was 1.38 times shorter than that of the conventional unidirectional freezing process, and 1.92 times shorter than that of the nondirectional freezing process, (Supplementary Fig. 10). With an increase of freezing distance, the freezing efficiency would be more significant.
This can be explained as the process intensification from decoupling the ice nucleation and growth processes, including: (i) Thin film freezing process on rotating cryogenic drum can effectively increase heat transfer efficiency; and (ii) Multi-point growth from 3D distribution of crystal nucleus can reducing crystallization path and accelerating the crystallization efficiency." The authors should provide a more convincing discussion.
Response: Thank you for your meaningful comments and suggestions. I am very sorry that we may not clearly describe the structure of the lamellar sponge (Nat Commun 11, 3732 (2020); ACS Nano 12, 3103-3111 (2018).), which is stacked layer by layer. The axial and radial direction is completely different from the anisotropic aerogel ( Supplementary Fig. 15). The thermal conductivity of the lamellar sponges consists of three parts: thermal convection, thermal conduction and thermal radiation. The low thermal conductivity of the axial direction is mainly provided by submicro-fibrous building block, lamellar structure, and ultra-light characteristics, owing to the layer-bylayer air blocking effect, the multilayer diffuse reflection effect, and the thermal bridge inhibition effect 1 . However, in the radial direction, gas conduction and convection mainly contribute to the thermal conductivity in lamellar sponges owing to the large gaps between lamellas, while the nanofiber walls serve as channels for solid conduction.
We agree with the reviewer's viewpoint of anisotropic aerogels have better insulation performance than their isotropic counterparts, which is not contradictory to our results.
The thermal conductivity of the anisotropic aerogels in the radial direction is relativity lower than that of the axial direction, which is may also smaller than the porous structure formed by random freezing. When the thermal energy is propagated in the radial direction, lamellar aerogels prepared by unidirectional freezing could effectively block solid heat conduction since the interlamellar connection is missing. Moreover, the thermal convection is also restricted in the interlaminate, while the thermal radiation could be weakened by multiple refraction and emission between the regular lamellas, thus significantly reducing the total thermal conductivity 2 . In the axial direction, heat would transfer through the gap between lamellas, resulting in higher thermal conductivity as compared with that in radial direction. As stated by the referee, this It must be pointed out that, the manifestation of advantages must be reflected in the basically identical pore structure, including size and topology. In the literature description (Chem. Eng. J., 2020, 385), their random freezing is not crushed ice casting, ice crystals still grow along temperature gradients (from the outside to the inside), forming a seemingly isotropic structure. Columnar crystals dominate this structure, making it difficult to form a more refined separate cellular structure.
In terms of CIC process, ice crystals are artificially added to increase discrete nuclei.
All pre-existing and 3D random distributed ice crystals can act as primitive growth sites.
The adjacent crystals will unconstrainedly grow with no favorable location or orientations and ultimately form multi-domain bulk materials, along with the refinement of individual ice crystals. After the sublimation of ice crystals, the structure distributed along the ice crystal interface assembled by nanofibers with high tortuous is preserved. An isotropic aerogel with separate cellular networks can be readily obtained. The refined tortuous and complex pore structure would become effective barrier for thermal transport. Thus, the thermal conductivity of our structure is lower than that of anisotropic aerogel.
For anisotropic aerogels, heat will flow through parallel channels in the axial direction and be transferred to the outside of the material for heat dissipation. However, due to insufficient lateral dimensions, anisotropic materials are difficult to utilize in the radial direction for practical insulation applications. While for ASNFs with unrestricted available size, this separate cellular pore structure can effectively block heat in the cell cavity and prevent it from conducting. As the reviewer noted, we have provided more convincing discussion about the thermal conductivity of lamellar sponges in different direction.
Supplementary Fig. 15. Structure comparison of different porous materials. a SEM images of the ASNF lamellar sponges from the different view. b 3D reconstruction of the ASNF anisotropic aerogels from X-ray microtomography (upper).
SEM images from the top view at different magnifications (bottom). c 3D structure of the ASNF isotropic aerogels from X-ray microtomography.

Revise details:
 Page 12, Line 10, added "The low thermal conductivity of the axial direction was mainly provided by the the thermal bridge inhibition effect, layer-by-layer air blocking effect, and the multilayer diffuse reflection effect 1 . However, in the radial direction, gas conduction and convection mainly contributed to the thermal conductivity owing to the large gaps between lamellas, while the nanofiber walls serve as channels for solid conduction."  Response: Thank you for the comments. We are sorry for the careless mistakes. We have revised them.

Responses to Referee #2
Comment 1: I went through the revised version of the manuscript. Although the authors reply to the points raised, I found many of these arguments still unconvincing and/or based on unfair comparisons. Regarding the freezing process itself, for example, which is supposedly the central novelty of the paper, the key argument proposed by the author is that the overall freezing time is faster, and this should enable commercialization of these materials. I think the authors do not have sufficient understanding of the costs and hurdles of commercializing novel materials. Freezing is (energy-wise) relatively cheap, the expensive part is the drying, which is the same in both cases (previous work and current work), and the cost of the process is just a small part of the equation to decide whether a novel material can be commercialized successfully or not. Besides, the authors already mentioned in the introduction that freeze-casting is low cost. The main benefit of the novel approach (claimed by the authors) is that pieces of large dimensions can be processed, but the same dimensions can be obtained with standard freeze-casting.

Response:
We are thankful for these comments. We are sorry for our responses did not meet the point. We agree with the referee's viewpoint on the energy consumption of freezing is a small part of the freeze-drying process (~5%). Solving the freezing problem cannot solve the industrialization problem of our materials. However, it is worth mentioning that our work does own several innovative points more than the freezing time. We would like to emphasize our innovation to present this work's novelty and research focus.
(1) Decoupling the ice nucleation and growth processes in freezing process. We successfully combine two mechanisms to enhanced the freezing process through decoupling the ice nucleation and growth processes. To increase the proportion of equiaxed grains, mechanical vibration or ultrasound inducting were applied to form more discrete nuclei. Similar in the crushed ice casting process, ice crystals are artificially added to increase discrete nuclei, thereby promoting the formation of similar equiaxed crystals. In our process, the adjacent crystals will unconstrainedly grow with no favorable location or orientations and ultimately form multi-domain bulk materials, along with the refinement of individual ice crystals. After the sublimation of ice crystals, the structure distributed along the ice crystal interface assembled by nanofibers with high tortuous is preserved. An isotropic aerogel with separate cellular networks can be readily obtained. The refined tortuous and complex pore structure would become effective barrier for thermal, electronic, or acoustic transport. According to your comments, we have made the following revise:  Page 7, Line 30, we added, "However in CIC process, the setting temperature and slurry temperature were essentially consistent and no nucleation peak appeared, demonstrating the crystallization process does not require excessive external energy input to overcome the crystallization energy barrier (Fig. 2b). Therefore, we propose a coupling mechanism between nucleation and crystal growth based on the above theory. The added ice particles can serve as a source of nucleation for the subsequent freezing process, thereby saving nucleation time. At the same time, the introduction of ice particles increases the nucleation sites during the freezing process, thereby regulating the growth orientation of ice crystals and forming similar randowly distributed equiaxed ice crystals. Finally, a randomly distributed pore structure was obtained through freeze-drying. Based on the suggestions of the reviewers, we have rewritten the relevant paragraphs and added appropriate descriptions of nucleation. We have also attempted to clarify the theoretical basis for our proposed mechanism more clearly.

Revise details:
 In Page 7, Line 24, we added, "To probe the mechanism differences between unidirectional freezing and CIC, we in-situ observed the cooling process under an optical-fluorescence microscope after mixing with a small amount of fluorescent polystyrene microspheres. Two comparative tests have been conducted at a same cooling rate of 1°C min −1 during freezing. In unidirectional freezing process, the slurry temperature decreased linearly before ice nucleation. A sudden temperature rise could be observed when the slurry temperature reached 1.3 °C, which was attributed to the heat release upon nucleation. However in CIC process, the setting temperature and slurry temperature were essentially consistent and no nucleation peak appeared, demonstrating the crystallization process does not require excessive external energy input to overcome the crystallization energy barrier (Fig. 2b)."  In Page 9, Line 3, we added, "Next, we compared the overall freezing efficiency of different freezing methods. When the freezing distance was low, the CIC method had no significant advantage over directional freezing because two separate steps were required. However, when producing ice block with a high thickness of 3 cm, the total freezing time was 1.38 times shorter than that of the conventional unidirectional freezing process ( Supplementary Fig. 10). With an increase of freezing distance, the freezing efficiency will be more significant.
This can be explained as the process intensification from decoupling the ice nucleation and growth processes, including: (i) Thin film freezing process on rotating cryogenic drum can effectively increase heat transfer efficiency; and (ii) Multi-point growth from 3D distribution of crystal nucleus can reducing crystallization path and accelerating the crystallization efficiency. Moreover, the whole process time from raw materials to products was also much lower than the traditional fiber felt/silica aerogel composites (Supplementary Table   2), demonstrating the practicality of this method."  In Page 9, Line 16, we added, "In terms of resultant materials, the columnar and equiaxed crystals were main components in ice blocks from unidirectional freezing and CIC, respectively, derived from the difference in the growth orientation. In the manufacturing process of industrial ingots, columnar crystals seriously affect the strength and toughness of materias. To increase the proportion of equiaxed grains, mechanical vibration or ultrasound inducting were applied to form more discrete nuclei. In the CIC process, ice crystals were artificially added to increase discrete nuclei, thereby promoting the formation of similar equiaxed crystals. We investigated the influence of ice-slurry ratio on the crystal structure of obtained ice blocks. When the proportion of ice crystals was small, most ice crystals would form oriented columnar crystals along the direction of temperature gradient. Only a small portion of the area was affected by external ice crystals to form partial equiaxed crystals (Fig. 2d).
When the proportion of ice crystals reached a reasonable level (2 : 1), all preexisting and 3D randomly distributed ice crystals could act as primitive growth sites. The adjacent crystals would unconstrainedly grow with no favorable location or orientations and ultimately form multi-domain bulk materials, along with the refinement of individual ice crystals. After the sublimation of ice crystals, the structure distributed along the ice crystal interface assembled by nanofibers was preserved. An isotropic aerogel with separate cellular networks could be readily obtained. The refined tortuous and complex pore structure would become effective barrier for thermal, electronic, or acoustic transport."

Comment 3:
The key idea upon which the paper is based (preparing ice crystals and then mixing them with the solution) has already been reported in many papers: see for The detailed demonstration summary of its novelty and characteristics is as follows.
(1) In terms of production methods, these materials prepare initial ice templates by freezing pre-formed water droplets, and accurately control the shape and position of the initial ice crystals to regulate the pore structure of the final material. While the CIC method for producing ice crystals proposed in our work is more convenient, which can quickly and efficiently prepare crushed ice. Thereby a large amount of crushed ice can obtain in a short period of time.
(2) There are also differences in the composition of ice crystals. During the preparation process, the dispersed solution is directly prepared into crushed ice instead of deionized water, so the density of the final block will not change significantly with the proportion of ice added.
(3) There are differences in the mixing process between ice crystals and water. Due to the composition of ice crystals, there is no need to pre-freeze the mixed solution or add other reagents to avoid the melting of ice crystals during the mixing process.
We significantly simplified the overall preparation process by adjusting the ratio of ice to water.
(4) The application scopes are varied. The work reported mostly is to prepare specific materials, which has precise requirements for structure. As for cell culture, it is required to have a uniform and regular distribution of pores to ensure the uniform distribution of subsequent cells. The method we proposed is a universal process for preparing isotropic porous materials, which can be applied to various fiber or powder materials that can be prepared into slurries.
As the reviewer suggested, we have cited relevant literature. And more detailed description was provided on the differences from previous work and the novelty were added in the revised manuscript.

Revise details:
 In Page 3, Line 17, we added "Alternatively, dynamic freeze casting and ice particulate templating can achieve multi-point ice nucleation and growth [3][4][5] . The obtained uniform porous materials have been proven to be applicable to tissue scaffolds. These methods partly overcoming the issues of low nucleation and growth efficiency in freeze casting process. However, such behavior, combined with process intensification and structural evolution, has not been verified nor extended to different material systems. Most critically of all, the fundamental question of how ice crystal transformation can allow for such a large alteration in structure remains unanswered."

Comment 4:
Overall, although the claimed novelty is on the process (bearing in mind my concerns above), it still represents only 20% of the paper or so, the rest is dedicated to different materials obtained by the process, I thus still find the paper poorly balanced.
Response: Thank you for your constructive comments and suggestions. We have deleted part of discussion on the characterization of material properties and placed the figures in supporting information, while adding a detailed description of the preparation process. More than 60% of the manuscript now revolves around preparation methods rather than material properties. As mentioned in Comment 1, we have conducted in-depth discussions on the novelty of the method.

Revise details:
 We deleted the part of "Temperature-invariant mechanical performances".
 In Page 7, Line 24, we added, "To probe the mechanism differences between unidirectional freezing and CIC, we in-situ observed the cooling process under an optical-fluorescence microscope after mixing with a small amount of fluorescent polystyrene microspheres. Two comparative tests have been conducted at a same cooling rate of 1°C min −1 during freezing. In the unidirectional freezing process, the slurry temperature decreased linearly before ice nucleation. A sudden temperature rise could be observed when the slurry temperature reached 1.3 °C, which was attributed to the heat release upon nucleation. However, in CIC process, the setting temperature and slurry temperature were essentially consistent and no nucleation peak appeared , demonstrating the crystallization process does not require excessive external energy input to overcome the crystallization energy barrier (Fig. 2b)."  In Page 9, Line 3, we added, "Next, we compared the overall freezing efficiency of different freezing methods. When the freezing distance was low, the CIC method had no significant advantage over directional freezing because two separate steps were required. However, when producing ice block with a high thickness of 3 cm, the total freezing time was 1.38 times shorter than that of the conventional unidirectional freezing process ( Supplementary Fig. 10). With an increase of freezing distance, the freezing efficiency would be more significant.
This can be explained as the process intensification from decoupling the ice nucleation and growth processes, including: (i) Thin film freezing process on rotating cryogenic drum can effectively increase heat transfer efficiency; and (ii) Multi-point growth from 3D distribution of crystal nucleus can reducing crystallization path and accelerating the crystallization efficiency. Moreover, the whole process time from raw materials to products was also much lower than the traditional fiber felt/silica aerogel composites (Supplementary Table   2), demonstrating the practicality of this method."  In Page 9, Line 16, we added, "In terms of resultant materials, the columnar and equiaxed crystals were main components in ice blocks from unidirectional freezing and CIC, respectively, derived from the difference in the growth orientation. In the manufacturing process of industrial ingots, columnar crystals seriously affect the strength and toughness of materias. To increase the proportion of equiaxed grains, mechanical vibration or ultrasound inducting were applied to form more discrete nuclei. Similar in the CIC process, ice crystals were artificially added to increase discrete nuclei, thereby promoting the formation of similar equiaxed crystals. We investigated the influence of iceslurry ratio on the crystal structure of obtained ice blocks. When the proportion of ice crystals was small, most ice crystals would form oriented columnar crystals along the direction of temperature gradient. Only a small portion of the area was affected by external ice crystals to form partial equiaxed crystals (Fig.   2d). When the proportion of ice crystals reached a reasonable level (2 : 1), all pre-existing and 3D randomly distributed ice crystals could act as primitive growth sites. The adjacent crystals would unconstrainedly grow with no favorable location or orientations and ultimately form multi-domain bulk materials, along with the refinement of individual ice crystals. After the sublimation of ice crystals, the structure distributed along the ice crystal interface assembled by nanofibers with high tortuous was preserved. An isotropic aerogel with separate cellular networks could be readily obtained. The refined tortuous and complex pore structure would become effective barrier for thermal, electronic, or acoustic transport." Comment 5: non-scientific expressions are still present throughout the manuscript (e.g. « salad-like », which does not mean anything, « ultralight feathers » ?) Response: Thank you for your valuable comments and suggestions. We have doublechecked the manuscript and deleted all the non-scientific expressions (such as "saladlike", "ultralight feathers") in the revised manuscript to ensure that all the clams revised on properties and performance are scientific and objective.

Comment 6:
My concerns about reproducibility have not really been addressed, I still don't see error bars in Fig S7a, for example.

Response:
Thank you for your constructive comments and suggestions. We are sorry for our responses did not addressee your concerns. We have added a detailed description of the preparation process and characterization of the sample to further validate repeatability.
First, the ratio of crushed ice to dispersion was discussed, and the rationality of Time consumption of crushed ice manufacturing (first step) and crushed ice re-casting steps (second step). c Freezing rate of crushed ice re-casting and uniderectional freezing.
Comment 7: non-standard tests (Fig 3a) are still present. These are uninformative since we cannot compare the behavior with that of previously reported materials.
Response: Thank you for your valuable comments and suggestions. The purpose of the test is to verify the impact elasticity of the aerogel materials. The recovery speed of ASNF aerogels was measured by rebounding a falling steel ball (2.26 g), which was calculated as 764 mm/s −1 , demonstrating the rapid rebound ability of the aerogels. This method has been used in many carbon or organic materials (e.g., carbon nanofibrous aerogels (860 mm s -1 ) 6 , chitosan-graphene oxide monoliths (580 mm s -1 ) 7 , ceramic nanofibrous aerogels (860 mm s -1 ) 8 , mullite sponges (1,233 mm s -1 ) 9 ). As the reviewer noted, it is not a standard test.
In order to test the impact resistance of materials, we used drop hammer impact testing and pendulum impact testing. However, due to the characteristics of nanofiber materials, the prepared aerogel shows flexibility, which is difficult to obtain data from the test. We also attempted Hopkinson rod testing, but the mechanical strength of this porous material is not sufficient to obtain experimental data. To make the paper more scientific, we have removed this test. Revise details:  We removed Fig 3a and Corresponding videos. Fig 4g and 4i proves that the thermal conductivity properties are essentially the same than previously reported, which is not surprising since macropores dominate the thermal behaviors in both cases. There are thus no improvements here.

Comment 8:
Response: Thank you for your important comments and suggestions. At present, there are many kinds of thermal insulation materials, and aerogel is a common material.
Because air is an ideal thermal insulator, the low thermal conductivities of aerogels mainly originate from the restricted heat transfer across the gas phase confined in the voids. The ASNFs aerogels prepared by crushed ice casting achieves low thermal conductivity through the intrinsic low thermal conductivity of ceramic materials and the structure with high porosity. And the high tortuosity channel structure is realized by columnar-equiaxed crystal transition, which further reduces the convective heat transfer of the gas phase.
(1) Using other different testing methods, some previous work reported very low thermal conductivity. For example, an infrared camera was used to record the heat-conducting process of the materials after applying laser heat source 10 . Then the thermal conductivity was calculated. In addition, testing in an argon or vacuum environment can result in lower thermal conductivity, as the heat transfer performance of argon or vacuum is lower than that of air. In this paper, we use the standard test method in the air atmosphere. Under the standard test condition, the air thermal conductivity at room temperature is 25.52 mW/m K. The ASNF aerogels has reached an ideal level in thermal conductivity.
(2) While maintaining low thermal conductivity, temperature resistance and mechanical properties are the areas worth breaking through in thermal insulation materials. We compared the lowest thermal conductivity and highest temperature resistance of our ASNF aerogels with those of other typical thermal insulation materials in an oxidizing atmosphere, such as nanocellulose/graphene oxide 11 , nanowood 12 , SiO2 aerogels 13 , carbon nanotube aerogels 14 , graphene/Al2O3 15 , hBN aerogels 16 , SiC@SiO2 nanowire aerogels 10 , SiO2 nanofiber aerogels 8 , and ceramic microfiber sponges 17 ( Fig. 3i and Supplementary Table 5). The ASNF aerogels showed a lower thermal conductivity than most of the reported thermally insulating materials, well below that of standing air (0.025 W m −1 K −1 ).
Furthermore, the long-term temperature resistance of ASNF aerogels was 1200 °C, which was the highest among the thermally superinsulating materials.
(3) Most elastic structures with low thermal conductivity made from polymers or carbonaceous materials cannot withstand high temperatures under ambient conditions. As ceramic families, ASNF aerogels are expected to possess temperature-invariant elasticity. Different viscoelastic properties, such as storage modulus, loss modulus, and damping ratio, were investigated over a broad temperature range of −100 to 500 °C at a constant frequency of 1 Hz. Indeed, the aerogels showed temperature-independent stable viscoelastic performances and a consistent small damping ratio of ∼0.1 (Supplementary Fig. 16a-c). A frequency dependency test (0.1 to 100 Hz) also showed stable viscoelastic properties over a wide temperatures range of −100 to 500 °C. Furthermore, compression tests were conducted by exposing the materials to a butane flame (1300 °C) and submerging it in liquid nitrogen (−196 °C) (Supplementary Fig. 16d and Video 8). The ASNF aerogels retained their resilience up to 80% compressive strain under extreme conditions. After several cycles, the aerogels fully recovered, with no obvious fracture. cathode has been considered the most complicated hazards among other battery TR events. Taking advantages of its stable mechanical and thermal insulation performance at ultrahigh temperatures, a thin layer of the aerogels can successfully prevent the deflagration propagation of a pack of high-energy Li-ion batteries, while one of the batteries caught on fire/explosion. As far as we know, no materials have been able to achieve this effect. Our materials provide new technique to solve the energy storage safety problems, which is the primary focus of the electric vehicle (EV) community.