Structurally Integrated Properties of Random, Uni-Directional, and Bi-Directional Freeze-Dried Cellulose/Chitosan Aerogels


 Aerogels as a thermally insulating material have attracted extensive consideration in recent years due to prevailing energy consumption in building, industrial sector, and aerospace applications. Cellulose contains all the favorable properties to be used as a thermally resistant substance but due to structural instability, it needs to be impregnated with other resilient substances. We report herein the preparation of cellulose nanofiber (CNF)/chitosan (CS) isotropic and anisotropic composite aerogels (CCSA) by employing three different: random, uni-directional, and bi-directional freezing techniques, ensuing freeze-drying process to investigate the structural modification effect on the thermal and mechanical properties. Aerogels behaved distinctively along different axis due to holding entirely eccentric porous micro-orientation along lateral (perpendicular to ice growth) and axial (parallel to ice growth) directions. Randomly frozen aerogels (r-CCSA) contained uneven porous networks because of random ice crystal formation within the microstructure and resulted in isotropic characteristics with thermal conductivity (λ) of 0.038Wm-1K-1 in both axial and radial direction. whereas unidirectional frozen aerogels (u-CCSA) exhibited anisotropic microstructure and performance with lamellas in axial and honeycombed in radial direction resulting in λ value of 0.040 Wm-1K-1 and 0.034 Wm-1K-1 respectively. Similarly, the controlled temperature gradient during the preparation of bidirectional aerogels (b-CCSA) presented an anisotropic sheet-like microstructure and resulted in ultra-low thermal conductivity along the axial and radial geometry with λ of 0.33 Wm-1K-1 and 0.027 Wm-1K-1 respectively because of the Knudsen effect. CCSA reported in this work resulted in ultra-low density (5 to 16 mgcm-3), high porosity (~99.6%), and robustness mechanically with an endurance of 60% strain. Such a composite bio-mass aerogel preparation method will offer a clear insight into the adoption of the right methodology according to target practical applications with consideration of environmental safety.


Introduction
Thermal insulation is an important factor for energy-e cient buildings, space applications, and greenhouse gas emissions ). Since thermally insulating materials need to be porous to minimize heat ow and environmentally friendly to address biocompatibility, plenty of commercial insulating materials have been fabricated e.g., expanded polystyrene with λ of ~ 0.035 Wm − 1 K − 1 , rock wool, and glass bers with thermal conductivity (λ) of 0.15 Wm − 1 K − 1 to 0.53 Wm − 1 K − 1 . However, it requires high thickness to meet the target insulation value. Vacuum insulation panels with λ of 0.030 Wm − 1 K − 1 to 0.040 Wm − 1 K − 1 , but can easily damage while handling (Alam et al. 2011). Conversely, aerogels are the materials obtained from the gel after supercritical drying to preserve gel morphology, composed of 95-99% air with average pore size < 20 nm, high speci c surface area, and lower density < 0.1 g/cm 3  Polysaccharides such as cellulose nano ber-based aerogels, due to their unique characteristics, such as high speci c surface area, low density, great mechanical strength, high porosity, low thermal conductivity, presence of a large number of functional groups, biodegradability, and natural renewability (Dufresne 2017; Lavoine and Bergström 2017; Ahankari et al. 2021a) has engrossed increasing interest as a potential and sustainable replacement for thermal insulation (Ahankari et al. 2021b; Han et al. 2021), environmental (Lavoine and Bergström 2017), energy , and biomedical applications (Zhang et al. 2020a). Thermal insulation is an attractive property to target nowadays for environmental interest, as studies have been conducted to reduce the thermal conductivity of CNF aerogels by making their composites with other substantial materials (Zhang et al. 2020a; Liu et al. 2021; Zuo and Yuan 2021). Such an approach helps in the reduction of heat radiation caused by carbonaceous materials. Another naturally most abundant bio-based polymer, chitosan (CS), holding many amino and hydroxyl groups, is a desirable material for plenty of application areas. However, low mechanical strength and functionalization can be reinforced with other high mechanical and thermally insulating materials Aerogels behave differently on varying structural morphology (Hu et al. 2020), and this structural orientation can diverge by modifying pore structure within aerogels' geometry after employing various freeze casting techniques and by changing freezing velocity and temperature. He et al. (2020a) fabricated boehmite hydrosols and alumina nanopowder aerogel by adopting different freezing temperatures and observed the relationship between ice front rate (ν) and structural wavelength (λ) variations. Random freeze aerogels exhibit randomly distributed porous microstructure, non-linear sensitivity, and unsatisfactory mechanical performance. Whereas, a controlled freeze-casting environment produces wellordered anisotropic aerogels, which behave speci cally along a different axis due to anisotropic behavior.
These aerogels have broad environmental applications according to obligatory end-use. This study adopted three different random, uni-directional, and bi-directional freezing strategies to fabricate CNF/CS-based isotropic and anisotropic composite aerogels to observe the structurally integrated characteristics along a different axis. Formaldehyde was used as a crosslinker for chemical bonds formation between CNF and CS in the hydrogel phase, and sublimation was done following the freeze-drying technique. As CNF acts as a structural reinforcing nano ller for the biopolymer chitosan aerogel, we investigated the effect of structural modi cation due to the adoption of each methodology on the nal aerogels' characteristics. We evaluated micro-orientation, physical properties, thermal behavior, compressibility, elasticity to demonstrate the structural modi cation effect on resultant properties. The lowest thermal conductivity value was obtained for anisotropic aerogels employing the bi-directional freezing technique was 0.027 Wm − 1 K − 1 in the radial direction, which is quite less than the isotropic thermal insulating aerogels. Therefore, the relative adoption of the best suitable preparation route according to practical applications would be easy to understand.

Preparation of hydrogel
The aqueous solution of CNF with a weight percent of 0.3% (C 3 ) and 0.7 wt% (C 7 ) was prepared by adding deionized water into TEMPO-CNF suspension, which dispersed and homogenized at 14,000 rpm with the aid of a high-speed homogenizer (Ultra-Turrax, IKA) for 30-35 min. Meanwhile, CS solutions have been prepared by mixing CS by 0.3 wt% (CS 3 ), 0.7 wt% (CS 7 ), 1 wt% (CS 10 ) wt%, and 1.2 (CS 12 ) wt% ratios with 0.1 M glacial acetic acid following continuous magnetic stirring in 80°C heated oil bath for 40 min to obtain a transparent and uniform suspension. CNF dispersions (C 3 & C 7 ) were further mixed individually with the prepared chitosan ratios (CS 3 , CS 7 , CS 10, and CS 12 ) while vigorous stirring for an additional 30 min to get a proper dispersed viscous suspension. The resultant composite solution named as C 3 CS 3 , C 3 CS 7 , C 3 CS 10 , C 3 CS 12 , C 7 CS 3 , C 7 CS 7 , C 7 CS 10 and C 7 CS 12 respectively ( Figure S1). Once thoroughly mixed, the pH was raised to 7 by adding a subsequent amount of NaOH. 0.2% (v/v) formaldehyde was added to the nal hydrogel and let the solution on magnetic attiring for 5h for crosslinking process and nal hydrogel formation.

Fabrication of aerogels following distinct freezing techniques
The CNF/CS hydrogel poured into a plastic mold and directly placed into liquid nitrogen (-196 ℃) bath for freezing, following lyophilizing process for 2 consecutive days at -50 ℃ using freeze dryer under 1Pa vacuum pressure, as prepared aerogels named as randomly freeze-dried CNF/CS aerogels (r-CCSA).
For directional freezing technique, the hydrogel suspension was transferred into PTFE (polytetra uoroethylene) mold while placing on a copper square bar in the liquid nitrogen-lled bath for 20 min and lyophilized for 60 hours, as prepared aerogels named unidirectional-CNF/CS aerogels (u-CCSA).
Surface bioinspired structure aerogels prepared by transferring complex suspension into a customized PDMS (polydimethylsiloxane) wedge square mold with a bottom slope angle of about 20° and immersed into liquid nitrogen for about 45 min ensuring absolute freezing. The frozen hydrogel was then lyophilized for 72 hours; aerogels were further kept in a vacuum oven at 100 ℃ for 60 min for well cross-linking.

Characterization
The microscopic architecture of CCSA was analyzed by SEM (Hitachi S-4800), the compression experiments were conducted using Instron 5944 universal machine with 500-N load cells using a strain rate of 2mm min − 1 with 60% strain. The weight of CCSA was obtained using an ultra-micro balance (MS105DU, Mettler Toledo). The chemical structure was observed conducting FTIR testing with spectra range 4000 to 400 cm − 1 in transmission mode on the resolution of 4 cm − 1 . Thermal conductivity obtained using Hot Disk thermal analyzer (TPS2500S) following transient plane source (TPS) method. Infrared images of aerogels were captured using a thermal infrared camera (Testo 868); thermal stability of aerogels was studied by a thermogravimetric analyzer (PerkinElmer Inc., USA) with a 20 ml/min nitrogen ow rate with a temperature range of 30 to 600 ℃ and heating rate 10 ℃ /min. Shrinkage of aerogels was obtained from volume to size of the mold ratio. Porosity values calculated using the following expression: Where P represents porosity, ρ 0 apparent density, and ρ skeletal density, here we use the average density of CNF (1.46 g/cm 3 ) and CS (1.75 g/cm 3 ) to get the nal skeletal density.

Structurally integrated functionality
CNF was blended with CS to fabricate isotropic (random) and anisotropic (uni-directional, bidirectional) composites aerogels by pouring CNF/CS suspension into the customized molds following the sol-gel method and freeze-drying, as shown in Fig. 1a-c. The adoption of three different freezing techniques generates three different cross-linked CNF/CS composite aerogel's (CCSA) interfacial assembly, as shown in Fig. 1d. CCSAs were manufactured using a lower and higher concentrated CNF suspension, with gradually increasing CS amount to study the properties and composition effect accordingly. To study the structural orientations and their ultimate effects on nal properties, we adopted three criteria while preparing CCSA: (I) randomly oriented pore structure-based aerogels (r-CCSA), (II) Controlled porous assembly oriented uni-directional anisotropic aerogels (u-CCSA), and (III) bio-inspired 3D ordered morphology attaining bi-directional composite aerogels (b-CCSA).
In this regard, three different freeze-drying techniques were employed. For attaining the rst criteria, we used the conventional freeze-drying method in which CNF/CS hydrogel poured into a at rounded plastic mold following direct freezing into liquid nitrogen ow; in this process, the temperature gradient was not speci ed, so the nal aerogels' morphology consisted of random pores distribution along the both axial (longitudinal) and radial (cross-sectional) direction. The structure of r-CCSA is illustrated in Fig. 2a-c. As shown in Fig. 2a, the low concentrated CNF/CS made aerogel exposed a less stable microstructure with wide pore size and uneven geometric distribution. On the other hand, C 7 CS 7 A structure (Fig. 2b-c) showed a compact morphology with an obvious pore wall in the sampling plane (Fig. 2 inset) with an average pore size of 12 to 20 µm in both axial and radial direction. The inset of Fig. 2c showing the evident presence of CNF on pore walls.
To meet the second criterion-the CNF/CS hydrogel was poured into PTFE mold and frozen by placing on a copper bar in liquid nitrogen, following the freeze-drying process. Because of a controlled temperature gradient, the ice crystals formation occurs, and a sublimation process left anisotropic morphology without affecting the parent structure. The longitudinal section of u-CCSA ( Fig. 2e and inset) exhibited parallel lamellar cell geometry with lamellas spacing around 20-30 µm. This geometry of aerogels formed due to water crystallization induction along the axis, which ultimately allowed nano bers and chitosan particles to form an organized sheet structure.
Chitosan is a substance that provided skeletal stability by forming bridge bands between lamellas ( Fig. 2f and inset ), which ultimately enhanced the mechanical strength of nal aerogels. However, the lower concentration of CS (C 7 CS 3 A) made aerogels displayed irregular geometric structure (Fig. 2d) with a lamellar distance around 35 to 55 µm. A nanoscale observation of u-CCSA exhibited micro bers of CNF on the walls of lamellas ( Figure S3a). It can also be noted that the lamellar structure of u-CCSA possesses a unique macro-anisotropic orientation within the cellular geometry which ultimately results in the anisotropic thermal conductivity and mechanical behavior of nal aerogels ( Figure S2).
Lastly, the bi-directional aerogels were prepared by pouring the sol-gel into a customized mold followed by freeze-drying and cross-linking. The structure made from this technique is presented in Fig. 2g-i; the parallel layered orientation of the structure is quite visible along the longitudinal direction and honeycombed along the cross-sectional. This speci c structure was induced because of a de ned temperature gradient along the length of aerogels (vertically by PDMS wedge) and with the PDMS wedge (horizontally); as a result, the ice crystallization behaved differently. Figure 2h and inset presents aligned lamellar structure of b-CCSA axially, containing a layer gap of about 20 to 30µm. While the crosssectional view also contains a clear presentation of individual layers and shows the anisotropic geometry of aerogels ( Fig. 2i and inset). A nanoscale observation of b-CCSA shows the evidence of cellulose nano bers on the wall of lamellas see Figure S3b. The lower concentration of cellulose and chitosan containing b-CCSA (C 7 CS 3 A) having non-solid geometry of lamellas with layers gap of about 35 to 40 µm (Fig. 2g).

Characterization of physical properties at room temperature
Composite CNF/CS aerogels has been prepared by mixing CNF with a concentration of 0.3% and 0.7% (C 3 and C 7 ) individually with CS solution of 0.3%, 0.7%, 1%, and 1.2% for C 3 CS 3 A, C 3 CS 7 A, C 3 CS 10 A, C 3 CS 12 A, C 7 CS 3 A, C 7 CS 7 A, C 7 CS 10 A and C 7 CS 12 A aerogels formation respectively. Mechanical behavior within random aerogels is mentioned with "r-CCSA," and anisotropic aerogels (uni-directional and bidirectional) represented with "u,b-CCSA" in Fig. 3 for comparison purposes. A digital photograph of lightweight CCSA with a volume of ~12cm 3 is shown in Fig. 3a, which indicates its ultra-low density. Fig. 3b displays a linear increase in densities with a gradual increase of CS ratio. The increase of CS concentration from 0.3% to 0.7% in CNF suspension increased the aerogel density value up to 30%. In contrast to C 7 CS 7 , upon 1% CS addition in C 7 CS 10 , the density decreased by 7% because of the formation of porous structure stability within aerogels. However, an appropriate amount of CS (0.7% to 1%) is suitable for better density to mechanical strength proportion achievement. The porosity value in r-C 3 CS 7 A and u-b-C 3 CS 7 A decreased with the addition of 1% CS in C 7 CS 10 . While the CS addition gradually increased the porosity value in C 7 CS 3 A, C 7 CS 7 A, and C 7 CS 10 A which decreased when the CS ratio exceeded up to 1.2% (Fig. 3c).
The optimum CS content ful lls the structural defects and stabilizes the porous structure of aerogels. A CS ratio of 1.2% enhances the solid content within the con ned geometry, which ultimately results in pores reduction inside the aerogel bril walls. The volume shrinkage was reduced to 37-48% on the addition of CS to 0.3% (C 3 CS 3 and C 7 CS 3 )) in pure CNF aerogels and further reduced to 24% with 0.7-1% CS addition. But when this ratio reaches 1.2%, the volume shrinkage increased up to 30% (Fig. 3d) because an optimum CS addition strengthened the interaction between micro bers and kept the 3D solid network structure locked onto the position through which porous structure was not affected. Figure 3e shows the FTIR spectra of CNF, CS, and CCSA. The characteristic peaks at 3351cm − 1 , 2893cm − 1 correspond to −OH, C−H bonds respectively (Peng et al. 2016). The peak at around 3351cm − 1 is stretching vibrations of −OH and −NH. The peak in chitosan aerogels at 1644cm − 1 corresponds to the amide group. However, in composite aerogels, the new peak at 1470cm − 1 formed after cross-linking reaction, corresponding to the amide bond formation between formaldehyde and chitosan. In composite aerogels, the sharp peak at ~ 1690cm − 1 is evidence of cross-linked aerogels corresponding to aldehyde group formation (Yagoub et al. 2019). Formaldehyde addition to aerogels develops structural stability and enhances the mechanical strength of nal aerogels by providing strengthening units within the micro-orientation of aerogels.

Thermal stability of the CNF/CS aerogels
Thermal stability is essential for thermal insulation applications. The TGA (Fig. 4a) and DTG (Fig. 4b) results display the thermal degradation of CNF, CS, and composite CCSA. There were four weight loss stages of CCSA aerogels. The rst weight-loss phase begins around 100 ℃, which could be attributed to the absorbed water from aerogels. The second stage was in the range of 200-350 ℃, the organic component's decomposition. The third weight-loss of raw CCSA was observed between 400-450 ℃, and the nal weight-loss occurs because of the removal of structural water and carbonization between 500-600 ℃. The pure CNF aerogels decomposed at 279 ℃ with mass retention of 12.6%. However, with an appropriate amount of CNF and CS ratio within aerogel, the degradation temperature (Tg) increased from 282 ℃ (C 3 CS 7 A) to 300 ℃ (C 7 CS 7 A), and the weight loss of CCSA was also reduced from ~ 90% to ~ 70%. Compared to the pure CNF aerogels, the mass retention rate increased up to 54% for C 7 CS 3 A and up to 84% for C 7 CS 7 A, which shows the CS content helped in the thermal stability of the CCSA. It can be seen in Table 1 the chitosan aerogels having great Tg, so with the addition of CS to CNF aerogels, a physical barrier formed within the structure consisting of nanosheets that prevent composite aerogel from combustion. The anisotropy of thermal conductivity (λ) correlates with the morphology of the porous structure formed as a result of the adoption of three different freezing processes (as mentioned in Fig. 1 earlier). The λ has been observed along all three directional aerogels (Fig. 5g, h, and i), and the values were compared for pure aerogels and composite aerogels on different ratios ( Figure S4). Schematic illustration of CCSA microstructure is represented in Fig. 5a-f represents the microstructure of all three directional aerogels individually. As can be seen, r-CCSA shows similar morphology along the radial (Fig. 5a) and axial directions (Fig. 5b). This random porous structure allows maximum thermal heat conduction through the pores, which ultimately maximizes the λ. While the directional freeze aerogels (uni and bi-CCSA) hold an anisotropic skeleton, which lowers the heat localization in a radial direction and controls heat transfer in the lateral one ( Fig. 5c-f). As the layered structure of anisotropic aerogels greatly minimizes the thermal conduction,(Su et al. 2020) its has been observed that for pure CNF-axial/lateral, the thermal conductivity was 0.031 Wm − 1 K − 1 /0.024 Wm − 1 K − 1 and CS-axial/lateral were 0.035 Wm − 1 K − 1 /0.031 Wm − 1 K − 1 (see Fig S4a,b). Conversely, the composite aerogels, r-CCSA (axial/lateral), u-CCSA (axial/lateral), and b-CCSA (axial/lateral) the λ were 0.038Wm − 1 K − 1 /0.038Wm − 1 K − 1 , 0.040 Wm − 1 K − 1 /0.034 Wm − 1 K − 1 and 0.033 SEM images of u-CCSA reveal the cellular structure of aerogels with long pore channel alignment with cross-bridges in longitudinal direction allows the heat to spread along with the spaces by preventing heat localization upon the accumulation of thermal energy in the sample. Whereas honeycombed pore orientation in the cross-sectional area provides a wide surface area for heat to oat in, which reduces solid heat conduction and prevents quick heat ow through the aerogels. However, the b-CCSA samples freeze in a customized mold with a dual temperature gradient shows the lowest thermal conductivity values compared to the previous two techniques. The lamellar structure reduces heat convection, conduction, and radiation, eventually reducing the whole thermal conductivity value. Whereas, along the longitudinal direction of b-CCSA, the heat can oat through the channels easily due to the absence of connecting medium between lamellas, resulting in more thermal conduction.
The concentration of cellulose and chitosan also affect the thermal conductivity values. The thermal conductivity of aerogels has been shown to have a direct relationship with their densities (Zou and Budtova 2021). So that with the lower concentration of CNF (C 3 ), composite aerogels C 3 CS 3 A, C 3 CS 10 A and C 3 CS 12 A expressed λ value of 0.027 Wm − 1 K − 1 , 0.025 Wm − 1 K − 1 and 0.036 Wm − 1 K − 1 respectively (see Fig S4c). Thermal conductivity value rst decreased to 8% with the addition of CS 10, and it increased to 44% when CS 12 content was enhanced from a certain level due to exhibiting a higher density of C 3 CS 12 A. Likewise higher CNF (C 7 ) comprising CCSA aerogels C 7 CS 3 A, C 7 CS 10 A and C 7 CS 12 A attained λ of 0.035 Wm − 1 K − 1 , 0.031 Wm − 1 K − 1 and 0.038 Wm − 1 K − 1 respectively. The λ value rst decreased by 11% and then increased to 22%; this increase in thermal conductivity can be attributed to C 7 CS 12 A's higher density.
The λ was (0.027 Wm − 1 K − 1 to 0.036 Wm − 1 K − 1 ), which is less than the highly concentrated composite aerogels (0.032 Wm − 1 K − 1 to 0.040 Wm − 1 K − 1 ) because of the density variation. Whereas with increasing chitosan ratio up to an optimum range (C 3 CS 10 A and C 7 CS 10 A), λ decreased from ~ 7% to ~ 13%, whereas with higher concentration (C 3 CS 12 A and C 7 CS 12 A), the λ increased drastically up to ~ 7% to ~ 18%. This variation of thermal conductivity occurred because CS minimized the structural defects up to a certain level, while the addition of more solid contents into the assembly affected the pore structure, which caused an increase in thermal conduction ).

Thermal Insulation Mechanism
Thermographic images of nano-composites aerogels have been captured using an infrared camera (Fig. 6a) by keeping the CNF, CS, and CCSA samples of 10 mm thickness on a hot stage while maintaining the temperature at 200 ℃. Infrared images from the side of samples were taken with each 5 min time intervals for 60 min, and the temperature difference (Δt) between the hot stage and aerogels' surface has been summarized in Fig. 6b. It was revealed that the temperature ow along the random aerogels was quite fast, with a temperature increase of 31% after 1-hour exposure to the heat. The corresponding temperature distribution along pure CNF and CS aerogels can be seen in Figure S5.
Conversely, along anisotropic aerogels, the heat ow in a radial direction is comparatively slower than the lateral ones, with a temperature increase of 7% and 28%, respectively. Moreover, by reducing the anisotropic CCSA's thickness from 20 mm to 5 mm, the surface temperature remains 64 ℃ with a stage temperature of 200 ℃, expressing the good heat resistance property of aerogels even on smaller thicknesses (Fig. 6c).
The heat transfer mechanism is described in Fig. 6d-i. The heat transfer in porous materials occurs in various modes, such as solid conduction, gas conduction, convection, and radiation. The heat transfer by convection is negligible because of the smaller pore size of aerogels (< 1 mm) (Apostolopoulou-Kalkavoura et al. 2020). Conversely, the radiation is negligible at ambient temperature and pressure for the lower density and porous materials. However, the gas conduction (due to collision of gas molecules) contributes to heat transfer along the aerogels at ambient atmosphere, and it depends on the mean free path and pore size of material, but gas conduction is relatively lower when the pore size is smaller than mean free path of air molecules as per Knudsen effect. In anisotropic aerogels (Fig. 6f&h), the CNF/CS particles are well-aligned within channels along the axial direction. The lateral heat conduction (brown dotted lines) within the aerogel assembly is much faster than the radial direction. Moreover, higher temperature and oxygen content make heat convection (green dotted lines) and heat radiation (red dotted line) quicker to ow along the axial path. Because of the wide opening of channels, most CNF and CS particles were exposed to high temperatures, which induced quicker heat ow in the axial direction than the lateral one. Figure 6g&i illustrates the heat ow of u & b-CCSA in an axial and lateral direction. The lamellas are at a distance in u-CCSA but connected with bridge bands which ultimately let the heat conduction transfer through this assembly. However, the micro-channels are not interconnected in b-CCSA radially and do not provide enough room to let heat through accessible space with any possible mean, which reduces the temperature ow eventually.

Mechanical property analysis
To evaluate the mechanical properties of CCSA, a compression test was conducted using the universal testing machine. The resulting stress-strain curve of CNF, CS, and their composite aerogels along the radial and axial direction is illustrated in Fig. 7. The stress-strain curve of aerogels tested along axial direction exhibits a typical compression behavior consisting of three distinctive regimes: (i) the linear elastic stage at low strain, (ii) plateau beyond the yield point, and (iii) nal densi cation region at compressive strain over 60% (Song et al. 2018;Wang et al. 2019). In random freeze aerogels, stress-strain behavior is similar along with both directions and exhibits low compressive strength because of disordered porous orientation. The compression along a radial direction of both uni and bi-directional aerogels ( Fig. 7c and e) do not exhibit an apparent linear regime before the plateau. However, a unique anisotropic mechanical behavior can be observed for bidirectional aerogels when compressed along the axial direction (Fig. 7d), because of its highly ordered microstructure. As presented in Fig. 7a, c, and e, the aerogels compressed in the direction parallel to lamellas express good elasticity. Along the direction parallel to lamellas, aerogels are di cult to bend and exhibit good compressive strength (see Figure S6a, b, and e).
As shown in Fig. 7f with increasing CNF and CS ratios, Young's modulus (E) and yield strength (Fig. 7g) increased signi cantly along an axial direction. Young's modulus increased to 175% on increasing CS content from 0.3 to 0.7 within the lower concentration of CNF suspension and up to 56% along axial direction when CS content increased up to an optimum range in higher concentrated CNF suspension (C 7 CS 7 A). While yield strength reached 100% for C 3 CS 7 A aerogels and 20% for C 7 CS 7 A relatively.
Furthermore, a cyclic compression test on 60% strain was conducted for anisotropic aerogels in the radial direction for 5 consecutive cycles (Fig. 7h); these aerogels could withstand the compressive strain of 60% and recovered to their original shape by 75% after the release of cyclic stress, Fig. 7i shows the recovery rate after each cycle completion, this result expresses the good elasticity and structural robustness of anisotropic CNF/CS composite aerogels.
For better understanding, the CNF/CS composite aerogels were prepared to adopt three different freezing techniques, which affected resultant morphology and properties. The CCSA aerogels fabricated by employing random freeze methodology consisted of compact brous assembly with irregular pore size and isotropic properties along with both axial and radial directions. Meanwhile, uni-directional CCSA grasped lamellar geometry with bridge support between, showed anisotropic micro-orientation and properties along an axial and radial direction. Whereas CCSAs prepared by adopting bi-directional freezing technique shown the detached layered assembly of CCSA after freeze-drying and cross-linking. However, b-CCSA presented the best physical, thermal, and compressive strength properties along the radial direction compared to other polymeric aerogels (Cai et

Conclusion
In summary, three different structurally oriented CNF/CS composite aerogels are fabricated at different ratios, using formaldehyde crosslinker by employing three random, uni-directional, and bidirectional freeze casting strategies. Chitosan served as a structural reinforcing component and reduced the volume shrinkage rate of aerogels from 40-16% while curtailing structural defects and improving the orientation of pores, which ultimately improved the macroscopic stability of CCSA with high porosity (~ 99.6%), lowest density (5-16 mg/cm 3 ) and 60% strain endurance. Optimized CNF to CS (C 7 CS 7 to C 7 CS 1 )) ratios de ned the stable microporous structure of CCSA over crosslinking by formaldehyde. The r-CCSA presented identical properties with the random pores distribution all over the structure holding pores size of 12 to 20 µm and comparatively high thermal conductivity (λ) along the axial (0.038Wm − 1 K − 1 ) and radial (0.038Wm − 1 K − 1 ) orientation. u-CCSA presents anisotropic performance with lamellas spacing of