Residual Strain and Nanostructural Effects during Drying of Nanocellulose/Clay Nanosheet Hybrids: Synchrotron X-ray Scattering Results

Cellulose nanofibrils (CNF) with 2D silicate nanoplatelet reinforcement readily form multifunctional composites by vacuum-assisted self-assembly from hydrocolloidal mixtures. The final nanostructure is formed during drying. The crystalline nature of CNF and montmorillonite (MTM) made it possible to use synchrotron X-ray scattering (WAXS, SAXS) to monitor structural development during drying from water and from ethanol. Nanostructural changes in the CNF and MTM crystals were investigated. Changes in the out-of-plane orientation of CNF and MTM were determined. Residual drying strains previously predicted from theory were confirmed in both cellulose and MTM platelets due to capillary forces. The formation of tactoid platelet stacks could be followed. We propose that after filtration, the constituent nanoparticles in the swollen, solid gel already have a “fixed” location, although self-assembly and ordering processes take place during drying.


Supporting Text Calculation of liquid evaporation coefficient of the wet mats
The liquid evaporation coefficient θ of the wet mats surface is described as: 1 where g h is amount of evaporated water per hour (kg h -1 ), A is water surface area (m 2 ), x s is maximum humidity ratio of saturated air at the same temperature as the water surface (kg kg -1 ) (kg H2O in kg dry air), is humidity ratio of the air (kg kg -1 ) (kg H2O in kg dry air).
x = 0.62198p w /(p a − p w ) where p ws is saturation pressure of water vapor, p w is the partial pressure of water vapor in moist air, p a is atmospheric pressure of moist air.
When the environment temperature is 25 °C, the parameters for the calculation and final θ values of the wet mats are listed in the Table S1. A and g h are measured from the drying experiment directly. The , and values are obtained from literatures. 2,3  When free water is evaporated, such as water evaporation from a swimming pool, the evaporation coefficient can be calculated as: 1 is the velocity of air above the water surface. In order to compare to drying of wet mats, we assume the velocity is 0, then ree has a value of 25 kg m -2 h -1 , which is in the same order of the values of wet mats drying. However, the values of ree is larger than ; this suggests that the CNF based wet mats have water retention properties.

Explanation of the MTM XRD peaks in Figure 2a
The intensity ratio of reflections 200:130 are reported to be of the order of 8:5, and the two contributions cannot be separated in the in-plane randomly oriented sample. Both peaks have asymmetric shape with tailing towards higher angle. This is related to the fact that the clay particle has highly anisotropic shape (individual crystal sheet is anisotropic due to the high aspect ratio, but in-plane orientation is random), and the diffraction spot is elongated in the reciprocal space perpendicular to the basal plane. Even when the center of the spot is not in the Bragg-condition, the tail of the spot overlaps the Ewald sphere giving rise to the intensity at higher angle.

Description of Video S1-S3
Video S1, S2 and S3 show the exposure patterns of CNF-water, C/M-water and C/M- The incident X-ray beam is perpendicular to the sample stage surface. At position 1, the beam is parallel to the film surface; at position 2, the beam is perpendicular to the film surface.

Explanation of SAXS curves in Figure S3
With increased drying time, the CNF-water sample shows slightly decreased intensity, due to reduction of the film thickness. The C/M-water sample shows a bump in the beginning since the contrast is between water and particles; with further drying, the particles aggregated, and the bump disappeared. The C/M-EtOH sample has larger porosity compared to the C/M-water sample, and the larger contrast between CNF/MTM and air thus results in higher SAXS intensity.  Explanation of 2D X-ray patterns in Figure S5 Within a thin film, CNF fibrils exhibit planar orientation while MTM platelets follow uniplanar orientation. 4 Herein, anisotropic X-ray patterns are observed. After drying, CNF-water sample shows tiny SAXS streaks ( Figure S5a

Explanation of 1D WAXS curves in Figure S6
In Figure S6a-b, we labelled the MTM and CNF peaks. The peaks in the low region (0.3-0.5 Å -1 ) is MTM 001 reflection, which can be used to calculate the inter-sheet distance of the stacked platelets. In Figure S6c and S6e, one can observe after drying for 100 min, the MTM 11 2c2 in the main text. With further drying, the peaks shifted to higher and then were relatively stable, corresponding to decreased and subsequent constant d-spacing values.
In C/M-EtOH sample, tactoids formed after solvent exchange from water to ethanol. The d-spacing of MTM in ethanol ( Figure S6f, red line) is larger than that in water (18-17 Å vs. 16-15 Å), this is because ethanol has lower dielectric constant, dipole moment and surface tension compared to water. Ethanol penetrates the interlayer space of the MTM platelets more easily than water. 5 With further drying, the d-spacing was reduced ~ 1 Å. At about 40-50 min, two peaks were observed (indicated by the arrows in Figure S6d

Explanation of the artefact of CNF orientation
The CNF 200 peaks become sharper due to the overlapping of MTM peaks, so the fitted CNF orientation index is an artefact. In this case, we should analyze the realistic orientation based on a theoretical estimation: First, we obtained MTM orientation index based on the fitting of MTM peak ~ 2.0 Å -1 (Figure 5b2). The CNF has a fibril feature which follows the planar orientation while the MTM has a platelet feature following the uniplanar orientation ( Figure   S15), 4 theoretically their orientation index should be the same. By utilizing the MTM orientation index, the CNF profile along the azimuthal angle direction can be simulated by a modeling method developed previously. 6 Figure S16 shows the simulation results, the simulated CNF profiles match well with the experimentally measured CNF profiles of CNFwater sample. This means the CNF orientation index in C/M-water is likely to be the same as for the CNF-water sample; the presence of MTM is not affecting CNF orientation.

Surface roughness characterization
The 3D topographic features of membrane surfaces were examined by an optical surface metrology confocal profilometer (Leica DCM8, Wetzlar, Germany). Typically, samples are attached to a coverslip with adhesive tapes and placed on the moving stage. The images were captured using a lens with 20×/100× magnification and images with minimal defects and asperities were chosen for further calculating.

Tensile testing method
The CNF/MTM nanocomposites films with a thickness around 30 µm were cut into a rectangle with 70 mm in length and 5 mm in width. The samples were cut by a LEICA Microtome blade after preconditioned in a 50 ± 2% relative humidity and 22 ± 1 °C room for at least 2 days. The films were tested by a Universal Testing Machine (Instron 5944, USA) equipped with a 500 N load cell and a video extensometer. The span length was set to 25 mm and the strain rate was 10% min -1 . All materials were tested using at least 5 specimens. The modulus was determined by fitting a linear curve from the initial elastic region.