Functionalized Cellulose Nanocrystals as Active Reinforcements for Light-Actuated 3D-Printed Structures

Conventional manufacturing techniques allow the production of photoresponsive cellulose nanocrystals (CNC)-based composites that can reversibly modify their optical, mechanical, or chemical properties upon light irradiation. However, such materials are often limited to 2D films or simple shapes and do not benefit from spatial tailoring of mechanical properties resulting from CNC alignment. Herein, we propose the direct ink writing (DIW) of 3D complex structures that combine CNC reinforcement effects with photoinduced responses. After grafting azobenzene photochromes onto the CNC surfaces, up to 15 wt % of modified nanoparticles can be introduced into a polyurethane acrylate matrix. The influence of CNC on rheological properties allows DIW of self-standing 3D structures presenting local shear-induced alignment of the active reinforcements. The printed composites, with longitudinal elastic modulus of 30 MPa, react to visible-light irradiation with 30–50% reversible softening and present a shape memory behavior. The phototunable energy absorption of 3D complex structures is demonstrated by harnessing both geometrical and photoresponsive effects, enabling dynamic mechanical responses to environmental stimuli. Functionalized CNC in 3D printable inks have the potential to allow the rapid prototyping of several devices with tailored mechanical properties, suitable for applications requiring dynamic responses to environmental changes.

. Schematic representation of the chemical modification of CNC. The grafting is represented here on C6, accounting for the higher probability that the reaction occurs in this position due to the reduced steric hinderance of this group. However, the reaction could also occur on the other OH groups of the AGU (Anhydroglucose Unit).

Note on XPS:
The DR1 is attached to the cellulose, since ref CNC samples do not show presence of N as contamination. Indeed, nitrogen seems to be present with 1% at. However, the N-O group tends to degrade under X-ray, becoming a N-H group. Longer acquisition times would be required to have a full quantification, but this would lead to destruction of the functional group and decreasing of the NO2 peak at 405-406 eV.

Estimation of Substitution fraction of CNC accessible hydroxyls from Solid State NMR:
The number of OH reacting with DR1 over the number of accessible OH was estimated from the NMR results in Figure  S3 by applying the following equation Where ∫ 1 corresponds to the area of the four aromatic carbons of DR1 at 152 ppm, while ∫ 1 corresponds to the area of the C1 of the anhydroglucose unite at 105 ppm. Each glucose unit bears 3 hydroxyl and for CNC with a transversal size of around 5 nm, only around 45 % (estimated from Sugiyama et al. 3 ) of the total amount of hydroxyl groups is accessible at the surface of the particles.

Degree of substitution from elemental analysis:
The DS per anhydroglucose unit was estimated directly from %N of CNC-DR1 assuming the grafted molecules to be composed of a DR1 and a Cyanuric Chloride unit as illustrated in Figure S1 and that the reference CNC do not contain nitrogen since the measured value is below the detection limit of the instrument. The following equation was applied: 4 Where MN is the nitrogen molecular weight, nAGU the number of nitrogen atoms in an anhydroglucose unit (AGU), MAGU the molecular weight of the AGU, Mcom is the molecular weight of the grafted compound and nN is the number of nitrogen atoms in the grafted compound. The substituted fraction of OH groups among those accessible at the surface was obtained by dividing the DS by a factor of 3 (three hydroxyls per AGU) and a factor of 0.45 (fraction of surface hydroxyls). 5 Note on elemental analysis: The degree of substitution here calculated is likely an over estimation arising from the assumption that one DR1 and one cyanuric chloride are linked together (according to the presence of Cl in the final product and the C-Cl bending band observed from FTIR). However, being a one pot reaction, there is the possibility that more cyanuric chlorides are attached to CNC than DR1. The calculated value remains an estimation.
(       Figure S10. Thermally activated shape memory effect of the printed composites for CNC (a,b,e,f) and d,g,h). Figure S11. Compression of a CNC-DR1 composite printed as negative stiffness honeycomb. The sample was illuminated for 5 minutes after each cycle in order to recover its initial shape.