**3. Experimental Section**

Phosphoric acid (99%), phosphorus pentoxide (98.5%), triethyl phosphate (99.8%, all Sigma-Aldrich, Schnelldorf, Germany), *n*-hexanol (98%, Acros Organics, Geel, Belgium), tetra-*n*-butylammonium fluoride (TBAF) trihydrate (≥97%, Sigma-Aldrich, Schnelldorf, Germany), dimethyl sulfoxide (DMSO, ≥99.9%, Merck, Darmstadt, Germany), molecular sieves 4 Å (Roth, Karlsruhe, Germany), and ethanol (96 vol.% and absolute, Merck, Darmstadt, Germany) were of highest available grade.

Cotton linters (CL) and total chlorine-free bleached hardwood (eucalyptus) pre-hydrolysis kraft pulp (hwPHK) were provided by collaboration partners of the COST E41 action. Their weight average molecular weight (MW) as well contents of carbonyl group were determined as detailed elsewhere [36,54]:


Prior to dissolution or phosphorylation, the cellulosic source materials were activated by disintegration in water (solid-to-liquid ratio 1:400, wt./wt.). The obtained slurry was freeze-dried and stored at +4 ◦C until further processing.

### *3.1. Phosphorylation of Cellulose*

Phosphorylation was accomplished as described elsewhere [39,42,55]. In brief, 250 g of phosphorus pentoxide was placed under argon atmosphere protection in a 1 L 3-necked round-bottom flask equipped with air-tight mechanical stirrer, condenser, and dropping funnel. Phosphoric acid (354 g) was added portion-wise under external ice-cooling. Triethyl phosphate (185 mL) was added slowly within 6 h. A clear, highly viscous solution was obtained after about 48 h of continued stirring. The thus prepared phosphorylation reagen<sup>t</sup> (PR) was stored in anhydrous atmosphere at +4 ◦C until further use. A defined amount of cellulosic source material (typically 1.0 g) was activated for phosphorylation by repeated short-time (10 s) disintegration in a large excess of water (1:400, *w*/*v*). This was followed by submersion in consecutive bathes of ethanol and *n*-hexanol (ca. 1:80, *w*/*v*; 2 times á 24 h for each of the organic solvents). About 40 mL of the supernatant *n*-hexanol was then decanted by gentle squeezing and replaced by an equal volume of PR. The flasks containing the reaction mixtures were then mounted onto a heated horizontal shaking device and left at 50 ◦C for 72 h. After that, phosphorylated cellulose was separated from the liquid phase and consecutively washed twice with *n*-hexanol, ethanol, and eventually deionized water. The last step was repeated until the filtrates showed a negative molybdenum blue reaction. The degree of phosphorylation (DSP) was analyzed as suggested previously [56]. An aliquot of the sample (100 mg) was subjected to microwave-assisted pressure digestion in HNO3/H2O2 using the following temperature program: 25 → 85 ◦C (3 ◦C min−1), 85 → 145 ◦C (ca. 1.5 ◦C min−1), 145 → 200 ◦C (ca. 1 ◦C min−1, 12 min hold), cooling to room temperature. The phosphorous content of the obtained digestion liquor was analyzed by ICP-OES and used to calculate the DSP according to Equation (1) [56].

$$DS\_p = \frac{162\text{x}}{3100 - 84\text{x}}\tag{1}$$

Finally, an aliquot of phosphorylated cellulose was subjected to solid-state 31P and 13C NMR experiments to confirm covalent introduction of monophosphate groups.

### *3.2. Preparation of the Cellulose Solvent System TBAF (16.6 wt.%)* / *H2O (0.95 wt.%)* / *DMSO*

TBAF trihydrate (157.55 g) was dissolved in anhydrous DMSO (342.45 g). Then, 325 g of freshly dried molecular sieves 4 Å (600 ◦C, 3 h, argon atmosphere) was added portion-wise over a period of 96 h to bind about 50% of the crystal water. When the remaining water content as determined by Karl Fischer titration reached about 1.5 wt.%—coinciding with the onset of a faint yellowish (λ = 420 nm) coloration [36]—the drying agen<sup>t</sup> was filtered o ff. The solution was stored at +4 ◦C until further use. Immediately before cellulose dissolution, the above-prepared TBAF/DMSO solution was diluted with anhydrous DMSO as described previously to obtain a final TBAF content of 16.6 wt.%. The final water content was 0.95 wt.% which ensures good dissolution performance and largely avoids decomposition of TBAF [36].

### *3.3. Cellulose Dissolution, Casting, Coagulation, Solvent Exchange, and scCO2 Drying*

The respective non-derivatized and/or phosphorylated cellulosic materials (3 g) were dissolved portion-wise in the pre-heated (60 ◦C) and continuously stirred solvent system (97 g) to give a 3 wt.% solution. While the obtained cellulose solutions were visually clear after two hours already, optical microscopy (magnification 100×) revealed that full dissolution requires 3–4 h. After four hours, the obtained dopes were cast into molds of cylindrical geometry (Ø = 10 mm, l = 20 mm). The molds were then immersed in 96 vol.% ethanol (EtOH) to initiate cellulose coagulation using a dope-to-ethanol volume ratio of 1:50. After 24 h of residence time, the molds were opened to expose the cylindrical, free-standing lyogels. The latter were transferred into consecutive bathes of 96 vol.% ethanol (3 × 24 h) and absolute ethanol (2 × 24 h) ensuring a gel-to-liquid ratio of 1:20 (*v*/*v*). Drying of the transparent gels was performed using supercritical carbon dioxide equipment (scCO2). Samples were loaded in a 500 mL autoclave equipped with two separators for carbon dioxide recycling (Separex SF1, Separex, France). After heating and pressurization to 40 ◦C and 10.5 MPa, the samples were dried at constant CO2 flow (2.5 kg <sup>h</sup>−1) for one hour following recommendations of previous studies [57,58]. It has been shown that increasing the pressure beyond 10 MPa slightly reduces the remaining volume [30]. This is similar for increasing the temperature, as hornification and loss of tightly bonded surface water reduce interfibril distances.

Once ethanol was quantitatively extracted from the voids, the system was slowly and isothermally depressurized to prevent pore collapsing and condensation by Joule–Thomson cooling [59]. Both volume (±0.1 mm3) and weight (±1 mg) of the samples (5 replicates) were recorded after each step of the sequential solvent exchange and after scCO2 drying.
