*2.7. Thermostability*

Thermogravimetric analysis (TGA) conducted in helium atmosphere revealed some interesting differences between the different sets of samples. This applies not only to the cellulosic source materials and the aerogels obtained thereof. It refers also to the comparison between aerogels prepared from non-derivatized cellulose and their phosphorylated analogues (*cf.* Figure 6, right). The two non-processed CL and hwPHK samples were virtually fully stable up to about 325 ◦C. Here, prominent polymer degradation sets in, exactly as reported for cotton linters before [50]. Thermal degradation is largely completed at 350 ◦C already when about 90% of the initial weight has been released as volatiles. Continued pyrolysis and carbonization up to 950 ◦C caused a further 50% reduction of the remaining mass.

Processing of the cellulosic source materials into lightweight open-porous aerogels gives rise to a series of differences in material properties. Compared to the cellulosic fibers of the source materials, the diameters of the aerogel network forming fibrils is much smaller [36] and the degree of crystallinity lower [32]. This translated into much higher specific surface areas and moisture sensitivity. Nanostructuration furthermore imparts thermal insulation properties to aerogels presumably considerably retarding heat-transfer. These properties are assumed to increase the sensitivity of respective aerogels towards temperature alterations and may promote kinetically less favored side reactions. This is supported by the following observations: (i) the above-described humidity-dependent adsorption of water is well reflected by smaller weight losses (ca. 2–4 wt.%) in the temperature range ≤ 150 ◦C; (ii) intra- and intermolecular condensation occurring between 150 and

240 ◦C [51] give rise to further weight loss of 1–2 wt.%; (iii) depolymerization starts at 300 ◦C already likely due to reduced crystallinity, and (iv) the mass remaining at 350 ◦C (36–37 wt.%) and 950 ◦C (24–25 wt.%) was much higher compared to the non-processed cellulosic source materials. The latter might be a result of stabilization by cross-linking and hornification.

**Figure 6.** Bulk density [%] of phosphorylated and non-derivatized aerogels after two and 84 days during conditioning at 65% RH (left). Thermogravimetric analysis of cellulose II aerogels from non-derivatized and slightly phosphorylated cotton linters (CL, CL-P) and hardwood prehydrolysis kraft pulp (hwPHK, hwPHK-P). TGA profiles of the unprocessed source materials were recorded for reference purposes (right).

The TGA profiles of the CL-P and hwPHK-P aerogels differed strongly from that of their non-derivatized analogues despite the low degrees of phosphorylation. While desorption of physically bonded water was negligible for the CL-P and hwPHK-P samples, prominent weight reduction started at 235 ◦C already. However, only about 36% of the initial weight was lost in this step, which is much lower compared to the cellulosic source materials (90 wt.%). Formation of volatiles continued at significantly lower rates until the final temperature of 950 ◦C and a residual weight of about 40% relative to the initial weight of the samples was reached. It is assumed that acidic phosphate groups as present in CL-P and hwPHK-P are capable to boost formation of low-molecular compounds. This includes levoglucosan, which has been shown to be involved in re-polymerization of volatile cellulose pyrolysis products and secondary char formation (*cf.* [52,53]). The latter could give rise to the formation of thin, heat-shielding char layers efficiently preserving a reasonably large weight fraction under enhanced pyrolytic conditions.
