Figure 1 illustrates the route from bacterial cellulose (BC) to carboxylated CNCs. Briefly, BC was washed and dried before being weighed and placed in a custom-built reactor for hydrolysis using gaseous HCl. The hydrolysis took place overnight at 1 bar, as reported previously.16
After washing thoroughly, hydrolyzed BC was obtained. Gel permeation chromatography showed that the molecular weight of the hydrolyzed BC is lower than that of the untreated BC (Fig. 2). The mass average molecular weight (Mw) decreased from 326,600 to 43,400 Da for BC and hydrolyzed BC, respectively. The degree of polymerization (DP) decreased from roughly 2,000 to 270 for BC and hydrolyzed BC, respectively, indicating successful hydrolysis.
The subsequent electrochemical oxidation was performed on a laboratory scale using 0.5 g of hydrolyzed bacterial cellulose in a 100 mL carbonate buffer solution with 2 mmol of TEMPO. The reaction occurred in a three-electrode setup consisting of a carbon foam working electrode, a titanium mesh counter electrode, and an Hg|Hg2SO4|K2SO4 reference electrode. The process was optimized for high efficiency, achieving a current efficiency of 98% with a stable potential of 0.5 V applied over 24 h. Hydrolyzed BC was electrochemically oxidized using this setup for up to 24 h at a constant potential of 0.5 V. Figure 3 shows a cyclic voltammogram (CV) obtained using our electrochemical setup containing a TEMPO catalyst in a pH 10 carbonated buffer measured against a reference electrode. The anodic peak in the CV shows the oxidation of the TEMPO radical to the N-oxoammonium ion, while the cathodic peak shows the reduction to the N-hydroxy form. Under these conditions, the reaction of the TEMPO catalyst is reversible.
We demonstrated the pH dependence of TEMPO during the electrochemical reaction (Figure S1) and its reversibility using different buffer systems (Figure S2) to identify the most efficient buffer system for our electrochemical oxidation. The use of high-surface-area electrodes reduced oxidation times from the previously reported 45 h21 to 12 h. Our electrochemical oxidation process is less labor-intensive than traditional TEMPO oxidation methods and enables precise and controlled oxidation. Our buffer system could also be reused twice after the initial electrochemical reaction upon removal of the newly carboxylated CNCs and the introduction of fresh substrate, which improves scalability and efficiency. Currently, we are optimizing the process to reduce the reaction time by adjusting electrode material and current settings, making it sustainable and suitable for large-scale production. More information on the system’s electrochemistry is provided in the SI.
The surface charge, and hence the carboxylate content of CNCs, was determined by conductometric titration. With increasing electrochemical oxidation time, a higher carboxylate content was obtained, reaching a maximum of 1.24 ± 0.15 mmol/g after 24 h for the hydrolyzed BC (Fig. 4). We also attempted the electrochemical oxidation of bacterial cellulose without prior hydrolysis, which resulted in a carboxylic acid content of 0.7 ± 0.2 mmol/g after 24 h. Therefore, hydrolysis appears to increase the accessibility of BC in CNC preparation.
As the formation of carboxylate groups proceeds via aldehyde intermediates, they were monitored by conversion to their oxime derivatives for elemental analysis, post electrochemical oxidation. This analytical protocol was adapted from procedures predominantly used in dialdehyde cellulose research.25 Conventional aldehyde-specific titration techniques for TEMPO-oxidised cellulose could not be performed on the small sample amounts available from the used electrochemical setup. After 3 h of electrochemical oxidation, the aldehyde content was higher than the carboxylate content, but with increasing time, the aldehyde content remained roughly the same while the carboxylate content increased (Fig. 4). Both aldehyde and carboxylate content were determined at least in triplicate for each oxidation time. We demonstrated that our method is effective for other cellulose sources, as evidenced by the successful oxidation of a cotton linter (1.35 ± 0.21 mmol/g) under the same conditions.
Traditional TEMPO oxidation (TEMPO/NaBr/NaClO system) for BC provided an aldehyde and carboxylate content of 0.1 and 1.05 mmol/g, respectively.26 In our electrochemical oxidation, we achieved a carboxylate content of 1.07 ± 0.13 mmol/g, which is consistent with values reported in the literature for BC after 12 h. Our findings correlate somewhat with previously reported carboxylate contents of Cladophora also produced using electromediated TEMPO-oxidations, which were 0.595 and 0.599 mmol/g after 1 or 3 days, respectively.22 The smaller surface charge in Cladophora is justified by the larger crystal size present in Cladophora CNCs.
The CNC yield after electrochemical oxidation was 75% across all oxidation times, indicating that no considerable solubilization of highly charged chains occurred. The remaining yield loss is predominantly assignable to handling losses during workup. Pääkkönen et al. reached a similar CNC yield of 80% using HCl gas hydrolysis and conventional TEMPO oxidation.16 Other processes using different cellulose sources typically reach CNC yields of around 15–50%, including hydrolysis using sulfuric acid (yields 15–50%),7 and esterification (yield 25%).27 Details on our CNC yield calculations can be found in the SI.
We evaluated whether the surface modification of our CNCs can be considered quantitative. Lacking a model crystallite, we used cotton linters as a reference, as they share a similar crystallite dimension to our CNCs.28 Using this model, we estimated a complete conversion of our CNCs, signifying that all surface-accessible sites were oxidized to carboxylic acid (detailed calculations are available in the SI).29, 30 The full conversion should be interpreted cautiously, as it may be influenced by the use of an inadequate crystallite model and the potential for overoxidation on crystallite end sites.31
The length and width of the electrochemically oxidized CNCs were investigated using atomic force microscopy (AFM, Fig. 5a and Figure S5) and transmission electron microscopy (TEM, Figure S3), respectively.
After 9 h of electrochemical oxidation, the average CNC length was 353 ± 277 nm (Figure S5). After 12 h, the average CNC length remained similar (349 ± 205 nm; Figure S5), but after 24 h had decreased slightly to 313 ± 155 nm (Fig. 5b). We determined the width of CNCs to be 13 ± 6 nm after 24 h of electrochemical oxidation (Figure S4). Vasconcelos et al. produced CNCs from BC under different hydrolysis conditions, showing lengths from 200–1000 nm (average 622 ± 100 nm) and a width ranging from 16–50 nm (average 34 ± 14 nm),32 in good agreement with our findings. Another report on CNCs from BC showed average lengths and widths of 855 nm and 17 nm, respectively.28 However, no standard deviation was provided, probably due to a small available sample pool. CNCs produced from BC using conventional TEMPO oxidation were reported to have lengths of 100–300 nm (average 170 nm).16 It should be noted, however, that only a small number of CNCs were measured in that study. Full length and width distributions and AFM and TEM images of all samples can be found in the SI (Figure S3-5). Essentially, the lower the charge of the CNCs, the more difficult it is to disperse them. The dispersion ability is seen in Figure S5a, where after 9 h of electrochemical oxidation, the samples had to be diluted by a factor of 10 to avoid agglomeration. Such dilution was not required after 12 or 24 h of electrochemical oxidation.
The entire process was examined by calculating its environmental impact factor (E factor). A higher E factor means more waste and, consequently, more significant adverse environmental impact. The ideal E factor is zero. More recently, the inventor of the E factor concept suggested using simple E factors (sEF) and complete E factors (cEF), depending on the process's development stage.33 The sEF does not consider solvents and water and hence assumes recycling of solvents. In contrast, the cEF accounts for all process materials, including solvents and water, assuming no recycling and is more appropriate for total waste stream analysis. We have calculated both sEF and cEF of our process. Both the sEF and cEF for our gaseous hydrolysis are 2, as no solvents or water was used. The sEF of the electrochemical oxidation is 2, whereas the cEF is 134. As the TEMPO catalyst and the buffer solution can be reused,23, 24 our entire process has a simple E-factor (sEF) of 2 for both HCl gas hydrolysis and electrochemical oxidation. Equations and calculations of sEF and cEF can be found in the SI.
A cost analysis of our CNC production process was conducted. All details on the calculations, including energy consumption, can be found in the SI (Table S2-S5). We want to emphasize that the absolute cost of these materials is exclusively for comparative purposes and only considers lab-scale production.
Our method distinguishes itself with a notably high yield, demonstrating its potential for efficient scalability in cellulose nanoparticle production compared to existing literature. We emphasize responsible practices by implementing stringent measures to minimize the use of harmful substances. Our approach contributes to the production of oxidized cellulose by reducing chemicals such as NaClO, NaBr, and NaOH, underscoring our commitment to responsible and conscientious nanocellulose production.
