*3.3. Mechanical properties*

Figure 7 compiles the tensile tests data, namely in terms of Young's modulus, tensile strength and elongation at break, determined from the stress-strain curves. Although the tensile tests were not performed for the cross-linked PMPC due to its lack of film-forming aptitude, the cooperative effect between PMPC and BNC relies on the mechanical properties of both nanocomposites. Overall, the Young's modulus and tensile strength of the two membranes increased with the increasing content of the cellulosic substrate (Figure 7A,B), on account of its good mechanical performance, namely Young's modulus of 6.6 ± 1.8 GPa and tensile strength of 221 ± 48 MPa. In fact, the former parameter increased from 430 ± 150 MPa for PMPC/BNC\_2 with 21 wt.% of BNC to 3.3 ± 0.8 GPa for PMPC/BNC\_1 with 46 wt.% of BNC (Figure 7A), whereas the tensile strength increased from 18 ± 4 MPa for PMPC/BNC\_2 to 69 ± 15 MPa for PMPC/BNC\_1 (Figure 7B). In contrast, the elongation at break decreased with the increasing content of BNC from 6.0 ± 1.4% for PMPC/BNC\_2 to 2.8 ± 0.3% for PMPC/BNC\_1, as shown in Figure 7C. This means that the nanocomposites are more pliable than the stiff BNC nanofibers with an elongation at break of 4.7 ± 1.0%.

The dependence of the membranes' mechanical performance on the amount of BNC is in tune with earlier studies of other BNC-based nanocomposites with polymers of low mechanical properties [14,35,36]. For example, Zhijiang et al. [14] prepared a chitosan/BNC-based hydrogel composite for dye removal, whose Young's modulus increased from 96.5 MP for pure chitosan (dry state) to 244 MPa after the incorporation of BNC nanofibers grafted with carbon nanotubes into the chitosan hydrogel. The same behavior was obtained for the tensile strength and elongation at break [14].

**Figure 7.** (**A**) Young's modulus, (**B**) tensile strength and (**C**) elongation at break of neat BNC and PMPC/BNC nanocomposites; the error bars correspond to the standard deviations; the asterisk (\*) denotes statistically significant differences with respect to the neat BNC (*p* < 0.05).

#### *3.4. In vitro antibacterial activity*

Materials with antibacterial activity are relevant for application in multiple fields [37,38] since they can inhibit the growth and simultaneously kill pathogenic bacteria that are harmful to human health [39]. The MPC polymer is known for having antimicrobial and antifouling properties [19–21], which can be a major benefit in reducing/avoiding bacterial growth in contaminated water. This hypothesis was validated by assessing the growth inhibition of gram-positive (*S. aureus*) and gram-negative (*E. coli*) bacteria. *E. coli* was selected for being frequently present in contaminated water, which is a strong indication of recent sewage or fecal contamination. *S. aureus* is not so frequently present in contaminated waters; however, di fferent strains have already been detected in urban wastewater, namely the methicillin-resistant *S. aureus* ST398 [40].

Figure 8 outlines the antibacterial activity of PMPC/BNC nanocomposites and of the neat BNC membrane for comparison purposes. The inoculation of both bacteria in culture media without any sample was used as an experimental control. The neat BNC membrane, along with the experimental control, do not a ffect the bacterial viability of both *S. aureus* (Figure 8A) and *E. coli* (Figure 8B). This was expected given that BNC is reported not to inhibit the growth of *S. aureus* [41,42], *E. coli* [36,41,42], and other microorganisms such as *Pseudomonas aeruginosa*, *Bacillus subtilis* [42] and *Candida albicans* [43]. In fact, BNC can even be used as a substrate for microbial cell culture [44].

The bacterial killing of *S. aureus* by the two PMPC/BNC nanocomposite membranes is markedly concentration-dependent, as portrayed in Figure 8A. The PMPC/BNC\_1 nanocomposite with 54 wt.% of cross-linked PMPC originated a significant reduction (*p* < 0.05) of bacterial concentration relatively to the control, causing a maximum of 2.5–log CFU reduction after 24 h of incubation. The PMPC/BNC\_2 with 79 wt.% of cross-linked PMPC reached a higher bacterial inactivation of 4.3–log CFU reduction after 24 h, which indicates that this membrane can be considered an e ffective antibacterial because according to the American Society of Microbiology (ASM), every new approach has to prove an efficacy of 3–log10 reduction of CFU before being considered antimicrobial or antibacterial [43]. This antibacterial activity is mainly attributed to the trimethylammonium cation that is known for imparting antimicrobial properties [45]. When comparing the activity of the PMPB/BNC membranes with literature, Bertal et al. [46] verified that the triblock copolymer containing PMPC originated an inhibitory zone up to six times greater than the corresponding control against *S. aureus* and a reduction of bacterial growth by 45% compared with the experiments carried out in the absence of PMPC-based copolymer. The authors also claimed that the addition of the copolymer to a 3D-skin model infected with *S. aureus* reduced bacterial recovery by 38% compared to that of controls over 24–48 h [46].

**Figure 8.** Effect of BNC, PMPC/BNC\_1 and PMPC/BNC\_2 on the bacterial killing (CFU) of ( **A**) *S. aureus* and (**B**) *E. coli* after 24 h of exposure; error bars represent the standard deviation (three independent experiments); the asterisk (\*) denotes statistically significant di fferences with respect to the control and neat BNC (*p* < 0.05).

Regarding the *E. coli* bacteria (Figure 8B), the picture is quite different and both nanocomposites exhibit a lower reduction with 1.3– and 1.8–log CFU reduction for PMPC/BNC\_1 and PMPC/BNC\_2, respectively. A similar behavior was reported by Fuchs et al. [47] that witnessed no antibacterial activity towards *E. coli* for one MPC copolymer. In fact, this could be expected given that *E. coli* is a gram-negative bacterium whose killing mechanism is more difficult to prevent due to the low permeability of their membranes as discussed previously in detail [48,49].

#### *3.5. Water-uptake and dye removal capacity*

Table 2 presents the water-uptake values for BNC and the two PMPC/BNC nanocomposite membranes after immersion in aqueous solutions of 0.01 M HCl (pH 2.1), phosphate buffer saline (pH 7.4) and 0.01 M NaOH (pH 12.0) for 48 h at RT. Overall, the water-uptake vividly increased with the increasing content of cross-linked PMPC. At pH 7.4, it increased from 101 ± 12% for neat BNC up to 639 ± 23% for PMPC/BNC\_1 (54 wt.% of PMPC) and 899 ± 44% for PMPC/BNC\_2 (79 wt.% of PMPC) (Table 2). In acidic aqueous solutions, PMPC/BNC\_1 can absorb 6.3 ± 0.4 g of water *per* g of membrane, while for PMPC/BNC\_2 the value is 9.1 ± 0.2 g of water *per* g of membrane. At pH 12, PMPC/BNC\_1 absorbs 6.4 ± 0.4 g of water *per* g of membrane, whereas for PMPC/BNC\_2 the water-uptake is 9.1 ± 0.3 g of water *per* g of membrane.

**Table 2.** Water-uptake (water-uptake) of neat BNC and the two PMPC/BNC nanocomposite membranes at different pH media for 48 h at RT.


a Measured after immersion in 0.01 M of HCl aqueous solution; b Measured after immersion in phosphate buffer solution; c Measured after immersion in 0.01 M of NaOH aqueous solution. All values are the mean of three replicates with the respective standard deviations.

The larger water-uptake of the nanocomposites is correlated with the hydrophilic nature of the phosphorylcholine moiety of the cross-linked PMPC. Additionally, water-uptake is not pH-dependent since there are no significant differences (the means difference is not significant at α = 0.05) for the individual membranes under the distinct conditions of acidity or basicity. This can be explained by the unique hydration state of the PMPC chains, where the phosphorylcholine moieties have a hydrophobic hydration layer that do not disturb the hydrogen bonding between the water molecules, as discussed by Ishihara et al. [18]. This is an important characteristic in the water remediation context given that contaminated water can have different pH values. Furthermore, the higher water-uptake of PMPC/BNC\_2 is an indication of a higher removal capacity of water-soluble dyes. After 48 h of immersion in aqueous solutions of different pH values, the two nanocomposites were oven dried (at 40 ◦C) and the final weights demonstrated that the polymer loss ranges between 1%–2%, which emphasizes the insignificant leaching of the cross-linked PMPC from the BNC network.

The removal of two model ionic organic dyes, namely methylene blue (MB) and methyl orange (MO), from water samples at room temperature after 12 h was assessed as a proof-of-concept. While MB is a heterocyclic cationic aromatic compound that is used either as a dye or a drug with for example antimalarial, antidepressant and anxiolytic activity [50], MO is a heterocyclic anionic aromatic compound that is widely used in the textile, pharmaceutical and food industries, and also as an acid-base indicator. Both azo dyes are potentially toxic towards humans and the environment [51].

Figure 9A shows that the PMPC/BNC membranes can indeed retain the model water pollutants as confirmed by the different color of the nanomaterials. This is further corroborated by the data shown in Figure 9B where the dye removal capacity is plotted for each membrane. The pure BNC can remove 0.55 ± 0.12 mg of MB and 0.50 ± 0.06 mg of MO *per* g of membrane. These low removal values

were expected, given the lack of binding sites in pure BNC for both cationic and anionic organic dyes. Furthermore, these values are comparable with the dyeability reported by Shim and Kim [52] in their study about the coloration of BNC fabrics with different dyes using *in situ* and *ex situ* methods.

**Figure 9.** Photographs (**A**) and (**B**) dye removal capacity of BNC, PMPC/BNC\_1 and PMPC/BNC\_2 after 12 h of immersion in the dye aqueous solution, and (**C**) photographs of the MB and MO aqueous solutions removal from paraffin oil by PMPC/BNC\_2 nanocomposite.

Concerning the nanocomposites, PMPC/BNC\_1 can remove 3.14 ± 0.19 mg g<sup>−</sup><sup>1</sup> of MB and 3.32 ± 0.31 mg g<sup>−</sup><sup>1</sup> of MO, whereas PMPC/BNC\_2 has a removal capacity of 4.44 ± 0.32 mg g<sup>−</sup><sup>1</sup> for MB and 4.56 ± 0.43 mg g<sup>−</sup><sup>1</sup> for MO. Comparing with pure BNC, the dye removal capacity of PMPC/BNC\_1 is 5.7 and 7.4 times higher for MB and MO, respectively, while PMPC/BNC\_2 removes 8.1 and 9.1 times more MB and MO, respectively, than pure BNC. The higher dye removal capacity of PMPC/BNC\_2 is consistent with its higher PMPC content (Table 1). Moreover, the two nanocomposites can remove both cationic and anionic dyes due to the zwitterionic nature of the cross-linked PMPC which can establish electrostatic interactions with either MB or MO model dyes. Worth mentioning is the fact that the PMPC/BNC nanocomposites can easily and quickly remove both MB and MO (25 mg mL−1) from the bottom of a paraffin oil container without the removal of any oil, as exemplified for PMPC/BNC\_2 in Figure 9C. This is a good indication of the lack of affinity of the nanocomposites towards the hydrophobic oil and affinity for water or aqueous solutions. A similar behavior was observed for MB (aqueous solution, 100 mg <sup>L</sup>−1) removal from silicone oil by sulfated-cellulose nanofibrils aerogels [53].

When compared with literature, the dye removal capacity of the PMPC/BNC nanocomposites is lower than that achieved for example with highly carboxylated (COO–) nanocrystalline cellulose with a maximum removal capacity of 101 mg g<sup>−</sup><sup>1</sup> for MB [54], or with the amino-functionalized cellulose nanofibrils-based aerogels with 266 mg g<sup>−</sup><sup>1</sup> for MO [55]. These higher removal capacities are most likely associated with the simultaneous high content of surface binding sites and specific surface area in the first case [54], and the aerogel structure in the second case, which translates into materials with very high porosity and low density [55]. Still, the dye removal values of the PMPC/BNC membranes prepared in the present study are comparable for instance with those achieved with the sulfated-cellulose nanofibrils aerogels that removed *ca*. 5 mg g<sup>−</sup><sup>1</sup> of MB at an adsorbent dosage of 16 mg mL−<sup>1</sup> [53].

Hence, the adsorbent nanocomposites developed in the present work present a customizable combination of properties, namely antibacterial activity, water-uptake and dye removal capacity, that depend on the amount of the individual components (*i.e*. PMPC and BNC), and that reveal their potential application in the context of water remediation.
