Modification of nanocellulose films in deep eutectic solvents using vinyl esters

Bleached birch kraft pulp sheets were used as a cellulose raw material for nanocellulose film production after wet disintegration. Imidazole (purity > 98.0%), TEMACl (purity > 98.0%), vinyl butyrate (stabilized with mequinol, MeHQ) (purity > 98.0%), vinyl octanoate (stabilized with MeHQ) (purity > 99.0%), vinyl laurate (stabilized with MeHQ) (purity > 99.0%), and vinyl palmitate (stabilized with MeHQ) (purity > 96.0%) were from Tokyo Chemical Industry Co. Ethanol (purity > 96%), and acetone (purity 100.0%) were from VWR International. Deionized water was used throughout the study, if not mentioned otherwise.

Cellulose was mechanically nanofibrillated to CNF using a super mass colloider (Masuko, MKCA6-2, Japan). First, pulp sheets were torn apart and diluted to water, and the suspension was mixed to achieve a homogenous mixture. Then, the fibers were fed multiple times through the grinder while reducing the gap between grinding discs. The grinding procedure was described in detail by Selkälä et al. (2018). The consistency of the obtained CNF suspension was 1.7 wt.%.

CNF films were prepared by measuring 0.3 g (dry) of nanofibrillated cellulose to a beaker glass and then diluting it with water to a total weight of 110 g (0.27 wt.%). The sample was degassed through the mild ultrasonic treatment (Elmasonic P, Elma Schmidbauer GmbH, Germany) for 10 min, followed by vacuum filtration on top of the polymer membrane (Durapore DVPP 0.65 μm, Merck Millipore Ltd., Ireland) using a negative pressure of approximately 800 mbar. After the film was formed and the excess water was removed, another membrane was placed on top of the wet CNF cake. The entire system was placed between two cardboards for further vacuum drying (Karl Schröder KG, Germany) at 93 °C using a negative pressure of 900 mbar for 10 min to obtain self-standing films with a basis weight of 72.4–75.8 g/m². The thickness of the films varied between 46.5 and 52.8 µm as measured with a thickness gauge (FT3, Hanatek Instruments, UK) as an average of three random points.

DES of imidazole and TEMACl (molar ratio of 7:3) was prepared by weighing the components into the decanting glass (total of 140 g) and heating the mixture in the oil bath at 80 °C. When approximately half of the liquid was formed, a magnetic stirrer was enabled and kept constantly mixing to speed up the formation of DES. When a clear liquid was obtained, the DES was used as a reaction medium to modify CNF films.

The UV/O₃ treatment was conducted using UVO Cleaner 42–220 (Jetlight Company Inc., USA). The device was equipped with a low-pressure mercury-vapor grid lamp, generating light at 253.7 nm with an average intensity of 28–32 mW/cm². The samples were illuminated from both sides for 10 min. After the UV/O₃ treatment, the samples were immediately processed further.

Before modification, the CNF film was washed with acetone or treated with UV/O₃ to remove possible impurities that could disturb the reaction between the reagent and cellulose film (Johansson et al. 2011; Österberg et al. 2013a). The vinyl reagent, i.e., vinyl butyrate, vinyl octanoate, vinyl laurate, or vinyl palmitate, was added into the DES and then submerged the film into the DES system horizontally. To prevent the magnetic stirrer from touching the film, a custom-made film protection design was used (Figure S24 in supporting material). The ratio between the reagent and film was 10:1 in weight. After 15 min reaction time at 80 °C, the modified film was washed four times (~ 5 min each time) with water to remove the DES and residues of the unreacted reagent from the film. Fresh washing liquid was used in every washing phase. Finally, the film was placed between two membranes (Durapore DVPP 0.65 μm, Merck Millipore Ltd., Ireland), which were further covered with two cardboards. Then, samples were dried in a vacuum dryer for 10 min (Karl Schröder KG, Germany).

The chemical characteristics of the films were analyzed using ATR-FTIR. The spectra were recorded from small film specimens using Bruker Hyperion 3000 (USA). The wavelength range of 600–4000 cm⁻¹ and a total of 32 scans at a resolution of 4 cm⁻¹ were used.

The chemical composition and bonding on the surface were analyzed through XPS using the Axis Supra from Kratos Analytical Ltd (UK). The instrument uses the Al Kα spectral line with a photon energy of 1486.6 eV. An electron flood gun was used for charge compensation. The acquisition of narrow regions of the spectrum was conducted with the pass energy of 20 eV. CasaXPS software was used for the spectra analysis. The U 2 Tougaard type background was subtracted from the measured data, and mixed Gauss–Lorentzian line shapes were used for spectra fitting. The spectra were calibrated to C–C/C–H peak of C 1 s at 285 eV binding energy.

The contact angle measurements were conducted using the DSA100 system (Krüss, Germany). The measurement system used a high-speed camera (1000 fps) to capture the drop behavior, which was consequently analyzed using drop analyzing software. A set of high-speed videos was recorded during the measurements for further analysis and contact angle extraction for various films. The Milli-Q water size of the droplet was retained at approximately 1 µL to reduce the influence of gravitational flattening. In all cases, the contact angles were extracted using the height–width method, where a rectangle enclosed by a contour line is interpreted as being the segment/part of a circle. For each sample, three measurements at different locations were conducted. The acquired results were averaged, and the standard deviations were calculated.

Water vapor permeability (WVP) of the films was determined according to ASTM Standard E96/E96M. All films were conditioned in a separate room with controlled conditions (23 °C ± 1 °C and 50% ± 2.5% RH) for 48 h, and then the measurements were conducted in the same conditions (room was equipped with a fan to ensure good circulation of air inside the room). The test sample was put between 100 mL Schott Duran^© glass bottle and a cap with a hole (area of 5 cm²) drill on it. The glass bottle was filled with water one-fourth of its volume, and a rubber seal was used to ensure a tight sealing. Weight loss of the bottles was recorded at 1 h intervals for 8 h in the course of 2 days. A weight loss–time curve was drawn, the slope of the strain hardening region was estimated, and the water vapor transmission rate (WVTR) was determined by dividing the slope by the exposed film area. WVP was calculated as follows:where S is the saturation vapor pressure at the test temperature (2800 Pa at 23 °C), R₁ is the RH in the bottle expressed as a fraction (1), R₂ is the RH in the room expressed as a fraction (0.5), and h is the thickness of the film sample.

Moisture absorption measurements were conducted by weighing small film specimens in 0% and 100% RH and recording the weight value of three separate specimens of each sample. For 0% RH, the samples were dried and kept in a desiccator to absorb moisture in the silica gel. For 100% RH, the samples were kept in a desiccator, where water-absorbing silica gel was replaced with water. Cumulative weight percentage increase of the specimens was calculated, and the results were reported as an average of three measurements per film.

Mechanical properties of the films were measured using a universal testing machine (Zwick Roell, Ulm, Germany) with a 2 kN load cell. The samples were prepared by cutting five to seven strips from each film with a width of 5 mm and length of 50–70 mm. The test was conducted in 50% RH with fully wetted samples at the wet state. The films were kept in a separate room with controlled conditions (23 °C ± 1 °C and 50% ± 2.5% RH) 48 h before analysis. For the wet state tensile measurements, the strips were immersed in water 1 h before analysis, and excess water was swept with tissue paper before testing. Before mechanical tests, the thickness of the strips was measured as an average of three random measuring points using a thickness gauge (FT3, Hanatek Instruments, UK). For tensile testing, the gauge length was set to 40 mm, and the strain was controlled with a speed of 5 mm/min with a prestrain value of 0.1 and 1 MPa for dry and wet films, respectively. Young’s modulus was calculated from the slope of the line from the linear region of the stress–strain curve, and an ultimate tensile strength was determined as the stress of specimen fracture. For dry tensile measurements (50% RH), the results are reported as an average of five tests and an average of three tests for wet tensile measurements.

FESEM (JEOL JSM-7900F, Japan) was used to study the surface and cross-section morphologies of the films. Images were taken at an accelerating voltage of 5.0 kV. The samples were sputtered with platinum (High-Resolution Sputter Coater, Agar Scientific, UK) before analysis using a sputtering time of 30 s and a current of 40 mA.

CNF films were modified with vinyl esters with different alkyl chain lengths using a DES of imidazole and TEMACl (molar ratio of 7:3) as a reaction medium. Four different vinyl esters, i.e., vinyl butyrate (three carbons in alkyl chain), vinyl octanoate (seven carbons in alkyl chain), vinyl laurate (11 carbons in alkyl chain), or vinyl palmitate (15 carbons in alkyl chain), were used. The modified films are named based on the vinyl ester used. Additionally, the role of film pretreatment in DES-mediated esterification reaction was revealed, i.e., films were washed with acetone or treated with UV/O₃ before modification (labeled as UV/O₃ treated films). Pure CNF film and UV/O₃ treated CNF film served as reference films.

In this study DES was used as a reaction medium and imidazole in DES could act as catalyst via formation of acyl imidazole with vinyl esters (Pires et al. 2015). DES based on imidazole and TEMACl was harnessed as it bears no similar chemical groups compared to cellulose (i.e. hydroxy groups) which could react with esterification agent (i.e. vinyl ester). Previously, pyridine has been used a reaction medium for cellulose modification, but pyridine is more volatile than the imidazole of the DES used in this study (vapor pressure of pyridine and imidazole are 2.4 kPa at 20 °C and 0.3 Pa at 20 °C, respectively). CNF films have been esterified using fatty acid chlorides, but their drawback is a long reaction time (90 min) and high reaction temperature (110 °C). In addition, fatty acid chlorides are more harmful than vinyl esters used in this study (Nawaz et al. 2013; Hinner et al. 2016; Balasubramaniam et al. 2020). Other esterification reagents have been studied, but most of them are utilized in homogenous media with solvent such as ionic liquid. However, dissolution of cellulose by ionic liquids prevents their use in current application (Kakko et al. 2017; Kostag et al. 2019).

The reaction between vinyl esters and hydroxy groups of cellulose occurred through an esterification route (Fig. 1). ATR-FTIR spectroscopy was used to analyze the chemical structure of the reacted films in terms of ester bond formation and elucidate the effect of UV/O₃ treatment on the modification. Figure 2 shows the ATR-FTIR spectra of pristine CNF and laurate-modified films and their UV/O₃-treated counterparts. The spectra of the other samples are presented in supporting material in Figure S1. Compared with reference CNF film, the modified films showed a new peak at a wavenumber of 1730 cm⁻¹, indicating the presence of O–C=O bond of the ester group and a successful esterification reaction of CNF film in the DES medium. Unexpectedly, the ester peak was not found with a reaction of vinyl palmitate using the UV/O₃-pretreated CNF film despite multiple attempts (Figure S1). This could be because of a low degree of substitution of the sample (see Table 1) or low resolution of ATR-FTIR. Later, it was clearly confirmed with contact angle measurements and XPS analysis that the reaction also occurred with vinyl palmitate.

The UV/O₃-pretreated CNF film showed a carbonyl peak similar to the modified films without adding any vinyl reagent. Ozone is a very powerful oxidation agent; therefore, it can introduce carboxylic acid or carbonyl groups at the material surfaces. Ozone mainly oxidizes the primary hydroxy group of cellulose, yet simultaneously secondary hydroxy groups can be oxidized. Ozone can also results in cleavage of glycosidic bonds between cellulose molecules. Typically, carbonyl groups are seen in ozone-treated polymer films in the 1700–1750 cm⁻¹ region (Hedenberg and Gatenholm 1996; Romero-Sánchez et al. 2005; Österberg et al. 2013a; Wen et al. 2020). Here the carbonyl peak was observed at a wavenumber of 1730 cm⁻¹; otherwise, the spectra of CNF and UV/O₃ CNF samples remained unchanged.

Furthermore, XPS analysis was conducted to support the findings from ATR-FTIR spectroscopy and further characterize the surface chemical composition of the CNF films. Table 1 presents the elemental concentration of carbon, oxygen, and nitrogen, and concentration of carbon bonds (XPS survey spectra and C1s deconvolution spectra of CNF, butyrate, laurate, and UV/O₃-treated counterparts are presented in supporting information in Figures S2–S7 and S8–S13, respectively. A full table of elemental concentration of the samples is presented in Table S1). All esterified films modified in DES exhibit a higher O–C=O bond concentration (2.3–4.2%) than the original CNF film (O–C=O bond concentration of 2.0%), supporting the findings from ATR-FTIR measurements about the successful esterification reaction. In alignment with these results, the C–C bond concentration of the DES-modified films (17.8%–27.2%) decreased more than the reference CNF film (28.7%).

The O–C=O bonds in pristine CNF film likely originate from the hemicelluloses of the pulp, or the bonds can be formed during the pulping and bleaching processes since the unmodified cellulose does not contain these bonds (Bayer et al. 2016). Additionally, the UV/O₃-treated CNF film exhibits a higher O–C=O bond concentration (3.1%) than the original CNF film, supporting the FTIR results of CNF oxidation. The oxidation of lignin on the pulp surface was earlier seen in the XPS data as a decrease in the C–C bond and an increase in the O–C=O bond concentration (Koljonen et al. 2003). Similarly, the C–C bond concentration decreased from 11 to 4%, and O–C=O bond concentration increased (percentages not reported) after treating CNF film samples with UV/O₃ (Österberg et al. 2013a). This result is consistent with our results showing that UV/O₃-treated CNF film had a lower C–C bond concentration (24.2%) than the original CNF film (28.7%) and a higher O–C=O bond concentration (3.1% and 2.0%).

The O–C=O bond concentration decreased as a function of the alkyl chain length of the vinyl reagent. Vinyl butyrate with the shortest carbon chain resulted in the O–C=O bond concentration of 4.2%, whereas vinyl palmitate had an O–C=O bond concentration of 2.5%. This suggests that the reaction rate decreased with increasing alkyl chain length. A similar trend has previously been reported in the literature (Leszczyńska et al. 2018).

The UV/O₃ pretreatment of the CNF films increased the reaction efficiency of esterification only with vinyl laurate (O–C=O bond concentration with and without UV/O₃ was 3.4 and 2.8%, respectively). With vinyl butyrate, UV/O₃ treatment had no effect on O–C=O concentration, while it decreased with vinyl octanoate and palmitate after the UV/O₃ treatment.

XPS analysis was also used to elucidate whether the DES medium reacted with the films or there were any traces of the DES remaining in the films. The films were revealed in terms of their nitrogen content as both DES components, i.e., imidazole and TEMACl, contain nitrogen in their structure. Both the pristine CNF and UV/O₃-treated films exhibited 0.6% nitrogen content, whereas the DES-modified samples had a nitrogen amount ranging from 0.0 to 0.4% (Table 1). Therefore, the DES modification decreased the amount of nitrogen, suggesting that DES did not react with the films or there were not any DES traces left in the films. These findings are consistent with our previous experiments (Lakovaara et al. 2021).

The influence of esterification in DES and UV/O₃ treatment on the structure and morphology of the films was investigated through FESEM imaging. Figure 3 shows the surface (a–d) and cross-sectional FESEM images (e–h) of the pristine CNF, UV/O₃ CNF, butyrate, and UV/O₃ butyrate (FESEM images of other samples are presented in supporting information Figure S14–S23). Based on the images, the DES modification or UV/O₃ treatment did not affect the visual appearance of the films, and all films possessed a tightly packed and nonporous structure. The previous results also support that UV/O₃ treatment does not degrade the CNF film despite the strong oxidation efficiency of ozone (Österberg et al. 2013a). However, the UV-treated and DES-modified films have less rough surface structure than non-UV-treated and DES-modified films. Moreover, all cross-sectional images (Fig. 2e–h) display a typical layered structure of CNF films.

The role of esterification and UV/O₃ pretreatment in the alteration of surface characteristics of the films was evaluated through contact angle measurements (Fig. 4). All DES-modified films exhibit higher contact angles (77.0°–106.5°) than unmodified reference CNF film (37.8°); thus, the esterification increased the hydrophobicity of the films. These contact angles are substantially higher than those obtained previously with CNF films modified using n-octylsuccinic anhydride in DES of imidazole and TEMACl, where the film exhibited a contact angle of 51° (Lakovaara et al. 2021). The highest contact angle was obtained with the UV/O₃-treated and vinyl laurate-reacted film with a contact angle of 106.5°. Among non-UV/O₃-treated films, vinyl octanoate resulted in the highest contact angle of 94.7°.

The UV/O₃ treatment was conducted to clean and enhance the reactivity of the surface of the CNF films. When CNF films are exposed to air upon drying, carbonaceous contamination layers accumulate to the highly hydrophilic CNF surface (Johansson et al. 2011; Österberg et al. 2013a). The UV/O₃ treatment can remove carbon-based impurities and attach carboxylic acid or carbonyl groups at the material surfaces, thus making the films more hydrophilic (Österberg et al. 2013a). Supporting these previous findings, the contact angle of UV/O₃-treated CNF film was very low (13.3°), significantly lower than the reference CNF film (37.8°). For all DES-modified films, the UV/O₃ treatment increased the contact angle. With the films reacted with shorter alkyl chains, i.e., vinyl butyrate and octanoate, the contact angle increase was only a few degrees (for butyrate from 77.1° to 80.2° and octanoate from 94.7° to 98.2°). Meanwhile, the increase was more significant with longer alkyl chain samples (laurate and palmitate). The contact angle increased from 79.0° to 106.5° and 77.0° to 97.7° with laurate and palmitate, respectively.

The contact angle value seems to be affected by at least two factors: the length of the alkyl chain and reaction rate (i.e., the O–C=O bond content; see Table 1) (Leszczyńska et al. 2018). For example, butyrate has a higher reaction rate (4.24%) than palmitate (2.47%); however, the contact angles are the same (77°). Octanoate had a slightly lower reaction rate (4.05%) than butyrate, but a longer alkyl chain length of vinyl octanoate resulted in a higher contact angle of 94.7°. Laurate had a lower reaction rate (2.82%) than octanoate, with a contact angle of 79.0° despite a longer carbon chain length. The influence of the UV and ozone pretreatment on improvement in reaction rate was noted with vinyl laurate, which also increased its contact angle.

Figure 5 shows the water vapor permeabilities of the films. The reference CNF film exhibited the highest (i.e. the worst value) WVP of 18.67 g·µm/m²·d·kPa, whereas all the other films had a lower WVP ranging from 14.36 to 18.11 g·µm/m²·d·kPa, which is consistent with results reported in the literature (Wang et al. 2018). The UV/O₃-pretreated CNF film presented a WVP value of 15.68 g·µm/m²·d·kPa, which is second highest of the UV-treated samples. The lowest value (i.e., the best value) was with octanoate sample with 14.36 g·µm/m²·d·kPa.

Water vapor transmission through the film depends on many factors, such as film thickness, density, porosity, crystallinity, and hydrophobicity. Water vapor penetrates through the film in a multistep diffusion process where water molecules adsorb on the surface of the film first, then diffuse through the film, and finally evaporate from the surface of the film (Chi and Catchmark 2018). The enhanced WVP of the UV/O₃-treated film might be attributed to the cross-linking promoted by the carbonyl and carboxyl groups. These bonds can probably resist better the bond breakage induced by the moisture than the hydrogen bonds between pristine cellulose chains (Klein et al. 2008; Torres et al. 2010). Similar to contact angle values, the influence of the UV/O₃ treatment on WVP was observed with samples modified with short alkyl chain samples, i.e., butyrate and octanoate.

Figure 6 shows the cumulative moisture absorption of the films in the 0–100% RH range. All UV/O₃-treated films displayed a lower moisture absorption than their nonpretreated counterparts. The lowest moisture absorption was with UV/O₃ butyrate with a 16.6 wt.% weight increase, and the second-lowest was with UV/O₃ CNF with a 17.8 wt.% weight increase. For other UV/O₃-treated films, the moisture absorption increased as the function of alkyl chain length. For non-UV/O₃-treated samples, the values were close to each other, i.e., laurate had the lowest moisture absorption of 21.9 wt.%, whereas the other films exhibited moisture absorption between 23.4 and 24.4%. Overall, the moisture absorption was presumably associated with the interfibrillar bonding of CNFs as the UV/O₃ can promote the cross-linking, whereas the hydrogen bonding reduces as the alkyl chain length increases (Torres et al. 2010).

Mechanical properties of the films were evaluated using a tensile test in 50% RH and wet state. Figure 7 shows the stress–strain curves of the films. Table 2 summarizes the tensile strength (at break), strain, and modulus values. The original CNF film had the highest tensile strength of 175 MPa at 50% RH, and all esterified films without UV/O₃ treatment displayed a slightly lower tensile strength ranging from 158 to 171 MPa, similar to the values reported in the literature (Henriksson et al. 2008; Klemm et al. 2011; Lindström 2017). The UV/O₃ treatment slightly reduced the film strength with almost all samples. This reduction in strength properties may be attributed to the oxidation of cellulose and further breakage of interfibrillar hydrogen bonds, as noted earlier (Taniguchi and Okamura 1998; Torres et al. 2010; Österberg et al. 2013a; Benítez et al. 2013).

The esterification improved the strength properties of the films at the wet state as the original CNF film displayed the lowest tensile strength of 9.6 MPa. Meanwhile, the modified films possessed a wet tensile strength between 11.5 and 18.9 MPa. This improved wet tensile strength was more pronounced with samples modified with a longer alkyl chain length, i.e., laurate and palmitate. The tensile strengths and modulus were further increased at the wet state when the films were exposed with the UV/O₃ treatment. However, these samples were stiffer, and they presented lower strain values. This phenomenon is seen from Fig. 7b, where stress–strain curves of the UV/O₃-treated samples are separated from those of non-UV/O₃-treated ones. Overall, the UV/O₃ palmitate exhibited the highest wet tensile strength of 18.9 MPa. Unexpectedly, the UV/O₃-treated CNF film had the second-highest tensile strength of 17.6 MPa at wet state, which can be attributed to the UV/O₃-induced cross-linking (Torres et al. 2010) and decrease in the amount of hydroxy groups.

Water had a significant impact on the films since wet films had a notable lower tensile strength and Young’s modulus and increased strain than dry films. This behavior is very common for nanocellulose films since water breaks the internal hydrogen bonds and promotes interfibrillar slippage. These factors allow the film stretch better under stress (Benítez et al. 2013; Sehaqui et al. 2014; Benítez and Walther 2017). The effect of water on the films is also displayed by the stress–strain curves. The stress–strain curves at wet conditions contain a short or nonexistent linear elastic region, and the curve transforms straight into a strain hardening region (Sethi et al. 2019).