A Direct Preparation of Cellulose Nanocrystals by ZnCl₂-Based Deep Eutectic Solvent

Abstract

Here, we demonstrated a direct method to produce cellulose nanocrystals (CNCs) with a rod-like shape from microcrystalline cellulose by a ZnCl₂-based deep eutectic solvent (DES) with a high yield (~80.1%). We obtained CNCs, crystalline index (68.9%), with a width of ~30–50 nm and a length of 200–400 nm. Importantly, we were able to functionalize the CNCs with an acetyl, -(CO)CH₃, group, which could potentially modulate the hydrophobic property of the CNCs. We attributed the formation of the CNCs to the Lewis acid effect of ZnCl₂, which can hydrolyze the amorphous cellulose regime. Our study opens a new path to directly isolate cellulose nanocrystals with several functional groups on the surface of CNCs.

1. Introduction

Cellulose nanomaterials are considered one of the most promising building blocks [1,2,3,4,5], and have wide range of applications, for example, as wound dressings [6,7,8], antimicrobials [9,10], or as ionic conductors in energy storage and conversion [11,12,13]. For example, Burger et al. reported zinc chloride treatment for partial cellulose amorphization, which then enables functionalization of cellulosic fibers to obtain flame-retardant and catalytic properties [14]. Up to now, cellulose nanocrystals (CNCs) are generally extracted from microcrystalline cellulose using a hydrolysis reaction in strong, concentrated mineral acid such as a mixture of HCl and citric acid, or H₃PO₄. The process involves a hydrolysis reaction of glycosidic bonds between glucose units and an esterification reaction of hydroxy groups on the cellulose surface [9]. The less-ordered regions of cellulose undergo hydrolysis rapidly, Simultaneously, esterification happens to a fraction of the surface hydroxy groups, turning them into sulfate half-ester groups that are anionic under practical working solution conditions.

Despite the widespread industrial production of CNCs by mineral acid, the hydrolysis of cellulose usually faces some challenges such as its hydrophilic character, the lack of thermal stability, and environmental problems due to the strongly corrosive and non-recycled chemical agents. As a result, many alternative chemical isolation methods based on oxidizing agents, ionic liquids, organic acid and molten salt [15] have emerged to produce the longer, fibrillated form of nanocellulose, namely cellulose nanofibrils (CNFs).

Deep eutectic solvents (DESs), consisting of hydrogen bond donors (HBDs) and acceptors (HBAs), have emerged as innovative green solvents according to 12 principles of green chemistry [16]. DESs have been used to pre-treat microcrystalline cellulose to obtain pre-treated microcrystalline cellulose, and then post-treatment such as mechanical treatment or ultrasonic treatment is needed to produce cellulose nanomaterial [17,18,19,20,21,22]. For example, Sirviö et al. reported that DESs can be used to produce cellulose nanofibers (CNFs) with the support of microfluidizer [23,24,25] and cellulose nanocrystals (CNCs) after mechanical disintegration [26]. Yu et al. reported that choline chloride (ChCl)/oxalic acid DES was applied to produce CNFs and CNCs with the support of mechanical disintegration [27] or microwave radiation [28]. Recently, Shi et al. reported the production of carboxylated cellulose nanofibers (yield of 90.12%) with a diameter of ~3.4 nm, high aspect ratio up to 2500, and a high carboxyl content of 1.5 mmol g⁻¹ [22]. In general, many studies demonstrate the potential application of DESs in producing cellulose nanomaterials, but a simple, one-step method to produce nanocrystal cellulose with a functional group is still lacking.

In this study, we investigated the possible use of DES based on ZnCl₂ (as a Lewis acid) to extract cellulose nanomaterials. We observed that DES based on ZnCl₂ can directly produce nanocrystal cellulose without post-treatments. The cellulose nanocrystals have a length of 200–400 nm and a width of ~30–50 nm, with the crystalline structure of cellulose I. Importantly our method allows us to functionalize nanocrystal cellulose with the carboxyl, -(CO)CH₃, group. Our results further showed that DES can be applied to produce cellulose nanocrystals with functional groups.

2. Materials and Methods

2.1. Chemicals

a-cellulose powder (size 100–150 µm) was purchased from Macklin Viet Nam, Ha Noi, Vietnam. Zinc chloride (ZnCl₂), oxalic acid, choline chloride (ChCl) and glycerol were purchased from Sigma–Aldrich, Merck Vietnam Ltd., Ha Noi, Vietnam. All the chemicals were used directly without further treatment or purification.

2.2. Preparation of Deep Eutectic Solvents (DESs)

Deep eutectic solvents (DESs) of ZnCl₂ and choline chloride, namely ZnCl₂-Ch, were prepared by mixing zinc chloride and choline chloride, molar ratio 2:1 respectively, at 130 °C for 1 h to obtain a transparent solvent. Deep eutectic solvents (DESs) of ZnCl₂ and glycerol, namely ZnCl₂-Gly, were prepared by mixing zinc chloride and glycerol, molar ratio 1:3 respectively, at 90 °C for 1 h to obtain a transparent solvent. Lastly, oxalic acid and choline chloride, molar ratio 2:1 respectively, were mixed at 110 °C for 1 h to obtain the DES, Oxa-Ch. All DESs obtained were homogeneous, colorless and transparent mixtures. They were cooled down to 90 °C for the reaction with cellulose.

2.3. Extraction of Nanocrystals Cellulose

The extraction of nanocrystals cellulose by DES is presented in Scheme 1. Briefly, a-cellulose was added to the ZnCl₂-Ch or Oxa-Ch DES with a weight ratio of 1:30, respectively. For the ZnCl₂:Gly DES, to maintain the ratio of ZnCl₂ and a-cellulose, the weight ratio was modified to 1:60. The mixtures were then stirred in sealed containers at 90 °C for 16 h. To stop the reaction, deionized water (DI) was added to the mixture, and then the mixture was centrifuged at 10,000 rpm for 15 min to remove the DES. The step was repeated at least 4 times to obtain the CNCs. Finally, the cellulose pastes were vacuum dried at 20 °C and kept in storage or diluted in DI water for further characterization. The reaction yield was calculated by the mass difference of the sample before and after the reaction.

2.4. Material Characterization

The overall changes in particle size were observed by optical microscope (Olympus BX53M, Olympus Vietnam Co. LTD., Dong Nai, Vietnam). Dynamic light scattering (Zetasizer Nano ZS90 Malvern, DKSH Vietnam Co., Ltd., Ha Noi, Viet nam) was used to measure the cellulose particle size. The measurement was performed in deionized (DI) water and followed the Brookhaven guide for sample preparation.

Field Emission Scanning Electron Microscopy (FE-SEM). The cellulose suspension was spin coated on SiO₂ (100) wafer, and then dried at 40 °C under vacuum overnight. Then, 20 nm of Ti thin layer was deposited on top of the sample by e-beam evaporator (Kurt J. Lesker PVD-75, Kurt J. Lesker Company, Shanghai, China) at vacuum 10⁻⁶ torr and depositing rate 0.2 nm·s⁻¹. The observation was immediately performed after sample preparation by FE-SEM (JEOL, JSM-IT800, JEOL ASIA PTE. LTD, Singapore) at a low kV. For EDS evaluation, the accelerating voltage was set at 20 kV with an accumulation time of 1 min and the working distance was fixed at 8 mm.

X-ray Diffraction. A D8 ADVANCE X-ray powder diffractometer (Bruker, Berlin, Germany) was employed to obtain the XRD patterns of cellulosic fibers before and after treatments using Ni-filtered Cu K radiation (λ = 0.15406 nm) at 40 kV and 30 mA. The diffraction data were collected scattering 2 angles from 10° to 70° at the rate of 0.03°/0.7 s in ambient conditions. Segal’s method, expressed as the formula below, was applied to quantify the crystallinity index (CrI) of each sample [29].

CrI (%) = A₂₀₀/(A₂₀₀ + A_am) × 100%

(1)

where A₂₀₀ is the peak area at characteristic plane (200) of cellulose crystalline domain, and A_am is the peak area at the characteristic (101) plane of the amorphous domain in cellulose.

Fourier transform infrared spectroscopy (FT/IR Nicolet iS50, Thermo Scientific, Ha Noi, Vietnam) equipped with the attenuated total reflectance (ATR) accessory was used to analyze the chemical structure of the samples. The characteristic spectra of dried samples were obtained after 32 scans with resolution of 4 cm⁻¹ in the wavenumber range of 500 to 4000 cm⁻¹. Thermogravimetric analysis (TGA, NETZSCH STA 449F3, LMS Technologies Vietnam Co., LTD, Ho Chi Minh, Vietnam) was performed between ambient temperature and 900 °C with a heating rate of 10 °C/min, under a 2.5 L/min flow rate of inner gas. The obtained data were used to study the thermal stability and degradation behavior of raw and treated samples.

3. Results

First, we performed the reaction of a-cellulose with ZnCl₂-Ch DES in a weight ratio of 30:1 at 90 °C for 16 h. After working out the reaction, we obtained a white cellulose paste and part of it was suspended in water for optical observation. Figure 1 presents the optical image observed for fresh a-cellulose and the cellulose paste. As shown, the α-cellulose has a rod-like shape with a width of 20 to 50 μm and a length of 100 to 300 μm, whereas the cellulose paste has no well-defined shape and much smaller dimensions (less than 2 μm). The optical result provided evidence that during the chemical treatment with DES, the cellulose undergoes a hydrolysis reaction.

A particle size histogram of cellulose nanomaterials dispersed in aqueous suspension was estimated by dynamic light scattering (DLS) analysis, as presented in Figure 2A. The result shows that the cellulose particle is on a nano scale with an average size of 250 ± 75 nm. The effect of the reaction time on the particle size is presented in Figure 2B. As shown in Figure 2B, the particle dimension dropped to ~250 nm after only 4 h of reaction; however, we still observed some microcrystalline cellulose after the reaction by the optical microscopy. As a result, we kept the reaction time of 16 h to be sure that all the microcrystalline cellulose converted to nanocrystal cellulose. The yield of the reaction was calculated by the weight difference before and after the reaction, and we obtained a yield of ~80.1%.

To further confirm the size of the cellulose particles, we then performed observation by scanning electron microscopy. We first dropped the cellulosic colloidal suspension on the SiO₂ (100) substrate, and then after air drying, 20 nm of Ti layer was deposited on top of the cellulose layer by e-beam evaporation [30]. The role of the 20 nm Ti layer is to improve the conductivity of the cellulose layer but also to avoid the degradation of the cellulose material during SEM observation. In Figure 3A, the SEM image shows that the cellulose particle has a rod-like shape, with a length of 200–400 nm and a width of 20–50 nm, which leads to the small aspect ratio (less than 20).

Next, we performed a similar chemical reaction using ZnCl₂-gly and Oxa-Ch DES and then performed the SEM observation as shown in Figure 3B and Figure 3C, respectively. We notice here that ZnCl₂-gly and ZnCl₂-Ch DES consist of ZnCl₂ (Lewis acid) acting as HBA whereas in Oxa-Ch DES, oxalic acid (Brønsted acid) acted as HBA.

As shown, the dimensions of cellulose nanocrystals obtained from ZnCl₂-gly are similar to those from ZnCl₂-Ch, which was expected. Importantly, we observed that the cellulose nanocrystals obtained from Oxa-Ch also have a rod-like shape with a length of 100–300 nm and a width of ~10 nm, which is significantly smaller than those from ZnCl₂ DES, as shown in Figure 3C. We can attribute the dimension difference to the effectivity of the hydrolysis reaction, caused by the Lewis acid and Brønsted acid.

The chemical structures of the CNCs obtained from DES treatment were characterized by infrared spectroscopy, as presented in Figure 4A. First, we observed the vibrational bands between 899 and 1163 cm⁻¹ in all obtained spectra. These absorption peaks are usually assigned to β-(1–4) glycosidic ether links, β-glycosidic linkages, C-O-C glycosidic stretching, C-OH stretching vibration, and C-O stretching, indicating the presence of cellulose structure [29]. Two bands at 1426 and 1315 cm⁻¹ can be assigned to -CH₂- symmetric bending groups of cellulose. The vibration at 1740 cm⁻¹ also appeared in the spectra of the CNCs obtained from Oxa-Ch, associated with the vibrational mode of the C=O linkage in the carboxylic group, which provides support for the functionalization of the COOH group to the CNCs by oxalic acid. In general, information from FT-IR spectra demonstrated that the chemical structure of the CNCs was maintained similarly to that of a-cellulose and importantly, the CNCs produced by Oxa-Ch were functionalized with the -COOH group.

X-ray diffractograms of all samples, as presented in Figure 4B, were employed to study the crystalline structure and crystallinity. The peaks at 14–17°, 22.3° and 34° were observed in all obtained samples, which are attributed, respectively, to (110), (101), (200), and (400) in cellulose I structure [29,31]. In the diffractogram of CNCs^Oxa-Ch, the intensity of the (200) peak at 22.3° is significantly greater than that of a-cellulose and the CNCs^ZnCl2-Ch, which suggests the effective removal of most cellulose amorphous regimes. We noted here that ZnCl₂ molten salt was used to treat the cellulose, but it usually led to the change in the crystal structure of cellulose (from cellulose I into cellulose II) [32,33,34]. The cellulose crystallinity of all samples was compared through the crystallinity index (CrI), which was determined quantitively using Segal’s method. The CrI of the a-cellulose was 68.5%, similar to the CrI (68.8%) of the CNCs^ZnCl2-Ch, and the CrI of the CNCs^Oxa-Ch increased to 81.9%. Here, the XRD analysis indicates (i) the preservation of crystalline structure (Cellulose I) after the reaction with DES solution, and (ii) that removal of the amorphous regime in the CrI of the CNCs^Oxa-Ch is greater compared to that in the CNCs^ZnCl2-Ch, which is also consistent with the dimension analysis.

TGA/DTG analysis was performed in the range of 35 °C to 770 °C to determine the thermal stability of all samples, as presented in Figure 5. The decomposition of the obtained CNCs occurred in three main steps as expected for lignocellulose materials. At T^onset ~110 °C, the cellulose material initially lost a small amount of weight (<2%), which could be assigned to the amount of moisture in all samples. The crystalline region was then decomposed through gas phase transition and tar formation in the range of 220 °C to 450 °C, which led to the second step of degradation, with a significant weight loss (~80%). The T^onset of the second step, ~300 °C for a-cellulose and the CNCs^ZnCl2-Ch, respectively, is also significantly higher than that (~200 °C) of the CNCs^Oxa-Ch generated from the reaction with Oxa-Ch DES. During the last stage above 450 °C, the cellulosic crystalline structure was completely decomposed into volatiles and tar. We also note that the tar (~23%) obtained from Oxa-Ch DES was two times greater than from the others (~11%). The lower thermal stability and higher tar concentration could be explained by the existence of protons implemented in the CNCs, which could catalyze the reaction of the tar formation.

4. Discussion

It is known that deep eutectic solvents (DESs) can be applied to produce cellulose nanofibers or nanocrystal shapes with post-treatment including mechanical or ultrasonic treatment as summarized in Table S1. In addition, in most cases, DES was made of Brønsted organic acid, which usually results in the CNCs functionalized by the hydrophilic -COOH group, as was also observed in our study [22]. In the current study, by using DES made of ZnCl₂, we observed that first, cellulose rod-like shapes can be obtained from the direct treatment of microcrystalline cellulose with the DES and second, our CNCs with a width of ~50 nm and a length of 200–400 nm, have greater dimensions in comparison with the CNCs produced by proton-based DES and mineral acid [30]. The greater dimensions, especially in the length of the CNCs, indicated the less efficient hydrolysis of microcrystalline cellulose, which was supported by the XRD result since the crystalline index (CrI), 68.8%, of CNCs obtained from ZnCl₂-based DES was similar to that of the raw microcrystalline cellulose (68.5%), while the CrI of the CNCs from Oxa-Ch was 81.9%. We could attribute less efficient hydrolysis to the weaker acidity, as shown in Figure S1 from the Supplementary Matarial. In addition, the viscosity of ZnCl₂-based DES was approximately in the range of 150 to 200 mPa·s, which is also two times greater than that of Oxa-Ch DES (80 to 120 mPa·s) [35]. As a result, the H⁺ in Oxa-Ch transport is more favorable than that or Zn²⁺ in ZnCl₂-based DES.

To further diagnose the hydrolysis mechanism, we prepared the DES of ZnCl₂ and acetic acid (namely, ZnCl₂-Ac) with the molar ratio (1:2) [36] and then performed the reaction with MCC at 90 °C for 16 h. We also obtained cellulose nanocrystals with dimensions similar to other ZnCl₂-based DES as presented in Figure 6A, and more importantly the CNCs were functionalized with the -(C=O)-CH₃ group as depicted at the vibrational mode at 1720 cm⁻¹ in Figure 6B [37,38]. Our results are also consistent with the report made by Tamaddon et al. which proved that the DES of ZnCl₂ and acetic acid can be used for acetylation of alcohols, especially glucose [36].

Based on our results, we proposed the mechanism of the cellulose hydrolysis by ZnCl₂-based DES as follows. First, the Zn²⁺ (or H⁺) in DES diffuses into the network of microcrystal cellulose and activates the oxygen atom at the β-(1–4) glycosidic ether links of the amorphous regime, and then H₂O moisture existing within the cellulose framework or another component such as acetic acid supports the leverage of the β-(1–4) glycosidic ether links, which can break down the microcrystalline cellulose into smaller particles with a functional group, and eventually form rod-like shape nanoparticles. However, further experiments and also chemical quantum calculation are needed to provide insights into the mechanism of cellulose hydrolysis.

5. Conclusions

Here, we demonstrated a direct method to produce cellulose nanocrystals with a rod-like shape from microcrystalline cellulose by ZnCl₂-based DES without mechanical posttreatment. We obtained CNCs, crystalline index (68.9%), which were highly thermally stable with a width of ~50 nm and a length of 200–400 nm. Importantly, we were able to functionalize the CNCs with an acetyl group, which could potentially modulate the hydrophobic or hydrophilic property of the CNCs. We attributed the formation of the CNCs to the Lewis acid effect of ZnCl₂, which can hydrolyze the β-(1–4) glycosidic ether links of the amorphous cellulose regime. Our study opens a new path to directly isolate cellulose nanocrystals with functional groups on the surface of CNCs.