1. Introduction
Cellulose nanofibrils (CNFs) are high-aspect ratio nanoparticles formed by bundles of cellulose chains that are a succession of glucose subunits linked by β-1-4 glycosidic bonds. CNFs are produced from a cellulosic raw material, usually wood, the most abundant and renewable polymer available on earth. CNFs are produced by a combination of chemical/enzymatic pretreatments and mechanical treatment, usually using a homogeniser [
1], a microfluidiser [
2], or a grinder [
3]. The variety of existing pretreatments [
4] allows for a variety of surface chemistries, making CNF materials suitable for many applications [
5]. CNFs pretreated in the presence of (2,2,6,6-tetramethyl-piperidin-1-yl)oxyl, also known as TEMPO, proposed by Saito et al., 2006, which consists of the regioselective oxidation of C6 primary hydroxyls of cellulose to C6 carboxylate groups, have been considered in a wide variety of applications due to its carboxyl content and reduced size [
6,
7]. CNFs are generally used in two different forms: either as films/nanopapers or as gels and can be used as rheology modifiers or emulsion stabilisers and additives in many applications. Films are obtained by solvent casting [
8,
9,
10], and nanopapers are obtained by filtration [
11,
12]. Three types of gels can be identified: hydrogels, cryogels obtained by freeze-drying, and aerogels obtained by supercritical drying [
13,
14]. For many applications, water sorption properties are important. The TEMPO-oxidised cellulose nanofibrils (toCNF), like cellulose, are hygroscopic materials, which means that they can attract and retain water molecules from their environment by absorption or adsorption [
10]. The impact of process parameters on cryogels mechanical properties have been previously studied [
14,
15,
16,
17,
18,
19], highlighting the importance of density and preparation method on the mechanical properties.
As a natural, biodegradable, and abundant polymer with reactive surface chemistry and good biocompatibility, nanocellulose is a promising material within the medical field. In recent years, applications in wound healing [
20,
21], drug delivery [
9], and tissue engineering [
22] have been investigated. In tissue engineering, the scaffold should stimulate cells to differentiate, proliferate, and form tissue. The interplay between the matrix and cells should be driven by the action of signals, which can be a mechanical stimulation, chemical compounds, or growth factors (usually proteins) [
23]. To be suited for tissue engineering applications, scaffolds needs to exhibit stiffness similar to the natural extra-cellular matrix (ECM) of the tissue to be repaired and measured by the elastic modulus E. Typical values of stiffness of ECM are 0.1–1 KPa for brain tissue, 8–17 KPa for muscle tissue, and 25–40 KPa for the cross-linked collagen matrix [
24,
25]. Another crucial aspect for tissue engineering application is the scaffold architecture. A high porosity is needed to promote the cellular penetration and an adequate diffusion of nutrients to the cells [
26].
The utilisation of wood-based CNFs for tissue-engineering applications is encouraged by recent studies that confirmed the safety of CNFs [
27,
28,
29,
30], the construction of cell-friendly porous structures [
14], and the control of mechanical properties [
31,
32]. In addition, CNFs, in the form of a highly entangled network, have shown the ability to retain the active principle ingredient up to several months [
9]. However, major challenges for biomedical applications are yet to be overcome, such as increasing the bioavailability of drugs, as most new drugs are described as poorly soluble [
33], and controlling the delivery kinetics of active principle ingredient (API). To address these issues, this research study proposes the use of cyclodextrins (CD).
Cyclodextrins are cyclic oligosaccharides consisting of glucose subunits linked by α-1-4 glycosidic bonds. Due to their conformation, with a hydrophobic interior and a hydrophilic exterior, these macromolecules exhibit cage-like properties and can form an inclusion complex with hydrophobic compounds [
34,
35]. These properties have led to their use in various fields, such as cosmetics, food, environment, and medicine [
36,
37,
38]. Regarded as safe, they are widely used as an excipient in the pharmaceutical field [
39,
40]. For such applications, β-cyclodextrin (βCD), a cyclodextrin with seven glucose subunits, and its derivatives are the most commonly used [
33,
41]. βCD are also of a great interest for tissue engineering applications, with their properties to encapsulate lipophilic compounds proven to improve the performance of scaffolds [
42,
43,
44]. This property could also lead to the immobilisation of the growth factor [
45] or drug delivery [
39] during the cell growth to optimise the effect of the scaffold. Previous studies reported the association of cyclodextrin with various cellulose derivatives [
46,
47]. The association between CDs and CNFs or cellulose nanocrystals (CNCs) has been attempted in a very few and recent studies, summarised in
Table 1. To the best of our knowledge, no study presents the impact of βCD on both the sorption and the mechanical properties of toCNF substrates (films or cryogels).
The aim of the present study is to modify toCNF with βCD (preferably with covalent linkage) and to see what kind of effect this surface functionalisation has on the sorption and mechanical properties. Thus, a comparison with the same structures using unmodified toCNF is necessary. For that purpose, two suspensions of toCNF with different charge contents were prepared. Fibre modification with cyclodextrins was carried out at neutral and acidic (pH 2.5) to study the effect of the dissociation of the carboxylic acid group. Films, processed by casting/evaporation at 40 °C and cryogels, processed by freeze-drying were prepared from βCD-modified toCNF and compared with reference samples of unmodified toCNF. Water sorption was evaluated gravimetrically for both films and cryogels. The impact of density on the mechanical properties of the cryogels was assessed for cryogels obtained from unmodified toCNF and prepared by freeze-drying from suspensions at different dry matter contents for both charge contents. Compression tests in the dry and swollen state were performed on cryogels from all suspensions, and microscopic observation (SEM) was carried out to link the mechanical behaviour to the macroscopic structure of the materials.
3. Materials and Methods
3.1. Materials
A mixture of bleached and never-dried spruce (picea abies ca. 75%) and pine (pinus sylvestries, ca. 25%) cellulose pulp from Södra (Växjö, Sweden) was used as raw material. All chemicals used in this study were of laboratory-grade quality purchased from Sigma-Aldrich, St. Louis, MO, USA.
3.2. Preparation of toCNF With Two Different Charge Contents
TEMPO-oxidised cellulose nanofibrils were produced according to a protocol adapted from [
6]. Never-dried cellulose (110 g of cellulose content) was suspended in water (3 L) and stored overnight at 4 °C. The suspension was dispersed with a blender and mixed with a solution (400 mL) containing TEMPO (1.375 g) and sodium bromide (13.75 g). Water was added to obtain a total volume of 8250 mL (75-mL/g cellulose). TEMPO-mediated oxidation of cellulose was started by adding different amounts of 13% NaClO: 2.5 mmol/g cellulose for a charge content of 750 µmol/g and 3.3 mmol/g cellulose for a charge content of 1100 µmol/g. NaClO were added gradually, and the pH was maintained at 10.5 by adding 0.5M NaOH. The slurry was stirred for 15 min after the complete addition of NaClO, and the pH was then dropped to 7 with 0.1-M HCl. Methanol (100 mL) was then added to the slurry. The product was thoroughly washed with water by filtration until the conductivity of the filtrate was below 5 µS/cm. Homogenisation was conducted using a Rannie 15 type 12.56 × homogeniser (APV, SPX Flow Technology, Silkeborg, Denmark). The suspension was diluted to 1.2 wt% and dispersed with an electric mixer. The fibres underwent two passes in the homogeniser at 600 bar and 1000 bar, respectively. The final suspensions were stored at 4 °C.
3.3. Determination of the Charge Content
The carboxyl group content was determined by conductometric titration as described in previous studies (e.g., [
7,
32,
65]). NaCl (5 mL 0.1M) was added to a toCNF dispersion with 0.2-g solid content in 450 mL. The pH was adjusted to approximately 2.5 by addition of 0.1-M HCl and further diluted with water to a total volume of 500 mL. The dilution was titrated with 0.05-M NaOH solution added at a rate of 0.15 mL/min under stirring up to a pH of 11. An automatic titrator (902 Titrando, Methrom AG, Herisau, Switzerland) was used, and the conductivity of the sample was automatically measured (856 Conductivity Module, Methrom AG, Herisau, Switzerland) for increments of 0.02 mL. Data were recorded by Tiamo Titration software. The carboxyl content was calculated from the titration curve using the Gran plot. Duplicates were made for both suspensions and NaOH titration (control).
3.4. Material Processing
3.4.1. Film Processing
Dry toCNF (0.25 g) was weighed and diluted with water to a total volume of 50 mL. The suspension was dispersed for 2 min at 7000 rpm with an UltraTurrax (IKA-Werke, Staufen, Germany), at room temperature. βCD (0.025 g) and 0.1-M HCl (2 mL) were added to the relevant samples. The suspensions were magnetically stirred for 1 h and placed in an ultra-sonic bath for 3 min. The suspension was then cast in petri dishes (9-cm diameter) and stored in an oven at 40 °C for 18 h. The resulting films were stored in closed petri dishes at room temperature.
3.4.2. Cryogel Processing
Impact of pH, cyclodextrin, and comparison dry/swollen: Fifty millilitres of 0.8 wt% toCNF suspensions were prepared and dispersed 2 min at 7000 rpm with an UltraTurrax. βCD (0.04 g) and 0.1-M HCl (3 mL) were added if required. The suspensions were magnetically stirred for 1 h and placed in an ultra-sonic bath for 3 min. The suspensions were poured into a 24-well plate (3 mL per well) and freeze-dried for 24 h at −20 °C and 0.3 mbar (BK FD12S, Biobase Biodustry, Jinan, China). The resulting cryogels were stored in closed well plates.
Impact of density: Fifty millilitres of L-toCNF and H-toCNF suspensions at 1 wt%, 0.8 wt%, 0.6 w%, and 0.4 wt% were prepared and dispersed for 2 min at 7000 rpm with an UltraTurrax. The suspensions were magnetically stirred for 1 h and placed in an ultra-sonic bath for 3 min. The suspensions were then poured into a 24-well plate (3mL per well) and put in a freezer at −20 °C for 24 h before freeze-drying (ALPHA 2-4 LDplus, Christ ®, Osterode am Harz, Germany).
3.5. Water Sorption Analysis
Water sorption tests on films were carried out gravimetrically in a Percival climatic chamber at 25 °C and 90% RH (relative humidity). The samples were weighted every hour at the beginning of the experiment and at selected times thereafter. Water sorption experiments were conducted after 48 h, with at least 3 replicates for each sample. The samples were put in a desiccator for 16 h prior to the experiment. The water sorption was characterised by the weight change between the initial sample weight (m
0) and the weight after a certain time t (m
t), according to Equation (1):
Water sorption test on cryogels were conducted gravimetrically. The cryogels were weighted and immersed in distilled water and then removed at different times. Excess water was removed before weighting. Water sorption was calculated using Equation (1).
3.6. Microscopy
Atomic force microscopy images were recorded on a Dimension icon® (Bruker, Billerica, MA, USA). The concentration of the suspension was adjusted to 10−3 wt% by diluting the CNF dispersion using the high shear mixer Ultra-Turrax. A drop of this suspension was deposited on a freshly cleaved mica plate before drying overnight under a fume hood at room temperature. The acquisition was performed in tapping mode using a silica-coated cantilever (OTESPA® 300 kHz-42 N/m, Bruker, Billerica, MA, USA). Zones of 1.1*1.1 μm2 were analysed.
Scanning electron microscopy images were performed with ESEM (Quanta 200, FEI, Japan). Film and cryogel cross-sections were cut with a razor blade. SEM observation was carried out on cross-sections after carbon sputter coating of 5 nm, with a tension of 10 kV and a spot size of 3.5. The working distance was set between 9.5 mm and 11.5 mm depending on the sample.
For both microscopy techniques, at least 5 different images were performed to check the consistency in various zones of the sample, and the most representatives were selected for the discussion.
3.7. Mechanical Characterisation
Impact of density: Compression tests were performed using a TA Instruments RSA 3 (New Castle, DE, USA) dynamic mechanical analyser fitted with a 100-N load cell. Samples prepared as cylinders were individually measured and compressed with a crosshead speed of 0.1 mm/s at room temperature. At the least, triplicates were performed, and the average is presented.
Impact of pH, cyclodextrin, and comparison dry/swollen: Compression tests were performed with a Stable Micro Systems TA-XT2 texture-analyser (Stable Micro Systems, Godalming, UK), equipped with a P/35 probe and with a crosshead speed of 0.1 mm/s, as previously described by Heggset et al., 2018 [
66].
A minimum of 6 cryogels were tested for each sample. The compression modulus was calculated in the elastic region at half the strain of the beginning of the plateau region, and the stress at 70% strain was directly read from the data. The normalised compression modulus was calculated by dividing the compression modulus by the cryogel density. The cryogel density ρ was determined by dividing the mass of each cryogel by its volume. The volume of produced cryogels was measured from height and diameter measurements using a calliper. For each sample, the two extreme values were removed. The relative density of the cryogels was calculated from the ratio ρ/ρ
c, where ρ
c is the density of cellulose, 1.5 g/cm
3 [
67]. The porosity was calculated from Equation (2):
3.8. Fourier Transform Infrared Spectroscopy
Infrared spectra were recorded in attenuated total reflectance (ATR) mode, using a Perkin Elmer Spectrum 65 (Perkin Elmer, Wellesley, MS, USA). Spectra were recorded between 4000 and 600 cm−1, with 16 scans and a resolution of 4 cm−1. Since this technique is used to determine the possible esterification between the cyclodextrins and the toCNF, and given the proximity between the carboxylic peak and the ester peak (respectively, ≈1720 cm−1 and ≈1750 cm−1), each cryogel was dipped in 0.05 M NaOH for 10 s to convert carboxylic acid groups to carboxylate groups (1600 cm−1) and dried in the oven for 30 min prior to analysis. As the control sample, the neat samples and neat samples after 30 min drying in the oven were also analysed to ensure that esterification was only due to the freeze-drying process. At least 5 different zones of the sample were analysed, and the most representative spectra were used for discussion.