Dialdehyde cellulose crosslinked poly(vinyl alcohol) hydrogels: Influence of catalyst and crosslinker shelf life

Carbohydrate Polymers 198 (2018) 181–190

Contents lists available at ScienceDirect

Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol

Dialdehyde cellulose crosslinked poly(vinyl alcohol) hydrogels: Influence of catalyst and crosslinker shelf life Lukáš Münster, Jan Vícha, Jiří Klofáč, Milan Masař, Anna Hurajová, Ivo Kuřitka

T

⁎

Centre of Polymer Systems, University Institute, Tomas Bata University in Zlin, Tr. Tomase Bati 5678, 760 01 Zlín, Czech Republic

A R T I C LE I N FO

A B S T R A C T

Keywords: Poly(vinyl alcohol) Crosslinking Dialdehyde cellulose Aging Glutaraldehyde Network parameters

Dialdehyde cellulose (DAC) derived from α-cellulose by periodate oxidation was solubilized and utilized as a suitable crosslinking agent for poly(vinyl alcohol) (PVA). The crosslinking occurs between reactive aldehyde groups of DAC on the C2 and C3 carbons of anhydroglucose unit and hydroxyl groups on PVA backbone in the presence of acidic catalyst. Two catalyst systems based on diluted hydrochloric or sulfuric acid were tested. Their influence on the PVA/DAC network has been investigated by solid-state 13C NMR, XRD analysis and in the terms of network parameters and mechanical properties. Because DAC undergoes structural changes and decays with time, the role of DAC solution age (1, 14 and 28 days old) on material properties of formed PVA/DAC samples was studied as well. Outlined, even after 28 days after solution preparation, DAC exhibited the capability to act as an efficient crosslinker for PVA. The resulting material properties of PVA/DAC hydrogels were found to be dependent on the molecular weight of solubilized DAC closely related to its age and the choice of catalyst system. Furthermore, the DAC potential for PVA crosslinking was investigated in a broad concentration range. Besides, the DAC crosslinking efficiency was also compared to that of common crosslinking agent glutaraldehyde. The results showed different network topology of prepared hydrogels and exceptional crosslinking potential of DAC in comparison to glutaraldehyde, which is most likely related to DAC macromolecular character.

1. Introduction

Anseth, 2009), to introduce new functional groups for enhanced performance or to modify the biodegradability profile (Liu, Kost, Yan, & Spiro, 2012). In pharmaceutical sector, the hybrid biopolymer-based hydrogels find utilization as dressing materials containing active substance for wounds healing, drug delivery systems with variable release profiles dependent on external stimuli (pH, temperature), scaffolds for improved tissue regeneration, various body implants (i.e. cartilages) or as biosensors for specific molecules (Baker, Walsh, Schwartz, & Boyan, 2012; Carpi, 2011; Kamoun, Chen, Mohy Eldin, & Kenawy, 2015; Qiu & Park, 2001). One of the synthetic polymers suitable for the formation of hydrogels is poly(vinyl alcohol), PVA. The PVA-based hydrogels can be prepared using different routes, directly determining the nature of formed network and physico-chemical properties. These include (i) physical route of crosslinking by cyclical freezing and thawing or high temperature/energy treatment (Bolto, Tran, Hoang, & Xie, 2009) and (ii) chemical route utilizing crosslinking agents such as epichlorhydrine, various aldehydes (formaldehyde, glutaraldehyde, benzaldehyde etc.), anhydrides (EDTA dianhydride) or boric acid (Bolto et al., 2009; Lee & Mooney, 2001, p. 2; Miyazaki et al., 2010; Thomas et al., 2013). The

Cellulose is an abundant natural polymer, which exhibits valuable properties such as biocompatibility, good chemical modifiability and hydrophilicity beneficial for a plethora of applications in many different fields (Thomas et al., 2013). Standing alone, compounded with synthetic polymer or used in small concentrations as additives, they demonstrate water retention capability which is one of the key properties of hydrogels. The water retention capability of hydrogels arises due to the presence of three-dimensional network of crosslinked macromolecules containing hydrophilic groups (Wichterle & Lím, 1960). Crosslinked hydrogel matrix should contain (i) covalently bonded macromolecules of at least one polymeric substance, and/or (ii) physically crosslinked units constituted of macromolecular entanglements, (iii) hydrogen bonds or strong van der Waals interactions between chains, or (iv) at least two macromolecular chains joined in crystallites (Peppas, 2000). Biopolymer-based hydrogels are particularly favourable for applications in the field of biomedical sciences as they are able to mimic the structure of a living tissue (Chen & Hunt, 2007; Ma, 2008; Tibbitt &

⁎

Corresponding author. E-mail addresses: [email protected] (L. Münster), [email protected] (J. Vícha), [email protected] (J. Klofáč), [email protected] (M. Masař), [email protected] (A. Hurajová), [email protected] (I. Kuřitka). https://doi.org/10.1016/j.carbpol.2018.06.035 Received 22 December 2017; Received in revised form 24 May 2018; Accepted 2 June 2018 0144-8617/ © 2018 Elsevier Ltd. All rights reserved.

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

first approach has undeniable advantage in the absence of any kind of additive, although the crosslinking process has relatively poor definition in the exact terms of chemical reaction mechanism. The low molecular weight crosslinkers used in the second approach have typical set of features such as synthetic origin (except boric acid) and relatively high toxicity. The small size of molecules enables them to penetrate readily through various portals of entry into a living organism. Therefore, for medical applications, it is more desirable to utilize crosslinking agents derived from naturally available biopolymers, which have a potential to become lower-toxicity alternative to common toxic low-molecular weight crosslinkers. In this manner, Genipin (substance extractable from the fruits of Gardenia Jasminoides) was introduced as a low toxicity natural crosslinker (Bispo et al., 2010). However, limited availability and related high cost restricts its use on larger scale. This issue can be overcome by the choice of abundant, easily available and modifiable biopolymers as an original material for polymer analogue reaction imparting the product functional crosslinking side groups. One of the most accessible biopolymers is cellulose and its derivatives such as dialdehyde cellulose (DAC). In the recent study of Kim et al., solubilized DAC was utilized as a crosslinking agent for chitosan (Kim, Lee et al., 2017). Besides its crosslinking capability, the cytotoxicity of DAC was shown to be 6–12x lower than commonly used crosslinking agent glutaraldehyde. In this study, DAC prepared from α-cellulose by sodium periodate oxidation of hydroxyl groups on C2 and C3 to aldehydes (see Fig. S1) was investigated as a crosslinking agent for PVA. In order to utilize DAC in solution-based processes, it has to be solubilized using prolonged heating procedure (Kim, Wada, & Kuga, 2004). As has been shown in our recent study, solubilization step of DAC cannot be considered just as a simple dissolution as it leads to significant loss of molecular weight of final solubilized DAC product (Münster et al., 2017). Nevertheless, recent studies have shown potential of solubilized DAC towards the purpose of crosslinking (Kim, Lee et al., 2017; Kim, Kim, Choi, Kimura, & Wada, 2017). To obtain crosslinked PVA/DAC hybrid hydrogel, acidic catalyst must be added to initiate crosslinking reactions between aldehyde groups of DAC and hydroxyl groups of PVA. Two catalyst systems based on (i) hydrochloric acid or on (ii) sulfuric acid were introduced in this study and their effects on xerogel/hydrogel characteristics were investigated. These catalysts represent common systems used for crosslinking process employing reactions of aldehyde moiety (Jamnongkan, Wattanakornsiri, Wachirawongsakorn, & Kaewpirom, 2014; Kim, Lee, & Han, 1994; Saallah et al., 2016; Tang, Saquing, Harding, & Khan, 2010; Yeom & Lee, 1996). Moreover, the influence of aging of solubilized DAC (1, 14 and 28 days old) on properties of prepared xerogels and hydrogels was studied. The investigation of DAC solution shelf life is particularly important, because DAC in solution is known to be highly reactive and to undergo aging rapidly in terms of reactive aldehyde content loss and progressive molecular weight decrease (Kim, Kuga, Wada, Okano, & Kondo, 2000, 2004), and it is normally recommended to use always freshly made solutions (Sirviö, Liimatainen, Visanko, & Niinimäki, 2014). However, according to our recent work dedicated to characterization of the DAC solution composition in time (Münster et al., 2017), considerable stabilization of aldehyde content for extended period of time was observed, when DAC solutions were kept at low pH (3.0–3.5). It was shown that low pH reduces the degradation processes and slows down the reaction kinetics of β-elimination. Possible applicability of older (14+ days old) DAC solutions as crosslinking agents was also suggested. Therefore, crosslinking capabilities of the acidic DAC solutions of different age combined with the two catalyst systems were investigated in current study of PVA/DAC blends. Besides the evaluation of crosslinking capability of DAC during its aging, it is important to compare the crosslinking potential of this novel crosslinker to the most commonly used crosslinking agent. Thus, comparative study focusing on crosslinking of PVA by DAC or by glutaraldehyde including use of different concentrations of crosslinking agents was conducted.

2. Experimental 2.1. Materials Mowiflex TC 232 (Kuraray Specialities Europe GmbH) was used as source of poly(vinyl alcohol) (PVA), density ρp = 1.3 g/cm3, number average molecular weight Mn = 23 500 g/mol (estimated by GPC). Dialdehyde cellulose (DAC) was prepared by periodate oxidation of αcellulose (Sigma Aldrich Co.) by sodium periodate (NaIO4) (PENTA, Czech Republic). Oxidation reaction was terminated by addition of ethylene glycol (PENTA Czech Republic). The degree of oxidation was estimated by conversion of DAC to oxime by reaction with hydroxylamine hydrochloride (Sigma Aldrich Co.) and subsequent consumption sodium hydroxide (PENTA, Czech Republic). All chemicals used in crosslinking reaction including 35% hydrochloric acid (HCl), 96% sulfuric acid (H2SO4), 99.8% methanol (CH3OH) (PENTA, Czech Republic), 99.8% acetic acid (CH3COOH) and 50% water solution of glutaraldehyde (Sigma Aldrich Co.), were of analytical purity (p.a.) and used as received without further purification. Demineralized water was used throughout the experiment. 2.2. Preparation of crosslinked PVA/DAC and PVA/GA hydrogels In the first step, periodate oxidation of α-cellulose was carried out. Detailed preparation and properties DAC solutions and dried samples are presented elsewhere (Münster et al., 2017). In brief, 10 g of α-cellulose was suspended in 250 mL of water containing 16.5 g of NaIO4 and stirred in dark for 72 h. Oxidation was stopped by adding of excess of ethylene glycol. Never-dried product was flushed on filter and then solubilized at 80 °C for 7 h under reflux, purified by centrifugation and filtration to remove residual solids and diluted to 200 mL with water. The pH of the solution was kept between 3.2 ± 0.1 all the time. Solubilized DAC was analysed in the terms of reactive aldehyde group content over 28 days. This was achieved by the oxime reaction of prepared solubilized DAC of different age with hydroxylamine hydrochloride. Typically, sample solution containing 0.1 g of solubilized DAC was diluted to fixed volume by demineralized water and pH was set to 4.00 using 0.1 M HCl. Next, 20 ml of solution containing 0.43 g of hydroxylamine hydrochloride was adjusted to pH 4.00 by 0.1 M NaOH. These two solutions were mixed together and stirred for 24 h at 150 RPM in closed containers. The reactive aldehyde group content was determined by the consumption of 0.1 N NaOH (Kim et al., 2004; Veelaert, de Wit, Gotlieb, & Verhé, 1997). The second step involved preparation of PVA crosslinked samples using DAC of various age. PVA/DAC mixtures were prepared by dissolution of 4.95 g PVA in 70 mL of water and addition of defined volume of fresh (1 day old) DAC solution containing 0.05 g (1 wt%) of solubilized DAC. To form a crosslinked network, acidic catalyst was introduced. Two acidic catalyst systems were chosen giving rise to two types of PVA/DAC blends (referred as type or series A and B in further text). In type A, 1.5 mL of 1.33 M HCl solution served as the catalyst and pH buffer. In B type, 0.25 mL of 10%vol H2SO4 was used as the catalyst, together with 0.75 mL of 10 vol% solution of CH3COOH (pH buffer) and 0.5 mL of 10 vol% CH3OH (quencher). Whole process was repeated using two- and four-weeks old DAC solution. In the comparative crosslinking study, fresh DAC and conventional crosslinking agent glutaraldehyde (GA) was utilized for PVA crosslinking in the set of chosen concentrations. The concentration range of used DAC crosslinker was from 0.0625 to 5 wt%. For example, PVA/ DAC blend crosslinked using 1 wt% DAC contained 0.05 g of solubilized DAC. Since the fresh DAC was estimated to have content of reactive aldehyde groups 11.7 ± 0.3 mmol/g, the particular amount of reactive aldehyde groups used for crosslinking of this sample was 585 ± 13 μmol. To achieve comparable crosslinking of PVA, the volume of GA crosslinking solution containing equal amount of reactive aldehyde groups per sample was used. Content of reactive aldehyde 182

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

Table 1 Designation of PVA/DAC samples prepared using aged DAC (upper part). Designation of PVA/ DAC and PVA/GA samples prepared within comparative crosslinking study along with amount of reactive aldehyde group content per sample (lower part).

PVA crosslinking via aged DAC DAC age

1 day

14 days

28 day

PVA/DAC blend A

A01

A02

A03

PVA/DAC blend B

B01

B02

B03

Comparative crosslinking study DAC crosslinker n-CHO per sample (ȝmol)

DAC (wt%)

PVA/DAC

GA (ȝL)

PVA/GA

2924.3

5

DTC-A

233

GTC-A

1754.6

3

DTC-B

139.8

GTC-B

877.3

1.5

DTC-C

69.9

GTC-C

584.9

1

DTC-D

46.6

GTC-D

146.2

0.25

DTC-E

11.6

GTC-E

73.1

0.125

DTC-F

5.8

GTC-F

36.6

0.0625

DTC-G

2.9

GTC-G

500.13 MHz (1H) and 125.77 MHz (13C), equipped with 4 mm solidstate dual (BB-1 H) CP/MAS probe. Rotor was spun at 10 kHz and the relaxation delay was set to 4 s. The 13C NMR chemical shifts were referenced to crystalline α-glycine as a secondary reference (δst = 176.03 ppm for carbonyl carbon) (Bouzková, Babinský, Novosadová, & Marek, 2013). Spectra of PVA and PVA/DAC samples were measured with contact time of 1 ms, which provided best signal-to-noise ratio, consistently with previous studies (Capitani, De Angelis, Crescenzi, Masci, & Segre, 2001). The proton spin-lattice relaxation time in rotating frame, T1ρ(1H), in PVA and PVA/DAC samples was determined from exponential fit of carbon signal decay as a function of contact time (from 1 to 8 ms). The carbon spin-lattice relaxation time in rotating frame, T1ρ(13C), was obtained from exponential fit of the carbon signal decay as a function of the variable 13C spin-lock time (0.5–7 ms) following the cross-polarization time (set to 1 ms in all cases) (Kolodziejski & Klinowski, 2002; Voelkel, 1988).

groups of GA solution (11.1 ± 0.5 mmol/g) was experimentally estimated by oxime reaction as described above. Within this comparative study, only catalyst system type A (1.5 mL of 1.33 M HCl per sample) was used for the preparation of PVA/DAC and PVA/GA blends. All PVA/DAC and PVA/GA blends were cast on 20 × 10 cm frame, dried in laboratory oven at 30 °C and placed in desiccator. Samples were subsequently thoroughly washed to remove uncrosslinked material components and dried at 30 °C until constant weight. Such samples along with the input materials were analysed by methods described below. PVA/DAC and PVA/GA samples from the comparative crosslinking study were analysed only in the terms of their network parameters (see Section 2.6). Designation of the prepared PVA/DAC and PVA/GA samples is listed in the Table 1. 2.3. Solid-state

GA crosslinker

13

C NMR

Powdered samples suitable for solid-state 13C NMR analysis were prepared from PVA/DAC xerogels using ultra-centrifugal mill Retch ZM 200 equipped with 0.12 mm ring sieve. Samples were initially nitrogencooled to prevent heat degradation and milled at 10,000 rpm to obtain finely ground powder. After milling procedure, the powdered samples were further dried and kept in desiccator. The CP-MAS 13C NMR spectra were recorded at room temperature on Bruker Avance 500 MHz spectrometer operating at frequencies

2.4. XRD analysis X-ray diffraction analysis of PVA/DAC xerogels prepared using aged DAC and pure PVA, α-cellulose and DAC were performed on a Rigaku MiniFlex600 X-ray diffractometer in the diffraction 2θ angle range 5–95°, CoKα (λ = 1.789 Å) radiation source was used with Kβ line filter.

183

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

Gel fraction (%) =

M0 × 100 Minit

(3)

• Average molecular weight between crosslinks (M ), where M

n is the c number average of molecular weight of initial uncrosslinked polymer, ν is the specific volume of polymer, V1 is the molar volume of water, V2,s is the polymer volume fraction, and χ1 is the polymersolvent interaction parameter (Flory & Rehner, 1943):

1 2 = − Mc Mn

( ) [ln(1 − V ν V1

2, s )

+ V2, s + χ1 (V2, s )2]

1

⎡ (V2, s ) 3 − ⎣

V2, s 2

⎤ ⎦

(4)

• Polymer volume fraction (V

2,s), where MS/M0 is the weight swelling ratio of swollen hydrogel at equilibrium, ρp is the polymer density, and ρw is the density of water (Peppas, Huang, Torres-Lugo, Ward, & Zhang, 2000):

−1 ρp Ms ⎛ − 1⎞ ⎤ V2, s = ⎡1 + ⎢ ρw ⎝ M0 ⎠⎥ ⎦ ⎣ ⎜

Fig. 1. The CP-MAS 13C NMR spectra of PVA and selected PVA/DAC xerogels (left), comparison of PVA/DAC xerogels 13C NMR spectra of samples prepared using various catalyst systems (see Table 1) and acidic DAC solution (pH 3.2 ± 0.1) of different age (1, 14, 28 days).

⎟

(5)

• Crosslink density (ρ ) can be calculated as (Flory & Rehner, 1943; c

Omidian, Hasherni, Askari, & Nafisi, 1994):

ρc =

1 νMc

(6)

2.5. Tensile measurements 3. Results and discussion Tensile measurements were conducted on PVA/DAC xerogels prepared using aged DAC according to the EN ISO 37 (621436) on tensile testing machine M350 5CT (Labor Chemie, Czech Republic). A typical specimen of length 35 mm tempered at 25 °C and 40% humidity was strained at initial rate of 1 mm/min to 7.5% strain for Young’s modulus evaluation. The testing continued at the strain rate of 50 mm/min.

The crosslinking capability of fresh and aged DAC towards PVA was initially studied to prove applicability of aged DAC for this purpose. The suitability of defined catalyst system was evaluated. Subsequently, comparative crosslinking study between DAC and GA over the broad range of crosslinker concentrations was performed.

2.6. Network parameters

3.1. Structural analysis of PVA/DAC films

In order to assess network parameters, disc samples (d = 8 mm) were cut from dried unwashed crosslinked PVA/DAC and PVA/GA xerogels. These were washed and left to swell completely (1 week) and subsequently gently dried at 30 °C until constant weight. Weight of these samples was recorded before and after each step using high precision balance. The average molecular weight between crosslinks (Mc ) and crosslink density (ρc) were calculated with the aid of equilibrium swelling theory suggested by Flory and Rehner (1943). Following equations were used to calculate network parameters:

3.1.1. Solid-state NMR Cross-polarization/magic-angle-spinning (CP-MAS) 13C NMR spectroscopy was used to study the degree of crosslinking in washed PVA/ DAC samples prepared using fresh and aged DAC (Lai et al., 2003). Formation of PVA/DAC xerogels produced easily recognizable changes in PVA signals, see Fig. 1 and Table S1. The signals of DAC itself (located between 110–80 ppm for C1–C3 and 80–55 ppm for C4–C6) (Varma, Chavan, Rajmohanan, & Ganapathy, 1997) were very weak and partially obscured by PVA resonances. This is not surprising given the PVA:DAC ratio (99:1) in PVA/DAC samples and the expected variety in possible crosslinking modes due to complex composition of DAC in solution (Münster et al., 2017). As a result, only PVA signals in pure PVA and PVA/DAC samples are discussed in following. Assignment of 13C signals in pure PVA is based on published data (Lai et al., 2002; Terao, Maeda, & Saika, 1983), see Fig. 1 and Table S1. Because the presence of intramolecular hydrogen bonds is known to cause deshielding of 13C NMR resonances in PVA spectra (Lai et al., 2003; Terao et al., 1983), signal of methine group at 77.1 ppm (noted as I in Fig. 1) is attributed to the mm (isotactic) triad with two intramolecular hydrogen bonds, signal at 71.5 ppm (II in Fig. 1) to the mm and mr (heterotactic) triads with single intramolecular hydrogen bond and signal at 65.4 ppm (III in Fig. 1) to mm, mr and rr (syndiotactic) triads without intramolecular hydrogen bonds. Only one signal of CH2 group at 45.3 ppm is present in all spectra. Signal at 22.5 ppm belong to residual acetate groups in PVA chains (corresponding carbonyl resonance is situated at 173.0 ppm) and it is not further discussed.

(i) Percentage of swelling, where Mt is the weight of swollen polymer at time t, and M0 is the initial weight of washed and dry hydrogel (Gulrez, Al-Assaf, & Phillips, 2011):

Percentage of swelling (%) =

Mt − M0 × 100 M0

(1)

• The equilibrium water content (EWC), describing maximum amount

of water absorbed by hydrogel. Ms is the weight of sample swelled at equilibrium conditions and M0 is the weight of dry washed hydrogel (Tighe, 1986):

EWC (%) =

Ms − M0 × 100 Ms

(2)

• Gel fraction, where M

0 refers to the weight of dried hydrogel after extraction of soluble fraction in hydrogel and Minit is the weight of dry and unwashed hydrogel (Gulrez et al., 2011):

184

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

The T1ρ(13C) in B03 are even shorter than in pure PVA (4.4 ± 0.2 ms vs. 7.8 ± 0.5 ms for B03 and PVA respectively), which mean high restriction of local molecular motion and loss of chain flexibility. Because the hydrogen bond network in B03 is clearly not restored to any significant level (T1ρ(1H) values are constant, see Table S2), we assume that the restriction of molecular motion is caused by significant increase of crosslink density. This implies actual increase of reactivity of aged pH-stabilized DAC solution for crosslinking reactions with the use of sulfuric acid as catalyst, instead of expected decrease (accompanied by aldehyde content loss) previously reported for pH neutral DAC solutions (Kim et al., 2004; Sirviö et al., 2014). The explanation can be found in our previous work (Münster et al., 2017), where we investigated the composition of DAC solution in time. Although the total aldehyde content of pH-stabilized DAC solution remained almost intact during the observed time (equivalent to a decrease of the degree of oxidation by 6 percentage points in 28 days within this study, see left part of Fig. 2), the composition of the solution was changing as reflected in Mn of DAC. In brief, we have detected two competing processes in the acidic DAC solution. Initially (0–14 days), the formation of intermolecular hemiacetals connecting individual DAC chains takes place and large (about 20%) increase in Mn is observed. Subsequently, the second process, splitting of DAC chains into smaller molecular weight fragments, becomes dominant. This is reflected in decrease in Mn of DAC after 28 days (almost by 40% compared to value after 14 days) resulting in Mn even below that of fresh solution, see Fig. 2 (right). The changes of Mn of DAC solution in time correlate with T1ρ(13C) results (the right part of Fig. 2) implying the direct influence of changes in DAC Mn on resulting hydrogel structure and properties. The consequence of these changes on macromolecular level is more pronounced than decrease of DAC reactive aldehyde group content in time. Likely, smaller number of larger and less mobile fragments present after 14 days do not crosslink PVA chains as efficiently and densely as larger number of smaller fragments present in 28 days old solution. Therefore, the flexibility of PVA chains reaches maximum for sample prepared using 14 days old DAC solution (maximum Mn ), while the most rigid structure is obtained using 28 days old DAC (lowest Mn ). The structural conclusions derived from 13C NMR study of PVA/ DAC samples correlate well with crystallinity obtained from XDR analysis, and are further reflected in network parameters and mechanical properties of PVA/DAC xerogels, namely Young’s modulus. XRD analysis. Examples of diffractograms recorded for input materials and PVA/DAC samples from B are shown in Fig. S2. Detailed results from XRD analysis are summarized in Table S3. As can be seen in Fig. S2 (left), prepared DAC showed significant loss of crystallinity compared to original α-cellulose. This is a direct result of the opening of

The use of different reaction catalyst has minimal influence on the CP-MAS 13C NMR spectra of PVA/DAC xerogels. There are, however, significant changes between the PVA/DAC xerogel spectra and those of pure PVA. In PVA/DAC spectra, signal I is no longer observable, while II is shifted upfield from 71.5 ppm to 69.7 ppm. As a result, the signal II almost completely overlaps with somewhat deshielded signal III found between 66.6–67.2 ppm, depending on the DAC age, see Table S1. Observed changes imply considerable disruption of hydrogen bond network in PVA/DAC xerogels, which is consistent with formation of PVA-DAC crosslinks (larger PVA inter-chain separation due to bulky crosslinking agent is likely to disrupt nearby hydrogen bonds). To obtain more information about PVA/DAC structure and properties, rather qualitative conclusions derived from changes in NMR chemical shifts were correlated with results from T1ρ(1H) and T1ρ(13C) studies, see Table S2. The T1ρ(1H) depends on density of nearby spin-reservoirs, with individual spins interacting via 1H/1H dipolar mechanism (Voelkel, 1988). Thus, T1ρ(1H) is not very sensitive to motion of individual molecular groups, but can be treated as a “phase” quantity, describing the efficiency of spin-diffusion processes in given range around observed nucleus, which is related to the number of neighbouring spins and their individual distances (Lai et al., 2002, 2003). Considerable increase of T1ρ(1H) between PVA and PVA/DAC xerogels (see Table S2) agrees with conclusions derived from NMR chemical shift changes, i.e. the dilution of nearby 1H spin reservoirs, likely due to suspected disruption of hydrogen bond network and larger separation of PVA chains. Note, that T1ρ(1H) values for PVA/DAC xerogels prepared using various conditions are comparable mostly within experimental error, suggesting similar spin-diffusion efficiency in all samples and thus approximately the same hydrogen bond network density. The only exception is A03 sample with somewhat shorter T1ρ(1H) (4.1 ± 0.2 ms), indicating increased density of nearby spins, which may be connected to partial restoration of hydrogen bond network. The T1ρ(13C), which allows to study local site-specific motions on kilohertz timescale (Garbow & Stark, 1990), was used to obtain information about local dynamics of polymeric chains in pure PVA and PVA/DAC samples, see Table S2. Observed increase of T1ρ(13C) between PVA and PVA/DAC samples reflects increased flexibility of polymeric domains in PVA/DAC as compared to PVA, again supporting the assumption about hydrogen bond network dilution during crosslinking. In contrast to T1ρ(1H) values, the T1ρ(13C) values more sensitively reflect sample preparation conditions. The T1ρ(13C) values are in general higher for samples prepared using 14 days old DAC in both A and B series. Further, the difference between A and B series is increasing considerably with age of DAC solution, see Table S2 and Fig. 2 (right).

Fig. 2. Decrease of reactive aldehyde group content with DAC aging (left), correlation between Mn of solubilized DAC (Münster et al., 2017) and T1ρ(13C) values of resulting PVA/DAC hydrogels in dependence on aging time (right). The full line in the left graph shows the weakly decreasing trend. The lines connecting points in the right graph are only guides for eyes.

185

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

crystallinity (XA02 = XB02 = 16%) in xerogel state. On the other side, it can be expected for the both A03 and B03 samples to be less crystalline due to the denser crosslinked structure induced by the oldest and thus most fractioned DAC crosslinking agent. Obtained data reflect this expectation, which is particularly manifested for the B03 sample as its crystallinity drops down to 6%. This goes hand in hand with the lowest T1ρ(13C) value implying the densest crosslinking of PVA/DAC sample. Similar, however less pronounced, trend is observed for A03 sample. Schematic representation of the two extreme structural situations in xerogels is given in Fig. 4. The determined structural characteristics of PVA/DAC samples are further associated and correlated with mechanical properties (Young’s modulus) and network parameters. 3.2. Mechanical properties and network parameters of PVA/DAC samples prepared by aged DAC 3.2.1. Tensile measurements Mechanical properties were investigated for untreated PVA and simultaneously for PVA samples crosslinked using fresh and aged DAC and different catalyst system. Because the samples of pure DAC were very brittle, it was impossible to measure their tensile characteristics. Examples of stress vs. strain curves of PVA and crosslinked PVA/ DAC xerogel are shown in the left part of Fig. S3. In must be noted, that the apparent second yield point is an artefact caused by the change of strain rate applied during tensile testing (see carefully the experimental section). The different mechanical behaviour of all PVA/DAC samples in comparison to pure PVA (decreased elongation, yield point and observable necking) implies presence of crosslinked network. Fig. S3 (right) shows the dependence of Young’s modulus (E) of PVA/DAC samples on DAC solution age and chosen catalyst system. Detailed results of mechanical properties from tensile tests are given in Table S4. The effect of crosslinked polymer network presence within all of the PVA/DAC xerogels can be seen in the increase of Young’s Modulus to values above 2 GPa, ten times higher compared to untreated PVA (see Table S4). Thus, formed stiff polymer network is of higher contribution to Young’s modulus opposed to the contribution of fairly amorphous hydrogen bond structure of untreated PVA. As can be seen in Fig. S3 (right) and Table S4, xerogel samples prepared using 14 day old DAC exhibit the lowest average values of Young’s Modulus. This is not surprising as these samples have the highest values of network flexibility according to NMR. This also correlates with the development in macromolecular size of used DAC mentioned earlier. PVA/DAC samples prepared using oldest DAC solution (28 days) exhibit higher Young’s modulus value, namely B03

Fig. 3. Correlation between crystallinity of prepared PVA/DAC materials prepared using DAC of various age and carbon spin-lattice relaxation time in rotating frame, T1ρ(13C), corresponding to polymeric chain segments flexibility as obtained from NMR study. The lines connecting points in the graphs are only guides for eyes.

glucopyranose rings (Fig. S1), which led to destruction of ordered macromolecular packing as reported earlier (Kim et al., 2004; Münster et al., 2017). The graph in the right part of Fig. S2 illustrates noticeable crystallinity changes exemplified on diffractograms recorded for samples from B series. The trends in crystallinity of both PVA/DAC series correspond with trends in polymeric domains flexibility determined by T1ρ(13C) values obtained from NMR study (see Fig. 3) with relation to Mn of DAC in solution. It must be noted also, that the crystalline phase is only a minor component within prevailing amorphous crosslinked network and has small contribution to the flexibility (or stiffness) of chains examined by NMR. The most flexible structure and thus presumably sparsely crosslinked PVA macromolecular chains in the both A02 and B02 samples (14 days old DAC of highest Mn ) enables to form larger crystalline sequences between PVA/DAC crosslinking spots, which leads to increased

Fig. 4. Schematic representation of the structure of the respective xerogels within PVA/DAC series B with their relation between flexibility, crystallinity and Mn of DAC. The size of orange coils corresponds to Mn of DAC, the arrays of blue lines represent polymer chain crystallites, and the light grey lines represent the amorphous phase of the polymer with physical entanglements. The size, ratio, and number of depicted features is intentionally exaggerated for better understanding. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

186

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

sample. The presence of smaller DAC macromolecules (Mn ∼2000 g/ mol) (Münster et al., 2017) positively affects stiffness via denser crosslink network. To summarize, tensile study (Young’s Modulus and peak elongation) indicates presence of crosslinked polymer network with stiffness of all crosslinked samples ranging from 2 to 2.5 GPa and reveals trends within prepared PVA/DAC series corresponding to NMR and XRD findings. 3.2.2. Network parameters Samples of PVA/DAC prepared using aged DAC were characterized in the terms of their network parameters described in the Section 2.6. Fig. 5 shows the dependence of age of used DAC on the percentage of swelling, equilibrium water content (EWC), gel fraction, average molecular weight between crosslinks (Mc ) and crosslink density (ρc) of PVA/DAC samples from both prepared series. Observed quantities are related in following manner. The denser the crosslinked polymer network (higher ρc) is, the larger is the amount of gel fraction. Next, if there is an increase in the swelling capacity, values of EWC and Mc rise while gel fraction and ρc values drop and vice versa. Thus, higher water uptake indicates higher swelling capacity because there is more space available for water between the longer macromolecular chains segments bonded in sparsely distributed crosslink network. The trends in the network parameters between the two series of PVA/DAC prepared using the two distinct catalysts are similar. However, samples within series A (catalyst HCl) show somewhat more uniform behaviour. The properties are less dependent on the age (molecular weight) of DAC when compared to the series B (catalyst H2SO4), where relatively steep change of properties is noticeable for the sample prepared using 28 day old DAC (B03). The swelling capacity of PVA/ DAC samples prepared with 1 wt% of aged DAC is ranging approximately from 400 to 450% with the exception of B03 sample. As a result of combination of oldest (the most fractioned) DAC and the catalyst system containing sulfuric acid, acetic acid, and methanol, this sample exhibits the lowest swelling capacity slightly below 370%, EWC of 78.6%, highest amount of gel fraction (68.9%) and the highest crosslink density (ρc = 0.5 mmol/g) of all samples prepared using aged DAC. 3.2.3. Correlation between DAC aging and characteristics obtained for samples crosslinked by aged DAC Indeed, the trends in network parameters of the samples prepared using DAC of different age are in good agreement with trends in properties analysed by NMR and XRD. In order to investigate the effect of DAC aging, the properties of crosslinked materials are compared with properties of aged DAC. Reactive aldehyde group content loss during aging is almost negligible and co-linear with aging time, while molecular weight distribution of DAC changes considerably with aging of DAC in solution (Münster et al., 2017), here represented by the number average molar mass (Mn ), for both see Fig. 2. Rearranging of data by plotting the dependence of measured values on the Mn of aged DAC in the two upper graphs in Fig. 6 reveals, that the crosslinking effect of DAC is closely correlated to its molecular weight. 3.2.4. Role of catalyst Correlations between various characteristics of PVA/DAC samples series A and B prepared using aged DAC are given in Fig. 6 to better illustrate the difference between effects of investigated catalytic systems also. Aging time and Mn as important parameters are indicated by labels in the middle and two lower graph panels. It can be clearly seen hat the samples within A series (HCl catalyst) exhibit almost linear dependence of all characteristic properties. For example, as the Mn of DAC grows, the polymeric domain flexibility T1ρ(13C) and crystallinity linearly increases while crosslink density (ρc) linearly decreases. Moreover, as shown in the bottom graph in Fig. 6 (dependence of T1ρ(13C) on ρc), both prepared PVA/DAC series show strongly linear correlation between two independent methods defining the network properties. Both dependencies have apparently the same slope but the

Fig. 5. Influence of DAC age used for crosslinking on the network parameters of prepared PVA/DAC hydrogels. The graphs show percentage of swelling, equilibrium water content (EWC), gel fraction, average molecular weight between crosslinks (Mc ) and crosslink density (ρc) of both prepared series of PVA/DAC samples crosslinked by fresh (1 day) and aged DAC (two and four weeks). The lines connecting points in the graphs are only guides for eyes.

187

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

shift of the series B down from the line belonging to the series A indicates difference in distribution of crosslinks in the network. Series A has higher T1ρ(13C) than series B at the same ρc which might be due to more uniform distribution of crosslinks in network of samples B than in samples A. Since the DAC used for preparation of the corresponding pairs od A and B samples was always taken the same stock solution, we hypothesize that the catalytic system containing sulfuric acid can be more sensitive to low molecular fragments of DAC, or may cause degradation of DAC, or it may influence the formation of physical crosslinking by entanglements. In other words, there are different mechanisms of crosslinking when using the two catalytic systems resulting into products with principally different networking. Moreover, some of the observed correlation trends for B series show non-linear behaviour. Among possible effects of the catalytic system containing sulfuric acid the DAC degradation would be the worst situation that has to be avoided. Even without a clear interpretation of this issue it can be concluded that the mixture of a volatile quencher, sulfuric and acetic acid is not a suitable catalyst for reported preparation method. Due to better overall coherence of results and simplicity of chemical composition, in the following comparative crosslinking study only HCl catalyst was employed. 3.3. Network parameters of PVA/DAC and PVA/GA – comparative crosslinking study In order to evaluate crosslinking effectivity and efficiency of DAC, network parameters of prepared PVA/DAC hydrogels were compared with corresponding hydrogels prepared using GA (PVA/GA) as it is one of the most common PVA crosslinking agents (Jamnongkan et al., 2014; Kim et al., 1994; Tang et al., 2010). The concentrations of fresh solubilized DAC were set relative to amount of PVA in range from 0.0625 to 5 wt%. Based on experimentally estimated content of reactive aldehyde groups of fresh DAC (11.7 ± 0.3 mmol/g), the chemical amounts of aldehyde groups per sample (n−CHO) were calculated (Table S5). DAC has been found to be effective crosslinking agent for PVA with chosen catalyst (HCl) in the whole range of used concentrations. Resulting network parameters of PVA/DAC samples are shown in Table S5. As expected, with the decreasing amount of DAC, the swelling capacity, EWC, Mc increases and gel fraction and ρc decreases. For instance, swelling increases from ∼100 to ∼6500%, while gel fraction decreases from ∼84 to ∼14% when DAC concentration is decreased from 5 to 0.0625 wt%, respectively. In other words, as the concentration of crosslinker decreases, the polymer network becomes sparsely crosslinked and the size of PVA macromolecules between crosslink grows larger, leaving more space for water uptake. Similarly, decreased concentration of crosslinking agent leaves more of free uncrosslinked PVA macromolecules in the polymer matrix, increasing the material weight loss during washing. Table S6 shows the calculated network parameters for PVA/GA samples prepared using volumes of GA crosslinking solution of equivalent amount of reactive aldehyde groups per sample as in the case of PVA/DAC samples. PVA/GA samples with codenames GTC-F and GTC-G (n−CHO = 73.1 and 36.6 μmol of−CHO crosslinker groups per sample respectively) are not mentioned in Table S5 as it was not possible to measure them (samples teared apart or even dissolved during the washing process). The poor crosslinking effectivity of GA in this lowest concentration range is the main difference in comparison to DAC under identical conditions. For the higher concentration of crosslinker, PVA/GA hydrogels exhibited similar trends in network parameters as PVA/DAC. Direct comparison of network parameters of DAC and GA crosslinked hydrogels is given in Fig. 7 and Tables S5 and S6. Samples A–E showed comparable swelling capacity regardless on crosslinker, with PVA/DAC showing slightly lower values in concentrations between 5–1.5 wt%. This suggests higher crosslinking efficiency of DAC in this concentration range, which is reflected in EWC, Mc and ρc values.

Fig. 6. Correlation between data measured by various methods for PVA/DAC prepared using aged DAC and different catalyst system. The lines connecting points in the graphs are only guides for eyes. The datapoints are labelled by corresponding age and number average molecular weight for one series (A) only in each graph. The second series (B) has the same order of datapoints from left to right always. 188

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

On the contrary, gel fraction is higher in PVA/GA samples compared to PVA/DAC. This is likely due to ability of low molecular crosslinkers such as GA to form a network with relatively homogenously distributed crosslinks (see Fig. 8, right part). In contrast to molecular compound, macromolecular DAC with crosslinking groups present only on the polymer chain generates principally different network topology (see Fig. 8, left part) composed of (i) regions with high local crosslink density due to large amount of reactive aldehyde groups on DAC macromolecules and (ii) chemically uncrosslinked regions comprised of sizable sections of free PVA chains, which are physically entangled only. Some of these chains link the regions (i) together. Relatively large separation of high-density crosslink regions results in higher percentage of uncrosslinked PVA chains, which causes somewhat higher material weight loss during washing of hydrogel. To summarize, samples prepared using higher concentrations of DAC and GA crosslinking agents showed relatively comparable network parameters. However, slightly increased crosslink density was observed for PVA/DAC samples in the span of used DAC concentration from 1.5 to 5 wt%. Better crosslinking efficiency likely connected to the macromolecular character of DAC enables its use at very low concentrations (0.0625 wt%, i.e. 36.6 μmol of−CHO groups per sample) compared to GA, allowing to prepare hydrogels with very high swelling capacity. 4. Conclusion It was found that the solubilized DAC stabilized by low pH of the solution retains its crosslinking capability towards PVA even after 28 days from preparation. Thus there is no need to prepare fresh DAC daily for potential industrial usage, if pH stabilized DAC solution is used. Furthermore, it was found that the properties of PVA/DAC hydrogels are governed by the number average molecular weight (Mn ) of solubilized DAC in acidic solution and by selected catalyst system rather than by the reactive aldehyde groups content, which does not change during DAC aging significantly. The changes of DAC Mn are directly reflected in the distinct trends found in number of measured and observed qualities such as polymeric domain flexibility expressed by T1ρ(13C), crosslink density (ρc), swelling capacity, equilibrium water content (EWC), average molecular weight between crosslinks (Mc ), amount of crystalline phase and average values of Young’s modulus. This implies interesting potential for “tuning” the properties of resulting hydrogel without need for any other additives/catalysts. The DAC with required Mn can be cost-effectively obtained for example by optimization of the DAC solubilization conditions. The trends in hydrogel characteristics and properties remain similar regardless on the choice of catalyst system. The largest difference between catalysts was observed for samples prepared using the oldest (most fractioned) solubilized DAC. In this case, sulfuric acid seems to act as a more efficient catalyst, causing highest measured degree of crosslinking, while the use of hydrochloric acid results into roughly similar characteristics as before. Thus, besides other advantages, the utilization of hydrochloric acid within crosslinking reactions between DAC and PVA produces hydrogels of comparable qualities during observed timescale. Unlike GA, DAC is capable to form stable crosslinked polymer networks even at very low concentrations. Such hydrogels exhibit very high swelling capacity (about 6 500%). The explanation to this can be found in the macromolecular character of DAC as it forms highly crosslinked regions adjacent to each polymer coil bearing reactive groups. These regions are embedded in a matrix formed by PVA macromolecules. Some of these PVA chains link the highly crosslinked regions together preserving the hydrogel integrity. This topology is also responsible for specific properties of PVA/DAC hydrogels with respect to the PVA/GA hydrogels prepared at equivalent crosslinker concentrations. Together with recently reported lower toxicity of DAC compared to GA, this cellulose derivative was demonstrated as a particularly interesting hydrogel crosslinking agent.

Fig. 7. The results from comparative crosslinking study showing dependence of network parameters on the used DAC or GA crosslinker amount defined by chemical amount of reactive aldehyde groups per sample. The equivalency between DAC and GA reactive group concentrations is expressed by the two bottom x-axes. The lines connecting points in the graphs are only guides for eyes. 189

Carbohydrate Polymers 198 (2018) 181–190

L. Münster et al.

Fig. 8. Network topology of PVA/DAC is shown in the left part. It is composed of (i) regions containing high local crosslink density adjacent to DAC macromolecules (orange small coils) embedded in (ii) larger regions comprised of free, chemically unbound, PVA chains (light grey) which can be only physically entangled. Some of these chains link the regions (i) together. The right part of the figure shows the homogeneous network topology of PVA/GA crosslinked by GA (dark blue dots). The size, ratio, and number of depicted features is intentionally exaggerated for better understanding. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Acknowledgments

crystalline cellulose. Biomacromolecules, 1(3), 488–492. Kim, U.-J., Wada, M., & Kuga, S. (2004). Solubilization of dialdehyde cellulose by hot water. Carbohydrate Polymers, 56(1), 7–10. Kim, U.-J., Kim, H. J., Choi, J. W., Kimura, S., & Wada, M. (2017). Cellulose-chitosan beads crosslinked by dialdehyde cellulose. Cellulose, 24(12), 5517–5528. Kim, U.-J., Lee, Y. R., Kang, T. H., Choi, J. W., Kimura, S., & Wada, M. (2017). Protein adsorption of dialdehyde cellulose-crosslinked chitosan with high amino group contents. Carbohydrate Polymers, 163, 34–42. Kolodziejski, W., & Klinowski, J. (2002). Kinetics of cross-polarization in solid-state NMR: A guide for chemists. Chemical Reviews, 102(3), 613–628. Lai, S., Casu, M., Saba, G., Lai, A., Husu, I., Masci, G., et al. (2002). Solid-state 13C NMR study of poly(vinyl alcohol) gels. Solid State Nuclear Magnetic Resonance, 21(3–4), 187–196. Lai, S., Locci, E., Saba, G., Husu, I., Masci, G., Crescenzi, V., et al. (2003). Solid-state 13C and 129Xe NMR study of poly(vinyl alcohol) and poly(vinyl alcohol)/lactosilated chitosan gels. Journal of Polymer Science Part A: Polymer Chemistry, 41(20), 3123–3131. Lee, K. Y., & Mooney, D. J. (2001). Hydrogels for tissue engineering. Chemical Reviews, 101(7), 1869–1880. Liu, L. S., Kost, J., Yan, F., & Spiro, R. C. (2012). Hydrogels from biopolymer hybrid for biomedical, food, and functional food applications. Polymers, 4(4), 997–1011. Ma, P. X. (2008). Biomimetic materials for tissue engineering. Advanced Drug Delivery Reviews, 60(2), 184–198. Miyazaki, T., Takeda, Y., Akane, S., Itou, T., Hoshiko, A., & En, K. (2010). Role of boric acid for a poly(vinyl alcohol) film as a cross-linking agent: Melting behaviors of the films with boric acid. Polymer, 51(23), 5539–5549. Münster, L., Vícha, J., Klofáč, J., Masař, M., Kucharczyk, P., & Kuřitka, I. (2017). Stability and aging of solubilized dialdehyde cellulose. Cellulose, 24(7), 2753–2766. Omidian, H., Hasherni, S.-A., Askari, F., & Nafisi, S. (1994). Swelling and crosslink density measurements for hydrogels. Iranian Journal of Polymer Science and Technology, 3(2), 115–119. Peppas, N. A. (2000). Hydrogels in pharmaceutical formulations. European Journal of Pharmaceutics and Biopharmaceutics, 50(1), 27–46. Peppas, N. A., Huang, Y., Torres-Lugo, M., Ward, J. H., & Zhang, J. (2000). Physicochemical foundations and structural design of hydrogels in medicine and biology. Annual Review of Biomedical Engineering, 2, 9–29. Qiu, Y., & Park, K. (2001). Environment-sensitive hydrogels for drug delivery. Advanced Drug Delivery Reviews, 53(3), 321–339. Saallah, S., Naim, M. N., Lenggoro, I. W., Mokhtar, M. N., Abu Bakar, N. F., & Gen, M. (2016). Immobilisation of cyclodextrin glucanotransferase into polyvinyl alcohol (PVA) nanofibres via electrospinning. Biotechnology Reports, 10, 44–48. Sirviö, J. A., Liimatainen, H., Visanko, M., & Niinimäki, J. (2014). Optimization of dicarboxylic acid cellulose synthesis: Reaction stoichiometry and role of hypochlorite scavengers. Carbohydrate Polymers, 114, 73–77. Tang, C., Saquing, C. D., Harding, J. R., & Khan, S. A. (2010). In situ cross-linking of electrospun poly(vinyl alcohol) nanofibers. Macromolecules, 43(2), 630–637. Terao, T., Maeda, S., & Saika, A. (1983). High-resolution solid-state carbon-13 NMR of poly(vinyl alcohol): Enhancement of tacticity splitting by intramolecular hydrogen bonds. Macromolecules, 16(9), 1535–1538. Thomas, S., Durand, D., Chassenieux, C., & Jyotishkumar, P. (Eds.). (2013). Handbook of biopolymer-based materials: From blends and composites to gels and complex networks. Weinheim, Germany: Wiley-VCH Verlag GmbH. Tibbitt, M. W., & Anseth, K. S. (2009). Hydrogels as extracellular matrix mimics for 3D cell culture. Biotechnology and Bioengineering, 103(4), 655–663. Tighe, B. J. (1986). The role of permeability and related properties in the design of synthetic hydrogels for biomedical applications. British Polymer Journal, 18(1), 8–13. Varma, A. J., Chavan, V. B., Rajmohanan, P. R., & Ganapathy, S. (1997). Some observations on the high-resolution solid-state CP-MAS 13C-NMR spectra of periodateoxidised cellulose. Polymer Degradation and Stability, 58(3), 257–260. Veelaert, S., de Wit, D., Gotlieb, K. F., & Verhé, R. (1997). Chemical and physical transitions of periodate oxidized potato starch in water. Carbohydrate Polymers, 33(2–3), 153–162. Voelkel, R. (1988). High-resolution solid-state13C-NMR spectroscopy of polymers [new analytical methods(37)]. Angewandte Chemie, International Edition in English, 27(11), 1468–1483. Wichterle, O., & Lím, D. (1960). Hydrophilic gels for biological use. Nature, 185(4706), 117–118. Yeom, C.-K., & Lee, K.-H. (1996). Pervaporation separation of water-acetic acid mixtures through poly(vinyl alcohol) membranes crosslinked with glutaraldehyde. Journal of Membrane Science, 109(2), 257–265.

This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic - Program NPU I (LO1504), Czech Science Foundation grant 16-05961S to JV. and an internal grants from TBU in Zlin no. IGA/CPS/2015/005, no. IGA/CPS/2016/006, and no. IGA/ CPS/2017/007. The internal grants were funded by financial support for specific research by the university. CIISB research infrastructure project LM2015043 funded by MEYS CR is gratefully acknowledged for the financial support of the NMR measurements at the NMR CF at CEITEC MU. This work was also supported by Operational Program Research and Development for Innovations co-funded by the European Regional Development Fund (ERDF) and national budget of Czech Republic, within the framework of project CPS - strengthening research capacity (reg. number: CZ.1.05/2.1.00/19.0409). Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.carbpol.2018.06.035. References Baker, M. I., Walsh, S. P., Schwartz, Z., & Boyan, B. D. (2012). A review of polyvinyl alcohol and its uses in cartilage and orthopedic applications. Journal of Biomedical Materials Research Part B: Applied Biomaterials, 100B(5), 1451–1457. Bispo, V. M., Mansur, A. A. P., Barbosa-Stancioli, E. F., & Mansur, H. S. (2010). Biocompatibility of nanostructured chitosan/poly(vinyl alcohol) blends chemically crosslinked with genipin for biomedical applications. Journal of Biomedical Nanotechnology, 6(2), 166–175. Bolto, B., Tran, T., Hoang, M., & Xie, Z. (2009). Crosslinked poly(vinyl alcohol) membranes. Progress in Polymer Science, 34(9), 969–981. Bouzková, K., Babinský, M., Novosadová, L., & Marek, R. (2013). Intermolecular interactions in crystalline theobromine as reflected in electron deformation density and 13 C NMR chemical shift tensors. Journal of Chemical Theory and Computation, 9(6), 2629–2638. Capitani, D., De Angelis, A., Crescenzi, V., Masci, G., & Segre, A. (2001). NMR study of a novel chitosan-based hydrogel. Carbohydrate Polymers, 45(3), 245–252. Carpi, A. (Ed.). (2011). Progress in molecular and environmental bioengineering—From analysis and modeling to technology applicationsInTech.. Retrieved from http://www. intechopen.com/books/progress-in-molecular-and-environmental-bioengineeringfrom-analysis-and-modeling-to-technology-applications. Chen, R., & Hunt, J. A. (2007). Biomimetic materials processing for tissue-engineering processes. Journal of Materials Chemistry, 17(38), 3974. Flory, P. J., & Rehner, J. (1943). Statistical mechanics of crosslinked polymer networks II. Swelling. The Journal of Chemical Physics, 11(11), 521–526. Garbow, J. R., & Stark, R. E. (1990). Nuclear magnetic resonance relaxation studies of plant polyester dynamics. 1. Cutin from limes. Macromolecules, 23(10), 2814–2819. Gulrez, S. K. H., Al-Assaf, S., & Phillips, G. O. (2011). Hydrogels: Methods of preparation, characterisation and applications. In A. Carpi (Ed.). Progress in molecular and environmental bioengineering—From analysis and modeling to technology applicationsInTech.. Retrieved from http://www.intechopen.com/books/progress-in-molecular-and-environmentalbioengineering-from-analysis-and-modeling-to-technology-applications/hydrogelsmethods-of-preparation-characterisation-and-applications. Jamnongkan, T., Wattanakornsiri, A., Wachirawongsakorn, P., & Kaewpirom, S. (2014). Effects of crosslinking degree of poly(vinyl alcohol) hydrogel in aqueous solution: Kinetics and mechanism of copper(II) adsorption. Polymer Bulletin, 71(5), 1081–1100. Kamoun, E. A., Chen, X., Mohy Eldin, M. S., & Kenawy, E.-R. S. (2015). Crosslinked poly (vinyl alcohol) hydrogels for wound dressing applications: A review of remarkably blended polymers. Arabian Journal of Chemistry, 8(1), 1–14. Kim, K.-J., Lee, S.-B., & Han, N.-W. (1994). Kinetics of crosslinking reaction of PVA membrane with glutaraldehyde. Korean Journal of Chemical Engineering, 11(1), 41–47. Kim, U.-J., Kuga, S., Wada, M., Okano, T., & Kondo, T. (2000). Periodate oxidation of

190