3.1. Xanthan Gum Oxidation
After periodate oxidation and dialysis, the product was subject to hydroxylamine hydrochloride titration to confirm and quantify the aldehyde content. Figure 1
shows a representative example of the pH evolution of the titrated liquid and its first derivative.
The degree of oxidation computed from Equation (1) was 65% ± 8% for 2 h reaction time, based on the measurements from 10 distinct oxidation runs. Serrero and co-workers found similar degrees of oxidation for potato starch under analogous reaction conditions [24
]. For 24 h reaction time we obtained an aldehyde substitution of 83% ± 3%. This increase in the degree of oxidation with reaction time was expected. However, the more reasonable reaction time of 2 h was adopted for this work.
NMR was used to study the modifications occurring in the biopolymer’s primary structure after oxidation. Figure 2
presents the chemical structure of xanthan gum, where letters A to E identify the saccharide units in study. Previous studies on xanthan gum and other polysaccharides were helpful in the interpretation and discussion of the obtained results [36
]. The 1
H NMR spectra, obtained at 55 °C, for XG and XGox are presented in Figure 3
. Table 1
contains the spectroscopic data for the main signal attributions discussed.
In the range between 3 and 4 ppm, it is possible to find most of the chemical shifts of the protons from the saccharide ring, except for the anomeric protons that appear between 4.5 and 5.5 ppm. However, those are difficult to attribute due to peak overlap from the different saccharide units. On the XG spectrum, two distinct peaks from the methyl group of pyruvated (E) (1.458 ppm) and acetated (C) (2.134 ppm) units are visible. At 55 °C, we have quantified the amount of pyruvated and acetated units, since its known that not all units C and E have acetate or pyruvate groups, respectively. Comparing the known amount of internal reference present, we estimated that 43% of C units have an acetate group and 45% of E units have a pyruvate group. This is in accordance with the literature, which indicates the presence of one pyruvate and one acetate group per two polymer side chains [37
]. Other works present different ratios of pyruvate and acetate content, which can be due to the bacterial strain and growth conditions, such as temperature and nutrients present [44
]. In the XGox spectra, it is possible to note the appearance of two new signals in the same region (1.4–2.5 ppm), attributed to the groups present in the oxidized saccharide units. Interestingly, when comparing the signal for the methyl group for both oxidized and non-oxidized xanthan, it is also apparent that oxidation leads to a loss of pyruvate groups. This shows that oxidation with periodate not only affects the C–C cleavage between adjacent –CHOH groups, but also leads to a loss of terminal groups in the side chain, and therefore results in a polymer with lower molecular weight. Li, et al. also reported that an aggressive oxidation approach can lead to the degradation of the oxidized polysaccharide units, either in acidic or alkaline media [22
]. To better understand these phenomena, interpretation of other signals was taken into consideration. The proton NMR spectra also exhibits new signals in the region of 9 ppm, attributed to the presence of aldehyde groups formed during the oxidation process. The chemical shift at 9.103 ppm is attributed to an aldehyde proton in the vicinity of a C=C bond. This interpretation was based on the 2D heteronuclear correlations (HSQC and HMBC) obtained for the oxidized polymer. From the signals obtained, we can propose the formation of new structures at different stages of oxidation (Figure 4
Structure E2 was predicted following the correlations obtained in HSQC and HMBC mode. The aldehyde group (HC=O, 9.103 ppm/191.210 ppm) was revealed to be in the vicinity of an ethylenic carbon. HMBC data revealed a carbon signal at 152.060 ppm, meaning the presence of a quaternary carbon (qC) that was also neighbouring a carbon (126.686 ppm) linked to a proton with signal at 6.354 ppm. With this information, we were able to predict structure E2, assuming the previous form (E1) of the pyruvated mannose.
Correlating the values obtained for the proton signals at 9.103 ppm (aldehyde) and 6.354 ppm (HC=qC) for the presented structure (E2) with the ones obtained for the methyl group of the non-oxidized (E) and oxidized (E1) pyruvated unit, we are able to say that these three structures represent the total amount of pyruvated saccharide units in the original xanthan gum. In structure E1, the saccharide ring opened between C-2
and formed two aldehyde groups. In some repetitive units, this structure is stable and was confirmed by the shift to lower field of the CH3
-pyruvated mannose (Figure 3
). However, some of the E1 structures suffered further attack due to the slightly acidic conditions. After the formation of the aldehyde groups, a β-elimination reaction could occur, leading to the formation of the third structure (E2) with the release of pyruvic acid, which is then removed during dialysis. Although β-elimination is more likely to happen in alkali conditions [46
], Veelaert, et al. reached similar results studying physicochemical transitions on oxidized potato starch at acidic pH. They also verified a decrease of the molecular weight of the oxidized polymer, which was attributed to acid hydrolysis that may result in chain scission [48
]. This phenomenon was not studied by us, but is expected to occur.
On structure E2, the loss of pyruvic acid was suggested by the lower field shift of 2H-6 (4.482 ppm, dd, J = 3.6 and 19.2 Hz; 4.598 ppm, dd, J = 2.4 and 19.2 Hz) when compared to the respective signals of XG (3.7–4.0 ppm), as well as by the lack of other CH3 pyruvated signal. The coupling constants (dd, J = 2.4, and 3.6 Hz) of the ethylenic proton at 6.354 ppm (C-4) suggest that this proton has the ability to establish an allylic coupling with 2H-6.
The low intensity of proton signals in the region of 9–10 ppm can be explained by the fact that such aldehyde groups can be in equilibrium with hemiacetal groups, as proposed by other authors [31
]. Another aspect found in the literature is the behaviour of terminal mannose units, where three hydroxyl groups can be found in subsequent carbons. The oxidation pattern expected for those groups is represented in Figure 4
(E3 and E4 structures). The periodate attack to the mannose unit E3 happens in two locations: between C-2
. As suggested in the literature, the double cleavage leads to the release of C-3
as formaldehyde, which was later removed by dialysis [51
]. The C-6
carbon (60.740 ppm) identified in unit E3 is linked to two protons. Through 2D HMBC correlation, one of those protons (3.654 ppm) is correlated to the aldehyde group (HC=O, 9.223 ppm/194.137 ppm) attributed to the terminal carbonyl group of the non-pyruvated mannose unit at the end of the side chain. The terminal mannoses were more visible, and therefore easily identifiable due to the mobility of the molecules at the end of the side chain. The high molecular weight and complexity of the polymer reduce the internal mobility of relevant chemical groups, and make the identification of other relevant signals present in the oxidized molecule difficult.
3.3. Tensile Strength
The tensile strength results obtained for the glued cork samples with different adhesives at different mass concentrations in solution are shown in Figure 5
. In all cases, there is a tendency for improving strength with increasing concentration. After immersion in water for 24 h, joints glued with xanthan gum (XG) exhibit null bond strength for all concentrations tested. Oxidized xanthan gum (XGox), on the other hand, shows significantly better water resistance, in agreement with the TSM results, especially at the highest concentration. Interestingly, XGox at 6% yielded the best tensile strength result of all glues. The aldehydes present in the oxidized gum are probably able to react with hydroxyls in the cork structure, and to form a crosslinked adhesive structure upon heating. Chitosan glue (CS) exhibits a behavior similar to XGox, but with slightly higher water resistance. Chitosan’s known adhesive performance can be attributed to two main factors. On one hand, it is able to establish electrostatic interactions, hydrogen bonding, and van der Waals forces between d
-glucosamine and the adherend. On the other hand, it has good wetting capability, due to low surface tension and a high dispersive component of the surface energy [54
XG, XGox, and CS were combined in different ratios in order to evaluate the possibility of a synergistic effect resulting from physico/chemical interaction between the two. Figure 6
represents the tensile strength and water resistance of mixtures for two different gum:chitosan mass ratios. The results are compared to the synthetic polyurethane adhesive traditionally used for bonding cork.
Mixtures of XG and CS show worse performance than CS alone, indicating poor interaction between chitosan and xanthan gum. The combination of XGox and CS leads to better results, but still below the performance of CS alone in terms of water resistance.
In the tensile strength tests, chitosan and the synthetic adhesive show similar behavior. Since CS shows total solubility as a standalone film, the good water resistance of the glued joint is probably associated with interactions with reactive groups present on the cork surface, such as addition to epoxide ring. Oxidized xanthan gum proved to be a good adhesive for cork. Even though we could not obtain direct evidence of this, it is expected that aldehyde groups present in XGox would react with hydroxyls on the cork cell walls, forming hemi-acetals. As with chitosan, this interaction may justify why XGox films show poor water resistance (TSM test), but achieve very good results in the tensile strength tests of glued joints after immersion in water. Combination of oxidized xanthan gum with chitosan had the potential to improve the adhesion properties due to crosslinking of the aldehydes with the amino groups to form an imine linkage (Schiff base). This reaction, however, competes with the interaction of aldehydes with cork surface, and the net result is therefore uncertain. In this work, we observed no improvement in performance for this mixture.