3.1 Effect of irradiation in aqueous solution
Xyloglucan extracted from Tamarind seeds consists of a b-(1, 4)-D-glucan backbone, which is the same as in cellulose (Nishinari et al., 2007), partially substituted by a-(1, 6)-D-xylose. Additionally, a number of the xylose residues are substituted by b-(1, 2)-D-galactoxylose (Chen et al., 2012). The chemical structure of one of the possible repeating units of the polymer is reported in Figure 2A. The cellulose backbone promotes inter-chain association. The hydrophilic galactosylated side branches present in the structure prevent crystallization and provide the polymer with colloidal stability in water. At low concentration, XG readily dissolves in water at room temperature; the aqueous system appears clear and homogeneous, yet supramolecular, flat, ribbon-like aggregates are always present (Dispenza et al., 2020; Phillips, 2009). Therefore, any molecular weight value obtained from chromatographic analysis or static light scattering measurement should be considered an apparent molecular weight.
Xyloglucan was irradiated as air-saturated aqueous solutions to total doses ranging from 10 to 1000 Gy. The xyloglucan concentration was 0.1 %w, that is approximately the critical chain overlap concentration for this polymer (Freitas et al., 2005). Considering that water is by large the most abundant component of the system, under these irradiation conditions, any observed changes must be attributed to the reactions of polymer chains and their colloidal aggregates with aqueous radiolysis products, mainly hydroxyl radicals, hydrogen atoms and hydrated electrons. In the presence of molecular oxygen, hydrogen atoms and hydrated electrons are rapidly scavanged to produce HO2• and O2•-, respectively (Spinks and Woods, 1990).
Non-irradiated and irradiated XG solutions were analyzed using GFC. The chromatograms expressed as molecular weight, determined using pullulan standards, are presented in Figure 3A. The weight and number average Mw as function of the dose are shown in Figure 3B. The unirradiated XG solution presents a polydisperse profile with an apparent weight average molar mass of 2.57 MDa and a polydispersity index of 1.79. The chromatograms of the irradiated solutions show a significant shift towards lower molecular weights, indicating that chain scission is the dominating effect. Already a dose of 10 Gy caused a detectable reduction in the average molecular weight; it decreased by almost an order of magnitude at a dose of around 250 Gy, and at the highest dose, 1 kGy, it fell well below 10 kDa. This change in molecular weight roughly corresponds to a G-value (radiation chemical yield) for scission in the range of 3-6 x 10-8 mol J-1 based on the simple assumption that one scission per chain reduces the number average molecular weight by a factor of 2. This corresponds to ca 10 % of the G-value for hydroxyl radicals in the g-radiolysis of water, i.e., one hydroxyl radical out of ten produced in the aqueous solution yields scission of XG. Considering the very low molar concentration of XG in solution (< 1 mM), the scission yield is surprisingly high. Hydroxyl radicals can induce cleavage of glycosidic bonds at any location in the polymer chain, close to the middle of the macromolecule as well as very close to the end of the chain. Cleavage of small fragments will not induce significant modification of the molecular weight distribution obtained by GFC. Hence, the scission yield can actually be higher than the one roughly estimated on the basis of the measured changes in apparent molecular weight. The polydispersity index significantly increases with increasing absorbed dose. This implies that radiolysis of XG in dilute aqueous solution does not produce the same effects on all chains. Probably, chains or segments that are fully hydrated scavenge the hydroxyl radicals more effectively and undergo multiple scissions, whereas segments that are aggregated in condensed domains are less accessible to the primary radicals and, hence, less affected by fragmentation. The importance of hydration has been demonstrated in the study by Long et al. (Long et al., 2019) that investigates the degradation of insoluble β-Glucan irradiated in water (5 g of polymer dispersed in 100 ml of water). b-Glucans are linear homopolysaccharides composed of D-glucopyranosyls residues linked via a mixture of b-(1->3) and b-(1->4) linkages. Due to extensive intermolecular interactions the high molecular weight variants of these polymers are insoluble in water. Water-solubility increases in alkaline media, due to the ionization of hydroxyl groups, disrupting the ordered structures formed. In their study, much higher doses are required to achieve a significant reduction in polymer molecular weight and formation of significant amounts of soluble fractions than for XG and the observed effects are pH-dependent, hence water solubility-dependent. At pH 7, a dose of about 50 kGy is required to achieve a 50% reduction in Mw and this dose is reduced to less than 20 kGy when the pH is increased to 9.
We can compare the results for XG to those obtained upon irradiation of HPC at the same doses and concentrations. HPC is a semi-synthetic polysaccharide that shares the cellulosic backbone with XG. It is obtained by partial substitution of hydroxyl groups of cellulose with iso-hydroxypropyl groups (see Figure 2C). The slightly hydrophobic hydroxypropyl substituents give the polymer amphiphilic properties and a lower critical solution temperature (LCST) in water at 40°C. At temperatures below the LSCT, HPC is readily soluble in water. At temperatures above the LCST, the solutions phase-separate into more and less concentrated regions. The phase separation is connected to conformational transition from random coil to compact globule (Bulone et al., 2023). HPC is known to undergo mainly degradation when exposed to γ-radiation, either in solid state or in relatively dilute aqueous solution at temperatures below the LCST, both in air and in absence of oxygen (Wach et al., 2002). When we compare aqueous XG to aqueous HPC irradiated at the same concentration, dose rate and dose, we see that degradation is more pronounced for XG than for HPC. The polydispersity of HPC is almost constant with the irradiation dose. We can argue that the iso-hydroxypropyl substituents of HPC offer competitive reaction sites for hydroxyl radicals. Unlike the sites on the cellulose backbone, reactions of hydroxyl radicals with sites on the iso-hydroxypropyl substituent do not induce chain scission. The radicals formed upon reaction between hydroxyl radicals and the iso-hydroxypropyl substituent can undergo radical—radical combination to form inter- or intra-molecular crosslinks depending on the conditions. This can clearly be seen at moderate and high concentrations (10-30%w), where inter-molecular crosslinking of HPC dominates and macrogel formation is observed upon irradiation (Wach et al., 2002).
3.2 Effect of irradiation in the solid physical form
Preliminary studies conducted with XG in solid physical form, irradiated in air or in a nitrogen atmosphere, showed a dose-dependent reduction in molecular weight over a significantly higher dose range (0-40kGy) than for irradiation in aqueous solutions, with no appreciable influence of the gas atmosphere (data not shown). When XG was purified prior to irradiation, by extensive dialysis with double-distilled water followed by lyophilization, the purified XG was more prone to chain scission than the unpurified XG. One possible explanation was that the unpurified polymer contained some protective impurities that were successfully removed by dialysis. Another possible explanation is that the two polymers differed in their moisture content, which was assumed to be lower in the freeze-dried polymer than in the powder stored in air. The latter hypothesis prompted us to carry out a more systematic investigation of the effect of moisture content, considering two extreme cases, one obtained by conditioning the polymer for one week at 25°C in a desiccator (dry conditioning) and the other by conditioning in a water-saturated atmosphere at the same temperature, for the same time (wet conditioning). The moisture content was quantified by TGA analysis. To find out whether the state of the water, which can be strongly or weakly bound to the polymer or unbound and fill the pores in the solid, also has an effect, the same pretreatments were performed on HPC and CXG, as these two polymers, by virtue of their different substituents, should have different moisture adsorption characteristics from XG. The water adsorption properties are a result of their adsorption sites, the exposed OH groups present in all three polymers, and carboxyl groups for CXG, but also of their specific surface area and porous structure.
The results of the TGA analysis are illustrated in Figure 4. Interestingly, XG and CXG conditioned to be dry appear to have a water content of about 5 %w and 8 %w, respectively, and cannot be considered dry in the strict sense. HPC has a much lower water content after dry conditioning (1-2 %w). After wet conditioning, the water content is fairly high for all three polymers. For XG and HPC, it is around 40 %w, while for CXG it is above 60 %w. The higher water uptake of CXG may reflect the presence of carboxyl groups in addition to hydroxyl groups and the reduction of molecular weight and probably intermolecular association, favoring the water molecules to diffuse from the surface into the bulk polymer structure. Although XG and HPC have more or less the same water content, the state of the adsorbed water differs significantly. For XG, desorption occurs at temperatures around 80°C, while for HPC we observe several desorption phenomena, starting from about 40°C to well above 100°C. The wider temperature interval points to a variety of water release mechanisms acting and possible different states of water. At the lower end of the temperature range, water desorption can be associated with the typical coil-to-globule transition of HPC that releases water molecules and is prodromal to condensation and eventually phase separation. At the higher end of the temperature range, the largest fraction of water is released. Considering that temperatures are above 100°C, this water is strongly bound to the polymer. At the intermediate temperatures, the water released can be considered “free” or only loosely bound, i.e. water filling the pores of the solid powder.
The three polymers, preconditioned in dry and wet atmospheres, were subjected to the same irradiation doses (10 kGy, 20 kGy and 40 kGy) in air. After irradiation, the powders were dissolved in water and the solutions were analyzed using GFC. The chromatograms were further analyzed through deconvolution analysis. As for the irradiation in aqueous solutions, there is no indication of radiation-induced crosslinking in the irradiated samples.
The chromatograms, the relative occurrence of the different size fractions used for the deconvolution of the chromatograms as a function of dose and the Mn, Mw and PDI plots as function of dose, for both dry and humid XG powders are presented in Figure 5. The molecular weight fractions used in the deconvolutions were determined from the best fit to the chromatograms of the corresponding unirradiated polymers. The same fractions were used for the analysis of all the irradiated samples. At higher doses, additional lower molecular weight fractions were added if deemed necessary. In general, the fits were of high quality, as shown in Figure S2 for dry powders and Figure S3 for wet powders, where the original chromatograms, in green, and the curves resulted from gaussian deconvolutions, in yellow, mostly overlap. The chromatograms (Figure 5A, A’) show an evident shift of the peak towards the lower molecular weights with increasing dose. For each system, we also observe that Mw and Mn decrease almost to the same extent; in other words, the polydispersity increases to a much smaller extent as compared to irradiations in solution, especially for the irradiated dry powder, which points to random backbone scission as a result of the direct effect of irradiation (Kempner, 2011). The deconvolutions of the chromatograms (Figure 5B, B’) show that the two largest size fractions (dark and light blue lines) are progressively consumed, while the intermediate size fractions first increase, being fed by fractionation of the longer chains, then decrease and/or reach a plateau, while the abundance of smaller size fractions steadily increase. The largest size fraction is consumed faster in the dry conditioned XG compared to the humid sample. This would imply that adsorbed water has a slight protective effect. For the other size fractions, a comparison is more difficult to make. However, the conditioning has an insignificant impact on the radiation-induced change in Mw and Mn (Figure 5C).
When comparing the XG results to the other two polymers, CXG (Figure 6) and HPC (Figure 7), we observe the same general trend, with molecular degradation increasing with dose.
Judging from the chromatograms of CXG (Figure 6), the difference between the dry and humid conditions is not significant. However, a slight protective effect of adsorbed water can be observed at the highest dose. From the chromatograms of HPC (Figure 7), the protective effect of moisture is clear.
We can also compare the results of the dry irradiations in more absolute terms. For XG and HPC the average molecular weight decreases by a factor of 4.5-5 at 40 kGy while for CXG the average molecular weight only decreases by a factor of 2 at 40 kGy. XG and HPC have similar initial molecular weights, ca. 3 times higher than that of CXG. The difference in initial average molecular weight can to a large extent explain the apparent difference in degradation between CXG and the other polymers. For lower molecular weights, a higher number of scissions are required for a given reduction in molecular weight compared to higher molecular weights. This is simply due to the fact that a given mass of longer chains requires less scissions than the same mass composed of shorter chains in order to reduce the average molecular weight by a factor of 2. Accounting for the initial average molecular weight and the overall change in molecular weight upon irradiation we can estimate the G-values for scission using a very simplified approach (based on the assumption that one scission per chain reduces the average molecular weight by a factor of 2 and that for every additional reduction in molecular weight by the same factor 2n scissions are required, where n is the number of scissions leading to a reduction by a factor of 2 in the previous step). The resulting G-values for scission are in the order of 6-9 x 10-8 mol J-1 for all three polymers implying that the differences in structure play a very minor role in dry-irradiation of XG, CXG and HPC.
A protective effect has been observed before for other macromolecules and is often attributed to energy transfer from the macromolecule to the adsorbed water molecules (or in some cases other small molecules adsorbed to the macromolecule) (Ehrenberg et al., 1957). This would apply mainly to excited states originating from the absorption of γ-radiation. By transferring the excess energy to the adsorbed water molecules, scission of the excited states is avoided. From the TGA, it is clear that HPC has the most strongly bound water. It is also evident that pre-conditioning in humid atmosphere has the largest effect on radiation induced fragmentation in the case of HPC as compared to XG and CXG. This would imply that energy transfer is more efficient to strongly bound water.