Showing papers in "Journal of Polymer Science Part C: Polymer Symposia in 1965"

Effect of component removal upon the porous structure of the cell wall of wood

Abstract: It has been shown by nitrogen adsorption that the removal of lignin and hemicellulose from the middle lamella and cell wall of spruce wood produces pores with a median size in the 20–40 A. range throughout the dissolution process. During both kraft and sulfite pulping, these pores have a total volume which from 90–100% yield equals, from 55–90% yield exceeds, and below 55% yield is less than, the volume of material removed from the wall. The results therefore indicate a swelling of the wall at intermediate yields followed by a contraction below the initial dimensions towards the end of pulping. In accordance with a previously postulated theory of cell-wall structure based on a multilamellur model, the pores produced by the dissolution of wood components are the slit-like spaces between adjacent lamellae. Specific surface measurements show that, on this model, the cell wall of water-swollen wood consists of about 5 lamellae, which subdivide into a total of about 350 lamellae each 35 A. thick at the point of maximum swelling (65–70% yield). Thereafter, towards lower yields, the lamellae start to reaggregate. This technique applied to groundwood pulp shows that the process of splitting the wall into layers may be accomplished to some extent by purely mechanical action.

Crystalline character in polymers

Abstract: The original fringed micelle model of a crystalline polymer solid assumed two phases, submicroscopical bundled chain crystals embedded in an amorphous matrix. Every macromolecule passed alternately through crystalline and amorphous regions. This model ideally represented the crystallization with highly stretched rubber, but failed to explain spherulites and lamellar single crystals. The folded chain model of the lamellae grown from dilute solution postulated in a polymer solid the inclusion of all the material in single crystals (one phase model). The amorphous x-ray diffraction pattern and the density defect were due to crystal defects, particularly to the chain folds at the crystal surfaces. Later development has brought both models closer together. In spite of rather regular chain folds with immediate re-entry there are some chains folding back in a switchboard manner. In a sample crystallized from melt there are still more chains incorporated in two different crystals so that the intervening portion is prevented from crystallization. Together with uncrystallizable components, such tied molecules are constituting an amorphous phase of the very same type as assumed by the fringed micelle model. The basic improvement of the new model is just the comprehension of the important role of chain folds in establishing a discontinuity between the crystals and the amorphous material. A particularly important problem in crystalline polymers is the plastic deformation and the resulting high orientation in drawn fibers and films. From the experiments with single crystals, one learned the basic mechanism of deformation, i.e., phase transformation, twinning, chain tilt, and slip which finally leads to complete break up of original crystals into smaller blocks which are incorporated in the pulled out fibrils. The tied molecules in such highly oriented material are markedly strained and aligned which show up in a decrease of entropy and enthalpy as compared with a completely relaxed amorphous phase. The decrease in enthalpy quantitatively explains the conspicuous depression of sorption, dyeability, diffusion, and permeability of drawn fibers and films.

The thermal expansion of cellulose, hemicellulose, and lignin

Abstract: The thermal expansion of wood constituents has been measured dilatometrically both in the dry and the water-swollen state. Dry measure meats were made with mercury as the confining liquid on samples pressed into dense pellets at elevated temperatures and pressures. Between −30 and 30°C, the thermal expansion coefficients for various celluloses and hemicelluloses ranged from 5.1 to 6.0 × 10−4 ml./g./°C. For lignin, the coefficient increased to about 10 × 10−5 ml./g./°C. A second order transition was noted in all samples at 19–33°C. A further transition at 8G-110°C. was detected for the dioxane lignin, periodate lignin, and Avory cellulose. All transitions in the dry state produced an increase in the expansion above the transition temperature. In the water-swollen state, samples were in the form of powders or broken pellets, with distilled water as the confining liquid. The thermal expansion coefficients of the water-swollen samples were several times as great as the expansion coefficients measured in the dry state. The second order transition near 20°C. persisted, but in water the transition was negative with expansion decreasing above the transition temperature. Analogous results were obtained for glucose and cellobiose, both in the dry state and in solution. The transition in the dry state is attributed to increased thermal motion caused by the rupture of weak hydrogen bonds. The large and interesting discrepancy found in the expansion of the water-swollen materials is interpreted as arising from the perturbation of the water structure by the hydrophilic surfaces of woody macromolecules, and a quantitative assessment of this effect is attempted.

The acidic groups in wood and pulp as measured by ion exchange

Abstract: The ion-exchange capacities were determined for a variety of wood samples consisting of both softwood and hardwood. The experiments revealed that considerably more acidic groups are present in heartwood than in sapwood. The capacity (carboxyl) values found for all samples were substantially lower than would be expected from the estimated uronic acid contents. This difference was more striking for hardwood than for softwood. After alkali treatment (pH 11.5, 64°C), the exchange capacity of the wood was increased considerably and this effect was more pronounced for hardwood than for softwood. This observation either indicates the presence of carboxylic acid estera and lactones or is attributable to an increased accessibility of carboxyls. Such an increase may occur in hardwood as a consequence of the deacetylation of xylan. In another series of experiments the effect of sulfonation was studied. Wood samples (Picea abies, Betula verrucosa, and Betula pubescens) were sulfonated at various pH values at 80 and 90°C. so that only small quantities of the wood substance were dissolved. In addition, the number of acidic groups was determined for a series of spruce and birch pulps with varying lignin contents. It was possible to distinguish between the carboxyl and sulfo groups by means of an ion-exchange method developed for that purpose. The experiments revealed that the accessibility of carboxyl groups is considerably increased even after sulfonation at a low temperature. It also became apparent that most of the carboxyl groups are accessible even in pulps with relatively high lignin contents. The carboxyl values together with xylose analyses showed that the acidity of xylan remaining in pulp is substantially decreased during cooking. This decrease is more pronounced for birchwood than for sprucewood. Finally, it is evident from these experiments that sprucewood lignin is sulfonated to a markedly higher degree than birchwood lignin.

The sorption of glucomannan and xylan on α‐cellulose wood fibers

Abstract: The rates of sorption of spruce glucomannan and of a modified birch xylan (containing less than 3% uronic anhydride) upon never-dried α-cellulose wood fibers in 0.02N alkali have been measured at 4 temperatures ranging from 5–40°C. By labeling the glucomannan with carbon-14 and the xylan with tritium it was possible to measure the rates of sorption of each hemicellulose alone or in solution together at the same concentration. In both cases the rates of sorption of the glucomannan were about twice those for the xylan. Average activation energies (kcal./mole) derived from the rates were as follows: glucomannan alone, 3.8; glucomannan in admixture with xylan, 3.1; xylan alone or in admixture with glucomaunan, 6.3. The initial rates of sorption, estimated graphically by a methods which is shown to have severe theoretical limitations, gave higher values for the activation energies, viz.: glucomannan alone, 6.6; glucomannan in admixture with xylan, 5.8; xylan alone, 10.0; xylan in admixture with glucomannan, 10.1 kcal./mole. All these values suggest that the sorption of both hemicelluloses is physical. Although the initial rate of sorption of the xylan was not significantly affected by the presence of the glucomannan, its total sorption after 10 min was reduced by an average of 18% when glucomannan was present. The lower rate of sorption of the xylan is ascribed to the presence of its uronic acid carboxylic groups. A small percentage of these groups can apparently override the conformational similarity of xylan and cellulose.

The number‐average molecular weight of native hardwood xylans

Abstract: A 4-O-methylglucuronoxylan has been isolated in a yield of 90% by direct alkaline extraction of trembling aspen (Populus tremuloides) wood. The number-average molecular weights of the acetate and valerate derivatives were determined in chloroform-ethanol and of the xylan itself in dilute cupriethylenediamine. There was no evidence for the occurrence of molecular association of the acetate in chloroform-ethanol. Tetra-chloroethane, dimethyl sulfoxide, and cadoxen were unsuitable solvents. The results showed that the polysaccharide had a number-average molecular weight of 31,500, corresponding to a degree of polymerization of 200. The constants K and a in the Staudinger-Mark-Houwink equation were determined for cupriethylenediamine solutions of a series of celluloses and xylans, the values obtained being 2.2 × 10−2 for K in both cases, and 0.78 and 0.72 for a, respectively. It was concluded that cellulose and xylan probably have the same conformation in aqueous alkali, intermediate between a compact and a free-draining coil. Since the xylan framework is known to contain two branches per average molecule, the results would probably also imply that the branches are relatively short.

Derivation of possible chain conformations for poly‐β‐1,4‐anhydroxylose

Abstract: An oriented fiber diagram of deacetylated O-acetyl xylan from white birch has been obtained. The quality of the x-ray pattern, while better than any reported to date, is still inferior to the various cellulose patterns. Only 14 intensity maxima are recorded and available for analysis. The repeat distance parallel to the chain direction has been directly calculated from the x-ray data as 14.84 ± 0.3 A. Systematic meridional absences indicate a threefold screw axis along the chain. Using these data, and assuming x-ray structure parameters for small molecules to apply to the monomer unit in the polymer, the basic repeat unit has been described mathematically. All possible relative settings of two monomer rings consistent with a threefold screw axis and a given repeat distance for a left- and right-handed helix have been calculated and examined. This approach is equivalent to that used for cellulose where the C1OC4,′ angle is allowed to vary in order to satisfy the above conditions. Choice of probable conformations is made by simultaneously considering existing infrared dichroic data for oriented xylan films and generally accepted stereochemical concepts. The problem has been programmed for and run on an IBM 1620 data processing system. The choice of a unique conformation is not possible but evidence seems to favor the left-handed conformation with a C1OC4 bond angle of about 116°, for which an intramolecular hydrogen bond (O5…O3) is possible. The H1… H4′ distance for this conformation is 2.4 A., while the CH2 transition moment vector is strongly perpendicular (88° to the chain axis) as required by the infrared data. None of the right-handed conformations allowed an (O5…O3′) intramolecular hydrogen bond and usually a serious overlap of van der Waals' radii for O2 and H4' atoms was found.

The polymerization of 1,4‐anhydro sugar derivatives

Abstract: 1,4-Anhydro-2,3,6-tri-O-methyt-D-galactose has been synthesized and positively identified for the first time. 1,4-Anhydro-2,3-di-O-methyl-L-arabinose was obtained in nearly pure form by an analogous reaction sequence. Both of these compounds undergo addition polymerization under the influence of Lewis acid catalysts to yield amorphous polymers with maximum degrees of polymerization of approximately 90. The chain propagation proceeds in each case with nearly equal amounts of 5- and 6-mernbered ring opening leading to a mixture of furanosidic and pyranosidic unite in the polymer backbone. The resulting configurations at the anomeric centers under various conditions of polymerization, as elucidated mainly by degradation studies, can be explained only by a carbonium ion rather than a trialkyl oxonium ion polymerization mechanism and with the assistance of conformations analysis. The C-5 substituent in the 1,4-anhydro-2,3,6-tri-O-methyl-D-gitlactose was found to influence the stereochemistry of the chain propagation.

Investigation of polymer order by X-ray diffraction

Abstract: The determination of per cent crystallinity in polymers by x-ray diffraction techniques is generally based on a two-phase system. Complications arise when intermediate phases are present reducing the significance of the value obtained, To overcome this difficulty, a procedure is described which yields the maximum information from the x-ray diagram of a polymer. The use of this procedure with oriented nylon 6 establishes the presence of a paracrystalline a phase. The effect of this phase on the density of oriented materials is emphasized. In conclusion, it is suggested that crystallinity, as a degree of order, can best be represented as a tensor rather than a scalar property.

Comparison of the configuration of β‐1,4 linked hexosans and wood xylan pentosans by means of the eizner‐ptitsyn viscosity equations

Abstract: The Porod-Kratky persistence lengths, monomer and polymer radii, and Flory coefficients of a number of β0–1,4 linked hexosans have been compared to each other and to the β-1,4 linked xylan pentosans of wood by applying the Eizner-Ptitsyn viscosity equations to intrinsic viscosity and molecular weight analyses from a number of sources. The analyses were made on fractions of cellulose acetate, cellulose nitrate, cellulose caproate, the diethylacetamide derivative of cellulose xanthate, cellulose in cadoxene and in the alkaline ferric tartrate solvent, FeTNa, the galactomannan triacetate of guar, and the glucomannan triacetate of orchid salep tuber. Xylans from spruce and birch, polystyrene, and a branched dextran were also investigated. The persistence lengths of the hexosan derivatives, except the cellulose trinitrate, were remarkably similar and were found to vary from 48 A. for the cellulose caproate determined under theta conditions to 58 A. for the galactomannan triacetate. The persistence lengths of cellulose in the two complexing solvents (72 A.) were nearly identical, confirming that these solvents do not increase the solution dimensions of cellulose greatly. The configuration of the polymers were ranked on a common graph by plotting the log of the polymer radii to contour length ratios vs. the log of the contour length. The investigation shows that the 0–1,4 linked hexosans have a common configuration which is independent of the length of the derivative side group or of single α-1,6 linked side groups, as in the galactomannan. The configuration of the hexosans is also independent of whether the main chain is composed of glucose or mannose, or a mixture of these sugars. The greater rigidity of the hexosans, in comparison to the pentosans, is explained in the former by hindrance between the C2 group of the one sugar with the C6 group of the next, which is absent in the pentosan that has no C6. Analogy is made to the hindered racemization in substituted biphenyls to explain the great rigidity of cellulose trinitrate, whose persistence length is 132 A.

Structure of cholla gum (Opuntia fulgida)

Abstract: The structure of cholla gum has been investigated using (a) the methylation procedure, (b) by oxidation of the polysaccharide with periodate (the Smith procedure), and (c), by its partial hydrolysis with the formation of sugars and oligosaccharides. It is composed of a main chain of D-galactopyranose residues fi-linked 1 3 and/or 1 6, to which are attached side chains. These branches contain D-galacturonic acid, D-galactose, and D-xylose, all in the pyranose form, and L-rhamnose and L-arabinofuranose residues. The chains are terminated by D-xylopyranose or L-arabinofuranose units. D-Xylose is present only in terminal residues.

Studies on alcohol-modified transition metal polymerization catalysts. II. Interaction of TiCl4 with cellulose and model compounds

Abstract: Experiments have been performed to obtain evidence concerning the chemical and physical reactions which may be expected to occur when TiCl4 is sorbed onto cellulose from nonswelling solvents. For model compounds such as 1,2-butanediol, cis- and trans-1,2-cyclohexanediol, one ionic bond per TiCl4 was the general rule. The presence of multiple electron donating groups in the diols leads to chelation and/or polymer formation, so that insoluble reaction products were obtained. Analysis of the reaction products permitted the assignment of characteristic formulae, which accounted for variations in composition and physical properties of the reactants as a function of diol: TiCl4 ratio. The foregoing studies have provided a base for interpreting, respectively, the homogeneous and heterogeneous reactions between ethyl cellulose (DS = 2.5) and cellulose fibers with TiCl4. It appears that all the oxygens of cellulosics (not just the hydroxyls) can take part in the completing, which otherwise is similar to the diol reaction. Drying history has an important effect on the titanium content of the cellulose after reaction with TiCl4.

Review of solution methods and certain other methods of polymer fractionation

Abstract: Discrimination in terms of solubility differences forms the basis of the most widely applicable methods of polymer fractional ion. These include fractional precipitation and extraction on a batch scale; the rapid column methods, namely, the elution method and precipitation chromatography; and turbidimetric titration, which is strictly an analytical procedure. The principles of solution fractionation are reviewed as applied to batch fractionation and to the column methods, and certain essential differences are pointed out between these two approaches. The importance of proper solvent selection in the several methods is considered, and examples are presented of fractionation on the basis of chemical and structural differences as well as in terms of molecular weight. A brief summary is also presented on the present status of certain other methods, including thermal diffusion, none refining, and adsorption chromatography, which utilize different physical principles as the basis of separation.

Polyethylene‐encapsulated cellulose

Abstract: An encapsulation process is described in which cellulose fibers were enveloped with sheaths of polyethylene by the use of transition-metal catalysts. Cellulose-catalyst interactions resulted in a fiber surface completely covered with initiating sites, and growth of polymer occurred from the fiber surface outward to form a shell around each individual fiber. Removal of moisture and in situ catalyst formation were found to be essential for successful encapsulation. Catalyst interactions during different stages of formation were studied by following the absorption of the various catalyst components on cellulose. Active sites on fiber surfaces were postulated to have been formed by interaction with TiCl4 and by reaction of reduced transition metal compounds with electron donor groups in the cellulose. Polymer growth was followed in a series of electron microphotographs and polyp-like spherical polyethylene particles were observed at an early stage in the polymerization. It was concluded that encapsulation was a chemical means of placing polymer around a fiber core and not a grafting reaction. Fiber properties, uses of encapsulated cellulose, and conversion into web structures by use of paper making techniques were also discussed.

The alkane dithioether and the polysulfide crosslinking of cellulose

Abstract: New crosslinked cellulose derivatives were obtained from the reactions of potassium disulfide and n-alkane-α-ω-dithiols with partial tosyl or brosyl esters of native cellulose. Treated cotton print cloth showed significant improvement in wet and dry wrinkle recovery over untreated. Alkane dithioether crosslinks, —S(CH2)nS—, of different lengths (2–10 CH2 groups) were introduced. Treatment with aqueous or ethanolic solutions of potassium disulfide resulted in the formation of polysulfide crosslinks, such as tetrasulfide. The effect of crosslink length on the physical properties of the modified cottons was studied, and it was found that fabric wrinkle recovery tended to decrease as crosslink length increased. It was also found that the tosyl and brosyl cellulose esters were degraded by alkali. This degradation is ascribed to the loss of primary sulfonyloxy groups with subsequent hydrofuranol ring formation in the case of the anhydroglucose units involved.

The optical rotation of wood xylans

Abstract: Previously derived equations relating the specific rotation and chemical composition of 4-O-methylghicurono-xylans and arabino-4-O-methylglucurono-xylans have been found to be partly incorrect because of the assignment of an erroneous value to the rotatory contribution of the 4-O-methyl-α-D-glucuronosyl xylose residue. The correct value was calculated from the known molar rotations of xylobiose and 22-(4-O-methyl-α-D-glucuronosyl) xylotriose. Approximately the same values were obtained from the specific rotations of five carefully purified 4-O-methylglucurono-xylans, containing different proportions of acid side chains, and from the rotations of an aldopentaouronic and an aldohexaouronic acid. A revised equation was used for estimating the specific rotation of several arabino-4-O-methylglucunmo-xylans from various softwoods. Contrary to the earlier results, a satisfactory agreement was found to exist in most cases between found and calculated values.

Polymerization on the surface of cellulose sheets: Formation of an impermeable film

Abstract: Polymerization by interfacial polycondensation was explored as a means of preparing materials of superior printability in the form of paper sandwiched between 2 sheets of polymer film and intimately attached to them. Reaction systems forming polyesters, polyurethanes, and polyureas were found to be unsatisfactory, because either the reactions weretooslow, or the polymers were weakand tacky. Nine combinations of diamines and diacid chlorides yielded promising polyamide films and were tested at various levels of concentration in a specially designed apparatus. Since papers containing either size or residual lignin were unsuitable for this coating system, Whatman No. 4 filter paper, representing a pure cellulose paper, was used as the substrate. As measured by air, water, and oil resistance, the best coated papers were from poly(1,6-hexanediamine sebacamide), poly(1,6-hexanediamine azelamide), poly(1,4-bisaminomethylcyclohexane sebacamide), and poly(m-xylylenediamine sebacamide). Some qualitative correlations were found between the physical properties of coated papers and the concentrations of the reactants, but the results were too variable to permit formal statistical analysis. Electron micrographs indicated that the strongest combinations were obtained when the polymer conformed closely to the fibers at the surface of the paper.

Warts in selected species of pine

Abstract: The warts found on the lumen walls of pine tracheids survive the standard pulping processes and are therefore present in finished chemical pulp fibers. The present work was done to supply more fundamental information concerning the warts in pine wood. Sections and replicas for electron microscopy were prepared from the cambial zone of growing slash and loblolly pine trees. The warts appear on the lumen walls at the end of the cell differentiation phase. A multitude of heterogeneous materials along with the warts can be found at this stage on the lumen walls. The majority appear to be resin droplets which can be removed by hot alcohol extraction. The tracheids in the mature wood have numerous, non-uniformly distributed particles on their lumen walls. Judging by the distribution patterns and the localities where they are usually found, the warts appear to be sedimentation products originating from the translocation stream of the decomposed protoplasts. Proteinaceous residues, whole cell organelles, microorganisms, and their spores are considered as possible constituents of the warty layer. The species of pine investigated can be rated in the following order according to the increasing number of warts contained: eastern white, sugar, slash, and loblolly pine. The extensive microscopy work revealed some interesting, unreported, morphological characteristics of pine wood.

Brightness reversion of birch sulfate pulp

Abstract: Three methods of improving brightness stability have been examined: (1) Avoidance of chlorination of extractive components and the residual resin. (2) Modification of the hemicellulose by high-temperature treatment of the pulp to remove xylan components of poor brightness stability. (3) Maximum reduction of the residual lignin in the pulp. The results of the three methods have been compared and an examination of their combined effect has thrown light on the main cause of brightness reversion in each case. The chlorinated extractives adversely affect the brightness stability during aging in a dry atmosphere, and the reversion in this case is due to degradation mainly of carbohydrates by the hydrogen chloride liberated from the chlorinated extractives. Different fractions of the extractives have been tested with respect to their ability to promote this effect. It would seem that the brightness stability of the pulp is dependent on its carbohydrate composition. An aqueous treatment of the pulp at a high temperature removes some hemicellulose components sensitive to reversion (particularly xylan, which has a high content of uronic acid). Under moist conditions, the hydrochloric acid released from the extractives is diluted and thus has much less influence on the degradation of the carbohydrates. The extractives then contribute to the reversion mainly through their own discoloration. The transformation of the carboxyl groups in the pulp to the hydrogen form by a sulfur dioxide treatment usually increases the brightness stability in a moist atmosphere and decreases it under dry conditions. Our results also strongly indicate that the brightness reversion is highly dependent on the degree of chlorination of residual lignin.

Degradation of wood pulp during acetylation

Abstract: A comparison of the DP distributions of wood pulps and the cellulose acetates prepared from them shows that the DP distributions of the acetates are essentially exponential and virtually independent of the DP distribution of the original pulp. The DP of a number of acetates prepared under identical conditions from different pulps were unrelated to the DP to the starting pulps. Although the pulp DP is thus not a factor which determines the acetate DP, there are other intrinsic properties of the pulp which do affect the final DP of the cellulose acetate. The nature of these properties is not understood, but it can be shown that they are altered by changes in wood pulp processing. The ratio of hemicellulose in pulp to hemicellulose in the acetate made from it varies from one pulp to another. The acetate solution properties of the cellulose are more closely related to the composition of the acetate flake than to the composition of the original pulp. The viscosity of the concentrated acetate dope can be accounted for almost entirely in terms of flake DP and flake mannan content. Pulp manufacturers and acetate producers should strive toward conditions which increase the rate of degradation of the he mi celluloses and decrease the rate of cellulose degradation during acetylation.

Estimation of Z‐average molecular weight by light scattering in good solvents

Abstract: The unimodal molecular weight distributions found in many synthetic polymers can be fairly well characterized by the measurement of three different average molecular weights: the number-, weight-, and Z-averages. The first two are easily measured; rhe third is more difficult to determine. This paper describes a light-scattering method for the estimation of the Z-average molecular weight which is applicable to random coil polymers in the molecular weight range between 0.2 × 106 and 2 × 106 and for which a suitable solvent is available. The solvent may be either a theta solvent or a good solvent for which the relation between intrinsic viscosity and molecular weight is known. The method depends on the fact, that, when angular dissymmetry of scattered light is observed, the Z-average mean-square radius of gyration of the polymer molecules, r2 can be determined. In a thela solvent, r2 is proportional to M, so that measurement of rz2 is equivalent to the determination of Mz. In good solvents, r2 = kMb, and the corresponding molecular weight, called Md, is slightly different. The parameters k and b are derived from the relation between intrinsic viscosity and molecular weight in the test solvent. Provided that the dispersion of molecular weight is not too great, Md does not differ widely from Mz, as b cannot exceed 1.2 for random coil molecules. Experimental tests in toluene on narrow molecular weight fractions of poly-a-methyl-styrene and binary combinations of them gave results in accord with theory.

Degradation by mechanical action as a means of studying cellulose‐water interaction

Abstract: A number of samples containing varying amounts of lignin, xylan, and glucomannan were prepared from spruce wood and subjected to vibratory ball-milling m water, methanol, and benzene. The results support the idea that the main factor determining the extent of cellulose degradation by mechanical action is the plasticity of the substrate It was found that the presence of lignin greatly increases the extent of degradation probably as a result of mechanical restriction of the deformability of the system. On the other hand, glucomannan reduces the extent of degradation, while xylan has little or no effect. This is believed to be due to the location of part of the glucomannan inside the microfibril, where it is able to facilitate the penetration of water. Xylan is located on the outside of the fibril where a similar effect could not occur. It m also shown that the moisture regain of the samples varies with the hemicellulose content, but that the contribution of the glucomannan is less than that of the xylan This indicates that the mobility of the glucomannan molecule is restricted, presumably by its close association with the cellulose in the microfibril.