Hydration Mechanism of Polysaccharides:A Comparative Study
S. DESPOND,1,2 E. ESPUCHE,1 N. CARTIER,2 A. DOMARD1
1Laboratoire des Materiaux Polyme`res et des Biomateriaux, Unite Mixte de Recherche 5627, Batiment Institut Sciences etTechniques de IIngenieur de Lyon, 15 Bd A. Latarjet, 69622 Villeurbanne Cedex, France
2Ahlstrom Research, Zone Industrielle de lAbbaye, Impasse Louis Champin, 38780 Pont-Eveque, France
Received 18 March 2004; revised 23 July 2004; accepted 25 July 2004DOI: 10.1002/polb.20277Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: Water sorption was studied at 20 C on lms composed of different naturalpolymers. Three polysaccharides were investigated: chitosan, cellulose, and alginate.The major differences between these polymers, from a structural point of view, lay inthe substitution of an OH group by an NH2 function for chitosan and by an ionicCOONa group for alginate. An analysis of the experimental water sorption iso-therms, expressed as the number of water molecules sorbed per repeating unit in theamorphous phase, associated with an analysis of the enthalpy prole related to thewater sorption allowed us to propose a water sorption mechanism in two steps for allthe polymers: water sorption on polymer-specic sites in the rst step and waterclustering around the rst sorbed water molecules in the second step. It was deter-mined that two water molecules interacted with the polymer chains for cellulose andchitosan, whereas four water molecules were bonded to alginate chains. The specicsorption sites were identied as OH groups for cellulose, OH and NH2 groups forchitosan, and ionic and OH groups for alginate. A systematic reduction of the half-sorption time was observed in the activity range corresponding to this rst sorptionstep, and it was explained by a water plasticization effect. On the other hand, anincrease in the half-sorption time was observed in the second sorption step, at a highactivity (0.8), for chitosan and alginate. A modelization associating the GuggenheimAndersonde Boer model and the clustering theory, applied to our systems, allowed usto relate the occurrence of this last phenomenon to the formation of water clusterscontaining more than two water molecules. 2004 Wiley Periodicals, Inc. J Polym Sci PartB: Polym Phys 43: 4858, 2005Keywords: alginate; cellulose; chitosan; water sorption; hydrophilic polymers; biode-gradable polymers
Biopolymers are in strong demand for applica-tions such as edible packaging,1,2 membranes,3,4
and aqueous efuent treatments.5,6 Their biode-
gradable character, associated with the presenceof specic interaction sites in their structure,make them very attractive. Indeed, the polargroups and even in some cases the ionic groups7
present in these polymers are the reason for thehigh cohesive energy density, which represents adetermining parameter for some applications, es-pecially packaging. All these functional groupscan also be used to modify the polymers and in-crease their interaction capacity toward well-de-
Correspondence to: E. Espuche (E-mail: email@example.com)Journal of Polymer Science: Part B: Polymer Physics, Vol. 43, 4858 (2005) 2004 Wiley Periodicals, Inc.
ned species. This possibility is particularly inter-esting for water depollution, for example.8,9 How-ever, these functions are also the reason for thewell-known hydrophilic character of many naturalpolymers. In most applications, the ability of a poly-mer to swell in water and the hydration kinetics areof great importance. Fringant et al.10 reported thatthe water uptake greatly depends on the polymerstructure, but no data were reported concerning thewater sorption kinetics. It seems, therefore, essen-tial to have a better understanding of the watersorption phenomenon from thermodynamic and ki-netic points of view.
In this work, we have chosen to compare thewater sorption behavior of polysaccharide sam-ples differing in their chemical structures. Cellu-lose, chitosan, and alginate were all investigatedin the form of lms. The sorption kinetics and thesorption isotherms were determined and ana-lyzed as a function of the polymer structure.
Water sorption in hydrophilic polymers is usu-ally a nonideal process leading to plasticizationand/or clustering phenomena and, as a result,to complex BrunauerEmmettTeller (BET)type III11,12 or sigmoidal isotherms.13,14 Somemodels have already shown their ability to de-scribe such experimental isotherms.1520 Amongthem, we have chosen the GuggenheimAnder-sonde Boer equation1820 because it has beenalready used for natural polymers21 and can pro-vide information about clustering.11
Materials and Film Processing
Table 1 shows the structures of the different poly-mers used in this study: cellulose, chitosan with alow degree of acetylation (DA 1.6%), and algi-nate with a mannuronic/guluronic residue ratio of0.8. Cellulose and chitosan belong to the family oflinked ,1,4-polysaccharides, whereas alginate iscomposed of both ,1,4- and ,1,4-linkages. Themajor differences between these polymers, from astructural point of view, lie in the substitution ofan OH group by an NH2 function for chitosan andby a COONa group for alginate.
The main characteristics of the different poly-mers and the process used to prepare the lmsare described next.
Cellulose was studied in the form of parchmentpaper. The paper was composed of pure cellulosic
wood bers without any additive included in thenal brous dispersion (i.e., 100% cellulose modelpaper). Blotting paper was rst realized, and thenparchment paper was obtained by the immersionof this initial paper in a concentrated sulfuric acidsolution. Fibers were then partially dissolved,and this formed a cellulose gel that precipitatedand partially lled the paper pores. Figure 1 pre-sents scanning electron microscopy pictures of theinitial blotting paper and the correspondingparchment paper. The thickness of the papersheet was 40 m. We focused our study on theparchment paper because it was less porous thanthe blotting paper and, therefore, correspondedbetter to the lms obtained with the other poly-mers.
Chitosan was provided by France Chitin. Theakes were puried through dissolution in a stoi-chiometric amount of acetic acid, and the solution(1%) was ltered through 0.22-m-pore mem-branes (Millipore). Aqueous ammonia was thenadded to precipitate the polymer. After severalwashings with deionized water up to a neutralpH, the product was dried overnight under re-duced pressure. DA was determined by 1H NMRspectroscopy according to the method proposed byHirai et al.22 and was found to be 1.6%.
The average molecular weights were deter-mined by means of size exclusion chromatography(SEC) with an online multi-angle laser light scat-tering detector. The weight-average molecularweight (Mw) was 200,000 g mol
1, and thenumber-average molecular weight (Mn) was165,000 g mol1 (with a precision of 5%).
Chitosan lms were prepared via the casting ofa 1% (w/w) aqueous acetic acid solution contain-ing chitosan onto a polystyrene plate. After dry-ing at room temperature, the obtained 20-m-thick lms were immersed in an ammonia/meth-anol solution for 15 min for neutralization. Theywere then rinsed with water and dried. Chitosanlms in the free amine form were thus obtained.The crystallinity of the lms was determined byX-ray diffraction with ltered Cu K radiationgenerated at 30 kV and 5 mA. The degree ofcrystallinity was 40 5%.
This polysaccharide was composed of polymerblocks of 1,4-poly(-D-mannuronic acid), 1,4-
HYDRATION MECHANISM OF POLYSACCHARIDES 49
poly(-L-guluronic acid), and segments of alter-nating D-mannuronic and L-guluronic acid resi-dues.
The sodium alginate was provided by FMCBiopolymer AS. The composition of the copolymer[80% mannuronic sequences (M)] was determinedby 1H NMR. The guluronic acid units (G) wereessentially located in sequenced MGMGMGblocks. Mw and Mn were determined by SEC: Mw 182,000 g mol1 and Mn 85,500 g mol
1.Alginate lms were prepared via the casting of
a 1% (w/w) aqueous sodium alginate solution ontoa polystyrene plate. After drying at room temper-ature, the lms were analyzed by X-ray diffrac-tion. The degree of crystallinity was 35 5%.
Water Sorption Apparatus
The apparatus used for the water sorption studiesconsisted of a Setaram B92 microbalance and a
microcalorimeter. The balance precision was5 g.
Two samples of exactly the same weight wereused. One sample was added to the microbalance,and the other was added to the microcalorimeter.After desorption in vacuo (2.106 mbar) at a con-stant temperature (20 1 C), a partial pressureof water was established within the apparatus bymeans of an evaporator placed at temperature T.The water uptake at time t (Mt) was followed untilthe equilibrium sorption (M) was attained, andthe thermal changes associated with the wateruptake were recorded. The exothermic peak wasintegrated and corrected with values of the en-ergy obtained under the same conditions withempty boats. The ratio of the obtained energy tothe water uptake gave the interaction enthalpy(kJ mol1). It was representative of the internalenthalpy change of one molecule from the gaseousstate to the sorbed state.
Table 1. Chemical Structures of the Polymers
50 DESPOND ET AL.
The partial pressure was then increased in suc-cessive steps through discrete changes in temper-ature T from 12 to 20 C, and so over the entirerange of activity, the kinetics of water sorption,the sorption isotherm, and the interaction en-thalpy prole were obtained.
Methodology Used To Analyze the Water SorptionData
The water sorption results were examined as afunction of the activity and the considered poly-mer in three main ways.
First, the sorption data were tted to an em-pirical equation:
The n value indicates the type of diffusion mech-anism; there are three possible cases. For the rst
one, corresponding to Fickian transport, the rateof diffusion is much lower than the rate of relax-ation, and n is equal to 0.5. For the second one,the diffusion is very fast, contrary to the rate ofrelaxation, and n is 1. The third case correspondsto anomalous diffusion with n values lying be-tween 0.5 and 1.
Second, the sorption rates were estimated inthe range of all partial pressures via the half-sorption time (t1/2).
Third, the water isotherms were determinedand analyzed with respect to the enthalpy prole.Considering, as generally assumed, that the crys-talline parts were impermeable to small mole-cules, we also presented the isotherms for eachpolymer as the number of water molecules sorbedper repeating unit in the amorphous phase. Thisrepresentation allowed us to discuss the role ofthe polymer chemical composition in the watersorption mechanism. At last, the experimentalisotherms were modeled with the Guggenheim-Anderson-de Boer (GAB) equation.1820 Thisequation has been widely used to describe watersorption in foods and natural polymers21,23,24:
c aGAB cP,GAB kx1 kx
11 cP,GAB 1kx
According to this equation, the water concentra-tion (c) is related to the Guggenheim constant(cP,GAB), to the water concentration correspondingto the saturation of all primary adsorption sitesby one water molecule (formerly called the mono-layer in BET theory; aGAB), and to a factor cor-recting the properties of the multilayer moleculeswith respect to the bulk liquid k value. Bizotsmethod25 was used for calculating the parametersof the model. The curve-tting efciency was es-timated from the residual sum of squares (RSS):
RSS(yexp ycalc)2 (3)where yexp and ycalc are the reported experimentalvalues and the corresponding calculated values,respectively.
The GAB model is also interesting because itcan provide more information about water clus-tering. A positive deviation of M from Henryslaw sorption is generally interpreted by a cluster-ing tendency of the penetrant in the polymer ma-terial. The cluster integral G11 can be calculatedwith the following equation26:
Figure 1. Scanning electron microscopy photographsof (a) blotting paper and (b) parchment paper.
HYDRATION MECHANISM OF POLYSACCHARIDES 51
where 1, 1, and a1 are the molecular volume,volume fraction, and activity of component 1, re-spectively, and P and T are the pressure andtemperature, respectively.
The average number of solvent molecules in acluster (Nc) is dened as follows:
For an ideal solution, there is no clustering, andNc is equal to 1.
According to Zhang et al.s study,11 Nc valuescan be deduced from the GAB model parameterswith the following equation:
Nc 1 1 1aGAB cP,GAB (2 cP,GAB kx 2kx cP,GAB2)1 (6)
Nc values were determined and examined for allthe polymers.
Diffusion Mechanism and Sorption Rates
The n values representative of the sorption mech-anism were determined at various activities forchitosan and alginate lms (Table 2).
For chitosan, n values close to 0.5 were ob-tained for activities lower than 0.3. In this do-main, the diffusion mechanism can be consideredFickian. For higher activity, n increased progres-sively from 0.5 to 0.85, and the diffusion mecha-nism became anomalous.
For alginate, n values were higher than 0.5even at very low activities. As for chitosan, nincreased with the activity and reached a plateauaround 0.85 for activities higher than 0.43. Thus,a Fickian diffusion domain could not be deter-mined for alginate.
In conclusion, anomalous sorption mechanismswere observed over a wide range of activities forthese natural polymers.
As we demonstrated that the diffusion mecha-nism was not Fickian over the entire activityrange, we did not calculate the diffusion coef-cients but instead chose to compare t1/2. For thecomparison of the results, and because the thick-ness of the two lms (chitosan and alginate) wasequal to 20 m, all the data were expressed forthis thickness. The results are presented inFigures 2(a) and 3(a) for chitosan and alginateand in Figure 4(a) for paper.
Two main observations can be drawn from theexperimental curves with respect to chitosan andalginate:
1. For each of these polymers, the variation oft1/2 as a function of the activity is not mo-notonous, and three domains can be deter-mined. In the rst domain, correspondingto a low activity (domain I), a decrease int1/2 can be observed. For an intermediatevalue of the water activity aw (domain II),no great variations of t1/2 can be observed,and at a high activity (domain III), an in-crease in t1/2 can be observed.
2. In domains I and II, the t1/2 values mea-sured for chitosan and alginate are quitesimilar. On the other hand, in domain III,the t1/2 values measured for alginate lmare higher than those determined for chi-tosan.
A particular behavior has been noticed for pa-per. The kinetics are very fast, and only a smalldecrease in t1/2 can be observed as the activityincreases [Fig. 4(a)].
Sorption Isotherm and Interaction Energy
The water sorption isotherms are BET type II forall the st...