Introduction

Polymer hydrogels are macromolecular networkswhich imbibe large amounts of water without dis-solving, and because of this they have found a largenumber of applications, ranging from agriculture tomedicine and pharmacy [1–3]. Their good biocompat-ibility and water permeation properties are the basis ofthese applications. The main handicap of hydrogels istheir poor mechanical properties, mainly in the swollenstate. Consequently, reinforcement is needed for theirapplications. Increasing the cross-linking density,copolymerization with a more resistant hydrophobicnetwork and crystallization (when possible) are amongthe ways to reinforce hydrogels as well as the forma-

tion of interpenetrating polymer networks [4], inter-penetrated polymer networks (IPNs), with twopolymers: a hydrophilic one and a hydrophobic one,which is the option chosen in the present work. It isknown that combination of a hydrophilic and ahydrophobic polymer forming microphase-separatedstructures is a way for reinforcing hydrogels [5]; thismethod takes advantage of the behaviour of hydro-phobic polymer in the presence of water. The hydro-phobic domains act as additional linking points in thehydrogel network, improving mechanical properties.Besides, it has been shown [6] that the alternatingmicrodomain structure shows excellent antithrombo-genic properties, and blood compatibility of theresulting materials is enhanced.

G. Gallego Ferrer

J. M. Soria Melia

J. Hernandez Canales

J. M. Meseguer Duenas

F. Romero Colomer

M. Monleon Pradas

J. L. Gomez Ribelles

P. Pissis

G. Polizos

Poly(2-hydroxyethyl acrylate) hydrogelconfined in a hydrophobous porous matrix

Received: 6 May 2004Accepted: 16 August 2004Published online: 22 October 2004� Springer-Verlag 2004

Abstract A series of interpenetratedpolymer networks (IPNs) in whichthe first component is a porouspoly(ethyl methacrylate) (PEMA)hydrophobic network and the sec-ond one is a poly(2-hydroxyethylacrylate) (PHEA) hydrophilic net-work were synthesized. Equilibriumsorption isotherms can be reduced toa single master curve for all the IPNswhen the water absorbed is ex-pressed per gram of PHEA in them.The equilibrium water sorption inimmersion is always much smallerthan that of pure PHEA. This fea-ture is due to the confining effect ofthe stiff PEMA matrix. The plasti-cizing effect of the absorbed wateron the PHEA phase was character-ized using thermally stimulated

depolarization currents, dynamic-mechanical analysis and dielectricrelaxation spectroscopy. The resultsshow that the shift of the mainrelaxation peak towards lowertemperatures is unaffected by thepresence of the PEMA matrix, andonly depends on the water contentper gram of PHEA in the IPN.

Keywords Interpenetrated polymernetworks Æ Hydrogels ÆWater sorption isotherms ÆDynamic-mechanical analysis ÆPlasticization

Colloid Polym Sci (2005) 283: 681–690DOI 10.1007/s00396-004-1208-y ORIGINAL CONTRIBUTION

G. Gallego Ferrer Æ J. M. Soria MeliaJ. Hernandez CanalesJ. M. Meseguer DuenasF. Romero ColomerM. Monleon PradasJ. L. Gomez Ribelles (&)Centre for Biomaterials,Universidad Politecnica de Valencia,Camino de Vera s/n, 46071 Valencia, SpainE-mail: jlgomez@ter.upv.esTel.: +34-96-3877275Fax: +34-96-3877276

P. Pissis Æ G. PolizosDepartment of Physics,National Technical University of Athens,Zografou Campus, 15780 Athens, Greece

The structural heterogeneity of the IPNs commentedabove is due to the development of thermodynamicalinstability of the homogeneous mixture during poly-merization process, as a consequence of the growingmolecular weight of at least one of the components inthe case of sequential IPNs, those synthesized in thepresent work. The phase morphology of an IPN dependson several factors: miscibility of the components, com-position, cross-linking density and the kinetic details ofthe reaction. Although all these factors are importantwhen a sequential IPN is prepared, it has been verifiedthat the cross-linking density of the first network is thedeterminant factor in the morphology that is finallyobtained [7, 8]. When one of the networks is a hydro-philic polymer, the resulting IPN can form hydrogelswhen swollen in water, and the properties sought for inapplications, which rest upon water uptake and trans-port inside the hydrogel, may be affected by the phasemicromorphology of the IPN. In this paper we study asequential IPN formed by a first hydrophobous porouspoly(ethyl methacrylate) (PEMA) network and a secondhydrophilic poly(2-hydroxyethyl acrylate) (PHEA) net-work polymerized in the presence of different quantitiesof solvents (PEMA(sponge)-i-PHEA IPNs). We con-sider the effect of changing IPN composition upon waterdiffusion and sorption, and on the phase organizationand the morphology of the first porous network, whichcan be inferred from these data and from thermallystimulated depolarization currents (TSDC), dynamic-mechanical analysis and dielectric relaxation spectros-copy (DRS) at different water contents in the samples.

Previous work on hydrogel-forming networks refersmostly to poly(hydroxyethyl methacrylate) [9], poly-acrylamide, poly(acrylic acid) [10] and poly(vinyl alco-hol). PHEA is more hydrophilic than poly(hydroxyetylmethacrylate). The innovation of this study in relationto the previous work is that the hydrophilic component,the PHEA network, is confined in a rigid (at roomtemperature) sponge prepared with different quantitiesof diluent, and it has been possible to relate the mor-phology of the confining sponge with the sorptionproperties of the resulting IPNs.

Materials and methods

Different sequential IPNs of porous PEMA,PEMA(sponge), and PHEA were polymerized with thesame quantity of cross-linker in both networks (1% ofethylene glycol dimethacrylate, EGDMA, relative tomonomer weight). The first network was polymerized ina porous form using different quantities of ethanol asporogen. Two types of solutions of EMA monomer(Aldrich 99% pure) with a 1% by weight of EGDMA ascross-linking agent, a 2% by weight of benzoin asphotoinitiator and two different ethanol contents 40 and

60% by weight (relative to monomer/diluent mixture)were prepared and polymerized at room temperatureunder ultraviolet radiation for 24 h between two glassplates. The resulting PEMA porous networks wereboiled in ethanol for 1 day to extract the remaining lowmolecular weight substances. The rinsed samples wereintroduced and swollen to equilibrium in three types ofsolutions: the first one contained HEA monomer (Al-drich 96% ethanol), a 1% by weight of EGDMA and a0.2% of benzoin; the second and third ones consisted ofa mixture of HEA monomer (with the same quantitiesof cross-linker and initiator as the first one) and a 60 or80% by weight (relative to monomer/diluent mixture) ofethanol, respectively. Then, polymerization of the sec-ond network took place as explained for the first one.Low molecular weight substances were extracted inboiling ethanol for 1 day and the resulting IPNs weredried in vacuum to constant weight firstly at 120 �C andafterwards at 170 �C. The weight fraction of the PHEAnetwork in the IPNs, xPHEA, is shown in Table 1.

Water sorption isotherms for the IPNs were deter-mined at 45 �C by the standard method [11, 12] ofequilibrating the samples in ambients of controlled andknown relative humidities (RH): the samples wereallowed to equilibrate to constant weight in variousthermostatized sealed dessicators over saturated watersolutions of different salts in an oven. A Sartorius A200Sbalance with 10)4 g sensitivity was employed for thesemeasurements.

Water diffusion in immersion in liquid water for theIPNs was analysed at 45 �C by placing the dry samplesin liquid water and weighing them at selected times aftercarefully drying their surface with filter paper.

Thermally stimulated depolarization currents of thehydrated samples were measured using a Keithley 610electrometer [13, 14]. Small discs of the samples of15 mm in diameter and around 1 mm in thickness wereinserted at room temperature between the plates of acapacitor and polarized by the application of a DCelectric field between 300 and 600 V (depending on thethickness of the sample) at room temperature for 5 min.With the electric field still applied, the sample wasquenched to )170 �C and short-circuited. The depolar-ization current was measured on heating at a rate

Table 1 Composition of the PEMA(sponge)-i-PHEA IPNs

IPN Weight fraction ofethanol in thepolymerization ofthe PEMA network

Weight fractionof ethanol in thepolymerization ofthe PHEA network

Weight fractionof PHEA in theIPN, xPHEA

40/80 0.4 0.8 0.1540/60 0.4 0.6 0.4540/– 0.4 0 0.7160/80 0.6 0.8 0.3460/60 0.6 0.6 0.6460/– 0.6 0 0.81

682

ranging between 3 and 4 �C min)1 until roomtemperature.

Dynamic-mechanical spectroscopy (DMS) of the dryand swollen samples was performed in a Seiko DMS 210apparatus at the frequency of 1 Hz in the tension mode.The temperature dependence of the storage modulus(E¢), loss modulus (E¢¢) and loss tangent (tan d) wasmeasured from )130 to 170 �C at a rate of 2 K min)1.The samples were prismatic, approximately 10 · 4.5 ·0.9 mm3.

Dielectric relaxation spectroscopy was performedwith a Schlumberger Frequency Response Analyser FRASI1 260. The sample was placed between two goldenelectrodes in a dielectric cell, which was introduced insidea cryostatic Novocontrol system. The samples weremeasured between 0.1 and 106 Hz every 10� in theappropriate temperature range for each sample.

Results

Figure 1a shows the experimental sorption isotherms(water uptake of the polymer, expressed as mass ofwater gained per unit mass of dry polymer, w=m1/m2,where subindex 1 corresponds to water and subindex 2to the polymer) as a function of the water activity in thegel, a, for the PEMA(sponge)-i-PHEA IPNs and for thePHEA homonetwork. As explained above the experi-mental sorption isotherms are determined by equili-brating the samples in ambients of controlled andknown RH and measuring their water uptake gravi-metrically. The relative humidity in air atmosphere atpressure p and temperature T is the ratio of the partialpressure of the water vapour in the atmosphere, y1Æp,and the pressure of pure water at T, p1

v (T),

RH ¼ y1ppv1

(y1 is the molar fraction of water in the atmosphere).The sorption experiments are conducted at atmosphericpressure, p=1 atm. At this pressure the gas mixtureforming the atmosphere can be regarded as ideal, andthe partial pressure of whatever its components is itsfugacity: y1p ¼ f v

1 ð1 atm; T Þ: On the other hand, at1 atm and T the fugacity of pure liquid water coincideswith its vapour pressure: f1

L=p1v(T), and thus the rel-

ative humidity turns out to be the activity of water in theatmosphere, with reference to liquid water

RH ¼ f v1

f L1

¼ av1:

Equilibrium of the gel in the gaseous environmentdemands equality of the chemical potential of water inthe gel and in the environment:

lgel1 ¼ lv

1;

and this equality turns, by l1 ¼ l1 þ RT ln a1; into an

equality between activities agel1 ¼ av

1; such that,

agel1 ¼ RH;

and the experimental isotherm determines, at eachtemperature, the relationship between composition andactivity

w ¼ wða; T Þ;

with a ¼ agel1 ; for the sake of clarity.The sorption isotherms in Fig. 1a exhibit the convex

shape characteristic of type III isotherms in the Bru-nauer classification [15]. For each constant water activitythe water content in the IPNs increases with increasingmass fraction of the PHEA component in them.

a

w

0

0.1

0.2

0.3

0.4

0

0.1

0.2

0.3

0.4

0 0.2 0.4 0.6 0.8 1

0 0.2 0.4 0.6 0.8 1

a

w'

(a)

(b)

Fig. 1 Sorption isotherms at 45 �C for the PEMA(sponge)-i-PHEAIPNs (a). The Fig. b shows the isotherms displayed as mass fractionof sorbed water per unit PHEA mass in the IPN, w¢, against wateractivity, a. The samples are identified by their PHEA mass fractionequal to: 1 PHEA network (filled diamond), 0.81 (open square), 0.71(open diamond), 0.64 (filled triangle), 0.45 (open circle), 0.34 (filledcircle), 0.15 (open triangle) and 0 PEMA network (·)

683

In order to study the water transport properties in thehydrogels dynamic sorption, experiments in immersionin liquid water were performed and are shown in Fig. 2,where the water uptake of the polymer, w, is representedas a function of time, t, in a logarithmic scale. The plotshows that the equilibrium water content increases as themass fraction of PHEA in the IPNs is increased.

Thermally stimulated depolarization current experi-ments were carried out on all IPNs with different watercontents, and the results for all samples were similar. Asan example Fig. 3 shows the spectra of the IPN withPHEA mass fraction, xPHEA, equal to 0.81. The depo-larization current has been normalized to

In ¼I � d

V � A � b ;

where I is the depolarization current in pA, d is thethickness of the sample in mm, V is the voltage in Volt,A the area of the sample in mm2 and b the heating rate inK min)1. This normalization makes it possible to com-pare thermograms obtained on samples of differentthicknesses, with different voltages, areas and heatingrates [14]. These spectra are very similar to those of purePHEA [16]. For the lower water mass fractions threerelaxations are observed, which have been called, inorder of increasing temperature, the c, bsw and a relax-ations following the references [17, 18]. As the magni-tude of the single relaxation observed in the PEMAhomopolymer in the temperature range of measure-ments in Fig. 3 (results not shown) is very small whencompared with those of PHEA homonetwork [16] andthose of the PEMA(sponge)-i-PHEA IPNs, these three

relaxations have been attributed to the PHEA phase inthe IPNs. The a relaxation, which occurs at highertemperatures, is ascribed to the main relaxation of thePHEA phase in the IPNs. This peak shifts to lowertemperatures as the water content increases, and for thelargest water contents it splits into two peaks, a and q.The origin of this new q peak is thought to be due tospace charges, and its intensity depends on numerousfactors such as the type of electrodes, impurities in thesample, etc. [14]. The c relaxation, which shows uparound 130 K, has been attributed to internal motionsin the side chain of PHEA phase, which orient the dipoleof the hydroxyl group [19]. The magnitude of thisrelaxation decreases with increasing water content, andfrom a certain humidity on it completely disappears. Itstemperature shows no dependence on water content andremains constant. The bsw relaxation happens at tem-peratures immediately higher than the c, around 175 Kfor the scan with the lowest water content of the gel.With increasing amounts of water in the sample, thispeak increases in magnitude and shifts to lower tem-peratures, until it completely overruns the c peak.

The dynamic-mechanical experiments show in all thesamples the presence of two a relaxations correspondingto the PEMA and PHEA phases. As a representativeexample, Fig. 4 shows the temperature dependence ofthe real part of the elastic modulus (Fig. 4a) and the losstangent (Fig. 4b) of the IPN with PHEA mass fraction,xPHEA, equal to 0.81 at different water contents. Whenthe sample is nearly dry the peak of tan d correspondingto the a relaxation of the PEMA phase appears at 102 �Cwhile that corresponding to PHEA appears at 39 �C.

0.0

0.2

0.4

0.6

0.8

1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07t (s)

w

Fig. 2 Water uptake in immer-sion as a function of time (in alogarithm scale) at 45 �C for thePEMA(sponge)-i-PHEA IPNs,identified by their PHEA massfraction equal to: 0.81 (opensquare), 0.71 (open diamond),0.64 (filled triangle), 0.45 (opencircle), 0.34 (filled circle), 0.15(open triangle) and 0 PEMAnetwork (·)

684

At lower temperatures the secondary c and bsw relax-ations take place.

The samples containing absorbed water were rapidlycooled down inside the equipment to )130 �C. Noevaporation of water from the sample up to room tem-perature during the measuring scan is expected, but athigher temperatures water is rapidly lost by the sampleand the results in the temperature interval of the mainrelaxation of PEMA are not representative of the wetsamples.

As the amount of water increases the a relaxationcorresponding to the PHEA phase shifts towards lowertemperatures due to the plasticization of this phase.However, at the highest water contents, above 0.3 g ofwater per gram of PHEA in the sample, the temperatureof the peak starts increasing. The loss tangent plots showa shoulder at the low-temperature side of the main peakof the PHEA phase that must be ascribed to the bsw

relaxation: the overlapping of the a relaxation and thescattering of the results in some cases prevents a detailedanalysis of the secondary dynamic-mechanical relax-ations, although they play an important role in themechanical behaviour of the IPNs, as proved by thesignificant decrease of the elastic modulus with increas-ing temperature below )50 �C (Fig. 4a). A shoulder inthe high-temperature side of the PHEA a relaxation canbe due to the loss of water. Again, the E¢ plots are onlysignificant up to room temperature in the wet samples.They show the plasticization of the PHEA phase up to a

limit of water content, and then the main relaxationshifts towards higher temperatures and at the same timethe values of the modulus in the glassy state are clearlyhigher than in samples with lower water contents. In thecase of the sample whose spectra are shown in Fig. 4a,this behaviour corresponds to the sample containing0.25 g of water per gram of sample, or 0.31 g of waterper gram of PHEA.

Figure 5 shows the dielectric results obtained in anIPN with PHEA mass fraction equal to 0.15. The studyof the influence of water on the a relaxations of thesystems is difficult because of the overlapping of theionic conductivity with the dipolar contributions corre-sponding to the conformational motions. In fact in allthe samples containing more than 15 wt% of PHEA, thea relaxation of both phases was completely covered bythe space charge contributions. In the case of the sampleof Fig. 5 the plasticization of the PHEA phase is clearlyobserved, and the peak shown by the imaginary part ofthe dielectric permittivity shifts clearly towards lowertemperatures as the water contents increases; but even inthis case it is difficult to determine the temperature of themaximum because of the conductivity contribution atthe high-temperature side of the peak. At low tempera-tures the bsw shows the same behaviour described abovefor the TSDC results; these results correspond to thesample with PHEA mass fraction, xPHEA, equal to 0.81,because the effect of water to the main and to the bsw

relaxations is more clearly seen in this IPN than in that

1.E-06

1.E-05

1.E-04

1.E-03

100 150 250

T (K)

I n (

pA

. min

. V-1

. mm

-1. K

-1)

200

Fig. 3 TSDC thermogram ofthe IPN with PHEA mass frac-tion equal to 0.81 (PEMAsponge synthesized with 60%ethanol and immersed in PHEAmonomer) with different watercontents: w=0.0066 (filled cir-cle), w=0.0290 (open square),w=0.0349 (*), w=0.0366 (opendiamond), w=0.0814 (filled tri-angle), w=0.1288 (open trian-gle), w=0.2521 (open circle)

685

with lower PHEA mass fractions. The height of the peakincreases and the temperature diminishes with increasingwater content.

Discussion

Figure 1b shows the equilibrium sorption isotherm ofthe IPNs when the water content is referred to the massof PHEA in them, that is w¢=w/xPHEA. This plot shows

that the different experimental isotherms of the IPNssuperpose on the isotherm of pure PHEA. This reduc-ibility of the isotherms has been found in other systems[20] and is interpreted to mean that, first, water is sorbedessentially in the hydrophilic phase of the IPNs, and,second, that the behaviour of this phase in the IPNs isthat of the pure PHEA homonetwork. In order for thisbehaviour to be possible, it is necessary to have phaseseparation, otherwise the behaviour of the hydrophilicphase would not be that of the pure PHEA, due to thehydrophobic interaction with the PEMA chains. In factas the PEMA network is produced with a porous mor-phology, the phase separation is expected to happen ona larger scale than in conventional IPNs [11, 20], whichmeans that great hydrophilic domains will exist in thepores of the first network, with a behaviour similar tothat of the PHEA network. The reduced curves inFig. 1b start to diverge for water activities around 0.8:from that activity on the IPNs absorb less water thandoes the pure PHEA homopolymer. In the PEMA(s-ponge)-i-PHEA IPNs the PHEA phase is confined by arigid PEMA network, which hinders the PHEA expan-sion and does not allow the PHEA phase to accommo-date the same amounts of water.

The dynamic sorption process in immersion in liquidwater is not expected to be a Fickian process due to thegreat volume increase of the samples during the sorp-tion experiment. Nevertheless, an apparent diffusion

5

6

7

8

9

10

-130 -80 -30 20 70 120 170

Temperature (°C)

log

(E

' /

Pa

)

0.01

0.1

1

-130 -30 70 170

Temperature (°C)

tan

δ

(a)

(b)

Fig. 4 Temperature dependence of the real part of the elasticmodulus (a) and the loss tangent (b) measured at 1 Hz in the IPNwith PHEA mass fraction equal to 0.81 (PEMA sponge synthesizedwith 60% ethanol and immersed in PHEA monomer) with differentwater contents: (·) w=0.011, (open circle) w=0.026, (open triangle)w=0.069, (+) w=0.130, (open square) w=0.180, (filled square)w=0.434, (filled diamond) w=0.865

0.001

0.01

0.1

1

-100 -50 0 50 100

T (°C)

ε"

Fig. 5 Imaginary part of the complex dielectric permittivity e¢¢measured at 100 Hz in the IPN with PHEA mass fraction equal to0.15 (PEMA sponge synthesized with 40% ethanol and immersedin the mixture PHEA diluted with 80% of ethanol) with differentwater contents: (filled circle) w=0, (open triangle) w=0.0074, (filledtriangle) w=0.0194, (open square) w=0.0290, (filled square)w=0.0513. The curve corresponding to the dry pure PEMAnetwork polymerized with 40% ethanol as diluent is also shown forcomparison (open circle)

686

coefficient, Dap, can be calculated according to Fick’sdiffusion equation in order to characterize numericallythe rate of the sorption process in the different IPNs.According to Fick’s equation

@c@t¼ Dap

@2c@x2

;

where c is the concentration of water in the sample, x isthe distance and t is the time. The solution of thisequation in the case of diffusion through a sheet withthickness l for short times corresponding to Dmt/Dm¥<0.6 (where Dmt and Dm¥ are the mass gains of thesample at time t and at equilibrium) can be approxi-mated by [21]

Dmt

Dm1¼ 4� Dap � t

p � l2

� �12

: ð1Þ

Thus, the plot of Dmt/Dm¥ against t1/2/l must be lin-ear in the initial stage of the sorption process. If K de-notes the slope of this linear plot, from Eq. 1

Dap ¼p16� K2:

Figure 6 shows the apparent diffusion coefficients forthe IPNs as a function of the mass fraction of PHEA inthem determined by this procedure. Two apparent dif-fusion coefficients are represented for xPHEA=0; one ofthem corresponds to the PEMA sponge polymerizedwith a 40% of ethanol and the other to the PEMAsponge polymerized with a 60% of ethanol. The plotshows a clear increasing tendency of the value of theapparent diffusion coefficient as the mass fraction ofPHEA in the IPNs increases. This means that the morehydrophilic the sample, the more rapid is the diffusionprocess of liquid water in the sample. For PHEA massfraction lower than 0.4, the apparent diffusion coefficientof the IPNs is lower than that of the corresponding purePEMA sponge; this means that although these IPNs aremore hydrophilic than pure PEMA, the sorption processin immersion in water is slower. This behaviour suggeststhat the hydrophilic phase in these IPNs is in the form ofdispersed domains inside the PEMA sponge, and theloss of continuity of the hydrophilic phase makes thediffusion process more difficult.

As discussed above, the main peak of the TSDCthermograms of the IPNs corresponds to the mainrelaxation of the PHEA phase in them. The plasticiza-tion effect of water is manifested: the temperature of thispeak diminishes as the water content of the sample in-creases. In Fig. 7a the temperature of the a peak of theIPNs is represented as a function of the water contentreferred to the mass of PHEA in them, w¢=w/xPHEA.Again, a single reduced curve is obtained, which dem-onstrates that all hydrophilic phases of the IPNs are

equally plasticized. This result supports the hypothesisthat the system is a phase-separated one in which wateris mainly absorbed by the hydrophilic phase in it.

The same conclusion can be reached from thedynamic-mechanical results obtained with samples witha water content below 0.3 g of water per gram of PHEA,as shown in Fig. 7b. Further amounts of water do notcontribute to the plasticization of the PHEA chains.Differential scanning calorimetry (DSC) experiments[16, 22, 23] show that in pure PHEA network above thiswater content, a crystallization peak appears in coolingscans. Thus, at the beginning of the DMS measuringscan, the samples with w¢ ‡ 0.3 contain nanometric puresolid water domains dispersed in the polymer matrix. Attemperatures below 0 �C these domains are in the solidphase and act as a reinforcing disperse filler from themechanical point of view, whose effect is to shift theglass transition to higher temperatures and to increasethe elastic modulus. Immediately before the drop of E¢in the main transition, the value corresponding to thesample containing solid water domains can be up to fivetimes the modulus of the sample with a water contentaround w¢=0.3. This result strongly supports the con-clusions of the DSC experiments on the phase diagramof the PHEA-water system [16, 22, 23]. By comparingthe TSDC and the DMS results with each other (Fig. 7a,b) it is interesting to note the shift of the DMS data tohigher temperatures due to the higher frequency ofmeasurements (1 Hz against about 1 mHz [13, 14]).However, the range of temperature shift due to plasti-cization is the same by both techniques, as it should beexpected, approximately 90 �C.

The secondary relaxations of PHEA and theirevolution with increasing water content in the networkhave been explained elsewhere [11, 13, 19]. The same

0.0

1.0

2.0

3.0

0 0.2 0.4 0.6 0.8 1xPHEA

Dap

. 107 (

cm2 s-1

)

Fig. 6 Apparent diffusion coefficient of water in the IPNs as afunction of IPN composition (PHEA mass fraction in it, xPHEA)measured on an immersion experiment. The two points representedat xPHEA=0 correspond to the two types of the PEMA sponges

687

behaviour is found here for the hydrophilic phase of thePEMA(sponge)-i-PHEA IPNs. The temperature of the crelaxation is not affected by water content; this relaxa-tion is due to internal motions in the –CH2–CH2–OHgroups and shows up as long as there remain suchgroups free from association with water molecules. Theincorporation of water into the PHEA phase leads to theoccurrence of a new relaxation, labeled bsw, which wasinterpreted in the literature, is due to the motion of anew lateral group formed by the association of one

water molecule and two side-chain hydroxyl groups [16,17, 19, 24]. As shown in Fig. 3, a rapid increase of themagnitude of the bsw relaxation is observed at low watercontents, when the water molecules are mainly at thefirst sorption layer of the polymer. As it is only the firstsorbed water molecules that can associate with thehydroxyl groups, consequently, a slower increase isobserved for higher water fractions. The increase of thenumber of water molecules is probably accompanied bya weakening of interactions between lateral chains,which would thus confer more mobility to the relax-ational unit, shifting the temperature of the bsw relaxa-tion towards lower values. Figure 8 shows the evolutionof the temperature of this peak in all the IPNs as afunction of the water content referred to the mass ofPHEA in them, w¢. Once again a reduced single curve isobserved for all IPNs, which supports the hypothesisthat the behaviour of the hydrophilic phase in the IPNsis not affected by the presence of the hydrophobic phase.

If the PHEA phase behaviour of the IPNs is notaffected by the presence of the PEMA sponge when thewater activity of the samples is lower than one, a dif-ferent situation shows up when the samples are equili-brated in immersion in liquid water. The importance ofthe morphology of the PEMA sponge on the sorptioncapacity and expansion of the hydrophilic phase can beremarked in these experiments. Figure 9 shows theequilibrium water content of the IPNs referred to thePHEA phase when immersed in liquid water, w¢**,compared with that of pure PHEA when immersed in

170

190

210

230

250

270

0 0.1 0.2 0.3 0.4

w'

TαT

SD

C (K

)

200

220

240

260

280

300

320

0 0.2 0.4 0.6 0.8 1w'

TαD

MA

(K)

(a)

(b)

Fig. 7 Temperature of the TSDC peak (a) and DMS loss tangentpeak (b) corresponding to the main relaxation of PHEA, a, as afunction of water content of the hydrogel referred to PHEA weightin the sample, w¢. The samples are identified by their PHEA massfraction equal to: 0.81 (open square), 0.71 (open diamond), 0.64(filled triangle), 0.45 (open circle), 0.34 (filled circle), 0.15 (opentriangle)

135

145

155

165

175

0 0.1 0.2 0.3 0.4w'

Tβs

wT

SD

C (K

)

Fig. 8 Temperature of the TSDC peak corresponding to the bswsecondary relaxation of the PHEA phase in the IPNs as a functionof water content of the hydrogel referred to PHEA weight in thesample, w¢. The samples are identified by their PHEA mass fractionequal to: 0.81 (open square), 0.71 (open diamond), 0.64 (filledtriangle), 0.45 (open circle), 0.34 (filled circle), 0.15 (open triangle)

688

liquid water, w**PHEA, and in equilibrium in a saturatedvapour, w*PHEA (dotted lines). The difference found forw*PHEA and w**PHEA for the same value of wateractivity has been explained elsewhere [25]. The first valuecorresponds to saturation of the gel with its micelles orcollapsed heterogeneities unavailable for sorptionbecause the expansion of the network is insufficient toopen and disentangle those structures. The second valuecorresponds to saturation of the gel in a state wherethose structures have loosened and opened, and thenetwork can lodge a pure water phase in it because theelastic energy of network expansion in immersion inliquid water is enough to produce this effect. For IPNcomposition lower than xPHEA<0.5, the PHEA networkis thought to be dispersed in the PEMA sponge, and itswater sorption capacity in immersion does not exceedthe water content in saturated vapour equilibrium of thehomonetwork. When xPHEA>0.5 both phases (hydro-philic and hydrophobic) are thought to be cocontinuous,but, there exists a clear difference between the behaviourof the samples where the PEMA sponge is prepared witha 60% of ethanol and those synthesized with a 40% ofethanol. In the former the equilibrium water content inimmersion is higher than that of pure PHEA in equi-librium in saturated vapour, which means that thePHEA phase in the IPNs can expand to some extent,micellized chain segments and micropores can open anda liquid water phase can be lodged in the hydrogelphase. This means that the expansion of the network hasbeen able to overcome the confining effect of the PEMAsponge. This situation is only possible if the PEMAsponge is formed by a loosely connected structure of

merged microspheres; in the presence of this morphol-ogy, the expansion of the PHEA phase can distort thePEMA sponge and expand. When the amount of diluentused in the polymerization of the PEMA sponge is only40%, the expansion of the PHEA network is quite small,since the water content in immersion is similar to that ofpure PHEA in equilibrium in saturated vapour. Thissuggests that the morphology of the PEMA sponges issimilar to a honeycomb-like structure, as found also inother systems [26], which is much more rigid and pre-vents the expansion of the hydrophilic phase to someextent. This structural difference can also be related withthe spectra of the dynamic-mechanical measurementsfor the pure PEMA sponges (results not shown). Indeedthe rubber mechanical storage modulus for the PEMAsponge polymerized with a 60% of ethanol is lower thanthat for the sample prepared with a 40% of ethanol.Also the magnitude of tan d for the most porous spongeis higher after the main relaxation due to an additionalfriction mechanism when the pores are opened [26].

Conclusions

The PEMA(sponge)-i-PHEA IPNs are phase-separatedsystems as revealed by the reducibility of the equilibriumsorption isotherms when the water content is referred tothe PHEA mass fraction in them. The existence of this‘universal’ sorption isotherm means that water is sorbedessentially in the hydrophilic phase of the IPNs, and thebehaviour of this phase in the IPNs is similar to that ofthe pure PHEA homonetwork. However, in equilibriumin a vapour environment at the highest water activities, aconfining effect of the PEMA network is manifested, andthe water content of the hydrophilic phase in the IPNs islower than in the pure hydrophilic homonetwork. Theincreasing tendency of the value of the apparent diffu-sion coefficient as the mass fraction of PHEA in theIPNs increases shows that the more hydrophilic thematerial the more rapid is the diffusion process of liquidwater in the sample. For PHEA mass fraction lowerthan 0.4, the hydrophilic phase in the IPNs forms dis-perse domains inside the PEMA sponge as revealed bythe lower diffusion coefficients compared to that of thePEMA homonetwork. The evolution of the temperatureof the main relaxation, measured by TSDC and DMS,of all IPNs as a function of the water content in thesample referred to the mass fraction of PHEA in themalso reduces to a single curve, which confirms thebiphasic nature of the IPNs. DMS measurements showthat the samples with w¢ ‡ 0.3 contain nanometric puresolid water domains (crystallized at temperatures below0 �C) that act as a reinforcing dispersed filler from themechanical point of view, whose effect is to shift theglass transition to higher temperatures and to increasethe elastic modulus. When the samples are equilibrated

0

0.4

0.8

1.2

1.6

0 0.2 0.4 0.6 0.8 1xPHEA

w' *

, w' *

*

w*PHEA

w**PHEA

Fig. 9 Equilibrium water content of the PHEA phase in the IPNsin immersion in liquid water, w¢**, as a function of their massfraction of PHEA, xPHEA. The symbols represent: (filled circle) aseries of IPNs whose PEMA first network was polymerized with a60% of ethanol and (open triangle) a series of IPNs whose PEMAfirst network was polymerized with a 40% of ethanol. Thehorizontal lines represent the equilibrium sorption of pure PHEAnetwork by immersion in liquid water (w**PHEA) and in saturatedvapour (w*PHEA)

689

in immersion in liquid water, the importance of themorphology of the PEMA sponge on the sorptioncapacity and expansion of the hydrophilic phase can beremarked. For IPN composition lower than xPHEA<0.5,the PHEA network is dispersed in the PEMA sponge,and its water sorption capacity in immersion does notexceed the water content in saturated vapour equili-brium of the homonetwork. When xPHEA>0.5 bothphases (hydrophilic and hydrophobic) are thought to becocontinuous, but there is a clear difference between thebehaviour of the samples where the PEMA sponge isprepared with a 60% of ethanol and those synthesizedwith a 40% of ethanol. In the former, the PEMA sponge

is formed by a connected set of merged microspheres,and the PHEA phase in the IPNs can expand to someextent and overcome the confining effect of the PEMA.When the amount of diluent used in the polymerizationof the PEMA sponge is 40%, the morphology of thePEMA sponges is more similar to a honeycomb-likestructure, and the expansion of the PHEA network isalmost the same as the expansion of the PHEA homo-network when equilibrated in saturated vapour.

Acknowledgements This work was supported by the GeneralitatValenciana through project CTIDIA/2002/046 and by NTUAthrough the programme Thales.

References

1. Tsuruta T (1996) Adv Polym Sci 126:12. Hoffman AS (2002) Adv Drug Deliv

Rev 43:33. Kazanskii KS, Dubrovskii SA (1992)

Adv Polym Sci 104:974. Klempner D, Sperling LH, Utracki LA

(1994) Interpenetrating polymer net-works, Adv Chem Ser 239. AmericanChemical Society, Washington

5. Stoy V, Kliment C (1996) Hydrogels:speciality plastics for biomedical andpharmaceutical applications.Technomic Pu, Basel

6. Roh HW, Song MJ, Han DK, Lee DS,Ahn JH, Kim SC (1999) Biomater Sci,Polym Ed 10:123

7. Donatelli AA, Sperling LH, ThomasDA (1976) Macromolecules 9:671

8. Salmeron M, Gallego Ferrer G, Tor-regrosa Cabanilles C, Meseguer DuenasJM, Monleon Pradas M, GomezRibelles JL (2001) Polymer 42:10071

9. Eschbach FO, Huang SJ (1994) In:Klempner D, Sperling LH, Utracki LA(eds) Interpenetrating polymernetworks, Adv Chem Ser 239. Ameri-can Chemical Society, Washington,p 205

10. Kim SJ, Lee KJ, Kim IY, An KH, KimSI (2003) J Appl Polym Sci 90:1384

11. Gomez Ribelles JL, Monleon PradasM, Gallego Ferrer G, Peidro Torres N,Perez Gimenez V, Pissis P, Kyritsis A(1999) J Polym Sci B Polym Phys37:1587

12. Greenspan L (1977) J Res Nat BurStand (US) 81A:89

13. Kyritsis A, Pissis P, Gomez Ribelles JL,Monleon Pradas M (1994) J Polym SciB Polym Phys 32:1001

14. Sessler GM (1980) Electrets. Topics inApplied Physics, vol 33. Springer, BerlinHeidelberg New York

15. Adamson AW (1990) Physicalchemistry of surfaces, 5th edn. Wiley,New York

16. Kyritsis A, Pissis P, Gomez Ribelles JL,Monleon Pradas M (1994) J Non-CrystSolids 172–174:1041

17. Janacek J, Kolarık J (1965) CollectCzech Chem Commun 30:1507

18. Bares J, Janacek J (1965) Collect CzechChem Commun 30:1604

19. Gomez Ribelles JL, Meseguer DuenasJM, Monleon Pradas M (1988) Polymer29:1124

20. Gallego Ferrer G, Monleon Pradas M,Gomez Ribelles JL, Pissis P (1998)J Non-Cryst Solids 235–237:692

21. Crank J, Parks GS (1968) Diffusion inpolymers. Academic, London

22. Rault J, Lucas A, Neffati R, MonleonPradas M (1997) Macromolecules30:7866

23. Salmeron Sanchez M, Gallego FerrerG, Monleon Pradas M, Gomez RibellesJL (2003) Macromolecules 36:860

24. Kolarık J (1982) Adv Polym Sci 46:11925. Gallego Ferrer G, Monleon Pradas M,

Gomez Ribelles JL, Salmeron SanchezM (2004) Polymer (in press)

26. Monleon Pradas M, Gomez RibellesJL, Serrano Aroca A, Gallego Ferrer G,Suay Anton J, Pissis P (2001) Polymer42:4667

690