Effect of cross-linking agents on the dynamic mechanical properties of hydrogel blends of poly(acrylic acid)-poly(vinyl alcohol-vinyl acetate)

Biomoterids 17 (1996) 2259-2264 0 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved ELSEVIER

PI1

014%9612/96/$15.00

SO142-9612(96)00058-O

Effect of cross-linking agents on the dynamic mechanical properties of hydrogel blends of poly(acrylic acid)poly(viny1 alcohol-vinyl acetate) J.V. Cauich-Rodriguez, IRC in Biomedical

Materials,

S. Deb and R. Smith

Queen Mary and Westfield College, Mile End Road, London E7 4NS, UK

A range of hydrogels

were prepared by blending aqueous solutions of poly(vinyl alcohol-vinyl acetate) with poly(acrylic acid) in various proportions. The effects of two cross-linking agents (glyoxal and glutaraldehyde) and subsequent thermal treatment on the properties of the blends are discussed. Dynamic mechanical analysis (DMA) of the xerogels indicated complete miscibility of the various blends which was evident from the appearance of a single glass transition temperature (T,) in the presence of either glyoxal or glutaraldehyde at all thermal treatments studied. A 50150% wtlwt blend was found to have the highest storage modulus and was thus selected for further study. Hydrogels prepared with glutaraldehyde without subsequent thermal treatment exhibited higher storage modulus values than those prepared using glyoxal when tested isothermally at 20°C in a water bath. A further increase in the storage modulus was observed when these hydrogels were thermally treated at 120 or 150°C. In a non-isothermal study on the cross-linked hydrogels, no variation in storage modulus was observed. Broad peaks were observed in tan 6 plots, these peaks shifting towards higher frequencies as the degree of cross-linking increased in the hydrogel. 0 1966 Elsevier Science Limited. Keywords:

Hydrogels,

Received 7 September

polyacrylic

acid, poly(vinyl

acetate),

dynamic

mechanical

analysis

1995; accepted 19 March 1996

Hydrogels are an interesting and unusual class of materials that are rapidly gaining importance as biomaterials. The unique properties that hydrogels display originate from a network structure which allows the retention o’f a considerable amount of water without dissolution of the polymer itself. For biomedical uses, this high water content gives rise to minimal interfacial tension with surrounding biological fluids, gas permeation, diffusion of low-molecularweight compounds and reduced mechanical and frictional irritation to tissue. Hydrogels are ge:nerally characterized by their hydrophilicity, the most important aspect being the amount of water that they are able to absorb, often defined as the equilibrium water content (EWC). In biomedical applications, high water content is an related to attractive attribute higher oxygen permeability and low cell adhesion and protein absorption. However, there are some disadvantages associated with high water content, the most important being a decrease in mechanical strength. A number of approaches have been adopted to improve the mechanical properties of hydrogels including

Correspondence

alcohol-vinyl

increasing the cross-linking density, copolymerization with bulky hydrophobic monomers and grafting to rigid substrates. Improved mechanical properties were achieved by Peppas and Merrill1 by electron beam irradiation of poly(viny1 alcohol) (PVA) solutions and reinforcement by a dehydration-annealing process in cross-linked network. Watase and the already Nishinari’ and Urushizaki et ~1.~ have demonstrated that the storage modulus in PVA hydrogels is increased with the number of freeze-thaw cycles and the concentration of PVA in the gel. Bo4 claimed to improve the mechanical properties of PVA hydrogels by cross-linking with epichlorohydrin and subsequent annealing. We have recently prepared hydrogels with poly(viny1 alcohol-vinyl acetate) cross-linked with glutaraldehyde and reinforced with poly(viny1 pyrrolidone)5. Most mechanical property data reported in the literature for hydrogels are obtained in tension/ compression. Dynamic mechanical measurements on hydrogels are rare2,3S6-8, although this technique has been widely used to analyse structural and intrinsic property changes in a variety of materials. Considering that many applications of hydrogels involve the material being subjected to cyclic loads of varying magnitudes, the importance of dynamic mechanical

to Dr R. Smith.

2259

Biomaterials

1996, Vol. 17 No. 23

2260

Dynamic

mechanical

analysis (DMA) cannot be underestimated. The measurement of parameters such as dynamic modulus, E’, loss modulus, E”, and the damping or internal friction, tan 6, enables relationships between structural parameters and the surrounding environment to be establishedg. In general, as the water content of hydrogels is increased, the storage modulus decreases and the glass transition temperature ( Tg) also decreases; however, with completely swollen networks, the variation in properties is not so clear. In the present study, hydrogel blends prepared from water-soluble polymers of poly(viny1 alcohol-vinyl acetate) (PVAA) and poly(acrylic acid) (PAA) in the presence of glutaraldehyde or glyoxal as cross-linking The dynamic mechanical agent are described. properties of xerogels and their hydrogels with different degrees of swelling are also described. The effect of thermal treatment on the dynamic mechanical properties of hydrogel blends with and without crosslinking agents is reported.

EXPERIMENTAL Poly(acrylic acid) (PAA, M, = 1250 000) and vinyl copolymer (PVAA, M, = alcohol-vinyl acetate 125000, 87-89s hydrolysed) were supplied by Aldrich and BDH respectively and used as received without further purification. Glutaraldehyde (BDH) (50% aqueous solution) and Glyoxal (Aldrich) (40% aqueous solution) were used as cross-linking agents at a concentration of 5% w/w in the presence of lactic acid as catalyst. Aqueous solutions (1%) of PVAA and PAA were mixed in order to obtain a final concentration of 50% wt/wt of polymer pair. Lactic acid (1 ml) was added as catalyst followed by the addition of either glutaraldehyde or glyoxal (5% wt), maintaining the temperature at 60°C for 1 h. Films of the resulting polymer blends were obtained by casting onto polystyrene Petri dishes cross-linking was and drying at 60°C. Further introduced by thermal treatment of the films after 30 min drying at 100°C. After drying, the temperature was increased to 120”C, over a period of about 15 min, and some blends were removed (referred to as thermal treatment at 120°C). The rest of the samples were heated for 30 min at 120°C after which the temperature was increased to 150”C, over a period of 20 min (referred to as thermal treatment at 150°C). After the thermal treatment, the samples were placed in a desiccator and allowed to attain room temperature. The equilibrium water upake (EWC) was calculated using the following relationship: EWC (%) = [(w, - w,J/w,] x 100 where w, is the weight of the swollen gel and wd is the weight of the dry gel after swelling. DMA was performed in tension using a Perkin-Elmer DMA-7, applying sufficient force to produce a deformation of less than 2% in the hydrogel. Strips measuring 15 x 3 x 0.1 mm were used for both blends and the completely swollen hydrogels. Blends were tested in a temperature scan mode from 20 to 180°C at a Biomaterials

1996, Vol. 17 No. 23

properties

of PAA-PVAA

hydrogels:

J.V. Cauich-Rodriguez et al.

heating rate of 5°C min-’ and a frequency of 10 Hz using helium as a purge gas. Hydrogels were tested in a hot water bath from 20 to 60°C at a heating rate of 3°C mini’ and a frequency of 1Hz. A stainless steel serpent was placed around the DMA probe holder in which hot water was circulating. A small amount of water was used for the experiments in order to minimize the time required to equate the temperature within the serpent and the medium. For comparison purposes measurements were also obtained isothermally at 20°C (in water) at a frequency of 10Hz. In the frequency scan mode, hydrogels were tested from 0.1 to 50 Hz. Mechanical properties in tension were also determined for comparative purposes using dumb-bell specimens of 3 mm neck with a gauge length of 25 mm. Hydrogels were kept wet by spraying with distilled water. For blends, an In&on model 6025 with a load cell of 1 kN was used. For hydrogels, an In&on 4464 with a load cell of 2.5 N was used. A cross-head speed of 5mm rnin~l was chosen for both blends and hydrogels.

RESULTS AND DISCUSSION Blend characterization DMA analysis of the PAA-PVAA system showed a single Ts for all compositions, indicating complete miscibility of the blend. These findings are in good agreement with previous data reported from DSC and I,&jE’O, 11 measurements. The strong hydrogen bonding between the hydroxyl groups of PVAA and the carbonyl group from PAA is believed to be responsible for this behaviour. Table z presents the dependence of the T, with composition for these blends as determined by DMA. Blends of different composition exhibited a drastic change in modulus when Tg was reached. Although a drop in modulus from 10’ to 10” Pa was observed in all cases, the initial modulus was not the same. shown in Figure 1, indicate a These results, maximum modulus at 50% composition. This blend was therefore chosen for cross-linking studies. Figure z shows the tan6 variation with temperature for all compositions studied. The transition peak is displaced towards higher temperatures with increasing PAA content, due to the higher Tg of PAA. The presence of a secondary relaxation in PAA and some blends may arise from local motions of either the main chain or pendant groups (Figure 2). However, it is quite unlikely that this should correspond to either a double bond or anhydride formation because these changes appear with severe thermal treatments (120 Table 1 Dependence blends Blend composition PVAA 100 PVAA 75-PAA PVAA 50-PAA PVAA 25-PAA PAA 100

25 50 75

of the

(% wt)

J, on the composition

of PAA-PVAA

J, PMN (“C) 79.0 99.3 106.8 120.8 134.0

Dynamic mechanical

20

40

properties

60

80

of PAA-PVAA hydrogeis:

106

120

TEMPERATURE

146

160

180

(“C)

Figure 1 Storage modulus as a function of temperature for PAA-PVAA blends: a, F’VAA 10% wt; b, PVAA 75-PAA 25; c, PVAA SO-PAA 50; d, PVAA 25-PAA 75; e, PAA 100%.

2.5

1

Glutaraldehyde or glyoxal will cross-link PVAA by inter/intramolecular acetal formation. The effect of this reaction on the properties of the xerogels is shown in Figures 3 and 4. Again, lowering of the intensity of the displacement towards higher tan6 peak and temperatures with a broadening of the peak were observed. The T, of the blends increased up to 115 and 132°C for thermal treatment at 120 and 150°C respectively, when glyoxal was used as a cross-linking agent. The Tg further increased up to 125 and 153°C when glutaraldehyde was employed. The modulus of the xerogels tested both dynamically and on the In&on generally increased with thermal treatment. The glyoxal-cross-linked xerogel exhibited a modulus almost twice as high as glutaraldehyde-crosslinked xerogels. However, the glutaraldehyde-crosslinked xerogels, after swelling in water, had a superior mechanical performance which may be explained by the presence of the longer aliphatic chain lengths of glutaraldehyde. The dynamic mechanical properties of the PM-PVAA 50% xerogels are summarized in Table 2.

,-.d

2-I 1

,’

.j

\,”

I.

i\‘,, 1.5 / 2

2261

J.V. Cauich-Rodriguez et a/.

’

--, ; ,,’ X’-

‘E‘

l-

\ ‘\

\ ’

120

140

160

\‘:

180

TEMPERATURE CC,

Figure 2 Tan 6 as a function of temperature for PAA-PVAA blends: a, PVAA 100% wt, b, PVAA 75-PAA 25, c, PVAA 50PAA 50, d, PVAA 25-PAA 75, e, PAA 100%.

For the :DMA studies the samples were mildly thermally treated at 100°C during 30 min. Similarly, a secondary transition has also been reported for PVA at --20”C1’. or 150°C).

Effect of thermal treatment and cross-linking agent Thermal treatment of the 50% w/w PAA-PVAA xerogel showed that dynamic modulus increases with increasing temperature of treatment. As expected, the effect of the thermal treatment was to reduce the solubility of the xerogel. At 120°C and above, sequences of double bonds and ether formation are produced which may be attributed to dehydration of the PVA13. At 15O”C, PAA undergoes anhydride formation (ring structure)14 and in both cases, crosslinked structures are formed. DSC traces of these blends up to 250°C did not show melting endotherms around 190-22o”C as evidence of crystallinity15 indicating that ther:mal treatment did not induce crystallization. The degree of cross-linking is further increased by using cross-linking agents such as dialdehydes.

lTTTy&!_ ‘,,’ 01 20

1

50

,

,

70

100

,

130

7

150

180

TEMPERATURE CC)

Figure 3 Effect of the cross-linking agent and thermal treatment on tan 6. PAA 50-PVAA 50 xerogeis cross-linked 0 min, 12O”C, ---; 0 with giyoxal: 30 min, lOO”C, -; min, 150°C . . . . .

"1

73

100

130

TEMPERATURE (“C)

Figure 4 Effect of the cross-linking agent and thermal treatment on tan 6. PAA 50-PVAA 50 xerogeis cross-linked with giutaraidehyde. 30 min, 100°C -; 0 min. 120°C - - ; 0 min, 150°C ...... Biomaterials

1996, Vol. 17 No. 23

2262

Dynamic

Table 2 Dynamic xerogels

mechanical

Xerogel

120°C 150°C 120°C 150°C 120°C 150°C

Glyoxal Glyoxal Glutaraldehyde Glutaraldehyde

properties

of PAA-PVAA

Storage modulus

Loss modulus

(GPa)

(GPa)

1.31 3.16 4.47 4.37 2.15 2.48

f f f f f f

0.35 1.39 1.2 1.6 1 0.8

mechanical

0.47 0.55 0.12 0.7 0.33 0.32

f f f f f f

properties

0.37 0.28 0.28 0.16 0.15 0.13

f f f f f *

0.24 0.18 0.22 0.08 0.05 0.03

Hydrogel characterization The mechanical properties in tension mode as conducted on the In&on of the various xerogels and hydrogels are presented in Table 3. In the case of xerogels, the modulus increased with thermal treatment for each system studied. For thermally treated xerogels that contained no cross-linking agents, the ultimate tensile strength (UTS) increased initially and then decreased as expected, a phenomenon associated with development of sub-micro cracks due to internal stresses. However, xerogels formed with the addition of cross-linking agents did not exhibit a substantial decrease in UTS with increasing thermal treatments. The strain to failure for the xerogels was generally found to decrease with the introduction of cross-linking agents (Table 3). For hydrogels, it was observed that while the elastic modulus increased with thermal treatment and chemical cross-linking, the breaking strain was considerably reduced. The elastic behaviour of these hydrogels was strongly dependent on the amount of water embedded in the matrix. Hydrogels with lower EWC values, such as those containing glutaraldehyde, exhibited lower breaking strain (8-9%) in comparison with glyoxal-cross-linked blends that had higher water contents (1148%). The Young’s modulus and the tensile strength of the hydrogels increased with thermal treatment; the maximum values were obtained for the

Table 3 Mechanical

properties

Treatment

of PAA-PVAA

50% xerogels

J.V. Cauich-Rodriauez

hydrogels:

Dynamic mechanical analysis In an isothermal DMA study of hydrogels the modulus remained constant for 15min at a frequency of 10Hz. Under these conditions, hydrogels crosslinked with glutaraldehyde exhibited a higher compared to glyoxal-cross-linked and modulus hydrogels without cross-linking agents. Additional thermal treatment at 150°C improved the modulus in each case. These results are in good agreement with modulus measurements obtained statically5. The storage modulus was lower for all blends when a low frequency (1 Hz) was used, which was independent of the cross-linking agent. Table 4 summarizes the dynamic mechanical properties of these hydrogels at 10Hz. The effect of temperature on the storage modulus at 1 Hz is presented in Figures 5 and 6. No significant change in the modulus was observed up to 60°C. Error bars are the standard deviations observed for at least five samples. In some samples, an initial

and hydrogels Hydrogels

Xerogels

E (GPa)

UTS (GPa)

Breaking

strain

E (MPa)

Breaking

UTS (MPa)

(%) 120°C 150°C 120°C 150°C 12O”C, 15O”C,

Glyoxal Glyoxal Glutaraldehyde Glutaraldehyde

Table 4 Dynamic

5.3 9.2 9.2 9.7 8.5 9.5

mechanical

f 2.2 f 2.7 f 2.0 f 0.7 z!z2.0 * 3.0

0.26 0.08 0.11 0.14 0.07 0.075

properties

f f zt f f *

0.03 0.03 0.03 0.02 0.02 0.03

of PAA-PVAA

5.7 6.1 2.4 3.0 1.2 1.3

f f f It * +

1.9 2.3 0.8 1.0 0.5 0.8

0.76 12.0 0.3 4.3 20 20

f f f f zt +

0.6 3.5 0.15 1.0 6.0 3.3

0.18 zt 1.4 & 0.11 f 0.4 f 1.5f 0.47 f

0.04 0.85 0.07 0.2 1.3 0.31

170 f 13.0 * 68 + 11 i 8.3 f 3.9f

E’ (MPa)

E” (MPa)

TanG

120°C 150°C 120°C 15O”C, 120°C 150°C

8.74 5.33 6.47 5.10 19.5 29.2

2.72 4.44 2.08 4.5 17.2 25.3

8.29 2.92 6.12 2.3 9.0 14.5

3.05 0.64 2.90 0.50 0.56 0.55

Biomateriak

1996, Vol. 17 No. 23

1.80 0.92 2.00 0.54 8.80 8.00

EWC (%)

36 4 28 6.5 2.6 1.7

76.06 41.88 85.1 36.9 32.6 21.9

50% hydrogels

E* (MPa) f f f f + zt

strain

(%)

Hvdroael

Glyoxal Glyoxal Glutaraldehyde Glutaraldehyde

et al.

glutaraldehyde cross-linked system. This behaviour is also associated with the low water content of this system. It is evident that water uptake is reduced with cross-linking achieved either by thermal treatment or by the use of a cross-linker; hence the mechanical properties of the hydrogel are improved, the exception being the glyoxal cross-linked at 120°C hydrogel where the EWC exceeded that of the untreated blend. The PAA-PVAA blends have strong hydrogen bonding present in the network which is disrupted during crosslinking and any moiety with a higher hydrophilicity renders it more hydrophilic. The same effect is not observed at elevated temperatures as the degree of cross-linking offsets the effects of higher hydrophilicity. The mechanical properties of these hydrogels are superior compared to PVA-epichlorohydrin-crosslinked hydrogels4 but are inferior to those reported by Corkhill et al.16 where good mechanical behaviour was observed in hydrogels with EWC of 70450%.

50%

Tan 6

0.09 0.23 0.11 0.5 0.1 0.1

of PAA-PVAA

f f f f f f

0.60 0.54 0.56 0.46 8.31 6.80

f 1.70 f 0.88 & 2.00 zt 0.29 *3.12 * 4.46

f f f * * f

EWC (%) 0.12 0.12 0.26 0.01 0.09 0.06

76.06 41.88 85.10 36.90 32.60 21.90

Dynamic mechanical

properties

of PAA-PVAA hydrogels:

J.V. Cauich-Rodriguez

1

oh

10

20

I

I

30

40

’

50

I

I

60

70

TEMPERATURE CC,

Figure 5 Temperature dependence of the storage modulus for PAA 50-PVAA 50 xerogels cross-linked with glyoxal. Unfilled markers correspond to hydrogels thermally treated at 120°C and the filled at 150°C.

2263

et al.

implies that the viscoelastic behaviour was greater with higher water content (Figures 7 and 8). The frequency dependence of PAA-PVAA hydrogels cross-linked with glutaraldehyde is shown in Figure 9. A plot of modulus versus frequency revealed a continuous decay (exponential) in the modulus as the frequency changed from 50 to O.lHz. At higher frequencies the material is relatively stiff, while at lower frequencies typical rubbery behaviour is observed. However, a logarithmic plot showed that the more highly cross-linked hydrogels possessed higher storage modulus at frequencies lower than 10Hz. Interestingly, if tan6 is plotted instead of the modulus (Figure 10) broad peaks are observed. These peaks were found to shift towards higher frequencies with increasing cross-linking density in the hydrogel. It is well known that the degree of swelling increases when a swollen rubber is under tensile stress and the appearance of a broad peak in the rubbery state (equivalent to a swollen hydrogel) may be related to changes in the degree of swelling and not to a molecular transition.

16 3.5 ‘3

PAA-PVAA

+ GU

:

PAA-PVAA

+ GY

3/ 1

0 A

PAA-PVAA

8

0+-r10

I

20

30

’

I

40

TEMPERATURE

1-w

50

60

70

(“C)

Figure 6 Temperature dependence of the storage modulus for PAA 50-PVAA 50 xerogels cross-linked with glutaraldehyde. Unfilled markers correspond to hydrogels thermally treated at 120°C and the filled at 150°C.

increase in the modulus was detected as reported by Watase and Nishinari’ which has been explained as an increase in the concentration of long flexible chains which in turn increases the Brownian motion (thermally activated process). The presence of these types of chains in covalent cross-linked hydrogels indicates the format:ion of a heterogeneous network in which loose ends are present in variable proportions. PVA hydrogels prepared by freeze-thaw methods exhibit a reduction in the dynamic modulus when heated. Bao and Bagga7 have reported a decrease in modulus above a temperature of 45”C, attributed to disentanglement of flexible chains in PVA. Watase and Nishinari’ have related the modulus reaction to chain disentangling and melting of crystalline regions above 35°C. In the same way that water content controls values of the modulus, other related properties such as tan 6 and the phase angle are also affected by the equilibrium water content. Hydrogels with EWC higher than 70% showed high tan6 values as well as high phase angles, whereas highly cross-linked networks with low EWC exhibited low tan6 and phase angle values. This

Oi!Er-ri

50

EQUILIBRIUM

60

I

,

1

70

80

90

WATER CONTENT (%)

Figure 7 Tan6 variation with EWC for PAA 50-PVAA 50 hydrogels with different cross-linking agents, glutaraldehyde (GU) and glyoxal (GY). Unfilled markers correspond to hydrogels thermally treated at 120°C and the filled at

ml 70m h‘ 2

60-

3

50-

D z

40

0

PAA-PVAA + GU

A

PAA-PVAA + GY

n

PAA-PVAA

0

n

. 30-. 20 20

0 .

1 30

I 40

Tm-1m--7pY 50 60

70

80

90

EQUILIBRIUM WATER CONTENT (9%)

Figure 6 Phase angle variation with EWC for PAA 5&PVAA 50 hydrogels with different cross-linking agents, glutaraldehyde (GU) and glyoxal (GY). Unfilled markers correspond to hydrogels thermally treated at 120°C and the filled at 150°C. Biomaterials 1996. Vol. 17

No.23

2264

Dynamic mechanical

properties

of PAA-PVAA hydrogels:

J.V. Cauich-Rodriguez

for buoyancy effects corrected measurements are carried out in water.

when

et al. the

ACKNOWLEDGEMENTS The authors thank the EPSRC (UK) and (Mexico) for support to this programme.

CONACYT

REFERENCES ,--- ~~

0

10

20

30

or --

40

50

60

1

FREQUENCY (Hz) Figure 9

Storage modulus versus frequency for PAA 50PVAA 50 hydrogel cross-linked with glutaraldehyde: 30 min, lOO”C,--;Omin, 12O”C,---; Omin, 150°C ......

2

3 1.4-

4 5

6

07 0

10

20

30

40

50

60

7

FREQUENCY (Hz) Figure 10

hydrogel lOO”C, -;

Tans versus frequency for PAA 50-PVAA 50 cross-linked with glutaraldehyde. 30 min, 0 min, 12O”C, ---; 0 min, 150% ......

8

9

CONCLUSIONS The complete miscibility of blends of poly(acrylic acid)poly(viny1 alcohol-vinyl acetate) copolymer has been demonstrated using DMA. The criterion of a single Ts was observed for blends of different composition and after different chemical and thermal treatments. Because of the miscibility of the two polymers in the blend, isotropic hydrogels may be formed. Blends of PAA and PVAA behave as hydrogels when they are chemically cross-linked with a dialdehyde and thermally treated at 120 or 150°C. Hydrogels of PAA and PVAA cross-linked with glutaraldehyde exhibited higher modulus values when compared to those cross-linked with glyoxal. The use of glutaraldehyde as a cross-linking agent resulted in hydrogels of lower water contents with improved mechanical properties. Hydrogels crosslinked with either glutaraldehyde or glyoxal did not show any variation in storage modulus when tested in water from 20 to 60°C. The results obtained from DMA measurements compare well with modulus values obtained statically. Care must be taken in direct comparisons because DMA values need to be Biomaterials 1996, Vol. 17 No. 23

10

11

12

13

14

15

16

Peppas NA, Merrill EW. Development of semicrystalline poly(viny1 alcohol) hydrogels for biomedical applications. J Biomed Mater Res 1977; 11: 423434. Watase M, Nishinari K. Rheological and DSC changes in poly(viny1 alcohol) gels induced by immersion in water. JPolym Sci Polym Phys Ed 1985; 23: 1803-1811. Urushizaki F, Yamaguchi H, Nakamura K, Numajiri S, Sugibayashi K, Morimoto Y. Swelling and mechanical properties of poly(viny1 alcohol) hydrogels. Int J Pharm 1990; 58:135-142. Bo J. Study on PVA hydrogel crosslinked by epichlorohydrin. JAppl Polym Sci 1992; 46: 783-786. Cauich-Rodriguez JV, Deb S, Smith R. Characterisation of hydrogel blends of poly(viny1 pyrrolidone) and poly(viny1 alcohol-vinyl acetate). J Mater Sci: Mater Med 1996; 7: 269-272. Lustig SR, Caruthers JM, Peppas NA. Dynamic mechanical properties of polymer-fluid systems: characterisation of poly(%hydroxyethyl methacrylate) and poly(‘& hydroxyethyl methacrylate-co-methyl methacrylate). Polymer1991; 32:3340-3353. Bao QB, Bagga CS. The dynamic mechanical analysis of hydrogel elastomers. Thermochim Acta 1993; 26: 107113. Lazzeri L, Barbani N, Gascone MG, Lupinacci D, Giusti P, Laus M. Physico-chemical and mechanical characterisation of hydrogels of poly(viny1 alcohol) and hyaluronit acid. J Mater Sci Mater Med 1994; 5: 862-867. Murayama T. Dynamic Mechanical Analysis of Polymeric Materials. Amsterdam: Elsevier, 1978: 1. Vazquez-Torres H, Cauich-Rodriguez JV, Cruz-Ramos CA. Poly(viny1 alcohol)/poly(acrylic acid) blends: miscibility studies by DSC and characterisation of their thermally induced hydrogels. J AppJ PoJym Sci 1993; 50: 777-792. Zhang H, Takegoshi T, Hikichi K. Phase separation and alcohol)/ thermal degradation of poly(viny1 poly(methacrylic acid) and poly(viny1 alcohol)/ poly(acrylic acid) systems by 13C c.p.1m.a.s. n.m.r. Polymer 1992; 33:718-724. Sarti B, Scandola M. Viscoelastic and thermal properties of collagen/poly(vinyl alcohol) blends. Biomaterials 1995; 16: 785-792. Immelman E, Sanderson RD, Jacobs EP and Van Reenan AJ. Poly(viny1 alcohol) gel sublayers for reverse osmosis I. Insolubilisation by acid-catalysed membranes. dehydration. JAppl Polym Sci 1993; 50: 1013-1034. Maurer JJ, Eustace DJ, Ratcliff CT. Thermal characterisation of poly(acrylic acid). Macromolecules 1987; 20: 196-202. Peppas NA, Merrill EW. Differential scanning calorimetry of crystallised PVA hydrogels. J Appl Polym Sci 1976; 20: 1457-1465. Corkhill PH, Trevett AS, Tighe BJ. The potential of hydrogels as synthetic articular cartilage. Proc Inst Mech Eng, Part H: J Eng Med 1990; 204: 147-155.