NMR Imaging of Water Sorption into Poly(hydroxyethyl methacrylate

Kristofer J. Thurecht, David J. T. Hill, and Andrew K. Whittaker ... A. Chowdhury, David J. T. Hill, Andrew K. Whittaker, Michael Braden, and Mangala ...
0 downloads 0 Views 105KB Size
Biomacromolecules 2001, 2, 504-510

504

NMR Imaging of Water Sorption into Poly(hydroxyethyl methacrylate-co-tetrahydrofurfuryl methacrylate) Phuong Y. Ghi,† David J. T. Hill,*,† and Andrew K. Whittaker‡ Polymer Materials and Radiation Group, Department of Chemistry, and Centre for Magnetic Resonance, The University of Queensland, Brisbane, Australia 4072 Received December 15, 2000

The diffusion of water into a series of hydroxyethyl methacrylate, HEMA, copolymers with tetrahydrofurfuryl methacrylate, THFMA, has been studied over a range of copolymer compositions using NMR imaging analyses. For polyHEMA the diffusion was found to be consistent with a Fickian model. The mass diffusion coefficient of water in polyHEMA at 37 °C was determined from the profiles of the diffusion front to be 1.5 × 10-11 m2 s-1, which is less than the value based upon mass uptake, 2.0 × 10-11 m2 s-1. The profiles of the water diffusion front obtained from the NMR images showed that stress was induced at the interface between the rubbery and glassy regions which led to formation of small cracks in this region of the glassy matrix of polyHEMA and its copolymers with mole fractions of HEMA greater than 0.6. Water was shown to be able to enter these cracks forming water “pools”. For copolymers of HEMA and THFMA with mole fractions of HEMA less than 0.6 the absence of cracks was attributed to the ability of the THFMA sequences to undergo stress relaxation by creep. Introduction HEMA copolymers have acceptable biocompatibility,1-3 so they have found numerous medical applications. The copolymers of HEMA have been identified as having potential use in controlled release drug delivery systems,4-6 as well as in medical implants and other medical products.7,8 In these applications, the diffusion of water in the polymer is of fundamental importance as its presence controls the properties of the swollen polymer. In controlled delivery systems based on hydrogels, the release of the drug occurs following diffusion of body fluids into the polymer. The diffusion of body fluids (water) into a glassy polymer causes it to become rubbery, thus allowing the drug molecules to diffuse out of the polymer. The glass transition temperature of the polymer is lowered by the absorbed water in the rubbery regions, resulting in a significant increase in the segmental mobility of the polymer chains. Recently, Braden and co-workers7-11 have shown that copolymers of 2-hydroxyethyl methacrylate (HEMA) with tetrahydrofurfuryl methacrylate (THFMA) have excellent prospects as biomaterials for use in medicine and dentistry, including drug delivery applications. The rate of diffusion of water into a copolymer, as well as the equilibrium amount of water which can be absorbed by the matrix, is dependent on the nature of the comonomers present and on the comonomer composition. In the present study, water diffusion into copolymers of HEMA and THFMA has been * To whom correspondence should be sent. † Polymer Materials and Radiation Group, Department of Chemistry, The University of Queenslandand. ‡ Centre for Magnetic Resonance, The University of Queensland.

Figure 1. Chemical structures of HEMA and THFMA.

investigated using magnetic resonance imaging of the water diffusion front. The distinguishing features of these monomers are the hydrophilic hydroxy side chain of the HEMA monomer and the more hydrophobic side chain of the THFMA monomer. The objective of the study was to examine the profile of the water diffusion front in polyHEMA, to estimate the best value of the diffusion coefficient, and to examine the effect of increasing proportions of THFMA on the nature of the diffusion of water into the HEMA copolymers. The structures of the HEMA and THFMA repeat units in the polymers are presented in Figure 1. Experimental Section Stabilized THFMA (DAJAC) and HEMA (Ubichem Ltd) were purified immediately before use by vacuum distillation (pressure ≈ 102 Pa) with only the middle fraction used experimentally. The purity of the monomers was confirmed by NMR analysis. Preparation of Copolymer Cylinders for Diffusion Measurements. Cylindrical samples of the copolymers,

10.1021/bm000146q CCC: $20.00 © 2001 American Chemical Society Published on Web 06/11/2001

Biomacromolecules, Vol. 2, No. 2, 2001 505

NMR Imaging of Water Sorption

Figure 3. NMR image of the diffusion front for water penetration into polyHEMA at 37 °C.

Figure 2. Mass uptake plot for polyHEMA at 37 °C showing the experimental points and the best fit curve for D ) 2.00 × 10-11 m2 s-1. Table 1. Diffusion Coefficients for THFMA/HEMA Copolymers Reported by Ghi et al.10 for a Fickian Model at 37 °C as a Function of the HEMA Mole Fraction, FHEMA

FHEMA

D × 1011/m2s-1

FHEMA

D × 1011/m2s-1

1.0 0.9 0.8 0.7 0.6 0.5

2.00 1.63 1.28 0.99 0.86 0.73

0.4 0.3 0.2 0.1 0.0

0.61 0.53 0.50 0.60 0.93

polymerized to high conversion, were prepared for diffusion studies in cylindrical Teflon molds. Cylindrical samples were chosen because they fit neatly into the NMR resonator used for imaging studies of the diffusion process. The monomers were weighed into a 25 mL Pyrex glass flask in the required mole ratios and benzoyl peroxide, BPO, was added to yield a 0.05 M solution. The mixture was shaken until the BPO had dissolved, and then it was poured into the Teflon mold which had an internal diameter of ≈5 mm and a length of ≈20 mm. The cylindrical mold was closed with a cap containing a small hole, to allow excess mixture to drain out of the cylinder. The polymerization was forced to complete conversion of monomer to polymer in a vacuum oven using the following temperature/time protocol: the samples were held initially at 50 °C for 20 h and then at 80 °C for 2 h. This polymerization protocol leads to the formation of a polymer cylinder without the generation of excessive heat, which could result in the formation of bubbles or the loss of optical clarity. After polymerization, the polymer cylinders were removed from the Teflon molds, the absence of monomer was confirmed by FT-NIR analysis, and the ends were ground to a smooth, flat finish. NMR Imaging Measurements. The copolymer cylinders prepared to 100% conversion were placed in stoppered tubes filled with distilled water, in a water bath at 37 ( 1 °C. Prior to the sorption measurements the cylinders were stored in a vacuum oven for 1 week at 37 °C and then at 80 °C for

10 min to remove stress and any absorbed moisture. At periodic times the cylinders were removed for imaging. A Bruker AMX300 spectrometer was used to obtain the images with a standard Bruker spin echo, three-dimensional, SE3D, pulse sequence. The images were acquired using a 90° pulse of a duration of 50 µs and echo and repetition times of 2.67 ms and 1.0 s, respectively. The read gradient was 1.5 T m-1 and the images consisted of 128 × 128 × 8 voxels with an in-plane resolution of 78 × 78 µm and a slice thickness of 3.75 mm in a field of view of 1 × 1 × 3 cm. Two averages were added over an acquisition time of 35 min. After removal of the cylinders from the water for NMR measurements, any water adhering to the outer surface of the cylinders was removed by drying with a paper towel. This process, along with some evaporation of water from the outer edges of the cylinder during the acquisition of the NMR image, in some cases led to the formation of small anomalous features arising from a loss of signal intensity for the water protons at the outer edges of the NMR images. Results and Discussion The diffusion of water into polymers has been reported to occur between two limiting cases: namely Fickian (or case I) diffusion and case II diffusion.12 For Fickian diffusion, the concentration gradient is the driving force for diffusion and occurs when the rate of diffusion of the penetrant is much slower than the rate of relaxation of the polymer chains. In case II diffusion, the rate of relaxation is slow relative to the rate of penetrant diffusion, so the relaxation (or mobility) of the polymer chains is the controlling factor for diffusion. In case II diffusion, the rate of relaxation of the polymer chains under stress induced by the increase in volume on swelling is slow relative to the rate of penetrant diffusion, and so is the controlling factor for diffusion. A behavior between the extremes of cases I and II is known as case III diffusion. PolyHEMA and its copolymers have been previously reported to follow Fickian or case I diffusion, principally on the basis of mass uptake measurements.4,13-18 A coefficient, D, for diffusion of water into a very lightly cross-

506

Biomacromolecules, Vol. 2, No. 2, 2001

Ghi et al.

Figure 5. Decay in the transverse magnetization of the outer (b) and inner (2) features in the image profile for polyHEMA at a fractional water uptake of 0.36 with increasing echo time, TE, at 37 °C.

Figure 4. Image profiles for polyHEMA at 37 °C for Mt/M∞ equal to (A) 0.18, (B) 0.43, (C) 0.55, (D) 0.75, (E) 0.96, and (F) 1.00. Simulated curves (--) for D ) 1.5 × 10-11 m2 s-1.

linked polyHEMA of 4.78 × 10-10 m2 s-1 at 34 °C has been reported by Peppas et al.4 Gerhke et al.13 have carried out a detailed investigation into polyHEMA sheets, and reported that the diffusion is Fickian at low mass uptake. They determined the diffusion coefficients for water in polyHEMA in the absence of added cross-linker over a range of temperatures from 4 to 88 °C. An interpolated value of the diffusion coefficient of 1.5 × 10-11 m2 s-1 at 37 °C was determined from an Arrhenius fit to the diffusion coefficient data of Gerhke et al.13 More recently we have reported that the diffusion of water into HEMA polymers at 37 °C follows the Fickian model,14-17 from which we have obtained diffusion coefficients for water in polyHEMA in the range (1.5-2.0) × 10-11 m2 s-1, depending on the cross-link density. The copolymers studied here are amorphous, so isotropic diffusion can be assumed to apply. It has been shown12 that for such a matrix the mass uptake of penetrant by diffusion into an infinite cylinder of radius a can be represented by eq 1, and the concentration profile across the diameter of Mt M∞



)1-

∑ n)1

4 a Rn 2

2

exp(- DRn2t)

(1)

the cylinder can be represented by eq 2, where t is the time Cr,t C0,∞

)1-

2



(

J0(rRn)

exp - DRn2t ∑ a n)1 J (aR ) 1

n

)

(2)

for which penetrant diffusion has occurred, Mt and M∞ are the mass uptakes at time t and at equilibrium, respectively,

Cr,t is the concentration of the penetrant at distance r at time t, C0 is the constant surface concentration, and Rn is the positive root of J(aRn) ) 0. Jo(x) is a Bessel function of the first kind of order zero, and J1(x) is a Bessel function of the first order. Mass Uptake Measurements of Diffusion Coefficients. We have previously reported15 the time dependence of the mass uptake and measurement of the diffusion coefficients for the series of HEMA/THFMA copolymers using the relationship given in eq 1. A typical mass uptake curve for polyHEMA and the best fit curve is shown in Figure 2. The best value of the diffusion coefficient obtained from the curve fit for polyHEMA shown in Figure 2 was (2.00 ( 0.1) × 10-11 m2 s-1 at 37 °C, which is in close agreement with the literature values which range from 1.5 × 10-11 to 2.0 × 10-11 m2 s-1. However, an examination of the experimental data and the fitted curve in Figure 2 shows that there is structure in the deviations of the experimental points about the fitted curve. These deviations are consistent with what would be expected if there was a change in behavior at the point where the glassy core disappears, which is at a mass uptake ratio of approximately 0.60, as described by Gehrke et al.13 Therefore, a second data analysis was performed using only the mass uptake data obtained prior to the disappearance of the glassy core. This yielded a value of D of 1.90 ( 0.1 × 10-11 m2 s-1. The lower value obtained from data below a mass uptake fraction of 0.6 is consistent with what has been observed previously for polyHEMA and other HEMA copolymers.13-18 Ghi et al.15 have reported values of D obtained from mass uptake data for a range of HEMA/THFMA copolymers, and these values are summarized in Table 1. With increasing proportions of THFMA in the copolymers, the diffusion coefficients decrease, reflecting the influence of the compositional microstructure of the copolymers on the diffusion of water into the cylinders. As the concentration of HEMA in the copolymers decreases, the HHH triad fraction decreases, and as a result, the polar HEMA units more frequently have a much less polar THFMA neighbor. This has the effect of inhibiting the diffusion of the polar water molecules through the copolymer matrix.

NMR Imaging of Water Sorption

Figure 6. Image profiles of the water diffusion front for HEMA/THFMA copolymers at various diffusion times: (A) XHEMA ) 0.80, 36 h; (B) XHEMA ) 0.70, 36 h; (C) XHEMA ) 0.50, 36 h; (D) XHEMA ) 0.40, 38 h; (E) XHEMA ) 0.30, 38 h; (F) XHEMA ) 0.20, 38 h at 37 °C.

While the aspect ratio for the cylinder is relatively high (ratio ≈ 4), effects associated with water diffusion through the ends of the cylinders will play a role, albeit a small one, in measurements of the mass of water absorbed. Hence the values of the diffusion coefficients for water obtained from analyses of mass uptake measurements on cylinders such as those used here will be somewhat over-estimated. NMR Imaging of Diffusion in PolyHEMA Cylinders. NMR imaging experiments can provide details about the nature of the water diffusion front as it penetrates the cylinder. Figure 3 shows an NMR image of the concentration of water protons across the diameter of a HEMA cylinder which has absorbed water at 37 °C. Since this slice has been taken in the central region of the cylinder, it is representative of the regions of the cylinder where end effects can be truly ignored. The figure shows that the concentration of the water is greatest at the outer edge of the cylinder and decreases toward the center of the cylinder which remains in a glassy state in this image. The outer swollen region is made rubbery by the water which acts as a plasticizer. It should be noted that the spin-spin relaxation time of the polymer protons is short compared to the imaging echo time, and hence the polymer protons do not contribute to the image. This was confirmed by spectroscopic measurements of the polymers swollen in deuterium oxide.

Biomacromolecules, Vol. 2, No. 2, 2001 507

In the image, a cross section of the water concentration across the diameter of the cylinder can be obtained in which the concentration is normalized to that of the water in the surrounding medium. A series of such profiles is presented in Figure 4 for different penetration times for water sorption. An examination of these profiles shows that the concentration of the water decreases monotonically toward the center of the cylinder in matrices where the glassy core has disappeared, and is clearly not of the form predicted by the case II model for diffusion. The profiles for which a glassy core remains are also characterized by an underlying Fickian profile, but they contain peaks in the water concentration at the interface between the inner edge of the diffusion front and the core. The origin of these seemingly anomalous peaks lies in the formation of cracks in the glassy core as a result of stress associated will the swelling of the outer region of the matrix by the water.14 Water molecules can enter these cracks, forming small “pools” of relatively mobile water. Figure 5 shows the variation with increasing spin echo time of the intensities of the outer and the inner features of the image profile for polyHEMA at a relative mass uptake of 0.36. The rate of decay in intensity of the outer feature is much greater than that of the inner feature, which suggests the presence of two types of water. From the data in Figure 5, an estimate of the spin-spin relaxation time T2 ) 0.98 ms was calculated for the water contributing to the outer feature and T2 ) 1.81 ms for the water contributing to the inner feature. Water “pools” in cracks would be characterized by a longer relaxation time than water which interacts more intimately with the polymer. Thus, because of the difference in the relaxation times, the proportion of water associated with the inner peaks in the T2-weighted profiles in Figure 4 is exaggerated, and in reality the concentration of this water is smaller than the profile of the image in Figure 4 would suggest. However, the cracking of the glassy core and the subsequent entry of the water into the cracks plays an important role in the diffusion process. As the diffusion front moves through the cylinder, so too does the annulus of the cracks. However, the water which enters these cracks appears to “dissolve” the adjacent polymer chains in such a way that in the rubbery section of the matrix the water molecules appear to become more homogeneous from an NMR point of view. As a result, the peak in the NMR image associated with the water in cracks appears to move through the cylinder along with the leading edge of the diffusion front. The cracks appear to be healed behind the interface between the glassy core and the rubbery region. This is suggested to be due to swelling of the polymer chains removed from the constraining glassy core. Thus, the cracks are closed with the concurrent loss of water characterized by the longer relaxation time. Once the glassy core of the cylinder disappears, at a mass uptake fraction of ≈ 0.6, there is no further evidence for any anomalous features. To examine the underlying nature of the diffusion of water, the concentration profiles in Figure 4 were analyzed using eq 2. A value of D ) 1.50 × 10-11 m2 s-1 was found to describe best the profiles over the whole regime of diffusion

508

Biomacromolecules, Vol. 2, No. 2, 2001

Ghi et al.

Figure 7. Image profiles of the water diffusion front for HEMA/THFMA copolymers at various diffusion times: (A) XHEMA ) 0.90, 7 h; (B) XHEMA ) 0.80, 5 h; (C) XHEMA ) 0.70, 15 h; (D) XHEMA ) 0.60, 15 h; (E) XHEMA ) 0.50, 15 h; (F) XHEMA ) 0.30, 18 h at 37 °C.

times up to the equilibrium point. Representative fits of the underlying concentration profiles with this value of D are shown in Figure 4. The good representation of the profiles provides convincing evidence of an underlying Fickian diffusion mechanism. This value of D is somewhat lower than that found from the mass uptake measurements, D ) 2.00 × 10-11 m2 s-1 from Table 1, perhaps due to the influence of finite end effects and crack formation on the latter measurements. NMR Imaging of Diffusion in Copolymer Cylinders. In Figure 6 a series of representative image profiles for HEMA/THFMA copolymers is presented for diffusion times

for which the glassy core of the cylinders has disappeared. The shapes of these profiles are clearly consistent with adherence of the diffusion process to a Fickian model. These matrix profiles can be analyzed by curve fitting and the results obtained for the diffusion coefficients for the copolymers mirrored the observations described above for polyHEMA. In Figure 7, a series of profiles is shown for a range of copolymer compositions for diffusion times at which some glassy core remains in the cylinders. These profiles show clear evidence for the presence of crack formation in the copolymer samples over the HEMA mole fraction range 1.0

NMR Imaging of Water Sorption

Biomacromolecules, Vol. 2, No. 2, 2001 509

coefficient for water in THFMA at 37 °C obtained from curve fits to the profiles was consistent with that obtained from the mass uptake study15 of 9.3 × 10-12 m2 s-1. Conclusions

Figure 8. Image profiles of the water diffusion front for polyTHFMA at various diffusion times: (A) 18 h, (B) 28 h at 37 °C.

to ≈ 0.6. However, for the copolymers with mole fractions of HEMA less than 0.6, no evidence for the formation of cracks adjacent to the diffusion front was observed. PolyTHFMA is a much more ductile polymer than polyHEMA,19 and the polymer chains in the glassy core near the diffusion front in polyTHFMA, and copolymers of THFMA and HEMA containing higher contents of THFMA can undergo creep, thus relaxing the stress on the core in this region exerted by the adjacent swollen rubbery region during diffusion of water into the polymer cylinders. This observation is in agreement with the explanation proposed by Braden et al.19 to explain the slow uptake of water after an initial pseudo equilibrium water uptake15 is reached by polyTHFMA and HEMA/THFMA copolymers rich in THFMA. NMR Imaging of Diffusion in PolyTHFMA Cylinders. The NMR images of two representative profiles for the diffusion of water into polyTHFMA are shown in Figure 8. The profile taken after 18 h was obtained while the glassy core is still present, and that at 28 h was obtained after the glassy core had disappeared. The noise in the profiles is much greater than that for the polyHEMA profiles because the concentration of water present in the polymer cylinders is much less. However, the profiles are clearly consistent with Fickian diffusion of water in the cylinders, with no evidence for crack formation at the diffusion front. The diffusion

The diffusion of water into cylinders of polyHEMA and its copolymers with THFMA was found to be predominantly Fickian-type diffusion. The best value for the diffusion coefficient for polyHEMA obtained by curve fits to the NMR image profiles using a Fickian model was 1.50 × 10-11 m2 s-1, compared to a value of 2.00 × 10-11 m2 s-1 from mass uptake. The T2-weighted image profiles also showed evidence for the presence of an annulus of cracks in the glassy core of the cylinders at the leading edge of the water diffusion front. Water was observed to enter these cracks to form “pools” of water which were characterized by a larger spin-spin relaxation time than that of the water which interacts more closely with the polymer chains. These cracks were found to be also present in the copolymer cylinders, but the degree of crack formation decreased with increasing THFMA content. This was interpreted in terms of the ductile nature of the THFMA component. As the diffusion front progresses through the matrix of the cylinder, the water in the pools appears to “dissolve” the adjacent polymer chains, so that the cracks appear to become incorporated into the rubbery matrix behind the front. Once the glassy core disappeared as a result of continued diffusion of water, there was no evidence for the presence of water in cracks in the T2-weighted water concentration profiles. Acknowledgment. The authors wish to acknowledge financial support for this work from the Australian Research Council. The contributions through discussions with Michael Braden, Mangala Patel and Sandra Parker of the IRC in Biomedical Materials at Queen Mary and Westfield College of the University of London are also acknowledged. References and Notes (1) Choudray, M. S.; Varma, I. K. Macromol. Sci. Chem. 1983, A20, 771. (2) Wichterle, O.; Lim, D. Nature 1960, 185, 177. (3) Jeyanthi, R.; Pandurang Rao, K. Biomaterials 1990, 11, 238. (4) Peppas, N. A.; Moynihan, H. J. In Hydrogels in Medicine and Pharmacy; Peppas, N. A., Ed.; CRC Press: Boca Raton, FL, 1987; Vol II, p 49. (5) Tighe, B. J. In Hydrogels in Medicine and Pharmacy; Peppas, N. A. Ed.; CRC Press: Boca Raton, FL, 1987; Vol III, p 53. (6) Braden, M.; Clarke, R. L.; Nicholson, J.; Parker, S. Polymeric Dental Materials; Springer: London, 1997. (7) Patel, M. P.; Braden, M. Biomaterials 1991, 12, 645. (8) Patel, M. P.; Braden, M. Biomaterials 1991, 12, 645. (9) Scotchford, C. A.; Sims, B.; Downs, S.; Braden, M. Biomaterials 1997, 38, 3869. (10) Patel, M. P.; Swai, H.; Davy, K. W. M.; Braden, M. J. Mater. Sci: Mater. Med. 1999, 10, 147. (11) Rigga, P. D.; Braden, M.; Patel, M. Biomaterials 2000, 21, 345. (12) Crank, J. The Mathematics of Diffusion; University Press: Oxford, U.K., 1964; Chapter 5. (13) Gerhke, S. H.; Biren, D.; Hopkins, J. J. J. Biomater. Sci., Polym. Ed. 1994, 6, 375. (14) Ghi, P. Y.; Hill, D. J. T.; Maillet, D.; Whittaker, A. K. Polym. Commun. 1997, 38, 3985.

510

Biomacromolecules, Vol. 2, No. 2, 2001

(15) Ghi, Y. P.; Hill, D. J. T.; J. H.; Whittaker, A. K. J. Polym. Sci, Part B 2000, in press. (16) Hill, D. J. T.; Lim, M. C. H.; Whittaker, A. K. Polym. Int. 1999, 48, 1046. (17) Hill, D. J. T.; Moss, N.; Pomery, P. J.; Whittaker, A. K. Polymer 2000, 41, 1287.

Ghi et al. (18) Franson, M.; Peppas, N. A. J. Appl. Polym. Sci. 1983, 28, 1299. (19) Riggs, P. D.; Braden, M.; Tilbrook, D. A.; Swai, H.; Clarke, R. L.; Patel, M. P. Biomaterials 1999, 30, 435.

BM000146Q