Biomacromolecules 2004, 5, 971-976
971
NMR Imaging of the Diffusion of Water at 37 °C into Poly(2-hydroxyethyl methacrylate) Containing Aspirin or Vitamin B12 Mohammad A. Chowdhury,† 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 QLD 4072, Australia Received November 21, 2003; Revised Manuscript Received February 16, 2004
The ingress of water into poly(2-hydroxyethyl methacrylate), PHEMA, loaded with either one of two model drugs, vitamin B12 or aspirin, was studied at 37 °C using three-dimensional NMR imaging. PHEMA was loaded with 5 and 10 wt % of the drugs. From the imaging profiles, it was observed that incorporation of vitamin B12 into PHEMA resulted in enhanced crack formation on sorption of water and the crack healing behind the diffusion front was slower than for PHEMA without added drug. This was accounted for by the anti-plasticization of PHEMA by vitamin B12. Crack formation was inhibited in the PHEMA-aspirin systems because of the plasticizing effect of the aspirin on the PHEMA matrix. All of the polymers were found to absorb water according to an underlying Fickian diffusion mechanism. For PHEMA loaded with 5 wt % of aspirin or vitamin B12, the best values of the water diffusion coefficients were both found to be 1.3 ( 0.1 × 10-11 m2 s-1 at 37 °C, while the values for the polymer loaded with 10 wt % of the drugs were slightly higher, 1.5 ( 0.1 × 10-11 m2 s-1. Introduction NMR imaging (magnetic resonance imagin, MRI) has been developed as a very powerful tool in medical diagnostics and materials science,1 and in recent years it has also been used to study the controlled release of drugs from hydrophilic polymer matrixes.2 The technique can provide information on either the penetrant sorption or drug release processes at the molecular level.2 For example, MRI has recently been used to observe the ingress of penetrants into solid systems3-9 and to provide two- or three-dimensional, time-dependent images of the density of these highly mobile components in a material. Because it provides direct information on the nature of the penetrant diffusion front during sorption, it provides direct information on the penetrant diffusion mechanism. Ghi et al.8-11 have reported the use of mass uptake and MRI to study the sorption of water into poly(2-hydroxyethyl methacrylate), PHEMA. They found that the underlying profile of the water diffusion front in this polymer was Fickian in nature, but that cracks formed in the glassy core near the interface between the glassy core and the rubbery region. These cracks were found to contain water that had a long T2 NMR relaxation time, characteristic of water in larger pools.8,10 Cracks formed in the polymer matrix during the sorption of a penetrant have an influence on the overall water sorption process. To complement the previous studies on PHEMA in the absence of added drugs9 and a more recent study on PHEMA * To whom correspondence should be sent. E-mail: hill@chemistry. uq.edu.au. † Department of Chemistry, The University of Queensland. ‡ Centre for Magnetic Resonance, The University of Queensland.
when model drugs are present,15 a MRI study of the water diffusion front during sorption of water into two PHEMAdrug systems has been undertaken. The two drugs chosen for study were vitamin B12 and aspirin, which are soluble in water and which vary significantly in molecular size, having molecular masses of 1355 and 180 g mol-1, respectively. The aim of the present study was to investigate further the processes that control the diffusion of water in PHEMA systems and to observe the impact of the presence of the drugs during polymerization on the behavior of the polymer, the water diffusion front, and the water diffusion coefficent. The systems studied contained the two model drugs at concentrations of 5 and 10 wt %. Experimental Section Materials. 2-Hydroxyethyl methacrylate, HEMA, was obtained from Rocryl or Sigma Aldrich and was stabilized with monomethyl ether hydroquinone. To remove the stabilizer, the HEMA was passed through an anhydrous alumina column and then further purified by distillation. The colorless middle fraction with a boiling point of 76-78 °C (6.6-6.7 × 102 Pa) was collected for use. The purified monomer was stored in dark brown containers at 4 °C and used within 1 week of distillation. The initiator, dibenzoyl peroxide, was purified by recrystallization twice from a mixture of chloroform and methanol. The crystals were filtered, dried under a vacuum, and stored in a dark brown container at 4 °C. Acetyl salicylic acid (aspirin) and cyanocobalamin (vitamin B12) were obtained from ICN biomaterials, and they were used without further purification.
10.1021/bm030079a CCC: $27.50 © 2004 American Chemical Society Published on Web 03/30/2004
972
Biomacromolecules, Vol. 5, No. 3, 2004
Preparation of the Polymers and Drugs Loading. A uniform molecular dispersion of the model drugs in PHEMA was accomplished by incorporating the drugs in the monomer before polymerization. The required amounts of monomer, initiator, and model drug were mixed in a volumetric flask to give an initiator concentration of 0.08 wt %, and the desired drug concentration of either 5 or 10 wt %. Samples of the mixtures were subsequently transferred into a cylindrical Teflon mould with an internal diameter of 4.5-4.7 mm and a length of ≈50 mm. A lid on the mould prevented the loss of monomer, and a small hole in the lid facilitated the release of any dissolved gas from the mixture on evacuation. The mould was then placed in a vacuum oven, the air was removed to degas the mixture, and then the oven was filled with dry nitrogen gas. Thermal initiation of the polymerization of the monomer was effected by setting the oven temperature to 50 ( 2 °C for 20 h followed by 2 h at 80 ( 2 °C to ensure full conversion. After the cylinders of polymer were removed from the oven, they were checked for complete conversion over their length by Fourier transform near-infrared spectroscopy as described previously.11 No evidence was found for the presence of any unreacted double bonds, which have a near-infrared absorbance at 6170 cm-1. The polymer cylinders were rigid and glassy and showed no evidence for the presence of any water-soluble polymer, even though the presence of the aspirin and vitamin B12 may influence the polymerization processes and reduce the average kinetic chain lengths through chain transfer reactions. A water mass uptake study15 on the cylinders showed that they behaved very similarly to PHEMA cylinders prepared in the absence of any drugs. MRI Study of the Polymers. Samples with cylinder diameters of 4.5-4.7 mm fit neatly into the NMR resonator used for the imaging studies. The advantages of a choice of a cylindrical geometry have been described elsewhere.9,12 The samples were placed in distilled water at 37 °C and were removed periodically, their surface dried with a soft paper tissue, and they were placed in the NMR resonator. To detect the ingress of the water into the polymer and to identify the nature of the diffusion front, proton images of the samples were obtained using an AMX 300 NMR spectrometer. The temperature within the NMR cavity was 295 K, and during the time period of a measurement, the inner water profile of the image was stable. T2-weighted NMR images consisting of 128 × 128 × 8 voxels were obtained across the central region of the cylinders. The images were obtained using a threedimensional spin-echo method with the following parameters: read gradient strength 0.303 T m-1; excitation pulse 90°; duration 14.0 µs; refocusing pulse 180° of 28.0 µs; echo time 3.531 ms; and recovery time 3 s. All the images have an in-plane resolution of 78 × 78 µm and a slice thickness of 5.00 mm in a field of view of 1 × 1 × 4 cm. Depending on the circumstances, 2-28 averages were co-added to increase the signal-to-noise ratio. The MRI images were analyzed using the NIH image13 program to extract the waterconcentration profile across the cylinders.
Chowdhury et al.
Figure 1. MRI of the water protons in a PHEMA cylinder containing 10 wt % aspirin after water sorption at 37 °C for (A) 6.3 h and (B) 18.1 h.
Results and Discussions During the course of penetrant diffusion into a polymer cylinder, different aspects of the diffusion phenomena are observed at different stages of the sorption process. In the first stage, an outer rubbery region of the polymer is present along with a glassy core. During this stage, the radius of the cylinder increases slightly but the length of the cylinder remains fixed at a value determined by the length of the glassy core. Longitudinal and radial stresses develop in the rubbery and glassy regions in this regime. In the second stage, the sorption process has proceeded to the extent that the glassy core of the sample disappears. Hence, the longitudinal and radial stresses are reduced, the length of the cylinder starts to increase, and the radius decreases as these stresses relax. In the third stage, as the water uptake continues toward equilibrium, for some polymers the diffusion characteristics change and a sorption overshoot phenomenon is observed as the polymer chains relax. In these latter cases, water is ejected from the gel as equilibrium is established. Finally, in the fourth stage, the polymer gel is at its equilibrium level. For simple Fickian diffusion, the profile for the concentration of the penetrant across a cylinder is given by eq 1.14 This equation was used by Ghi et al.8,9 to fit the underlying water concentration profile during the sorption of water into a cylinder of PHEMA of radius a in the absence of added drugs: Cr,t C0,∞
)1-
2
∞
∑
1 J0(rRn)
an)1Rn J1(aRn)
exp(-DRn2t)
(1)
where t is the time for which penetrant diffusion has occurred, 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, J0(x) is a Bessel function of the first kind of zero order, and J1(x) is a Bessel function of first order. To obtain a “best value” of the diffusion coefficient from a curve fit to the time-dependent, MRI water concentration profiles using eq 1, Ghi et al. used a summation over 150 roots of the Bessel function and a χ2 test of the goodness of fit.9 A similar curve fitting procedure was adopted in the current work. Figure 1 shows two typical T2-weighted MRI images of the water protons in a cylinder of PHEMA containing 10 wt % aspirin obtained during the water sorption process. To
Diffusion of Water into PHEMA
Figure 2. Relative concentration of water (C/C0) profile across the cylinder after water sorption at 37 °C for 6.3 h as determined from the image in Figure 1A.
carry out an analysis of the distance and time dependence of the water concentration in the system, a cross section of the water concentration along the image diameter was used. A typical cross section of this type is shown in Figure 2, which was taken from the profile in Figure 1 obtained after 6.3 h. The water concentration profile in Figure 2 has been normalized using the concentration of water at the surface of the cylinder.
Biomacromolecules, Vol. 5, No. 3, 2004 973
PHEMA-Vitamin B12 Systems. In Figure 3, a series of representative T2-weighted water concentration profiles are presented for the sorption of water at 37 °C at various sorption times into PHEMA-containing 5 wt % vitamin B12. The water concentration profiles in Figure 3A,B were obtained at sorption times before the loss of the glassy core of the cylinder. They clearly show an initial decreasing water concentration with distance into the cylinder, which is typical of the profile expected for a Fickian diffusion mechanism. (For a case II diffusion mechanism the water concentration in this initial region of the water uptake would be constant.) However, near the interface between the region swollen by the water and the glassy core, inner features are present in the concentration profiles shown in Figure 3A,B. The profiles in this regime of the water sorption are similar to those observed previously for PHEMA in the absence of any occluded drugs, for which the observed inner features have been identified as being characteristic of water in cracks.8,9 Because the water concentration profiles obtained by MRI are weighted by the T2 of the local water protons in the matrix and the water in cracks behaves somewhat like bulk water which has a longer T2 relaxation time than the protons of “bound” water resident in the rubbery region, the water concentration profiles in the vicinity of cracks will be distorted9 (increased), as observed in Figure 3A,B.
Figure 3. Profiles of the relative concentration of water on sorption at 37 °C into a PHEMA cylinder containing 5 wt % vitamin B12 after (A) 1.5 h, M/M∞ ) 0.23; (B) 6.1 h, M/M∞ ) 0.46; (C) 10.1 h, M/M∞ ) 0.59; and (D) 83.7 h, M/M∞ ) 1.00. Dashed curve for D ) 1.3 × 10-11 m2 s-1.
974
Biomacromolecules, Vol. 5, No. 3, 2004
Chowdhury et al.
Figure 4. Profiles of the relative concentration of water on sorption at 37 °C into a PHEMA cylinder containing 10 wt % vitamin B12 after (A) 5.1 h, M/M∞ ) 0.49; (B) 9.2 h, M/M∞ ) 0.66; and (C) 12.9 h, M/M∞ ) 0.76. Dashed curve for D ) 1.5 × 10-11 m2 s-1.
As shown previously for PHEMA,9 the interfacial cracks are “repaired” in the rubbery region of the swelling matrix. This occurs once the local water concentration reaches a value sufficient to lower the polymer Tg to below the measurement temperature, 37 °C. Thus, the inner features in the concentration profiles appear to progress through the matrix along with the diffusion front. When the glassy core of the polymer disappears, at a mass uptake of approximately 0.5-0.6 relative to the equilibrium value, there are no inner features in the concentration profiles (see Figure 3C,D). In this region of the sorption process, the concentration profiles are clearly Fickian-like. The dashed curve in Figure 3B has been generated for a sorption time of 6.1 h using eq 1 and a value of the diffusion coefficient of 1.3 × 10-11 m2 s-1, which is similar to that obtained from an analysis of the mass uptake data at 37 °C, 1.4 × 10-11 m2 s-1.15 The curve clearly shows that the underlying water concentration profiles can be predicted well using this value of the diffusion coefficient, thereby confirming that the water diffusion process in the PHEMA-vitamin B12 systems is Fickian. Similar analyses for a range of different sorption times were carried out, and a diffusion coefficient of 1.3 ( 0.1 × 10-11 m2 s-1 could describe the data for all of the profiles obtained prior to the loss of the glassy core. Observations made for the PHEMA-vitamin B12 system containing 10 wt % of the drug were similar to those found for the system with 5 wt % B12, as demonstrated in Figure 4. In Figure 4A, a profile obtained at a relative mass uptake of 0.49, prior to the loss of the glassy core, is shown. Again, the presence of water in cracks at the interface between the rubbery region and the glassy core is evident in the profile. After disappearance of the glassy core, the concentration profile shows no anomalous features and is clearly Fickianlike up to the attainment of equilibrium, as shown in Figure 4B,C. A best-fit Fickian curve was generated with eq 1 using the data in the outer region of the profile with a diffusion coefficient of 1.5 × 10-11 m2 s-1. The curve provides an excellent representation of the underlying water concentration in the rubbery region (see Figure 4A), again confirming that a Fickian mechanism provides a good representation of the water diffusion process in this PHEMA-drug system.
The diffusion coefficient obtained from analyses of profiles for a range of sorption times prior to the loss of the glassy core was 1.5 ( 0.1 × 10-11 m2 s-1, which is in close agreement with the value of 1.6 ( 0.1 × 10-11 m2 s-1 obtained for this PHEMA-drug system at 37 °C from a mass uptake study.15 We have previously observed9 that the bestfit value of the underlying diffusion coefficient obtained from the MRI water concentration profiles is slightly smaller than the value of the diffusion coefficient obtained from mass uptake data. This is believed to be due to the influence of end effects and crack formation which are not accounted for adequately in determining the coefficient from the mass uptake data. MRI of PHEMA-Aspirin Systems. NMR imaging experiments during the diffusion of water into PHEMAaspirin systems (5 and 10 wt % of aspirin) were carried out, and some typical results for the water concentration profiles are presented in Figures 5 and 6. The results for diffusion of water into a PHEMA cylinder containing 5 wt % aspirin at 37 °C are shown in Figure 5. The shapes of the profiles for the PHEMA-aspirin system prior to the loss of the glassy core in the swelling polymer, shown in Figure 5A,B, are clearly different from those observed during the equivalent regime for pure PHEMA9 and for PHEMA containing vitamin B12. The profiles in Figure 5A,B show no evidence for the presence of significant inner features similar to those observed in Figure 3. However, the profile in Figure 5B appears to be slightly convex in shape near the outer edge of the cylinder, which is not in accord with the predictions of a Fickian diffusion model, which may be due to some interfacial cracking. The dashed line in Figure 5B shows the profile generated using eq 1 for Fickian diffusion with a diffusion coefficient of 1.3 × 10-11 m2 s-1, which is consistent with the value obtained from a gravimetric sorption analysis,15 1.4 ( 0.1 × 10-11 m2 s-1. The dashed line demonstrates the poor fit of the predicted curve to the experimental profile near the edges. On the other hand, the concentration profile shown in Figure 5C, obtained after the glassy core has disappeared, does adhere more closely to that expected for Fickian diffusion, and rules out the possibility of case II type diffusion of water into the matrix. (The slight bulge in the
Diffusion of Water into PHEMA
Biomacromolecules, Vol. 5, No. 3, 2004 975
Figure 5. Profiles of the relative concentration of water on sorption at 37 °C into a PHEMA cylinder containing 5 wt % aspirin after (A) 4.5 h, M/M∞ ) 0.40; (B) 9.8 h, M/M∞ ) 0.57; and (C) 14.4 h, M/M∞ ) 0.71. Dashed curve for D ) 1.3 × 10-11 m2 s-1.
Figure 6. Profiles of the relative concentration of water on sorption at 37 °C into a PHEMA cylinder containing 10 wt % aspirin after (A) 1.2 h, M/M∞ ) 0.25, and (B) 6.3 h, M/M∞ ) 0.48.
profile in Figure 5C at the outer edges is caused by a small loss of near-surface water, either during drying or by evaporation during the MRI measurement. This does not affect the inner part of the water profile.) Thus, the observations suggest that the deviations of the experimental data from the predictions for Fickian diffusion in Figure 5B have their origins in the formation of cracks in the matrix but that the extent of crack formation is less than that found for PHEMA in the absence of any drugs and PHEMA-B12 systems. This proposal is consistent with the previously demonstrated plasticizing effect of the aspirin on the PHEMA matrix,15 which would be expected to reduce the tendency for cracking to occur and to enhance the crack repair processes. However, plasticization of the matrix may not be the only reason for the absence of crack formation, and for example, the presence of the drug during polymerization may also play a role. On addition of 10 wt % aspirin to the PHEMA matrix, again there is no evidence of any significant anomalous inner features in the water concentration profiles obtained prior to the disappearance of the glassy core (see Figure 6A,B). Hence, there is no evidence for significant crack formation during water sorption by this system, thereby confirming the observations made for PHEMA containing 5 wt % aspirin.
Conclusions NMR imaging of the water diffusion front has confirmed that the Fickian diffusion mechanism provides the best representation of the diffusion process for water into the PHEMA-drug systems studied herein. The PHEMA-B12 systems displayed extensive crack formation at the interface between the rubbery region and the glassy core during water sorption. This contrasted with the PHEMA-aspirin systems for which crack formation was not as significant as that observed for PHEMA in the absence of added drug. These observations were attributed to the anti-plasticization effect of the B12 and the plasticization effect of aspirin on PHEMA. For PHEMA loaded with aspirin and vitamin B12 at a level of 5 wt % the water diffusion coefficient at 37 °C was found to be 1.3 ( 0.1 × 10-11 m2 s-1. This value of the diffusion coefficient is similar to that for water sorption into PHEMA at 37 °C 9 in the absence of drugs, 1.5 ( 0.1 × 10-11 m2 s-1, which was also obtained from curve fits to water concentration profiles. The corresponding value of the water diffusion coefficient for the PHEMA-drug systems containing 10 wt % of the drugs at 37 °C was found to be 1.5 ( 0.1 × 10-11 m2 s-1, which is slightly higher than for 5 wt % of drugs. Acknowledgment. The authors wish to acknowledge the financial support for this work from the Australian Research Council.
976
Biomacromolecules, Vol. 5, No. 3, 2004
References and Notes (1) Watanabe, T. Nucl. Magn. Reson. 2001, 30, 453-476. (2) Fyfe, C. A.; Grondey, H.; Blazek-Welsh, A. I.; Chopra, S. K.; Fahie, B. J. J. Controlled Release 2000, 68, 73-83. (3) Knorgen, M.; Arndt, K.-F.; Richter, S.; Kuckling, D.; Schneider, H. J. Mol. Struct. 2000, 554, 69-79. (4) Riggs, P. D.; Kinchesh, P.; Braden, M.; Patel, M. P. Biomaterials 2001, 22, 419-427. (5) Webb, A. G.; Hall, L. D. Polymer 1991, 32, 2926-2938. (6) Grinsted, R. A.; Koenig, J. L. Macromolecules 1992, 25, 12291234. (7) Hyde, T. M.; Gladden, L. F.; Mackley, M. R.; Gao, P. J. Polym. Sci., Part A: Polym. Chem. 1995, 33, 1795-1806. (8) Ghi, P. Y.; Hill, D. J. T.; Maillet, D.; Whittaker, A. K. Polymer 1997, 38, 3985-3989. (9) Ghi, P. Y.; Hill, D. J. T.; Whittaker, A. K. Biomacromolecules 2001, 2, 504-510.
Chowdhury et al. (10) Ghi, P. Y.; Hill, D. J. T.; Whittaker, A. K. Biomacromolecules 2002, 2, 991-997. (11) Ghi, P. Y.; Hill, D. J. T.; Whittaker, A. K. J. Polym. Sci., Polym. Phys. Ed. 2000, 38, 1939-1946. (12) Chowdhury, M. A. Diffusion and Drug Release Study from 2-Hydroxyethyl Methacrylate (HEMA) - Based Methacrylate Polymers. Ph.D. Thesis, The University of Queensland, Brisbane, Australia, 2003. (13) NIH Image, version 1.61; Research Services Branch, National Institute of Mental Health, National Institutes of Health: Washington, D.C., 2000. (14) Crank, J.; Park, G. S. Methods of Measurement. In Diffusion in Polymers; Crank, J., Park, G. S., Eds.; Academic Press: London, 1968; Chapter 1, pp 1-39. (15) Chowdhury, M. A.; Hill, D. J. T.; Whittaker, A. K. Polym. Int. 2003, submitted for publication.
BM030079A
