- Email: [email protected]
Microgel particles containing methacrylic acid: pH-triggered swelling behaviour and potential for biomaterial application
Journal of Colloid and Interface Science 316 (2007) 367–375 www.elsevier.com/locate/jcis
Microgel particles containing methacrylic acid: pH-triggered swelling behaviour and potential for biomaterial application Sarah Lally a , Paul Mackenzie a , Christine L. LeMaitre b , Tony J. Freemont b , Brian R. Saunders a,∗ a Manchester Materials Science Centre, School of Materials, The University of Manchester, Grosvenor Street, Manchester, M1 7HS, UK b Tissue Injury and Repair Group, School of Medicine, Stopford Building, The University of Manchester, Oxford Road, Manchester, M13 9PT, UK
Received 22 June 2007; accepted 13 August 2007 Available online 19 August 2007
Abstract pH-responsive microgels are crosslinked polymer particles that swell when the pH approaches the pKa of the ionic monomer incorporated within the particles. In recent work from our group it was demonstrated that the mechanical properties of degenerated intervertebral discs (IVDs) could be restored to normal values by injection of poly(EA/MAA/BDDA) (ethylacrylate, methacrylic acid and butanediol diacrylate) microgel dispersions [J.M. Saunders, T. Tong, C.L. Le Maitre, T.J. Freemont, B.R. Saunders, Soft Matter 3 (2007) 486]. In this work we report the pH dependent swelling and rheological properties of poly(MMA/MAA/EGDMA) (methylmethacrylate and ethyleneglycol dimethacrylate) microgel dispersions. This system was investigated because it contains monomers that are already used as biomaterials. The poly(MMA/MAA/EGDMA) particles exhibit pH-triggered volume swelling ratios of up to ca. 250. The swelling onset for these particles occurs at pH values greater than ca. 6.0. A pKa for these particles of ca. 6.7 is consistent with titration and swelling data. Fluid-to-gel phase diagrams for concentrated poly(MMA/MAA/EGDMA) dispersions were determined as a function of polymer volume fraction and pH using tube-inversion measurements. The rheological properties for the gelled microgel dispersions were investigated using dynamic rheology measurements. The elastic modulus data for the poly(MMA/MAA/EGDMA) gelled dispersions were compared to data for poly(EA/MAA/BDDA) microgels. A similar pH-dependence for the elastic modulus was apparent. The maximum elastic modulus was achieved at a pH of about 7.0. The elastic modulus is an exponentially increasing function of polymer volume fraction at pH 7.0. Preliminary cell challenge experimental data are reported that indicate that gelled poly(MMA/MAA/EGDMA) microgel dispersions are biocompatible with cells from human intervertebral discs. However, the duration over which these experiments could be performed was limited by gradual redispersion of the gelled microgel dispersions. Based on the results presented it is suggested that poly(MMA/MAA/EGDMA) microgel would be a good candidate as a biomaterial for structural support of soft connective tissues. © 2007 Elsevier Inc. All rights reserved. Keywords: Microgel; Responsive polymer; Gel; Biocolloid; Intervertebral disc
1. Introduction For the present work pH-responsive microgels can be defined as cross-linked polymer particles that swell when the pH approaches the pKa of the incorporated methacrylic acid units. Microgels have been reviewed elsewhere [1,2]. The first waterdispersible responsive microgel was reported by Pelton and Chibante [3] in 1986. In previous work from our group [4] it was established that the height and mechanical properties of * Corresponding author.
E-mail address: [email protected] (B.R. Saunders). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.030
degenerated intervertebral discs (IVDs) could be restored by injection of poly(EA/MAA/BDDA) microgel dispersion followed by pH triggered swelling. (EA, MAA and BDDA are ethylacrylate, methacrylic acid and butanediol diacrylate, respectively.) Poly(EA/MAA/BDDA) microgel was originally developed by Rodriguez et al. [5] In this study we replaced EA and BDDA with components which should be more acceptable for potential use in the body. The comonomers chosen were MMA and EGDMA, which are methylmethacrylate and ethyleneglycol dimethacrylate, respectively. The aims of this study were to investigate the pH-dependent swelling behaviour of two poly(MMA/MAA/EGDMA) microgels and compare the data with those for poly(EA/MAA/BDDA) microgel to determine
368
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
whether the former systems would be good candidates for potential use as injectable structural supports in the body. There are several criteria that an injectable biomaterial for load support in the body must meet. It must flow during injection, fill the irregular shape of the internal cavity and then subsequently form an immobilised solid, preferably a gel when required. It should also be biocompatible [6]. In order to provide structural support for biomechanical loads the gel must have a high swelling pressure. We previously investigated poly(EA/MAA/BDDA) microgels that contain high concentrations of MAA groups [4]. The dispersions had relatively low viscosity at low pH. The particles swelled considerably as the pKa of the MAA groups was approached and the dispersion formed a gelled microgel dispersion. The steric confinement of neighbouring swollen particles is responsible for the fluid-togel transition. MMA and EGDMA were selected for use within the pHresponsive microgels in this study because they have been used in bone cement and contact lenses [7,8]. There has been very little work reported in the literature for poly(MMA/MAA/ EGDMA) microgel dispersions. The pH-dependent swelling properties of poly(MMA/MAA/EGDMA) microgel particles containing 17 mol% MAA were reported by Saunders et al. [9] in 1997. They studied the osmotic deswelling of the particles in the presence of non-adsorbing polymer. Their microgel particles, which were prepared using surfactant-free emulsion polymerisation, exhibited much lower swelling ratios than those for the poly(MMA/MAA/EGDMA) particles studied in this work. The properties of concentrated dispersions were not considered in that work. The poly(MMA/MAA/EGDMA) microgels dispersions considered in this work were prepared using seedfeed emulsion polymerisation and have twice the level of MAA present. They exhibit much greater swelling ratios than those prepared in the earlier work. pH-responsive microgel dispersions have been the subject of considerable interest within the microgel literature. A large proportion of the work has focused on microgels containing carboxylic acid groups. Surprisingly, there has been very little work reported on the rheological properties of gelled microgel dispersions. Rodriguez et al. [5] were the first to report the pH-dependent swelling behaviour of poly(EA/MAA/BDDA) microgel dispersions. Their studies were confined to dilute dispersions. They concluded that the particles had a core-shell structure with a lightly crosslinked shell. Tan et al. [10] studied the behaviour of poly(EA/MAA/DAP) microgels (DAP is diallyl phthalate) using static rheology. They reported that the swelling ratio scaled with the MAA content raised to the power of 1.5. Hoare and Pelton [11] reported an insightful study of the effect of carboxylic acid containing monomers on the structure and swelling of poly(NIPAM) (N -isopropylacrylamide) microgels. They identified a link between the apparent pKa of the acid groups and the reactivity ratios of the co-monomers. Reactivity ratios that favoured formation of blocks of acid groups within the microgel network were proposed to cause an increase in the local pKa due to the polyelectrolyte effect. The latter refers to the effect whereby deprotonation of one carboxylate group causes increased electrostatic repulsion between OH− groups
Table 1 Nominal compositions of microgels No.
Composition
MMA or EA (wt%)
MAA (wt%)
EGDMA or BDDA in monomer feeda (wt%)
1 2 3
Poly(MMA/MAA/EGDMA) Poly(MMA/MAA/EGDMA) Poly(EA/MAA/BDDA)
64.0 64.1 66.0
35.7 35.7 33.0
0.4 1.0 1.0
a The seed contained 1.0 wt% EGDMA or BDDA with respect to total monomer mass in each case. The total monomer mass used to during the seed stage was about 8–9 wt% of that used to prepare the particles (see Section 2). Note that EGDMA and BDDA have the same molar mass.
and the charged portion of chain which opposes subsequent deprotonation of the neighbouring carboxylate group. Our previous work using poly(EA/MAA/BDDA) microgel dispersions showed that conditions that gave the maximum elastic modulus, as judged by dynamic rheology, also gave the greatest extent of structural support for degenerated IVDs containing injected microgel. In this work we compare the swelling and rheological data for two new poly(MMA/MAA/EGDMA) microgels to that for poly(EA/MAA/BDDA). Uniaxial compression of IVDs containing poly(MMA/MAA/EGDMA) microgels was not studied in this work. In this work we report colloidal characterisation data in order to improve the structure– property understanding for this class of responsive polymer colloids. By comparison of the rheological and swelling data to the reference system [poly(EA/MAA/BDDA)] we seek to establish design rules for constructing colloidal biomaterials with potential for load-bearing applications within the body in the future. 2. Experimental 2.1. Preparation of microgels Table 1 shows a summary of the compositions for the microgels discussed in this work. All microgels were prepared using a seed-feed method. Two variations of this method were used. For Microgel 1 the feed had a lower weight fraction of EGDMA than the seed. A seed monomer mixture containing MMA (139.3 g), MAA (60.1 g) and EGDMA (2.1 g) was prepared and 25.2 g (8.8% of total monomer) of the mixture added to a nitrogen purged, stirred, solution of SDS (1.75 g in 500 g of water), which had been heated to 80 ◦ C. K2 HPO4 (2.55 g of 7 wt% solution in water) and ammonium persulfate (APS) (4.50 g of 3 wt% solution in water) were immediately added whilst maintaining a nitrogen atmosphere. A feed monomer mixture containing MMA (273.1 g), MAA (134.0 g) and EGDMA (1.7 g) was prepared. After the appearance of a slight blue turbidity, 264 g (91.2% of total monomer) of the feed monomer mixture was added at a constant rate over a 90 min period. Additional APS (2.73 g of 3% solution in water) was added after the feed and the temperature maintained for a further 2 h. The dispersion was extensively dialysed against Milli-Q quality water. Microgel 2 was prepared in the presence of reduced SDS concentration and added NaCl in order to increase the particle
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
size. A monomer mixture containing MMA (125.6 g), MAA (58.7 g) and EGDMA (1.94 g) was prepared. A total of 7.5% of this mixture was used as the seed. This was added to a nitrogen purged, stirred, solution of SDS and NaCl (0.14 g and 0.44 g, respectively, in 200 g of water), which had been heated to 80 ◦ C. K2 HPO4 (2.56 g of 7 wt% solution in water) and APS (2.57 g of 5 wt% solution in water) were immediately added whilst maintaining a nitrogen atmosphere. The remainder of the monomer mixture (92.5% of total) was added to a stirred solution of SDS (0.17 g in 300 g of water) to form an emulsion phase. After one hour of reaction, the emulsified feed monomer mixture, under constant stirring, was added to the reaction at a constant rate over a 105 min period. The reaction temperature was maintained for a further 2 h. The microgel was extensively centrifuged and washed with Milli-Q quality water. Microgel 3 was prepared according to the method reported by Rodriguez et al. [5] and is described fully elsewhere [4]. The method used was similar to that described above for Microgel 1. The weight fraction of BDDA used in the seed and feed were the same (Table 1). It was also purified by dialysis. 2.2. Physical measurements Photon correlation spectroscopy measurements were performed using a BI-9000 Brookhaven light scattering apparatus (Brookhaven Instrument Cooperation), fitted with a 20 mW He–Ne laser and the detector was set at 90◦ scattering angle. The measured data were analysed using cumulant analysis. The swelling ratio (q) was calculated using equation: d , q= (1) dc where d is the hydrodynamic diameter and dc is the average diameter of collapsed particles. In this work we use the numberaverage diameter determined from SEM to estimate dc . The errors (one standard deviation) for the d and q values were estimated as 5 and 15%, respectively. SEM measurements were obtained using a Philips FEGSEM instrument. The samples were prepared using samples at pH values of ca. 4. This pH is well below the estimated pKa for these particles (below). These conditions were used to avoid the possibility of particle deformation of swollen particles during the drying process. The effect of pH on the average diameter of the particles, as measured by SEM, was not investigated in this work. At least 250 particles were counted to obtain number-average diameters. Fluid-to-gel phase diagrams were constructed by mixing known ratios of microgel with NaOH solution and then tested using tube inversion. Gels were considered to be present if the samples did not flow when the tube was inverted. Rheology measurements were performed using a Rheometrics RMS-800. A cone and plate measurement geometry was used. The diameter, cone angle and gap were 25 mm, 0.1 radians and 1.3 mm, respectively. A frequency range of 0.01–15.9 Hz was employed. The strain used for these measurements was 10%. Titration data were obtained from Synthomer Ltd. (Harlow Essex) using automated titration and standardised
369
NaOH solution. The pKa value was taken at the pH at which 50% of the MAA groups were neutralised. 2.3. Cell challenge experiments Agar (1.5 wt%) was dissolved in water with heating and the solution cooled to form a gel. The latter was used as a control. The gelled microgel dispersions were prepared at pH 7 and φp = 0.10. Adherent cultured human annulus fibrosus (AF) cells were used for the cell challenge experiments. They were washed free of media using PBS. The cells were incubated for about 2 min with trypsin which was deactivated with 5 ml of complete media (DMEM). The AF cells were concentrated and then samples of gelled microgel place on top of the cells in the presence of medium. This ensured contact of the gel and cells. The cells were view from beneath the well plate using optical microscopy. 3. Results and discussion 3.1. pH-triggered swelling of microgel particles containing methacrylic acid In this work we synthesised two new poly(MMA/MAA/ EGDMA) microgels (Microgels 1 and 2, see Table 1). One difference between Microgels 1 and 2 is that Microgel 1 was prepared using reduced crosslinking monomer weight fraction (with respect to total monomer mass) during the feed stage. Microgel 2 was prepared using conditions designed to give relatively large particles (cf. Microgels 1 and 3). It can be seen from Table 2 that Microgels 1 and 3 had comparable collapsed diameters. The diameter for Microgel 2 was significantly larger. This was due to the presence of added NaCl and the lower concentration of SDS used during microgel synthesis. These conditions promoted a greater extent of aggregation during the particle formation stage. SEM micrographs for Microgels 1 and 2 are shown in Fig. 1. The SEM data showed that the collapsed particles were spherical with little tendency to deform during the drying process. This means that average diameters calculated from SEM data can be used to estimate the collapsed diameter (dc ). Table 2 shows the average diameters of the particles measured by SEM and those measured by PCS (d) at pH 4. It can be seen that dc is less than or equal to d for each system. This is additional support for a lack of particle deformation to a pancake-like structure as a result of SEM Table 2 Particle size and swelling data for the microgels Microgel No.
dc (nm)a
d (nm)b
qc
1 2 3d
82 ± 9 210 ± 23 65 ± 7
105 ± 5 245 ± 10 75 ± 3
210 ± 32 250 ± 38 105 ± 16
a Number-average particle size from SEM data. The coefficient of variation for the particle size distributions had the same value (11%). b Hydrodynamic diameter measured at pH 4. c The swelling ratio calculated at pH 9 and using Eq. (1). d The data shown for Microgel 3 were obtained from Ref. [4].
370
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
ratios than for Microgel 3. Because of the synthetic method used to prepare these particles (seed-feed emulsion polymerisation) it is reasonable to suggest that Microgels 1 and 2 have core–shell structures. A core-shell structure has been shown to be present for poly(EA/MAA/BDDA) microgel particles [5]. Titration data published by Rodriguez et al. [5] for poly(EA/ MAA/BDDA) prepared in the same manner as Microgel 3 show that at pH 7.0 the degree of neutralisation was 50%. Microgel 3 was titrated for the present study (data not shown) and a pH of 6.7 was evident at 50% neutralisation. Due to the similarity in the pH-dependence of the swelling transitions (Fig. 2) we assume that the pKa for Microgels 1, 2 and 3 are the same, i.e., about 6.7. Fig. 2 also shows the ionicity (i), which is the fraction of MAA units that are deprotonated. These values were calculated using a pKa of 6.7 and the following equation i=
Fig. 1. Scanning electron micrographs of (a) Microgel 1 and (b) Microgel 2. The sample plane was viewed at an angle of ca. 45◦ for the image shown in (b). The horizontal line represents 1 µm and applies to both images.
Fig. 2. Variation of hydrodynamic diameter with pH for Microgels 1, 2 and 3. The ionicity calculated using Eq. (2) and pKa is shown. The error bars for the Microgels 1 and 3 data are smaller than the symbols.
sample preparation because such a process would increase the apparent diameter. Fig. 2 shows the variation of the hydrodynamic diameter with pH for each of the microgels. Swelling ratios for each microgel are shown in Table 2. These values were calculated using the hydrodynamic diameters measured at pH 9. It can be seen from Fig. 2 that each of the microgels exhibits a pH-dependent swelling transition with an onset pH estimated to be about 6.0. The region of maximum diameter increase is between pH values of 6.0 and 8.0. Approximately 50% of the volume increase has occurred at pH 7.5. Little if any further swelling increase occurs after the pH has reached 8.0. It can be seen from the data shown in Table 2 that the swelling ratios for Microgels 1 and 2 are comparable. Interestingly, both of the microgels had significantly greater swelling
1 . 1 + 10pKa -pH
(2)
The use of a pKa of 6.7 for all three microgels appears to be a reasonable approximation. The pKa for the MAA groups within poly(EA/MAA/BDDA), and most likely the poly(MMA/MAA/EGDMA) microgels, is within the range of pKa ’s reported for poly(MAA) by Spencer et al. [12] Hoare and Pelton [11] used co-monomer reactivity ratios to explain the variation of pKa of ionic poly(NIPAM)based microgels. They argued that the high pKa value for poly(NIPAM/MAA) microgels (in the region of 6.2) was due to formation of MAA blocks within the polymer network. The Q and e values for MMA and MAA have been published [13]. Using the Alfrey–Price equation the reactivity ratios were calculated as rMMA = 0.87 and rMAA = 1.09. (These are the same values used by other groups [14].) If these reactivity ratios were present during emulsion polymerisation then the probabilities of adding MMA or MAA to a growing chain end would be similar. Under those circumstances, and considering the approximate two-to-one molar MMA-to-MAA stoichiometry used during microgel preparation, blocks of MAA groups should be present. Accordingly, a pKa similar to the range reported for poly(MAA) is expected for poly(MMA/MAA/EGDMA) microgel particles. An additional explanation for the apparent high pKa values for these microgels is that the MAA groups are buried within the particles. It is well known [11] that such groups have an increased pKa . This possibility cannot be ruled out based on the current data available. The effect of added NaNO3 on the hydrodynamic diameter for the microgels was investigated in order to obtain additional information about electrostatic interactions and their role in governing the swelling and dispersion stability. Data are shown in Fig. 3 for Microgel 2 obtained using pH 5 and 9. At pH 5 the hydrodynamic diameter decreased slightly from 275 to 240 nm when the NaNO3 concentration was increased from 10−3 to 10−1 M. At higher salt concentrations (0.15 M) the hydrodynamic diameter increased significantly. This is due to aggregation as a consequence of the salt concentration exceeding the critical coagulation concentration (CCC). The CCC is
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
371
(a) Fig. 3. Variation of hydrodynamic diameter for Microgel 2 with concentration of NaNO3 at pH 5 and 9. The lines are guides for the eye.
estimated to be between 0.10 and 0.15 M for Microgel 2 at pH 5. It can be further seen from Fig. 3 that the hydrodynamic diameter decreases considerably (at pH 9) when the salt concentration increases from 0.10 and 0.15 M. This is suggestive of electrostatic screening between ionic groups of the microgel network (e.g., RCOO− ). There was no evidence of flocculation for these samples at salt concentrations less than or equal to 0.5 M. The particles would contain about 97% water at pH 9 in the presence of 0.15–0.5 M NaNO3 . This improved dispersion stability in the presence of high salt concentrations at high pH (cf. pH 5) is attributed to a combination of the low effective Hamaker constant for swollen microgel particles [15] and steric stabilisation due to expanded peripheral chains. 3.2. pH-triggered fluid-to-gel transitions of concentrated microgel dispersions In order to assess the gel-forming properties of the microgel dispersions fluid-to-gel phase diagrams were constructed (Fig. 4). The boundaries that define the transition between the fluid and gelled dispersion state were based on tube inversion measurements. The figure shows that gel formation occurs over a wide range of pH and polymer volume fraction, φp , values for these microgel dispersions. The gels that form are attributed to swelling of the particles to the point where the critical jamming volume fraction (due to the swollen particles) is exceeded [16]. This is taken as the point where the effective volume fraction occupied by the microgel particles, φeff(m) , exceeds the value ∗ . The particles then become sterically confor jamming, φeff(m) fined by their neighbours. Dilution of the gels with water of the same pH caused a gel-to-fluid transition occur. The following trends are evident from the data shown in Fig. 4. (a) The positions of the boundaries for the fluid-to-gel transitions for Microgels 1 and 2 are similar. (b) Gelled microgel dispersions form at pH values greater than 6.4 provided φp 0.10.
(b) Fig. 4. Fluid-to-gel phase diagrams for (a) Microgel 1 and (b) Microgel 2.
(c) Gelled microgel dispersions form at φp values greater than 0.035 provided pH 7.0. In the previous work [4] fluid-to-gel boundaries were established for Microgel 3 at a constant φp of 0.10. In that case the onset for gelation occurred at a pH of 6.2, which is slightly lower than observed for Microgels 1 and 2. Thus, the onset pH values for the fluid-to-gel transitions for all of the microgels are similar when φp = 0.10. This shows that the occurrence of the fluid-to-gel transition is not dependent on the particle size or the maximum q value measured for dilute dispersions (Table 2). The fluid-to-gel transitions that occur for φp 0.10 and pH ∗ 6.2–6.4 must result from φeff(m) exceeding φeff(m) . The value for φeff(m) will increase with q according to the following equation φeff(m) = qφp .
(3)
Fig. 5 shows q values as a function of pH for each of the microgels calculated using Eq. (1) and the data shown in Table 2 and Fig. 2. Although not many data points were measured within the pH 6.2 to 6.4 range it is likely that swelling becomes signif∗ icant in this pH region which results in φeff(m) being exceeded. ∗ Even though we cannot be certain of the exact value for φeff(m) it can be estimated to be in the range of 0.4–0.6 for Microgels 1 and 2 from interpolation of the data shown in Fig. 5 and application of Eq. (3).
372
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
(a)
Fig. 5. Variation of swelling ratio with pH for Microgels 1, 2 and 3. The vertical box shows the onset pH range for gelation when φp = 0.10. The curve at the top shows the ionicity calculated using Eq. (2) and pKa = 6.7.
(b)
(a)
(b) Fig. 6. Variation of (a) elastic modulus and (b) tan δ with oscillation frequency and pH for Microgel 1. The volume fraction was 0.10.
It is our working hypothesis that the response of the gelled microgel dispersions to oscillatory shear will provide a guide to their behaviour when compressed under biomechanical loads within degenerated IVDs. This is based in part on the fact that the uniaxial compression modulus (K) and high frequency shear modulus (G∞ ) for flocculated dispersions normally have comparable values [17]. The relationship between K and the modulus of degenerated IVDs containing microgels remains to
Fig. 7. Variation of (a) elastic modulus and (b) tan δ with oscillation frequency and volume fraction for Microgel 1. The pH of the gels was 7.0. The volume fractions of polymer are shown in the legend.
be determined. It was apparent from the earlier study [4] that the best load support occurred for pH values corresponding to gels with the highest G values. We assume that conditions where G (storage modulus) values for Microgels 1 or 2 are found to match those for Microgel 3 provide support for the respective microgels (1 or 2) being considered as candidates suitable for future load-bearing studies within degenerated IVDs. The variation of G and tan δ (= G /G , where G is the loss modulus) as a function of oscillation frequency (ω) is shown for Microgel 1 in Fig. 6. Comparable data were obtained for Microgel 2 (not shown). It can be seen from Fig. 6a that the gradient of the log G vs log ω plot generally decreases as the pH increases beyond 6.2. In line with other related studies [18] we extracted the frequency exponent (α) using the following equation: G = βωα .
(4)
Equation (4) is an empirical equation which appeared to fit the microgel rheological data reasonably well. In this work gelled dispersions of Microgels 1 and 2 had α values in the range 0.15 to 0.25. Previous work using Microgel 3 gave [4] α values less than 0.10. English et al. [18] found that concentrated solutions of neutralised linear hydrophobically modified poly(MAA-EA) polyelectrolytes exhibited α values in the range 0.2 to 0.4. Car-
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
(a)
373
(a)
(b)
(b)
Fig. 8. Variation of (a) elastic modulus and (b) tan δ with pH for Microgels 1, 2 and 3. The volume fraction was 0.10. The data were obtained using ω = 10 rad/s. The curves are guides for the eye.
Fig. 9. Variation of (a) elastic modulus and (b) tan δ with polymer volume fraction at pH 7 for Microgels 1 and 3. The data were obtained using ω = 10 rad/s.
bopol dispersions have also been reported [19] to give G values which are almost independent of ω. The behaviour observed in Fig. 6 is generally consistent with rheological data published elsewhere for gelled pH-responsive microgel systems. Fig. 7 shows the variation of G and tan δ with oscillation frequency for Microgel 1 measured using different φp values. The α values for these data were also in the range of 0.15–0.25. These values were independent of φp over the range of 0.037– 0.16. This parameter does not appear to be affected by particle concentration for these gelled microgel dispersions. Fig. 8 shows the variation of G and tan δ as a function of pH for the gelled microgel dispersions. Data for Microgels 1 and 2 were obtained from samples used to construct the phase diagrams (Fig. 4). Data for Microgel 3 are extracted from our previous work [4]. Although there is variability in the data it is suggested that all of the samples follow a general trend. It can be seen that for all systems G increases and tan δ decreases sharply as the pH increases from low values to approach 7.0. This is rheological support for pH-triggered fluid-to-gel transitions identified by tube inversion. The data also indicate that a maximum G value and minimum tan δ values occur when the pH is in the vicinity of 7.0. This is a good pH range for biomedical applications. It follows that for these gelled microgel dispersions the greatest elastic modulus values occurs when i is ca. 0.5 (i.e., when pH pKa).
Fig. 9 shows the variation of G and tan δ as a function of φp at pH 7.0 for Microgels 1 and 3. (Unfortunately, data were not obtained for Microgel 2 at pH 7.0.) The data show that G increases exponentially with φp . Given that it is the high φp values (e.g., greater than ca. 0.03) that are expected to be important for potential application in the context of load support these data indicate that the elastic modulus (and presumably the uniaxial compression modulus) of the gelled microgel dispersions can be controlled in a predictable manner using φp . Microgel 3 had sufficiently good mechanical properties to provide structural support for degenerated IVDs under compression using biomechanically meaningful loads [4]. Consideration of the data shown in Figs. 5, 8 and 9 show that pHtriggered swelling and rheological behaviour for Microgels 1, 2 and 3 are comparable under similar conditions. For both Microgels 1 and 2 conditions can be found where G values were similar or greater to values obtained for Microgel 3 at similar φp values. Provided gelled poly(MMA/MAA/EGDMA) microgel dispersions have good biocompatibility they should be good candidates for future testing of their ability to support load within degenerated IVDs. 3.3. Preliminary biocompatibility studies A preliminary investigation was conducted in order to investigate the behaviour of cells from the IVD in the presence
374
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
Fig. 10. Optical micrographs of annulus fibrosus cells in contact with (a) agar, (b) Microgel 1, (c) Microgel 2 and (d) Microgel 3. The micrographs were taken after 24 h of contact.
of gelled microgel dispersions. We selected AF cells because the introduced gel would be in contact with viable AF. Gelled microgel dispersions were prepared by adjusting the pH to 7.0 and AF cells were placed in direct contact with the gel. Because the microgels slowly redispersed with time it was not practical to change the nutrient media. (This tendency for redispersion is being addressed in a separate study.) Therefore, these cell challenge experiments were not able to be continued past five days. The morphology of the AF cells in the presence of the gelled dispersions was investigated as a function of time using optical microscopy. Fig. 10 shows representative micrographs. Compared to the control (agar) the morphology of the AF cells was normal (non-spherical) when in contact with Microgels 1 and 2 after 24 h of contact (Fig. 10). Most of the cells also had normal morphology in the presence of Microgel 3. By comparison when the microgel contained residual SDS the AF cells became spherical and died within the first hour of contact. Other experiments indicated that the AF cells retained a normal morphology for up to five days (not shown). These results provide support for the view that Microgels 1 and 2 have good biocompatibility and long term potential for use in the IVD. Future work will involve live-dead assays using cells from human IVDs. 4. Conclusions This work has investigated the pH-triggered swelling behaviour of MAA containing microgel particles, fluid-to-gel transitions of concentrated dispersions and the rheological
properties of the gelled dispersions. New results have been presented for poly(MMA/MAA/EGDMA) dispersions and these have been compared to those obtained previously for poly(EA/MAA/BDDA) dispersions. There are a lot of similarities between the two types of microgels in terms of their colloidal properties. Both appear to have similar pKa values, which is about 6.7 based on titration and swelling data. The swelling onset for the microgel particles is in the vicinity of 6.0. The critical pH for the fluid-to-gel transitions is about 6.2 to 6.4 for these dispersions when φp is 0.10. The maximum G value and minimum tan δ values for gelled microgel dispersions occurs when the pH is about 7.0 at φp = 0.10. The rheological properties for the gelled poly(MMA/MAA/EGDMA) and poly(EA/MAA/BDDA) dispersions were shown to be comparable. Furthermore, the data suggest that the elastic modulus can be controlled using φp . Because of the strong similarities in the rheological behaviour for gelled poly(EA/MAA/BDDA) and poly(MMA/MAA/EGDMA) dispersions the key parameter controlling the elastic behaviour would seem to be the weight fraction of MAA used. Preliminary cell challenge experiments showed that gelled poly(MMA/MAA/EGDMA) microgel dispersions did not cause pronounced cell death for AF cells over short contact times (days). The physical and cell-challenge data obtained in this work indicate that poly(MMA/MAA/EGDMA) microgels are good candidates for further investigation in the context of injectable load supporting biomaterials for soft tissue. The ability of gelled poly(MMA/MAA/EGDMA) microgel dispersions to restore the mechanical properties of degenerated IVDs will be investigated in future work.
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
Acknowledgments B.R.S. and A.J.F. are grateful to the University of Manchester for funding part of this work. We are also grateful to Professor P. Lovell for access to the PCS instrument. B.R.S. is grateful to Dr. Bob Groves (Synthomer Ltd.) for conducting titrations of Microgel 3. References [1] B.R. Saunders, B. Vincent, Adv. Colloids Interface Sci. 80 (1999) 1. [2] B.R. Saunders, B. Vincent, in: A. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, p. 4544. [3] R.H. Pelton, P. Chibante, Colloids Surf. 120 (1986) 247. [4] J.M. Saunders, T. Tong, C.L. Le Maitre, T.J. Freemont, B.R. Saunders, Soft Matter 3 (2007) 486. [5] B.E. Rodriguez, M.S. Wolfe, M. Fryd, Macromolecules 27 (1994) 6642. [6] K.E. Uhrich, S.M. Cannizzaro, R.S. Langer, M.K. Shakesheff, Chem. Rev. 99 (1999) 3181.
375
[7] H. Hiratani, C. Alvarez-Lorenzo, Biomaterials 25 (2004) 1105. [8] X.S. Zhang, P.A. Revell, S.L. Evans, M.A. Tuke, J.P. Gregson, J. Biomed. Mater. Res. 46 (1999) 279. [9] B.R. Saunders, H.M. Crowther, B. Vincent, Macromolecules 30 (1997) 482. [10] B.H. Tan, K.C. Tam, Y.C. Lam, C.B. Tan, Adv. Colloids Interface Sci. 113 (2005) 111. [11] T. Hoare, R.H. Pelton, Langmuir 22 (2006) 7342. [12] H.G. Spencer, J. Polym. Sci. 56 (1962) S25. [13] J. Brandrup, E.H. Immergut, E.A. Grulke, A. Abe, R.D. Bloch, Polymer Handbook, John Wiley and Sons, 1999. [14] W.C. Chen, Y. Chuang, W.Y. Chiu, J. Appl. Polym. Sci. 79 (2001) 853. [15] B. Vincent, J. Colloids Interface Sci. 42 (1973) 270. [16] V. Prasad, V. Trappe, A.D. Dinsmore, P.N. Segre, L. Cipelletti, D.A. Weitz, Farday Discuss. 123 (2003) 1. [17] R. Buscall, Colloids Surf. 5 (1982) 269. [18] R.J. English, S.R. Raghavan, R.D. Jenkins, S.A. Khan, J. Rheol. 43 (1999) 1175. [19] J.-Y. Kim, J.-Y. Song, E.-J. Lee, S.-K. Park, Colloids Polym. Sci. 281 (2003) 614.
Microgel particles containing methacrylic acid: pH-triggered swelling behaviour and potential for biomaterial application Sarah Lally a , Paul Mackenzie a , Christine L. LeMaitre b , Tony J. Freemont b , Brian R. Saunders a,∗ a Manchester Materials Science Centre, School of Materials, The University of Manchester, Grosvenor Street, Manchester, M1 7HS, UK b Tissue Injury and Repair Group, School of Medicine, Stopford Building, The University of Manchester, Oxford Road, Manchester, M13 9PT, UK
Received 22 June 2007; accepted 13 August 2007 Available online 19 August 2007
Abstract pH-responsive microgels are crosslinked polymer particles that swell when the pH approaches the pKa of the ionic monomer incorporated within the particles. In recent work from our group it was demonstrated that the mechanical properties of degenerated intervertebral discs (IVDs) could be restored to normal values by injection of poly(EA/MAA/BDDA) (ethylacrylate, methacrylic acid and butanediol diacrylate) microgel dispersions [J.M. Saunders, T. Tong, C.L. Le Maitre, T.J. Freemont, B.R. Saunders, Soft Matter 3 (2007) 486]. In this work we report the pH dependent swelling and rheological properties of poly(MMA/MAA/EGDMA) (methylmethacrylate and ethyleneglycol dimethacrylate) microgel dispersions. This system was investigated because it contains monomers that are already used as biomaterials. The poly(MMA/MAA/EGDMA) particles exhibit pH-triggered volume swelling ratios of up to ca. 250. The swelling onset for these particles occurs at pH values greater than ca. 6.0. A pKa for these particles of ca. 6.7 is consistent with titration and swelling data. Fluid-to-gel phase diagrams for concentrated poly(MMA/MAA/EGDMA) dispersions were determined as a function of polymer volume fraction and pH using tube-inversion measurements. The rheological properties for the gelled microgel dispersions were investigated using dynamic rheology measurements. The elastic modulus data for the poly(MMA/MAA/EGDMA) gelled dispersions were compared to data for poly(EA/MAA/BDDA) microgels. A similar pH-dependence for the elastic modulus was apparent. The maximum elastic modulus was achieved at a pH of about 7.0. The elastic modulus is an exponentially increasing function of polymer volume fraction at pH 7.0. Preliminary cell challenge experimental data are reported that indicate that gelled poly(MMA/MAA/EGDMA) microgel dispersions are biocompatible with cells from human intervertebral discs. However, the duration over which these experiments could be performed was limited by gradual redispersion of the gelled microgel dispersions. Based on the results presented it is suggested that poly(MMA/MAA/EGDMA) microgel would be a good candidate as a biomaterial for structural support of soft connective tissues. © 2007 Elsevier Inc. All rights reserved. Keywords: Microgel; Responsive polymer; Gel; Biocolloid; Intervertebral disc
1. Introduction For the present work pH-responsive microgels can be defined as cross-linked polymer particles that swell when the pH approaches the pKa of the incorporated methacrylic acid units. Microgels have been reviewed elsewhere [1,2]. The first waterdispersible responsive microgel was reported by Pelton and Chibante [3] in 1986. In previous work from our group [4] it was established that the height and mechanical properties of * Corresponding author.
E-mail address: [email protected] (B.R. Saunders). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.08.030
degenerated intervertebral discs (IVDs) could be restored by injection of poly(EA/MAA/BDDA) microgel dispersion followed by pH triggered swelling. (EA, MAA and BDDA are ethylacrylate, methacrylic acid and butanediol diacrylate, respectively.) Poly(EA/MAA/BDDA) microgel was originally developed by Rodriguez et al. [5] In this study we replaced EA and BDDA with components which should be more acceptable for potential use in the body. The comonomers chosen were MMA and EGDMA, which are methylmethacrylate and ethyleneglycol dimethacrylate, respectively. The aims of this study were to investigate the pH-dependent swelling behaviour of two poly(MMA/MAA/EGDMA) microgels and compare the data with those for poly(EA/MAA/BDDA) microgel to determine
368
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
whether the former systems would be good candidates for potential use as injectable structural supports in the body. There are several criteria that an injectable biomaterial for load support in the body must meet. It must flow during injection, fill the irregular shape of the internal cavity and then subsequently form an immobilised solid, preferably a gel when required. It should also be biocompatible [6]. In order to provide structural support for biomechanical loads the gel must have a high swelling pressure. We previously investigated poly(EA/MAA/BDDA) microgels that contain high concentrations of MAA groups [4]. The dispersions had relatively low viscosity at low pH. The particles swelled considerably as the pKa of the MAA groups was approached and the dispersion formed a gelled microgel dispersion. The steric confinement of neighbouring swollen particles is responsible for the fluid-togel transition. MMA and EGDMA were selected for use within the pHresponsive microgels in this study because they have been used in bone cement and contact lenses [7,8]. There has been very little work reported in the literature for poly(MMA/MAA/ EGDMA) microgel dispersions. The pH-dependent swelling properties of poly(MMA/MAA/EGDMA) microgel particles containing 17 mol% MAA were reported by Saunders et al. [9] in 1997. They studied the osmotic deswelling of the particles in the presence of non-adsorbing polymer. Their microgel particles, which were prepared using surfactant-free emulsion polymerisation, exhibited much lower swelling ratios than those for the poly(MMA/MAA/EGDMA) particles studied in this work. The properties of concentrated dispersions were not considered in that work. The poly(MMA/MAA/EGDMA) microgels dispersions considered in this work were prepared using seedfeed emulsion polymerisation and have twice the level of MAA present. They exhibit much greater swelling ratios than those prepared in the earlier work. pH-responsive microgel dispersions have been the subject of considerable interest within the microgel literature. A large proportion of the work has focused on microgels containing carboxylic acid groups. Surprisingly, there has been very little work reported on the rheological properties of gelled microgel dispersions. Rodriguez et al. [5] were the first to report the pH-dependent swelling behaviour of poly(EA/MAA/BDDA) microgel dispersions. Their studies were confined to dilute dispersions. They concluded that the particles had a core-shell structure with a lightly crosslinked shell. Tan et al. [10] studied the behaviour of poly(EA/MAA/DAP) microgels (DAP is diallyl phthalate) using static rheology. They reported that the swelling ratio scaled with the MAA content raised to the power of 1.5. Hoare and Pelton [11] reported an insightful study of the effect of carboxylic acid containing monomers on the structure and swelling of poly(NIPAM) (N -isopropylacrylamide) microgels. They identified a link between the apparent pKa of the acid groups and the reactivity ratios of the co-monomers. Reactivity ratios that favoured formation of blocks of acid groups within the microgel network were proposed to cause an increase in the local pKa due to the polyelectrolyte effect. The latter refers to the effect whereby deprotonation of one carboxylate group causes increased electrostatic repulsion between OH− groups
Table 1 Nominal compositions of microgels No.
Composition
MMA or EA (wt%)
MAA (wt%)
EGDMA or BDDA in monomer feeda (wt%)
1 2 3
Poly(MMA/MAA/EGDMA) Poly(MMA/MAA/EGDMA) Poly(EA/MAA/BDDA)
64.0 64.1 66.0
35.7 35.7 33.0
0.4 1.0 1.0
a The seed contained 1.0 wt% EGDMA or BDDA with respect to total monomer mass in each case. The total monomer mass used to during the seed stage was about 8–9 wt% of that used to prepare the particles (see Section 2). Note that EGDMA and BDDA have the same molar mass.
and the charged portion of chain which opposes subsequent deprotonation of the neighbouring carboxylate group. Our previous work using poly(EA/MAA/BDDA) microgel dispersions showed that conditions that gave the maximum elastic modulus, as judged by dynamic rheology, also gave the greatest extent of structural support for degenerated IVDs containing injected microgel. In this work we compare the swelling and rheological data for two new poly(MMA/MAA/EGDMA) microgels to that for poly(EA/MAA/BDDA). Uniaxial compression of IVDs containing poly(MMA/MAA/EGDMA) microgels was not studied in this work. In this work we report colloidal characterisation data in order to improve the structure– property understanding for this class of responsive polymer colloids. By comparison of the rheological and swelling data to the reference system [poly(EA/MAA/BDDA)] we seek to establish design rules for constructing colloidal biomaterials with potential for load-bearing applications within the body in the future. 2. Experimental 2.1. Preparation of microgels Table 1 shows a summary of the compositions for the microgels discussed in this work. All microgels were prepared using a seed-feed method. Two variations of this method were used. For Microgel 1 the feed had a lower weight fraction of EGDMA than the seed. A seed monomer mixture containing MMA (139.3 g), MAA (60.1 g) and EGDMA (2.1 g) was prepared and 25.2 g (8.8% of total monomer) of the mixture added to a nitrogen purged, stirred, solution of SDS (1.75 g in 500 g of water), which had been heated to 80 ◦ C. K2 HPO4 (2.55 g of 7 wt% solution in water) and ammonium persulfate (APS) (4.50 g of 3 wt% solution in water) were immediately added whilst maintaining a nitrogen atmosphere. A feed monomer mixture containing MMA (273.1 g), MAA (134.0 g) and EGDMA (1.7 g) was prepared. After the appearance of a slight blue turbidity, 264 g (91.2% of total monomer) of the feed monomer mixture was added at a constant rate over a 90 min period. Additional APS (2.73 g of 3% solution in water) was added after the feed and the temperature maintained for a further 2 h. The dispersion was extensively dialysed against Milli-Q quality water. Microgel 2 was prepared in the presence of reduced SDS concentration and added NaCl in order to increase the particle
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
size. A monomer mixture containing MMA (125.6 g), MAA (58.7 g) and EGDMA (1.94 g) was prepared. A total of 7.5% of this mixture was used as the seed. This was added to a nitrogen purged, stirred, solution of SDS and NaCl (0.14 g and 0.44 g, respectively, in 200 g of water), which had been heated to 80 ◦ C. K2 HPO4 (2.56 g of 7 wt% solution in water) and APS (2.57 g of 5 wt% solution in water) were immediately added whilst maintaining a nitrogen atmosphere. The remainder of the monomer mixture (92.5% of total) was added to a stirred solution of SDS (0.17 g in 300 g of water) to form an emulsion phase. After one hour of reaction, the emulsified feed monomer mixture, under constant stirring, was added to the reaction at a constant rate over a 105 min period. The reaction temperature was maintained for a further 2 h. The microgel was extensively centrifuged and washed with Milli-Q quality water. Microgel 3 was prepared according to the method reported by Rodriguez et al. [5] and is described fully elsewhere [4]. The method used was similar to that described above for Microgel 1. The weight fraction of BDDA used in the seed and feed were the same (Table 1). It was also purified by dialysis. 2.2. Physical measurements Photon correlation spectroscopy measurements were performed using a BI-9000 Brookhaven light scattering apparatus (Brookhaven Instrument Cooperation), fitted with a 20 mW He–Ne laser and the detector was set at 90◦ scattering angle. The measured data were analysed using cumulant analysis. The swelling ratio (q) was calculated using equation: d , q= (1) dc where d is the hydrodynamic diameter and dc is the average diameter of collapsed particles. In this work we use the numberaverage diameter determined from SEM to estimate dc . The errors (one standard deviation) for the d and q values were estimated as 5 and 15%, respectively. SEM measurements were obtained using a Philips FEGSEM instrument. The samples were prepared using samples at pH values of ca. 4. This pH is well below the estimated pKa for these particles (below). These conditions were used to avoid the possibility of particle deformation of swollen particles during the drying process. The effect of pH on the average diameter of the particles, as measured by SEM, was not investigated in this work. At least 250 particles were counted to obtain number-average diameters. Fluid-to-gel phase diagrams were constructed by mixing known ratios of microgel with NaOH solution and then tested using tube inversion. Gels were considered to be present if the samples did not flow when the tube was inverted. Rheology measurements were performed using a Rheometrics RMS-800. A cone and plate measurement geometry was used. The diameter, cone angle and gap were 25 mm, 0.1 radians and 1.3 mm, respectively. A frequency range of 0.01–15.9 Hz was employed. The strain used for these measurements was 10%. Titration data were obtained from Synthomer Ltd. (Harlow Essex) using automated titration and standardised
369
NaOH solution. The pKa value was taken at the pH at which 50% of the MAA groups were neutralised. 2.3. Cell challenge experiments Agar (1.5 wt%) was dissolved in water with heating and the solution cooled to form a gel. The latter was used as a control. The gelled microgel dispersions were prepared at pH 7 and φp = 0.10. Adherent cultured human annulus fibrosus (AF) cells were used for the cell challenge experiments. They were washed free of media using PBS. The cells were incubated for about 2 min with trypsin which was deactivated with 5 ml of complete media (DMEM). The AF cells were concentrated and then samples of gelled microgel place on top of the cells in the presence of medium. This ensured contact of the gel and cells. The cells were view from beneath the well plate using optical microscopy. 3. Results and discussion 3.1. pH-triggered swelling of microgel particles containing methacrylic acid In this work we synthesised two new poly(MMA/MAA/ EGDMA) microgels (Microgels 1 and 2, see Table 1). One difference between Microgels 1 and 2 is that Microgel 1 was prepared using reduced crosslinking monomer weight fraction (with respect to total monomer mass) during the feed stage. Microgel 2 was prepared using conditions designed to give relatively large particles (cf. Microgels 1 and 3). It can be seen from Table 2 that Microgels 1 and 3 had comparable collapsed diameters. The diameter for Microgel 2 was significantly larger. This was due to the presence of added NaCl and the lower concentration of SDS used during microgel synthesis. These conditions promoted a greater extent of aggregation during the particle formation stage. SEM micrographs for Microgels 1 and 2 are shown in Fig. 1. The SEM data showed that the collapsed particles were spherical with little tendency to deform during the drying process. This means that average diameters calculated from SEM data can be used to estimate the collapsed diameter (dc ). Table 2 shows the average diameters of the particles measured by SEM and those measured by PCS (d) at pH 4. It can be seen that dc is less than or equal to d for each system. This is additional support for a lack of particle deformation to a pancake-like structure as a result of SEM Table 2 Particle size and swelling data for the microgels Microgel No.
dc (nm)a
d (nm)b
qc
1 2 3d
82 ± 9 210 ± 23 65 ± 7
105 ± 5 245 ± 10 75 ± 3
210 ± 32 250 ± 38 105 ± 16
a Number-average particle size from SEM data. The coefficient of variation for the particle size distributions had the same value (11%). b Hydrodynamic diameter measured at pH 4. c The swelling ratio calculated at pH 9 and using Eq. (1). d The data shown for Microgel 3 were obtained from Ref. [4].
370
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
ratios than for Microgel 3. Because of the synthetic method used to prepare these particles (seed-feed emulsion polymerisation) it is reasonable to suggest that Microgels 1 and 2 have core–shell structures. A core-shell structure has been shown to be present for poly(EA/MAA/BDDA) microgel particles [5]. Titration data published by Rodriguez et al. [5] for poly(EA/ MAA/BDDA) prepared in the same manner as Microgel 3 show that at pH 7.0 the degree of neutralisation was 50%. Microgel 3 was titrated for the present study (data not shown) and a pH of 6.7 was evident at 50% neutralisation. Due to the similarity in the pH-dependence of the swelling transitions (Fig. 2) we assume that the pKa for Microgels 1, 2 and 3 are the same, i.e., about 6.7. Fig. 2 also shows the ionicity (i), which is the fraction of MAA units that are deprotonated. These values were calculated using a pKa of 6.7 and the following equation i=
Fig. 1. Scanning electron micrographs of (a) Microgel 1 and (b) Microgel 2. The sample plane was viewed at an angle of ca. 45◦ for the image shown in (b). The horizontal line represents 1 µm and applies to both images.
Fig. 2. Variation of hydrodynamic diameter with pH for Microgels 1, 2 and 3. The ionicity calculated using Eq. (2) and pKa is shown. The error bars for the Microgels 1 and 3 data are smaller than the symbols.
sample preparation because such a process would increase the apparent diameter. Fig. 2 shows the variation of the hydrodynamic diameter with pH for each of the microgels. Swelling ratios for each microgel are shown in Table 2. These values were calculated using the hydrodynamic diameters measured at pH 9. It can be seen from Fig. 2 that each of the microgels exhibits a pH-dependent swelling transition with an onset pH estimated to be about 6.0. The region of maximum diameter increase is between pH values of 6.0 and 8.0. Approximately 50% of the volume increase has occurred at pH 7.5. Little if any further swelling increase occurs after the pH has reached 8.0. It can be seen from the data shown in Table 2 that the swelling ratios for Microgels 1 and 2 are comparable. Interestingly, both of the microgels had significantly greater swelling
1 . 1 + 10pKa -pH
(2)
The use of a pKa of 6.7 for all three microgels appears to be a reasonable approximation. The pKa for the MAA groups within poly(EA/MAA/BDDA), and most likely the poly(MMA/MAA/EGDMA) microgels, is within the range of pKa ’s reported for poly(MAA) by Spencer et al. [12] Hoare and Pelton [11] used co-monomer reactivity ratios to explain the variation of pKa of ionic poly(NIPAM)based microgels. They argued that the high pKa value for poly(NIPAM/MAA) microgels (in the region of 6.2) was due to formation of MAA blocks within the polymer network. The Q and e values for MMA and MAA have been published [13]. Using the Alfrey–Price equation the reactivity ratios were calculated as rMMA = 0.87 and rMAA = 1.09. (These are the same values used by other groups [14].) If these reactivity ratios were present during emulsion polymerisation then the probabilities of adding MMA or MAA to a growing chain end would be similar. Under those circumstances, and considering the approximate two-to-one molar MMA-to-MAA stoichiometry used during microgel preparation, blocks of MAA groups should be present. Accordingly, a pKa similar to the range reported for poly(MAA) is expected for poly(MMA/MAA/EGDMA) microgel particles. An additional explanation for the apparent high pKa values for these microgels is that the MAA groups are buried within the particles. It is well known [11] that such groups have an increased pKa . This possibility cannot be ruled out based on the current data available. The effect of added NaNO3 on the hydrodynamic diameter for the microgels was investigated in order to obtain additional information about electrostatic interactions and their role in governing the swelling and dispersion stability. Data are shown in Fig. 3 for Microgel 2 obtained using pH 5 and 9. At pH 5 the hydrodynamic diameter decreased slightly from 275 to 240 nm when the NaNO3 concentration was increased from 10−3 to 10−1 M. At higher salt concentrations (0.15 M) the hydrodynamic diameter increased significantly. This is due to aggregation as a consequence of the salt concentration exceeding the critical coagulation concentration (CCC). The CCC is
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
371
(a) Fig. 3. Variation of hydrodynamic diameter for Microgel 2 with concentration of NaNO3 at pH 5 and 9. The lines are guides for the eye.
estimated to be between 0.10 and 0.15 M for Microgel 2 at pH 5. It can be further seen from Fig. 3 that the hydrodynamic diameter decreases considerably (at pH 9) when the salt concentration increases from 0.10 and 0.15 M. This is suggestive of electrostatic screening between ionic groups of the microgel network (e.g., RCOO− ). There was no evidence of flocculation for these samples at salt concentrations less than or equal to 0.5 M. The particles would contain about 97% water at pH 9 in the presence of 0.15–0.5 M NaNO3 . This improved dispersion stability in the presence of high salt concentrations at high pH (cf. pH 5) is attributed to a combination of the low effective Hamaker constant for swollen microgel particles [15] and steric stabilisation due to expanded peripheral chains. 3.2. pH-triggered fluid-to-gel transitions of concentrated microgel dispersions In order to assess the gel-forming properties of the microgel dispersions fluid-to-gel phase diagrams were constructed (Fig. 4). The boundaries that define the transition between the fluid and gelled dispersion state were based on tube inversion measurements. The figure shows that gel formation occurs over a wide range of pH and polymer volume fraction, φp , values for these microgel dispersions. The gels that form are attributed to swelling of the particles to the point where the critical jamming volume fraction (due to the swollen particles) is exceeded [16]. This is taken as the point where the effective volume fraction occupied by the microgel particles, φeff(m) , exceeds the value ∗ . The particles then become sterically confor jamming, φeff(m) fined by their neighbours. Dilution of the gels with water of the same pH caused a gel-to-fluid transition occur. The following trends are evident from the data shown in Fig. 4. (a) The positions of the boundaries for the fluid-to-gel transitions for Microgels 1 and 2 are similar. (b) Gelled microgel dispersions form at pH values greater than 6.4 provided φp 0.10.
(b) Fig. 4. Fluid-to-gel phase diagrams for (a) Microgel 1 and (b) Microgel 2.
(c) Gelled microgel dispersions form at φp values greater than 0.035 provided pH 7.0. In the previous work [4] fluid-to-gel boundaries were established for Microgel 3 at a constant φp of 0.10. In that case the onset for gelation occurred at a pH of 6.2, which is slightly lower than observed for Microgels 1 and 2. Thus, the onset pH values for the fluid-to-gel transitions for all of the microgels are similar when φp = 0.10. This shows that the occurrence of the fluid-to-gel transition is not dependent on the particle size or the maximum q value measured for dilute dispersions (Table 2). The fluid-to-gel transitions that occur for φp 0.10 and pH ∗ 6.2–6.4 must result from φeff(m) exceeding φeff(m) . The value for φeff(m) will increase with q according to the following equation φeff(m) = qφp .
(3)
Fig. 5 shows q values as a function of pH for each of the microgels calculated using Eq. (1) and the data shown in Table 2 and Fig. 2. Although not many data points were measured within the pH 6.2 to 6.4 range it is likely that swelling becomes signif∗ icant in this pH region which results in φeff(m) being exceeded. ∗ Even though we cannot be certain of the exact value for φeff(m) it can be estimated to be in the range of 0.4–0.6 for Microgels 1 and 2 from interpolation of the data shown in Fig. 5 and application of Eq. (3).
372
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
(a)
Fig. 5. Variation of swelling ratio with pH for Microgels 1, 2 and 3. The vertical box shows the onset pH range for gelation when φp = 0.10. The curve at the top shows the ionicity calculated using Eq. (2) and pKa = 6.7.
(b)
(a)
(b) Fig. 6. Variation of (a) elastic modulus and (b) tan δ with oscillation frequency and pH for Microgel 1. The volume fraction was 0.10.
It is our working hypothesis that the response of the gelled microgel dispersions to oscillatory shear will provide a guide to their behaviour when compressed under biomechanical loads within degenerated IVDs. This is based in part on the fact that the uniaxial compression modulus (K) and high frequency shear modulus (G∞ ) for flocculated dispersions normally have comparable values [17]. The relationship between K and the modulus of degenerated IVDs containing microgels remains to
Fig. 7. Variation of (a) elastic modulus and (b) tan δ with oscillation frequency and volume fraction for Microgel 1. The pH of the gels was 7.0. The volume fractions of polymer are shown in the legend.
be determined. It was apparent from the earlier study [4] that the best load support occurred for pH values corresponding to gels with the highest G values. We assume that conditions where G (storage modulus) values for Microgels 1 or 2 are found to match those for Microgel 3 provide support for the respective microgels (1 or 2) being considered as candidates suitable for future load-bearing studies within degenerated IVDs. The variation of G and tan δ (= G /G , where G is the loss modulus) as a function of oscillation frequency (ω) is shown for Microgel 1 in Fig. 6. Comparable data were obtained for Microgel 2 (not shown). It can be seen from Fig. 6a that the gradient of the log G vs log ω plot generally decreases as the pH increases beyond 6.2. In line with other related studies [18] we extracted the frequency exponent (α) using the following equation: G = βωα .
(4)
Equation (4) is an empirical equation which appeared to fit the microgel rheological data reasonably well. In this work gelled dispersions of Microgels 1 and 2 had α values in the range 0.15 to 0.25. Previous work using Microgel 3 gave [4] α values less than 0.10. English et al. [18] found that concentrated solutions of neutralised linear hydrophobically modified poly(MAA-EA) polyelectrolytes exhibited α values in the range 0.2 to 0.4. Car-
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
(a)
373
(a)
(b)
(b)
Fig. 8. Variation of (a) elastic modulus and (b) tan δ with pH for Microgels 1, 2 and 3. The volume fraction was 0.10. The data were obtained using ω = 10 rad/s. The curves are guides for the eye.
Fig. 9. Variation of (a) elastic modulus and (b) tan δ with polymer volume fraction at pH 7 for Microgels 1 and 3. The data were obtained using ω = 10 rad/s.
bopol dispersions have also been reported [19] to give G values which are almost independent of ω. The behaviour observed in Fig. 6 is generally consistent with rheological data published elsewhere for gelled pH-responsive microgel systems. Fig. 7 shows the variation of G and tan δ with oscillation frequency for Microgel 1 measured using different φp values. The α values for these data were also in the range of 0.15–0.25. These values were independent of φp over the range of 0.037– 0.16. This parameter does not appear to be affected by particle concentration for these gelled microgel dispersions. Fig. 8 shows the variation of G and tan δ as a function of pH for the gelled microgel dispersions. Data for Microgels 1 and 2 were obtained from samples used to construct the phase diagrams (Fig. 4). Data for Microgel 3 are extracted from our previous work [4]. Although there is variability in the data it is suggested that all of the samples follow a general trend. It can be seen that for all systems G increases and tan δ decreases sharply as the pH increases from low values to approach 7.0. This is rheological support for pH-triggered fluid-to-gel transitions identified by tube inversion. The data also indicate that a maximum G value and minimum tan δ values occur when the pH is in the vicinity of 7.0. This is a good pH range for biomedical applications. It follows that for these gelled microgel dispersions the greatest elastic modulus values occurs when i is ca. 0.5 (i.e., when pH pKa).
Fig. 9 shows the variation of G and tan δ as a function of φp at pH 7.0 for Microgels 1 and 3. (Unfortunately, data were not obtained for Microgel 2 at pH 7.0.) The data show that G increases exponentially with φp . Given that it is the high φp values (e.g., greater than ca. 0.03) that are expected to be important for potential application in the context of load support these data indicate that the elastic modulus (and presumably the uniaxial compression modulus) of the gelled microgel dispersions can be controlled in a predictable manner using φp . Microgel 3 had sufficiently good mechanical properties to provide structural support for degenerated IVDs under compression using biomechanically meaningful loads [4]. Consideration of the data shown in Figs. 5, 8 and 9 show that pHtriggered swelling and rheological behaviour for Microgels 1, 2 and 3 are comparable under similar conditions. For both Microgels 1 and 2 conditions can be found where G values were similar or greater to values obtained for Microgel 3 at similar φp values. Provided gelled poly(MMA/MAA/EGDMA) microgel dispersions have good biocompatibility they should be good candidates for future testing of their ability to support load within degenerated IVDs. 3.3. Preliminary biocompatibility studies A preliminary investigation was conducted in order to investigate the behaviour of cells from the IVD in the presence
374
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
Fig. 10. Optical micrographs of annulus fibrosus cells in contact with (a) agar, (b) Microgel 1, (c) Microgel 2 and (d) Microgel 3. The micrographs were taken after 24 h of contact.
of gelled microgel dispersions. We selected AF cells because the introduced gel would be in contact with viable AF. Gelled microgel dispersions were prepared by adjusting the pH to 7.0 and AF cells were placed in direct contact with the gel. Because the microgels slowly redispersed with time it was not practical to change the nutrient media. (This tendency for redispersion is being addressed in a separate study.) Therefore, these cell challenge experiments were not able to be continued past five days. The morphology of the AF cells in the presence of the gelled dispersions was investigated as a function of time using optical microscopy. Fig. 10 shows representative micrographs. Compared to the control (agar) the morphology of the AF cells was normal (non-spherical) when in contact with Microgels 1 and 2 after 24 h of contact (Fig. 10). Most of the cells also had normal morphology in the presence of Microgel 3. By comparison when the microgel contained residual SDS the AF cells became spherical and died within the first hour of contact. Other experiments indicated that the AF cells retained a normal morphology for up to five days (not shown). These results provide support for the view that Microgels 1 and 2 have good biocompatibility and long term potential for use in the IVD. Future work will involve live-dead assays using cells from human IVDs. 4. Conclusions This work has investigated the pH-triggered swelling behaviour of MAA containing microgel particles, fluid-to-gel transitions of concentrated dispersions and the rheological
properties of the gelled dispersions. New results have been presented for poly(MMA/MAA/EGDMA) dispersions and these have been compared to those obtained previously for poly(EA/MAA/BDDA) dispersions. There are a lot of similarities between the two types of microgels in terms of their colloidal properties. Both appear to have similar pKa values, which is about 6.7 based on titration and swelling data. The swelling onset for the microgel particles is in the vicinity of 6.0. The critical pH for the fluid-to-gel transitions is about 6.2 to 6.4 for these dispersions when φp is 0.10. The maximum G value and minimum tan δ values for gelled microgel dispersions occurs when the pH is about 7.0 at φp = 0.10. The rheological properties for the gelled poly(MMA/MAA/EGDMA) and poly(EA/MAA/BDDA) dispersions were shown to be comparable. Furthermore, the data suggest that the elastic modulus can be controlled using φp . Because of the strong similarities in the rheological behaviour for gelled poly(EA/MAA/BDDA) and poly(MMA/MAA/EGDMA) dispersions the key parameter controlling the elastic behaviour would seem to be the weight fraction of MAA used. Preliminary cell challenge experiments showed that gelled poly(MMA/MAA/EGDMA) microgel dispersions did not cause pronounced cell death for AF cells over short contact times (days). The physical and cell-challenge data obtained in this work indicate that poly(MMA/MAA/EGDMA) microgels are good candidates for further investigation in the context of injectable load supporting biomaterials for soft tissue. The ability of gelled poly(MMA/MAA/EGDMA) microgel dispersions to restore the mechanical properties of degenerated IVDs will be investigated in future work.
S. Lally et al. / Journal of Colloid and Interface Science 316 (2007) 367–375
Acknowledgments B.R.S. and A.J.F. are grateful to the University of Manchester for funding part of this work. We are also grateful to Professor P. Lovell for access to the PCS instrument. B.R.S. is grateful to Dr. Bob Groves (Synthomer Ltd.) for conducting titrations of Microgel 3. References [1] B.R. Saunders, B. Vincent, Adv. Colloids Interface Sci. 80 (1999) 1. [2] B.R. Saunders, B. Vincent, in: A. Hubbard (Ed.), Encyclopedia of Surface and Colloid Science, Dekker, New York, 2002, p. 4544. [3] R.H. Pelton, P. Chibante, Colloids Surf. 120 (1986) 247. [4] J.M. Saunders, T. Tong, C.L. Le Maitre, T.J. Freemont, B.R. Saunders, Soft Matter 3 (2007) 486. [5] B.E. Rodriguez, M.S. Wolfe, M. Fryd, Macromolecules 27 (1994) 6642. [6] K.E. Uhrich, S.M. Cannizzaro, R.S. Langer, M.K. Shakesheff, Chem. Rev. 99 (1999) 3181.
375
[7] H. Hiratani, C. Alvarez-Lorenzo, Biomaterials 25 (2004) 1105. [8] X.S. Zhang, P.A. Revell, S.L. Evans, M.A. Tuke, J.P. Gregson, J. Biomed. Mater. Res. 46 (1999) 279. [9] B.R. Saunders, H.M. Crowther, B. Vincent, Macromolecules 30 (1997) 482. [10] B.H. Tan, K.C. Tam, Y.C. Lam, C.B. Tan, Adv. Colloids Interface Sci. 113 (2005) 111. [11] T. Hoare, R.H. Pelton, Langmuir 22 (2006) 7342. [12] H.G. Spencer, J. Polym. Sci. 56 (1962) S25. [13] J. Brandrup, E.H. Immergut, E.A. Grulke, A. Abe, R.D. Bloch, Polymer Handbook, John Wiley and Sons, 1999. [14] W.C. Chen, Y. Chuang, W.Y. Chiu, J. Appl. Polym. Sci. 79 (2001) 853. [15] B. Vincent, J. Colloids Interface Sci. 42 (1973) 270. [16] V. Prasad, V. Trappe, A.D. Dinsmore, P.N. Segre, L. Cipelletti, D.A. Weitz, Farday Discuss. 123 (2003) 1. [17] R. Buscall, Colloids Surf. 5 (1982) 269. [18] R.J. English, S.R. Raghavan, R.D. Jenkins, S.A. Khan, J. Rheol. 43 (1999) 1175. [19] J.-Y. Kim, J.-Y. Song, E.-J. Lee, S.-K. Park, Colloids Polym. Sci. 281 (2003) 614.