J. CHEM. SOC. FARADAY TRANS., 1994, 90(13), 1961-1966 Micellisation and Gelation of Triblock Copolymer of Ethylene Oxide and r-Caprolactone, CLnEmCLn in Aqueous Solution Luigi Martini, David Attwood,* John H. Collett, Christian V. Nicholas, Siriporn Tanodekaew, Nan-Jie Deng, Frank Heatley and Colin Booth Manchester Polymer Centre, Departments of Chemistry and Pharmacy, University of Manchester, Manchester, UK M 13 9PL Three triblock copolymers of ethylene oxide and E-caprolactone, nominally CL,E,,CL, , CL4EgOCL4 and CL,E,,CL, have been prepared and characterised. The micellar and surface properties in aqueous solution of the copolymers with CL-block lengths of four and six were investigated as a function of temperature and concen- tration using surface tension and static and dynamic light scattering techniques.Reversible gelation on cooling solutions of CL&g&L, was observed at critical concentrations and temperatures ranging from 130 g kg-' at 25°C to 300 g kg-' at 80°C. The micellisation and gelation properties of triblock co-poly(oxyethylene/oxypropylene/oxyethylene)s, E,P,E, (where E represents an oxyethylene unit and P an oxypropylene unit) have been extensively studied in recent years.'-9 In an extension of this the oxypropylene unit was replaced with the more hydrophobic oxybutylene unit, B, to form a series of E,B,E, copolymers, the use of which avoids many of the problems related to sample purity and repro- ducibility which are associated with the anionic polymeris- ation of propylene oxide.6 Members of both series of triblock copolymers form micelles in dilute aqueous solution and thermally reversible gels at higher concentrations. Gelation results essentially from the packing of micelles acting as hard spheres and can be effected either by heating from a low tem- perature or cooling from a high temperature, so-called 'hot' and 'cold' ge1ati0n.I~ Gelation on heating occurs whilst micelles are being formed and is a consequence of the close packing of the micelles at sufficiently high concentration to prevent micellar translational motion.The influential factor governing gelation on cooling is the significant increase in hydration of the oxyethylene units of the micellar fringe as the temperature is decreased.', COpOlymerS such as E106P6,E106 (F127), which form gels at temperatures between ambient and body temperature, have been widely studied for their potential in the formula- tion of implants for the controlled delivery of drugs (see, e.g.ref. 14). These systems offer the possibility of implant forma- tion in situ by the subcutaneous injection of a mobile solu- tion, thus avoiding the necessity for surgical implantation. However, their poor biodegradation characteristics might necessitate surgical removal of the implant after drug release. In an attempt to address this problem, block copolymers have been synthesised incorporating poly(ecapro1actone) which is known to be subject to degradation in viuo by hydrolytic chain scission involving the ester linkages.' 5*16 We report here an investigation of the solution properties of a series of CL,E,CL, copolymers (where CL represents E-caprolactone).Although a limited number of compounds of this type have been synthesised previ~usly,~~-'~ their associ- ation properties in aqueous solution have not been reported. In view of the role of micelles in the gelation of block copoly- mers, light scattering methods have been used to examine micellisation in dilute aqueous solution and changes in micel- lar properties on increase of temperature and concentration. Experimental Materials To remove any hydroxy acid formed by ring opening, E-caprolactone (Aldrich 99%) was stirred over 2,4-diisocyanate- 1-methylbenzene (Fluka) for at least 24 h, then fractionally distilled under reduced pressure (bp 96-98 "C, 5 mmHg) directly onto fresh CaH, .Polyethylene glycol 4000 (Fluka) was evacuated at 80 "C (1 mmHg) for 75 h to remove residual traces of water. Cop1ymerisation The block copolymers were prepared by using polyethylene glycol 4000 [or-hydro, o-hydroxypoly(oxyethylene), E,,] to initiate the polymerisation of ecaprolactone at 180 "C in the absence of added catalyst; the method used was that of Cerrai et a1.I9 The E-caprolactone (CL), was inserted by syringe into an ampoule containing a pre-weighed quantity of dried E90 under dry N, and sealed after evacuation. Poly- merisation was carried out at 180°C for 30 h, after which the product was recovered from the ampoule using dichloro- methane which was subsequently removed by evacuation.Purification was effected by extracting repeatedly with hexane to remove any poly(CL) initiated by traces of mois- ture. The product was subsequently subjected to prolonged evacuation to remove any remaining volatile substances. The overall compositions of the copolymers were deter- mined by 'H and 13C NMR (Bruker Spectrospin AC-300E, 75 or 300 MHz, CDCl, solvent) and their molar masses by gel permeation chromatography (GPC) [tetrahydrofuran solvent ; columns calibrated with poly(oxyethy1ene) standards]. The three samples are denoted CL,E,,CL,, where n = 2,4, or 6. These formulae are based on overall composition from NMR combined with an E-block length of 90 units.Details of the molecular characterisation of the samples by GPC and NMR are given below. GPC curves obtained for the three copolymers contained narrow peaks, corresponding to M,, in the range 5000-6000 g mol-', and A?w!l@n in the range 1.10-1.15 (see Table 1). These values of MW//l@"are similar to that found for the precursor PEG4000, showing that no degradation (chain scission) occurred under the conditions used for copolymer- isation. The GPC curve of sample CL,E,,CL, also con-tained a small peak (< 5% by area) centred on Mpk= 450 g mol-',which was assigned to homopoly(CL) impurity, initi- ated by moisture. No homopolymer was detected in the other two samples. Assignments for relevant 'H and 13C resonances are listed in Table 2.'H NMR spectra were used to measure the overall composition (see Table 1) by comparison of the inte- grated resonances of the methylene protons of the E block with those of the CL block. A similar analysis of the inte- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 1 Molecular characteristics of the copolymers mol% E "MR) copolymer MwIMn (GW 'H 3c wt.% E (NMR)(average) mol% E* (I3CNMR) Mn (13CNMR) An (formula) CL2E90CL2 1.12 95.8 96.3 90.5 25 4420 4430 CL4E90CL4 1.15 90.7 91.5 79.8 6 4870 4690 CL6E90CL6 1.15 87.5 87.2 72.8 3 5330 5420 grated resonances of the methylene carbons in the I3C NMR num plate suspended from a calibrated torsion balance. The spectra confirmed these compositions (see Table 1).usual precautions were taken to ensure cleanliness and the The I3C spectra of the three copolymers contained clear water was triply distilled from alkaline permanganate solu- evidence of oxyethylene hydroxy ends (E*, resonances a* and tion. The accuracy of measurement was checked by frequent b* of Table 2) as well as carboxypentamethylenehydroxy measurement of the surface tension of water (71.0 nN m-'; ends (CL*, resonances 5* and 6* of Table 2). Comparison of expected at 30 "C,71.2 mN m- I). Solution temperatures were integrals gave the mole percentages of E* ends listed in Table maintained at 30 k 1"C by means of a thermostatted water 1. If all ends (E* and CL*) have equal reactivity towards jacket. Surface tension readings were measured at intervals addition of CL units, then the probability of finding an E* until consistent values were obtained.end after reaction with 2n CL units should be simply 0.5": i.e. mol% = 25, 6 and 1.6 for n = 2, 4 and 6, respectively. Given Light Scattering the experimental uncertainty of determination of the mol% of E* ends (+2), the results obtained for the copolymer samples Static light scattering was measured at temperatures in the are as expected. range 30-50 "C (_+1 "C) by means of a Malvern PCSlOO Integrals of the resonances of end-group carbons and back- instrument with vertically polarised incident light of wave- bone carbons were used to obtain the overall number- length 488 nm supplied by a 2 W argon-ion laser (Coherent average molar masses listed in Table 1.These values of &%, Innova-90). Solutions were clarified by repeated ultrafiltra- are in good agreement with the values expected from the con- tion through 0.1 pm filters, the final filtration being directly ditions of preparation, as indicated by their nominal formu- into the cleaned scattering cell. The intensity scale was cali- lae. brated against filtered benzene. Refractive index increments The mol% of triblock copolymers in the samples, as calcu- were measured over the temperature range 30-50 "C lated from mol% of E* ends, were ca. 94, 88 and 50% for ( & 0.5 "C) using an Abbe 60/ED precision refractometer CL,E,,CL,, CL,E,,CL, and CL,E,,CL, , respectively. In (Bellingham and Stanley Ltd). view of the low purity of CL,E,,CL,, this copolymer was Dynamic light scattering measurements were made by not studied further.means of the Malvern instrument described above combined with a Malvern K7027 auto-correlator using 60 linearly spaced channels with a far-point delay of 1024 sample times. Cloud Point Measurement of scattered light was at an angle of 90" to the Aqueous solutions of the copolymers (2 wt.%) in small tubes incident beam. The data were analysed either by the were slowly heated (0.5 K min-') in a water bath to 100°C CONTIN method to obtain information on the distribution and clouding was observed visually. The cloud points quoted of decay times or, if appropriate, by a single-exponential fit of below were the mean of two determinations. the correlation curve.Surface Tension Gelation Characteristics Surface tensions (7) of dilute aqueous solutions were mea- A solution of copolymer of known concentration was sured by the Wilhelmy plate method using a roughened plati- enclosed in a small tube and observed over the temperature Table 2 Assignments for 'Hand NMR spectroscopy: E,CL, and CL,E,CL, a* 61.5 b* -72.4 C 3.6 70.4 d -69.0 e 4.1 63.2 2 2.2 33.9 3 1.5 25.2 4 1.2 24.4 5 1.5 28.2 6 3.9 63.9 5* -32.1 6* 3.6 62.2 HOCH,CH2[OCH,CH21,-,0CH2CH20[C0CH,CH2CH2CH2CH20]~-ICOCH2CH,CH2CH,CH,0H(diblock) a* b* c c de 23456 2 3 4 5* 6* . . .OCH ,CH 2[OCH ,CH 23, -,OCH ,CH,O [COCH ,CH ,CH ,CH ,CH,O], -COCH ,CH ,CH ,CH ,CH ,OH (triblock) ed cc de 23456 2 3 4 5* 6* J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 range 5-50°C. The change from sol to gel (or uice versa), determined by inverting the tube, defined the gel temperature to within f1 K. Results and Discussion Preliminary measurements of the in vitro degradation of CL~E~OCL~in aqueous solution at 37°C have shown a decrease of molar mass (from GPC) of ca. 20% over a period of 16 weeks. The extent of the reduction of molar mass over the timescale of the measurements of this study (not exceed- ing 2 days from the preparation of a polymer solution) was less than 0.4% and it was considered that any influence of degradation on the results was negligible. Cloud Points The cloud points of 2 wt.% aqueous solutions of the copoly- mers were 65 and 55 "C for CL,E,,CL,, and CL&,oCL6.Surface Tension Plots of surface tension, y, of solutions at 30°C against log c (where c is concentration in g dmP3) are shown in Fig. 1. Equilibrium readings were obtained within 8 h for all solu- tions. The plot obtained for CL&,oCL6 shows a single clear transition point at c = 0.010 g dm-3, at which the surface tension attained a constant value (yc.m,c.= 50.6 mN m-'), assumed to be the critical micelle concentration (c.m.c.). The corresponding plot for CL,E,,CL, is less easily interpreted. Behaviour of this type has been observed previo~sly~,~~ and attributed to a range of compositions within the sample. In the present case the composition range will be wide, including as it does diblock copolymers as well as a range of CL chain lengths in the triblock copolymers (n = 1-10, average n = 4).The surface tension became roughly constant (y = 50.9 mN m-') at a critical concentration of c = 0.050 g drnp3. The unfiltered aqueous solutions of this copolymer were clear, suggesting that the homopoly(CL) impurity, thought to be present in CL,E,,CL,, was solubilised by the copolymer micelles. As a consequence, the onset of micelle formation will occur at a lower concentration and the critical concen- tration quoted above should be regarded as a minimum value. The slopes of the y us. log c plots for CL,E,,CL, and CL~E~OCL~well below the critical concentrations were used 60 0 0 c I2 55 + 0 E--. 02-50 -3 -2 -1 0 1 log(c/g d~n-~) Fig. 1 Surface tension us.log c for aqueous solutions of block copolymers at 30 "C : (@) CL,E,,CL,; (+) CL,E,,CL, 1963 to calculate approximate values of excess surface concentra- tion, r, by use of the simple form of the Gibbs absorption isotherm r = -(l/RT) (dy/d In C) (1) and hence the area per molecule in the full monolayer, a, from a = i/m, (2) where N, is Avogadro's constant. The values of Q were 1.5 and 1.4 nm2 for CL,E9,CL, and CL&&L6, respectively, which are similar to those of other triblock copolymers with long oxyethylene chains. For example, copolymers E,,P,,E,, and E,,B, ,E5, in aqueous solution have values of 1.4 and 1.3 nm2, respectively, at 30 0C.10,21 Dynamic Light Scattering Dilute solutions (2-4 g drn-,) of each of the copolymers were examined at 30°C by dynamic light scattering.The results were analysed by the constrained regularisation CONTIN technique developed by Provencher22 to provide information on the nature of any association. The size distribution of any aggregates present in solution was determined from the dis- tribution of decay rates and hence of apparent diffusion coef- ficients, Dapp,by application of the Stokes-Einstein equation rh, app = kB T/(6nqDapp) (3) where rh,app is the radius of the hydrodynamically equivalent hard sphere corresponding to Dapp,k, is the Boltzmann con- stant and q is the viscosity of water at temperature T. The results from the CONTIN analysis are presented in Fig. 2 in the form of (a) intensity and (b) weight distributions of log(?-,,app).The distributions for both CL,E,,CL, and CL,E,,CL6 were narrow, indicative of a closed association. There was evidence in the intensity distribution of CL,E,'CL, of an extremely small proportion of material in the form of large particles; the presence of these particles could not be detected when the dynamic scattering data were expressed in the form of a weight distribution of log rh,app). Their influence was considered to be negligible. Values of Dapp for CL,E,,CL6 at concentrations below 20 g dm- and over the temperature range 30-50 "C were deter- mined from single exponential fits of the correlation curves. Extrapolation of the data to infinite dilution yielded the Do values listed in Table 3.The radii of the hydrated micelles, rh (rh = [( l/rh)=)]-'), as calculated from the Stokes-Einstein equation, were constant between 30 and 50 "C (see Table 3). Static Light Scattering Plots of the scattering function K'c/(S -S,) against c for solutions of CL,E,,CL, at 30, 40 and 50°C are shown in Fig. 3; S is the intensity of light scattered from a solution at 90" relative to the scattering from benzene and S, is the corre- sponding value for water. The constant K' was calculated using the values of refractive index increments, dn/dc, in Table 3. The temperature derivative of dn/dc (-2 x lov4 cm3 8-l K-I) of this copolymer was identical to that obtained previou~ly'~.~~ for aqueous solutions of di- and tri- block copolymers of oxyethylene and oxybutylene.The dis- symmetries, Z, of the scattering envelopes, as determined by the ratio of intensities at scattering angles of 45" and 135", were close to unity (2< 1.10 & 0.05) indicative of small par- ticles. The pronounced curvature of the plots of Fig. 3 is a conse- quence of interparticle interference which was accounted for, as in previous work,"-' by treating the micelles as 2923*24 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I I I I I I B 0 1 2 3 lodr,, app/nm) Fig. 2 Dynamic light scattering from aqueous solutions of block copolymers at 30 "C. Normalised, A, intensity distributions and, B, weight distributions of the logarithm of apparent hydrodynamic radius obtained for (a)CL,E,,CL,, c = 4 g dm-3; (b) CL,E,,CL4, c = 5 g dm-j.hard spheres and using the scattering equation in the form K'c/(S -S,) 1: l/QA?, (4) where is the mass-average molar mass of the micelles and Q is the interparticle scattering (the intraparticle scattering factor being unity for small spheres). For moderate concen- trations of uniform hard spheres, Vrij25 has suggested that Q can be calculated from the Carnahan-Starling equation26 1/Q = "1 fw2-V(44 -+2)1(1-(5)+)r4 where 4 is the volume fraction of equivalent thermodynamic uniform spheres calculated from the actual volume fraction of copolymer in micelles in the system by applying a volume swelling factor (6, = swollen volume relative to dry volume). The adjustable parameters are A?, and 6,.In applying eqn. (5)concentrations were converted to a volume basis assuming a density of dry polymer, p, of 1.10 g dm-3, irrespective of temperature. Fig. 3 shows approximate fits to the experimen- I I I I I D I 1 I I I 0 20 40 60 80 100 c/g dm-3 Fig. 3 Static light scattering function K'c/(S -S,) us. c for aqueous solutions of CL,E,,CL, at (e)30, (A) 40 and (7)50°C and (+) CL,E,,CL, at 30°C. The curves were calculated using the Carnahan-Starling equation as described in the text. tal data obtained with 6, values of 2.8, 2.5 and 2.3 at 30, 40 and 50 "C, respectively. Data points for the lowest concentra- tions were possibly affected by micellar dissociation and were given less weight than those for higher concentrations (c 3 10 g dmr3).Thermodynamic hard-sphere radii, rt , were calcu- lated from 6, and A? using I, = (36, MW/47CN,p)f'3 (6) and are listed in Table 3 together with the association numbers of the micelles calculated from N = NJmicelle)/A? w( molecule) (7) The results in Table 3 show an increase of the molar masses and thermodynamic radii (rt) of micelles of CL~E,OCL~with temperature rise but a constancy of their hydrodynamic dimensions as measured by rh . Similar results have been reported for a number of oxyethylene/ oxypropylene and oxyethylene/oxybutylene triblock copoly- mers.1-5,10,12,27 The effect is due to the dehydration of the micelles with increase of temperature which counteracts the increase in anhydrous size due to micellar growth in such a way that the hydrodynamic radius does not show any signifi- cant change with temperature.Table 3 shows the decrease with temperature of the hydrodynamic swelling factor, 6, [calculated from the ratio of the volumes of swollen (vh = 47cnr,3/3) to anhydrous (V, = A?,/N, p) spheres] which pro- vides evidence for micellar dehydration. Previous workers have shown similar decreases of micellar hydration by this means'-4 and also from changes in the intrinsic viscosity2' Table 3 Micellar properties" of aqueous solution of block copolymers CL,E,CL, T dnldc ', DO 'h 'h. apg NWcopolymer /"c /cm3 g-' 10-5Mw ,hm cm2 s-l /nm /nm 6, 30 0.144 -----8.6 -CL4E90CL4 CL6E90CL6 30 0.143 2.07 38 6.0 2.9 9.8 9.9 13 40 0.138 3.13 58 6.7 3.4 10.4 -10 -750 0.134 4.16 77 7.1 4.3 10.1 ~~ -~-a dnldc = refractive index increment (k0.002 cm3 g-'); M, = molar mass (k10% g mol- I); N, = association number (k15%); rt = thermodynamic radius (+ 1 nm); Do = limiting diffusion coefficient (k0.5x lo-' cm2 s-'); rh = hydrodynamic radius from Do (k1 nm); 6, = hydrodynamic swelling factor ( f2).Intensity-average value estimated by the CONTIN method from results on dilute solutions (4-5 gdmP3). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 and in the fluorescence quantum yields of polarity-sensitive probes.28 Light scattering measurements on aqueous solutions of CL,E,,CL, were restricted to 30 "C and to concentrations of less than 20 g dm-3. In view of the scattered data in a limited concentration range and the evidence of dissociation at the lowest concentrations (Fig.3), it was not meaningful to evalu- ate 6, and M,,,by application of the Carnahan-Starling equa- tion. Gelation Characteristics The sol-gel diagram for solutions of CL6EgoCL, is illus- trated in Fig. 4. The dashed line through the data points defines approximately the sol-gel transition on cooling solu- tions of given concentration, or alternatively the gel -+ sol transition on heating. No lower gel -+sol transition was observed on further cooling the solutions of concentrations represented by the data points of Fig. 4to 0°C. Previous work".' 2,24,29*30 has shown that the gelation of oxyethylene/oxypropylene and oxyethylene/oxybutylene copolymers can be broadly explained by assuming that the critical gel concentration (c.g.c) is equivalent to the critical concentration for close packing of the micelles acting effec- tively as hard spheres, i.e.C.g.C. = (2"2/8ri,)( 1024M/NA)= 0.292M/& (8) In eqn. (8), M is the molar micellar mass in g mol- ',rhs is the equivalent hard sphere radius in nm, and c.g.c. is in g dmP3. Application of the equation to the gelation of CL,E,,CL6, assuming that the values of A?, of Table 3 can be directly related to hf,yields rhs values in reasonable agreement with the r, values from static light scattering. For example, the c.g.c. for CL~E~OCL~ at 30°C is 165 g dm-3 which leads to an equivalent hard sphere radius of rhs = 7.2 nm compared with an r, value from light scattering of r, = 6.0 nm.Gelation of solutions of CL&oCL6 by cooling from a high temperature, commonly known as hot gelation,' will be accompanied by a negative standard enthalpy of gelation. If, as is thought, gelation is a consequence of the aggregation of spherical micelles, then it is possible to determine the ther- modynamic quantities for the gelation process. The standard states for gelation are micelles in solution at unit molarity and the micelles in their gel state. The equations by which the I I I I I / / /.80 / / 9 60 / /L-/ / /sol / / / // e'/ 20 / 1 I I I I 100 150 200 250 300 clg kg-' Fig. 4 Gel-sol diagram for aqueous solutions of CL,E,,CL, thermodynamic quantities can be obtained are AgelGe = R (c.g.t.)ln(c/N) (9) and AgelH* = R[d In c/d (c.g.t.)-'] (10) where c.g.t.is the critical gel temperature of a solution of concentration c [expressed in (mol chain) dmW3], and N is the aggregation number. The units of the thermodynamic quantities are J (mol micel1es)- ' or J (mol chains)-' on divid- ing by N. Gibbs energies of gelation from eqn. (9) for CL,EgoCL6 at 30, 40 and 50°C were -17.7, -19.1 and -20.4 kJ (mol mice1les)-'. AgelHo,as calculated from the gradient of the linear plot of In c against (c.g.t.)-' was -17.3 kJ (mol mice1les)-'. Values of similar magnitude were reported for hot gelation of oxyethylene/oxybutylene triblock copoly-mer~.~~,~'Conversion of these values to units of J (mol chain)-' using the aggregation numbers of Table 3 yields very low values, i.e.the gelation of a micellar solution of a triblock copolymer of caprolactone and oxyethylene, as with that of micellar solutions of oxyethylene/oxypropylene31.32 and oxyethylene/~xybutylene~~copolymers, is almost an athermal process. This work was generously supported by the Wellcome Trust and SERC. Research studentships were provided by the Science and Engineering Research Council (L.M.), the Insti- tute for the Promotion of Teaching Science and Technology, Thailand (S.T.) and the Colloid Fund of the University of Manchester (N.J.D.).Professor Provencher kindly provided copies of the CONTIN programme for analysis of the dynamic light scattering data.Dr. R. H. Mobbs and Mr. K. Nixon gave valuable assistance with the preparation and characterization of the copolymers. References 1 Z-K. Zhou and B. Chu, Macromolecules, 1988,21, 2548. 2 Z-K. Zhou and B. Chu, J. Colloid Znterface Sci., 1988,126, 171. 3 G. Wanka, H. Hoffman and W. Ulbricht, Colloid Polym. Sci., 1990,268,101. 4 W. Brown, K. Schillen, M. Almgren, S. Hvidt and P. Bahadur, J. Phys. Chem., 1991,95, 1850. 5 N. K. Reddy, P. J. Fordham, D. Attwood and C. Booth, J. Chem. Soc., Faraday Trans., 1990,86, 1569. 6 L. Yang, A. D. Beddells, D. Attwood and C. Booth, J. Chem. Soc., Faraday Trans., 1992,88, 1447. 7 G-E. Yu, Y-L. Deng, S. Dalton, Q-G. Wang, D. Attwood, C. Price and C. Booth, J. Chem. Soc., Faraday Trans., 1992, 88, 2537.8 W. Brown, K. Schillen and S. Hvidt, J. Phys. Chem., 1992, 95, 6038. 9 P. Linse and M. Malmsten, Macromolecules, 1992,25, 5434. 10 Y-Z. Luo, C. V. Nicholas, D. Attwood, J. H. Collett, C. Price and C. Booth, Colloid Polym. Sci., 1992,270, 1094. 11 C. V. Nicholas, Y-Z. Luo, N-J. Deng, D. Attwood, J. H. Collett, C. Price and C. Booth, Polymer, 1993,34, 138. 12 Y-Z. Luo, C. V. Nicholas, D. Attwood, J. H. Collett, C. Price, C. Booth, Z-K. Zhou and B. Chu, J. Chem. Soc., Faraday Trans., 1993,89, 539. 13 I. R. Schmolka, J. Biomed. Mater. Rex, 1972,6, 571. 14 C. J. Tait, J. B. Houston, D. Attwood and J. H. Collett, J. Pharm. Pharmacol., 1987,39,57. 15 C. G. Pitt, A. R. Jeffcoat, R. A. Zweidinger and A. Schindler, J.Biomed. Mater. Res., 1979, 13, 497. 16 Y. H. Bae and S. W. Kim, Adv. Drug Delivery Rev., 1993,11,109. 17 R. Perret and A. Skoulios, Makromol. Chem., 1972,156, 143. 18 R. Perret and A. Skoulios, C. R. Acad. Sci. (Paris), 1969, 268, 230. 19 P. Cerrai, M. Tricoli, F. Andruzzi, M. Paci and M. Paci, Polymer, 1989,30,338. 1966 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 20 21 Z. Yang, S. Pickard, N-J. Deng, R. J. Barlow, D. Attwood and C. Booth, Macromolecules, submitted. K. N. Prasad, T. T. Luong, A. T. Florence, J. Paris, K. Vaution, M. Seiller and F. Puisieux, J. ColIoid Interface Sci., 1979, 69, 28 29 J. C. Gilbert, C. Washington, M. C. Davies and J. Hadgraft, Int. J. Pharm., 1987,40,93. P. Bahadur, P. Li, M. Almgren and W. Brown, Langmuir, 1992, 8, 1903. 22 23 24 25 26 27 225. S. W. Provencher, Makromol. Chem., 1979,180,201. A. D. Bedells, R. M. Arafeh, Z. Yang, D. Attwood, F. Heatley, J. C. Padget, C. Price and C. Booth, J. Chem. Soc., Faraday Trans., 1993,89, 1235. S. Tanodekaew, N-J. Deng, S. Smith, Y-W. Yang, D. Attwood and C. Booth, J. Phys. Chem., 1993,97,11847. A. Vrij, J. Chem. Phys., 1978,69, 1742. N. F. Carahan and K. E. Starling, J. Chem. Phys., 1969,51,635. D. Attwood, J. H. Collett and C. J. Tait, Int. J. Pharm., 1985, 26, 25. 30 31 32 A. D. Bedells, R. M. Arafeh, Z. Yang, D. Attwood, J. C. Padget, C. Price and C. Booth, J. Chem. Soc., Faraday Trans., 1993, 89, 1243. G-E. Yu, Y-L. Deng, S. Dalton, Q-G. Wang, D. Attwood, C. Price and C. Booth, J. Chem. SOC., Faraday Trans., 1992, 88, 2537. Y-L. Deng, G-E. Yu, C. Price and C. Booth, J. Chem. Soc., Faraday Trans., 1992,88, 1441. Paper 4/0813H ;Received 9th February, 1994