Sm[Fms aFsm Fms 1 akp[o

allows simplification of Eq. (23) to the following:

In principle kppp depends to some extent on monomer conversion (it depends on a = Vm/V, i.e., on the volume of microspheres Vm that might change during polymerization). However, there are several limiting cases when Eq. (25) could be used to obtain interesting and important information.

1. For polymerizations with propagation in microspheres much faster than diffusion of monomer into and outside of particles (1 - a)kppm[I]o >> Sm[Fm, + a(Fsm - Fm,s)] the rate monomer conversion becomes diffusion controlled and independent of kapp,mp and of the concentration of propagating species:

When microspheres occupy a small fraction of the total volume of reaction mixture (a << 1), Eq. (25) could be simplified further to d[M]/dt = -SmFsm[M].

2. For polymerizations with very slow propagation in microspheres (1 - a)

kppm << Sm[Fm,s + a(Fs,m - Fm,s)], kpp! could be expressed as:

In the case when coefficients characterizing flux of monomer into and outside of microspheres are equal (Fsm = Fm,s), the propagation rate constant in dispersion should be equal to the propagation rate constant in microspheres

For polymerizations with very slow propagation in microspheres and with coefficients characterizing flux of monomer into particles that is much higher than the corresponding coefficient for monomer flux from microspheres into solution (so high that relation aFsm >> Fm,s is fulfilled, i.e., particles are highly swollen with monomer) Eq. (27) is reduced to:

In this case, the apparent propagation rate constant strongly depends on the volume fraction occupied by growing microspheres. Generally, a depends on monomer conversion; however, for particles highly swollen with monomer (i.e., when almost all monomer becomes located inside of particles) the volume fraction of microspheres may change very little during polymerization. In a subsequent section some experimental results for dispersion polymerization of e-caprolactone will be discussed and analyzed.

The kinetics of dispersion polymerization of e-caprolactone was investigated for systems with initiators of pseudoanionic [(CH3CH2)2AlOCH2CH3] [53,54,58] and anionic [(CH3)3SiONa] [54,58] polymerization. The initial monomer concentration in these studies was kept in a narrow range (0.41 ± 0.02 mol/L) and initiator concentrations were varied from 1.4 x 10-3 to 5.4 x 10-3 mol/L for (CH3)3-SiONa and from 3.4 x 10-3 to 5.4 x 10-2 mol/L for (CH3CH2)2AlOCH2CH3. Polymerizations were carried out at room temperature in 1,4-dioxane/heptane (1:9 v/v) mixtures in a presence of poly(dodecyl acrylate)-g-poly(e-caprolactone) surface-active copolymer. Concentration of poly(dodecyl acrylate)-g-poly(e-caprolactone) in polymerizing mixture was 2.2 g/L. Poly(dodecyl acrylate)-g-poly(e-caprolactone) with Mn = 28,800 contained 15 wt% of poly(e-caprolac-tone) grafts with Mn = 3600 and Mw/Mn = 1.1. Dodecane (0.6 vol%) used as an internal standard for GPC determination of residual monomer was also present in the mixture. Polymerizations were carried out under argon with stirring at 60 revolutions/min. At chosen times small samples of polymerizing mixture were withdrawn and added to THF containing acetic acid (concentration 4 x 10-4

mol/L) that terminated propagation. Samples with dissolved polymer were frozen in liquid nitrogen and evacuated for 10 min. Thereafter, they were melted and the whole liquid contents distilled to evacuated vials. The amount of unre-acted monomer was determined using GPC (from ratio of signal integrals of e-caprolactone and dodecane using the separately prepared calibration curve). Obtained data were formally analyzed using Eq. (25) in which kpp was considered to be a time-independent parameter.

In Eq. (25) no monomer equilibrium concentration was taken into account. For polymerization of e-caprolactone values of thermodynamic parameters (in THF, AHp = -28.8 kJ/mol, ASp = -53.9 J/molK [58,59]) yield equilibrium monomer concentration equal to 5.9 x 10-3 mol/L. Thus, for initial concentration of e-caprolactone equal to 0.4 mol/L such simplification could be justified for monomer conversion up to about 90%.

Solution of Equation (25) leads to the relation:

FIG. 14 Kinetics of the dispersion polymerization of e-caprolactone. Conditions of the polymerization: [e-caprolactone]o = 4.3 X 10-1 mol/L, [(CH3CH2)2AlOCH2CH3]0 = 1.6 X 10-2 mol/L. (From Ref. 53.)

in which [M]0 denotes the initial monomer concentration and t denotes time. Thus, from plots of ln([M]0/[M]) vs. time, kppp was evaluated dividing slope by [I]0.

A typical kinetic plot is shown in Fig. 14. Experimental data in the kinetic plot could be perfectly fitted with a straight line with the slope 2.02 x 10-3 L/s yielding the corresponding kppp = 1.25 x 10-1 L/(mol-s). The short induction period has been attributed to the stage when microspheres were nucleated.

Rate constants for dispersion polymerizations of e-caprolactone initiated with (CH3CH2)2AlOCH2CH3 and (CH3)3SiONa initiators are given in Table 3. Data in the table indicate that rates of dispersion polymerization depend on the chemical nature of propagating species and on their concentration. Thus, the possibility that propagation rates are controlled by monomer diffusion into growing particles should be ruled out. It has to be noted, however, that kppp depends on concentration of active centers. Plots of kppp as a function of active center concentrations (assumed to be equal to initial concentrations of initiators) are shown in Fig. 15. Dependencies of the apparent propagation rate constants on initiator concentration have been observed earlier for many ionic polymerizations in solution, including solution polymerizations of lactones [12,18,30,38,60-63]. They reflect the presence of equilibria between various physical forms of ionic

TABLE 3 Rates and Apparent Propagation Rate Constants of Dispersion Pseudoanionic [Initiated with (CH3CH2)2AlOCH2CH3] and Anionic [Initiated with (CH3)3SiONa] Polymerizations of e-Caprolactone

[Initiator]0 (mol/L)

[£ -caprolactone] 0 (mol/L)

Rate of polymerization (L/s)

kPPPPP

[1/(mol X s)]

(CH3CH2)2AlOCH2CH3

3.4 X 10-3

3.9 X 10-1

1.63 X

10-3

4.97 X 10-1

4.8 X 10-3

3.9 X 10-1

1.50 X

10-3

3.13 X 10-1

1.0 X 10-2

4.2 X 10-1

1.98 X

10-3

1.98 X 10-1

1.6 X 10-2

4.3 X 10-1

2.02 X

10-3

1.25 X 10-1

2.6 X 10-2

4.1 X 10-1

1.69 X

10-3

6.50 X 10-2

(CH3)2SiONa

1.4 X 10-3

4.0 X 10-1

3.00 X

10-1

2.14

1.4 X 10-3

4.0 X 10-1

4.51 X

10-1

3.22

2.0 X 10-3

4.0 X 10-1

2.56 X

10-1

1.28

2.0 X 10-3

4.0 X 10-1

2.62 X

10-1

1.31

2.4 X 10-3

4.0 X 10-1

2.86 X

10-1

1.19

3.5 X 10-3

4.0 X 10-1

7.20 X

10-1

1.80

5.4 X 10-3

4.0 X 10-1

6.12 X

10-1

1.53

Source: Data from Ref. 55.

Source: Data from Ref. 55.

FIG. 15 Dependence of the apparent propagation rate constants (^5) in pseudoanionic [initiated with (CH3CH2)2AlOCH2CH3] and anionic [initiated with (CH3)3SiONa] dispersion polymerizations of e-caprolactone on initiator concentration. Initial monomer concentration 0.41 ± 0.02 mol/L, reaction medium 1,4-dioxane/heptane 1:9 v/v mixture, room temperature. (Plots based on data from Ref. 55.)

and pseudoionic active species (ions, ion pairs, ionic aggregates, covalent species, and covalent species aggregates) propagating with different rate constants. The positions of these equilibria (molar fractions of various physical forms) strongly depend on the overall concentration of active centers.

It is interesting to compare the apparent propagation rate constants for polymerization of e-caprolactone in dispersed systems (kppd) with the corresponding propagation rate constants (on active centers with the same chemical structure) for polymerizations in solution (kppf). Figure 16 illustrates dependence of the kppp/kppf ratio on concentration of initiator. The plot indicates that for anionic polymerization kppJVkppf is essentially independent from concentration of active centers and that the apparent propagation rate constant for polymerization in dispersion is about 10 times higher than for the corresponding propagation in solution. In the case of pseudoanionic polymerization initiated with (CH3-C^^AlOC^CH kppp/kppp decreases from 21 to 12 when concentration of active centers increases from 3.4 x 10-3 to 2.6 x 10-2 mol/L. Apparently, pseudoanionic

FIG. 16 Dependence of the ratio of apparent propagation rate constants for pseudoan-ionic and anionic polymerizations in dispersion and in solution (k^/k^) on initiator concentration. Conditions of polymerization: initiators (CHjC^^AlOC^CHj, and (CH3)3-SiONa for pseudoanionic and anionic polymerizations, respectively, initial monomer concentration 0.41 ± 0.02 mol/L, reaction media 1,4-dioxane/heptane {1:9 v/v mixture) and THF for dispersion and solution polymerizations, respectively, room temperature. {Plots based on data from Refs. 54 and 55.)

FIG. 16 Dependence of the ratio of apparent propagation rate constants for pseudoan-ionic and anionic polymerizations in dispersion and in solution (k^/k^) on initiator concentration. Conditions of polymerization: initiators (CHjC^^AlOC^CHj, and (CH3)3-SiONa for pseudoanionic and anionic polymerizations, respectively, initial monomer concentration 0.41 ± 0.02 mol/L, reaction media 1,4-dioxane/heptane {1:9 v/v mixture) and THF for dispersion and solution polymerizations, respectively, room temperature. {Plots based on data from Refs. 54 and 55.)

active centers confined in microspheres aggregated efficiently with increased overall (averaged over the while reaction volume) concentration.

One could expect that high values of kapp,dp are due to high local monomer concentration within microsheres. Indeed, in investigated dispersed system microspheres occupied a small fraction of the total volume (a == 0.04). Thus, assuming that monomer diffusion is fast and that Eq. (28) can be used, it was possible to estimate kppm (kppm = a kppp). For example, for pseudoanionic polymerization initiated with (CH3CH2)2AlOCH2CH3 (initial initiator concentration 3.4 x 10-3 mol/L), kppm was estimated to be 1.9 x 10-2 L/(mol-s), whereas the corresponding propagation rate constant in for polymerization solution was equal to 2.3 x 10-2 L/(mol-s). Similar estimation for anionic dispersion polymerization and for the polymerization in solution ([(CH3)3SiONa]0 = 3.5 x 10-3 mol/L) gave kppm = 7.2 x 10-2 L/(mol-s) and kppp = 2.0 x 10-1 L/(mol-s). Thus, the propa gation rate constants inside microspheres are lower that the corresponding rate constants for propagation in solution, probably due to higher local viscosity and more efficient aggregation of active centers due to their confinement in small particles.

Direct evidence that poly(e-caprolactone) microspheres in 1,4-dioxane/hep-tane (1:9 v/v) mixed medium are efficiently swollen with a monomer has been obtained by determination of monomer partition between microspheres and liquid phase [55]. Figure 17 shows dependence of £-caprolactone concentration in liquid phase and in particles plotted as a function of the monomer concentration ([e-caprolactone]av) averaged over the whole volume (volume of liquid phase plus volume of microspheres). The volume fraction of microspheres in this system was 0.037.

Some investigators [12,38,61,64] indicate that in the anionic and pseudoan-ionic polymerizations of e-caprolactone in solution concentration of active centers does not change during the polymerization process (absence of termination). In dispersion polymerizations of this monomer linear dependencies of ln([M]0/ [M]) on time (cf. Fig. 14, slope equal to the product of kppp and concentration

FIG. 17 Partition of e-caprolactone between poly(e-caprolactone) microspheres and 1,4-dioxane/heptane (1:9 v/v) liquid phase. Volume fraction of poly(e-caprolactone) microspheres 2.06%. (From Ref. 55.)

of propagating species) indicate that also in these processes significant termination is absent, at least for monomer conversion up to about 90%. In pseudoanionic polymerization of e-caprolactone in solution initiated with CH3CH2)2AlOCH2-CH3 it was also found that chain transfer reactions (inter- and intramolecular transesterification) are reduced to such an extent that during time needed for monomer conversion up to 95% these side reactions could be practically neglected [30,38,62,65]. On the contrary, in the anionic polymerization of e-capro-lactone in solution with alkali metal counterions, transesterification reactions (including back biting leading to macrocycles) play a significant role and products contain substantial fraction of cyclic oligomers [12,33,34]. For anionic and pseudoanionic dispersion polymerizations of e-caprolactone [initiated with (CH3)3-SiONa and (CH3CH2)2AlOCH2CH3, respectively], good agreement between the measured and calculated (assuming absence of chain transfer and quantitative initiation) values of the average molecular weight for obtained polymers [Mn and Mn(calcd)] has been observed (cf. Fig. 18). The plot also indicates that samples with higher molecular weight that were synthesized with lower initiator concentration had lower molecular weight polydispersity down to Mw/Mn = 1.15. Apparently, at lower initiator concentrations when formation of microspheres is completed at lower monomer conversions (microspheres are formed when Mn is

FIG. 18 Relation between measured Mn, Mw/Mn, and calculated molecular weight [Mn (calcd)] of poly(e-caprolactone) for pseudoanionic and anionic dispersion polymerization. (From Ref. 56.)

between 500 and 1000) the subsequent propagation proceeds without significant contribution of side transesterification reactions.

Absence of transesterification reactions, in particlular intramolecular back biting leading to formation of cyclics, in dispersion polymerization of e-capro-lactone with Na + cations was clearly manifested in GPC trace of synthesized polymers. GPC traces of poly(e-caprolactone) synthesized in anionic solution polymerization in which back biting is important display a very broad signal of linear polymer and signals of cyclic oligomers at long elution times (Fig. 19). On the contrary, GPC trace of poly(e-caprolactone) synthesized in dispersion polymerization with Na + counterion contains only a narrow signal of a linear polymer. Signals of cyclics are absent in these chromatograms (cf. Fig. 20, Ref. 55).

The following reasons might contribute to this difference. It is known that due to thermodymanic relations the fraction of cyclic oligomers is lower for polymerizations with higher initial monomer concentration [12,33,34,37]. Since microspheres are swollen with monomer the initial concentration in particles could be close to that in bulk. At room temperature the total concentration of monomeric units in cyclics is close to 0.25 mol/L. Thus, for polymerization in

Elution volume, ml

FIG. 19 GPC tracing of the product of anionic polymerization of e-caprolactone in THF solution. Polymerization conditions: [e-caprolactone]0 = 6.1 X 10-1 mol/L, [(CH3)3-SiONa]0 = 6.1 X 10-3 mol/L, temperature 20°C.

FIG. 20 GPC tracing of the product of anionic dispersion polymerization of e-caprolac-tone. Polymerization conditions: [e-caprolactone]0 = 4.0 X 10- mol/L, [(CH3)3SiONa]0 = 5.1 X 10-4 mol/L, room temperature. From a calibration curve obtained using poly(e-caprolactones) with narrow molecular weight distribution Mn = 106, 600, Mw/Mn = 1.15. (From Ref. 55.)

FIG. 20 GPC tracing of the product of anionic dispersion polymerization of e-caprolac-tone. Polymerization conditions: [e-caprolactone]0 = 4.0 X 10- mol/L, [(CH3)3SiONa]0 = 5.1 X 10-4 mol/L, room temperature. From a calibration curve obtained using poly(e-caprolactones) with narrow molecular weight distribution Mn = 106, 600, Mw/Mn = 1.15. (From Ref. 55.)

bulk (initial monomer concentration equal to 9.0 mol/L) the fraction of cyclics would amount to only as high as 2.8%. It is also possible that relatively low mobility of polymer chains inside of microspheres (in comparison with chain mobility in solution) decreases rates of transesterification reactions. Thus, in anionic dispersion polymerization of e-caprolactone, both thermodynamic and kinetic factors could contribute to practical elimination of cyclics up to the moment of full monomer conversion.

IX. PHASE TRANSFER OF POLY(e-CAPROLACTONE) AND POLY(LACTIDE) MICROSPHERES FROM HYDROCARBONS TO THE WATER-BASED MEDIA

Poly(lactide) and poly(e-caprolactone) microspheres were synthesized in 1,4-dioxane/heptane mixed media. After transfer to pure heptane these particles form stable suspensions (if they settle down they can be resuspended by gentle shaking). However, for any biological or medical applications such particles must be transferred to water or water-based media. It was reported (cf. Ref. 65) that such transfer was accomplished in the following way. A sample of suspension of poly(e-caprolactone) and/or poly(lactide) microspheres (volume 1-1.7 mL) was added to the ethanol solution of a given surfactant [Triton X-405, ammonium sulfobetaine-2 (ASB), and/or sodium dodecyl sulfate (SDS); chemical structures of these surfactants are given in Fig. 21]. Total volume of obtained suspension was 7.5 mL. Concentrations of surfactants down to about 0.5% (wt/ vol) could be used.

Thereafter, to the suspension 0.6 mL of 1 M KOH was added. Poly(e-capro-lactone) microspheres were incubated in this medium for 10 min, poly(lactide) particles for 1 h. Thereafter, the microspheres were isolated by centrifugation and redispersed in ethanol solution of any of the above-mentioned surfactants (1% wt/vol). This procedure was repeated at least once more and then particles isolated by centrifugation were transferred to 1 mL of the aqueous surfactant solution. Eventually, pH of suspending medium was adjusted to 11 by addition of the appropriate amount of KOH solution. Suspensions of microspheres were

FIG. 21 Chemical structures of Triton X-405, sodium dodecyl sulfate (SDS), and ammonium sulfobetaine-2 (ASB) surfactants.

FIG. 21 Chemical structures of Triton X-405, sodium dodecyl sulfate (SDS), and ammonium sulfobetaine-2 (ASB) surfactants.

kept at 4°C. The lowest concentrations of surfactants in water providing sufficient stabilization of suspensions of microspheres were 1.5 x 10-1, 1.7 x 10-1, and 4.7 x 10-1% (wt/vol) for ABS, Triton X-405, and SDS, respectively. Stability of suspensions was monitored by quasi-elastic light scattering. When microspheres begun to coalesce the averaged diameter of light scattering objects in suspension increased [65].

Particles in the above-described buffered water suspensions were stabilized electrostatically and sterically. Electrostatic repulsion between particles resulted from presence of carboxylic anions in their surface layers. These groups were formed by basic hydrolysis of poly(e-caprolactone) induced by KOH. Steric stabilization was due to adsorption of surfactants (ABS, Triton X-405, and/or SDS). It has to be noted that alone neither electrostatic stabilization (due to KOH treatment) nor the steric one (using the above-listed surfactants) were sufficient for stabilization of poly(lactide) and/or poly(e-lactide) microspheres.

Hydrolysis of polyester molecules in surface layer of poly(e-caprolactone) and/or poly(lactide) microspheres obviously should lead to changes in the overall molecular weight of polymer in particles. One should also take into account changes in microsphere diameters resulting from hydrolysis. In Table 4 are given values of averaged diameters (Dn) and diameter polydispersity factors (Dw/Dn) for particles that were stored in heptane and after their transfer to the buffer. The molecular weights of polymers before and after transfer to water media are given in Table 5.

Diameters of particles from suspension in heptane and from the buffer were determined from SEM images. In addition, diameters of microspheres suspended in the buffer were determined by quasi-elastic light scattering (this method was not suitable for measurements in heptane since particles from suspensions in hydrocarbons were adsorbed onto the cell's surface).

TABLE 4 Diameters and Polydispersity Indexes of Poly(e-caprolactone) and Poly(l,l-lactide) Microspheres Before and After Transfer from Heptane to Water/ABS System

Before transfer to water/ASB

After transfer to water/ASB

Measurements

Measurements Measurements by SEM

by SEM by QELS

Microspheres

Poly(e-caprolactone) Poly(l,l-lactide)

TABLE 5 Molecular Weight and Molecular Weight Distribution of Poly(e-caprolactone) and Poly(l,l-lactide) Microspheres Before and After Their Transfer from Heptane to the Water/ASB System

Before transfer After transfer from heptane to water/ASB

Microspheres Mn Mw/Mn Mn Mw/Mn

Poly(e-caprolactone) 42,600 1.52 8,350 1.57

Poly(l,l-lactide) 38,400 1.12 19,300 1.53

Data in Table 4 revealed that the procedure used for transferring micro-spheres from heptane to buffers did not change the diameters of microspheres. However, hydrolysis constituting an important step in transfer of particles to water substantially decreased the molecular weight of microspheres and, in the case of particles composed of polymer with low Mw/Mn, also significantly increased the molecular weight polydispersity (cf. Table 5).

Hydrolysis led to formation of carboxylic groups in microsphere surface layer. In media with sufficiently basic pH these groups would be ionized providing particles that are negatively charged. Ionic surfactants (SDS and ASB) also modify the charge at surfaces of microspheres. Surface charge densities (combining charge of polyester carboxyl end groups, ionic groups of adsorbed surfactants, and counterions embedded into loose surface layer) for both kinds of microspheres and for all kinds of surfactants were determined on the basis of the measurements of microsphere electrophoretic mobility [65]. The following relation holds for surface charge density of microspheres (a) and their electrophoretic mobility (pe)

In Eq. (30), n denotes viscosity of the medium and k the reciprocal of the Debye length given by the following expression:

Was this article helpful?

0 0

Post a comment