Siloxane-based copolymers for use in two-phase partitioning bioreactors.

Amsden, B.G. ; Lau, A.

INTRODUCTION

Two-phase partitioning bioreactors (TPPBs) have been demonstrated to be an effective technology for the bioremediation of toxic and/or poorly soluble organic compounds (Daugulis, 2001). TPPBs consist of a cell-containing aqueous phase, and a second immiscible phase that contains toxic and/or hydrophobic substrates that partition to the cells at sub-inhibitory levels in response to the metabolic demand of the organisms. The delivery phase in TPPBs has traditionally been a hydrophobic solvent, which needed to possess a variety of important properties including biocompatibility, nonbioavailability to the organism used to degrade the xenobiotic, low volatility, and low cost. Relatively few liquids are available that meet all these criteria. Recently, it has been demonstrated that the organic solvent phase can be replaced by a solid polymer phase that functions in a similar fashion as organic solvents (Amsden et al., 2003; Daugulis et al., 2003; Prpich and Daugulis, 2004, 2005, 2006). This approach has several potential advantages. The solid polymers are inaccessible as substrates for the organisms, they can be used with consortia and thus the bioreactor can be much more effective and efficient, they are easier to handle, complete polymer recovery from the bioreactor is possible through conventional separation techniques such as filtration, and there is no potential for adsorption to or absorption of the polymer phase into reactor materials such as rubber gaskets, tubing, and seals.

A major requirement of the polymer used in this technology is the ability to rapidly absorb a large quantity of xenobiotic. The absorption rate is controlled by the both diffusivity of the solute within the polymer and the affinity of the solute for the monomer(s) comprising the polymer. A number of commodity polymers have been screened for use in the removal of various xenobiotics (Table 1) (Amsden et al., 2003; Prpich and Daugulis, 2004). However, relying on commodity polymers, while useful, may not result in the most effective polymer being used. It was the objective of this work to prepare a polymer that would be more effective than the polymers examined to date. To accomplish this, the following criteria were considered. To achieve a large solute-polymer diffusion coefficient, the polymer should be amorphous and have a very low glass transition temperature (Crank and Park, 1968). The glass transition temperature of a polymer is determined to a great extent by the flexibility of the backbone. Siloxanes have very flexible backbones and correspondingly low glass transition temperatures (Brandrup and Immergut, 1999; Mark, 1999). They are also generally amorphous. Moreover, a wide variety of siloxane monomers are available. The most commonly used siloxane is poly(dimethylsiloxane). Although the solubility of low molecular weight compounds, such as oxygen and hydrogen, through poly(dimethylsiloxane) are generally high, it does not absorb most organic compounds to an appreciable extent (Brandrup and Immergut, 1999). However, substitution of the methyl groups with phenyl groups, as in poly(dimethyldiphenylsiloxane) (PDMDPS), confers some polarizability and results in an increase in the absorbance of more polar compounds (Kiridena et al., 2001). Also, siloxanes are generally bioinert (Lukasiak et al., 2004). With this in mind, divinyl terminated PDMDPS was chosen as the basis for the development of the desired polymer. In an attempt to improve on the absorption capacity of the polymer, various water-soluble monomers were copolymerized with PDMDPS. It was reasoned that the inclusion of water-soluble monomers would improve the rate and amount of absorption of more polar compounds by increasing water absorption of the polymer in the aqueous medium, thereby lowering its glass transition temperature as well as increasing the overall surface area for absorption. As a model xenobiotic, phenol was used. Phenol was chosen because it has been investigated previously in demonstrations of the proof of concept of the polymer-based TPPB approach (Amsden et al., 2003; Prpich and Daugulis, 2004), and because it is a priority waste water pollutant (Kujawski et al., 2004; Veeresh et al., 2005).

MATERIALS AND METHODS

Poly(dimethyldiphenylsiloxane) divinyl terminated (9300 g/mol ([M.sub.n]); 84:16 mole ratio of dimethylsiloxane/diphenylsiloxane) (PDMDPS), 1,3-dimethyltetravinyldisiloxane (DMTVDS), acrylic acid (AA), ethylene glycol dimethacrylate (EGDMA), Nvinylpyrrolidone (NVP), 1,1'-azobis(cyclohexanecarbonitrile) (ACBN), phenol, toluene, and acetonitrile were all obtained from Aldrich, Canada.

Polymer Preparation and Characterization

Poly(dimethylsiloxane-co-diphenylsiloxane) divinyl terminated and DMTVDS were mixed in varying ratios by weight (Figure 1). The monomers, AA, EGDMA, or NVP were added to this mixture in various amounts as a weight percentage of the PDMDPS-DMTVDS mixture. Polymer mixtures were crosslinked for 18 h at 110.C in sealed Pyrex glass tubes (2.5 mm internal diameter) using 1.5 wt% ACBN initiator dissolved in toluene (1 mg/15 [micro]L). The resulting polymer rods had a diameter of 2.3 [+ or -] 0.1 mm, and were cut with a razor blade to a length of 2.5 [+ or -] 0.1 cm.

[FIGURE 1 OMITTED]

The glass transition temperatures of the polymers were recorded using a Seiko 5200 differential scanning calorimeter. Samples of the polymers (10 mg) were preheated to 80[degrees]C and cooled to -140[degrees]C before being heated to 250[degrees]C under a nitrogen gas flow. The temperature was held at each stage for 10 min, and heating/cooling rates of 10[degrees]C/min were used. All measurements were recorded from the second thermal scan. The water-soluble component (sol) and swelling capacity of the polymer rods were determined by immersing them in Type 1 deionized, distilled water. The water was removed and replaced five times over 24 h. The polymer samples were then dried in vacuo in the presence of desiccant for 48 h. The organic sol content of the polymers was measured by immersing them in toluene, which was replaced five times over 24 h. The samples were then dried in vacuo for 48 h. The sol content is expressed as:

sol = [m.sub.i] - [m.sub.f]/[m.sub.i] (1)

wherein [m.sub.i] is the initial and [m.sub.f] the final dry mass of the polymer, respectively.

Absorption and Release Studies

Prior to the absorption studies, the polymer rods were soaked in Type I water overnight. To characterize phenol absorption, the rodswere patted dry and immersed in 4.2 mL of a 2 mg/mL phenol solution in 1 dram glass vials. The vials were placed on a rotary mixer rotating at 15 rpm at room temperature (25[degrees]C). Samples (200[micro]L) were taken from the vials at frequent time points. Release studies were performed immediately after an absorption study. Polymer rods loaded with phenol were patted dry and immersed in 4.2 mL of Type I water. After 30 min, 1, 2, 4, and 24 h, the phenol concentration in the release media was sampled and replaced entirely with Type I water. Solution samples were analyzed for phenol on an analytical HPLC unit (Waters), using a Symmetry [C.sub.18] column (4.6 x 150 mm 5 [micro]m particle diameter). Samples (200 [micro]L) were diluted to a total volume of 1 mL before analysis to generate sufficient volume for analysis. The flow rate was 1 mL/min and 100 [micro]L of sample was injected. Peaks were detected by a UV detector at [lambda] = 270 nm. The mobile phase consisted of water (A) and acetonitrile (B). The following gradient was used: initially 80% A and 20% B, 55% A and 45% B for 0-11.25 min, 20% A and 80% B for 11.25-14.25 min, 80% A and 20% B for 14.25-19.5 min, 80% A and 20% B for 19.5-25 min. The retention time of phenol was approximately 8.8 min.

The diffusivity of phenol in the polymers, D, was calculated from a least-squares fit of the following equation, valid for solute uptake by a cylinder immersed in a solution of finite volume, V, to the absorption data (Crank, 1975):

[M.sub.t]/[M.sub.[infinity] = 1 - [[infinity].summation over (n=1)] 4[alpha] (1 + [alpha)/4 + 4 [alpha] + [[alpha].sup.2] [q.sup.2.sub.n] exp (-D[q.sup.2.sub.n] t/[r.sup.2] (2)

wherein [M.sub.t] is the mass absorbed in time t, [M.sup.infinity] the total mass absorbed at infinite time, r the radius of the cylinder, [q.sub.n] s the positive, non-zero roots of:

[alpha] [q.sub.n] [J.sub.o] ([q.sub.n]) = 0 (3) and [alpha] is given by:

[M.sub.infinity]/[VC.sub.0] = 1/1 + [alpha] (4)

In Equation (3), [J.sub.0] and [J.sub.1] are Bessel functions of the first kind of order zero and one, respectively. [C.sub.0] the initial solute concentration in the aqueous medium. The fitting was performed using the first 30 terms in the summation of Equation (2).

In the release experiments, the concentration of phenol in the aqueous phase was kept at conditions approximating an infinite sink. Under these conditions, the release rate of phenol from the rods is given by (Crank, 1975):

[M.sub.t]/[M.sub.[infinity] = 1 - [[infinity].summation over (n=1)] 4/[r.sup.2] [[alpha].sup.2.sub.n] exp (-D [[alpha].sup.2.sub.n] t/[r.sup.2] (5)

wherein [[alpha.sub.n] s are the positive roots of [J.sub.0] (r[[alpha].sub.n]) = 0.

The polymer/water partition coefficient, [K.sub.p/w], was calculated as a ratio of the total mass phenol absorbed/mass polymer to the initial total mass of phenol/mass water.

All experiments were performed in triplicate and the data were presented as the average value obtained, while the error bars represent one standard deviation about this average.

RESULTS AND DISCUSSION

Initial polymerizations were performed with the PDMDPS and DMTVDS in various ratios (10, 15, 20, 25, 30 wt% DMTVDS) alone. Useable polymers were only obtained with DMTVDS ratios of 25 and 30 wt%; with 10 and 15% DMTVDS, polymerization was unsuccessful, and with 20% DMTVDS, the polymer was gummy. The glass transition temperatures of the 25% and 30% DMTVDS containing polymers were essentially the same at -102[degrees]C. Therefore, in subsequent work, all polymers were prepared with PDMDPS/DMTVDS weight ratios of 70:30. The polymers will be referred to using the abbreviation for the macromer, with the weight percent and type of comonomer added following, if used.

The measured polymer properties are listed in Table 2. The polymers obtained were all non-porous and rubbery at room temperature and the glass transition temperatures were all very low, as would be anticipated for polysiloxanes, ranging from -104[degrees]C for PDMDPS, to an essentially constant value of about -98[degrees]C for the polymers containing hydrophilic comonomers. The 20% AA containing copolymer exhibited a secondary glass transition temperatures at -61[degrees]C. The secondary glass transition temperature indicated the presence of partially miscible poly(AA)rich phases in this particular copolymer, as the glass transition temperature of poly(AA) is approximately 110[degrees]C (Mark, 1999).

The use of DMTVDS monomer to prepare the polymers resulted in all of the polymers possessing significant gel fractions, the portion of which varied with the nature of the comonomer. Without added hydrophilic comonomer, the organic sol fraction of PDMDPS was about 30 wt%. Of the polymers prepared with hydrophilic comonomers, only those containing NVP demonstrated sol contents that were similar to those of the PDMDPS; the sol contents of polymers containing EGDMA and AA were about 45 and 57%, respectively. Conversely, very little of the material was extractable into water; the extractable portion was consistently below 2.5%. Thus, most of the water-soluble monomers were incorporated into the copolymer.

Without the addition of a hydrophilic comonomer, the polymers did not swell appreciably in water. As the amount of hydrophilic comonomer added was increased, the swelling increased. The degree of swelling in water, however, was dependent on the comonomer used. PDMDPS containing EGDMA did not swell markedly, even with 40 wt% EGDMA incorporated into the polymer, reaching only 3.2 [+ or -] 0.3%. PDMDPS containing 20% AA swelled to 10.9 [+ or -] 0.3%, while PDMDPS containing 20% NVP swelled to only 2.9 [+ or -] 0.3%. Increasing the NVP content to 30% was necessary to achieve equivalent swelling extents as with 20% AA. EGDMA swelled to a lesser extent because it is difunctional, and thus produced tightly crosslinked hydrophilic regions within the polymer formed. As the crosslinking density of a polymer increases, the ability of the polymer to swell decreases. For the PDMDPS polymers prepared with mono-functional monomers, those containing AA swelled to a greater extent than those with NVP due to the dissociation of AA monomers to generate negative charges along the polymer backbone. The negatively charged monomers exert additional electrostatic repulsion forces that act to increase the swelling of the polymer. Thus, a number of different siloxane polymers can be prepared having varying degrees of water absorption capacity.

[FIGURE 2 OMITTED]

The phenol absorption capability of the prepared polymers was determined by monitoring the uptake of phenol from deionized distilled water. The results can be seen in Figure 2. The fastest rate of uptake of phenol was obtained with the PDMDPS polymer, while the rate of uptake of phenol decreased as water-soluble comonomers were incorporated into the polymer. This absorption rate was reflected in the values of the diffusion coefficients calculated from a least-squares fit of Equation (2) to this data (Table 3). The diffusion coefficient was highest in PDMDPS at 5.83([10.sup.-7]) [cm.sup.2]/s and lowest in PDMDPS containing 40% EGDMA at 0.44(10-7) [cm.sup.2]/s. PDMDPS containing 20% hydrophilic comonomers had comparable diffusion coefficients of between 1.5 and 2([10.sup.-7]) [cm.sup.2]/s. The diffusion coefficient of phenol in PDMDPS represents almost a fourfold increase over that of the highest diffusion coefficient found previously in Hytrel. Diffusivity in a polymer is determined primarily by the flexibility of the polymer chains, reflected in the glass transition temperature. PDMDPS possessed the lowest glass transition temperature and the highest diffusivity for phenol. Incorporation of more polar comonomers decreased the glass transition temperature by only 6-7[degrees]C, and yet the diffusivity decreased by up to 10 times. The larger decrease in diffusivity than expected from the change in glass transition temperature is likely due to the enhanced hydrogen bonding between the hydroxyl group of Phenol and the polar groups of the hydrophilic comonomers.

[FIGURE 3 OMITTED]

The capacity of the polymers for phenol is reflected by their [K.sub.p/w], given in Table 3. Although it possessed the greatest phenol diffusion coefficient, PDMDPS exhibited the lowest average [K.sub.p/w]. Incorporation of the more polar monomers into the polymer improved the phenol absorption capacity, with EGDMA conferring the greatest increase in [K.sub.p/w], reaching 9.4 [+ or -] 1.6 g phenol absorbed/g phenol in solution, while NVP provided a [K.sub.p/w] of 5.1 [+ or -] 0.6. The greater absorption capacity achieved with these two comonomers can be attributed to the enhanced hydrogen bonding between phenol and the comonomers. Incorporation of AA into the polymer did not provide significant improvement to phenol absorption capacity, perhaps due to the fact that phenol is a weak acid, and thus would experience some electrostatic repulsion from dissociated AA in the polymer.

Using the copolymer possessing the greatest absorption capacity for phenol (20% EGDMA), the rates of phenol absorption and desorption were compared. Moreover, the rate of desorption was calculated using Equation (5) and compared to the experimental data. The results are given in Figure 3. Desorption into an approximate infinite sink occurs more slowly than absorption from a concentrated phenol solution because the concentration gradient driving force is lower. However, the desorption rate is closely predicted using the diffusivity obtained from a curve fit of the absorption data and the appropriate diffusion equation, indicating that the phenol is not bound within the polymer. A mass balance on the phenol absorbed versus desorbed indicated that essentially all the phenol was released during the desorption experiment, within the limits of experimental error. This result indicates that the siloxane-based copolymer can be re-used, as has been shown previously for other polymers such as ELVAX (Amsden et al., 2003) and Hytrel (Prpich and Daugulis, 2004).

CONCLUSIONS

Including a water-soluble comonomer into a siloxane-based polymer does confer water-swellability to the polymer, and ultimately leads to an increase in the amount of absorbable phenol. However, the increase in amount absorbed is achieved at the cost of a reduction in phenol diffusivity within the copolymer. Comparison of the absorption capability of the siloxane-based copolymers to the commodity polymers previously examined, shows that PDMDPS containing 20% EGDMA is equivalent to Hytrel 8206. Nevertheless, despite not being able to absorb as much phenol/g polymer, PDMDPS without a hydrophilic comonomer can absorb phenol at a greater rate than Hytrel 8206. Thus, some improvement can be obtained in the rate of phenol removal by absorption from aqueous waste streams using PDMDPS.

ACKNOWLEDGEMENTS

Funding for this study was provided by a grant from the Natural Sciences and Engineering Research Council of Canada.

Manuscript received April 17, 2007; revised manuscript received June 19, 2007; accepted for publication June 21, 2007.

REFERENCES

Amsden, B. G., J. Bochanysz and A. J. Daugulis, "Degradation of Xenobiotics in a Partitioning Bioreactor in which the Partitioning Phase is a Polymer," Biotechnol. Bioeng. 84, 399-405 (2003).

Brandrup, J. and E. H. Immergut, "Polymer Handbook," Wiley, New York (1999).

Crank, J., "The Mathematics of Diffusion," Oxford University Press, New York (1975).

Crank, J. and G. S. Park, "Diffusion in Polymers," Academic Press, New York (1968).

Daugulis, A. J., "Two-Phase Partitioning Bioreactors: A New Technology Platform for Destroying Xenobiotics," Trends Biotechnol. 19, 457 (2001).

Daugulis, A. J., B. G. Amsden, J. Bochanysz and A. Kayssi, "Delivery of Benzene to Alcaligenes Xylosoxidans by Solid Polymers in a Two-Phase Partitioning Bioreactor," Biotechnol. Lett. 25, 1203-1207 (2003).

Kiridena, W., W. Koziol, C. Poole and M. Nawas, "Influence of Diphenylsiloxane Composition on the Selectivity of Poly(dimethyldiphenylsiloxane) Stationary Phases for Open-Tubular Column Gas Chromatography," Chromatographia 54, 749-756 (2001).

Kujawski, W., A. Warszawski, W. Ratajczak, T. Porebski, W. Capala and I. Ostrowska, "Removal of Phenol from Waste-water by Different Separation Techniques," Desalination 163, 287 (2004).

Lukasiak, J., A. Dorosz, M. Prokopowicz, P. Rosciszewski and B. Falkiewicz, "Biodegradation of Silicones," in "Miscellaneous Bioppolymers, Biodegradation of Synthetic Polymers," S. Matsumura and A. Steinbuchel, Eds., Wiley-VCH (2004), p. 539.

Mark, J. E., "Polymer Data Handbook," Oxford University Press, Toronto, ON, Canada (1999).

Prpich, G. P. and A. J. Daugulis, "Polymer Development for Enhanced Delivery of Phenol in a Solid-Liquid Two-Phase Partitioning Bioreactor," Biotechnol. Prog. 20, 1725 (2004).

Prpich, G. P. and A. J. Daugulis, "Enhanced Biodegradation of Phenol by a Microbial Consortium in a Solid-Liquid Two Phase Partitioning Bioreactor," Biodegradation 16, 329 (2005).

Prpich, G. P. and A. J. Daugulis, "Biodegradation of a Phenolic Mixture in a Solid-Liquid Two-Phase Partitioning Bioreactor," Appl. Microbiol. Biotechnol. 72, 607 (2006).

Veeresh, G. S., P. Kumar and I. Mehrotra, "Treatment of Phenol and Cresols in Upflow Anaerobic Sludge Blanket (UASB) Process: A Review," Water Res. 39, 154 (2005).

B. G. Amsden * and A. Lau

Department of Chemical Engineering, Queen's University, Kingston, ON, Canada K7L 3N6

* Author to whom correspondence may be addressed. E-mail address: brian.amsden@chee.queensu.ca

DOI 10.1002/cjce.20003

Table 1. Phenol absorption capacity, polymer/water partition
coefficient ([K.sub.p/w]), and diffusion coefficient in previously
investigated polymers

Polymer Type Absorption capacity [K.sub.p/w]
 (mg/g polymer)

Hytrel (a) 8206 19 9.5
ELVAX (b) 40% Vinyl acetate 12.4 6.1
 23% Vinyl acetate 7.7
Nylon 6 9.9
 4/6 4.2

Polymer Diffusivity Reference
 ([cm.sup.2]/s)

Hytrel (a) 1.54 x [10.sup.-7] Prpich and Daugulis (2004)
ELVAX (b) 3.53 x [10.sup.-9] Amsden et al. (2003)

Nylon Prpich and Daugulis (2004)

[K.sub.p/w] is reported at an initial aqueous phase concentration of
2000 mg/L

(a) Poly(butylene terephthalate)-block-poly(butylene ether glycol
terephthalate)

(b) Poly(ethylene-co-vinyl acetate)

Table 2. Glass transition temperature ([T.sub.g]), organic sol content,
and degree of swelling in water of siloxane polymers

Polymer [T.sub.g] Sol Swelling
 ([degrees]C) (a) (wt%) (b) (wt%) (c)

PDMDPS -104 31.2 2.5
PDMDPS + 20% EGDMA -98 47.7 1.9
PDMDPS + 40% EGDMA -98 42.3 3.2
PDMDPS + 10% AA -98 56.6 2.3
PDMDPS + 20% AA -98, 61 57.1 10.9
PDMDPS + 20% NVP -99 34.5 2.9
PDMDPS + 30% NVP -97 34.6 10.7

(a) [+ or -] 1[degrees]C

(b) [+ or -] 2%

(c) [+ or -] 0.3%

Table 3. Phenol polymer/water partition coefficient and the diffusion
coefficient of phenol within the polymers

Polymer [K.sub.p/w] D (x [10.sup.7] SSR (b)
 ([+ or -] SD (a)) [cm.sup.2]/s)

PDMDPS 3.2 (0.8) 5.83 0.015
PDMDPS + 20% EGDMA 9.4 (1.6) 1.60 0.001
PDMDPS + 20% AA 3.7 (0.7) 1.51 0.007
PDMDPS + 20% NVP 5.1 (0.6) 2.05 0.010
PDMDPS + 40% EGDMA 7.5 (0.3) 0.44 0.011

(a) SD, standard deviation

(b) SSR, sum of squares of the residuals obtained from the curve fit
used to obtain D