J. BIOMED. MATER. BES.
VOL. 10, PP. 33-45 (1976)
Surface Hydroxylation of Styrene-Butadiene-Styrene Block Copolymers for Biomaterials MICHAEL V. SEFTON* and EDWARD W. MERRILL, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 ,
Summary This work pertains to the development of high strength elastomers potentially useful as nonthrombogenic cardiovascular prostheses. Triblock copolymers of the styrene-butadiene-styrene type have been subjected to surface hydroxylation which provide reactive sites a t the surface for the subsequent coupling of heparin while retaining the unique mechanical properties of the SBS copolymers. Curves of hydroxyl content versus the copolymer film thickness demonstrate the effect of swelling in the surface region on the product distribution and on the time dependence of the hydroxylation process. In addition, the effect of time, temperature, and the composition of the reaction bath on the diffusion/reaction process is shown. Finally, the general applicability of this surface modification scheme to the development of biomaterials is discussed.
INTRODUCTION Over the past several years many implantable devices have been devised which aid the circulation of blood in diseased patients. One of the simplest of these is the vascular prosthesis, designed to act as a replacement or bypass for arteries or veins which have become occluded. While Dacron velour is most commonly used for these prostheses, the use of a velour limits the use of the prosthesis to those portions of the circulatory system (principally the larger arteries) where the blood flow rate is sufficiently fast to prevent buildup of thrombin to levels adequate to activate fibrin formation. The de-
* Present address: Department of Chemical Iihgineering, University of Toronto, Toronto, Ontario b15S 1A4, Canada. 33 @ 1976 by John Wiley & Sons, Inc.
34
SEFTON AND MERRILL
velopment of a nonthrombogenic material suitable for these prostheses would significantly extend the applicability of vascular replacement surgery. I n order to prepare a suitable high strength elastomeric substrate potentially useful as nonthrombogenic cardiovascular prostheses, certain triblock copolymers of the styrene-butadiene-styrene (SBS) type have been subjected to surface hydroxylation. As a consequence of this treatment, a n elastomcric material modeling the laminate structure of the arterial wall has been developed, namely, a soft hydrophilic layer bound to a tough elastomeric substrate. The hydrophilic layer may be subsequently rendered nonthrombogenic by fixation of heparin therein.
MATERIALS AND METHODS
SBS as a Biomaterial The unique properties of the styrene-butadiene-styrene block copolymers makes them particularly useful as biomaterials. First, because of the requirements of the “living-polymer” anionic polymerization technique by which they are synthesized, they are ultrapure and contain no elutable contaminants which might render them toxic t o the human system. The work of Nyilas e t al.’ demonstrated the strong effect that trace contaminants and defects in the microstructure have on the blood compatibility of silicones. Second, as a result of the thermodynamic incompatibility of the polystyrene and polybutadiene blocks, there is a phase separation in the solid phase, with the formation, in a SBS containing 25% polystyrene, of glassy domains of polystyrene in a continuum of the rubbery polybutadiene (Fig. 1). However, because of the bonds between the two phases, the polystyrene domains act to keep the polybutadiene entanglements in place, in addition to acting as particles of reinforcing filler to give the copolymer, a t body temperature, the properties of a high strength elastomer (initial modulus, 650 psi; tensile strength, 3900 psi) .z Thus, without any further potentially contaminating crosslinking or strength inducing steps, the SBS copolymer containing approximately 257, wt polystyrene satisfies all the mechanical requirements of a vascular prosthesis. I n addition, because the dimensional stability of the SBS copolymer is the result of a polystyrene phase formation which may be simply
STYRENE-BUTADIENE-STRENE
BLOCK COPOLYMERS /
35
POLYBUTADIENE
\
POLYSTYRENE
Fig. 1. Morphology of a styrene-butadiene-styrene block copolymer containing approximately 25% polystyrene; surface morphology also shown.
reversed by addition of a solvent for polystyrene or by raising the temperature above polystyrene’s glass transition temperature (lOO°C), the elastomer can be easily extruded or molded to give the required final form. This easy processing is an added advantage of these materials.
Surface Reaction Scheme The most common treatment proposed for preparing nonthrombogenic surfaces involves the coupling of heparin to the polymer. Ionically bound heparin3r4is useful for the preparation of surfaces with short-term compatibility, but ion exchange of the heparin with blood components renders this approach unsuitable for long-term implants. However, the covalent coupling of heparin to an alcohol polymer through an acetal bridge results in materials which retain long-term blood compatibility in arterio-venous shunts in dogs.5 I n order to utilize this technique the surfaces of our SBS copolymers have been hydroxylated to provide reactive sites for such a coupling procedure. By modifying only the surface of the triblock copolymers, their unique mechanical properties can be retained. Hydroxylation of polybutadiene copolymers in solution is a wellestablished procedure that has, in fact, been used on SBS and SIS copolymers to prepare b i ~ m a t e r i a l s ,but ~ . ~since the ultimate purpose was to prepare sulphonated block copolymers, results reported therein
SEFTON AND ~ 1 E 1 2 R I L I ~
36
are of limited applicability to this study. These previous attempts to use solutlion hydroxylated copolymers for biomedical devices were unsuccessful principally because of the poor physical properties of the resulting materials. Hydroxylation in solution results in scission of the polymer chain, the appearance of large amounts of diblock polymer, and, as expected, the deterioration of the mechanical properties of the material.I0 By modifying only the surface of the triblock copolymers, a crosslinked surface layer is generated which is firmly attached t o the underlying substrate.ll This surface gel would bear a negligible fraction of any applied load and its weakness would not adversely affect the tensile strength or elastic modulus of the whole. It is established that each double bond in the polybutadiene block is oxidized by peracetic acid to form an epoxide product which is immediately cleaved by the acetic acid in the reaction bath, in the presence of mineral acid, to form a hydroxyacetate addition product (Fig. 2 ) . The acetate group is then replaced by a second hydroxyl group by base hydrolysis, the net result being a 1 , 2 glycol structure. I n this work, the SBS copolymer is in the solid phase during reaction and not in solution, consequently side reactions take on added importance. Aside from considerations of yield, intermolecular ethers resulting from epoxide-epoxide reactions are crosslinks which strongly affectthe mechanical and swelling properties of the resulting material. Another consideration common to both the solution and surface reactions is the different reactivities of the three double bond isomers in the polybutadiene block. The cis 1 , 4 and trans 1 , 4 isomers are 8v9
ALKENE
€POX IDE
HYDROXYLACETATE
GLYCOL
Fig. 2 . Hydroxylation reaction scheme.
STYRENE-BUTADIENEGSTYRENE BLOCK COPOLYMERS
37
approximately 25 times as reactive as the vinyl isomer resulting from 1 , 2 addition during the polymerization. (The microstructure of the polybutadiene block is controllable, to a certain ext$nt, by the choice of polymerization solvent.2) However, it might be supposed that due to a reduction in steric hindrance, the pendant double bonds from 1 , 2 addition would provide a pair of alcohol groups, one group being secondary and one primary, that might be more suitable for the subsequent coupling of heparin.
EXPERIMENTAL Experimental work was principally directed toward the determination and quantitative evaluation of the content of hydroxyl and other chemical groups in films of the copolymer as a function of time, temperature, composition of the reaction bath, and film thickness. Films of SBS (Shell Experimental Block Copolymer TR-41-2443, see Table I), solvent cast on mercury, were suspended in a flack in a reaction medium composed of peracetic acid (FMC, 40% peracetic acid), HzSO4, and varying amounts of acetic acid and water, all maintained a t a fixed temperature (30-45°C). After sufficient time had elapsed, the films were removed from the bath, washed free of acid, and then the acetate groups were hydrolyzed in 2 N KOH in a separate bath. The films were dried between microfiber glass disks under pressure. With a Perkin-Elmer model 521 grating infrared spectrometer, the spectrum between 3,800-3000 cm-1 of each of the dried films was TABLE I Characterization of SBS TR-41-244312
Block Molecular Weights 16,000-85,000- 17,000 Composition Polystyrene 27.7y0 Triblock 100 % Polybutadiene Microstructure cis 1 , 4 40% trans 1 , 4 49% 1,2 11%
SEFTON AND MIEIIRILL
38
recorded. The area under the OH stretching peak centered around 3,430 cm-I was calculated using the formula of R a n ~ s a y . ~ With ~ the absorptivity of the same peak determined from films of poly(viny1 alcohol), the values of
Lb
C Od.l:~ (0
=
film thickness, em; C ~ I = I
local concentration of hydroxyl groups, mol/cm3; 5 = position) were determined using the Lambert-Beer law. ( N o other chemical groups present absorb in this wave1engt)hregion.)
The values of b
1
C"H--
niol of hydroxyl groups per unit area of film (area concentration) were then plotted against the film thickness.
RESULTS Evaluation of Diffusion/Reaction Process By analysis of these curves of the extcnt of the reaction versus the film thickness (Figs. 3 , 4, 7, 8), the depth of penetration was deterininrd and the various parameters governing the diffusion/reaction process were evaluated. All the curves of hydroxyl content (area concentration) versus film thickness were found to show the same behavior as in Figure 3 : an initial linear increase in hydroxyl content followed by a gradual decrease beyond a certain maximum value. The linear portion of the curve, AH, defines the range of filnis over which the peracetic acid,
2t OO
1
I 1- 1-1-__ 40
80 120 160 FILM THICKNESS (microns)
200
Fig 3. Kxtent of reaction (area concentlation of hydioxyl groups) vs. film thickness, 3O0C, 180 min, 70' acetic arid
STY~~ENE-BUTADIENE-STYRENEBLOCK COPOLYMERS
39
under the given experimental conditions, has diffused through the complete film and all potentially reactive double bonds are presumed to have reacted. The slope of this portion is independent of time and temperature and only slightly dependent on the composition of the reaction bath. The value of the slope corresponds to a 24% conversion of the double bonds to the glycol structure in the 7Oy0 acetic acid-24% water reaction bath. While some of the double bonds are unreactive due to steric or electronic effects (e.g., vinyl groups), many of the double bonds reacted to form intramolecular or intermolecular ethers, as confirmed by qualitative analysis of the complete infrared spectrum. However, the maximum a t C in Figure 3 and the subsequent decrease was not t o be expected. In a simple sorption experiment for a given set of experimental conditions, there is a particular film thickness a t which the weight gain per unit area would become independent of film thickness, this film thickness being the depth of penetration of the permeant. In larger films, the polymer beyond this depth of penetration would have no effect on the amount of permeant being absorbed.I4 To understand the anomolous behavior in our system, the total oxygen content of some of our samples was determined (Fig. 4). To the extent that the assumption of a constant stoichiometric ratio (constant moles oxygen to moles double bond reacted) is valid, the total oxygen content reflects the number of double bonds reacted. This curve, as expected, asymptotically approaches a much higher level and a t a film thickness greater than
-I
Fig. 4. Extent of reaction (area concentration of oxygen and hydroxyl gloiips) vs. film thickness, 30"C, 180 min, 704; acetic acid.
SEFTON AND MEHRILL
40
that a t point C . Thus, it appears that the peracetic acid diffuses in a normal manner through the films, but a change in product distribution occurs in the thicker films. According t o this reasoning, by addition of oxygen to the polymer, the polymer is converted from a hydrophobic t o a hydrophilic material; the reacted surface region of the polymer therefore, swells in the acetic acid-water reaction bath. However, this swelling behavior is modified by the presence of the unreacted core (Fig. 5), since the swelling stresses near the surface are transferred by the elastic network t o the higher modulus unreacted portion of the polymer. This stress transfer is dependent on the ratio of the size of the surface region to that of the total material. Hence, the surface region of the thinner films swells more than that of the thicker films and as a result, the slower diffusing epoxide cleavage agents (acetic acid, water, sulphuric acid) are retarded in their permeation of the polymer. This decreases the lag between the advancing fronts of epoxide formation and hydroxyacetate formation in thinner films and a greater fraction of the epoxy groups introduced into the polymer are converted t o hydroxyl groups in these films than in the thicker films. Detailed quantitative analysis of the infrared spectra of these films1' confirms this hypothesis (Fig. 5). I
I
-
I
I
---
30°C ~
~
o
~
~ --• a
~ I
I
I
I
Hydroxyl Epoxy Cyclic c , Ether d Acyclic Ether
--*
-
........t..... ...........
.
......
-
bpC._r_._.-.-.-.-.------.-.-.-.-.~
I I
I
50
100
I
150
I
200
Film Thickness
I
I
I
250
300
350
(microns)
Fig. 5 . Composition of surface hydroxylated films as a function of film thickness; 30"C, 180 min, 70% acetic acid. f2 is the fraction of oxygen-containing functional groups that are of one type: hydroxyl, epoxy, acyclic ether, or tetrahydrofuran, mathematically defined as Ci/ C,. L
STYRENE-BUTADIENE-STYRENEBLOCK COPOLYMERS
41
OH
REACTED SURFACE
I
I
UNREACTED REACTED CORE SURFACE b
-
Fig. 6. Concentration profile of hydroxyl groups across film. S1, S , are solubilities of reaction mixture in reacted and unreacted regions, respectively (8, > Sz).
The model of Figure 6 also helps to explain the time dependence of the surface hydroxylation step. Figure 7 shows a series of hydroxyl content versus film thickness curves for films reacted a t various times under otherwise the same conditions. There is a 60 min ‘(induction period” before there is any evidence of hydroxyl groups in the infrared spectrum. The depth of penetration of the reaction 2 calculated from the OH absorptivity is, therefore, less than ~ . p (as and the minimum peak area distinguishable in the spectrum) with -T-
I
0
50
100 150 200 250 300 350 400 450 Film Thickness
(microns)
Fig. 7. Area concentration of hydroxyl groups vs. film thickness for various times; 40°C,71 % acetic acid.
SISFTON
42
mri
MERRILL
50 min of reaction, but is increased to 30 1 in the next 15 min. This induction period is also due to swelling in the surface region. As the surface region swells, the solubility of the peracetic acid in the polymer increases and the driving force for diffusion is increased commensurately. Hence, the diffusion rate increases with a subsequent increase in the penetration depth and the degree of swelling a t the surface which further increases the peracetic acid solubility. This “autoacceleration” behavior can be verified by increasing the concentration of acetic acid in the reaction bath. Acetic acid, being a better swelling agent for the reacted polymer, would decrease the duration of the induction period. This is shown in Figure 8, where the extent of reaction is greater after only 20 niin in 937, acetic acid than after 65 min in 71y0acetic acid.
Degradation I n the use of polybutadiene based elastomers as vascular prosthcses, particular attention must be paid t o the effects of long-term aging (approximately two to five years) in the biological environment on the elastomer. Potential aging mechanism^'^ are degradation by enzymatic action and thermal oxidation. Due to the absence of high temperatures, ozone, and ultraviolet light in the body, pure thermal degradation, ozonolysis, and oxidative photodegradation are not relevant here (but would be during the processing of these copolymers). __ 32 -
!
I
- 1 7 - i - 1
40°C
T
7
STYRENE-BUTADIENE-STYRENg;
BLOCK COP0LYMF:ItS
43
I n considering enzymatic action, no a przori remarks can be made regarding the stability of SBS copolymers in the presence of the enzymes found in blood a t 37°C and pH 7.4. However, from tests of elastomers using microbes not commonly found in humans it was concluded that the material susceptible to microbial attack is not the elastomer molecule itself, but rather the other material in the formulation.16 This problem was avoided here by using these ultrapure elastomers. Incubation of a few grams of SBS 1’13-41-2443 with soil microbes failed to result in any significant growth of the microbes. Although this is not conclusive with regard t o the biological stability of SBS copolymers i t is an encouraging observation. Thermal autooxidation is a free radical process involving the addition of molecular oxygen a t relatively low temperatures to a polymer, the net result of which, in unsaturated polymers, is chain scission and hence, degradation of mechanical properties.” Initiation may be by a variety of free radical generating processes, although the only possible ones here might be mechanical stress or thermal deconiposition of weak bonds (e.g., peroxides in the surface hydroxylated layer). While the SBS copolymers used in this study needed t o be stabilized with an antioxidant to prevent thermal oxidation initiated by ultraviolet light,6it was not clear that in the absence of ultraviolet light (as in the body) and with the presence of the surface hydroxylated layer, the oxidative stability of the polybutadiene block would be critical. Similar to the enzymatic stability, the oxidative stability must be studied further. Rlediating against any potential degradation process, whether i t be enzymatic or oxidative, is the fact that polystyrene is generally unaffected by these processes. As such, the reinforcement function (stress redistribution) of the polystyrene domains would remain intact and act t o minimize the effect of chain scission in the polybutadiene block on the bulk mechanical properties. I n effect, the process of aging would be prolonged.
CONCLUSION The general applicability of this laminate structure should be mentioned. By virtue of the polymerization process, SBS copolymers can be prepared with varying total and/or block molecular weights and thus with varying polystyrene contents. While the CQ-
44
SEFTON AND MERRILL
polymer containing 25% polystyrene is a high strength elastomer, a copolymer containing 90% polystyrene has the properties of a high impact strength polystyrene. Coupled with the ability t o vary the microstructure of the polybutadiene block, there is an extremely wide variation in the properties of SBS copolymers. There is, therefore, a high probability that a SBS copolymer can be found which would satisfy the mechanical requirements for any given biomaterial use. To make this material nonthrombogenic, the simple surface hydroxylation scheme developed here can be used t o provide reactive sites for whatever final treatment is necessary, whether i t be the covalent coupling of heparin as described here, the immobilization of fibrinolytic enzymes118or the binding of antiplatelet drugs.19 As such, by a suitable combination of a triblock substrate, surface modification and nonthrombogenic treatment any set of biomaterial requirements can be satisfied.
Nomenclature b Con S,
S2 z
film thickness, cm local concentration of hydroxyl groups in film, g mol/cm3 solubility of reaction mixture in reacted polymer solubility of reaction mixture in unreacted polymer position in diffusion direction
The authors express thanks for the support of this work under U.S. Public Health Grant No. NIH-5-POl-HLl4322 and are grateful to t,he Shell Chemical Company for supplying the materials used in this work. They also thank A. J . Grauer and I). L. Traiit for help in the development of the experimental proeedure. This paper was presented a t AIChE 77th National Meeting, Pittsburgh, June 3, 1974.
References 1. E. Nylias, E. L. Kupski, P. Burnett,, and It. &I. Waag, J . Biornecl. Muter. Res., 4, 371 (1970). 2 . 15. 13. Bradford and Id. D. McKeever, Progr. Polym. Sci., 3, 109 (1971). 3. G. Grode, I
VOL. 10, PP. 33-45 (1976)
Surface Hydroxylation of Styrene-Butadiene-Styrene Block Copolymers for Biomaterials MICHAEL V. SEFTON* and EDWARD W. MERRILL, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 ,
Summary This work pertains to the development of high strength elastomers potentially useful as nonthrombogenic cardiovascular prostheses. Triblock copolymers of the styrene-butadiene-styrene type have been subjected to surface hydroxylation which provide reactive sites a t the surface for the subsequent coupling of heparin while retaining the unique mechanical properties of the SBS copolymers. Curves of hydroxyl content versus the copolymer film thickness demonstrate the effect of swelling in the surface region on the product distribution and on the time dependence of the hydroxylation process. In addition, the effect of time, temperature, and the composition of the reaction bath on the diffusion/reaction process is shown. Finally, the general applicability of this surface modification scheme to the development of biomaterials is discussed.
INTRODUCTION Over the past several years many implantable devices have been devised which aid the circulation of blood in diseased patients. One of the simplest of these is the vascular prosthesis, designed to act as a replacement or bypass for arteries or veins which have become occluded. While Dacron velour is most commonly used for these prostheses, the use of a velour limits the use of the prosthesis to those portions of the circulatory system (principally the larger arteries) where the blood flow rate is sufficiently fast to prevent buildup of thrombin to levels adequate to activate fibrin formation. The de-
* Present address: Department of Chemical Iihgineering, University of Toronto, Toronto, Ontario b15S 1A4, Canada. 33 @ 1976 by John Wiley & Sons, Inc.
34
SEFTON AND MERRILL
velopment of a nonthrombogenic material suitable for these prostheses would significantly extend the applicability of vascular replacement surgery. I n order to prepare a suitable high strength elastomeric substrate potentially useful as nonthrombogenic cardiovascular prostheses, certain triblock copolymers of the styrene-butadiene-styrene (SBS) type have been subjected to surface hydroxylation. As a consequence of this treatment, a n elastomcric material modeling the laminate structure of the arterial wall has been developed, namely, a soft hydrophilic layer bound to a tough elastomeric substrate. The hydrophilic layer may be subsequently rendered nonthrombogenic by fixation of heparin therein.
MATERIALS AND METHODS
SBS as a Biomaterial The unique properties of the styrene-butadiene-styrene block copolymers makes them particularly useful as biomaterials. First, because of the requirements of the “living-polymer” anionic polymerization technique by which they are synthesized, they are ultrapure and contain no elutable contaminants which might render them toxic t o the human system. The work of Nyilas e t al.’ demonstrated the strong effect that trace contaminants and defects in the microstructure have on the blood compatibility of silicones. Second, as a result of the thermodynamic incompatibility of the polystyrene and polybutadiene blocks, there is a phase separation in the solid phase, with the formation, in a SBS containing 25% polystyrene, of glassy domains of polystyrene in a continuum of the rubbery polybutadiene (Fig. 1). However, because of the bonds between the two phases, the polystyrene domains act to keep the polybutadiene entanglements in place, in addition to acting as particles of reinforcing filler to give the copolymer, a t body temperature, the properties of a high strength elastomer (initial modulus, 650 psi; tensile strength, 3900 psi) .z Thus, without any further potentially contaminating crosslinking or strength inducing steps, the SBS copolymer containing approximately 257, wt polystyrene satisfies all the mechanical requirements of a vascular prosthesis. I n addition, because the dimensional stability of the SBS copolymer is the result of a polystyrene phase formation which may be simply
STYRENE-BUTADIENE-STRENE
BLOCK COPOLYMERS /
35
POLYBUTADIENE
\
POLYSTYRENE
Fig. 1. Morphology of a styrene-butadiene-styrene block copolymer containing approximately 25% polystyrene; surface morphology also shown.
reversed by addition of a solvent for polystyrene or by raising the temperature above polystyrene’s glass transition temperature (lOO°C), the elastomer can be easily extruded or molded to give the required final form. This easy processing is an added advantage of these materials.
Surface Reaction Scheme The most common treatment proposed for preparing nonthrombogenic surfaces involves the coupling of heparin to the polymer. Ionically bound heparin3r4is useful for the preparation of surfaces with short-term compatibility, but ion exchange of the heparin with blood components renders this approach unsuitable for long-term implants. However, the covalent coupling of heparin to an alcohol polymer through an acetal bridge results in materials which retain long-term blood compatibility in arterio-venous shunts in dogs.5 I n order to utilize this technique the surfaces of our SBS copolymers have been hydroxylated to provide reactive sites for such a coupling procedure. By modifying only the surface of the triblock copolymers, their unique mechanical properties can be retained. Hydroxylation of polybutadiene copolymers in solution is a wellestablished procedure that has, in fact, been used on SBS and SIS copolymers to prepare b i ~ m a t e r i a l s ,but ~ . ~since the ultimate purpose was to prepare sulphonated block copolymers, results reported therein
SEFTON AND ~ 1 E 1 2 R I L I ~
36
are of limited applicability to this study. These previous attempts to use solutlion hydroxylated copolymers for biomedical devices were unsuccessful principally because of the poor physical properties of the resulting materials. Hydroxylation in solution results in scission of the polymer chain, the appearance of large amounts of diblock polymer, and, as expected, the deterioration of the mechanical properties of the material.I0 By modifying only the surface of the triblock copolymers, a crosslinked surface layer is generated which is firmly attached t o the underlying substrate.ll This surface gel would bear a negligible fraction of any applied load and its weakness would not adversely affect the tensile strength or elastic modulus of the whole. It is established that each double bond in the polybutadiene block is oxidized by peracetic acid to form an epoxide product which is immediately cleaved by the acetic acid in the reaction bath, in the presence of mineral acid, to form a hydroxyacetate addition product (Fig. 2 ) . The acetate group is then replaced by a second hydroxyl group by base hydrolysis, the net result being a 1 , 2 glycol structure. I n this work, the SBS copolymer is in the solid phase during reaction and not in solution, consequently side reactions take on added importance. Aside from considerations of yield, intermolecular ethers resulting from epoxide-epoxide reactions are crosslinks which strongly affectthe mechanical and swelling properties of the resulting material. Another consideration common to both the solution and surface reactions is the different reactivities of the three double bond isomers in the polybutadiene block. The cis 1 , 4 and trans 1 , 4 isomers are 8v9
ALKENE
€POX IDE
HYDROXYLACETATE
GLYCOL
Fig. 2 . Hydroxylation reaction scheme.
STYRENE-BUTADIENEGSTYRENE BLOCK COPOLYMERS
37
approximately 25 times as reactive as the vinyl isomer resulting from 1 , 2 addition during the polymerization. (The microstructure of the polybutadiene block is controllable, to a certain ext$nt, by the choice of polymerization solvent.2) However, it might be supposed that due to a reduction in steric hindrance, the pendant double bonds from 1 , 2 addition would provide a pair of alcohol groups, one group being secondary and one primary, that might be more suitable for the subsequent coupling of heparin.
EXPERIMENTAL Experimental work was principally directed toward the determination and quantitative evaluation of the content of hydroxyl and other chemical groups in films of the copolymer as a function of time, temperature, composition of the reaction bath, and film thickness. Films of SBS (Shell Experimental Block Copolymer TR-41-2443, see Table I), solvent cast on mercury, were suspended in a flack in a reaction medium composed of peracetic acid (FMC, 40% peracetic acid), HzSO4, and varying amounts of acetic acid and water, all maintained a t a fixed temperature (30-45°C). After sufficient time had elapsed, the films were removed from the bath, washed free of acid, and then the acetate groups were hydrolyzed in 2 N KOH in a separate bath. The films were dried between microfiber glass disks under pressure. With a Perkin-Elmer model 521 grating infrared spectrometer, the spectrum between 3,800-3000 cm-1 of each of the dried films was TABLE I Characterization of SBS TR-41-244312
Block Molecular Weights 16,000-85,000- 17,000 Composition Polystyrene 27.7y0 Triblock 100 % Polybutadiene Microstructure cis 1 , 4 40% trans 1 , 4 49% 1,2 11%
SEFTON AND MIEIIRILL
38
recorded. The area under the OH stretching peak centered around 3,430 cm-I was calculated using the formula of R a n ~ s a y . ~ With ~ the absorptivity of the same peak determined from films of poly(viny1 alcohol), the values of
Lb
C Od.l:~ (0
=
film thickness, em; C ~ I = I
local concentration of hydroxyl groups, mol/cm3; 5 = position) were determined using the Lambert-Beer law. ( N o other chemical groups present absorb in this wave1engt)hregion.)
The values of b
1
C"H--
niol of hydroxyl groups per unit area of film (area concentration) were then plotted against the film thickness.
RESULTS Evaluation of Diffusion/Reaction Process By analysis of these curves of the extcnt of the reaction versus the film thickness (Figs. 3 , 4, 7, 8), the depth of penetration was deterininrd and the various parameters governing the diffusion/reaction process were evaluated. All the curves of hydroxyl content (area concentration) versus film thickness were found to show the same behavior as in Figure 3 : an initial linear increase in hydroxyl content followed by a gradual decrease beyond a certain maximum value. The linear portion of the curve, AH, defines the range of filnis over which the peracetic acid,
2t OO
1
I 1- 1-1-__ 40
80 120 160 FILM THICKNESS (microns)
200
Fig 3. Kxtent of reaction (area concentlation of hydioxyl groups) vs. film thickness, 3O0C, 180 min, 70' acetic arid
STY~~ENE-BUTADIENE-STYRENEBLOCK COPOLYMERS
39
under the given experimental conditions, has diffused through the complete film and all potentially reactive double bonds are presumed to have reacted. The slope of this portion is independent of time and temperature and only slightly dependent on the composition of the reaction bath. The value of the slope corresponds to a 24% conversion of the double bonds to the glycol structure in the 7Oy0 acetic acid-24% water reaction bath. While some of the double bonds are unreactive due to steric or electronic effects (e.g., vinyl groups), many of the double bonds reacted to form intramolecular or intermolecular ethers, as confirmed by qualitative analysis of the complete infrared spectrum. However, the maximum a t C in Figure 3 and the subsequent decrease was not t o be expected. In a simple sorption experiment for a given set of experimental conditions, there is a particular film thickness a t which the weight gain per unit area would become independent of film thickness, this film thickness being the depth of penetration of the permeant. In larger films, the polymer beyond this depth of penetration would have no effect on the amount of permeant being absorbed.I4 To understand the anomolous behavior in our system, the total oxygen content of some of our samples was determined (Fig. 4). To the extent that the assumption of a constant stoichiometric ratio (constant moles oxygen to moles double bond reacted) is valid, the total oxygen content reflects the number of double bonds reacted. This curve, as expected, asymptotically approaches a much higher level and a t a film thickness greater than
-I
Fig. 4. Extent of reaction (area concentration of oxygen and hydroxyl gloiips) vs. film thickness, 30"C, 180 min, 704; acetic acid.
SEFTON AND MEHRILL
40
that a t point C . Thus, it appears that the peracetic acid diffuses in a normal manner through the films, but a change in product distribution occurs in the thicker films. According t o this reasoning, by addition of oxygen to the polymer, the polymer is converted from a hydrophobic t o a hydrophilic material; the reacted surface region of the polymer therefore, swells in the acetic acid-water reaction bath. However, this swelling behavior is modified by the presence of the unreacted core (Fig. 5), since the swelling stresses near the surface are transferred by the elastic network t o the higher modulus unreacted portion of the polymer. This stress transfer is dependent on the ratio of the size of the surface region to that of the total material. Hence, the surface region of the thinner films swells more than that of the thicker films and as a result, the slower diffusing epoxide cleavage agents (acetic acid, water, sulphuric acid) are retarded in their permeation of the polymer. This decreases the lag between the advancing fronts of epoxide formation and hydroxyacetate formation in thinner films and a greater fraction of the epoxy groups introduced into the polymer are converted t o hydroxyl groups in these films than in the thicker films. Detailed quantitative analysis of the infrared spectra of these films1' confirms this hypothesis (Fig. 5). I
I
-
I
I
---
30°C ~
~
o
~
~ --• a
~ I
I
I
I
Hydroxyl Epoxy Cyclic c , Ether d Acyclic Ether
--*
-
........t..... ...........
.
......
-
bpC._r_._.-.-.-.-.------.-.-.-.-.~
I I
I
50
100
I
150
I
200
Film Thickness
I
I
I
250
300
350
(microns)
Fig. 5 . Composition of surface hydroxylated films as a function of film thickness; 30"C, 180 min, 70% acetic acid. f2 is the fraction of oxygen-containing functional groups that are of one type: hydroxyl, epoxy, acyclic ether, or tetrahydrofuran, mathematically defined as Ci/ C,. L
STYRENE-BUTADIENE-STYRENEBLOCK COPOLYMERS
41
OH
REACTED SURFACE
I
I
UNREACTED REACTED CORE SURFACE b
-
Fig. 6. Concentration profile of hydroxyl groups across film. S1, S , are solubilities of reaction mixture in reacted and unreacted regions, respectively (8, > Sz).
The model of Figure 6 also helps to explain the time dependence of the surface hydroxylation step. Figure 7 shows a series of hydroxyl content versus film thickness curves for films reacted a t various times under otherwise the same conditions. There is a 60 min ‘(induction period” before there is any evidence of hydroxyl groups in the infrared spectrum. The depth of penetration of the reaction 2 calculated from the OH absorptivity is, therefore, less than ~ . p (as and the minimum peak area distinguishable in the spectrum) with -T-
I
0
50
100 150 200 250 300 350 400 450 Film Thickness
(microns)
Fig. 7. Area concentration of hydroxyl groups vs. film thickness for various times; 40°C,71 % acetic acid.
SISFTON
42
mri
MERRILL
50 min of reaction, but is increased to 30 1 in the next 15 min. This induction period is also due to swelling in the surface region. As the surface region swells, the solubility of the peracetic acid in the polymer increases and the driving force for diffusion is increased commensurately. Hence, the diffusion rate increases with a subsequent increase in the penetration depth and the degree of swelling a t the surface which further increases the peracetic acid solubility. This “autoacceleration” behavior can be verified by increasing the concentration of acetic acid in the reaction bath. Acetic acid, being a better swelling agent for the reacted polymer, would decrease the duration of the induction period. This is shown in Figure 8, where the extent of reaction is greater after only 20 niin in 937, acetic acid than after 65 min in 71y0acetic acid.
Degradation I n the use of polybutadiene based elastomers as vascular prosthcses, particular attention must be paid t o the effects of long-term aging (approximately two to five years) in the biological environment on the elastomer. Potential aging mechanism^'^ are degradation by enzymatic action and thermal oxidation. Due to the absence of high temperatures, ozone, and ultraviolet light in the body, pure thermal degradation, ozonolysis, and oxidative photodegradation are not relevant here (but would be during the processing of these copolymers). __ 32 -
!
I
- 1 7 - i - 1
40°C
T
7
STYRENE-BUTADIENE-STYRENg;
BLOCK COP0LYMF:ItS
43
I n considering enzymatic action, no a przori remarks can be made regarding the stability of SBS copolymers in the presence of the enzymes found in blood a t 37°C and pH 7.4. However, from tests of elastomers using microbes not commonly found in humans it was concluded that the material susceptible to microbial attack is not the elastomer molecule itself, but rather the other material in the formulation.16 This problem was avoided here by using these ultrapure elastomers. Incubation of a few grams of SBS 1’13-41-2443 with soil microbes failed to result in any significant growth of the microbes. Although this is not conclusive with regard t o the biological stability of SBS copolymers i t is an encouraging observation. Thermal autooxidation is a free radical process involving the addition of molecular oxygen a t relatively low temperatures to a polymer, the net result of which, in unsaturated polymers, is chain scission and hence, degradation of mechanical properties.” Initiation may be by a variety of free radical generating processes, although the only possible ones here might be mechanical stress or thermal deconiposition of weak bonds (e.g., peroxides in the surface hydroxylated layer). While the SBS copolymers used in this study needed t o be stabilized with an antioxidant to prevent thermal oxidation initiated by ultraviolet light,6it was not clear that in the absence of ultraviolet light (as in the body) and with the presence of the surface hydroxylated layer, the oxidative stability of the polybutadiene block would be critical. Similar to the enzymatic stability, the oxidative stability must be studied further. Rlediating against any potential degradation process, whether i t be enzymatic or oxidative, is the fact that polystyrene is generally unaffected by these processes. As such, the reinforcement function (stress redistribution) of the polystyrene domains would remain intact and act t o minimize the effect of chain scission in the polybutadiene block on the bulk mechanical properties. I n effect, the process of aging would be prolonged.
CONCLUSION The general applicability of this laminate structure should be mentioned. By virtue of the polymerization process, SBS copolymers can be prepared with varying total and/or block molecular weights and thus with varying polystyrene contents. While the CQ-
44
SEFTON AND MERRILL
polymer containing 25% polystyrene is a high strength elastomer, a copolymer containing 90% polystyrene has the properties of a high impact strength polystyrene. Coupled with the ability t o vary the microstructure of the polybutadiene block, there is an extremely wide variation in the properties of SBS copolymers. There is, therefore, a high probability that a SBS copolymer can be found which would satisfy the mechanical requirements for any given biomaterial use. To make this material nonthrombogenic, the simple surface hydroxylation scheme developed here can be used t o provide reactive sites for whatever final treatment is necessary, whether i t be the covalent coupling of heparin as described here, the immobilization of fibrinolytic enzymes118or the binding of antiplatelet drugs.19 As such, by a suitable combination of a triblock substrate, surface modification and nonthrombogenic treatment any set of biomaterial requirements can be satisfied.
Nomenclature b Con S,
S2 z
film thickness, cm local concentration of hydroxyl groups in film, g mol/cm3 solubility of reaction mixture in reacted polymer solubility of reaction mixture in unreacted polymer position in diffusion direction
The authors express thanks for the support of this work under U.S. Public Health Grant No. NIH-5-POl-HLl4322 and are grateful to t,he Shell Chemical Company for supplying the materials used in this work. They also thank A. J . Grauer and I). L. Traiit for help in the development of the experimental proeedure. This paper was presented a t AIChE 77th National Meeting, Pittsburgh, June 3, 1974.
References 1. E. Nylias, E. L. Kupski, P. Burnett,, and It. &I. Waag, J . Biornecl. Muter. Res., 4, 371 (1970). 2 . 15. 13. Bradford and Id. D. McKeever, Progr. Polym. Sci., 3, 109 (1971). 3. G. Grode, I
