- Email: [email protected]
poly(ethylene glycol)
Biomaterials 23 (2002) 2353–2358
Phase behavior and miscibility in blends of poly(sebacic anhydride)/ poly(ethylene glycol) Cheng-Kuang Chan, I-Ming Chu* Department of Chemical Engineering, National Tsing Hua University, Hsinchu 300, Tainan 30043, Taiwan Received 25 May 2001; accepted 29 October 2001
Abstract In this research, poly(sebacic anhydride) was synthesized via melt-condensation. The polymer was then blended with poly(ethylene glycol) in various ratios by solvent casting, to obtain the desired polymer blends. A differential scanning calorimeter was employed to investigate the crystalline behavior of the blends. Blends with under 10% poly(ethylene glycol) were found to consist of two partially miscible polymers in the amorphous phase. This compatibility of the two polymers may induce the crystallization of the poly(sebacic anhydride) component from decreasing the glass transition temperature of the poly(sebacic anhydride) component in the blending system due to the presence of poly(ethylene glycol) chain segments. Furthermore, phase separation occurred and the crystallinity of poly(sebacic anhydride) diminished when at least 20% poly(ethylene glycol) was present in the blends. These results were verified by the variation in the absorptions of carbonyl groups in infrared spectra; these spectra exhibit the changes in crystallinity of poly(sebacic anhydride) in the blends. r 2002 Published by Elsevier Science Ltd. Keywords: Poly(sebacic anhydride); Poly(ethylene glycol); Crystalline; Blending; Phase separation
1. Introduction In recent years, biodegradable polymeric materials have been found useful for drug delivery, bone and cartilage repairing, and tissue engineering applications. A number of novel approaches including poly(a-hydroxy acids), poly(aliphatic esters), poly(orthoesters), poly (phosphate esters), synthetic peptide-based polymers, and polysaccharides have been synthesized and applied as biodegradable polymers for biomedical application usages. Moreover, the integration of these polymeric materials and their interactions with host tissues have been studied [1,2]. During the past two decades, interest in a class of polymers, termed as polyanhydrides for biomedical applications, has grown tremendously. A number of papers have provided information on the synthesis of these polymers for the controlled delivery of drugs with surface-eroding behavior thereby providing a sustained release rate over a long period of time [3–9]. In the 1980s, Langer et al. illustrated that polyanhydrides have *Corresponding author. Tel.: +886-3-571-3704; fax: +886-3-5715408. E-mail address: [email protected] (I.-M. Chu). 0142-9612/02/$ - see front matter r 2002 Published by Elsevier Science Ltd. PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 3 7 0 - 2
advantages due to the delivery of bioactive substances, either locally or systematically, in vivo to achieve desired pharmacology effect. These materials are not only degradable but also highly biocompatible, as have been demonstrated by tissue response and toxicological study [1]. The highly hydrophobic polyanhydrides exhibit surface-eroding properties, and their rates of degradation can be modified by structure alterations in the polymer backbone. Another development in drug delivery is the use of poly(amino-acids) [10–12] or poly(ethylene glycol) (PEG) [13] to modify the properties of the delivery system, where these additional materials are used in the forms of pre-polymer or copolymer. In this research, blending of PEG with polyanhydride material was studied. In this study, poly(sebacic anhydride) (PSA) was synthesized from sebacic acid via melt-condensation. This material was then blended with PEG, which may adjust the releasing rate and the hydrophilic properties of the controlled drug delivery system. In general, higher crystallinity may enhance the strength of the materials. However, when used as a drug delivery system, higher crystallinity of the matrix may result in a slower releasing rate. On the other hand, phase behavior is also an important factor for drug release, which affects
2354
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
physical properties of the materials and may also cause different releasing rates for drugs with different hydrophilicities or possibly cause biphasic or two-step releasing profile, if drugs are partitioned into the separated matrixes [14]. Hence, melting points and heats of fusion of the distinct components in blends were investigated by differential scanning calorimeter to study crystalline and phase behavior, which can provide some critical information for further application of these materials in the drug delivery system or other biomedical application usages. The variation of amorphous and crystalline phases was further studied by infrared spectrophotometer analysis, and the relation between the blending ratio and phase behavior of the blends was determined.
2. Experimental 2.1. Materials Sebacic acid and acetic anhydride were purchased from Riedel-de Ha.en, and PEG (MW=2000) from Merck, respectively. Dichloromethane (HPLC grade, Fisons Scientific Eq.), toluene (Mallinckrodt), petroleum ether (anhydrous, J.T. Baker), and ethyl ether (anhydrous, J.T. Baker), were all HPLC grade and were used as received without further purification.
silane as internal standard for the examination of the synthesized polyanhydride. The molecular weight of the synthesized PSA was determined by gel permeation chromatography, which was constructed with a pump (JASCO PU-1580) and detector (JASCO RI-1530), using polystyrene (Shodex standard SL-105, Showa Denko) as the standard. In addition, the eluent used was chloroform with an elution rate of 1.5 ml/min. The samples of neat PSA, neat PEG, and PSA/PEG blends with various blending ratios were prepared from synthesized PSA and PEG by solvent casting. Using dichloromethane as a mutual solvent, solutions of both the polymers were mixed in the desired weight ratio and subsequently, cast onto aluminum dishes. Samples were then placed into a 351C vacuum oven for 20 min to remove the solvent for further investigations. A differential scanning calorimeter (DSC, TA 2010) was used for investigating the melting point and the heat of fusion of the blends. DSC thermograms of these materials were conducted under a dry-nitrogen atmosphere at a heating rate of 101C/min from 01C to 1251C. Besides, infrared spectra of these materials were obtained from Fourier transform infrared spectrophotometer (FTIR, Nicolet Avatar 320) at 2 cm1 resolution with 64 scans in spectral range of 4000–400 cm1.
3. Results and discussion
2.2. Synthesis of PSA
3.1. Characterization
After raising the temperature to 1401C, 24 g sebacic acid and 248 g acetic anhydride were charged into a drynitrogen-purged reactor. The reaction was processed by rigorous mixing for 20 min. The mixture was then placed in a rotary evaporator to remove acetic acid and unreacted acetic anhydride. The crude pre-polymer was recrystallized from dry toluene. Subsequently, the crystals were immersed in the solvent (1:1 of dry petroleum ether and ethyl ether) to extract traces of acetic anhydride and toluene. The refined pre-polymer was then obtained after being dried under vacuum [3]. The refined pre-polymer (3.0 g) was then placed into the reactor under a 1801C oil bath and vacuum for 120 min to carry out the melt-condensation process [1,3,15]. The final product was dissolved into dichloromethane first, and was purified by precipitation in dry petroleum ether. The precipitate was further extracted with ethyl ether anhydrous to give the purified polymer product.
The synthesized PSA was characterized by a solution NMR to confirm the structure. Figs. 1 and 2 present the 13 C- and 1H-NMR spectra, respectively, of pre-polymer and polyanhydride with assignments for distinct carbon and hydrogen atoms. The chemical shifts were all well resolved in these figures. The chemical shifts for the end group of the molecular 13C- and 1H-NMR spectra were both diminished after melt-condensation, revealing the success of the synthesis. The molecular weight of the desired polymer was obtained by gel permeation chromatography. The average molecular weight of the synthesized polyanhydride was 9900, and the polydispersity index was 2.74. Table 1 presents the chemical shifts obtained from NMR investigation and the molecular weight measured by GPC.
2.3. Methods and measurements 1
H- and 13C-NMR spectra were recorded on a 500 MHz instrument (Varian unity INOVA 500NMR spectrometer) using CDCl3 as solvent and tetramethyl-
3.2. DSC study of crystalline behavior Following the solvent casting, the crystallinity of both the neat polymer and the polymer blends was examined by a DSC. The profiles of the DSC thermogram indicate that heats of fusion of the two semicrystalline polymers can be easily determined by integrating the melting peaks. The heats of fusion of the neat PSA and PEG
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
Fig. 1.
2355
13
C-NMR spectra of: (a) pre-polymer and (b) PSA. The assignments for distinct carbon were made on the figure.
Fig. 2. 1H-NMR spectra of: (a) pre-polymer and (b) PSA. The assignments were made on the figure.
Table 1 Analytical data of synthesized pre-polymer and poly(sebacic anhydride) Materials
Pre-polymer PSA
13
C-NMR (CDCl3, TMS) (ppm)
21.82 (CH3), 24.08 (2CH2), 28.73 (2CH2), 35.17 (2CH2), 169.56 (2COO) 24.10 (2CH2), 28.65 (2CH2), 35.18 (2CH2), 169.56 (2COO)
1
H-NMR (CDCl3, TMS) (ppm)
1.30 2.19 1.30 2.42
(s, 8 H), 1.62 (4 H), (3 H), 2.41 (t, 4 H) (s, 8 H), 1.63 (4 H), (t, 4 H)
GPC results MW
PDI
F
F
9900
2.74
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
250
PEG
Heat of fusion (J/g)
200
150
100
50
PSA 0 0
20
40
60
80
100
WPEG (%)
Fig. 3. Dependences of fusion heats of PSA and PEG on the weight percent of PEG in the PSA/PEG blends. Linear decrease for the crystallinity of PSA and PEG with the blending ratio was also made on the figure.
were 132.0 and 220.6 J/g, respectively. Fig. 3 presents the heats of fusion of these two components in the blends, for each composition, as specified by the weight percentage of PEG. The dotted lines corresponding to PSA and PEG represent the product of the heats of fusion for the distinct components, and the corresponding blending ratio. These determined heats of fusion for the specific blending ratios were theoretical values, indicating that the polymers did not affect the crystallinity of any other component. These lines were compared with the data points for the crystallinity of individual components in the blending system, and thereby to provide information regarding the change in crystallinity that followed the variation of the blending ratio. The heat of fusion of the PEG component in the blends clearly shows systematic negative deviations from the dotted line. At a PEG content of 10%, the heat of fusion of the PEG component exhibits a large deviation from the dotted line, with an extremely low value of 0.76 J/g. No crystalline PEG component was observed in the DSC thermograms for samples with o5% PEG. This behavior is attributable to the presence of the other polymer component, PSA, which restricts the crystal growth of the PEG component. Interestingly, however, the heat of fusion of the PSA component exhibits a different behavior, positively, but deviates from the dotted line for a PEG content of 0–20%, although the deviation is small. A similar phenomenon has been reported for the poly(butane-1,4-diyl adipate)/poly(vinylidene fluoride) blending system [16]. In this study, the crystallinity of the poly(vinylidene fluoride) component (harder chain segments) increased from approximately 50% for the unblended material to approximately 60% for the blend, as the blending ratio of poly(butane-1,4diyl adipate) increased. However, the crystallinity of the
poly(butane-1,4-diyl adipate) component (softer chain segments) significantly declined with increase in the poly(vinylidene fluoride) content, and dropped to zero at approximately 90% poly(vinylidene fluoride) in the system, whose decrease was similar to that of the crystallinity of the PEG component in this study. The mechanism by which the addition of poly(butane-1,4diyl adipate) improves the crystallization of poly(vinylidene fluoride) was suggested to rely on the blending of softer chain segments to lower the glass transition temperature (Tg ) of harder polymer chains. Decrease in Tg may promote molecular chain mobility and enhance the crystallization of harder polymer chains in the blends. The crystallinity of PSAFhigher in the blend with 5% PEG (137.7 J/g) than in the neat PSA (132.0 J/ g)Fand the positive deviation for heats of fusion for PSA components of the blends containing 0–20% PEG reveals that the same phenomenon occurs in the PSA/ PEG system. This enhancement effect has also been reported for poly(ethylene oxide)/polycarbonate blending systems [17]. However, the crystalline behavior of PEG was affected by PSA, very differently from where the chain mobility of PEG was confined by harder polymer chains of PSA; the growth of crystals was thus restricted. With over 20% PEG blended in the system, heats of fusion of the PSA component exhibited a slight negative deviation from the dotted line. Accordingly, the fundamental phase behavior of PSA/PEG blends changes as PEG content is increased from 30%. The mechanism and the nature of these changes are not presently known. The melting temperature of the individual components of the blends may provide some information concerning the phase behavior. The variation in melting temperatures, shown in Fig. 4, indicates that increasing the blending ratio of either the PSA or the PEG component did not cause the melting temperature
85
PSA
80
Melting temperatures ( oC)
2356
75
55
50
PEG
45
40 0
20
40
60
80
100
WPEG (%)
Fig. 4. Dependences of melting temperatures of PSA and PEG on the weight percent of PEG in the PSA/PEG blends.
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
PSA/PEG 25/75
PSA/PEG 50/50
PSA/PEG 80/20
Absorption
of the other polymer component to decrease linearly. The melting temperature of PSA in the 5% PEG blend was 76.61C, apparently lower than the mean values. The melting temperature (42.41C) of PEG in the blend with 10% PEG was much lower than the averaged value. These findings indicate the partial miscibility of two polymers in the amorphous phase, and that the PEG chain segments not only increase the crystallinity of the PSA component, as well as hinder the thickening of the PSA crystal. Hence, the melting temperature of PSA with 5% PEG, decreased. However, the melting temperatures of both PSA and PEG components were raised to the temperature of the neat components when PEG content was increased to >20%, revealing phase separation in the amorphous phase, and denoting that the crystal growth for each of the two individual components was no longer influencing that of the other component.
2357
PSA/PEG 90/10 PSA/PEG 95/05
semicrystalline PSA
3.3. FTIR spectroscopy analysis The absorption peaks of saturated aliphatic anhydrides were at 1840–1800 and 1780–1740 cm1, due to the asymmetric and symmetric stretching vibrations of the carbonyl groups (CQO), respectively [18]. In the blending system, both bands were distinct and did not overlap the absorption peaks of PEG, thus, the crystalline behavior can be deduced from the variations of infrared absorptions. Fig. 5 reveals that the IR absorptions for neat PSA annealed above the melting temperature are different from those for semicrystalline PSA. Distinct peaks at 1818 and 1748 cm1 correspond to the asymmetric and symmetric stretching absorptions of the carbonyl groups on amorphous PSA, respectively. However, absorption peaks of carbonyl groups on semicrystalline PSA shift to low wavenumbers, 1809 (Dn ¼ 9 cm1) and 1741 cm1 (Dn ¼ 7 cm1). This result is attributable to the reduction of molecular motion of the well-stacked crystalline PSA chain segments. Additionally, although the absorption for symmetric stretching vibrations of the carbonyl groups (1748 cm1) did not vary with increasing PEG content, the absorption for asymmetric stretching vibrations of the carbonyl groups (ca. 1810 cm1) was indeed affected by the blending of PEG chain segments. These variations in absorption shifting may provide some information concerning the crystalline behavior of the blends. The figure demonstrates that all the absorption peaks for asymmetric stretching vibration of the carbonyl groups of PSA in the blends were at approximately 1809 cm1, except that in the blend with 75% PEG, which resulted in higher wavenumbers, it was around 1813 cm1. This result may follow from decreases in the crystallinity of the PSA components, as confirmed by the DSC results, discussed in Section 3.2.
amorphous PSA
1900
1850
1800
1750
1700
1650
-1
Wavenumbers (cm )
Fig. 5. Infrared spectra analysis of carbonyl groups at wavenumbers of 1900–1650 cm1 on amorphous and semicrystalline PSA and PSA/ PEG blends. The shoulders on the absorption for asymmetric stretching vibrations of carbonyl groups have been denoted by arrows.
For blends with 5–10% PEG content, an apparent shoulder was observed at approximately 1795 cm1, with a frequency shift of around Dn ¼ 23 cm1 to the absorption of the carbonyl group on amorphous PSA. This shift greatly exceeded that of Dn ¼ 9 cm1 for the asymmetric stretching absorptions of carbonyl groups on the crystalline PSA. This shoulder was prominent for a PEG content from 5% to 10%, and disappeared at a higher PEG content >20% in the blends. This frequency shifting does not follow from a dipole–dipole interaction, whose rather weak interaction may only cause frequency shifts in the range, Dn ¼ 3–6 cm1. Hence, the formation of the shoulder may be related to the formation of hydrogen bonds between the carbonyl group of PSA and the hydroxyl group on the end of PEG dissolved in the PSA phase [19–22]. These results strongly suggest that highly compatible PSA and PEG molecular chains exist with a PEG blending ratio of 5– 10%, furthermore, crystallinity of PSA increases as PEG content increases from 0% to 10%. It can be also found that phase separation occurred in the blending system when PEG content was at least 20%. The molecular chain segments of PEG aggregated from the PSA matrix to form PEG domains and the amount of PEG chains in the PSA matrix therefore declined. Consequently, the
2358
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
crystallinity of PSA decreased to the ordinary value of neat PSA, without enhancing molecular motion due to the presence of PEG chain segments. These results are consistent with the results from DSC studies.
4. Conclusions In this research, poly(sebacic anhydride) was successfully synthesized via melt-condensation. The synthesized polyanhydride was then blended with poly(ethylene glycol) to study the crystalline behavior. The blending system was found to be miscible in the amorphous phase for two polymers with under 10% of blended PEG. The crystallinity of the PSA component in such blends was enhanced by decreasing the glass transition temperature of the system that included the PEG blend. However, the crystal growth of PEG segments was suppressed by the presence of PSA. In addition, phases separated with at least 20% PEG in the blends. Consequently, without enhancing the molecular motion due to the presence of PEG chain segments in PSA, the crystallinity of PSA decreased to the ordinary value of neat PSA as the PEG content increased. The melting temperatures for both PSA and PEG components returned to the normal value of neat materials. These results were verified by infrared spectra of these blends, which revealed variations in asymmetric stretching absorptions for carbonyl groups, and the formation of hydrogen bonds in the blends with 5–10% PEG. Acknowledgements The authors are grateful to Dr. Hsin-Lung Chen of Department of Chemical Engineering, National Tsing Hua University, for the use of DSC and infrared spectrophotometer. References [1] Peppas NA, Langer RS. Biopolymers I. Heidelberg: Springer, 1993. [2] Hollinger JO. Biomedical application of synthetic biodegradable polymers. Boca Raton: CRC Press, 1995. [3] Domb AJ, Langer R. Polyanhydrides. I. Preparation of high molecular weight polyanhydrides. J Polym Sci: Polym Chem 1987;25:3373–86. [4] Wu MP, Tamada JA, Brem H, Langer R. In vivo versus in vitro degradation of controlled release polymers for intracranial surgical therapy. J Biomed Mater Res 1994;28:387–95.
[5] Shieh L, Tamada J, Chen I, Pang J, Domb A, Langer R. Erosion of a new family of biodegradable polyanhydrides. J Biomed Mater Res 1994;28:1465–75. [6] Mylonas CC, Tabata Y, Langer R, Zohar Y. Preparation and evaluation of polyanhydride microspheres containing gonadotropin-releasing hormone (GnRH), for inducing ovulation and spermiation in fish. J Control Release 1995;35:23–34. [7] Park E-S, Maniar M, Shah J. Effects of model compounds with varying physicochemical properties on erosion of polyanhydride devices. J Control Release 1996;40:111–21. [8] Park E-S, Maniar M, Shah JC. Influence of physicochemical properties of model compounds on their release from biodegradable polyanhydride devices. J Control Release 1997;48:67–78. [9] Anastasiou TJ, Uhrich KE. Novel polyanhydrides with enhanced thermal and solubility properties. Macromolecules 2000;33: 6217–21. [10] Hartmann M, Schulz V. Synthesis and characterization of polyanhydrides containing amido group. Makromol Chem 1989;190:2133–40. [11] Uhrich KE, Gupta A, Thomas TT, Laurencin CT, Langer R. Synthesis and characterization of degradable poly(anhydride-coimides). Macromolecules 1995;28:2184–93. [12] Hanes J, Chiba M, Langer R. Synthesis and characterization of degradable anhydride-co-imide terpolymers containing trimellitylimido-l-tyrosine: novel polymers for drug delivery. Macromolecules 1996;29:5279–87. [13] Jiang HL, Zhu KJ. Preparation, characterization and degradation characteristics of polyanhydrides containing poly(ethylene glycol). Polym Int 1999;48:47–52. [14] Jeong B, Bae TH, Kim SW. Drug release from biodegradable injectable thermosensitive hydrogel of PEG–PLGA–PEG triblock copolymers. J Control Release 2000;63:155–63. [15] Domb AJ, Maniar M. Absorbable biopolymers derived from dimer fatty acids. J Polym Sci: Polym Chem 1993;31: 1275–85. [16] Penning JP, St John Manley R. Miscible blends of two crystalline polymers. 1. Phase behavior and miscibility in blends of poly(vinylidene fluoride) and poly(1,4-butylene adipate). Macromolecules 1996;29:77–83. [17] Brus J, Dybel J, Schmidt P, Kratochvil J, Baldrain J. Order and mobility in polycarbonate-poly(ethylene oxide) blends studied by solid-state NMR and other techniques. Macromolecules 2000;33:6448–59. [18] Socrates G. Infrared characteristic group frequencies. Great Britain: A Wiley–Interscience Publication, 1980. [19] Wu H-D, Ma C-CM, Chu PP, Tseng H-T, Lee C-T. The phase behaviour of novolac type phenolic resin blended with poly(adipic ester). Polymer 1998;39:2859–65. [20] Wang F-Y, Ma C-CM, Wu H-D. Hydrogen bonding in polyamide toughened novolac type phenolic resin. J Appl Polym Sci 1999;74:2283–9. [21] Chan C-K, Chu I-M, Lee W, Chin W-K. Preparation and properties of organic–inorganic hybrid materials based on poly(butyl methacrylate-co-(3-(methacyloxypropyl)) trimethoxysilane. Macromol Chem Phys 2001;202:911–6. [22] Chan C-K, Chu I-M. Effect of hydrogen bonding on the glass transition behavior of poly(acrylic acid)/silica hybrid materials prepared by sol–gel process. Polymer 2001;42:6089–93.
Phase behavior and miscibility in blends of poly(sebacic anhydride)/ poly(ethylene glycol) Cheng-Kuang Chan, I-Ming Chu* Department of Chemical Engineering, National Tsing Hua University, Hsinchu 300, Tainan 30043, Taiwan Received 25 May 2001; accepted 29 October 2001
Abstract In this research, poly(sebacic anhydride) was synthesized via melt-condensation. The polymer was then blended with poly(ethylene glycol) in various ratios by solvent casting, to obtain the desired polymer blends. A differential scanning calorimeter was employed to investigate the crystalline behavior of the blends. Blends with under 10% poly(ethylene glycol) were found to consist of two partially miscible polymers in the amorphous phase. This compatibility of the two polymers may induce the crystallization of the poly(sebacic anhydride) component from decreasing the glass transition temperature of the poly(sebacic anhydride) component in the blending system due to the presence of poly(ethylene glycol) chain segments. Furthermore, phase separation occurred and the crystallinity of poly(sebacic anhydride) diminished when at least 20% poly(ethylene glycol) was present in the blends. These results were verified by the variation in the absorptions of carbonyl groups in infrared spectra; these spectra exhibit the changes in crystallinity of poly(sebacic anhydride) in the blends. r 2002 Published by Elsevier Science Ltd. Keywords: Poly(sebacic anhydride); Poly(ethylene glycol); Crystalline; Blending; Phase separation
1. Introduction In recent years, biodegradable polymeric materials have been found useful for drug delivery, bone and cartilage repairing, and tissue engineering applications. A number of novel approaches including poly(a-hydroxy acids), poly(aliphatic esters), poly(orthoesters), poly (phosphate esters), synthetic peptide-based polymers, and polysaccharides have been synthesized and applied as biodegradable polymers for biomedical application usages. Moreover, the integration of these polymeric materials and their interactions with host tissues have been studied [1,2]. During the past two decades, interest in a class of polymers, termed as polyanhydrides for biomedical applications, has grown tremendously. A number of papers have provided information on the synthesis of these polymers for the controlled delivery of drugs with surface-eroding behavior thereby providing a sustained release rate over a long period of time [3–9]. In the 1980s, Langer et al. illustrated that polyanhydrides have *Corresponding author. Tel.: +886-3-571-3704; fax: +886-3-5715408. E-mail address: [email protected] (I.-M. Chu). 0142-9612/02/$ - see front matter r 2002 Published by Elsevier Science Ltd. PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 3 7 0 - 2
advantages due to the delivery of bioactive substances, either locally or systematically, in vivo to achieve desired pharmacology effect. These materials are not only degradable but also highly biocompatible, as have been demonstrated by tissue response and toxicological study [1]. The highly hydrophobic polyanhydrides exhibit surface-eroding properties, and their rates of degradation can be modified by structure alterations in the polymer backbone. Another development in drug delivery is the use of poly(amino-acids) [10–12] or poly(ethylene glycol) (PEG) [13] to modify the properties of the delivery system, where these additional materials are used in the forms of pre-polymer or copolymer. In this research, blending of PEG with polyanhydride material was studied. In this study, poly(sebacic anhydride) (PSA) was synthesized from sebacic acid via melt-condensation. This material was then blended with PEG, which may adjust the releasing rate and the hydrophilic properties of the controlled drug delivery system. In general, higher crystallinity may enhance the strength of the materials. However, when used as a drug delivery system, higher crystallinity of the matrix may result in a slower releasing rate. On the other hand, phase behavior is also an important factor for drug release, which affects
2354
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
physical properties of the materials and may also cause different releasing rates for drugs with different hydrophilicities or possibly cause biphasic or two-step releasing profile, if drugs are partitioned into the separated matrixes [14]. Hence, melting points and heats of fusion of the distinct components in blends were investigated by differential scanning calorimeter to study crystalline and phase behavior, which can provide some critical information for further application of these materials in the drug delivery system or other biomedical application usages. The variation of amorphous and crystalline phases was further studied by infrared spectrophotometer analysis, and the relation between the blending ratio and phase behavior of the blends was determined.
2. Experimental 2.1. Materials Sebacic acid and acetic anhydride were purchased from Riedel-de Ha.en, and PEG (MW=2000) from Merck, respectively. Dichloromethane (HPLC grade, Fisons Scientific Eq.), toluene (Mallinckrodt), petroleum ether (anhydrous, J.T. Baker), and ethyl ether (anhydrous, J.T. Baker), were all HPLC grade and were used as received without further purification.
silane as internal standard for the examination of the synthesized polyanhydride. The molecular weight of the synthesized PSA was determined by gel permeation chromatography, which was constructed with a pump (JASCO PU-1580) and detector (JASCO RI-1530), using polystyrene (Shodex standard SL-105, Showa Denko) as the standard. In addition, the eluent used was chloroform with an elution rate of 1.5 ml/min. The samples of neat PSA, neat PEG, and PSA/PEG blends with various blending ratios were prepared from synthesized PSA and PEG by solvent casting. Using dichloromethane as a mutual solvent, solutions of both the polymers were mixed in the desired weight ratio and subsequently, cast onto aluminum dishes. Samples were then placed into a 351C vacuum oven for 20 min to remove the solvent for further investigations. A differential scanning calorimeter (DSC, TA 2010) was used for investigating the melting point and the heat of fusion of the blends. DSC thermograms of these materials were conducted under a dry-nitrogen atmosphere at a heating rate of 101C/min from 01C to 1251C. Besides, infrared spectra of these materials were obtained from Fourier transform infrared spectrophotometer (FTIR, Nicolet Avatar 320) at 2 cm1 resolution with 64 scans in spectral range of 4000–400 cm1.
3. Results and discussion
2.2. Synthesis of PSA
3.1. Characterization
After raising the temperature to 1401C, 24 g sebacic acid and 248 g acetic anhydride were charged into a drynitrogen-purged reactor. The reaction was processed by rigorous mixing for 20 min. The mixture was then placed in a rotary evaporator to remove acetic acid and unreacted acetic anhydride. The crude pre-polymer was recrystallized from dry toluene. Subsequently, the crystals were immersed in the solvent (1:1 of dry petroleum ether and ethyl ether) to extract traces of acetic anhydride and toluene. The refined pre-polymer was then obtained after being dried under vacuum [3]. The refined pre-polymer (3.0 g) was then placed into the reactor under a 1801C oil bath and vacuum for 120 min to carry out the melt-condensation process [1,3,15]. The final product was dissolved into dichloromethane first, and was purified by precipitation in dry petroleum ether. The precipitate was further extracted with ethyl ether anhydrous to give the purified polymer product.
The synthesized PSA was characterized by a solution NMR to confirm the structure. Figs. 1 and 2 present the 13 C- and 1H-NMR spectra, respectively, of pre-polymer and polyanhydride with assignments for distinct carbon and hydrogen atoms. The chemical shifts were all well resolved in these figures. The chemical shifts for the end group of the molecular 13C- and 1H-NMR spectra were both diminished after melt-condensation, revealing the success of the synthesis. The molecular weight of the desired polymer was obtained by gel permeation chromatography. The average molecular weight of the synthesized polyanhydride was 9900, and the polydispersity index was 2.74. Table 1 presents the chemical shifts obtained from NMR investigation and the molecular weight measured by GPC.
2.3. Methods and measurements 1
H- and 13C-NMR spectra were recorded on a 500 MHz instrument (Varian unity INOVA 500NMR spectrometer) using CDCl3 as solvent and tetramethyl-
3.2. DSC study of crystalline behavior Following the solvent casting, the crystallinity of both the neat polymer and the polymer blends was examined by a DSC. The profiles of the DSC thermogram indicate that heats of fusion of the two semicrystalline polymers can be easily determined by integrating the melting peaks. The heats of fusion of the neat PSA and PEG
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
Fig. 1.
2355
13
C-NMR spectra of: (a) pre-polymer and (b) PSA. The assignments for distinct carbon were made on the figure.
Fig. 2. 1H-NMR spectra of: (a) pre-polymer and (b) PSA. The assignments were made on the figure.
Table 1 Analytical data of synthesized pre-polymer and poly(sebacic anhydride) Materials
Pre-polymer PSA
13
C-NMR (CDCl3, TMS) (ppm)
21.82 (CH3), 24.08 (2CH2), 28.73 (2CH2), 35.17 (2CH2), 169.56 (2COO) 24.10 (2CH2), 28.65 (2CH2), 35.18 (2CH2), 169.56 (2COO)
1
H-NMR (CDCl3, TMS) (ppm)
1.30 2.19 1.30 2.42
(s, 8 H), 1.62 (4 H), (3 H), 2.41 (t, 4 H) (s, 8 H), 1.63 (4 H), (t, 4 H)
GPC results MW
PDI
F
F
9900
2.74
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
250
PEG
Heat of fusion (J/g)
200
150
100
50
PSA 0 0
20
40
60
80
100
WPEG (%)
Fig. 3. Dependences of fusion heats of PSA and PEG on the weight percent of PEG in the PSA/PEG blends. Linear decrease for the crystallinity of PSA and PEG with the blending ratio was also made on the figure.
were 132.0 and 220.6 J/g, respectively. Fig. 3 presents the heats of fusion of these two components in the blends, for each composition, as specified by the weight percentage of PEG. The dotted lines corresponding to PSA and PEG represent the product of the heats of fusion for the distinct components, and the corresponding blending ratio. These determined heats of fusion for the specific blending ratios were theoretical values, indicating that the polymers did not affect the crystallinity of any other component. These lines were compared with the data points for the crystallinity of individual components in the blending system, and thereby to provide information regarding the change in crystallinity that followed the variation of the blending ratio. The heat of fusion of the PEG component in the blends clearly shows systematic negative deviations from the dotted line. At a PEG content of 10%, the heat of fusion of the PEG component exhibits a large deviation from the dotted line, with an extremely low value of 0.76 J/g. No crystalline PEG component was observed in the DSC thermograms for samples with o5% PEG. This behavior is attributable to the presence of the other polymer component, PSA, which restricts the crystal growth of the PEG component. Interestingly, however, the heat of fusion of the PSA component exhibits a different behavior, positively, but deviates from the dotted line for a PEG content of 0–20%, although the deviation is small. A similar phenomenon has been reported for the poly(butane-1,4-diyl adipate)/poly(vinylidene fluoride) blending system [16]. In this study, the crystallinity of the poly(vinylidene fluoride) component (harder chain segments) increased from approximately 50% for the unblended material to approximately 60% for the blend, as the blending ratio of poly(butane-1,4diyl adipate) increased. However, the crystallinity of the
poly(butane-1,4-diyl adipate) component (softer chain segments) significantly declined with increase in the poly(vinylidene fluoride) content, and dropped to zero at approximately 90% poly(vinylidene fluoride) in the system, whose decrease was similar to that of the crystallinity of the PEG component in this study. The mechanism by which the addition of poly(butane-1,4diyl adipate) improves the crystallization of poly(vinylidene fluoride) was suggested to rely on the blending of softer chain segments to lower the glass transition temperature (Tg ) of harder polymer chains. Decrease in Tg may promote molecular chain mobility and enhance the crystallization of harder polymer chains in the blends. The crystallinity of PSAFhigher in the blend with 5% PEG (137.7 J/g) than in the neat PSA (132.0 J/ g)Fand the positive deviation for heats of fusion for PSA components of the blends containing 0–20% PEG reveals that the same phenomenon occurs in the PSA/ PEG system. This enhancement effect has also been reported for poly(ethylene oxide)/polycarbonate blending systems [17]. However, the crystalline behavior of PEG was affected by PSA, very differently from where the chain mobility of PEG was confined by harder polymer chains of PSA; the growth of crystals was thus restricted. With over 20% PEG blended in the system, heats of fusion of the PSA component exhibited a slight negative deviation from the dotted line. Accordingly, the fundamental phase behavior of PSA/PEG blends changes as PEG content is increased from 30%. The mechanism and the nature of these changes are not presently known. The melting temperature of the individual components of the blends may provide some information concerning the phase behavior. The variation in melting temperatures, shown in Fig. 4, indicates that increasing the blending ratio of either the PSA or the PEG component did not cause the melting temperature
85
PSA
80
Melting temperatures ( oC)
2356
75
55
50
PEG
45
40 0
20
40
60
80
100
WPEG (%)
Fig. 4. Dependences of melting temperatures of PSA and PEG on the weight percent of PEG in the PSA/PEG blends.
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
PSA/PEG 25/75
PSA/PEG 50/50
PSA/PEG 80/20
Absorption
of the other polymer component to decrease linearly. The melting temperature of PSA in the 5% PEG blend was 76.61C, apparently lower than the mean values. The melting temperature (42.41C) of PEG in the blend with 10% PEG was much lower than the averaged value. These findings indicate the partial miscibility of two polymers in the amorphous phase, and that the PEG chain segments not only increase the crystallinity of the PSA component, as well as hinder the thickening of the PSA crystal. Hence, the melting temperature of PSA with 5% PEG, decreased. However, the melting temperatures of both PSA and PEG components were raised to the temperature of the neat components when PEG content was increased to >20%, revealing phase separation in the amorphous phase, and denoting that the crystal growth for each of the two individual components was no longer influencing that of the other component.
2357
PSA/PEG 90/10 PSA/PEG 95/05
semicrystalline PSA
3.3. FTIR spectroscopy analysis The absorption peaks of saturated aliphatic anhydrides were at 1840–1800 and 1780–1740 cm1, due to the asymmetric and symmetric stretching vibrations of the carbonyl groups (CQO), respectively [18]. In the blending system, both bands were distinct and did not overlap the absorption peaks of PEG, thus, the crystalline behavior can be deduced from the variations of infrared absorptions. Fig. 5 reveals that the IR absorptions for neat PSA annealed above the melting temperature are different from those for semicrystalline PSA. Distinct peaks at 1818 and 1748 cm1 correspond to the asymmetric and symmetric stretching absorptions of the carbonyl groups on amorphous PSA, respectively. However, absorption peaks of carbonyl groups on semicrystalline PSA shift to low wavenumbers, 1809 (Dn ¼ 9 cm1) and 1741 cm1 (Dn ¼ 7 cm1). This result is attributable to the reduction of molecular motion of the well-stacked crystalline PSA chain segments. Additionally, although the absorption for symmetric stretching vibrations of the carbonyl groups (1748 cm1) did not vary with increasing PEG content, the absorption for asymmetric stretching vibrations of the carbonyl groups (ca. 1810 cm1) was indeed affected by the blending of PEG chain segments. These variations in absorption shifting may provide some information concerning the crystalline behavior of the blends. The figure demonstrates that all the absorption peaks for asymmetric stretching vibration of the carbonyl groups of PSA in the blends were at approximately 1809 cm1, except that in the blend with 75% PEG, which resulted in higher wavenumbers, it was around 1813 cm1. This result may follow from decreases in the crystallinity of the PSA components, as confirmed by the DSC results, discussed in Section 3.2.
amorphous PSA
1900
1850
1800
1750
1700
1650
-1
Wavenumbers (cm )
Fig. 5. Infrared spectra analysis of carbonyl groups at wavenumbers of 1900–1650 cm1 on amorphous and semicrystalline PSA and PSA/ PEG blends. The shoulders on the absorption for asymmetric stretching vibrations of carbonyl groups have been denoted by arrows.
For blends with 5–10% PEG content, an apparent shoulder was observed at approximately 1795 cm1, with a frequency shift of around Dn ¼ 23 cm1 to the absorption of the carbonyl group on amorphous PSA. This shift greatly exceeded that of Dn ¼ 9 cm1 for the asymmetric stretching absorptions of carbonyl groups on the crystalline PSA. This shoulder was prominent for a PEG content from 5% to 10%, and disappeared at a higher PEG content >20% in the blends. This frequency shifting does not follow from a dipole–dipole interaction, whose rather weak interaction may only cause frequency shifts in the range, Dn ¼ 3–6 cm1. Hence, the formation of the shoulder may be related to the formation of hydrogen bonds between the carbonyl group of PSA and the hydroxyl group on the end of PEG dissolved in the PSA phase [19–22]. These results strongly suggest that highly compatible PSA and PEG molecular chains exist with a PEG blending ratio of 5– 10%, furthermore, crystallinity of PSA increases as PEG content increases from 0% to 10%. It can be also found that phase separation occurred in the blending system when PEG content was at least 20%. The molecular chain segments of PEG aggregated from the PSA matrix to form PEG domains and the amount of PEG chains in the PSA matrix therefore declined. Consequently, the
2358
C.-K. Chan, I.-M. Chu / Biomaterials 23 (2002) 2353–2358
crystallinity of PSA decreased to the ordinary value of neat PSA, without enhancing molecular motion due to the presence of PEG chain segments. These results are consistent with the results from DSC studies.
4. Conclusions In this research, poly(sebacic anhydride) was successfully synthesized via melt-condensation. The synthesized polyanhydride was then blended with poly(ethylene glycol) to study the crystalline behavior. The blending system was found to be miscible in the amorphous phase for two polymers with under 10% of blended PEG. The crystallinity of the PSA component in such blends was enhanced by decreasing the glass transition temperature of the system that included the PEG blend. However, the crystal growth of PEG segments was suppressed by the presence of PSA. In addition, phases separated with at least 20% PEG in the blends. Consequently, without enhancing the molecular motion due to the presence of PEG chain segments in PSA, the crystallinity of PSA decreased to the ordinary value of neat PSA as the PEG content increased. The melting temperatures for both PSA and PEG components returned to the normal value of neat materials. These results were verified by infrared spectra of these blends, which revealed variations in asymmetric stretching absorptions for carbonyl groups, and the formation of hydrogen bonds in the blends with 5–10% PEG. Acknowledgements The authors are grateful to Dr. Hsin-Lung Chen of Department of Chemical Engineering, National Tsing Hua University, for the use of DSC and infrared spectrophotometer. References [1] Peppas NA, Langer RS. Biopolymers I. Heidelberg: Springer, 1993. [2] Hollinger JO. Biomedical application of synthetic biodegradable polymers. Boca Raton: CRC Press, 1995. [3] Domb AJ, Langer R. Polyanhydrides. I. Preparation of high molecular weight polyanhydrides. J Polym Sci: Polym Chem 1987;25:3373–86. [4] Wu MP, Tamada JA, Brem H, Langer R. In vivo versus in vitro degradation of controlled release polymers for intracranial surgical therapy. J Biomed Mater Res 1994;28:387–95.
[5] Shieh L, Tamada J, Chen I, Pang J, Domb A, Langer R. Erosion of a new family of biodegradable polyanhydrides. J Biomed Mater Res 1994;28:1465–75. [6] Mylonas CC, Tabata Y, Langer R, Zohar Y. Preparation and evaluation of polyanhydride microspheres containing gonadotropin-releasing hormone (GnRH), for inducing ovulation and spermiation in fish. J Control Release 1995;35:23–34. [7] Park E-S, Maniar M, Shah J. Effects of model compounds with varying physicochemical properties on erosion of polyanhydride devices. J Control Release 1996;40:111–21. [8] Park E-S, Maniar M, Shah JC. Influence of physicochemical properties of model compounds on their release from biodegradable polyanhydride devices. J Control Release 1997;48:67–78. [9] Anastasiou TJ, Uhrich KE. Novel polyanhydrides with enhanced thermal and solubility properties. Macromolecules 2000;33: 6217–21. [10] Hartmann M, Schulz V. Synthesis and characterization of polyanhydrides containing amido group. Makromol Chem 1989;190:2133–40. [11] Uhrich KE, Gupta A, Thomas TT, Laurencin CT, Langer R. Synthesis and characterization of degradable poly(anhydride-coimides). Macromolecules 1995;28:2184–93. [12] Hanes J, Chiba M, Langer R. Synthesis and characterization of degradable anhydride-co-imide terpolymers containing trimellitylimido-l-tyrosine: novel polymers for drug delivery. Macromolecules 1996;29:5279–87. [13] Jiang HL, Zhu KJ. Preparation, characterization and degradation characteristics of polyanhydrides containing poly(ethylene glycol). Polym Int 1999;48:47–52. [14] Jeong B, Bae TH, Kim SW. Drug release from biodegradable injectable thermosensitive hydrogel of PEG–PLGA–PEG triblock copolymers. J Control Release 2000;63:155–63. [15] Domb AJ, Maniar M. Absorbable biopolymers derived from dimer fatty acids. J Polym Sci: Polym Chem 1993;31: 1275–85. [16] Penning JP, St John Manley R. Miscible blends of two crystalline polymers. 1. Phase behavior and miscibility in blends of poly(vinylidene fluoride) and poly(1,4-butylene adipate). Macromolecules 1996;29:77–83. [17] Brus J, Dybel J, Schmidt P, Kratochvil J, Baldrain J. Order and mobility in polycarbonate-poly(ethylene oxide) blends studied by solid-state NMR and other techniques. Macromolecules 2000;33:6448–59. [18] Socrates G. Infrared characteristic group frequencies. Great Britain: A Wiley–Interscience Publication, 1980. [19] Wu H-D, Ma C-CM, Chu PP, Tseng H-T, Lee C-T. The phase behaviour of novolac type phenolic resin blended with poly(adipic ester). Polymer 1998;39:2859–65. [20] Wang F-Y, Ma C-CM, Wu H-D. Hydrogen bonding in polyamide toughened novolac type phenolic resin. J Appl Polym Sci 1999;74:2283–9. [21] Chan C-K, Chu I-M, Lee W, Chin W-K. Preparation and properties of organic–inorganic hybrid materials based on poly(butyl methacrylate-co-(3-(methacyloxypropyl)) trimethoxysilane. Macromol Chem Phys 2001;202:911–6. [22] Chan C-K, Chu I-M. Effect of hydrogen bonding on the glass transition behavior of poly(acrylic acid)/silica hybrid materials prepared by sol–gel process. Polymer 2001;42:6089–93.