ε-Decalactone - American Chemical Society

Article pubs.acs.org/Biomac

ε‑Decalactone: A Thermoresilient and Toughening Comonomer to Poly(L‑lactide) Peter Olsén, Tina Borke, Karin Odelius, and Ann-Christine Albertsson* Department of Fibre and Polymer Technology, School of Chemical Science and Engineering, KTH, Royal Institute of Technology, SE-100 44 Stockholm, Sweden S Supporting Information *

ABSTRACT: The renewable monomer ε-decalactone is an excellent partner to L-lactide, where their copolymers overcome inherent drawbacks of polylactide, such as low thermal stability and brittleness. ε-Decalactone is a seven-membered lactone that was successfully polymerized with Sn(Oct)2 and 1,5,7-triazabicyclo[4.4.0]dec-5-ene into both an amorphous homopolymer and copolymers with high molecular weight, low dispersity, and predicted macromolecular architecture. The thermoresilient nature of ε-decalactone is reflected in a high polymerization ceiling temperature and increased thermal stability for the prepared copolymers. The high ceiling temperature enables easy modulation of the polymerization rate via temperature while maintaining architectural control. The apparent rate constant was increased 15-fold when the temperature was increased from 110 to 150 °C. Copolymers of L-lactide and ε-decalactone, either with the latter as a central block in triblock polymers or with randomly positioned monomers, exhibited exceptionally tough material characteristics. The triblock copolymer had an elongation-at-break 250 times greater than that of pure poly(L-lactide). The toughness of the copolymers is attributed to the flexible nature of the polymer derived from the monomer ε-decalactone and to the segment immiscibility. These properties result in phase separation to soft and hard domains, which provides the basis for the elastomeric behavior.

■

INTRODUCTION In an ever-growing and changing world, material renewability and degradability are key factors for a sustainable future. An array of natural, renewable raw materials are available for polymer technology, and many of these come from the agricultural sector.1 A proven methodology for the utilization of this resource is through microorganism fermentation.2 This can provide a synthetic route to a very interesting class of compounds, the lactones, which ultimately can be used as monomers in the syntheses of renewable, degradable polymers. Numerous polymerization routes for the synthesis of aliphatic polyesters have been developed;3−5 however, the ring-opening polymerization (ROP) of lactones is the most commonly employed method.4 The ring size of monomer lactones is an important factor in their ability to ring-open; their propensity for ring-closing reactions, however, is equally important for achieving a high-molecular-weight polymer.6 The ideal ROP should exhibit a low depolymerization rate (a high equilibrium monomer concentration) and a high polymerization rate. Methods to provide increased flexibility of the main chain to enhance the properties of poly(L-lactide) are widely sought. A common strategy to yield these properties in a final polymer is through copolymerization with a complementing monomer, which, depending on the strategy employed, can yield block, random, or gradient copolymers. The structure of the initiator also plays an important role, and the careful design of the © XXXX American Chemical Society

initiator result in the formation of bifunctional, star-shaped or branched systems tailored to provide an optimal combination of properties in the copolymer products.7−11 In searching for new monomers that enhance ductility for poly(L-lactide), two critical aspects besides the flexibility of the monomer are the reactivities of the individual monomers and copolymer segmental miscibility. If the reactivities of the comonomers with one another is low or if the addition is performed in sequence, blocky structures will be produced. If the resultant blocks also are immiscible, phase separation into harder and softer domains will occur. Harder domains can subsequently act as physical cross-linking sites and result in a material with enhanced ductility.12 This concept has been widely applied to increase the toughness of L-lactide, mainly with ε-caprolactone but also with other flexible monomers, such as trimethylene carbonate (TMC)13,14 and 1,5-dioxepan2-one (DXO).15,16 ε-Caprolactone is virtually immiscible with lactide due to their different chemical natures,17 and also 6methyl-ε-caprolactone exhibits a similar behavior.18 Poly(δ-decalactone) has been shown to be amorphous and have a very low Tg value;19 both are properties that induce flexibility in the main chain of the polymer. The decalactones are a class of renewable lactones composed of 10 carbons, with Received: May 21, 2013 Revised: June 27, 2013

A

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

transferred into the reaction vessel using a new, disposable syringe in the glovebox. All reactions were stirred at a constant temperature that was maintained (±2 °C) using an IKAMAG RCT basic safety control magnetic stirrer. Aliquots for investigation with 1H nuclear magnetic resonance (NMR) and gel permeation chromatography (GPC) were withdrawn from the reaction vessels at regular time intervals using new, disposable syringes while the vessels were flushed with nitrogen gas. Polymerizations with Sn(Oct)2 were performed in bulk. The monomer(s), initiator, and Sn(Oct)2 ([M]/[Sn(Oct)2] ≈ 200; [M]/ [I] ≈ 200−400) were added to the reaction vessel under an inert atmosphere and immersed into a thermostated oil bath (110 °C). After completion of the reaction, the reaction mixture was cooled to room temperature, and the copolymers were dissolved in chloroform and precipitated three consecutive times in methanol. The precipitates were dried under reduced pressure for 4 days. Benzyl alcohol was used as the initiator in the syntheses of the poly(ε-decalactone) (PεDL) and poly(L-lactide) (PLLA) homopolymers. All copolymerizations were initiated with 1,4-butanediol. For the syntheses of block copolymers, the soft segment block was polymerized first and allowed to react completely before LLA was added to form the crystallizable end blocks. Polymerizations with TBD were performed in bulk in a manner analogous to that of the reactions with Sn(Oct)2. The monomer(s), initiator, and TBD ([M]/[TBD] ≈ 200; [M]/[I] ≈ 200) were added to the reaction vessel under an inert atmosphere, and the reaction vessel was immersed in a thermostated oil bath (110 °C). After 24 h, the reaction mixtures were cooled to room temperature and analyzed by GPC and 1H NMR. Sample Preparation. For the preparation of films, the polymers (∼3%, w/w) were dissolved in chloroform, filtered, and cast in glass Petri dishes. The solvent was evaporated, and the films were dried under reduced pressure at ambient temperatures for at least one week before analysis. Instruments. Nuclear Magnetic Resonance. 1H NMR (400.13 MHz) and 13C NMR (100.62 MHz) spectra were recorded on a Bruker Avance 400 spectrometer at 298 K. For the measurements, either ∼10 mg (1H NMR) or ∼100 mg (13C NMR) of the polymer was dissolved in 0.8 mL of CDCl3 in a 5 mm diameter sample tube. 1H NMR spectra were calibrated using the residual protons in CDCl3 (7.26 ppm), and 13C NMR spectra were calibrated against CDCl3 (77.0 ppm). Copolymer compositions and monomer conversions were determined from 1H NMR spectra through a comparison of the relative intensities of the monomer and polymer peaks (δ(LLA) = 5.01 ppm, δ(PLLA) = 5.17 ppm, δ(DL) = 4.21 ppm, δ(PDL) = 4.83 ppm). 13 C NMR was used to qualitatively determine the block sequences. Gel Permeation Chromatography. GPC was performed on a Verotech PL-GPC 50 Plus equipped with a PL RI detector and two PolarGel M organic columns, 300 × 7.5 mm (Varian, Santa Clara, CA), and the results were used to determine the number-average molecular weights (Mn) and polydispersity indices (PDIs) of the polymer blocks during and after polymerization. Samples were injected with a PL-AS RT autosampler (Polymer Laboratories) with chloroform as the mobile phase at 30 °C and a flow rate of 1 mL/min; toluene was used as an internal standard. The calibration was performed using polystyrene standards with narrow molecular weight distributions and with molecular weights that ranged from 160 to 371000 g/mol. Differential Scanning Calorimetry (DSC). The thermal properties of the materials were investigated under a nitrogen atmosphere using a Mettler Toledo DSC 820 module. The samples (4−7 mg) were held at −70 °C for 10 min under flowing nitrogen prior to heating. They were then heated from 70 to 200 °C at a rate of 10 °C/min, held at 200 °C for 2 min, cooled to −70 °C at a rate of 10 °C/min, and held at this temperature for 2 min. Finally, the samples were heated from −70 to +200 °C at a rate of 10 °C/min. The melting temperature (Tm) and the glass transition temperature (Tg) were derived from the second heating scan. The maximum value of the curve was noted as Tm, whereas Tg was taken as the midpoint of the glass transition. According to the results obtained by DSC, PεDL is totally amorphous.

an alkyl side chain at the ring-closing position. Consequently, the length of the side chain depends on the ring size of the lactone. Decalactones are commercially available, with ring sizes of five (γ-decalactone, γDL), six (δ-decalactone, δDL), or seven (ε-decalactone, εDL). These naturally occurring monomers can be produced through fungal technology and are widely used in the flavoring and fragrance industries;2,20 however, they have received little attention in the field of polymer science.19,21,22 The thermodynamics and kinetics of polymerization differ greatly between the different decalactones. From a thermodynamic perspective, the polymerization of ε-decalactone is the most favorable due to differences between the ceiling temperatures (Tc) of the six- and seven-membered lactones. The difference in T c between ε-caprolactone and δvalerolactone has been shown to be approximately 100 °C6,23 or more.24 When the δ-lactone is substituted on the ringclosing position, a further decrease in the Tc value is observed22,24 and results in Tc values for δ-substituted lactones just above room temperature. A Tc value this low seriously restricts the synthetic and commercial potential of these lactones. In this study, we aspire to synthesize a tough and more thermoresistant, fully degradable poly(L-lactide)-based material, solely from renewable resources. Our hypothesis is that the chemical structure ε-decalactone is ideal for finalizing this ambition. We have previously shown that the amorphous, flexible, hydrophilic polymer derived from DXO in copolymers with L-lactide significantly improves the toughness of the final material.15,16 Likewise, the anticipated flexible properties of εdecalactone should correspond to the DXO-based materials, but on the contrary being more hydrophobic and originating from renewable resources. The flexible amorphous and hydrophobic properties of εDL, in conjunction with the more rigid poly(L-lactide) should result in immiscibility, leading to phase separation into hard−rigid and soft−flexible domains in the material. This will provide the basis for a highly flexible material originating from renewable resources. In addition, εdecalactone should also display good polymerizability, i.e., a high equilibrium monomer concentration and a high ceiling temperature, properties originating from its ring size.

■

EXPERIMENTAL SECTION

Materials. εDL (≥99%, Sigma-Aldrich, Sweden) was dried over calcium hydride for at least 12 h and was subsequently distilled at reduced pressure under an inert atmosphere prior to use. L-Lactide (LLA) (≥99%, Boehringer Ingelheim) was recrystallized twice from toluene (HPLC grade, Fisher Scientific, Germany) and once from dry toluene (99.8%, Sigma-Aldrich, Sweden) and was dried in vacuo for at least 48 h prior to use. Stannous octoate (Sn(Oct)2) (Sigma-Aldrich, Sweden) was dried over molecular sieves (3 Å) before use. Benzyl alcohol (≥99%, Sigma-Aldrich, Sweden) and 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) (98%, Sigma-Aldrich, Sweden) were used as received, and 1,4-butanediol (99%, Sigma-Aldrich, Sweden) was distilled at reduced pressure and stored over 3 Å molecular sieves. Chloroform (HPLC grade, Fisher Scientific, Germany), methanol (general purpose grade, Fisher Scientific, Germany), and methylene chloride (HPLC grade, Fisher Scientific, Germany) were used without further purification. Chloroform-d (99.8%, with silver foil, Cambridge Isotope Laboratories) was used as received. Polymerizations. All reaction vessels were silanized before use. Typically, the desired amounts of the reactants were weighed into a two-neck, round-bottom flask (25 mL) under a nitrogen atmosphere in a glovebox (Mbraun MB 150-GI). Each flask was equipped with a magnetic stirring bar and sealed with a two-way valve and a septum. For solution polymerizations, the desired volume of dry solvent was B

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Scheme 1. Polymerization of ε-Decalactone with Sn(Oct)2 or TBD as the Catalyst and Benzyl Alcohol as the Initiator

Table 1. Comparison of the Polymerization of ε-Decalactone with That of Lactones with Similar Structures polymer

solvent

conversion (%)

time (h)

T (°C)

Mn (g/mol)

Mn(GPC) (g/mol)

PDI

PεDLSn(Oct)2

none

200:1:1

73a

48

110

17600b

8200c

1.14

PεDLSn(Oct)2

none

600:3:1

80a

360

110

77500b

40500c

1.21

110 110 rt rt rt

b

12200c

1.45

20800f 16500f

1.16 1.12

PεDLTBD PδDLTBD22 PδDLTBD22 PεCLTBD27 PδVLTBD27

none none none C6D6 C6D6

[M]0:[BnOH]:[cat.]

200:1:1 200:1:1 200:1:1 200:1:1 333:1:1.65

a

86 35d 80 52 77

24 10 8 0.5

23300 34100e 34100e 28800e 20000e

a

Determined by 1H NMR. bCalculated from 1H NMR by the integration relative ratio of the monomer and initiator times the monomer molecular weight. cDetermined by CHCl3-GPC with a polystyrene standard. dMaximum obtainable conversion at this temperature. eCalculated from the given initiator to monomer ratio. fDetermined by THF-GPC.

polymers. The homopolymer of ε-decalactone was sticky, amorphous, and oily and was able to flow at molecular weights as high as 40000 g/mol. In comparison to Sn(Oct)2, TBD was almost twice as active in the ROP of εDL. After 24 h at [M]0/ [I] = 200, the TBD-catalyzed polymerization had reached 86% conversion (Table 1), whereas the Sn(Oct)2-catalyzed polymerization reached only 48% conversion in the same reaction time (Figure 2a). The ROP of εDL can be easily followed by 1H NMR. Upon polymerization, a distinct change in the chemical shift occurred, which was especially obvious for the α- and ε-protons (Figure 1). The α-proton peak position shifted from 2.61 to 2.25 ppm, whereas the positions of the ε-protons shifted from 4.21 to 4.81 ppm. Notably, the molecular weight determined by gel permeation chromatography strongly underestimates the theoretical value calculated by 1H NMR, in contrast to what is normally seen in aliphatic polyesters.26 This underestimation is likely due to the

To calculate the approximate degree of crystallinity, we therefore assumed that only PLLA contributes to the heat of fusion. The degree of crystallinity (Xc) was calculated using the following equation: ⎛ ΔHf ⎞ Xc = ⎜ ⎟ × 100 ⎝ ΔHf °Wf ⎠

(1)

where Xc is the degree of crystallinity, wf is the weight fraction of PLLA in the copolymer, ΔHf is the heat of fusion of the sample, and ΔHf0 is the heat of fusion of 100% crystalline PLLA (93 J·g−1).25 Thermogravimetric Analysis (TGA). Thermal stabilities were evaluated by thermogravimetric analysis under a nitrogen atmosphere using a Mettler Toledo TGA/DSC 820 module with a sample mass of 4−6 mg and a heating rate of 10 °C/min. All samples were heated from 30 to 600 °C. Atomic Force Microscopy (AFM). The surface morphologies of the polymer films were determined by AFM in tapping mode, and phase imaging was performed using a NanoScope IIIa MultiMode atomic force microscope (Digital Instruments) via a 7850 EV scanner and a silicon-etched probe tip (RTESP, antimony-doped Si, Bruker AFM Probes) with a normal spring constant (k) of 20−80 N/m and a resonant frequency (f 0) of 306−366 kHz. The scan rate was in the range of 0.5−0.7 Hz. The scan size was 500 nm (512 × 512 pixels).

■

RESULTS AND DISCUSSION εDL can be produced through fungal biotechnology and is utilized by the flavor and fragrance industries;2,20 however, it is also representative of a very interesting class of monomers that can be used in the syntheses of aliphatic polyesters. The monomer is a seven-membered lactone with a butyl side group at the ε-position; hence, its structure closely resembles that of ε-caprolactone. Its homopolymer is anticipated to be completely amorphous due to the racemic stereochemistry of the butyl group; it is also expected to be more hydrophobic than ε-caprolactone. The low Tg and amorphous nature of this polymer should in conjunction with a rigid comonomer such as L-lactide yield a copolymer with a high flexibility. The ROP of εDL was explored using two commonly used catalysts, Sn(Oct)2 and TBD (Scheme 1). The polymerizations were performed at 110 °C in bulk with benzyl alcohol as the initiator unless otherwise described. Both catalysts were active in the ROP of ε-decalactone and yielded high-molecular-weight

Figure 1. 1H NMR with peak designation of the polymerization of εdecalactone with Sn(Oct)2 (mol 1%) as the catalyst and benzyl alcohol ([M]/[I] = 200) as the initiator. C

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Figure 2. Kinetic behavior with Sn(Oct)2 as the catalyst and benzyl alcohol as the initiator at 110 °C (a) and 150 °C (b). Squares and triangles indicate the conversion and ln([M]0/[M]) as functions of time. Plots c and d show the dependence of Mn and PDI on the conversion, where squares represent Mn and triangles represent the PDI.

caprolactone and δ-valerolactone (δVL).27,28 The main difference between the kinetic and thermodynamic behavior of sixand seven-membered unsubstituted lactones is the polymerization rate, which is higher for the six-membered than the seven-membered lactone, whereas the ceiling temperature is lower for the former.6,24 The ideal polymerization is fast and proceeds to full monomer conversion. Because the most straightforward strategy to increase the polymerization rate is to increase the reaction temperature, an acceptably high ceiling temperature is important for maintaining high monomer conversion. To assess the effect of the reaction temperature on the polymerization of εDL, reactions at 110 and 150 °C were evaluated with the Sn(Oct)2 catalyst. The polymerization of εDL proceeded in a controlled manner with a linear relationship between the conversion and the molecular weight at both 110 and 150 °C. As expected, the higher reaction temperature promoted a higher polymerization rate. The apparent rate constant of polymerization (kpapp) was calculated from the slope of the curve defined by the log of the monomer conversion ratio as a function of time (Figure 2). The apparent rate constant of εDL polymerization at 150 °C was approximately 15 times greater than that at 110 °C, with only a slight increase in the PDI. However, different reaction temperatures did not provide any measurable difference in the equilibrium monomer concentration. Even at 217 °C, 99% monomer conversion was observed. This result indicates a ceiling temperature so high that an approximate value could not be calculated. Similar trends have been reported for δvalerolactone and ε-caprolactone. The difference in the ceiling temperatures for δ-valerolactone and ε-caprolactone has been reported to be approximately 100 °C;6,23 however, other values reported for the ceiling temperature of ε-caprolactone suggest that the difference is even greateras much as 1700 °C.24 Although such a high ceiling temperature has no practical importance because the polymer would have degraded by other mechanisms long before this temperature was reached, it highlights both the problem with assessing the ceiling

hydrodynamic volume of the polymer chain becoming suppressed in chloroform because of its higher polarity than that of PεDL. Gel permeation chromatography in DMF showed the same trend and resulted in suppression of the determined molecular weight to an even greater extent. This result is consistent with DMF being more polar than chloroform. However, better control over the molecular weight distribution was achieved with Sn(Oct)2 (Table 1). The ROP of εDL can be easily followed by 1H NMR. Upon polymerization, a distinct change in the chemical shift occurred, which was especially obvious for the α- and ε-protons (Figure 1). The α-proton peak position shifted from 2.61 to 2.25 ppm, whereas the positions of the ε-protons shifted from 4.21 to 4.81 ppm. Notably, the molecular weight determined by gel permeation chromatography strongly underestimates the theoretical value calculated by 1H NMR, in contrast to what is normally seen in aliphatic polyesters.26 This underestimation is likely due to the hydrodynamic volume of the polymer chain becoming suppressed in chloroform because of its higher polarity than that of PεDL. Gel permeation chromatography in DMF showed the same trend and resulted in suppression of the determined molecular weight to an even greater extent. This result is consistent with DMF being more polar than chloroform. To place the reactivity of ε-decalactone in a larger context and provide a general overview, it was compared to those of several well-known monomers with similar structures (Table 1), although a completely accurate comparison is not possible due to the difference in solvent, time, and temperature among the different polymerizations. This indicated that the polymerization rates of decalactones are substantially lower than those of their unsubstituted analogues. The reduction in rate of the substituted lactones compared to that of ε-caprolactone (εCL) is believed to be due to the lower nucleophilic character of the propagating chain end that originates from a secondary alcohol. The TBD-catalyzed polymerization of δDL shows a higher rate compared to that of εDL (Table 1), as do the ROPs of εD

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Scheme 2. Depiction of the Two Different Monomer Addition Strategies for the Copolymerization of ε-Decalactone and LLactide Using Sn(Oct)2 as the Catalyst and 1,4-Butanediol as the Initiator

temperature for the ε-caprolactone ring and the sizable difference between the behaviors of six- and seven-membered lactones.22 We concluded that the ceiling temperature for εdecalactone is more than 170 °C higher than that of δdecalactone. Further experiments are needed to determine the exact value. The nucleophilicity of the propagating chain end is a very important aspect in the copolymerization of two monomers. The nucleophilicity for propagating alcohol chain ends roughly follows the trend primary > secondary > tertiary from high to low reactivity. This trend was evident in the copolymerization of LLA and ε-caprolactone when Al(OiPr)3 was used as the initiator. Upon simultaneous monomer addition, a gradient copolymer will be formed where the first part of the polymer mainly contains LLA and the second part mainly contains εcaprolactone, despite the fact that ε-caprolactone exhibits a much higher polymerization rate during homopolymerization than does LLA. A reversal of this addition preference is possible, however, through modulation of the chain end with bulky bidendate ligands.29 The secondary nature of the propagating chain for both LLA and εDL during polymerization, combined with the sterically unhindered catalyst Sn(Oct)2, should enable a translation of the rate of polymerization into the copolymerization system. To create two different architectures for copolymers that contain LLA and ε-decalactone, we used two different monomer addition strategies: simultaneous addition and sequential addition (Scheme 2). As determined from the chemical shift of the carbonyl carbons in the 13C NMR spectra, the sequential monomer addition of LLA to a macroinitiator of εDL yielded a very pure triblock copolymer. This result indicated a low degree of transesterification. In contrast, simultaneous monomer addition yielded a molecular structure that closely can be described as a “random” copolymer (Figure 3). To elucidate the reactivity preferences in the copolymerization of LLA and εDL, monomer conversion as a function of time was examined after simultaneous addition of both monomers (Supporting Information, Figures S1 and S2). As expected, LLA was much more reactive during copolymerization, reaching 90% conversion after just 30 min, in contrast to εDL, which reached the same conversion only after 24 h. This difference in reactivities during copolymerization was subsequently used as part of our strategy for forming triblock copolymers through sequential addition.

Figure 3. 13C NMR spectra of the carbonyl region of the homo- and copolymers of LLA and ε-decalactone.

The chemical shift of the carbonyl carbon is highly sensitive to its surrounding monomers; therefore,13C NMR can be used to determine the composition and structure of the copolymer chain.30 The spectra of PLLA and PεDL homopolymers contain distinct singlets at 169.6 and 173.3 ppm, respectively. With sequential monomer addition, only these singlet peaks were observed, which indicates that a pure triblock structure was formed. The formation of this structure was further supported by results obtained from GPC. The GPC results indicated that PεDL acted as a macroinitiator, which resulted in the disappearance of its peak after the addition of LLA. For the copolymers synthesized via simultaneous monomer addition, several carbonyl peaks between the homopolymer singlets were observed in the 13C NMR spectra (Supporting Information, Figures S3 and S4). These peaks originate from the numerous possible combinations of LLA and εDL triads, similar to what has been previously observed in the copolymerization of εCL and LLA.30,31 However, the greater intensities of the homopolymer peaks were interpreted as a relatively greater abundance of homopolymer sequences in the macromolecule. Hence, random copolymers of PLLA and PεDL have a somewhat “blocky” macromolecular structure (Figure 3). The differences in crystallinity, flexibility, and miscibility of PεDL and PLLA should result in very interesting properties of their copolymers. Three different copolymers were therefore synthesized to examine the differences between the copolymers and the homopolymers (Table 2). The feed ratio and final polymer composition were in good agreement with each other. E

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Table 2. Number-Average Molecular Weights and Polydisperity Indices for the Homo- and Copolymers of ε-Decalactone and LLactide polymera(mol %)

Mn,theoryb (g/mol)

Mnc (g/mol)

PDIc

LLA/εDLfeed (mol %)

LLA/εDLd (mol %)

P(εDL)100 PLLA20−PεDL60−PLLA20 P(εDL50-r-LLA50) PLLA40−PεDL20−PLLA40 PLLA100

81700 64000 63900 83400 53600

40500 32700 54100 36300 58700

1.21 1.40 1.50 1.34 1.46

0/100 40/60 50/50 80/20 100/0

0/100 40/60 51/49 82/18 100/0

Synthesized in bulk at 110 °C with Sn(Oct)2 as the catalyst. bDetermined by the monomer to initiator ratio with respect to conversion. Determined by GPC. dDetermined by 1H NMR.

a c

The copolymer synthesized through sequential addition, PLLA20−PεDL60−PLLA20, separates into randomized cylinders with widths of approximately 15 nm. The chemical similarity between the end blocks allows bridging between the harder domains, and hence, a dispersed microstructure is formed. In the AFM image of the copolymer synthesized through simultaneous addition, P(εDL50-r-LLA50), the randomized patchy pattern indicates the existence of pure block domains. The hard block domains of PLLA in P(εDL50-r-LLA50) act as cross-linking sites in the material, and this cross-linking results in a highly ductile material that showed an elongation-at-break greater than 600% in mechanical testing (Table 3). The thermal degradation of polyesters occurs in one or several steps and may involve thermohydrolysis, zipper-like depolymerization, thermo-oxidative degradation, or transesterification reactions.34 The degradation of both PLLA and PεDL homopolymers proceeded in a single step. The degradation temperature for PεDL was approximately 100 °C higher than that observed for PLLA (Figure 5) and is comparable to the literature value for ε-caprolactone, indicating the good thermal stability of the seven-membered lactones.35,36

To illuminate the microphase structures of the materials, AFM phase images were recorded in tapping mode. In the tapping mode, AFM provides chemical resolution of surfaces by sensing mechanical differences, such as differences in the elasticity of the domains.32 Observed phase shifts are a result of the different interactions of the tip with the polymer blocks. For multiblock copolymers, the phase behavior is controlled by the interactions between the blocks and by their relative lengths.33 If the polymer segments are largely immiscible, microphase separation will occur, and ordered structures will be formed. The appearance of these ordered structures will depend on the constitution and relative volume fractions of the different blocks. The AFM images showed microphase separation (Figure 4), which was expected because the PLLA and PεDL segments

Figure 4. AFM phase images of PLLA/PεDL triblock and random copolymers and neat PLLA recorded in tapping mode. The scan size is 500 × 500 nm2, and the scale is −20° to +20°. The images were flattened using the third derivation. The text insets describe the polymer architectures and compositions. Figure 5. Thermal decomposition patterns of the LLA and εDL homo- and copolymers.

should be immiscible. PLLA is more polar than PεDL due to more polar ester functionalities in its repeating unit. In addition, the hydrophobic butyl group on the ε-position of PεDL enhances this discrepancy. Furthermore, the differences in the solid states, i.e., semicrystalline versus amorphous, and in elasticity contribute to the incompatibility.

An elevated decomposition temperature, as compared to that of PLLA, was observed for all of the copolymers. The block structures of the triblock copolymers determine their thermal degradation patterns. The PLLA blocks degraded first, followed by the PεDL blocks, as a result of the inherent lower thermal

Table 3. Thermal and Mechanical Properties of the Homo- and Copolymersa polymer PLLA100 P(εDL)100 P(εDL50-r-LLA50) PLLA20−PεDL60−PLLA20 PLLA40−PεDL20−PLLA40 a

Tg,1b (°C) −53 −27 −51 −52

± ± ± ±

0.2 0.5 0.2 0.5

Tg,2b (°C)

Tm,1b (°C)

Xcc (%)

Emod (MPa)

strain at break (%)

51 ± na

168 ± 0.0

50

44 ± 0.8 45 ± 1.6 54 ± na

151 ± 0.0 148 ± 0.0 169 ± 0.1

19 52 60

4.4 ± 1.1 nff 2.0 ± 0.36 nff 0.87 ± 0.23

3.8 ± 0.28 nff 660 ± 76 nff 1060 ± 230

Abbreviations: na, not available; nff, no film formation. bDetermined from the second heating scan. cCalculated according to eq 1. F

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

■

Article

CONCLUSIONS We have successfully synthesized a tough and thermoresistant fully degradable poly(L-lactide)-based copolymer, solely from renewable resources. As predicted, these material properties are a consequence of the chemical structure of the comonomer, εDL, which induces both flexibility and thermoresilience to the polymeric chain. Also, the difference in reactivity and miscibility of the monomers during copolymerization enabled tailoring of material microstructures through the appropriate addition strategy. With sequential addition of L-lactide to the growing εDL chain, a triblock copolymer was synthesized and verified with 13C NMR and GPC, whereas the copolymer synthesized through simultaneous addition exhibited a random blocky structure that originates from the difference in the reactivities of the two monomers. The immiscible nature of the εDL and LLA regions was visualized with AFM, where hardness differences of the blocks showed phase separation down to the 20 nm scale. The thermal stability increased with increasing εDL composition in the copolymer. The copolymerization of εDL with LLA yielded a highly tough material with strain-at-break values more than 250 times greater than that of pure PLLA. These properties highlight the immense promise of the renewable monomer εDL. The εDL homopolymer synthesized using either Sn(Oct)2 or TBD as the catalyst yielded a completely amorphous, oily, and viscous polymer with a Tg value of −53 °C. The polymerization in bulk proceeded in a controlled manner with a linear relationship between conversion and molecular weight and with a low PDI. The rate constant of the polymerization at 150 °C was approximately 15 times greater than that at 110 °C, which resulted in 96% conversion in 8 h with a low PDI. The sevenmembered lactone ring of εDL has a higher ceiling temperature than δDL and a lower polymerization rate, following the trend of its unsubstituted analogues, ε-caprolactone and δ-valerolactone. The substitution of εDL at the ε-position decreases the polymerization rate; therefore, a high ceiling temperature is required to make use of conventional catalysis.

stability. This lower stability is supported by the parallels between the weight loss and the PLLA content of the polymer blocks and affirms the triblock structure of the copolymers. In contrast, the random copolymer exhibited one-step degradation. If the homoblocks in the random copolymer were sufficiently long, a multistep decomposition should have been observed in which the PLLA-rich regions degraded before the PεDL-rich blocks. Although 13C NMR shows the existence of reasonably long homoblocks of LLA and εDL, they are not sufficiently long to impart a multistep decomposition pathway in the copolymer. The 50/50 composition of the random copolymer should also yield a decomposition temperature between that of the two homopolymers. However, the overall decomposition temperature is closer to that of PεDL; hence, the εDL segments increase the LLA segment degradation temperature. The immiscibility of the blocks was further confirmed by the observation that the triblock copolymer exhibited two glass transitions: one in the temperature region that corresponds to PεDL and a second in the range that corresponds to PLLA. Interestingly, for the random copolymer, two glass transition temperatures were also found. One Tg at −27 °C corresponds to the εDL-rich phase, whereas another at 44 °C corresponds to the LLA-rich phase. Apparently, while the LLA- and εDL-rich segments determined by 13C NMR are not sufficiently long to affect the decomposition behavior, they are of sufficient length to affect the thermal transitions. The immiscibility of PεDL and PLLA results in an increase in the degree of crystallinity of the triblock copolymers (Table 3). The most probable explanation is the flexibility of εDL, which yields a much more flexible polymer chain and facilitates crystallization. In addition, the amorphous block of PεDL can also act as a filler in the amorphous phase in the material. The pure PεDL is a completely amorphous polymer with a Tg of approximately −53 °C, as determined by DSC (Table 3). Both PLLA40−PεDL20−PLLA40 and P(εDL50-r-LLA50) exhibited very high strain-at-break values in comparison to that of pure PLLA (Table 3). For the triblock copolymer, the strain-atbreak increased more than 250-fold compared to that of the LLA homopolymer, which shows the immense influence of the addition of the flexible central block of PεDL. The immiscible central block facilitated the formation of tie chains through “shielding”, a term meaning that the flanking PLLA segments are shielded from each other by the immiscible nature of the components and are forced into separate domains upon crystallization. Similar results have been reported for triblock copolymers of LLA with DXO or poly(but-2-ene-1,4-diyl malonate) (PBM) as central amorphous blocks, but not to the same extent.11,16 The large increase in strain-at-break of the random copolymer as compared to that of neat PLLA was more unexpected. However, the somewhat blocky structure determined by 13C NMR clearly results in the hard domains observed by AFM, and these domains have the ability to crystallize (DSC). These crystalline PLLA-rich domains act as physical cross-links and result in increased strain-at-break. The pure PεDL and PLLA20−PεDL60−PLLA20 lacked sufficient mechanical performance to form films. In summary, the copolymers of LLA and εDL exhibited higher degradation temperatures and greater toughness than pure PLLA. Also, by varying the addition sequence and block lengths, these properties might be tailored toward specific applications.

■

ASSOCIATED CONTENT

S Supporting Information *

Copolymerization kinetics and additional 13C NMR and copolymer 1H NMR spectra. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +46-8-790 82 74. Fax: +46-8-20 84 77. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge the European Research Council (ERC), ERC Advanced Grant, PARADIGM (Grant Agreement No. 246776), for the financial support of this work. We gratefully acknowledge Robertus Nugroho for assistance with the AFM measurements.

■

REFERENCES

(1) Pfaltzgraff, L. A.; De bruyn, M.; Cooper, E. C.; Budarin, V.; Clark, J. H. Green Chem. 2013, 15, 307−314.

G

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

(36) Sivalingam, G.; Karthik, R.; Madras, G. J. Anal. Appl. Pyrolysis 2003, 70, 631−647.

(2) Romero-Guido, C.; Belo, I.; Ta, T. M. N.; Cao-Hoang, L.; Alchihab, M.; Gomes, N.; Thonart, P.; Teixeira, J. A.; Destain, J.; Waché, Y. Appl. Microbiol. Biotechnol. 2011, 89, 535−547. (3) Undin, J.; Plikk, P.; Finne-Wistrand, A.; Albertsson, A.-C. J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 4965−4973. (4) Albertsson, A.-C.; Varma, I. Degradable Aliphatic Polyesters; Springer: Berlin, Heidelberg, 2002; Vol. 157. (5) Ajioka, M.; Suizu, H.; Higuchi, C.; Kashima, T. Polym. Degrad. Stab. 1998, 59, 137−143. (6) Duda, A.; Penczek, S. Polymers from Renewable Resources; American Chemical Society: Washington, DC, 2001; Vol. 764. (7) Kim, S. H.; Han, Y.-K.; Ahn, K.-D.; Kim, Y. H.; Chang, T. Makromol. Chem. 1993, 194, 3229−3236. (8) Stridsberg, K.; Albertsson, A.-C. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 1774−1784. (9) Odelius, K.; Albertsson, A.-C. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 1249−1264. (10) Löfgren, A.; Renstad, R.; Albertsson, A. C. J. Appl. Polym. Sci. 1995, 55, 1589−1600. (11) Målberg, S.; Plikk, P.; Finne-Wistrand, A.; Albertsson, A.-C. Chem. Mater. 2010, 22, 3009−3014. (12) Honeker, C. C.; Thomas, E. L. Chem. Mater. 1996, 8, 1702− 1714. (13) Andronova, N.; Albertsson, A.-C. Biomacromolecules 2006, 7, 1489−1495. (14) Zhang, Z.; Grijpma, D. W.; Feijen, J. Macromol. Chem. Phys. 2004, 205, 867−875. (15) Mathisen, T.; Albertsson, A. C. Macromolecules 1989, 22, 3838− 3842. (16) Ryner, M.; Albertsson, A.-C. Biomacromolecules 2002, 3, 601− 608. (17) Tsuji, H.; Ikada, Y. J. Appl. Polym. Sci. 1996, 60, 2367−2375. (18) Martello, M. T.; Hillmyer, M. A. Macromolecules 2011, 44, 8537−8545. (19) Martello, M. T.; Burns, A.; Hillmyer, M. ACS Macro Lett. 2012, 1, 131−135. (20) Endrizzi, A.; Pagot, Y.; Clainche, A. L.; Nicaud, J.-M.; Belin, J.M. Crit. Rev. Biotechnol. 1996, 16, 301−329. (21) Chuang, H.-J.; Chen, H.-L.; Huang, B.-H.; Tsai, T.-E.; Huang, P.-L.; Liao, T.-T.; Lin, C.-C. J. Polym. Sci., Part A: Polym. Chem. 2013, 51, 1185−1196. (22) Pitt, G. G.; Gratzl, M. M.; Kimmel, G. L.; Surles, J.; Sohindler, A. Biomaterials 1981, 2, 215−220. (23) Lecomte, P.; Jérôme, C. Synthetic Biodegradable Polymers; Springer: Berlin, Heidelberg, 2012; Vol. 245. (24) Save, M.; Schappacher, M.; Soum, A. Macromol. Chem. Phys. 2002, 203, 889−899. (25) Fischer, E. W.; Sterzel, H.; Wegner, G. Kolloid-Z. Z. Polym. 1973, 251, 980−990. (26) Kricheldorf, H. R.; Eggerstedt, S. Macromol. Chem. Phys. 1998, 199, 283−290. (27) Pratt, R. C.; Lohmeijer, B. G. G.; Long, D. A.; Waymouth, R. M.; Hedrick, J. L. J. Am. Chem. Soc. 2006, 128, 4556−4557. (28) Duda, A.; Kowalski, A.; Penczek, S.; Uyama, H.; Kobayashi, S. Macromolecules 2002, 35, 4266−4270. (29) Florczak, M.; Duda, A. Angew. Chem., Int. Ed. 2008, 47, 9088− 9091. (30) Kricheldorf, H. R.; Jonté, J. M.; Berl, M. Makromol. Chem. 1985, 12, 25−38. (31) Odelius, K.; Plikk, P.; Albertsson, A.-C. Biomacromolecules 2005, 6, 2718−2725. (32) McLean, R. S.; Sauer, B. B. Macromolecules 1997, 30, 8314− 8317. (33) Armin, K.; Robert, M.; Georg, K. J. Chem. Phys. 2004, 120, 1105−1116. (34) Södergård, A.; Stolt, M. Prog. Polym. Sci. 2002, 27, 1123−1163. (35) Persenaire, O.; Alexandre, M.; Degée, P.; Dubois, P. Biomacromolecules 2001, 2, 288−294. H

dx.doi.org/10.1021/bm400733e | Biomacromolecules XXXX, XXX, XXX−XXX