Article pubs.acs.org/JPCB
Multistep Kinetic Behavior in the Thermal Degradation of Poly(L‑Lactic Acid): A Physico-Geometrical Kinetic Interpretation Masahiro Yoshikawa,† Yuri Goshi,† Shuto Yamada,‡ and Nobuyoshi Koga*,† †
Chemistry Laboratory, Department of Science Education, Graduate School of Education, Hiroshima University, 1-1-1 Kagamiyama, Higashi-Hiroshima 729-8524, Japan ‡ Department of Applied Chemistry, National Defense Academy of Japan, 1-10-20 Hashirimizu, Yokosuka 239-8686, Japan ABSTRACT: A physico-geometrical kinetic interpretation of the thermal degradation of poly(L-lactic acid) (PLLA) is described based on the results of a kinetic study using thermogravimetry (TG) and the microscopic observation of the reaction process. From the physico-geometrical viewpoint, the reaction process is separated into two different stages characterized by a surface reaction of the molten PLLA in the initial reaction stage followed by continuous bubble formation and disappearance in the established reaction stage. The generally reported trend of variation in the apparent activation energy as the reaction advances is explained by the partial overlapping of these two reaction stages. The kinetic rate data obtained using TG were kinetically separated into those for the respective reaction stages by optimizing the kinetic parameters. The significance of the kinetic results is discussed in terms of the physico-geometrical characteristics of the reaction. Such systematic kinetic analyses demonstrate the importance of considering the physicogeometrical perspective when interpreting the kinetic results for the thermal degradation of polymers.
1. INTRODUCTION The idealized kinetic equation for a single step reaction has been applied to thermoanalytical curves of a range of chemical and physical processes of materials:1,2 ⎛ E ⎞ dα = A exp⎜ − a ⎟f (α) ⎝ RT ⎠ dt
molecular level, such as a series of random scission mechanisms.8 For example, the thermal degradation of poly(lactic acid) has been extensively studied because of its biodegradable properties. The reaction occurs in the molten state with the evolution of ethanal, 2-propenoic acid, 3,6dimethyl-1,4-dioxane-2,5-dione, and other compounds as gaseous products, from which the chemical reaction mechanisms, including random chain scission via a cis-elimination, cyclic rupture via intramolecular transesterification, and others have been deduced.9−13 Thus, in many kinetic studies, the random scission model has been applied as the kinetic model function.12,14−17 For a wide range of practical purposes, these studies are being further extended to the thermal degradation of polymer−inorganic composites, polymer blends, and so on.18−20 Although such a kinetic approach to the thermal degradation of polymers is promising, a successful kinetic study requires the coordination between the idealized kinetic equation, the actual reaction process, and the experimentally resolved kinetic curves obtained via thermal analysis; however, this is seldom discussed in the kinetic studies for the thermal degradation of polymers. This study focuses on the physico-geometrical characteristics of the thermal degradation of polymers as exemplified by the thermal degradation of poly(L-lactic acid) (PLLA). By observing the reaction process using an optical microscope, the physico-geometrical characteristics of the reaction process
(1)
where α, A, Ea, and f(α) are the fractional conversion, the Arrhenius preexponential factor, the apparent activation energy, and the kinetic model function that describe the physicochemical or physico-geometrical mechanism of the reaction, respectively. Although there have been many examples of successful kinetic characterization using thermal analysis, it is well recognized that discrepancies between the idealized kinetic equation, actual reaction process, and the experimentally resolved kinetic curves obtained via thermal analysis typically confuse the interpretation of the kinetic results with respect to the physical chemistry and chemical reaction mechanism.3 The thermal degradation of polymers is one of the most studied processes using thermoanalytical kinetics methodology. Many import kinetic calculation methods have been proposed in the past, as exemplified by the thermal degradation of polymers.4−7 The chemical reaction mechanism has also been revealed using thermoanalytical techniques involving evolved gas analysis. Thus, the thermal degradation of polymers is a chemical process for which interpreting the kinetic results in connection with the chemical reaction mechanism may be possible. For the correlation of the kinetic rate behavior and the chemical reaction mechanism, the kinetic model functions have also been derived from chemical reaction mechanisms at the © 2014 American Chemical Society
Received: July 20, 2014 Revised: August 29, 2014 Published: September 3, 2014 11397
dx.doi.org/10.1021/jp507247x | J. Phys. Chem. B 2014, 118, 11397−11405
The Journal of Physical Chemistry B
Article
cm3 min−1). Changes in the microscopic views were recorded with the reaction time and sample temperature data in movies.
occurring during the kinetic measurements were revealed. On the basis of the microscopic evidence for the physicogeometrical reaction behaviors, systematic kinetic analyses were then performed to demonstrate the physico-geometrical kinetic interpretation and the significance of the kinetic results.
3. RESULTS AND DISCUSSION 3.1. Characterization of Thermal Behavior. Figure 1 shows the TG−DTA and DSC curves for the PLLA samples in
2. EXPERIMENTAL SECTION 2.1. Samples. Commercially available PLLA samples with molecular weights of 325 000−460 000 and >700 000 were purchased from Polyscience Inc. The PLLA samples were purified by liquid−liquid extraction for removing possible contaminant of tin ions. PLLA solution (4% in weight) was obtained after dissolving each sample into dichloromethane (Wako Chem.). Using a separatory funnel, the liquid−liquid extraction was performed by adding aqueous hydrochloric acid (3% in weight) to the PLLA solution.9,21 The liquid−liquid extraction was repeated three times. After separating the dichloromethane layer, the PLLA was precipitated by adding methanol. The precipitate was washed with methanol and dried in a vacuum desiccator at room temperature for 24 h. The PLLA samples with molecular weights of 325 000−460 000 and >700 000 were termed PLLA3 and PLLA7, respectively. The thermal behavior of the samples (initial mass m0 = 3.0 mg weighed in an aluminum pan, 5 mm diameter × 2.5 mm height) were characterized via thermogravimetry−differential thermal analysis (TG−DTA; DTG-50, Shimadzu) with heating at a rate of β = 5 K min−1 in flowing N2 (80 cm3 min−1). Differential scanning calorimetry (DSC) curves were obtained for the samples (m0 = 10.0 mg weighed in an aluminum pan, 5 mm diameter × 2.5 mm height, and crimped with an aluminum lid) at β = 5 K min−1 in flowing N2 (50 cm3 min−1). The degradation products were identified via pyrolysis−gas chromatography/mass spectrometry (GC/MS) using an instrument with a single-shot pyrolyzer (PY-2020D, Frontier Lab) interfaced to a GC/MS (GCMS-QP2010; Shimadzu; GC column = UA5-30M-0.25F; MS ionization potential = 70 V; emission current = 60 μA), in which approximately 20 μg of a sample was heated at 773 K for 0.1 s in flowing He at a rate of 50 cm3 min−1. 2.2. Measurement of Kinetic Data. The mass loss process for the thermal degradation of the purified PLLA samples (flake, m0 = 3.00 ± 0.02 mg weighed in an aluminum pan, 5 mm diameter × 2.5 mm height) was traced using the above DTG-50 instrument in flowing N2 (80 cm3 min−1) under isothermal, linear nonisothermal, and modulated temperature conditions. The isothermal mass loss curves for the thermal degradation were recorded at different temperatures in the range of 548−578 K for both the purified PLLA samples. The mass loss curves under linear nonisothermal conditions were recorded at different β (0.5−5 K min−1). By applying the triangular wave of temperature modulations to the above isothermal measurements,22,23 the mass loss curves under modulated temperature conditions were also recorded using the temperature programs with basal constant temperatures at 563 or 568 K, an amplitude of modulation of 30 K, and a period of 10−30 min. 2.3. Microscopic Observation. The thermal degradation reaction process for the PLLA samples were observed under an optical microscope (BH2, Olympus) during the heating of each sample in a heating stage with an infrared image furnace (MSTPS, Yonekura). For the microscopic observations, 3.0 mg of each sample was weighed in a platinum cell (5 mm diameter × 2.5 mm height) and heated at β = 5 K min−1 in flowing N2 (80
Figure 1. Thermoanalytical characterization of the PLLA samples: (a) TG−DTA curves for the as-received PLLA samples, (b) TG−DTA curves for the purified PLLA samples, and (c) DSC curves for the purified PLLA samples.
flowing N2. The TG curves for the two as-received PLLA samples exhibited different mass loss behaviors (Figure 1a), although nearly 100% of mass loss due to the thermal degradation was observed for both the samples. The thermal degradation process of the as-received PLLA7 involved a multistep mass loss behavior that initiated at a temperature lower by approximately 50 K than that of the as-received PLLA3, which exhibited a smooth single-step mass loss. Effect of the residual tin ion added during the PLLA synthesis has been reported as the possible cause of the multistep thermal degradation behavior.9,21 In contrast, the thermal degradation of both the purified PLLA samples was comparable and consisted of a smooth mass loss curve with a single DTA endothermic peak (Figure 1b), with the mass loss initiated at a temperature higher by approximately 100 K than the complete melting temperature. Nearly 100% of the sample mass was lost during the thermal degradation process, although char 11398
dx.doi.org/10.1021/jp507247x | J. Phys. Chem. B 2014, 118, 11397−11405
The Journal of Physical Chemistry B
Article
accounting for ≤0.5 mass% of the sample was occasionally observed. The DSC curves of the purified samples were also very similar (Figure 1c), and the glass transition was observed at 330 K (middle point) followed by a detectable exothermic effect attributed to the cold crystallization of the amorphous phase in the temperature range of 340−385 K. The samples melted over a wide temperature range of 405−465 K, and the endothermic peaks were characterized by practically identical values for the extrapolated onset temperature (Teo = 447 ± 1 K), peak top temperature (Tp = 456 ± 1 K), and enthalpy change of melting (ΔmH = 49 ± 1 J g−1). The pyrograms of the purified samples were also identical as shown in Figure 2. The
Figure 3. Mass loss traces for the thermal degradation of the purified PLLA samples (m0 = 3.0 mg) under isothermal conditions at different temperatures in flowing N2 (80 cm3 min−1): (a) PLLA3 and (b) PLLA7.
Figure 2. Pyrograms of the purified PLLA samples.
major peaks are attributed to ethanal, 2-propenoic acid, and 3,6dimethyl-1,4-dioxane-2,5-dione, which is in agreement with the previous reports.9−13 Therefore, the purified PLLA samples used in this study exhibited practically the same thermal behaviors as those reported previously. 3.2. Formal Kinetic Analysis. Figures 3 and 4 show the mass loss traces for the thermal degradation of the purified PLLA samples under isothermal conditions at different temperatures and linear nonisothermal conditions at different β, respectively. Irrespective of the samples and heating conditions, smooth curves indicating a single mass loss process are observed during the thermal degradation process. The separate sharp peaks in the derivative mass loss curves obtained for the isothermal measurements at relatively high temperature were observed during linear heating to reach a programmed constant temperature and are due to the initiation of the reaction prior to reaching the constant temperature. Because of the observed systematic changes of the mass loss curves as a function of the reaction temperature under isothermal conditions and of β under linear nonisothermal conditions, the kinetic equation for the ideal single-step reaction, eq 1, was applied to universally analyze all the mass loss curves for each purified PLLA sample.24,25 Figure 5 shows an isoconversional kinetic approach based on eq 1 as applied universally to the kinetic data for the thermal degradation process under isothermal and linear nonisothermal conditions. Plots of ln(dα/dt) versus T−1 at a selected fractional reaction α among the series of kinetic data, known as Friedman plot,5 exhibit acceptable linear relations, as exemplified by the thermal degradation of the purified PLLA3 in Figure 5a. The trend in the variation of the apparent Ea values as a function of α was also similar for the two purified samples (Figure 5b). Although a large standard deviation of the slope of the Friedman plot was observed during the initial reaction stage, the Ea value tended to increase from approximately 130 to 175 kJ mol−1 when α
