Article pubs.acs.org/JPCB
Interplay between the Relaxation of the Glass of Random L/D‑Lactide Copolymers and Homogeneous Crystal Nucleation: Evidence for Segregation of Chain Defects René Androsch*,† and Christoph Schick‡ †
Center of Engineering Sciences, Martin Luther University Halle-Wittenberg, 06099 Halle/Saale, Germany Institute of Physics, University of Rostock, Wismarsche Str. 43−45, 18051 Rostock, Germany
‡
ABSTRACT: Random L-isomer rich copolymers of poly(lactic acid) containing up to 4% D-isomer co-units have been cooled from the molten state to obtain glasses free of crystals and homogeneous crystal nuclei. The kinetics of enthalpy relaxation and the formation of homogeneous crystal nuclei have then been analyzed using fast scanning chip calorimetry. It has been found that the relaxation of the glass toward the structure/enthalpy of the supercooled liquid state is independent of the presence of D-isomer co-units in the chain. Formation of homogeneous crystal nuclei in the glassy state requires the completion of the relaxation of the glass. However, nucleation is increasingly delayed in the random copolymers with increasing D-isomer chain defect concentration. The data show that the slower formation of homogeneous crystal nuclei in random L/D-lactide copolymers, compared to the homopolymer, is not caused by different chain-segment mobility in the glassy state but by the segregation of chain defects in this early stage of the crystallization process.
■
below Tg,14,15 enthalpy relaxation of the glass may then be followed by formation of homogeneous crystal nuclei and even crystal growth. The sequence of enthalpy relaxation preceding formation of homogeneous crystal nuclei has been predicted and confirmed early for inorganic Li2O−2 SiO2 glasses16 but recently also for polymers including poly(ε-caprolactone) (PCL),17 polyamide 6 (PA 6),18,19 isotactic poly(butene-1) (iPB-1),20 or isotactic polystyrene (iPS).21 It was suggested that the cooperative rearrangements responsible for the enthalpy relaxation below Tg and occurring on a length scale of a few nanometers overturn the formation of overcritical nuclei which result from sporadic fluctuations on a comparable or even smaller length scale. If the structure of the supercooled liquid state is achieved, then the driving force for the large-scale cooperative rearrangements vanishes and overcritical nuclei on a shorter length scale may survive.22 Evidence for formation of crystal nuclei in glassy PLLA has been provided recently, by analysis of the kinetics of cold-crystallization,23−26 employing Tammann’s two-stage crystal nuclei development method;27 even formation of ordered structures detectable by infrared spectroscopy and endothermic melting near Tg have been reported.28,29 The research presented here is about the establishment of the link between the kinetics of enthalpy relaxation and the onset of homogeneous crystal nucleation in the glass of PLLA and random L-isomer rich copolymers, containing a low amount of D-isomer co-units. In-depth analyses of the effect
INTRODUCTION Fast cooling of the melt of crystallizable polymers to below the glass transition temperature Tg may suppress both crystallization and formation of homogeneous crystal nuclei, leading then to a complete vitrification of the melt, and the formation of a glass free of any ordered domains/clusters. During annealing the glass, its specific volume and enthalpy decrease/ relax toward that of the liquid state at identical temperature,1,2 with the densification of the amorphous glass often leading to a characteristic change of ultimate properties, justifying intense research about the kinetics of enthalpy relaxation. This is in particular true for poly(L-lactic acid) (PLLA), which exhibits a glass transition temperature of about 60 °C, implying that relaxation of the glass at ambient temperatures occurs at a time scale interfering with its application as a material. Furthermore, PLLA is a rather slowly crystallizing polymer; that is, vitrification of the melt without prior crystallization occurs already on cooling at rather low rates of a few tens of Kelvin per minute, enhancing glass-relaxation effects on properties due to the absence of crystals. As such, intense research has been performed to describe the kinetics of enthalpy relaxation of PLLA, including its effect on properties. There have been previous quantitative analyses on the effects of annealing temperature, cooling rate, presence of D-isomer co-units in the PLLA chain, of the semicrystalline morphology, as well as of reinforcement and blending. Regarding enthalpy-relaxationinduced changes in properties, emphasis was placed to the analysis of the mechanical behavior, diffusion properties, or the biodegradation behavior.3−13 Since formation of crystal nuclei and crystal growth of polymers is assumed to cease only at a temperature 30−50 K © XXXX American Chemical Society
Received: March 23, 2016 Revised: April 24, 2016
A
DOI: 10.1021/acs.jpcb.6b03022 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
because it is necessary to vitrify the melt of all samples of different D-isomer content in the chain at well-defined and reproducible conditions of fast cooling, ensuring absence of both crystallization and formation of homogeneous crystal nuclei. These conditions have been evaluated recently and are reported elsewhere.38,42 Furthermore, FSC, with its instrumental time constant of the order of magnitude of milliseconds, allows conduction of isothermal annealing experiments for time periods less than a second, which, depending on the annealing temperature, may be needed to describe the kinetics of changes of the structure of glasses. Finally, FSC is required for application of Tammann’s two-stage crystal nuclei development method which implies formation of nuclei at high supercooling of the melt, or in the glass, and subsequent isothermal growth at higher temperature, utilizing the largely different temperatures of maximum rate of primary crystal nucleation and crystal growth.27 Besides the adjustment of well-defined states of amorphous structure free of crystals and nuclei at begin of the relaxation and nucleation experiments, FSC is then needed to ensure transfer of the nuclei-containing system to the temperature of growth such that formation of additional nuclei or even crystal growth during heating is avoided.43,44 FSC measurements were performed using a powercompensation Mettler-Toledo Flash DSC 1 connected to a Huber intracooler TC100, ensuring a constant sample-support temperature of −90 °C. Throughout the operation, the calorimeter was purged with nitrogen gas at a flow rate of 35 mL min−1. The FSC sensor was conditioned and temperaturecorrected according to the instrument operating instructions, and then the heatable area of the sample calorimeter was covered with a thin layer of highly viscous Wacker silicon oil AK 60 000, in order to improve the thermal contact between sensor and sample. Subsequently, the samples with a thickness and lateral dimension of 15 and 50−100 μm, respectively, were placed on the sensor and slowly heated to allow melting and establishment of a good thermal contact to the sensor. The sample mass was of the order of magnitude of 70−170 ng which can roughly be estimated by comparison of the measured heat-capacity increment at the glass transition of a fully amorphous sample in units of J K−1 with the mass-specific heatcapacity increment in units of J g−1 K−1, available in the literature.45 Precise information about the sample mass, required for quantitative estimation of the enthalpies of relaxation, were gained by an adjustment of experimental absolute heat capacities of fully liquid and solid PLLA, that is, of heat-capacity data collected at temperatures above the melting temperature and below the glass transition temperature, respectively, with specific heat-capacity data available in the literature.45
of the presence of D-isomer co-units in the PLLA chain on the crystallization behavior revealed that such chain defects cause a decrease of the maximum achievable crystallinity as well as a slowing of the crystallization process.30−33 It was furthermore reported that in such random L/D-lactide copolymers, the equilibrium melting temperature of crystals decreases, however, without an unequivocal opinion available whether the D-isomer co-units are incorporated in the crystalline phase, or rejected from crystallization.34−36 Importantly, in the context of the present study, it was also found that the rate of primary nuclei formation is similarly reduced as the crystal-growth rate, despite the largely different length scale of required transport processes due to the different size of crystals and nuclei.37,38 From this result, it was concluded that the D-isomer chain defects hinder nuclei formation already at this early stage of the crystallization process by their required segregation, however, assuming that these defects do not affect the chain dynamics in the amorphous state, as indicated by the unchanged glass transition temperature. In continuation of our prior research about crystallization of PLLA at high supercooling of the melt, and the nucleation kinetics at temperatures near and even lower than Tg, it is the goal of the present study to analyze and correlate the kinetics of enthalpy relaxation and crystal nucleation of PLLA, with emphasis placed on the evaluation of the effect of the D-isomer concentration in the chain. The results are considered as providing valuable insights about the mechanism behind the recently observed delay of formation of crystal nuclei in the PLLA glass in the case of the presence of chain defects. In other words, it is attempted to obtain information about whether the decrease of the nucleation rate is caused by a slower relaxation of chain segments or by the existence of an additional barrier to form a stable nucleus, which would be true in the case of a required filtering of defect-free chain segments to be included into the nucleus.
■
EXPERIMENTAL SECTION Material. The study was performed using PLLA extrusion grades from Sulzer Chemtec Ltd. (Switzerland) with a meltflow index of 6−7 g (10 min)−1 (190 °C, 2.16 kp) and a residual monomer content of less than 0.3%.39 The D-isomer content of the samples was lower than 0.3%, and 2 and 4%, determined by chiral gas chromatography.39 The mass-average molar mass and the polydispersity of around 120 kDa and 2, respectively, were identical for all samples and were determined by gel permeation chromatography as described elsewhere.38 Precise information about the molecular characteristics of the investigated samples are presented in Table 1. Instrumentation. The analysis of the kinetics of enthalpy relaxation and formation of homogeneous nuclei in glassy PLLA has been performed using fast scanning chip calorimetry (FSC).40,41 Application of this analysis tool is urgently required
■
RESULTS AND DISCUSSION Figure 1 shows the FSC temperature−time profile for analysis of the kinetics of enthalpy relaxation and homogeneous crystal nuclei formation in glassy samples of PLLA of different Disomer concentration. The samples were heated to a temperature higher than the equilibrium melting temperature Tm, 0 to obtain a relaxed melt and then cooled at a rate of 1000 K s−1 to below Tg; cooling the melt at this high rate suppresses both crystallization and formation of homogeneous nuclei.42 The glass free of crystals and homogeneous crystal nuclei was then annealed for different periods of time up to 10 000 s to gain information about changes of structure, that is, about its densification via enthalpy relaxation and about formation of
Table 1. List of Samples of the Present Work, Including Information about the D-Isomer Content (% D) in the Polymer Chain, the Mass-Average Molar Mass (Mw) and Polydispersity (PD)38,39 % 1 2 3
D
