11804
Langmuir 2007, 23, 11804-11811
Differential Water Sorption Studies on Kevlar 49 and As-Polymerized Poly(p-phenylene terephthalamide): Determination of Water Transport Properties Damian A. Mooney* and J. M. Don MacElroy School of Chemical & Bioprocess Engineering, Engineering and Materials Science Centre, UniVersity College Dublin, Belfield, Dublin 4, Ireland ReceiVed June 13, 2007. In Final Form: August 9, 2007 Water vapor sorption experiments have been conducted on Kevlar 49 at 30 °C over a range of water vapor pressures in 0-90% of saturation and on the as-polymerized form of the material at 30, 45, and 60 °C over a series of water vapor pressures of 0-60%, 0-25%, and 0-15%, respectively. For each of the differential steps in water vapor pressure, dynamic uptake curves were generated and analyzed according to a number of different mathematical models, including Fickian, Coaxial cylindrical, and intercalation models. The intercalation model was demonstrated to be the most successful model and considered two time-scales involved in the diffusion process, i.e., a penetrant-diffusive time-scale and a polymer-local-matrix-relaxation time-scale. The success of this model reinforces previously reported adsorption and desorption isotherms which suggested that water may penetrate into the surface layers of the polymer crystallite through a process known as intercalation.
1. Introduction The study of transport and diffusion in polymeric materials has important contributions in a number of areas, including membrane separation processes, microelectronic encapsulation devices, drug release, and materials design. Transport in materials is a function of the microscopic order that a diffusing species experiences, and therefore, differences in transport coefficients often provide an insight into the subtle differences within the same material. In the case of polymeric materials, the history that the material has been through is particularly important. This kind of information lends import to the study of transport, particularly in the light of the demands which are imposed upon new and emerging material technologies, where sophisticated materials are expected to operate in a predicable fashion under a variety of environmental and other operational conditions. Kevlar is DuPont’s trade-name for the commercial form of poly(p-phenylene terephthalamide) (PPTA), a rigid chain aromatic first synthesized by S. L. Kwolek, a DuPont research scientist, in 1965,1,2 and consists of alternating para-linked phenyl rings and amide linkages (see Figure 1). [*The microstructure and morphology of Kevlar is well documented and is summarized in the context of moisture sorption in a previous paper.3] The extraordinary mechanical and thermal properties exhibited by this material4 have resulted in its utilization in a number of high performance applications, particularly in the aviation industry where it is used in the construction of composite materials for aircraft components. For such an operational environment, materials must behave in a predictable manner under cyclic conditions of temperature and humidity, as well as dynamic loads, caused by changes in altitude. Extended exposure to such an environment necessarily requires a better understanding of the transport behavior of water under such conditions. * To whom all correspondence should be addressed. Phone +353 (0)1 716 1827. Fax: +353 (0)1 716 1177. E-mail:
[email protected]. (1) Kwolek, S. L. U.S. Patent 3,671,542, June 20, 1972. (2) Kwolek, S. L. U.S. Patent 3,819,587, June 25, 1974. (3) Mooney, D. M.; MacElroy, J. M. D. Chem. Eng. Sci. 2004, 59, 21592170. (4) Yang, H. H. KeVlar Aramid Fiber; John Wiley & Sons: Chichester, U.K., 1993.
Figure 1. Repeat structure of PPTA.
Early research into the sorption and diffusion of moisture in Kevlar aramid fibers was very limited in scope. Penn and Larsen,5 showed through integral sorption experiments that sorption was reversible, i.e., the adsorption-desorption cycle end-points were co-incident. However, the dynamics of adsorption were shown to be much more rapid than those of desorption. Morgan and Pruneda6 evaluated the activation energy for diffusion of moisture in Kevlar 49 to be 25 kJ mol-1 (compared to epoxides, 77 kJ mol-1) and concluded that this relatively low value could be attributed to diffusion occurring preferentially between the 60 nm fibrils of the spun fiber. The most extensive research to date on the sorption and diffusion of moisture in Kevlar was conducted by Fukuda and co-workers7-10 and Saijo et al.11 Diffusion measurements were made via gravimetric analysis using a helical quartz spring, with an optical reader to determine displacement from changes in weight. On samples of Kevlar 29, 49, and 149,7-9 a series of integral sorption steps ranging from 14% to 90% relative humidity were performed. Simple, short-time Fickian fits were found to be consistently too high at long times for both Kevlar 29 and 49 and consistently too low for Kevlar 149. A skin-core or coaxial cylinder model for the fiber was constructed9 based upon the observation of a skin-core morphology for Kevlar 29 and 49, and (5) Penn, L.; Larsen, G. J. App. Polym. Sci. 1979, 23, 59-73. (6) Morgan, R. J.; Pruneda, C. O. Polymer 1987, 28, 340-346. (7) Fukuda, M.; Ochi, M.; Miyagawa, M.; Kawai, H. Text. Res. J. 1991, 61, 668-680. (8) Fukuda, M.; Kawai, H. Polym. Eng. Sci. 1993, 34, 330-340. (9) Fukuda, M.; Kawai, H. Text. Res. J. 1993, 63, 185-193. (10) Fukuda, M. Polym. Eng. Sci. 1996, 36, 558-567. (11) Saijo, K.; Arimoto, O.; Hashimoto, T.; Fukuda, M.; Kawai, H. Polymer 1994, 35, 496-503.
10.1021/la7017538 CCC: $37.00 © 2007 American Chemical Society Published on Web 10/12/2007
Determination of Water Transport Properties
Langmuir, Vol. 23, No. 23, 2007 11805
for R2/R1 ) 1.2, which implies a skin thickness that is double to that observed experimentally,13,14 and the model’s requirement which suggests that Dcore/Dskin f 1 as P/PSat f 0. The ability of the skin-core model used by previous authors to fit moisture desorption is also unclear. This work, therefore, attempts to probe further the aforementioned observations by using a more rigorous differential moisture sorption apparatus, as opposed to previous integral sorption work, which limits the effects of concentration dependent diffusion coefficients, while providing an analysis of both moisture adsorption and desorption. This work also investigates the as-polymerized material which has not undergone spinning and, therefore, does not possess a skin-core morphology. It should therefore be possible to separate out the effects and instead focus on inherent transport properties of the polymer. More recent work by Fukuda10 studied moisture sorption and diffusion in PPTA films. Again, integral sorption experiments were conducted and the dynamic adsorption and desorption data were fitted to a model which considered the possibility that diffusion is affected by the adsorption of water at polymer amide groups (adsorption controlled). Reasonable success was achieved in fitting the data obtained. Figure 2. Schematic diagram of the grain boundary region with a definition of terms used in the intercalation model.
good fits were obtained by fixing the ratio of the fiber radius (R2) and the core radius (R1) to be equal to 1.2. Using Dskin,# and Dcore,# to denote the value of the diffusion coefficient in the skin and core, respectively, for Kevlar (where # equals to 29, 49, or 149), the magnitude of the diffusion coefficient of water in Kevlar (10-11 to 10-12 cm-2 s-1) is found to be considerably lower than that for nylon (10-8 cm-2 s-1) or for polyimide films (10-9 cm-2 s-1). Notably, all samples show a concentration-dependent diffusion coefficient, typical for hydrophilic systems, least marked for Dskin,49 and Dcore,149. In addition, the ratio Dcore/Dskin is greater than unity, except for Kevlar 149. These observations appear to be in agreement with the highly orientated skin structure of these fibers12,13 where the skin constitutes a barrier that hinders diffusion when compared to the morphology associated with its core (where diffusion takes place in the inter-crystalline and inter-fibril regions). These results also appear consistent with the fact that the additional heat treatment applied to Kevlar 49 enhances the alignment of fibrils within the core region, hindering water diffusion between fibrils). Patterns were observed to be different for Kevlar 149, consistent with the fact that the additional heat treatment this material undergoes leads to the loss of structure at the surface (reflected by the fact that the tensile strength of Kevlar 149 is less than both Kevlar 29 and 49), with increased crystallinity in the core region. The aforementioned studies also proposed that diffusion at low pressures is dictated by the adsorption of water to amide sites,9,10 whereas at higher values of coverage, the dynamics of the filling of microvoids becomes more important.10,11 In the case of Kevlar 149, an almost nonmonotonic behavior of the concentration dependence of Dcore,149 is seen. This is postulated to be a result of water clusters inhibiting diffusion. There are however a number of unresolved questions with regard to the proposed models, including the lack of a partition coefficient between skin and core regions (which would be expected for regions of different sorptive capacities), the value (12) Panar, M.; Avakian, P.; Blume, R. C.; Gardner, K. H.; Gierke, T. D.; Yang, H. H. J. Polym Sci. Polym. Phys. 1983, 21, 1955-1969. (13) Li, S. F. Y.; McGhie, J.; Tang, S. L. Polymer 1993, 34, 4573-4575.
2. Experimental Section 2.1. Equipment. For the purpose of studying the transport properties of water in Kevlar 49 and the as-polymerized material (AS-PPTA), a differential gravimetric sorption technique was performed using a specially constructed apparatus, with a Cahn D-200 microbalance at its heart and described elsewhere.3 Here, an equilibrated sample is exposed to a small (differential) change in penetrant pressure and the loss or gain in weight is monitored as a function of time until equilibrium. This differential step in penetrant pressure was achieved by using a control valve to admit vapor from a saturated environment (reservoir) to a very large control (buffer) volume to which the sample was exposed using a second control valve. During the course of this work, differential adsorption and desorption curves were measured for Kevlar 49 at 30 °C over a series of water vapor pressures from 0 to 90% of saturation, and on the as-polymerized form of the material at 30, 45, and 60 °C over a series of water vapor pressures of 0-60%, 0-25%, and 0-15%, respectively. From the dynamics of this uptake process, penetrant transport coefficients are determined using a series of mathematical models presented below, with the objective of gaining an insight into the mechanism of diffusion into these materials. 2.2. Materials. Both Kevlar 49 and the as-polymerized material were supplied by DuPont (U.K.) Ltd. Kevlar 49 samples were made up of short strands (∼130 mm) of 1667 dtex yarn. To remove any finishing oils present, the sample was washed three times in tetrachloroethylene and then dried in an oven at 110 °C for a period of 12 h. In the case of the as-polymerized (AS-PPTA) form, a size distribution analysis was first conducted in order to isolate a narrow range of particle sizes (+37 µm to 49 µm). Prior to being used, both types of sample were stored in a desiccator containing phosphorus pentoxide (PO5). The details of the sieving and other preparatory steps can be found elsewhere.3 2.3. Procedure. For each series of experiments (single sample, single temperature), samples are transferred from a desiccator (PO5 desiccant) to the Cahn D200 balance. Once equilibrated to the specified temperature, the system was evacuated at that temperature, and the sample dried under a continuous vacuum (
