J. Phys. Chem. B 2006, 110, 14279-14282
14279
Crystallization Behaviors of n-Nonadecane in Confined Space: Observation of Metastable Phase Induced by Surface Freezing Baoquan Xie,†,§ Haifeng Shi,† Shichun Jiang,‡ Ying Zhao,† Charles C. Han,† Duanfu Xu,† and Dujin Wang*,† Key Laboratory of Engineering Plastics, Joint Laboratory of Polymer Science and Materials, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100080, China, State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, China, and Graduate School of Chinese Academy of Sciences, Beijing 100080, China ReceiVed: May 25, 2006; In Final Form: June 7, 2006
Crystallization and phase transition behaviors of n-nonadecane in microcapsules with a diameter of about 5 µm were studied with the combination of differential scanning calorimetry (DSC) and synchrotron radiation X-ray diffraction (XRD). As evident from the DSC measurement, a surface freezing monolayer, which is formed in the microcapsules before the bulk crystallization, induces a novel metastable rotator phase (RII), which has not been reported anywhere else. We argue that the existence of the surface freezing monolayer decreases the nucleating potential barrier of the RII phase and induces its appearance, while the lower free energy in the confined geometry turns the transient RII phase to a “long-lived” metastable phase.
1. Introduction Normal alkanes [CnH2n+2, n-alkanes], consisting of linear chains of saturated hydrocarbons as the most fundamental organic series, can provide well-defined model systems for studying the complex crystallization behaviors of polymer materials, surfactants, lipids, and so forth.1 It is generally recognized that the condensed state of any molecules containing alkane moieties will be strongly influenced by the packing mode of n-alkanes. Thus, a detailed investigation of the alkane phase behavior is likely to aid in understanding the crystalline phase transition and ultimate properties of many polymer materials, especially polyolefins. Being such a fundamental and important system, the crystallization behaviors of n-alkanes have been extensively studied and a plethora of phases occurring between the isotropic liquid and full-crystallization states have been identified.2-5 The bulk n-alkanes exhibit two unique features of phase transitions, the first of which is the so-called “rotator phase”, exhibiting longrange order in the molecular axis orientation and the centerof-mass position but lacking long-range order in the rotational degree of freedom of the molecules around their long axis. Odd and even n-alkanes show different crystalline forms of the rotator phase. The face-centered orthorhombic modification called RI occurs in odd alkanes C11 through C25, while the rhombohedral modification with a hexagonal subcell called RII occurs in even alkanes C22 through C26.6-8 The second unique feature is the surface freezing phenomenon occurring for chain lengths ranging from n ) 15 to n ) 50, which has been widely investigated by X-ray reflectivity, grazing incidence X-ray diffraction, surface tension measurements, and molecular-dynamics simulations.9-13 * Author to whom correspondence should be addressed. E-mail:
[email protected]. † Key Laboratory of Engineering Plastics, Joint Laboratory of Polymer Science and Materials, Institute of Chemistry, Chinese Academy of Sciences. ‡ State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences. § Graduate School of Chinese Academy of Sciences.
A surface crystalline monolayer is formed at the temperature of up to ∼3 °C above the bulk crystallization temperature, and it is stacked into a planar hexagonal phase. This surface crystallization phenomenon is intimately related to the formation of the bulk rotator phase.14 Although great achievements have been obtained for the investigation of pure n-alkanes or mixed n-alkanes in the bulk phase, the phase behavior of n-alkanes in a confined system is more attractive, for example, encapsulated systems including emulsion, vesicle, micelle, and foam have been used to alter the crystallization kinetics and crystal morphology of n-alkanes from the bulk state.15-17 A generally accepted conclusion is that nucleation rate, crystal growth rate, and crystal morphology are remarkably influenced by droplet-droplet interaction, postnucleation growth, and interfacial membrane structures. Some interesting phenomena have been observed; for example, a new crystal phase of n-alkanes has been reported in oil-in-water emulsion droplets,18 and a rotator phase at the oil/water interface has been found to play a precursor role for bulk crystallization.19 These investigations enlighten a new strategy to design and understand the specific phase behaviors of n-alkanes in a soft microencapsulated environment provided by the thermodynamically unstable emulsions, foams, and so forth. Until now, however, less work has been done on the crystallization behavior of n-alkanes within a solid environment, for example, in microcapsules.20 The hard shell of microcapsules can afford a stable crystallization environment, in which the nucleation and crystal growth of n-alkanes are considered to be homogeneous and easily investigated; therefore, it is more attractive to explore the crystallization of n-alkanes in a confined space of microcapsules. In this paper, we prepared n-alkane-containing microcapsules with controllable size and morphology and investigated the crystallization behaviors of microencapsulated n-alkanes, aiming to further understand the complexity of phase stability and phase transition of n-alkanes. We presented new DSC evidence for the surface freezing in microencapsulated n-nonadecane, and
10.1021/jp063201j CCC: $33.50 © 2006 American Chemical Society Published on Web 07/04/2006
14280 J. Phys. Chem. B, Vol. 110, No. 29, 2006
Figure 1. SEM micrographs of the microcapsules containing nnonadecane prepared by in-situ copolymerization of melamine and formaldehyde, with sodium dodecyl sulfate (SDS) as emulsifier. n-Nonadecane is encapsulated inside the microcapsules, which content is calculated to be 28 wt % (DSC).
more interestingly, we found that the hexagonal surface crystalline monolayer induces the formation of a new bulk hexagonal rotator phase (RII), and to our best knowledge, this phenomenon has not been reported elsewhere for the n-nonadecane system. 2. Experimental Section n-Nonadecane was purchased from the Aldrich Company (purity > 99%) and used as received. Using melamineformaldehyde resin as shell material and n-nonadecane as core material, microcapsules were prepared by insitu polymerization.21 This polymerization method and subsequent sample treatment provided us with nearly monodispersed and high heat-resistant microcapsules, in which the crystallization and melting of n-nonadecane have been investigated. The particle size and surface morphology of the microcapsules were examined on a Hitachi S-4300 scanning electron microscope (SEM), fitted with a field emission source and operated at an accelerating voltage of 15 kV. DSC measurements were performed on a Mettler DSC822e calorimeter. Different cooling/ heating rates (0.1, 0.5, 1, and 2 °C/min) were adopted to evaluate the thermodynamic and kinetic properties of the phase transitions of both microencapsulated and pure n-nonadecane. Synchrotron radiation experiments were carried out at HASYLAB of DESY in Hamburg, Germany (beam line A2, λ ) 0.15 nm). There, WAXD and SAXS patterns can be measured simultaneously. The beam size at the sample is about 2 mm × 3 mm, and the scattering angles range between 0.1° and 50°. The samples with thickness of about 1 mm were enclosed in aluminum foil to keep them melted. The samples were first heated from 0 to 40 °C and kept for 3 min, and then they were cooled to 0 °C. The heating and cooling rates were all 0.5 °C/ min. The scattering angles were calibrated by standard PET and a rat-tendon-tail sample for WAXD and SAXS, respectively. 3. Results and Discussions The particle diameter of the microcapsules prepared in this work can be easily controlled from about 200 nm to 10 µm by adjusting the weight ratio of n-nonadecane to shell material, emulsifier concentration, as well as the stirring rate. The microcapsules with a diameter of about 5 µm have a smooth surface and narrow size distribution (Figure 1) and were used in this paper as research objects. Starting from a wall and core material of the same density,22 we can obtain the relationship between the radius of the core material and the ratio of the weight of the core material to that of the total: r1 )
Xie et al.
Figure 2. DSC traces of n-nonadecane during the heating and cooling processes: (a) pure n-nonadecane; (b) microencapsulated n-nonadecane. Specimens were heated at a rate of 1 °C/min from 0 to 50 °C, and then cooled to 0 °C at the same rate, followed by heating again to 50 °C. The first cooling and second heating thermograms were recorded. The experiments are reproducible.
[Wc/(WW + Wc)]1/3r2, where r1 and r2 correspond to the radius of the core and the microcapsule, and Wc and Ww represent the weight of the core and wall layer, respectively. With Wc/(Ww + Wc) ) 0.28 (DSC) and r2 ) 2.5 µm, r1 is calculated to be 0.65 µm. n-Nonadecane has 18 C-C single bonds with an average length of 1.25 Å in the all-trans conformational chain, so the volume percentage of surface monolayer molecules can be calculated approximately as follows: 18 × 1.25 Å × 4πr12/ (πr13 × 4/3) ≈ 0.01, where r1 takes the unit of Å. Such a result means that about 1% of the alkane molecules stand on the coreshell interface of the microcapsules, and consequently they exert a significant influence on the bulk crystallization behavior. The crystallization and melting behaviors of pure and microencapsulated n-nonadecane were first studied by the DSC method. It is a common sense that n-alkanes exhibit a rotator phase below the melting temperature. Odd n-alkanes (n e 21) exhibit only an orthorhombic rotator phase (RI) with molecules untilted with respect to the layers.6 Herein it was observed that with temperature changing, pure n-nonadecane shows two phase transitions between the isotropic liquid phase (L) and the stable orthorhombic phase (Figure 2a). In the cooling process, the sample becomes trapped in the face-centered orthorhombic RI, which converts to the stable crystal phase at 20.3 °C with 0.4 °C supercooling. The thermal behavior of microencapsulated n-nonadecane, however, is quite different. Four exothermic peaks appear in the cooling process, and three endothermic peaks appear in the heating process (Figure 2b). Above the bulk freezing temperature by 3 °C, a small sharp exothermic peak with normal enthalpy (∆Hs) of 0.11 J/g emerges during the cooling run, corresponding to the surface freezing of n-alkanes. In the heating process, the same small peak was also observed. This small peak is assigned to a thermodynamic first-order phase transition, taking the similar result of n-hexadecane in emulsified droplets as a reference.19 Previously, surface freezing of n-alkanes has been detected with the combination of surface tension and surface-grazing incidence diffraction (GID),12 whereas the direct measurement of the surface freezing of n-alkanes in hard microcapsules by thermal calorimetry is reported for the first time. According to Sirota and coauthors,8,12 the surface monolayer of n-alkanes tends to pack into a twodimensional hexagonal crystal, subsequently inducing the formation of the bulk rotator phase. As mentioned above, four peaks appear during the cooling process of the microencapsulated n-nonadecane. The big exothermic peak at 31 °C with negligible supercooling corresponds to bulk crystallization, the normal enthalpy (∆Hc) of which is 31.2 J/g, lower than that of the melt peak (∆Hm )
Crystallization Behaviors of n-Nonadecane
Figure 3. DSC traces of microencapsulated n-nonadecane at different cooling rates. Except the cooling rate, all the other measuring conditions are the same as described in Figure 2. The inserted figure on the right upper corner is the magnification of the surface freezing peaks at different cooling rates. The schematic illustration of the molecular stacking mode is considered to be a rotator II (RII) phase of n-nonadecane, as the reason stated below (also see XRD results in Figure 4).
39.4 J/g). The enthalpy difference may be attributed to the emergence of the new broad peak at 27.6 °C. The peak at 16 °C is assigned to the transition from RI to the stable orthorhombic phase, with a larger supercooling of 3.8 °C compared to that of pure n-nonadecane. We note that the enthalpy sum of the new broad peak (∆HR*) and the crystallization peak (∆Hc) is approximately equal to the normal melting enthalpy (∆Hm). This suggests that by slow cooling, the microencapsulated n-nonadecane is first trapped into a metastable phase from the isotropic liquid rather than directly into the RI phase, as is the case for pure n-nonadecane. With temperature further decreasing, this new metastable phase converts to the RI phase at ∼27.6 °C and finally turns into orthorhombic, the most stable crystal phase for n-nonadecane. To evaluate the thermodynamic and kinetic properties of the phase transitions of the microencapsulated n-nonadecane, DSC measurements were performed at different cooling rates. As can be seen from Figure 3, four exothermic peaks were observed, which correspond to the surface freezing and three transitions: isotropic liquid to new metastable phase, new metastable phase to RI phase, and RI phase to orthorhombic crystal, respectively. The cooling rate changing from 2 to 0.1 °C/min has little influence on the normal enthalpy of both surface freezing
J. Phys. Chem. B, Vol. 110, No. 29, 2006 14281 (∆Hs ) 0.11 J/g) and the transition from the RI phase to stable orthorhombic (∆HRI ) 10 J/g). The onset temperatures of the transitions both from isotropic liquid to the new metastable phase and from the new metastable phase to the RI phase are not perturbed by the cooling rate. These thermal behaviors indicate that the microencapsulated n-nonadecane is trapped into a new thermodynamically stable phase from the isotropic liquid with cooling. As the surface monolayer has been found to act as the precursor for the bulk rotator phase,18 the appearance of the new metastable phase is most likely induced by the twodimensional hexagonal crystal on the liquid surface of microencapsulated n-nonadecane, and it can be considered to be a bulk rotator II (RII) phase. To further test this speculation, XRD experiments were performed. Figure 4 shows the variation of different scattering vectors with decreasing temperature, which was measured with a synchrotron X-ray diffraction experiment. For pure nnonadecane (Figure 4a), above the melting temperature, the sample is in the isotropic liquid state, characterized by a single halo at q ) 1.38 Å-1. With sample cooling to 31 °C, the peaks of 200 at q ) 1.54 Å-1 and 110 at q ) 1.45 Å-1 lamellar reflections appear at the same time, indicating the appearance of the RI phase. The 200 lamellar reflection shows significant temperature dependence, but 110 does not. After cooling to 20 °C, the scattering vector of the 200 lamellar reflection increases to 1.62 Å-1 and stabilizes with temperature as that of 110 lamellar reflection, indicating that the phase transition from the RI to the orthorhombic crystal phase has been finished. As the microencapsulated n-nonadecane was cooled to 30 °C, however, only the 110 lamellar reflection appears, and its peak position q decreases from 1.49 to 1.45 Å-1 with further cooling to 27 °C (Figure 4b). The unique existence and large variation of the single peak position in the temperature region between 30 and 27 °C is indicative of a hexagonal arrangement, corresponding to a metastable RII phase. So, the synchrotron XRD results are in good agreement with DSC data in the cooling run, although the surface freezing peak was not detected due to the spherical character of the microcapsules. With further cooling to 27 °C, the 200 lamellar reflection also emerges. It is found that below 27 °C, the temperature dependence of the two lamellar reflections is the same as that of pure n-nonadecane. After the phase transition from RI to stable orthorhombic has been finished at about 16 °C, the peak positions of 110 and 200 lamellae reflections keep constant with temperature. Therefore, a conclusion can be drawn that although the surface freezing could not be detected by the XRD method, the RII phase does exist for the microencapsulated n-nonadecane.
Figure 4. The peak positions of the amorphous state (square) and 110 (rhombohedron) and 200 (triangle) crystal faces as a function of temperature for both (a) pure n-nonadecane and (b) microencapsulated n-nonadecane. The cooling rate was 0.5 °C/min. The interval between the data acquisition was 3 min.
14282 J. Phys. Chem. B, Vol. 110, No. 29, 2006
Xie et al. 4. Conclusions
Figure 5. Schematic illustration of the relative free energy versus temperature for (a) pure n-nonadecane and (b) microencapsulated n-nonadecane. The thin line (lower) indicates the heating process, and the broad line (upper) corresponds to the cooling process.
It was reported that n-alkanes (n g 23) exhibit the stable hexagonal rotator II phase with molecules untilted with respect to the layers.5 As for n-nonadecane (C19), the stable rotator phase is RI, which emerges during both heating and cooling processes, while the RII phase is not present in this compound. The transient RII phase does not emerge in pure n-nonadecane, because the energy barrier for its nucleation formation is too high (Figure 5a). In our microencapsulated n-nonadecane, the metastable RII phase has been detected by both DSC and XRD, which appears only upon cooling. It is speculated that the decrease of the nucleating barrier caused by the surface freezing induces the occurrence of the RII phase, while the lower free energy in the confined geometry provided by microcapsules makes the transient RII phase stay metastable for a longer time (Figure 5b). The above deduction can be supported by quantitative calculations. Because the radius of the inner surface of the microcapsules is as small as 0.65 µm, providing a large specific surface area, about 1% n-nonadecane molecules stand between the inner wall of the microcapsule and the bulk core material. To decrease the interfacial energy during the crystallization process, the surface n-alkane molecules first tend to stack into two-dimensional hexagonal arrays, which play the role of nucleating sites and subsequently induce the formation of the RII phase. In the case where there is a free surface in a semiinfinite system, the preference of the top layer to be RII has infinitely small effect on the free energy of the bulk to be in RII over RI. However, since in the microencapsulated space the “bulk” is finite, one can assume an interfacial energy cost to have the surface layer RII with the bulk RI. Thus, the free energy of the bulk being the RI with the RII at the surface is actually nonnegligibly higher than in the infinite size system. This will then reduce the driving free energy to transforming the RII to the RI and let it remain metastable for longer.
In summary, we have investigated the crystallization behaviors of microencapsulated n-nonadecane. The formation of a surface freezing monolayer was directly observed in the DSC measurements, which is attributed to the large specific area provided by the inner wall of the microcapsules. A novel metastable RII phase of microencapsulated n-nonadecane has been found in the cooling process. The formation of a surface freezing monolayer provides ideal nucleation sites and decreases surface free energy of the RII phase, turning this transient phase into a metastable one. Our present result will benefit the understanding of the role of confined space and surface freezing on the crystallization of polymer materials. Acknowledgment. We are thankful for the financial supports from the National Natural Science Foundation of China (No. 50573086) and Directional Project from Chinese Academy of Sciences (KJCX2-SW-H07). References and Notes (1) Small, D. M. The Physical Chemistry of Lipids; Plenum: New York, 1986. (2) Sirota, E. B.; Singer, D. M. J. Chem. Phys. 1993, 101, 10873. (3) Sirota, E. B.; King, H. E.; Hughes, G. J.; Wan, W. K. Phys. ReV. Lett. 1992, 68, 492. (4) Doucet, J.; Denicolo, I.; Craievich, A.; Collet, A. J. Chem. Phys. 1981, 75, 5125. (5) Dirand, M.; Bouroukba, M.; Chevallier, V.; Petitjean, D. J. Chem. Eng. Data 2002, 47, 115. (6) Sirota, E. B.; Herhold, A. B. Science 1999, 283, 529. (7) Sirota, E. B.; King, H. E.; Singer, D. M.; Shao, H. H. J. Chem. Phys. 1993, 98, 5809. (8) Sirota, E. B. Langmuir 1998, 14, 3133. (9) Wu, X. Z., et al. Science 1993, 261, 1018. (10) Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Ocko, B. M.; Deutsch, M. Phys. ReV. Lett. 1993, 70, 958. (11) Wu, X. Z.; Ocko, B. M.; Tang, H.; Sirota, E. B.; Sinha, S. K.; Deutsch, M. Phys. ReV. Lett. 1995, 75, 1332. (12) Ocko, B. M.; Wu, X. Z.; Sirota, E. B.; Sinha, S. K.; Gang, O.; Deutsch, M. Phys. ReV. E 1997, 55, 3164. (13) Prasad, S.; Dhinojwala, A. Phys. ReV. Lett. 2005, 95, 117801. (14) Sloutskin, E.; Sirota, E. B.; Kraack, H.; Ocko, B. M.; Deutsch, M. Phys. ReV. E 2001, 64, 1063. (15) Herhold, A. B.; Ertas, D.; Levine, A. J.; King, H. E., Jr. Phys. ReV. E 1999, 59, 6946. (16) Kraack, H. E.; Sirota, B.; Deutsch, M. J. Chem. Phys. 2000, 112, 6873. (17) Herhold, A. B.; King, H. E., Jr.; Sirota, E. B. J. Chem. Phys. 2002, 116, 9036. (18) Ueno, S.; Hamada, Y.; Sato, K. Cryst. Growth Des. 2003, 3, 935. (19) Shinohara, Y.; Kawasaki, N.; Ueno, N. S.; Kobayashi, I.; Nakajima, M.; Amemiya, Y. Phys. ReV. Lett. 2005, 94, 97801. (20) Fan, Y. F.; Zhang, X. X.; Wang, X. C.; Li, J.; Zhu, Q. B. Thermochim. Acta 2003, 413, 1. (21) Sun, G.; Zhang, Z. J. Microencapsul. 2001, 18, 593. (22) Sliwka, W. Angew. Chem., Int. Ed. Engl. 1975, 14, 539.
