J. Phys. Chem. B 2007, 111, 2473-2476
2473
Metastable Crystalline Lamella of Cetylpyridinium Chloride in the Krafft Transition Shigeo Sasaki* Department of Chemistry, Faculty of Sciences Faculty of Science, Kyushu UniVersity, 33 Hakozaki, Higashi ku, Fukuoka 812, Japan ReceiVed: October 9, 2006; In Final Form: January 11, 2007
The metastable crystalline lamella was found in the Krafft transition of aqueous cetylpyridinium chloride (CPC) solutions. Temperature-dependent profiles of small-angle X-ray scattering (SAXS) for the CPC solution incubated for 10 min at 5 °C exhibited the metastable lamella structure with a lattice spacing of dL ) 3.19 nm at temperatures below 12 °C and the stable lamella structure with a lattice spacing of dL ) 2.85 nm at temperatures between 12 and 19 °C. The former lamella structure, however, was not observed in the temperature scanning SAXS profiles of the CPC solution incubated for 24 h at 5 °C. The latter lamella structure observed in the SAXS profile mentioned above started melting at 18 °C. The electric conductance change of the CPC solution with a time elapsed after dropping the temperature showed the existence of the temperature-dependent induction period in the Krafft transition, indicating high activation energies for the transition. In the differential scanning calorimetry measurements over temperatures ranging from 5 to 30 °C, a single endothermic enthalpy peak at 19 °C observed for the CPC solution incubated at 5 °C for a longer period than 6 h was split into double peaks at 14 and 19 °C when the same solution was incubated at 5 °C for a shorter period than 6 h. The observed calorimetric behavior is explained by the existence of the metastable crystalline state that grows faster and melts at a lower melting temperature than the stable crystalline state.
Knowledge of the kinetics of phase transformation, leading to well-defined nanostructure, is fundamental in order to obtain materials with tailored properties1 and understand the mechanism of frequently phase transforming biological systems. A change in the physicochemical condition often induces the phase transformation, which starts with the kinetically favored metastable states.2 Supercooled (metastable) liquid water,3 which has not been fully understood yet, is frequently observed. Metastable conformations are often observed in the folding processes of the denatured proteins in the aqueous solutions.4 A metastable perforated lamellar structure has been found in the forming process of a gyroid phase in an aqueous solution of a diblock amphiphilic copolymer.5 Differential scanning calorimetry (DSC) curves of annealed materials are sometimes different from those of nonannealed ones.6 The metastable state appearing in the nonannealed materials could disappear with annealing for a time longer than the lifetime of the metastable state. The lifetime of the metastable state, however, changes with the physicochemical conditions of the materials. The physics of the metastable state in aqueous solutions has not been very developed so far due to lack of metastable states with lifetimes long enough to be investigated. Recently we found the longlived metastable state in the Krafft transition process of cetylpyridinium chloride (CPC) in the aqueous solution through measurements of small-angle X-ray scattering (SAXS), electric conductance, and calorimetry. A temperature scanning SAXS profile for the CPC solution incubated at 5 °C for 10 min revealed the existence of a transformation from one lamella structure to the other lamella at a temperature below the Krafft transition temperature. The change in the lattice spacing indicates the transformation from the ordinary bilayer lamella to the interdigitated lamella. The * To whom correspondence should be addressed. Phone/Fax: +81-92642-2609. E-mail:
[email protected].
SAXS profile for the CPC solution incubated at 5 °C for 24 h, however, exhibited the existence of only the interdigitated lamella structure in the solution at temperatures below the Krafft transition temperature. After the long incubation time of the CPC solution at 5 °C, the ordinary lamella structure melts and the interdigitated lamella structure forms. This indicates that the former is metastable and the latter stable. It was found in the present experiments that a time lag exists between the electric conductance drop and the temperature drop in the Krafft transition process of CPC. The time lag is an induction period of lamella structure formation. It was also found in the present DSC measurement that a single endothermic enthalpy peak observed for the CPC solution incubated at 5 °C for a longer time than 6 h was split into double peaks when the same solution was incubated at 5 °C for a shorter time than 6 h. The unusual calorimetry behavior of the aqueous CPC solution, which strongly depends on the incubation period, tInc, and the incubation temperature, TInc, can be explained by the existence of the metastable state. The CPC purchased from Tokyokasei Co. Ltd. was recrystallized from methanol/acetone and used. The SAXS experiments were carried out for the 100 mM CPC solution with a SAXS spectrometer of BL45XU (RIKEN Beamline I) installed at Spring8 of the Japan Synchrotron Radiation Research Institute, Hyogo, Japan. The observed scattering vectors, q, ranged from 0.07 to 5 nm-1. About 30 mg of the CPC solution was encapsulated in the aluminum cell for the SAXS experiment, the optical thickness of which was 2 mm. The cell windows were made of polyimide films. The cell was mounted on a temperature control stage (LTS-350 Japan High Tec. Inc.) slightly modified for the SAXS measurements. The time-sliced SAXS measurements were made with heating (rate ) 0.3 °C/ min). The quenching effect on the transition was examined by measuring the conductance as a function of time elapsed after
10.1021/jp066614t CCC: $37.00 © 2007 American Chemical Society Published on Web 02/16/2007
2474 J. Phys. Chem. B, Vol. 111, No. 10, 2007
Sasaki
Figure 1. SAXS profile for the 100 mM CPC solution incubated for about 24 h at 5 °C. The profiles are described as a function of temperature.
Figure 2. SAXS profile for the 100 mM CPC solution incubated for about 10 min at 5 °C. The profiles are described as a function of temperature.
dropping the temperature from 19.5 °C to a given temperature below the Krafft transition temperature. Measurements of the electric conductance were made with a conductivity meter (CM60V TOA Electrochemical Measurement Instruments Inc.) for the gently stirring CPC solutions in the cell inserted into the thermo bath, the temperature of which was controlled within 0.1 °C. A temperature change of the cell was accomplished with changing the bath. DSC measurements were carried out at a heating rate of 0.5 °C/min for a 100 mM CPC aqueous solution using a DSC calorimeter (DSC120 Seiko Inc.). About 60 mg of the CPC solution was encapsulated in an aluminum cell at a room temperature, incubated for a given tInc at a given TInc, and measured. It is well known that the Krafft transition is induced by melting of the hydrated solid of the CPC molecule7, which is mainly stabilized by van der Waals attraction between the alkyl chains. SAXS measurements while scanning temperature can reveal what nanostructure changes are made with the phase transformation. A SAXS 3-D profile for the CPC solution incubated for about 24 h at 5 °C is shown in Figure 1. A sharp scattering peak at q ) 2.20 nm-1 observed at a temperature below 18 °C transitionally disappears, and a broad peak at q ) 0.59 nm-1 is observed at a temperature above 18 °C. The sharp peak is due to the crystal-like and long-range ordered structure, and the broad peak is due to the liquid-like and short-range ordered structure. The peak intensity fluctuation seen in Figure 1 reflects the fluctuation of the crystal volume fraction in the optical volume. The structure of the hydrated solid is identified as a crystalline lamella with a lattice spacing dL ) 2.86 nm, which is indicated by the existence of a higher order peak at q ) 4.40 nm-1. It is noteworthy that dL is close to the stacking thickness of sheets in a crystal of cetyltrimethylammonium bromide (dL ) 2.6 nm), alkyl chains of which are interdigitated and arrayed with tilting to the direction parallel to the sheet.8 Figure 2 shows another SAXS 3-D profile for the CPC solution incubated for about 10 min at 5 °C after cooling it down from 25 °C at a rate of 5 °C/min. A sharp scattering peak at q ) 1.968 nm-1 starts diminishing at 12 °C and disappears at 15 °C, and another sharp peak at q ) 2.208 nm-1 starts growing at 12 °C, plateaus at 15 °C, and transitionally disappears at 19 °C. Higher order peaks at q ) 3.94 and 4.40 nm-1,
respectively, are also observed for the solutions at temperatures below 15 °C and between 12 and 19 °C, indicating that the structures of hydrated solids at lower and higher temperatures, respectively, are identified as crystalline lamellae with lattice spacings dL ) 3.19 and 2.85 nm. The former and latter lamellae, respectively, will be denoted as LM and LS hereafter. The LM is inferred a metastable structure from the fact that LM is not observed in the solution incubated for a long period at 5 °C as shown in Figure 1. It is remarkable that the observed thickness of LM is shorter than the bilayer thickness of CPC, two times the length of the stretched C16 molecule (about 4 nm). This indicates that the alkyl chains in the metastable lamella are also interdigitated and/or tilted. It should be pointed out that no gradual shift of the peak q value with temperature is observed. This indicates that shortening the lamella spacing does not induce transformation from the LM to the LS. Broad scattering peaks at q ) 0.488 and 0.611 nm-1 observed at a temperature above 19 °C in Figure 2 are intermicelle interference peaks, indicating the correlation lengths of pair distribution function of micelles ξC ) 12.8 and 10.3 nm. The broad peak (at q ) .59 nm-1) corresponding to ξC ) 10.6 nm is also observed in Figure 1 at a temperature above 19 °C for the solution incubated for a long period at 5 °C as described before. The uncertainty of ξC might reflect the large concentration fluctuation in the solution just after melting the LS. The degree of counterion dissociation of the surfactant molecules transitionally changes with the Krafft transition9 and can be evaluated by the electric conductance of the bulk solution, a decrease of which indicates the counterion binding to the surfactant molecule. Figure 3 shows the time evolution of the electric conductance of the 10 mM CPC solution when the temperature drops from 19.5 to 9.2, 9.8, 10.8, 11.6, or 12.9 °C. The monitored temperatures of the solution are very stabilized at a time longer than 200 s after changing the thermo bath as shown in Figure 3. It is noticeable that the drop of conductance lags long behind the temperature drop and that the conductance dropped once increases slightly with time elapsed. The degree of counterion binding to the surfactant molecule transitionally increases with the structural change from the micelle to the lamella at the Krafft transition. Accordingly, the conductance drops because of a transitional decrease in the
Metastable Crystalline Lamella
Figure 3. Time evolution of the electric conductivity (upper curves) of 10 mM CPC solutions in dropping the temperature (lower curves). The changing timing from the thermo bath at a temperature 19.5 °C to that at a destined temperature is t ) 0.
concentration of dissociated counterion.9 The time lag between the drops of temperature and conductance indicates the existence of the onset time or the activation barrier to overcome for formation of lamellae. It should be noticed that the onset or lag times increase with temperature. A small increase in the conductance at a time longer than 1000 s, shown in Figure 3, is caused by counterion dissociation from the lamella surface, which is induced by a decrease in the charge density of the surface. The results mentioned above can be explained by a structural change from the high charge density lamella LM to the low charge density lamella LS with the time elapsed on keeping the solution at temperatures below 10.8 °C. Therefore, it can be concluded that the LM is the ordinary tilt bilayer, which has a higher density of the head group on the layer surface than that of the interdigitated LS lamella. The entropy of unbinding counterion in the LS state can contribute more to decreasing the free energy than that in the LM state, since the free counterion concentration in the former state is higher than the latter state. The interdigitated lamella is more stable than the non-interdigitated lamella, which is consistent with the fact that the melting temperature of the LS, about 18 °C, is higher than that of the LM, about 12 °C. The nanostructure transformation indicated by the temperature-dependent SAXS profiles described above should induce the calorimetric transitions, since melting of the crystal is usually an endothermic process and crystallization is exothermic. Figure 4 shows the DSC curves of 100 mM CPC aqueous solutions, which were incubated at 5 or 14 °C for various periods. It is remarkable that the features of endothermic peaks are so different from each other and depend on the tInc at TInc ) 5 °C and that the total transition enthalpies calculated from integration areas of the curves are not very different from each other (3843 kJ/mol). A primary endothermic peak at 19 °C, which is observed for the solution incubated for 72 h at 14 °C (a curve 2 in Figure 4) or 24 h at 5 °C (a curve 3 in Figure 4), lowers with decreasing tInc. Another second endothermic peak at 14 °C is clearly observed for the solution incubated for 1 h at 5 °C and grows with shortening tInc (curve 4 in Figure 4). Parallel to the behavior mentioned above, the primary peak area decreases. It should be noticed that small endothermic trends between 10 and 14 °C are observed for the solutions incubated at 5 °C but not for the solution incubated at 14 °C. The transitional structure change at 18 °C as shown in Figure 1 is considered to induce the endothermic peak at 19 °C shown by curve 3 in Figure 4. The observed difference between the thermal and structural
J. Phys. Chem. B, Vol. 111, No. 10, 2007 2475
Figure 4. Endothermic DSC curves for the 100 mM CPC aqueous solutions incubated at 5 or 14 °C for various periods. The incubated times, temperatures, and total transition enthalpy are described in the figure. Vertical dotted lines A, B, and C, respectively, indicate temperatures of 10, 14, and 19 °C.
transition temperatures is due to melting relaxation of ordered stacking of alkyl chains, which gives the broad endothermic peak as shown by curve 3 in Figure 4. It should be mentioned that the small endothermic trend at temperatures between 10 and 14 °C seen in the curve 3 in Figure 4 is accompanied with no change in the structure over 1-100 nm as shown by Figure 1. It is inferred from the results shown in Figures 1 and 2 that melting temperatures of the metastable structure LM and the stable structure LS, respectively, are 12 and 18 °C. Thermal pretreatment of the DSC experiment shown by curve 5 in Figure 4 corresponds to that of the SAXS experiment shown in Figure 2. The nanostructure transformation indicated by the temperature-dependent SAXS profile shown in Figure 2 should be related to the calorimetric behavior exhibited by curve 5 in Figure 4. Melting a crystal is an endothermic process, and crystallization is exothermic. The endothermic enthalpy observed at temperatures between 14 and 19 °C can be explained by melting of the LM. The crystallization enthalpy of the LS is not observed as the exothermic peak in the DSC curve because the van der Waals interaction energy released in forming the LS is transiently transferred to other types of the energies, such as the electrostatic energy due to the decoupling of chloride and pyridinium ions and/or the hydrogen-bonding energy among water molecules intercalated in the LS. The endothermic enthalpy at temperatures between 19 and 22 °C shown by the curve 5 in Figure 4 might be a difference between the enthalpies in melting and forming the LS. At the low temperature where the van der Waals attraction exceeds the thermal kinetic energy of alkyl chains in the micelle, the alkyl chains are laterally stacked with each other to form the ordinary bilayer lamella, which is accompanied with counterion binding to the head groups of surfactant molecules. The onset time for changing from the micelle to the LM is shorter at lower temperature, as shown in Figure 3. This indicates that the activation energy for formation or nucleation of the LM increases with an increase in temperature as schematically illustrated in Figure 5. For interdigitating the oppositely directing alkyl chains, the stack of chains in the bilayer should be loosened. The activation energy for the loosening decreases with an increase of the defect portion of the stacking chain. The increasing temperature induces an increase of the defect portion and reduces the activation energy for interdigitating the alkyl
2476 J. Phys. Chem. B, Vol. 111, No. 10, 2007
Figure 5. Schematic figures of free energy map as a function of the characteristic length of structure (dL or a size of micelle) to illustrate the structural transition from the micelle, to LM, to LS at low and high temperatures. Free energy map as a function of the characteristic length of structure to illustrate the structural transition from the micelle, to the metastable bilayer lamella, to the melted lamella, to the stable interdigitated lamella.
chains. That is, the activation energy for changing from the LM to the LS decreases with increasing temperature as depicted in Figure 5. Thus, the changing period from the LM to the LS decreases with an increase of temperature as shown in Figure 3 in which the period for the conductance to increase gradually shortens with the change of incubation temperature from 9.2 to 10.8 °C. To obtain deeper insight into the nucleation and growth mechanisms of the LS, further kinetic experiments such as the time-sliced SAXS measurements at a constant temperature are needed. The high activation barriers among the micelle, the LM, and the LS as shown in Figure 5 require long periods for transforming from one to the other state. On cooling the solution to set the temperature below 15 °C, the LM forms first and transforms to the LS during the incubation time at this temperature. When the incubation time is longer than the period for transforming the structure from the LM to the LS, which is inferred about 4 h at 5 °C and about 24 h at 14 °C, the DSC curves with a
Sasaki single endothermic peak at 19 °C are observed as shown by curves 1, 2, and 3 in Figure 1. When the incubation time is shorter than the transforming period, part of the LM transforms to the LS and the other part of the LM remains in solution. In the heating process of this case, the SAXS experiment demonstrates that the existing LM melts while forming the LS at temperatures between 12 and 15 °C and the LS melts at 19 °C as shown Figure 2. The DSC experiment for the case described above shows, however, two endothermic peaks at 14 and 19 °C as shown by curves 4 and 5 in Figure 4. The former and latter, respectively, are the surplus endothermic enthalpies in melting of the LM and forming the LS and in forming and melting the LS during the heating process. The endothermic enthalpy in melting the LM is inferred to be almost the same as that in melting the LS from the fact that a sum of the endothermic enthalpies at low and high temperature is substantially invariant as shown in Figure 4. The ratio of the endothermic enthalpy at 14 °C to that at 19 °C is approximately proportional to the ratio of the LM amount to the LS amount in solution at the starting temperature of the DSC measurement. Acknowledgment. The SAXS experiments were performed at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute. References and Notes (1) Kitamura, M. Cryst. Growth Des. 2004, 4, 1153-1159. (2) Lebrum, N.; Foulon, M.; Gors, C.; Ferriol, M.; Cohen-Adad M. T. J. Cryst. Growth 1997, 160, 367-377. (3) Koop, T.; Luo, B.; Tsias, A.; Peter, T. Nature 2000, 406, 611614. (4) Religa, T. L.; Markson, J. M.; Mayor, U.; Freund, S. M. V.; Fersht, A. R. Nature 2005, 437, 1053-1056. (5) Hamley, I. W.; Castelletto, V.; Mykhaylyk, O. O.; Yang, Z.; May, R. P.; Lyakhova, K. S.; Sevink, G. J. A.; Zvelindovsky, A. V. Langmuir 2004, 20, 10785-10790. (6) Kodama, M.; Seki, S. AdV. Colloid Interface Sci. 1991, 35, 1-30. (7) Hato, M.; Shinoda, K. J. Phys. Chem. 1973, 77, 378-381. (8) Campanelli, A. R.; Scaramuzza, L. Acta Crystallogr. 1986, C42, 1380-1383. (9) Vautier-Gingo, C.; Bales, B. L. J. Phys. Chem. B 2003, 107, 53985403.
