Thermally Induced Phase Transitions and Morphological Changes in

3746

Langmuir 2004, 20, 3746-3758

Thermally Induced Phase Transitions and Morphological Changes in Organoclays M. Gelfer,† C. Burger,† A. Fadeev,‡ I. Sics,† B. Chu,† B. S. Hsiao,*,† A. Heintz,§ K. Kojo,§ S-L. Hsu,§ M. Si,| and M. Rafailovich| Chemistry Department, State University of New York at Stony Brook, Stony Brook, New York 11794-3400, Chemistry Department, Seton Hall University, South Orange, New Jersey 07009, Polymer Science and Engineering Department, University of Massachusetts at Amherst, Amherst, Massachusetts 01003, and Material Science and Engineering Department, State University of New York at Stony Brook, Stony Brook, New York 11794 Received July 25, 2003. In Final Form: February 18, 2004 Thermal transitions and morphological changes in Cloisite organoclays were investigated by differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), Fourier transform infrared (FTIR) spectroscopy, and in situ simultaneous small-angle X-ray scattering (SAXS) and wide-angle X-ray diffraction (WAXD) over the temperature range of 30-260 °C. On the basis of DSC and FTIR results, the surfactant component in organoclays was found to undergo a melting-like order-disorder transition between 35 and 50 °C. The transition temperatures of the DSC peaks (Ttr) in the organoclays varied slightly with the surfactant content; however, they were significantly lower than the melting temperature of the free surfactant (dimethyldihydrotallowammonium chloride; Tm ) 70 °C). FTIR results indicated that within the vicinity of Ttr, the gauche content increased significantly in the conformation of surfactant molecules, while WAXD results did not show any change in three-dimensional ordering. Multiple scattering peaks were observed in SAXS profiles. In the SAXS data acquired below Ttr, the second scattering peak was found to occur at an angle lower than twice that of the first peak position (i.e., nonequidistant scattering maxima). In the data acquired above Ttr, the second peak was found to shift toward the equidistant position (the most drastic shift was seen in the system with the highest surfactant content). Using a novel SAXS modeling technique, we suggest that the appearance of nonequidistant SAXS maxima could result from a bimodal layer thickness distribution of the organic layers in organoclays. The occurrence of the equidistant scattering profile above Ttr could be explained by the conversion of the bimodal distribution to the unimodal distribution, indicating a redistribution of the surfactant that is nonbounded to the clay surface. At temperatures above 190 °C, the scattering maxima gradually broadened and became nonequidistant again but having the second peak shifted toward a scattering angle higher than twice the first peak position. The changes in SAXS patterns above 190 °C could be attributed to the collapse of organic layers due to desorption and/or degradation of surfactant component, which was supported by the TGA data.

Introduction The intercalated organoclays formed by electrostatic complex formation between the layered silicates and the surfactants have long been an active research subject in the mineralogy community.1-7 In recent years, the subject of organoclays has further attracted a great deal of attention from the polymer community. Various organoclays have been synthesized and characterized with the goal to prepare new kinds of polymer composites.8-11 It * To whom correspondence should be addressed. Telephone: 631-632-7793. Fax: 631-632-6518. E-mail: [email protected]. sunysb.edu. † Chemistry Department, State University of New York at Stony Brook, Stony Brook. ‡ Chemistry Department, Seton Hall University. § Polymer Science and Engineering Department, University of Massachusetts at Amherst. | Material Science and Engineering Department, State University of New York at Stony Brook. (1) Walker, G. F. Clay Miner. 1967, 7, 129-143. (2) Johns, W. D.; Gupta, P. K. S. Am. Mineral. 1967, 52, 1706-1724. (3) Malik, W. U.; Srivastava, S. K.; Gupta, D. Clay Miner. 1972, 9, 369-382. (4) Lagaly, G.; Fitz, S.; Weiss, A. Clays Clay Miner. 1975, 23, 45-54. (5) Ijdo, W. L.; Pinnavaia, T. J. J. Solid State Chem. 1998, 139, 281289. (6) Lagaly, G. Solid State Ionics 1986, 22, 43-51. (7) Ijdo, W. L.; Pinnavaia, T. J. Chem. Mater. 1999, 11, 3227-3231. (8) Vaia, R. A.; Teukolsky, R. K.; Giannelis, E. P. Chem. Mater. 1994, 6, 1017-1022.

has been shown that the surfactant component can promote the exfoliation of layered silicates in the polymer matrix and form a “nanocomposite”, which would exhibit superior thermal stability and barrier and mechanical properties.12-15 However, the degree of clay exfoliation depends on the dispersion method used. As the melt mixing method represents the most practical way to compound nanocomposite samples,14,15 we are mainly concerned with thermal transition and phase behavior of organoclays in this study. This is because the melt mixing is often performed at elevated temperatures (above 150 °C), where thermally induced structural changes frequently occur in organoclays. This knowledge shall enable us to improve the control of clay dispersion in the polymer matrix. In this work, we have used several techniques, including differential scanning calorimetry (DSC), thermogravi(9) Xie, W.; Gao, Z. M.; Pan, W. P.; Hunter, D.; Singh, A.; Vaia, R. Chem. Mater. 2001, 13, 2979-2990. (10) Lee, J. W.; Lim, Y. T.; Park, O. O. Polym. Bull. 2000, 45, 191198. (11) Hackett, E.; Manias, E.; Giannelis, E. P. J. Chem. Phys. 1998, 108, 7410-7415. (12) Krishnamoorti, R.; Vaia, R. A.; Giannelis, E. P. Chem. Mater. 1996, 8, 1728-1734. (13) Vaia, R. A.; Giannelis, E. P. Macromolecules 1997, 30, 80008009. (14) LeBaron, P. C.; Wang, Z.; Pinnavaia, T. J. Appl. Clay Sci. 1999, 15, 11-29. (15) Alexandre, M.; Dubois, P. Mater. Sci. Eng. 2000, 28, 1-63.

10.1021/la035361h CCC: $27.50 © 2004 American Chemical Society Published on Web 04/02/2004

Morphological Changes in Organoclays

metric analysis (TGA), Fourier transform infrared (FTIR) spectroscopy, small-angle X-ray scattering (SAXS), and wide-angle X-ray diffraction (WAXD), to study the thermal transitions and corresponding morphological changes in a series of montmorillonite based organoclays with varying surfactant content. One major emphasis was made on the development of a reliable technique to interpret the SAXS data for the layered silicate systems. The large density difference between the mineral silicate and the organic component in organoclays has made X-ray scattering/ diffraction a natural technique to characterize the structure and morphology of organoclays and nanocomposite systems.1,2,8,9 It has been shown that the degree of delamination of layered silicates and the distribution of the interlayer distances in organosilicates can be determined from the SAXS data.16,17 This information is exceptionally useful for the analysis of nanocomposites because one can qualitatively evaluate the affinity between the polymer chains and the organoclay based on the changes in SAXS profiles from organoclays to composites.14,18 For example, the shift of the first scattering maximum toward a lower scattering angle can be attributed to the increase of the interlayer distance caused by the intercalation of organoclay stacks by polymer molecules, while the widening of the scattering peak and the decrease of the scattered intensity can be attributed to a beginning exfoliation of clay stacks.15 The fully exfoliated clay will not show any interlamellar peak but may show oscillations of the lamella particle scattering. The task, however, becomes much more complicated, if one wishes to determine the quantitative information about the state of intercalation and exfoliation in organoclays from the SAXS data. This is because there are many factors, including the defects in organic-layered silicate complexes and a heterogeneous distribution of interlayer periodicity, that can result in broadening and shifting of scattering peaks.12,17 Among these factors, the most dominating one is the heterogeneous interlayer thickness distribution (interstratification), which often has a polymodal characteristic. This effect has been pointed out by Reynolds16 and Brindley,17 where the positions of the scattering maxima in interstratified systems can significantly deviate from the equidistant ideal positions for perfectly periodic systems. Furthermore, the usually observed small lamellar stack heights can lead to a broadening of the peaks. The effects of surface defects in organic-layered silicate complexes and the breaking of clay stacks from the surfactant treatment, although relatively minor when compared with the effects of interstratification, can also result in the signal broadening.16,19,20 In this work, we have developed a novel approach for the quantitative interpretation of the SAXS data for the layered-silicate system, including organoclays and polymer-clay nanocomposites. The technique allowed us to account for the interlayer thickness distribution within the organoclay stacks as well as the stack size distribution. We assumed that the effects due to surface defects in organic-layered silicate complexes are negligible. The current state of understanding of the structure and morphology of organoclays formed via the electrostatic (16) Reynolds, R. C. Interstratified Clay Minerals; Brindley, G. W., Brown, G., Eds.; Mineralgical Society: London, 1980; pp 249-305. (17) Brindley, G. W. Order-Disorder in Clay Mineral Structures; Brindley, G. W., Brown, G., Eds.; Mineralgical Society: London, 1980; pp 125-197. (18) Giannelis, E. P.; Krishnamoorti, R.; Manias, E. Adv. Polym. Sci 1999, 138, 107-147. (19) Alexander, L. E. X-ray Diffraction Methods in Polymer Science; Wiley-Interscience: New York, 1969. (20) Hosemann, R.; Bagchi, S. N. Direct Analysis of Diffraction by Matter; North-Holland Publishing Co.: Amsterdam, 1962.

Langmuir, Vol. 20, No. 9, 2004 3747

complex formation between the organic surfactant component and clay (montmorillonite), in fact, is quite advanced. To fully appreciate the issues raised in this study and the efforts made to resolve them, a brief review of this subject is summarized as follows. Background on Organoclays Structural Heterogeneity in Montmorillonite Clays. The complex structure of montmorillonite-based organoclays can be partially attributed to the pronounced structural and energetic heterogeneity in pristine montmorillonite clays.21 On the basis of calorimetric data and adsorption measurements, it has been shown that clay minerals always contain several types of adsorption sites with different interaction energies.21,22 The energies associated with the basal and lateral surfaces in clay layers can differ greatly from each other and therefore result in different adsorption properties. The face energy alone on the basal surface may also vary at different locations. As a result, when montmorillonite clay is intercalated by small molecules such as water, the distribution of layer thickness can be quite nonuniform.21-23 If the long-chain cationic surfactants are used to complex with the clay surface, the nonuniformity in the distribution of layer thickness may be even more severe. The cationic surfactants can be adsorbed on the montmorillonite clay surface by two mechanisms, (1) cation exchange and (2) hydrophobic bonding.1,21,24 As a result, the amount of adsorbed surfactant may significantly exceed the ion exchange capacity.25-27 The excessive surfactant may be located either in the vicinity of the clay surface or between the adsorbed organic layers. Conformation of the hydrocarbon chains in the surfactant depends on the grafting density, the type of ion exchange process, and the thermal history utilized during the preparation of organoclays.1,2,4 FTIR and solid-state NMR measurements have confirmed that varying chain conformations with different trans-gauche ratios can coexist in the intercalated clay complexes.8,28-30 As the surfactant content approaches the overall ion exchange capacity of the clay, the hydrocarbon tails belonging to surfactant molecules adsorbed at opposing sides of the silicate gallery can interpenetrate.1,2 Obviously, the extent of the penetration will affect the thickness of the organic layers. As a result, a few types of distinct layers with different thicknesses may coexist in organoclays,24,31,32 which is the characteristic for the so-called interstratified systems.16,17,23 In this work, we have paid special attention to organoclays having relatively high content of a long-chain surfactant (21) Villieras, F.; Michot, L. J.; Cases, J. M.; Berend, I.; Bardot, F.; Francois, M.; Gerard, G.; Yvon, J. Static and dynamic studies of the energetic surface heterogeneity of clay minerals; Rudzinski, W. A. S. G. Z., Ed.; Elsevier: Amsterdam, 1997; Vol. 104, pp 573-623. (22) Cases, J. M.; Berend, I.; Besson, G.; Francois, M.; Uriot, J. P.; Thomas, F.; Poirier, J. E. Langmuir 1992, 11, 2730-2739. (23) Drits, V. A.; Tchoubar, C. X-ray Diffraction by Disordered Lamellar Structures; Theory and Application to Microdivided Silicates and Carbons; Springer-Verlag: New York, 1990. (24) Xu, S. H.; Boyd, S. A. Environ. Sci. Technol. 1995, 29, 30223028. (25) Hanley, H. J. M.; Muzny, C. D.; Butler, B. D. Langmuir 1997, 13, 5276-5282. (26) Patzko, A.; Dekany, I. Colloid Surf., A 1993, 71, 299-307. (27) Yui, T.; Yoshida, H.; Tachibana, H.; Tryk, D. A.; Inoue, H. Langmuir 2002, 18, 891-896. (28) Wang, L.-Q.; J. Liu; Exarhos, G. J.; Flanigan, K. Y.; Bordia, R. J. Phys. Chem. B 2000, 104, 2810-2816. (29) Kubies, D.; R.Jerome; Grandjean, J. Langmuir 2002, 18, 61596163. (30) Suga, K.; Rusling, F. Langmuir 1993, 9, 3649-3655. (31) Tahani, A.; Karroua, M.; Damme, H. V.; Levitz, P.; Bergaya, F. J. Colloid Interface Sci. 1999, 216, 242-249. (32) Xu, S. H.; Boyd, S. A. Langmuir 1995, 11, 2508-2514.

3748

Langmuir, Vol. 20, No. 9, 2004

(C16-C18), as these systems are often used for the preparation of polymer-clay nanocomposites by the meltblending process. Temperature Dependence of Morphology in Organoclays. Lagaly et al.4,33 investigated the complexes formed by montmorillonite-like bedellite and nontronite clays with n-tetradecylammonium and n-alkanols containing 8-22 CH2 groups. On the basis of the X-ray diffraction (XRD) data, they suggested that these compounds underwent a phase transition at high temperatures resulting in a steplike decrease in the basal spacing dL. They argued that the changes in the basal spacing during transition can be attributed to the formation of well-ordered gauche block structures rather than the simple melting-like order-disorder transition. In contrast, Brindley and Ray34 attributed the decrease in dL in the alkanol/calcium montmorillonite complex to a meltinglike order-disorder phase transition. Yui et al.27 and Okahata et al.35 proposed that the thermally induced transitions observed in layered silicate-surfactant complexes below 100 °C might be due to the changes from highly ordered pseudocrystalline structure of adsorbed surfactant bilayers to less ordered liquid crystal-like conformation. Lee at al.10 and Xie et al.9,36 investigated high-temperature behavior of organoclays. FTIR results from Lee et al.10 indicated that a higher gauche content in hydrocarbon tails of dimethyldihydrotallowammonium chloride present in thermally treated organoclay (Cloisite 20A) might be attributed to the surfactant loss at elevated temperatures. Xie et al.36 indicated that the onset of degradation might occur at a temperature as low as 130-150 °C. However, the maximum weight loss corresponding to the decomposition of surfactant usually occurred above 250 °C. Their combined TGA/mass spectrometry results suggested that the thermal degradation of “excessive” surfactant molecules residing between the surfactant layers adsorbed on the mineral surfaces might occur at the same or even higher temperatures than that of bound surfactant.9,36 This is because the degradation of bound surfactant can be facilitated by the proximity to the catalytically active alumosilicate sites. Experiment Samples and Preparation. Cloisite organoclays C6, C15, and C20 were obtained from the Southern Clay Company and used as received. The extracted organoclay C6extr was prepared by using the following procedures. The C6 organoclay was first immersed in boiling xylene in a Soxlet apparatus for 48 h. The recovered sample was subsequently vacuum-dried at 100 °C for 48 h. On the basis of the data provided by the Southern Clay Company, all organoclays contained the same mineral base (montmorillonite clay: Wyoming Cloisite) and an identical surfactant (dimethyldihydroditallowammonium chloride, DMDTA). DMDTA is a blend of surfactants prepared from natural products by Akzo Nobel. According to Akzo Nobel, the major component in this blend was dimethyldioctadecylammonium chloride (DMDOA), while minor components included (in the order of decreasing content) dimethyloctadecylhexadecylammonium chloride, dimethyldihexadecylammonium chloride, and a small ( C20), which is consistent with higher content of excess (adsorbed above CEC) surfactant. It is interesting to note that the rate of weight loss in the temperature range 400 °C < T < 500 °C is higher in nitrogen than that in air. This behavior may be related to the charring resulting from the oxidative dehydrogenation of the organic phase, promoted by the catalytic properties of the montmorillonite surface. The reaction can result in the formation of a thermally stable carbonaceous char from the organic component in organoclays.40,41 The purpose of the multiple-staged TGA experiment was to simulate the decomposition conditions used in the in situ SAXS study. It is interesting to note that in contrast to the ramp experiments, the weight loss in the multiplestaged TGA experiment is more pronounced in air than that in nitrogen (Figure 2c,d). We suggest that at lower temperatures used in temperature step measurements the oxidative charring processes are insignificant, so the oxidation of a surfactant yields volatile products resulting in the stronger weight loss in air. In both ramping and multiple-staged TGA thermograms (Figures 2c,d), the major loss of surfactant begins at 200 °C, indicating the escape of some free surfactant molecules. At 260 °C, all tested organoclays exhibit a weight loss accounting for 30-35 wt % of organic component (or 10-17 wt % total). This finding can be attributed to the escape of surfactant molecules due to the dissociation of electrostatic complex formation between surfactant molecules and clay surfaces and/or the thermal and oxidative degradation of surfactant molecules as demonstrated by Xie et al.36 Differential Scanning Calorimetry. DSC thermograms of the first and second heating scans for three organoclay samples (C6, C6-extract, and C20) are shown in Figure 3. All thermograms exhibit distinct endothermic peaks in the region of 45-55 °C, which represent meltinglike order-disorder transitions. The transitions, which are about 20 °C lower than the melting temperature of pure surfactant DMTA, are probably due to different mechanisms involving the conformation change as well as the dissociation and/or the thermal degradation of bound surfactant (with the clay surface) within the organoclay. This behavior has been extensively documented in the literature.4,33-36 The enthalpy change, ∆Ha, of the melting-like orderdisorder transitions increases with the surfactant content (C6 > C15 > C20 ∼ C6extr.). The values of ∆Ha for different organoclays and the surfactant in Figure 3 are listed in Table 2. It is interesting to see that the ∆Ha value obtained in the first heating scan is higher than that in the second scan (after being cooled from 180 to 5 °C). The ∆Ha value remains about constant during subsequent heatingcooling cycles over the same temperature range. As the TGA data did not exhibit appreciable surfactant loss in the range of 5-150 °C, the observed decrease in ∆Ha can be attributed to the irreversible conformation change or to the escape of free surfactant from the interlayer space to the space between clay tactoids after being heated and cooled (i.e., during the second heating run). Fourier Transform Infrared Spectroscopy. The FTIR technique is a useful tool to characterize the degree (40) Balogh, M.; Laszlo, P. Organic Chemistry Using Clays; SpringerVerlag: New York, 1993. (41) Gilman, J. W.; Jackson, C. L.; Morgan, A. B.; Harris, R.; Manias, E.; Giannelis, E. P.; Wuthenow, M.;Hilton, D.; Phillips, S. H. Chem. Mater. 2000, 12, 1866-1873.

3752


Gelfer et al.

Figure 2. (a) TGA thermograms during heating at 10 °C/min for different organoclays (C6, C15, and C20) in a nitrogen environment. (b) Comparison of TGA thermograms in air and in nitrogen for two organoclays (C6 and C20) (heating rate was 10 °C/min). (c) TGA thermograms collected using a multiple-staged temperature protocol in air. (d) TGA thermograms collected using a multiplestaged temperature protocol in nitrogen. (Multiple-staged temperature protocol: Multiple heating steps were taken from 30 to 260 °C at every 20 °C interval. The heating rate between each step was 30 °C/min, and the holding time during the isothermal step was 3 min. Upon reaching 260 °C and being held for 3 min, three cooling steps were also carried out at 200 °C, 160 °C, and 80 °C. The cooling rate between each step was also 30 °C/min.) Table 2. Enthalpy Changes during Melting Transitions for Several Cloisite Organoclays and Pure Surfactant material surfactant (dimethyldihydrotallowammonium chloride) organoclay C20 organoclay C6 extracted organoclay C15 organoclay C6

Figure 3. DSC thermograms for organoclays during heating at 10 °C/min.

of molecular order of the surfactant layers in organoclays. This is because the surfactant component consists of long alkyl chains, where the frequencies of CH2-stretching bands (2800-3000 cm-1) can reflect the state of conformational order in the adsorbed organic layers.42-46 For a (42) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559-3568. (43) Parikh, A. N.; Liedberg, B.; Atre, S. V.; Ho, M.; Allara, D. L. J. Phys. Chem. 1995, 99, 9996-10008. (44) Kojo, K.; Ge, S.; Takahara, A.; Kajiyama, T. Langmuir 1998, 14, 971-974.

∆H (J/g) first run

∆H (J/g) second run

155.0

N/A

14.9 18.61 22 32

8.1 10.1 16 17

completely disordered structure, such as liquid alkanes, the characteristic frequencies are 2924 cm-1 for CH2 symmetric stretching and 2856 cm-1 for CH2 asymmetric stretching. For well-ordered layers, such as crystalline paraffins, the characteristic frequencies are 2915-18 cm-1 for CH2 symmetric stretching and 2846-50 cm-1 for CH2 asymmetric stretching. The FTIR spectra between 2800 and 3000 cm-1 for the C6 sample at different temperatures are shown in Figure 4a. The temperature-dependent frequencies for CH2 symmetric stretching and asymmetric stretching in all organoclays are shown in parts a and b of Figure 5, respectively. In Figures 4 and 5, it is seen that the CH2 stretching frequencies are very close to those in liquid alkanes. This suggests that the organic phase in organoclays is rather loose and disordered. In Figure 5a, (45) Hair, M. L. IR spectroscopy in Surface Chemistry; Marcel Dekker: New York, 1967. (46) Snyder, R. G.; Strauss, H. L.; Elliger, C. A. J. Phys. Chem. 1982, 86, 5145-5150.


Langmuir, Vol. 20, No. 9, 2004 3753

Figure 4. Temperature-dependent FTIR spectra for C6 organoclay: (a) CH2 stretching region, (b) OH stretching (due to the adsorbed water).

the CH2 symmetric stretching frequency is found to increase in the order of C6 < C15 < C20. Even though C6 appears to exhibit the highest trans content which is indicative of higher molecular order in the organic layers, its structure is still significantly disordered. In parts a and b of Figure 5, the data indicate that the trans content in the organic layers decreases with increasing temperature (the corresponding frequencies become higher at high temperatures). The temperature dependence of the conformation changes will be discussed next. To further examine the temperature effect, the changes of two additional bands, CH2 bending (1400-1550 cm-1, Figure 5c) and CH2 rocking (700-750 cm-1, data not shown), were also characterized. All CH2 bands exhibit a relatively large change between 25 and 98 °C (Figures 4 and 5). For example, in symmetric and asymmetric CH2 stretching bands, large frequency shifts are found between 25 and 51 °C; in the CH2 bending band, the peak at 1468 cm-1 changes from “sharp” to “broad” between 24 and 98 °C; in the CH2 rocking band, the peak at 720 cm-1 also changes from “sharp” to “broad” between 24 and 73 °C. These data are consistent with DSC results, where endothermic transitions take place. In other words, these spectroscopic changes may be related to the melting-like order-disorder transitions. However, it is not clear whether the observed endotherms in DSC are due to a first-order transition or not, such as the melting of crystals. There are several reasons for this uncertainty. (1) It is well-know that in liquid paraffins, the gauche content would increase with temperature, resulting in frequency shifts of the C-H stretching bands. Moreover, the frequency of the C-H stretching band is not linearly proportional to the gauche content.46 Thus, the observed shifts of the CH2 stretching bands at a temperature range around the DSC endotherm region are expected, but they cannot exclusively indicate

Figure 5. Temperature dependence of the position of the scattering maximum for selected bands in FTIR for Cloisite organoclays: (a) CH2 symmetric stretching; (b) CH2 asymmetric stretching; (c) CH2 bending.

that these endotherms are due to the first-order transition. (2) While the position of the CH2 bending peak gradually shifts to lower frequencies as temperature increases, there is no appreciable change in the slope of the νCH2 bend vs T plot in the 30-60 °C range (Figure 5C). It is interesting to note that for all the test samples in the temperature range of 50-220 °C, the positions of CH2 stretching peaks can recover substantially after the heating/cooling cycle. However, after the heating/cooling cycle the CH2 stretching peaks at room temperature are always shifted 1-2 cm-1 toward higher frequencies than those of the original sample, indicating that the gauche content is increased after the heating-cooling cycle. A similar behavior was reported by Lee et al.,10 who suggested that the increase in the disorder might be related to the decrease in density of surfactant layers from thermal desorption of surfactant on the clay surface. This

3754


may also be the case here. However, we feel that there may be an additional possibility. Since the solution mixing method was used to prepare the as-received oroganoclays, it is likely that this process yields a nonequilibrium surfactant phase, having a higher content of trans conformation than that in the sample undergone the heating/cooling process. Figure 4b illustrates the temperature dependence of the shape of the 3416 cm-1 band, corresponding to OH stretching45 from adsorbed water molecules in the C6 sample. It is found that the OH stretching peak drastically weakens upon heating and disappears completely above 100 °C, indicating that the water molecules have escaped. However, upon cooling to room temperature in air, the OH stretching peak reappears. This indicates that the desorption and adsorption processes of water in organoclays are reversible. These observations further suggest that even with rather high surfactant loading, it is still not possible to transform the montmorillonite into an entirely “hydrophobic” material. Simultaneous Small-Angle X-ray Scattering and Wide-Angle X-ray Diffraction. Figure 6b exhibits selected SAXS profiles collected at different temperatures for two organoclays (C6 and C20). It is interesting to see that the SAXS profiles acquired at room temperature in different samples bear a similar trait. That is multiple peaks are seen but they do not appear at equidistant positions on the s scale. This feature is the characteristic of the so-called interstratified system, where multiple layers of different thicknesses can coexist in the same clay stack.16 With the increase in temperature, both SAXS profiles undergo drastic changes (Figure 6). The exact nature of these changes depends on the surfactant content. However, three temperature zones can be identified in these SAXS profiles for all tested systems: (1) 26-40 °C, below the endothermic DSC transition; (2) 40-180 °C, including and above the DSC transition but below the onset of thermal degradation; and (3) 180-260 °C, where significant surfactant loss is seen. In region 1, the SAXS peaks are not equidistant on the s scale. To be more specific, the position of the second scattering peak at s2 is less than twice of the position of the first scattering peak s1 (s2 < 2s1). In region 2, the scattering peaks become narrow and the positions of the peaks shift toward equidistant (s2 ) 2s1). In region 3, the scattering peaks broaden and become nonequidistant again. However, in this case, the position of the second peak (s2) shifts toward higher s values than twice that of the position of the first scattering peak s1 (s2 > 2s1). It is interesting to see that upon cooling from 260 to 30 °C (as in Figure 6a), the scattering peaks remain nonequidistant on the s scale, with s2 > 2s1 which indicates that the original layer structure in organoclays is permanently lost. This is in contrast with the FTIR results, which indicate that the average chain conformation of hydrocarbon chains before and after heating to 260 °C is quite similar (Figure 4). It should be noted that while at ambient temperature, the scattering profiles of C6 and C20 are different from each other; they become quite similar after heating to 260 °C and upon consequent cooling. The simultaneously measured WAXD profiles of two organoclays (C6 and C20) during heating and subsequent cooling (angular range 10° < 2θ < 35°) are shown in Figure 7. The observed major diffraction peaks are due to the structure of montmorillonite. The change of intensity in the measured angular range due to the contribution of organic component is minimal in all three temperature zones, which confirms that there is little long-range order

Gelfer et al.

Figure 6. Selected SAXS profiles for Cloisite organoclay: (a) C6 and (b) C20 at different temperatures. On the basis of the positions of scattering peaks, three temperature zones were determined.

Figure 7. Selected WAXD profiles for Cloisite organoclay C6 at different temperatures.

in the surfactant layers. On the basis of the procedures for SAXS analysis outlined earlier, we have characterized the temperature dependence of layer thickness distributions h(r) using the data from Figure 6. Comparisons of the measured SAXS profiles versus the calculated SAXS profiles at selected temperatures in each temperature zone for the C6 and C20 organoclays are shown in Figure 8, where the corresponding morphological parameters extracted from the analysis are displayed in Table 3 and Figure 9. In this analysis, we have assumed several conditions. (1) The mineral layers are rigid and uniform, having a thickness dmineral of 1.0 nm (Figure 1). In this


Langmuir, Vol. 20, No. 9, 2004 3755

Figure 8. Comparison of experimental vs calculated SAXS profiles for Cloisite organoclay: (a) C6 and (b) C20 at selected temperatures in different temperature zones.

case, the average thickness of the organic layer in the kth mode of a polymodal distribution h(r) can be calculated as rk,org ) rk - 1 (the unit in nm). (2) At elevated temperatures, some organic layers separating the mineral layers can escape or thermally degrade, resulting in the “fusion” of adjacent mineral layers into a dense sequence (tactoid), where the organic layer separating mineral layers is negligibly thin. In this case, the contribution of the fused mineral sequences in the overall h(r) distribution can be described by a delta function distribution δ(r - ro), centered at the position at ro ) 1.0 nm. (3) The overall thickness distributions h(r) for the organoclays (C6 and C20), depending on the temperature zone, can be modeled by one of the following morphological models: a unimodal layer thickness distribution (in Gaussian form), a bimodal layer thickness distribution (in Gaussian form), and a delta-function thickness distribution of the tactoid (eqs 13, 23, and 24), which may be described by eq 24.

h(r) ) c1gaush1gaus(r) + c2gaush1gaus(r) + cδhδ(r) c1gaus + c2gaus + cδ ) 1

(24)

Parameters for the h(r) distribution were calculated for the selected SAXS profiles, which included the mean distances between centers of silicate layers (average long period r1 and its standard deviation (σ1) for the unimodal Gaussian distribution, the two mean long period values (r1 and r2) and their standard deviations (σ1 and σ1) for the bimodal Gaussian distribution, as well as the corresponding fractions contribution of the first and second Gaussian modes and delta distribution for the fused tactoid (c1, c2, and cδ). In Figure 8, the comparisons of three sets of experimental SAXS profiles and calculated SAXS profiles using

the above modeling scheme show a very good qualitative agreement. The chosen SAXS profiles, representative to the one(s) in each temperature zone, including: (1) a lowtemperature curve, where s2 < 2s1; (2) two intermediate temperature curves, where s2 ) 2s1; (3) a high-temperature curve, where s2 > 2s1. For (1), the low-temperature scattering curve, a bimodal layer thickness distribution (rji ) 6.2 and 3.6 nm) was used to calculate the experimental scattering curves for both C6 and C20 organoclays. In this calculation, the average layer thicknesses rji as well as the content of the second mode of the layer distribution (with the higher rj value) c2 is larger in C6 than that in C20 (Figure 9, Table 3). Also it is seen that at ambient temperatures in C20 thinner and thicker organic layers coexist with fused mineral layers. This may be attributed to the structural heterogeneity of montmorillonite. Indeed, while overall surfactant content in C20 corresponds to ∼100% of cation exchange capacity in montmorillonite, the structural heterogeneity of a mineral21,22 may result in coexistence of nonintercalated mineral layers with layers having excessive surfactant content. Higher average layer thicknesses and absence of fused layers in C6 is consistent with higher overall surfactant content in this material (Table 1). For two intermediate temperature curves (collected at a temperature above the DSC transition at 60 and 160 °C) in (2), where s2 ) 2s1, a unimodal layer thickness distribution (rji ) 3.6 nm) was used for both C6 and C20 organoclays. Some differences were seen in these two systems. In C6, both scattering peaks become equidistant. In C20, while scattering peaks shift toward equidistant position, s2 remains less than 2s1. So the second phase with higher rj value persists above Ttr, and its content c2 (the second Gaussian mode) remains very low. In C6, the calculated rji value and its standard deviation (σι) value are slightly smaller (peaks are narrower) than those of the first layer thickness distribution calculated for the bimodal model in (1). In C20, rj1 slightly increases and σ1 decreases with increasing temperature, contribution of fused layers cδ, which is not zero even below Ttr, also increases with increasing temperature. At high temperatures (zone 3, s2 > 2s1), contribution of the second Gaussian mode c2 is zero for both C6 and C20, while the scattering curves in both materials become quite similar. Furthermore, the corresponding layer thickness distributions may be described as bimodal, consisting of a single Gaussian mode for the surfactant-intercalated layers and a delta thickness distribution (from the fused tactoid). It is interesting to see that the fractions of the fused tactoid (cfδ) becomes 30% for C6 and 35% for C20 at 260 °C, which is consistent with the corresponding weight losses detected in the staged TGA thermograms (Figure 2). Morphological Model for the Organoclays in Different Temperature Zones. On the basis of results obtained above, a semiquantitative morphological model can be established for organoclays in different temperature zones. Although the model was developed primarily for C6 organoclay, it is thought that this model is also valid for other montmorillonite-based organoclays, with organic surfactants bound to the clay surface by electrostatic interactions. To better understand this model, it is necessary to discuss some relevant physical parameters between the interactions of the clay surface and surfactant molecules first, as well as their consequences on the morphology in organoclays. The maximal surface density (Ds) of a monolayer formed by surfactant DMTA on the montmorillonite clay surface can be estimated by using the literature data47 as follows. The specific surface of montmorillonite, Sh, is 700 m2/g;

3756


Gelfer et al.

Table 3. Temperature Dependence of Morphology for Organoclays C6 and C20 As Determined from the Simulation of SAXS Data av no. of silicate layers N h

av long period 1st mode rj1 (nm)

std dev long period 1st mode σ1 (nm)

av long period 2nd mode rj2 (nm)

std dev long period 2nd mode σ1 (nm)

fraction of 2 organic mode c2

fraction of fused layers cδ

T (° C)

C6

C20

C6

C20

C6

C20

C6

C20

C6

C20

C6

C20

C6

C20

30 60 160 260

4 4 4 3

4 4 3 3

3.6 3.45 3.45 3.35

2.4 2.55 2.8 3.1

0.2 0.3 0.2 0.4

1.8 1.0 0.9 0.5

6.2

4.8 4.85 5.1

0.4

0.2 0.4 1.9

0.275 0 0 0

0.25 0.18 0.15 0

0 0 0.025 0.35

0.05 0.02 0.15 0.35

Figure 9. Temperature dependence of the fraction of fused layered silicates fδ calculated from the SAXS analysis.

while its cationic exchange capacity, Icec, on the surface is in the range (0.7-0.9) × 10-3 mol/g.47 For the fully exchanged system, the value of Ds (molecules/nm2) can be estimated using the following equation:

Ds )

IcecNA 109Sh

(25)

where NA ) 6.02 × 1023 mol-1 is the Avogadro number. The calculation thus yields Ds ∼ 0.8 molecules/nm2, representing a rather loosely packed monolayer on the montmorillonite surface. This evaluation is certainly consistent with the experimental observations made in this study: (1) the SAXS peaks were rather broad, indicating the presence of disordered layered structures; (2) the FTIR data also indicated the dominance of disordered conformation in surfactant chains. In other words, the conventional arguments that organoclays contain the formation of well-ordered predominantly trans organic layers27,35 or well-defined gauche blocks,4,33 are probably not correct for C6 and C20 organoclays. Another interesting observation is that the surfactant loadings in the tested organoclays (C6 and C20) are rather high. The loading in C20 is supposed to be equivalent to the maximum cation exchange capacity (CEC) of montmorillonite, while the loading in C6 significantly exceeds the CEC level of montmorillonite. As we conclude that the packing density of the surfactant monolayer on the clay surface is low, some free surfactant molecules, not complexed with montmorillonite, must exist. If these free surfactant molecules are excluded from the layered organoclay structures, they must form aggregates. These aggregates would manifest themselves as all-trans signals in FTIR. However, no evidence of surfactant aggregates was seen. Thus, we conclude that all surfactant molecules are likely to remain within the layered structures of organoclays. (47) Van Olphen, H. An Introduction to Clay Colloid Chemistry; Interscience Publishers: New York, 1963.

Results of our SAXS modeling suggest the coexistence of two distinct populations of organic layers, having average thicknesses rk,org1 ) r1 - 1 ) 2.6 and rk,org2 ) r2 - 1 ) 5.2 nm, in temperature zone 1. We propose that the thinner layer (rk,org1 ) 2.6 nm) represents the structure of a monolayer of surfactant molecules bound to the adjacent mineral surfaces (Figure 10). The monolayer structure is due to the hydrophobic aggregation of surfactant tails in the organic layers. The conformation of the tail structure is probably somewhat disordered due to the following reasons. The length of the C18 tail in DMTA for all-trans conformation is about 2.7 nm, which is larger than the measured thickness of the organic layer, including two hydrophilic surfactant heads. Thus, the surfactant chains in the monolayer do not consist of all-trans conformation (or the all-trans surfactant chains in the monolayer are tilted). Even if one considers the inclusion of the gauche block as suggested in the literature,4,33 the expected layer thickness would still be larger than the measured value (2.6 nm). Although we do not know the exact conformation of the surfactant chain in the organic layer, based on the experimental data (DSC and FTIR) and the simulation results, it cannot be completely disordered nor can it be in all-trans conformation. Similarly, we propose that the thicker organic layer (rk,org1 ) 5.2 nm) consists of two monolayers of surfactant molecules (we termed it the double-layer structure) with only about half of them bound to the mineral surfaces (Figure 10). The free (nonbound) surfactant molecules are held in the silicate gallery by van der Waals forces through aggregations of the hydrophobic tails.1,21,24 In temperature zone 2, we propose that the collapse of the double-layer structure leads to the formation of a unimodal thickness distribution in organic layers. This mechanism makes sense for several reasons. (1) The electrostatic interactions between the surfactant and clay surface are much stronger than the van der Waal forces between the surfactant layers. Thus, the complexation between the surfactant and clay surface should be much more thermally stable than the layer-layer interactions. (2) The proposed mechanism is certainly consistent with experimental results as follows. On the basis of the proposed model, the endothermic transitions observed in the first heating DSC scan can be attributed to the meltinglike order-disorder transition in surfactant tails. The melting-like order-disorder transition agrees well with FTIR results (the trans content significantly decreases in zone 2 as seen in Figure 5); it is also consistent with the appearance of a lower transition temperature in DSC (ca. 20 °C lower than the melting point of pure surfactant crystals). The order-disorder transition manifests itself in the decrease in a trans content (FTIR data) and a drastic decrease in the content of thicker organic layers (c2), while the thickness of thinner organic layer remains about constant. The “melting point” depression is probably due to the lower conformational order as well as the smaller size (i.e., one and two monolayers) of the trans domains formed by surfactant molecules within the organoclay


Langmuir, Vol. 20, No. 9, 2004 3757

Figure 10. The proposed morphological model for organoclays in three temperature zones.

galleries as compared with those in pure surfactant crystals. In Table 2, the total enthalpy of the endothermic transitions, ∆H, in the first heating scan of organoclays, is significantly lower than that of the free surfactant melting. This would also be expected, considering the lower conformational ordering in the as-received surfactant. It is interesting to note that the value of ∆H decreases almost two times at the second heating scan (Table 2); however, it remains about constant during subsequent heating/ cooling cycles. This can be explained by the decreased trans content after the heating-cooling cycle and redistribution of free surfactant molecules in organoclays resulting in their decoupling from bound surfactant. On the basis of Table 2, it is conceivable that the higher ∆H value represents a higher trans content formed during the preparation of organoclay from pristine clay via the ion exchange process in the solution. As the TGA data did not show appreciable weight loss during heating in the temperature zone 2, we suggest that the decoupled free surfactant may aggregate outside the layered structure; it must remain predominantly amorphous as paraffinlike structures cannot be detected by SAXS and WAXD neither in temperature zone 2 (above Ttr but below the melting temperature of a pure surfactant) nor after consequent cooling. In temperature Zone 3 (above 200 °C), the processes of thermal degradation as well as desorption of surfactant on the clay surface result in the loss of the organic phase and lead to the fusion of mineral layers (Figure 10C). It is interesting to note that while the content of the fused tactoid (cδ) increases rapidly with increasing temperatures the “fusion” process does not cause a drastic change in an average thickness of the remaining layered structure (Figures 9 and 10). We speculate that this is due to the structural heterogeneity of montmorillonite mineral base21,22 where the thermal stability of individual organic

layers in organoclays may vary significantly. Consequently, thermally less stable layers may be degraded and deplete the organic component significantly faster than the remainder of the material. Conclusions Morphological changes and phase transitions in Cloisite organoclay were investigated at different temperatures using TGA, DSC, FTIR, simultaneous SAXS, and WAXD techniques. In addition, a quantitative analysis of the SAXS data was carried out to model the corresponding organoclay structures. On the basis of experimental and modeling results, the structures of organoclays can be divided into three types, depending on the temperature regions. (1) In zone 1 temperature range (room temperature to about 40 °C, less than the endothermic DSC transition temperature, Ttr), a bimodal organic layer thicknesses distribution exists in organoclays. The thinner organic layer contains a monolayer with surfactant molecules bound to the adjacent clay surfaces. The thicker layer contains two monolayers of surfactant with only about half of them being bound to the clay surface. This assembly was termed the double layer structure with a thickness of exactly twice that of the thin layer thickness. The double layer content was found to increase with the overall surfactant loading. The FTIR data indicate that the organic layers are rather disordered even with high surfactant content and the conformation of hydrocarbon chains resembles that of liquid alkanes rather than that of paraffin crystals. (2) In zone 2 temperature range (40-180 °C), the content of thicker layers drastically decreases. At temperatures above Ttr, organoclays undergo a melting-like orderdisorder transition, resulting in the randomization of

3758


surfactant conformation. The thickness of the thinner organic layer is about the same as that of the thin layer in zone 1. (3) In the zone 3 temperature range (180-260 °C), the layer thickness distribution becomes bimodal again, organoclays containing disordered monolayers of bound surfactant and fused silicate layers. The drastic increase in the content of fused silicate layers above 200 °C can be attributed to thermal degradation and/or escape of bound surfactant molecules. Results from this study show that feasible interpretation of SAXS data from organoclays and nanocomposites

Gelfer et al.

must account for the complex character of layer thickness distribution in these systems. The values of average long periods determined from the position of the first scattering peak may not be considered reliable. Acknowledgment. This study was supported by NSF MSREC (DMR 0080604), NSF Inter-American Grant (DMR0302809), and NSF (DMR9984102) at Stony Brook. The SAXS synchrotron beamline X27C was supported by the Department of Energy (Grant DE-FG02-99ER 45760). LA035361H

Thermally Induced Phase Transitions and Morphological Changes in

Recommend Documents