5996
Langmuir 2003, 19, 5996-6003
Metastable and Stable Morphologies during Crystallization of Alkanes in Miniemulsion Droplets Rivelino Montenegro and Katharina Landfester* Max Planck Institute of Colloids and Interfaces, Research Campus Golm, 14424 Potsdam, Germany Received December 17, 2002. In Final Form: April 22, 2003 Undercooling and crystallization of different alkanes in stable nanodroplets with narrow size distributions were analyzed by using DSC measurements. The crystal structure was determined by X-ray measurements. Alkane droplets with a defined size between 100 and 500 nm in water were prepared by using the miniemulsion process. The required undercooling to obtain crystallization in such droplets is significantly increased compared to that for the bulk material since the nucleation mechanism is shifted from heterogeneous to homogeneous nucleation. For the even alkanes (C18-C24), a structure change from the triclinic in the bulk to orthorhombic structure in small droplets (100 nm) was detected and attributed to confinement effects inside the droplets. An intermediate rotator phase is of less relevance for the nanosized droplets. For odd alkanes, a strong temperature shift of the crystallization point compared to the bulk system was detected, but no structure change was observed. Both in bulk and in miniemulsion droplets an orthorhombic structure was formed.
Introduction The crystallization of alkanes has been exhaustively studied because of very important industrial consequences and applications, such as processing of oils, fats, and surfactants.1 Looking at the different n-alkane chain lengths, there is an unusual behavior of the crystallizing or melting for even and odd alkane chains, also known as the odd-even effect. As early as 1877, Baeyer stated that the melting point of the fatty acids with even numbers of carbon atoms are relatively higher than those with odd numbers.2 Although the phenomenon of the odd/even alternation has been known for a very long time, a plausible explanation of this pattern does not yet exist.3 The even-odd effect that is observed for the melting crystallization as well as for the intermediate phases can be detected several degrees before the melting. In 1932, Mu¨ller4 had shown by X-ray scattering the existence of an intermediate phase for some paraffins between the crystalline and the liquid phase where the molecular chains were in a state of more or less free rotation about their axes (the so-called rotator phase). He described the rotator phase as a layered structure in which each layer is formed by the hexagonal packing of the aliphatic chains with their long axes perpendicular to the layer planes. After that first observation of the rotator phase, other studies have been performed to better understand those metastable phases.5-11 While the structure of the crystalline phase of n-alkanes is characterized by a compact stacking of chain molecules, * To whom correspondence should be addressed. (1) Taggart, A. M.; Voogt, F.; Clydesdale, G.; Roberts, K. J. Langmuir 1996, 12, 5722-5728. (2) Baeyer, A. Ber. Chem. Ges. 1877, 10, 1286. (3) Boese, R.; Weiss, H.-C.; Bla¨ser, D. Angew. Chem., Int. Ed. 1999, 38, 988-991. (4) Mu¨ller, A. R. Soc. London, Ser. A 1932, 138, 514. (5) Mazee, W. M. Recl. Trav. Chim. Pays-Bas 1948, 67, 197. (6) Larsson, K. Nature 1967, 13, 383. (7) Ungar, G.; Masic, N. J. Phys. Chem. 1985, 89, 1036. (8) Doucet, J.; Denicolo, I.; Craievich, A. J. Chem. Phys. 1981, 75, 1523. (9) Doucet, J.; Denicolo, I.; Craievich, A.; Germain, C. J. Chem. Phys. 1984, 80, 1647. (10) Ungar, G. J. Phys. Chem. 1983, 87, 689. (11) Dorset, D. L. EMSA Bull. 1990, 20, 54.
the long molecular axes being perpendicular to the stacking planes in odd-numbered and tilted in evennumbered compounds,12 the rotator phases are lamellar crystals which lack long-range order in the rotational degree of freedom of the molecules about their long axes.13 There are five rotator phases reported.7-11,14 They differ in their symmetries, in-plane molecular packing, layering sequences, and the amount of molecular tilt with respect to the layer spacing.15 In general, the stable phase for n-alkanes is triclinic for 12 e n (even) e 26,1 orthorombic for 9 e n (odd) e 35,16 and monoclinic for 28 e n (even) e 36,1 while the rotator phase for n (odd) e 23 is orthorhombic and hexagonal for n (even) e 24.9 Those transition phases, once formed, may persist indefinitely and have a strong role in the crystallization process and in crystal morphology.17 It is well known that the melting temperatures and rotator transition phase temperatures increase with increasing alkane chain length.18 However, the difference between the melting temperature and the rotator equilibrium transition temperature in bulk systems is roughly constant as a function of chain length.19 It is the topic of the present contribution how the crystallization of different alkanes will occur in confined alkane nanodroplets. Emulsification tends to increase the supercooling required for crystallization over that of bulk liquids. This was related to the small volume of the droplets where the crystallization mainly takes part not by heterogeneous nucleation but by homogeneous nucleation which occurs at lower temperatures than heterogeneous nucleation.20-22 The decrease of the crystallization tem(12) Denicolo, I.; Craievich, A. F.; Doucet, J. J. Chem. Phys. 1984, 80, 6200-6203. (13) Sirota, E. B.; Herhold, A. B. Polymer 2000, 41, 8781-8789. (14) Sirota, E. B.; King, J.; Singer, D. M.; Shao, H. H. J. Chem. Phys. 1993, 98, 5809. (15) Herhold, A. B.; H. E. K., Jr.; Sirota, E. B. J. Chem. Phys. 2002, 116, 9036-9050. (16) Oliver, M. J.; Calvert, P. D. J. Cryst. Growth 1975, 30, 343-351. (17) Sirota, E. B.; Herhold, A. B. Science 1999, 283, 529. (18) Small, D. M. The Physical Chemistry of Lipids: from Alkanes to phospholipids, 1986. (19) Sirota, E. B.; H. E. K., Jr.; Shao, H. H.; Singer, D. M. J. Phys. Chem. 1995, 99, 798.
10.1021/la027019v CCC: $25.00 © 2003 American Chemical Society Published on Web 06/21/2003
Metastable and Stable Morphologies of Alkanes
perature, that is, the temperature where in a defined kinetic protocol crystallization occurs, is reported by McClements et al.23 and Kaneko et al.24 The reason for the shift of the crystallization temperature was associated with the number of foreign crystallization nuclei, which in the bulk phase usually causes heterogeneous nucleation, but which in the confined phase are now distributed among a large number of isolated droplets. The probability of any individual droplet containing a crystallization nucleus is, however, practically zero.25 The size of the droplets plays a very important role, since the crystallization temperature lowers with decreasing size as experimentally shown elsewhere for hexadecane.22 Another aim of this work is to study the influence of the nanosize droplets on the formation of the rotator phase of n-alkanes from C15H32 (C15) to C24H50 (C24). To obtain defined droplets of n-alkanes in the nanometer range, we have used the miniemulsion approach. Miniemulsions allows one to provide small, stable, and narrowly distributed nanodroplets with a controllable size in the range of 50-500 nm with partly adjustable surface tension and chemical surface structure. Their highly reproducible generation by high shear devices, the stabilization by combination of surfactants and osmotic pressure controlling agents, as well as some of their properties were recently reviewed.26 Experimental Section Materials. Pentadecane, hexadecane, and heptadecane (∼99% purity) were purchased from Aldrich. Sodium dodecyl sulfate (SDS) (>98% purity) was also purchased from Aldrich. The others n-alkanes C18-C24 (∼98% purity) and perfluorohexane were obtained from Fluka. All compounds were used as received. Preparation of the Miniemulsions. One gram of n-alkane together with 50 mg of perfluorohexane, which is used as ultrahydrophobe to suppress Ostwald ripening, was mixed with a solution of 30 mg SDS in 10 g of distilled water. The mixture was vigorously stirred for 1 h above the melting temperature of the alkane. After stirring for 1 h, the miniemulsion was prepared by ultrasonicating the emulsion with a Branson sonifier W450 (microtip) at an amplitude of 70%. To obtain different droplet sizes, we have used different sonication times. Analytical Methods. The particle sizes were measured using a Nicomp particle sizer (Model 370, PSS Santa Barbara, USA) at a fixed scattering angle of 90°. The DSC measurements were carried out using a Netzsch Thermal Analyze DSC 200. Sample masses between 10 and 20 mg were used. The cooling and heating rate of 5 K‚min-1 was kept for all experiments. Wide-angle X-ray (WAXS) diffraction was performed below the crystallization temperature of the droplets of the miniemulsions as well as of bulk n-alkanes using a Nonius CP120 Diffractometer.
Results and Discussions Even Alkanes. In the first set of experiments, the crystallization behavior of tetracosane (C24) droplets was compared to that of tetracosane in bulk, using DSC measurements. The crystal structure is determined by X-ray measurements. Cooling down the bulk system leads (20) Vonnegut, B. J. Colloid Sci. 1948, 3, 563. (21) Kelton, K. F. Solid State Phys. 1991, 45, 75. (22) Montenegro, R.; Antonietti, M.; Mastai, Y.; Landfester, K. J. Phys. Chem. 2003, accepted. (23) McClements, D. J.; Dungan, S. R.; German, J. B.; Simoneau, C.; Kinsella, J. E. J. Food Sci. 1993, 58, 1148. (24) Kaneko, N.; Horie, T.; Ueno, S.; Yano, J.; Katsuragi, T.; Sato, K. J. Cryst. Growth 1999, 197, 263. (25) Walstra, P.; Berensteyn, E. C. H. Neth Milk Dairy J. 1975, 29, 35. (26) Landfester, K. Macromol. Rapid Commun. 2001, 22, 896.
Langmuir, Vol. 19, No. 15, 2003 5997
to a first transition at 45.2 °C (see Figure 1a) to obtain the rotator phase as determined by X-ray scattering (see Figure 1b) and a second transition at 40.3 °C where a triclinic structure is formed. The characteristic peaks for this structure are (0 1 0) at 19.40°, (0 1 1) at 20.00°, (0 1 2) at 21.92°, (1 0 1) at 22.31°, (0 1 3) at 23.54°, and (1 1 1) at 24.97°. In the miniemulsions, the crystallization occurs in one single step at 30 °C. While heating the miniemulsion droplets, multiple transitions can be detected. For 560and 120-nm droplets, a first transition is seen at 43 °C and another large transition at 51 °C, but in both cases a third very weak transition at 46.3 °C is also visible. Therefore, the temperature at which (dynamic) crystallization of tetracosane in the miniemulsion occurs is lower than in bulk (shifted from 40 °C in bulk to about 30 °C in miniemulsion), whereas the melting process is not much influenced. This can be explained by the consideration of different nucleation mechanisms. In the bulk system, a few nuclei are sufficient to induce (heterogeneous) nucleation followed by crystal growth. In miniemulsion, 1016 to 1017 sites per liter have to nucleate separately, and crystal growth is limited to the dimension of the droplet. As already stated, the probability of nanodroplets to contain a “foreign” element acting as a substrate for heterogeneous nucleation is practically zero. This shifts the crystallization mechanism from heterogeneous nucleation to homogeneous nucleation. The fact that the melting point of the nanodroplets occurs roughly at the same point as in the bulk system, in contrast to the crystallization, is expected, since the release of energy of a droplet during the crystallization due to the undercooling is practically instantaneous because it occurs far from thermodynamic equilibrium. In the case of melting, the droplets absorb energy at a fixed melting point and the kinetics depend on the exchanges with the surrounding medium.27 However, it will be shown for alkanes with different chain length that the melting point can also be shifted significantly. The X-ray scattering measurements reveal that in the case of 150-nm droplets the stable phase has an orthorhombic structure (instead of the triclinic structure of the bulk phase). This means that there is a finite size effect due to confinement of crystallization in the droplets which influences the crystal morphology. In the 560-nm droplets, the system forms triclinic crystals as in the bulk phase which leads to the conclusion that for induction of structural changes the droplets must be very small. In the bulk system of tetracosane, the enthalpy of the transition into the rotator phase during cooling was -170 J‚g-1 and of the transition into the stable phase, -75 J‚g-1. While heating, the magnitude of enthalpy of the two transitions only slightly shifts (86 for the first and 159 J‚g-1 for the second) and the sums of ∆Hcryst1 + ∆Hcryst2 and ∆Hmelt1 + ∆Hmelt2 are constant. In the tetracosane miniemulsion, the enthalpy of the single transition in the cooling step, ∆Hcryst, is -240 J‚g-1 which nicely corresponds to the total enthalpy of both transitions in the bulk system. Evaluating the enthalpy in the heating process ∆Hmelt, the first transition has a ∆Hmelt of 32 J‚g-1 and the second large one has ∆Hmelt2 of 114 J‚g-1. A possible explanation for the difference between the enthalpy ratios ∆Hmelt1/ ∆Hmelt2 of the bulk and the droplet phases may be that, in the droplets, tetracosane needs less energy to reach the rotator phase than in the bulk, that is, the molecules in the droplets are more easily mobilized. The seemingly (27) Dumas, J. P.; Krichi, M.; Strub, M.; Zeraouli, Y. Int. J. Heat Mass Transfer 1994, 37, 737-746.
5998
Langmuir, Vol. 19, No. 15, 2003
Montenegro and Landfester
Figure 1. (a) DSC and (b) X-ray measurements for tetracosane droplets and bulk system.
smaller enthalpy during heating might also be attributed to a constant loss of the ordering during the heating process as recently discussed by Thurn-Albrecht et al.28 In the next sets of experiments, the behavior of shorter even alkanes, namely, C22, C20, C18, and C16, in droplets is examined and compared to the properties in bulk. The DSC graphs and the X-ray diagrams are shown in Figure 2 and the data extracted from the DSC curves are summarized in Tables 1 and 2. In the C22, C20, and C18 alkanes, the situation for crystal formation is very similar to that of C24. Whereas in the bulk systems of C22 and C20 throughout cooling a transition into the rotator phase and then into the stable phase is obtained, in the miniemulsion only one single transition is detected while lowering the temperature to the minimum temperature of the experiment which is given by the freezing of the continuous aqueous phase at a certain cooling rate (for the cooling rate of 5 K‚min-1 the limit temperature is -20 °C). Since, inside the droplets, the crystallization temperature is well below that for equilibrium melting, at the temperature of the observed (28) Ro¨ttele, A.; Thurn-Albrecht, T.; Sommer, J.-U.; Reiter, G. Macromolecules 2003, 36, 1257-1260.
crystallization, the equilibrium structure is already crystalline. However, for a short time a transient metastable rotator phase in the nucleation process may also exist.29 For the C18, the metastable phase was also not observed in the bulk, as this alkane crystallizes directly in the most stable phase, the triclinic one. When melting, it goes directly from the stable phase into the liquid phase. This may be the possible reason for the relatively higher melting points of the even alkanes, since the triclinic systems have greater packing efficiency, and thus stability, compared with monoclinic and orthorhombic systems.1 Gerson et al.30 and Roberts et al.31 have shown through alkane solubility measurements that the triclinic is the most stable solid phase. As seen earlier, the transition into the stable crystalline structure occurs in miniemulsion at much lower temperatures than the bulk phase which can be attributed to homogeneous nucleation. However, (29) Kraack, H.; Deutsch, M.; Sirota, E. B. Macromolecules 2000, 33, 6174-6184. (30) Gerson, A. R.; Roberts, K. J.; Sherwood, J. N.; Taggart, A. M. J. Cryst. Growth 1993, 128, 1176. (31) Roberts, K. J.; Sherwood, J. N.; Taggart, A. M. Proceedings from 12th Symposium on Industrial Crystallization 1993, 3, 63.
Metastable and Stable Morphologies of Alkanes
Langmuir, Vol. 19, No. 15, 2003 5999
Figure 2. DSC and X-ray measurements for some even alkane droplets and the bulk systems: docosane C22, eicosane C20, octadecane C18, and hexadecane C16.
6000
Langmuir, Vol. 19, No. 15, 2003
Montenegro and Landfester
Table 1. Transition Temperatures as Determined by DSC Measurements for Alkanes in Bulk and Miniemulsion Droplets cooling number of Cs 15 16 17 18 19 20 21 22 23 24
bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets: bulk droplets:
169 nm
TC2b (°C)
∆TCc (°C)
Tm1d (°C)
Tm2e (°C)
∆Tmf (°C)
3.7 -9.4
-6.9 -14.4 12.0 -4.7 5.6 -1.4 19.5 5.6 17.4 11.7 28.4 19.0 27.3 21.8 33.9 25.0 35.5 28.5 40.3 30.6
10.6 5.0
-1.7 -3.7
13.3 13.4
8.7 4.2
12.3 9.6
2.6 8.2 1.7 1.7
15.3 22.7 17.7
11.6 9.7 19.0 18.8 25.0 21.9 31.2 27.3 33.7 31.2 40.2 36.2 42.7 40.7 48.0 44.4 50.2 47.8 52.2 51.5
208 nm 140 nm 125 nm 111 nm
heating
TC1a (°C)
14.3 2.8 8.2 25.6 13.4 30.1
129 nm 34.2 126 nm 37.0 121 nm 41.6 128 nm 45.2 150 nm
6.9 3.1 6.1 4.9
24.1 33.1 30.0 44.8 35.1 41.6 39.2 48.7 43.2
12.7 12.3 12.0 11.0 13.5 12.1 9.6 10.7 3.2 9.3 8.6 8.6 3.5 8.3
a Transition from the liquid phase to the rotator phase during the cooling process. b Transition from the rotator phase to the more stable crystalline phase. c ∆TC ) TC1 - TC2: difference between these two crystallization steps. d Transition from the stable crystalline phase to the rotator phase during the heating process. e Transition to liquid state. f ∆Tm ) Tm1 - Tm2: difference between the beginning of the rotator phase and the melting point during the heating process.
Table 2. Transition Enthalpies of Alkanes in Bulk and Miniemulsion Droplets as Obtained by Integration of the Transition Peaks in the DSC Curves number of Cs 15 16 17
18
19
20 21 22 23 24
bulk droplets:210 nm droplets:169 nm bulk droplets: 308 nm 218 nm bulk droplets: 490 nm droplets: 153 nm droplets: 134 nm bulk droplets: 254 nm droplets: 153 nm droplets: 125 nm bulk droplets: 198 nm droplets: 135 nm droplets: 111 nm bulk droplets: 293 nm droplets: 129 nm bulk droplets: 193 nm droplets: 126 nm bulk droplets: 283 nm droplets: 121 nm bulk droplets: 196 nm droplets: 150 nm bulk droplets: 560 nm droplets: 150 nm
∆Hcryst1
∆Hcryst2
∆Hcryst1/∆Hcryst2
∆Hmelt1
∆Hmelt2
-161.2 -148.7 -151.1
-39.2 -22.4 -26.6 -220.8 -206.5 -205.7 -42.46 -31.67 -21.75 -26.0 -213.3 -8.30 -23.6 -39.1 -47.4 -47.4 -36.4 -30.0 -84.9 -248.2 -179.7 -51.0 -205.3 -197.7 -63.0 -199.3 -198.1 -43.6 -173.9 -165.7 -75.23 -242.2 -239.1
4.1 6.6 5.7
40.6
4.1 5.4 6.8 5.8
43.8 29.6 29.0 26.3
31.5 10.8 5.8 3.7 3.6 4.3 5.4 1.7
1.88 8.07 17.7 55.0 15.8 8.0 10.7
157.2 140.6 142.5 219.8 205.9 193.2 161.2 150.0 149.8 149.0 212.2 239.4 210.5 188.5 168.2 136.6 112.6 109.2 233.1 189.3 155.8 156.6 110.0 93.5 162.6 115.1 99.2 116.8 90.0 78.9 159.7 114.3 134.3
-175.7 -171.4 -147.3 -150.1 -261.5 -256.0 -225.4 -176.3 -169.6 -157.9 -163.0 -145.8 -170.4 -147.3 -133.4 -170.5
for the C22, C20, and C18 alkanes, the melting point is also significantly lowered by up to 4 K. However, heating the miniemulsion droplets consisting of those alkanes leads again to two transitions: the first transition into the rotator phase, which is very weak, and the second transition, which is strong. In the C18, there is a strong dependence on the droplet size of the ratio of ∆Hmelt1 to ∆Hmelt2. With decreasing droplet size, ∆Hmelt1 increases significantly which leads us to the conclusion
3.3 2.3 3.1 2.3
26.4
2.29 65.4 19.8 14.8 57.4 11.5 12.4 61.0 17.4 10.2 86.4 31.8 27.2
∆Hmelt2/∆Hmelt1 3.9 5.4
3.7 5.1 5.2 5.7 127.3 26.1 10.6 3.1 8.6 14.1 10.2 68.0 2.4 5.6 6.3 2.8 10.1 8.0 1.9 5.2 7.8 1.9 3.6 4.9
that the confinement in the droplets has a significant influence on the premelting transition. Comparable to the C24 miniemulsion, the sum of ∆Hmelt of the two transitions is significantly smaller than the sum of ∆Hcryst of the systems. The X-ray scattering measurements reveal that for small droplets (120 nm), an orthorhombic phase is formed. From the positions of the peak, it is seen that in this phase the crystal is better packed than in the rotator phase. The stable phase of the bulk system is triclinic.
Metastable and Stable Morphologies of Alkanes
Langmuir, Vol. 19, No. 15, 2003 6001
Figure 3. DSC and X-ray measurements for odd alkane droplets and bulk systems: tricosane C23, heneicosane C21, nonadecane C19, heptadecane C17, and pentadecane C15.
6002
Langmuir, Vol. 19, No. 15, 2003
Montenegro and Landfester
Table 3. Crystal Sizes (in nm) Obtained from the X-ray Measurements number of C atoms
bulk-rotator (110)
15 17 19 21 23 a
38 50 38 38 22
bulk-stable
(200) (110) (200) 43 43 34 27 25
37 43 38 50 24
38 38 38 43 34
miniemulsion (ca. 200 nm a droplets) rotator phase
miniemulsion (ca. 200 nm a) droplets) stable phase
miniemulsion (ca. 120 nm a) droplets) rotator phase
miniemulsion (ca. 120 nma droplets) stable phase
(110)
(200)
(110)
(200)
(110)
(200)
(110)
(200)
43 37 43 24 23
38 50 43 24 30
43 33 50 43 38
38 34 43 43 33
37 43 33 23 23
30 50 34 34 20
43 37 30 23 33
34 38 30 20 38
For exact sizes see Table 2.
This again shows nicely the influence of the droplet confinement on the crystal morphology. For larger droplets (about 300 nm), mixed structures (triclinic and orthorhombic) are obtained; in the C22 droplets, the orthorhombic structure is favored, and in the C20 and C18 droplets, the triclinic structure is favored. From these measurements, we can estimate that below a critical size of about 250 nm the confinement has a strong influence on the crystal structure. In addition, the finite size effect becomes less significant with decreasing chain length. In the C16 alkane, we have neither for the bulk nor for the miniemulsion detected a rotator phase, and direct crystallization into the triclinic structure occurs. While the peak (1 1 1) is more intense in the bulk, it decreases in the droplets, and the first two peaks (0 1 0) and (0 1 1) that are small in the bulk become very intense in the droplets. In addition, the (0 1 2) peak almost completely disappears in the droplets, indicating a preferential twodimensional growth and shape selectivity in the droplets. Although Sirota et al. reported that the difference between the melting temperature and the rotator equilibrium transition temperature is roughly constant as a function of chain length,19 we have seen (Table 2) that for the bulk odd alkanes this difference decreases with increasing the alkyl chain and that for the bulk even alkanes it increases from C20 on, as we did not observe the rotator phase in the even bulk alkanes before that length. Odd Alkanes. The size quantization effect on odd alkanes was qualitatively very different from that on even alkanes and so the experiments with odd alkanes are separately discussed in this section. The DSC and X-ray measurements for the odd alkane droplets and the corresponding bulk systems are shown in Figure 3. For odd alkanes in the bulk, one can clearly see a double transition in the cooling as well as in the heating step. In the cooling step, the first transition, TC1, corresponds to the transition from the liquid state to the rotator phase (orthorhombic), and the following one, TC2, corresponds to the final crystallization in the immobile phase, again an orthorhombic one. For the melting process, the first transition, Tm1, corresponds to the transition from the immobile orthorhombic to the mobile orthorhombic phase, and Tm2 corresponds to the transition from the rotator phase to the liquid state. In miniemulsions, the crystallization takes place also for the odd alkanes at lower temperature, but the structure in the droplets does not undergo any structure shift. As in the bulk system, the orthorhombic structure is formed. The X-ray diffractogram shows two characteristic peaks, the first (1 1 0) at 2θ ) 21.7° and the second (2 0 0) at 2θ ) 24.1°. As expected, the low-temperature phase shows more well-defined peaks than the rotator phase. The transition from the rotator to the low-temperature phase promotes a homogeneous strain on the crystals as indicated by the slight shift in the peak position. For all
Figure 4. Temperature shift for the crystallization and melting (only the transition in the stable state) between the bulk and the miniemulsions. Data taken from Table 1.
odd alkanes with a chain length between C15 and C23, neither the increase in chain length nor the different droplet sizes show a remarkable influence on the orthorhombic crystals. The crystallite size roughly decreases (see Table 3) with increasing alkyl chain length but does not show any direct relationship with droplet size, probably because of the stochastic nature of the nucleation. As well as the absence of structural differences between the bulk and the miniemulsion droplets of different sizes for the odd alkanes, the temperature shifts of the phase transitions are of interest. For the cooling step, the difference between TC1 and TC2 decreases for the odd alkanes in the bulk phase. For the miniemulsions of these alkanes, the difference between the first and second transition in the cooling step (∆TC ) TC1 - TC2) is drastically reduced in comparison to the bulk and decreases with increasing alkyl chain from C15 to C19. During the heating procedure, the nanodroplets and the bulk behave, however, very similarly, and ∆Tm (∆Tm ) Tm2 - Tm1) only decreases with increasing the alkyl chain length. This is indicative of the kinetic character of the observed temperature shifts. Figure 4 shows the differences of the transition temperatures between the bulk and miniemulsion for both the even and odd alkanes. For the cooling step, we took the difference between the transitions to the stable phase of the bulk and the miniemulsion, and for the heating step, the transitions to the liquid state. Looking at the differences of the transition temperature to the stable phases (TC2 and Tm2) between the bulk and the miniemulsion, we find that the crystallization and the melting in droplets is always shifted by the same factor for odd alkanes and is independent of the chain length, consistent with our observation that the confinement leads to a decrease of the crystallization and melting temperatures but does not influence the structure.
Metastable and Stable Morphologies of Alkanes
Langmuir, Vol. 19, No. 15, 2003 6003
ratios for ∆Hcryst1/ ∆Hcryst2 and ∆Hmelt2/∆Hmelt1. Here, with decreasing droplet size the ratios increase. Conclusion
Figure 5. Ratio of ∆Hcryst1/∆Hcryst2 (from cooling) and ∆Hmelt2/ ∆Hmelt1 (from heating) for the miniemulsions. Data taken from Table 2.
In comparison, even alkanes show in the cooling process that this temperature shift decreases with increasing the alkyl chain until C20, at carbon chains >C20, the difference increases again. For the melting process, one can see the inverse effect that the difference increases until C20 and decreases afterward. This is a consequence of a structural shift observed for even alkanes which additionally influences the transition to the stable phase. In Figure 5, the ratio of ∆Hcryst1/∆Hcryst2 (from cooling) and ∆Hmelt2/∆Hmelt1 (from heating) are presented for all alkane miniemulsions. For the even alkanes, the ratios show minimal values for small droplets and increase with increasing particle size, resulting in an infinite value of bulk (where only one transition is seen). The reverse case is seen for the odd alkanes, where the bulk shows small
Dynamic crystallization and melting experiments were performed in small, stable, and narrowly distributed nanodroplets consisting of different alkanes (C15-C24) in water. It was shown that the temperature of crystallization in such droplets is significantly decreased as compared to the bulk material. This is attributed to a very effective suppression of heterogeneous nucleation. A very different behavior was detected for odd and even alkanes. In even alkanes, the confinement in small droplets changes the crystal structure from a triclinic (in bulk) to an orthorhombic structure which is attributed to finite size effects inside the droplets. While cooling, an intermediate rotator phase is not detected for the miniemulsion droplets. Since the nucleation barrier causes crystallization well below the equilibrium melting point, at the temperature of observed crystallization, the equilibrium structure is crystalline. On (equilibrium) heating, the transition into a rotator phase is detected. For odd alkanes, only a strong temperature shift is observed compared to the bulk system but no structure change is observed, as both in bulk and in miniemulsion droplets, an orthorhombic structure is formed. Crystallization in miniemulsion droplets is indeed strongly influenced because of the finite size of the droplets which implies an expansion of the nanodroplet experiments to molecules with functionality, for example, dye molecules or liquid crystals. Acknowledgment. We would like to thank Markus Antonietti for many helpful discussions and Jeremy Pencer for carefully reading the manuscript. LA027019V
