Maltodextrin Mixture Studied

Biomacromolecules 2002, 3, 1208-1216

1208

Phase Separation in a Sheared Gelatin/Maltodextrin Mixture Studied by Small-Angle Light Scattering Michael F. Butler* Unilever Research and Development, Colworth House, Sharnbrook, Bedfordshire MK44 1LQ, U.K. Received May 13, 2002; Revised Manuscript Received October 2, 2002

The influence of shear on the structure of a gelatin/maltodextrin mixture was investigated using small-angle light scattering both during phase separation and after phase separation was allowed to occur quiescently. In all cases, phase separation occurred via spinodal decomposition to form a droplet morphology, and a characteristic length scale was formed in the structure that was prevalent during shear, as well as in quiescent conditions. Below the critical shear rate for droplet breakup, shear accelerated the coarsening rate of the droplets. A transient regime of rapid hydrodynamic coarsening was present when shear was initiated after phase separation and at late times in all cases once the droplets attained a certain size. At the critical shear rate for droplet breakup (1 s-1), the rapid repetition of breakup and coarsening was postulated to occur, which enabled a microstructure consisting of elongated droplets with a narrow size distribution to form. When the shear rate enabled droplets to extend to such an extent that a percolated structure could form (10 s-1), then the structure was relatively stable and changed very slowly over time. At very high shear rates (100 s-1), droplet breakup was suppressed and a highly fibrillar morphology formed that was stable only while the system was under shear. Cessation of shear at high rates led to fiber breakup and the formation of many small droplets. For a given shear rate, the final microstructure appeared to be independent of the time that shear was started when the structure consisted of discrete droplets or fibers. When a percolated structure could form, however, the shear history appeared to be important. Introduction Phase separation in polymer mixtures commonly occurs via the mechanism of spinodal decomposition.1-3 In quiescent conditions, this phenomenon is well understood. Mixtures near the critical composition possess an interconnected, bicontinuous structure, whereas off-critical mixtures contain droplets. In both cases, however, the morphology is characterized by the presence of a characteristic length scale corresponding to a domain size, L, the development of which can be separated into three regimes, the early, intermediate, and late stages. In the early stage, phase separation is diffusion-controlled. The characteristic length scale remains constant during this period, while the composition difference between the phases increases, and the interface between them sharpens. Nonlinear growth terms cause the characteristic length scale to increase during the intermediate stage, and by the late stage, the phases have attained their equilibrium compositions and growth of the domains is driven by the interfacial tension. Late-stage coarsening kinetics are characterized by a power-law dependence of the domain size, L ∝ tR, where R is typically 1/3 when growth is dominated by Ostwald ripening or droplet coalescence mechanisms and 1 when hydrodynamic flows are the major influence. Experiments have shown that the kinetics of spinodal decomposition in biopolymer systems can be well-explained by the theories that have been used to explain the behavior of model synthetic polymer systems.4,5 * E-mail: [email protected]. Telephone: +44 (0)1234 222958. Facsimile: +44 (0)1234 222757.

The influence of shear on phase separation is less well understood, however, although it is known to potentially alter phase-separation behavior in the following ways. First, shear can influence the mechanism of phase separation itself by affecting the composition fluctuations that lead to demixing.6-8 In near-critical fluids, shear suppresses the concentration fluctuations that lead to phase separation and therefore promotes mixing. In polymer solutions containing two components with very different viscoelastic properties, however, the concentration fluctuations cause stresses that result in diffusion in the direction of phase separation. Application of shear enhances these stresses, thus inducing demixing in this case. This effect becomes more pronounced as the polymer molecular weight increases. Second, shear alters the morphology of the phases that form upon demixing by inducing anisotropy in the structure.6,9-14 Droplets become extended to form ellipses and bicontinuous structures become elongated in the direction parallel to the shear flow, sometimes forming transient, chevron-shaped morphologies that become fibrillar at high shear rates.15 The extent of the anisotropy depends on the viscosity ratio, surface tension, and shear rate. In systems with very low surface tensions (such as polymer solutions), shear is believed to suppress the surface undulations that develop on extended droplets, thus preventing breakup and allowing extended cylindrical domains to form at high shear rates that are known as string phases.16,17 Experimental studies on string phases in polymer solutions have shown, however, that the diameter of the fibrils decreases with increasing shear rate until eventually

10.1021/bm0255645 CCC: $22.00 © 2002 American Chemical Society Published on Web 10/26/2002

Phase Separation and Shear in a Biopolymer Mixture

the fibril diameter equals the interfacial thickness and shearinduced homogenization therefore occurs.6 Third, shear influences the coarsening kinetics of the phase-separating domains. In quiescent phase separation, hydrodynamic flows accelerate the coarsening process, leading to the coarsening exponent of 1 in the power law describing the increase in domain size. Similar exponents have been measured during steady shear for the coarsening of domains parallel to the flow direction,8,9,11-13 implying that shear flow increases the likelihood of droplet coalescence. The domain size perpendicular to the flow direction is often found to be constant, however.8,9,12,13 Under certain conditions, the opposing effects of droplet breakup and coalescence can lead to the formation of dynamic steady-state morphologies and no coarsening of the microstructure is observed at all.6 Although a variety of biopolymer systems have been studied during quiescent phase separation,4,5,18-20 quantitative studies of the phase-separation behavior of these materials during shear are limited. The aim of the current study, therefore, was to use small-angle light scattering (SALS) to quantitatively investigate the effect of steady shear on the morphology development and coarsening kinetics in a liquid biopolymer mixture undergoing spinodal decomposition. The results will be contrasted with those obtained when the mixture undergoes spinodal decomposition under quiescent conditions. Experimental Section Materials and Sample Preparation. A mixture of gelatin and maltodextrin was used in the following experiments. The gelatin was a lime-treated gelatin (LH1e) supplied by SKW. Through the use of size-exclusion chromatography (SEC) coupled with light scattering, the weight- and number-average molecular weights were found to be Mw ) 146 000 and Mn ) 83 300. The maltodextrin was a DE2 grade (SA2) supplied by Avebe. SEC measurements revealed a broad molecular weight distribution, although the DE value allows an estimate of 9000 g mol-1 to be made for the value of Mn. Moisture contents of 12.4% and 10.0% for gelatin and maltodextrin, respectively, were taken into account when making solutions. To prevent biological degradation, all solutions were used only on the day that they were made. Care was taken throughout sample preparation to prevent dehydration. Solutions were prepared as follows: SA2 powder was first dispersed in cold water then mildly stirred at 95 °C for 30 min to form a solution. Gelatin solutions were prepared similarly but were not heated above 60 °C to prevent any thermal degradation. Before heating, sodium chloride was added to the gelatin solution to increase the total ionic strength of the final mixed solution to 0.3 M, (this figure includes the ionic contributions from the SA2 and the gelatin). After 30 min at 95 °C, the SA2 was cooled to 60 °C for 5 min and then mixed with the gelatin solution to form a final mixture containing 5 wt % gelatin and 3 wt % maltodextrin. The mix was then kept at 60 °C and stirred continuously. After mixing and while still at 60 °C, the solution had a pH (measured using a Jenway 3071 automatic temperature compensated pH meter) of approximately 5.6.

Biomacromolecules, Vol. 3, No. 6, 2002 1209

A pinch of rhodamine-B powder was added to the samples for use in the confocal microscopy experiments to provide additional contrast between the phases. Simultaneous Small-Angle Light Scattering and Steady Shear. Shear experiments were performed in a CSS450 shearing stage manufactured by Linkam Scientific Instruments Ltd. using a parallel-plate geometry. This device consisted of two parallel glass plates that were heated using Peltier devices and controllably cooled using a flow of liquid nitrogen. Holes, with a diameter of 3 mm, were placed in the Peltier elements to allow light to pass through the sample held between the glass plates. The top plate was fixed, and the bottom plate was rotated to obtain steady shear at rates between 0.1 and 100 s-1 of the liquid between the plates. It should be noted that, because the shear was performed in a parallel plate geometry, the shear rate actually varied slightly across the region of measurement. The measured data therefore represented an average behavior. Two types of experiment were performed. In the first, the mixture was sheared during cooling (at a rate of 30 °C/min) from 65 to 32 °C, at which temperature phase separation occurred. In the second, shear was started 20 min after the mixture was cooled from 65 to 32 °C, allowing the initial stages of phase separation and coarsening to occur under quiescent conditions. Small-angle light scattering patterns were obtained using a setup similar to one described previously5 but with a 0.5 mW helium-neon laser and a fast digital imager CCD camera supplied by Photonic Science. Patterns were recorded at 30 s intervals, and the scattering pattern from the liquid at 60 °C was subtracted from all subsequent patterns to remove any background scattering. To reduce the level of noise, which was significant for any individual SALS pattern, the data was averaged over several frames and on both sides of the two-dimensional scattering pattern. In addition, to provide a comparison of phase separation during shear with quiescent phase separation, a sequence of patterns was recorded for a sample that was not sheared. The intensity for the quiescently cooled sample was radially averaged to yield a plot of scattered intensity versus q vector. Confocal Laser Scanning Microscopy. Samples for study by confocal microscopy were made in the same way as for SALS. The sample was placed on a temperature-controlled stage (Linkam THMS600). The temperature profile used was the same as that for the SALS experiments, and the sample was held at 60 °C prior to beginning the profile. Micrographs of the mixtures were acquired using a Biorad MRC 600 CLSM. A 488 nm argon laser excited the rhodamine B, which provided the contrast between the gelatin-rich and maltodextrin-rich phases. Micrographs were captured using COMOS software. A ×20 lens was used with a digital zoom of ×1, ×5, or ×10. Results Quiescent Phase Separation. Confocal micrographs of the sample after cooling to 32 °C are shown in Figure 1. When the sample reached the quench temperature, phase separation in the form of a droplet morphology was observed

1210

Biomacromolecules, Vol. 3, No. 6, 2002

Butler

Figure 1. Confocal images showing coarsening droplets in the sample cooled to 32 °C and allowed to phase separate without shear. Images a-d were obtained after times of 180, 1080, 1620, and 1920 s after the onset of phase separation, respectively. The image width is 72 µm.

Figure 2. Sequence of radially averaged SALS patterns for the quiescently cooled sample. The peak height increases with time (as indicated by the arrow).

immediately. The droplets were relatively mobile and became significantly larger during the course of the experiment. After 19 min, which was the time when shear was imposed on the systems that were allowed to phase separate quiescently, the droplet size was about 4 µm. The droplet size was less than the droplet separation for the duration of the experiment (i.e., up to at least 2000 s). A sequence of radially averaged SALS patterns obtained for the sample cooled to 32 °C without shear is shown in Figure 2. At the onset of phase separation, a broad peak appeared in the SALS pattern that subsequently moved to lower scattering vectors with time as the structure coarsened. The evolution of the peak position is shown in Figure 3. The coarsening rate gradually increased until, after a time between 1100 and 1500 s after phase separation began, a final coarsening exponent, which was slightly less than 1/3, was reached. The evolution of the peak height, which increased rapidly after phase separation began and then moved more slowly toward a roughly constant value, is shown in Figure 4. The time at which the peak height reached a constant value was approximately the same as the time at which the final, highest coarsening exponent was measured.

Figure 3. Evolution of the peak position for the quiescently cooled sample. The theoretical exponent of 1/3 for coalescing droplets with a size/separation ratio of 1 is also shown.

Figure 4. Evolution of the peak height for the quiescently cooled sample.

Simultaneous Steady Shear and Phase Separation. Figure 5 shows SALS patterns measured from samples that were sheared while they were cooled to 32 °C. At the lowest shear rate, 0.1 s-1, the SALS pattern remained isotropic at all times. At the higher shear rates, however, the scattering pattern was anisotropic from the onset of phase separation


Figure 5. 2D SALS patterns for samples sheared while phase separation occurred at the shear rates shown. The arrow indicates the shear direction. The patterns shown were measured 1560 s after the onset of phase separation.

with the intensity concentrated perpendicular to the shear direction. The degree of anisotropy of the scattering patterns increased with increasing shear rate, becoming a highly oriented streak at the highest rate of 100 s-1. Figures 6-9 show the intensity of scattered light perpendicular to the shear direction for samples sheared at 0.1, 1, 10, and 100 s-1, respectively. Porod (double logarithmic) plots of the scattering intensity are shown on the right-hand side of each figure for the shear rates of 0.1, 10, and 100 s-1. For the shear rates above and below, but not including, 1 s-1, the intensity distribution consisted of a broad peak. The width of the peak increased with increasing shear rate. At 0.1 s-1, the peak continually moved toward lower scattering vectors, whereas at 10 s-1, the scattering pattern rapidly attained a constant appearance. At 100 s-1, the peak was intially very broad but became sharper with time as there was a reduction in the amount of light scattered at high scattering vectors. Linear regions were identified in the Porod plots in the limit of high scattering vectors for all of these rates with slopes, obtained by fitting the Porod region of the scattering function, of -4.1, -5.8, and -1.5 (with an uncertainty of around 20%) for the shear rates of 0.1, 10, and 100 s-1, respectively. When these slopes were extrapolated back to the ordinate, the value of the intercept decreased for the shear rate of 0.1 s-1, remained increased slightly, then remained constant for the shear rate of 10 s-1, and passed through a maximum for the shear rate of 100 s-1. The appearance of the scattered intensity perpendicular to the shear direction was, however, totally different for the sample sheared at 1 s-1. In this case, the scattered light was concentrated much more intensely at low scattering vectors


toward a maximum value at which the scattering vector equaled zero. There were, however, subsidiary maxima the intensity of which diminished rapidly with increasing scattering vector. Owing to the lower integration time necessitated by the much higher concentration of intensity near the beam stop at this shear rate, the data were significantly noisier than those for the other rates studied, and it was therefore impossible to accurately and quantitatively follow the detailed evolution of these subsidiary maxima. They did, however, appear to become better defined as time proceeded and possibly move toward slightly higher scattering vectors. Figure 10 shows the evolution of the peak position perpendicular to the shear direction for the samples sheared at 0.1, 10, and 100 s-1. In the case in which the particles remained spherical, that is, those subjected to a shear rate of 0.1 s-1, the coarsening rate was initially much greater than that in the nonsheared case and the coarsening law possessed an exponent of about -1/3. After a period during which the coarsening rate slowed to about the same value as that for the nonsheared case, there was a transition to a much more rapid coarsening regime in which the power law exponent was approximately -1. This transition occurred at about the same time as the transition to the higher coarsening exponent was observed in the nonsheared sample, that is, between 1100 and 1500 s. At 10 s-1, the peak position perpendicular to the shear direction increased slightly and then remained relatively invariant. At 100 s-1, the peak position increased slightly and then rapidly moved to lower scattering vectors with a power-law coarsening exponent of -1 before remaining relatively constant. The onset of the rapid coarsening regime corresponded to the time at which the scattering pattern became much sharper and the peak became better defined. Phase Separation Followed by Shear. The scattering patterns from the samples that were subjected to shear 19 min after the onset of phase separation were isotropic, as expected, until shear was started. Shown in Figure 11, the SALS pattern remained isotropic at the shear rate of 0.1 s-1, but for the higher shear rates, it immediately became anisotropic. In the latter cases, the SALS patterns were similar but not immediately identical, as shown by analysis of the intensity distribution perpendicular to the shear

Figure 6. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 0.1 s-1. The scattering patterns shown were obtained after 120, 810, 1110, and 1710 s.

1212


Figure 7. Scattered intensity perpendicular to the shear direction for the shear rate of 1 s-1 (at time 1290 s).

direction to those measured in the samples in which phase separation and shear occurred simultaneously. In fact, when shear was first imposed, transient structures were formed that lasted for a few seconds, as shown by the sequence of SALS patterns shown in Figure 12. In this case, which was for the application of a shear rate of 10 s-1 to an initially isotropic system, the initial transient SALS patterns (frames 3-5 in Figure 12) were characteristic of those from a steady-state structure sheared at a rate 10

Butler

times larger than the applied one. In the subsequent few seconds (frames 6-11), the streak became fainter and less extended and the azimuthal spread of intensity around the direction perpendicular to the shear rate began to increase. Finally, the SALS pattern acquired its final steady-state appearance (frames 9-12). The transient, highly elongated structures formed by stepping up the shear rate therefore lasted for times on the order of a few seconds. Figures 13-16 show the evolution of the intensity perpendicular to the shear direction for shear rates of 0.1, 1, 10, and 100 s-1, respectively. As for the samples in which shear and phase separation occurred simultaneously, there was one broad peak at nonzero scattering vectors in the intensity distribution for shear rates of 0.1, 10, and 100 s-1 but a series of subsidiary maxima leading to the maximum intensity at the origin for the shear rate of 1 s-1. As for the samples in which shear was applied at the same time as phase separation was initiated, linear regions were identified in the Porod plots in the limit of high scattering vectors. Before shear was applied, the Porod region had a slope of approximately -4 (the value obtained from curve fitting the Porod tail was -3.9 with an uncertainty of around 20%; the precise values after shearing are quoted in the following results after the shear rate; they also had the same degree of uncertainty). When the system was sheared, the gradient of the Porod region remained about -4 for the shear rates of

Figure 8. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 10 s-1. The scattering patterns shown were obtained after 30, 60, 240, and 3210 s.

Figure 9. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 10 s-1. The scattering patterns shown were obtained after 120, 510, 2190, and 3210 s.


Figure 10. Evolution of the peak position perpendicular to the shear direction for shear rates of 0.1, 10, and 100 s-1.

Figure 11. 2D SALS patterns for samples sheared 19 min after phase separation occurred at the shear rates shown. The arrow indicates the shear direction. The patterns shown were measured 1560 s after the onset of phase separation and 420 s after shear was applied.


0.1 (-4.2) and 1 s-1 (-4.1) but became about -6 for the shear rate of 10 s-1 (-5.8) and about -2 for the shear rate of 100 s-1 (-2.2). These values were very similar to those measured for the samples that were simultaneously sheared and phase-separated. Figure 17 shows the evolution of the peak position perpendicular to the shear direction for the shear rates of 0.1, 10, and 100 s-1 with the time at which shear was initiated marked by an arrow. In all cases, the application of shear caused an immediate increase in the coarsening rate and the coarsening exponent increased to a value approximately equal to -1. In the 0.1 s-1 sample, the coarsening exponent decreased to -1/3 after a while but eventually returned to the higher value of -1. For the shear rate of 10 s-1, the peak attained a steady position as the scattering pattern subsequently remained unchanged. At 100 s-1, the peak moved back to slightly higher scattering vectors toward the value that was measured when shear and phase separation occurred simultaneously. Relaxation of Sheared Structures. Figure 18 shows a sequence of SALS patterns acquired upon instantaneously stopping the shear in a sample that was sheared at 100 s-1. The highly elongated streak perpendicular to the shear direction gradually disappeared and was replaced by an isotropic scattering pattern. Although the isotropic scattering developed very quickly (within 1 s of the cessation of shear), the elongated streak took much longer (tens of seconds) to disappear. It coexisted with the isotropic scattering for up to two minutes after shear was stopped, becoming gradually fainter and less extended. Discussion

Figure 12. Transient SALS patterns upon step-up from 0 to 10 s-1. The patterns were collected at 0.5 s intervals. The arrow indicates the shear direction.

Quiescent Phase Separation. The presence of a peak in the SALS pattern that moved to lower scattering angles from the onset of demixing shows that in the gelatin/maltodextrin mixture studied in the current experiment phase separation occurred via the spinodal decomposition mechanism,21 forming a characteristic length scale. Previous work5 has shown that the binodal temperature of the current system was around 40 °C and the composition can be considered to be off-critical because a droplet morphology was formed.

Figure 13. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 0.1 s-1. The scattering patterns shown were obtained after 690, 1350, 2210, and 5400 s. The data at 690 s were obtained before shear was applied.

1214


Butler

Figure 14. Scattered intensity perpendicular to the shear direction for the shear rate of 1 s-1 (at time 1290 s after start of shear).

Figure 15. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 10 s-1. The scattering patterns shown were obtained after 690, 1350, 2210, and 5400 s. The data at 690 s were obtained before shear was applied.

Figure 16. Evolution of the averaged scattered intensity perpendicular to the shear direction for the shear rate of 100 s-1. The scattering patterns shown were obtained after 690, 2210, and 5400 s. The data at 690 s were obtained before shear was applied.

When droplet growth occurs via coalescence, a power law describes the coarsening law with an exponent that depends on the ratio of particle separation to particle size. When the particle size and separation are approximately the same, this exponent equals -1/3,22 but when the size is significantly smaller than the separation, much lower exponents are expected.23 The gradual increase in coarsening exponent that was measured was therefore a result of the small droplet size-to-separation ratio. Because this ratio did not reach 1

over the time scale of the experiment, the exponent of -1/3 was not reached. The Influence of Shear on Phase Separation and Coarsening Kinetics. Overview. The presence of the peak in the SALS pattern for the shear rates of 0.1, 10, and 100 s-1 shows that shear did not alter the phase-separation mechanism, which still occurred via spinodal decomposition. The major effect of shear was on the anisotropy of the structure, the coarsening rates, and, most dramatically, the


Figure 17. Evolution of the peak position perpendicular to the shear direction for shear rates of 0.1, 10, and 100 s-1. The vertical arrow indicates the time at which shear was applied to the system.

Figure 18. Transient SALS patterns upon step-down from 100 to 0 s-1. The times indicated are the times after which shear was stopped. The arrow indicates the shear direction.

formation of a “special” microstructure at the shear rate of 1 s-1. It is well-known that shear promotes coalescence, induces anisotropy, and, at a certain shear rate, causes elongated droplets to break up into smaller ones.24,25 The results will be discussed in the context of these phenomena in order of ascending shear rate. 0.1 s-1. The higher coarsening exponents that were measured in the samples sheared at 0.1 s-1 compared to the quiescently phase-separated sample best illustrated the promotion of coalescence by shear. Indeed, the exponent of -1, which was measured immediately from the time that shear was applied in the system that was first allowed to phase separate quiescently, is expected in a system in which hydrodynamic flow is the dominant factor responsible for the particle motion that leads to coalescence.26 These flow conditions are likely to be a transient effect resulting from the sudden onset of shear, however, because in that sample the coarsening exponent returned to -1/3 after a while and in the sample that was sheared during phase separation the exponent was -1/3 from the outset. That dramatic transient effects can occur was demonstrated by the step-up experiment that showed the brief formation of structures characteristic of a steady-state shear rate 1 order of magnitude higher than the actually imposed shear rate. The final stage of hydrodynamic coarsening once a certain value of qm had been reached in both cases probably occurred because the droplets had become large enough for convective flow fields established inside the droplets to become coupled to the motion of the matrix.5 It is therefore expected that, regardless of the time at which shear is imposed on the system, the rapid hydrodynamic coarsening mechanism will operate and ultimately lead to bulk phase separation. The slope of -4 that was measured in the Porod plots is consistent with an off-critical droplet morphology with sharp


interfaces.27 It indicates that even when the concentration fluctuations that led to phase separation were established during shear the phase separation kinetics were fast enough to lead to the formation of well-defined droplets. In a more viscous system, however, the lifetime of the concentration fluctuations would be much longer. A system could therefore be envisaged in which shear would have a large influence on the phase separation kinetics, even at relatively low rates. Because the shear rate of 0.1 s-1 was not high enough to cause droplet breakup, the only effect of shear was to promote coalescence. Ultimately, sedimentation leads to bulk phase separation in this situation, and the final morphology will be independent of the time at which shear was imposed on the system. At intermediate times, however, the droplet size will depend on when shear was started relative to the time at which the sample phase separated. 1 s-1. The scattering patterns that were observed at the shear rate of 1 s-1 were similar to the those measured in a previous study on a sheared, off-critical polymer mixture,28 in which they were considered to be reminiscent of the form factors of single elongated spheroids and were fitted to functions that described such structures. In that case, up to 13 subsidiary maxima were recorded perpendicular to the shear direction. The high quality of those results allowed them to be analyzed quantitatively, and it was concluded that the particle size was very close to the critical size for droplet breakup at the shear rates used. Consequently, the rapid repetition of droplet breakup and coalescence led to a particle size distribution that was extremely narrow. On the basis of the findings of that study and owing to the similarity between the present results and the previous study, it is proposed that a similar situation was present in the current experiment, that is, at 1 s-1 a narrow particle size distribution was formed. Through the use of literature values for the surface tension in the gelatin/maltodextrin system (50 µN m-1),29 the critical shear rate for breakup of droplets with the sizes measured in the current experiment by confocal microscopy was roughly estimated as 4 s-1, which is not too dissimilar to 1 s-1. Detailed structural information from fitting the data to structural models could not be obtained, however, owing to the lower quality of the data. 10 s-1. The structure that was formed at this shear rate was probably the most complicated one. The transition of the Porod slope from -4 to -6 at the time when shear was imposed on the system that was first allowed to phase separate quiescently indicates that a bicontinuous microstructure with sharp interfaces was formed.27 The combination of droplet elongation with the increased chance of coalescence parallel to the shear direction can lead to the formation of elongated percolated structures,28 and this is proposed as the reason for the observed scattering patterns measured at the shear rate of 10 s-1. This could not happen at 1 s-1 because in that case the droplets broke before they could lengthen to such an extent that a percolated structure could be formed. The relatively constant SALS patterns show that, compared to a percolated quiescent system in which the bicontinuous morphology will eventually break up to form droplets, the shear-induced percolated microstructure is relatively stable. Presumably, when breakup does occur the droplets thereby formed are elongated and rejoin the percolated network. That breakup and coalescence did not

1216


become fully equilibrated was shown by the gradual, albeit slow, change in the peak position for both samples. Furthermore, the increase in peak position when phase separation and shear occurred simultaneously suggests that the rate of breakup was greater than the coalescence rate, whereas the converse case for the sample that initially phase separated quiescently suggests that coalescence was the dominant process in that situation. Therefore, it would appear that the precise details of percolated structures that form by the action of shear depend on the shear history of the sample. Further work is required to elucidate the mechanism by which these structures form and evolve during shear. 100 s-1. At the highest shear rate of 100 s-1, the elongated streak that was present in the SALS pattern indicated the presence of highly elongated particles. These patterns are usually indicative of the formation of string phases that form at shear rates high enough to suppress the surface undulations that normally lead to droplet breakup.16,17 The unstable nature of these structures in the absence of shear was demonstrated by the relaxation of the elongated streak toward an isotropic scattering pattern when the shear was stopped, caused by the breakup of the elongated structures to form many small, spherical droplets. With the use of an equation for the characteristic relaxation time,31 τ (τ ) ηR/(0.465Γ), where η is the matrix phase viscosity, 0.1 Pa s; R is the fibril size, ∼100 µm; and Γ is the surface tension, 50 µN m-1), the relaxation time is calculated to be approximately 0.5 s, which is in general agreement with the rapid relaxation of the scattering pattern from the elongated streak to being isotropic. The slope of -2 that was measured in the limit of high scattering vectors may mean that the Porod regime was not attained, which could occur if the particle size distribution was very wide.13 It may, however, indicate that the interface between the phases was relatively diffuse. Because the wavelength of the radiation used to probe the structure was 633 nm, this implies an interfacial width in excess of 50 nm (∼λ/10). High shear rates are expected to promote mixing of the phases that is preceded by a gradual broadening of the interfacial profile between the phases.6 Possibly, 100 s-1 was not high enough to cause complete mixing but was sufficient to cause some mixing that led to an increase in the interfacial width. Further experiments are necessary to establish the rate at which shear-induced mixing occurs in this system. The convergence of the peak position to a common value, regardless of when shear started, suggests that, as for the droplet case, the final fibrillar microstructure was determined predominantly by the shear rate and not the initial phase-separation conditions. Conclusions Spinodal decomposition was the operative phase-separation mechanism in the gelatin/maltodextrin mixture studied in both quiescent conditions and those when shear was applied. Small-angle light scattering enabled shear-induced structural changes to be measured in situ. A characteristic length scale was retained in the microstructure regardless of the shear conditions. When the shear rate was too low (0.1 s-1) to cause droplet breakup, its main effect was to accelerate coarsening by enhancing the coalescence rate of the droplets. Regardless of the time at which shear was

Butler

started, coarsening eventually led to bulk phase separation in this case. A transient regime of rapid hydrodynamic coarsening was present when shear was initiated after phase separation and at late times in all cases once the droplets attained a certain size. When the shear rate was at a value (1 s-1) at which the droplet size formed by the spinodal decomposition mechanism was the critical size for droplet breakup, then the rapid repetition of breakup and coarsening formed a microstructure consisting of elongated droplets with a narrow size distribution. When the shear rate enabled droplets to extend to such an extent that a percolated structure could form (10 s-1), then the structure was relatively stable and changed very slowly over time. At very high shear rates (100 s-1), droplet breakup was suppressed and a highly fibrillar morphology formed that was stable only while the system was under shear. Cessation of shear at high rates led to fiber breakup and the formation of many small droplets. For a given shear rate, the final microstructure appeared to be independent of the time that shear was started when the structure consisted of discrete droplets or fibers. When a percolated structure could form, however, the shear history appeared to be important. Acknowledgment. Allan Clark, Ian Norton, and Bettina Wolf of Unilever R&D, Colworth, are thanked for their interest in and assistance with the work performed in this paper. References and Notes (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30)

Nose, T. Phase Transitions 1987, 8, 245. Hashimoto, T. Phase Transitions 1988, 12, 47. Bansil, R. J. Phys. IV 1993, 3, 225. Tromp, R. H.; Rennie, A. R.; Jones, R. A. L. Macromolecules 1995, 28, 4129. Butler, M. F.; Heppenstall-Butler, M. Biomacromolecules 2001, 2, 812. Onuki, A. J. Phys.: Condens. Matter 1997, 9, 6119. Criado-Sancho, M.; Casas-Va´zquez, J.; Jou, D. Phys. ReV. E 1997, 56, 1887. Chan, C. K.; Perrot, F.; Beysens, D. Phys. ReV. A 1991, 43, 1826. Berthier, L. Phys. ReV. E 2001, 63, 051503. La¨uger, J.; Laubner, C.; Gronski, W. Phys. ReV. Lett. 1995, 75, 3576. Gronski, W.; La¨uger, J.; Laubner, C. J. Mol. Struct. 1996, 383, 23. Chan, C. K.; Perrot, F.; Beysens, D. Phys. ReV. Lett. 1988, 61, 412. Qui, F.; Ding, J.; Yang, Y. Phys. ReV. E 1998, 58, R1230. Chen, Z. J.; Shaw, M. T.; Weiss, R. A. Macromolecules 1995, 28, 648. Hong, Z. J.; Shaw, M. T.; Weiss, R. A. Polymer 2000, 41, 5895. Hashimoto, T.; Matsuzaka, K.; Moses, E. Phys. ReV. Lett. 1995, 74, 126. Hobbie, E.; Kim, S.; Han, C. C. Phys. ReV. E 1996, 54, R5909. Loren, N.; Hermansson, A.-M. Int. J. Biol. Macromol. 2000, 27, 249. Emanuele, A.; Distefano, L.; Giacomezza, D.; Trapanese, M.; PalmaVitorelli, M. B.; Palma, M. U. Biopolymers 1991, 31, 859. Leone, M.; Sciortino, F.; Migliore, M.; Fornili, S. L.; Palma-Vitorelli, M. B. Biopolymers 1987, 26, 743. Strobl, G. In The Phyics of Polymers; Springer-Verlag: Berlin, 1996. Tanaka, H. J. Phys. D 2000, 12, 207. Tanaka, H.; Yokokawa, T.; Abe, H.; Hayashi, T.; Nishi, T. Phys. ReV. Lett. 1990, 65, 3136. Rallison, J. M. Annu. ReV. Fluid Mech. 1984, 16, 45. Stone, H. A. Annu. ReV. Fluid Mech. 1994, 26, 65. Siggia, E. D. Phys. ReV. A 1979, 20, 595. Furukawa, H. Phys. A 1984, 123, 497. Hashimoto, T.; Matsuzaka, K.; Fujioka, K. J. Chem. Phys. 1998, 108, 6963. Stokes, J. R.; Wolf, B.; Frith, W. J. J. Rheol. 2001, 45, 1173. Marrucci, G.; di Santo, S. Polym. Eng. Sci. 1997, 37, 1540.

BM0255645

Maltodextrin Mixture Studied

Recommend Documents