Grenzflächenspannung und Morphologie fließender zweiphasiger Polymermischungen

In der vorliegenden Arbeit wurde die Morphologie von zweiphasigen Polymermischungen unter Scherung in situ mit Hilfe einer Kombination aus optischer Scherzelle, Durchlichtmikroskop und computergestützten CCD-Kamera untersucht. Als Modellblends dienten die unverträglichen, bei Raumtemperatur flüssigen Polymersysteme Polyisobutylen (PIB)/Polydimethylsiloxan (PDMS) (I) und Poly(dimethyl-co-methylphenyl)siloxan/PDMS (II). Alle Komponenten verhalten sich bei den verwendeten Scherraten newtonisch.Eine der wichtigsten Einflussgrößen für die Blendmorphologie ist die Grenzflächenspannung gamma 12. Sie wurde für I und II mit Hilfe der Methode der Tropfenrelaxation (dynamisch) als Funktion der Zeit bestimmt. Diese Methode erlaubt die Messung von gamma 12 für Tropfen der Phase A in B sowie von Tropfen B in A. Bei der Methode des hängenden Tropfens (statisch) muss der Tropfen aus der Phase mit der höheren Dichte bestehen. Wo der Vergleich der beiden Methoden möglich ist, stimmen die Ergebnisse für beide Systeme sehr gut überein. Bei II sind die aus der Tropfenrelaxation erhaltenen gamma 12-Werte der beiden komplementären Zusammensetzungen im Rahmen des Fehlers gleich, bei I zeigt ein PIB-Tropfen in PDMS einen um 40 % niedrigeren Wert als ein PDMS-Tropfen in PIB, dies wird auf die Diffusion von kurzkettigen Anteilen des PDMS in die Grenzschicht zurückgeführt. Die Grenzflächenspannung hängt also unter Umständen auch bei binären Systemen deutlich von der Zusammensetzung ab.Für II wurde die Blendmorphologie über den gesamten Zusammensetzungsbereich untersucht. Die häufig beobachteten cokontinuierlichen Strukturen treten bei keiner Zusammensetzung auf. Die Phaseninversion erkennt man in einer sprunghaften Änderung der Tropfengröße zwischen phiPDMS <= 0,400 und 0,500; zudem lässt sich die Zeitabhängigkeit der Radien durch Auftragung gegen das Produkt aus der Deformation und dem Quadrat des Volumenbruchs der Tropfenphase für 0 <= phiPDMS <= 0,400 sowie 0,500 <= phiPDMS <= 1 normieren. Für I und II wurde die Morphologieentwicklung bei 25 °C nach Vorscherung bei 100 bzw. 50 s-1 und anschließendem Sprung der Scherrate auf deutlich niedrigere Werte als Funktion der Zeit verfolgt. Hierbei erhält man bei genügend langer Messdauer (mindestens 200 000-300 000 Schereinheiten) konstante Tropfengrößen. Zum einen handelt es sich dabei um pseudo-stationäre Werte, die nur durch Koaleszenz bestimmt sind, zum anderen um echte stationäre Radien, die durch gleichzeitig ablaufende Koaleszenz und Zerteilung entstehen. Für I liegen die stationären Mittelwerte auf der Zerteilungskurve, für II hingegen auf der Koaleszenzkurve.Der Einfluss eines grenzflächenwirksamen Additivs wurde anhand von I durch Zugabe des Blockcopolymer PIB-b-PDMS zu PIB untersucht. Der Vergleich des zeitlichen Verlaufs von gamma 12 mit der Morphologieentwicklung zeigt, dass das Additiv eine Stabilisierung der feinen Tropfen/Matrix-Struktur des Blends durch Hinderung der Koaleszenz und nicht durch Reduktion der Grenzflächenspannung bewirkt.

Patterns in and kinetics of phase separation in binary mixtures

This work focused mainly on two aspects of kinetics of phase separation in binary mixtures. In the first part, we studied the interplay of hydrodynamics and the phase separation of binary mixtures. A considerably flat container (a laterally extended geometry), at an aspect ratio of 14:1 (diameter: height) was chosen, so that any hydrodynamic instabilities, if they arise, could be tracked. Two binary mixtures were studied. One was a mixture of methanol and hexane, doped with 5% ethanol, which phase separated under cooling. The second was a mixture of butoxyethanol and water, doped with 2% decane, which phase separated under heating. The dopants were added to bring down the phase transition temperature around room temperature.rnrnAlthough much work has been done already on classical hydrodynamic instabilities, not much has been done in the understanding of the coupling between phase separation and hydrodynamic instabilities. This work aimed at understanding the influence of phase separation in initiating any hydrodynamic instability, and also vice versa. Another aim was to understand the influence of the applied temperature protocol on the emergence of patterns characteristic to hydrodynamic instabilities. rnrnOn slowly cooling the system continuously, at specific cooling rates, patterns were observed in the first mixture, at the start of phase separation. They resembled the patterns observed in classical Rayleigh-Bénard instability, which arises when a liquid continuously is heated from below. To suppress this classical convection, the cooling setup was tuned such that the lower side of the sample always remained cooler by a few millikelvins, relative to the top. We found that the nature of patterns changed with different cooling rates, with stable patterns appearing for a specific cooling rate (1K/h). On the basis of the cooling protocol, we estimated a modified Rayleigh number for our system. We found that the estimated modified Rayleigh number is near the critical value for instability, for cooling rates between 0.5K/h and 1K/h. This is consistent with our experimental findings. rnrnThe origin of the patterns, in spite of the lower side being relatively colder with respect to the top, points to two possible reasons. 1) During phase separation droplets of either phases are formed, which releases a latent heat. Our microcalorimetry measurements show that the rise in temperature during the first phase separation is in the order of 10-20millikelvins, which in some cases is enough to reverse the applied temperature bias. Thus phase separation in itself initiates a hydrodynamic instability. 2) The second reason comes from the cooling protocol itself. The sample was cooled from above and below. At sufficiently high cooling rates, there are situations where the interior of the sample is relatively hotter than both top and bottom of the sample. This is sufficient to create an instability within the cell. Our experiments at higher cooling rates (5K/h and above) show complex patterns, which hints that there is enough convection even before phase separation occurs. Infact, theoretical work done by Dr.Hayase show that patterns could arise in a system without latent heat, with symmetrical cooling from top and bottom. The simulations also show that the patterns do not span the entire height of the sample cell. This is again consistent with the cell sizes measured in our experiment.rnrnThe second mixture also showed patterns at specific heating rates, when it was continuously heated inducing phase separation. In this case though, the sample was turbid for a long time until patterns appeared. A meniscus was most probably formed before the patterns emerged. We attribute the reason of patterns in this case to Marangoni convection, which is present in systems with an interface, where local differences in surface tension give rise to an instability. Our estimates for the Rayleigh number also show a significantly lower number than that's required for RB-type instability.rnrnIn the first part of the work, therefore, we identify two different kinds of hydrodynamic instabilities in two different mixtures. Both are observed during, or after the first phase separation. Our patterns compare with the classical convection patterns, but here the origins are from phase separation and the cooling protocol.rnrnIn the second part of the work, we focused on the kinetics of phase separation in a polymer solution (polystyrene and methylcyclohexane), which is cooled continuously far down into the two phase region. Oscillations in turbidity, denoting material exchange between the phases are seen. Three processes contribute to the phase separation: Nucleation of droplets, their growth and coalescence, and their subsequent sedimentation. Experiments in low molecular binary mixtures had led to models of oscillation [43] which considered sedimentation time scales much faster than the time scales of nucleation and growth. The size and shape of the sample therefore did not matter in such situations. The oscillations in turbidity were volume-dominated. The present work aimed at understanding the influence of sedimentation time scales for polymer mixtures. Three heights of the sample with same composition were studied side by side. We found that periods increased with the sample height, thus showing that sedimentation time determines the period of oscillations in the polymer solutions. We experimented with different cooling rates and different compositions of the mixture, and we found that periods are still determined by the sample height, and therefore by sedimentation time. rnrnWe also see that turbidity emerges in two ways; either from the interface, or throughout the sample. We suggest that oscillations starting from the interface are due to satellite droplets that are formed on droplet coalescence at the interface. These satellite droplets are then advected to the top of the sample, and they grow, coalesce and sediment. This type of an oscillation wouldn't require the system to pass the energy barrier required for homogenous nucleation throughout the sample. This mechanism would work best in sample where the droplets could be effectively advected throughout the sample. In our experiments, we see more interface dominated oscillations in the smaller cells and lower cooling rates, where droplet advection is favourable. In larger samples and higher cooling rates, we mostly see that the whole sample becomes turbid homogenously, which requires the system to pass the energy barrier for homogenous nucleation.rnrnOscillations, in principle, occur since the system needs to pass an energy barrier for nucleation. The height of the barrier decreases with increasing supersaturation, which in turn is from the temperature ramp applied. This gives rise to a period where the system is clear, in between the turbid periods. At certain specific cooling rates, the system can follow a path such that the start of a turbid period coincides with the vanishing of the last turbid period, thus eliminating the clear periods. This means suppressions of oscillations altogether. In fact we experimentally present a case where, at a certain cooling rate, oscillations indeed vanish. rnrnThus we find through this work that the kinetics of phase separation in polymer solution is different from that of a low molecular system; sedimentation time scales become relevant, and therefore so does the shape and size of the sample. The role of interface in initiating turbid periods also become much more prominent in this system compared to that in low molecular mixtures.rnrnIn summary, some fundamental properties in the kinetics of phase separation in binary mixtures were studied. While the first part of the work described the close interplay of the first phase separation with hydrodynamic instabilities, the second part investigated the nature and determining factors of oscillations, when the system was cooled deep into the two phase region. Both cases show how the geometry of the cell can affect the kinetics of phase separation. This study leads to further fundamental understandings of the factors contributing to the kinetics of phase separation, and to the understandings of what can be controlled and tuned in practical cases. rn