- Email: [email protected]
Pluronic interfaces studied by a spinning drop tensiometer
Journal of Colloid and Interface Science 322 (2008) 669–674
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Note
Unusual behavior of PEG/PPG/Pluronic interfaces studied by a spinning drop tensiometer Jeffrey D. Martin, Sachin S. Velankar ∗ Department of Chemical Engineering, University of Pittsburgh, Pittsburgh, PA 15261, USA
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 24 October 2007 Accepted 28 March 2008 Available online 8 April 2008
The effects of surfactants on the interfacial tension driven retraction of elongated drops were studied in a spinning drop tensiometer. Experiments were conducted on polypropylene glycol (PPG) drops suspended in polyethylene glycol (PEG), with Pluronic block copolymers as surfactants. Two unusual observations are reported here. In the first, initially-elongated drops generated at high rotational speed were allowed to retract by reducing the rotational speed. Pluronic-laden drops would not retract completely, but would instead maintain strongly nonspherical shapes indefinitely. We attribute such “nonretraction” to an interfacial yield stress induced by the Pluronic surfactant. In the second, drops being heated while spinning at a constant speed would elongate sharply at some temperature, and subsequently breakup. Such “autoextension” and breakup indicate complex nonmonotonic changes in interfacial tension with time during heating. We propose that autoextension occurs because at low temperature, interfacially-adsorbed surfactant is crystallized and hence trapped at the interface at a concentration far above equilibrium. © 2008 Elsevier Inc. All rights reserved.
Keywords: Interfacial tension Interfacial viscoelasticity Pluronic Interfacial crystallization Interfacial yield stress Nonspherical drops
1. Introduction A cylindrical liquid drop suspended in a quiescent fluid can either retract back into a spherical shape or can break up by a Rayleigh instability into a series of drops [1]. Generally, drops with modest aspect ratio (< ∼15) retract, whereas very long drops (aspect ratio > ∼15) breakup. We are presently experimentally investigating the effect of surfactant on both these interfacial tensiondriven processes. We chose for these studies an experimental system composed of polypropylene glycol (PPG) drops suspended in a oplyethylene glycol (PEG) matrix with PPO–PEO block copolymer surfactants (mostly from the Pluronic family) at the interface. This Note reports on some remarkable observations about the dynamics of this system made during the course of our research. Due to its easy and cheap availability, transparency, and chemical stability, the PPG/PEG/Pluronic system has been used as an experimental model to study ternary systems composed of two homopolymers and the corresponding block copolymer [2–4]. The observations described here can be a complicating factor if such model experiments are conducted at room temperature.
PPG, PEG, and PPO–PEO block copolymer surfactants were obtained from BASF and used as received. Most experiments used
Corresponding author. Fax: +1 412 624 9639. E-mail address: [email protected] (S.S. Velankar).
0021-9797/$ – see front matter doi:10.1016/j.jcis.2008.03.050
ρω2 a3
σ
2. Experimental
*
the four materials listed in Table 1. The Pluronic L35 is a triblock copolymer of PEO–PPO–PEO. The WSB125 is quoted as being a diblock copolymer of PEO and PPO, with a butanol endgroup. Limited experiments with additional surfactants or other molecular weights of PPG and PEG will be mentioned in specific sections of this paper. Most experiments were conducted in a spinning drop tensiometer (SDT), an instrument generally used to measure the interfacial tension between two fluids [5–7] by the following procedure: the SDT tube is filled with the two fluids of interest, with the lower density fluid being in a minority. The tube is then held horizontally and spun about its axis causing the lower density fluid to centrifuge to the center and form an elongated drop (Fig. 1a). At equilibrium, the shape of this drop is a balance between interfacial stresses (∼σ /a) and centrifugal stresses (∼ρω2 a2 ), where ρ is the density difference between the phases, and all other quantities are defined in Fig. 1a. If the aspect ratio L /a of the drop, exceeds about 4, the drop is nearly cylindrical [5] and the equilibrium relationship is given by
©
2008 Elsevier Inc. All rights reserved.
= 4 provided
L a
> 4.
(1)
This equation can be used to obtain the interfacial tension σ . If the equilibrium between interfacial and centrifugal stress is disturbed subsequently, the drop shape will change in response as illustrated in Figs. 1b–1d. For example, an increase in rotational speed or decrease in σ will cause the drop to reduce a, and hence increase L by stretching. In contrast, a decrease in rotational speed or an increase in interfacial tension will cause the
670
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Table 1 Materials used Material
MW (g/mol)
PPO (wt%)
PPG PEG Pluronic L35: PEO–PPO–PEO triblock Pluracol WSB125 PEO–PPO diblock
3500 400 1800 4000
100 0 50 50a
a
From 1 H NMR experiments. All other data was provided by BASF.
Fig. 2. Retraction of a PPG drop suspended in PEG in the SDT. Initial shape at (a) 7620, (b) 550 rpm.
of drop shape, and hence only uncorrected images will be presented. 3. Observations Experiments on the PPG/PEG/Pluronic system revealed two unusual observations as described in the following two sections. 3.1. Nonretraction Fig. 1. (a) Schematic of SDT experiment. (b–d) Possible evolutions of drop shapes in a SDT. (b) Increase in ω or decrease in σ causes a decrease in diameter, and hence stretching, (c) decrease in ω or increase in σ causes retraction if the initial drop has a low aspect ratio, (d) decrease in ω or increase in σ can induce capillary instability if the initial drop has a high aspect ratio.
drop to retract into a new equilibrium shape (if L /a is less than about 15) or breakup by a capillary instability (if L /a is much larger than 15) [8]. Interpreting such shape changes is crucial to this paper. The SDT used in this research has a sample tube with inner diameter of 12.7 mm and a length of about 12 cm. The tube can be rotated at speeds ranging from about 300 rpm to 12,000 rpm. Following the design of Joseph et al. [7], the tube spins in an electrically-heated oven allowing its temperature to be raised up to 100 ◦ C. The temperature is measured by thermocouples inserted into the oven. The sample temperature cannot be monitored directly. Experiments using temperature-sensitive pellets (Omega, Inc.) placed in the sample tube show that at steady state, the sample is at the same temperature as the oven. However, during heating, the sample temperature may lag behind the oven temperature. Such lag has not been characterized. For all experiments, the desired amount of PPG, typically 0.2 to 1% by weight (depending on the desired drop size), and the desired amount of surfactant were first dispersed as droplets in the PEG. This was to ensure that there were no diffusion limitations for surfactant adsorption onto the PEG/PPG interface. The blend was then poured into the SDT tube and the tube allowed to stand upright until the dispersed PPG droplets rose to the top and coalesced into a few large drops. The tube was then spun in the SDT where the few large drops centrifuged to the center and coalesced into a single elongated drop. In some cases, a diagonal stripe pattern was placed in the background to highlight the edges of the drops. This paper does not perform quantitative analysis of drop dimensions. However, we note that the drops appear distorted (e.g., spheres appear ellipsoidal) due to the lensing effect of the cylindrical sample tube. This can be corrected using a spherical plastic bead, imaged in the same fashion as a drop, for calibration. This paper is only concerned with qualitative aspects
A moderately elongated drop of fluid suspended in another fluid tends to retract due to interfacial tension into a spherical shape. Indeed, this is the basis of a popular method to measure interfacial tension between immiscible polymers [9–11]. Drop retraction can be induced in the SDT by the following procedure [11]: The drop is spun at high speed in order to draw it into an elongated cylindrical shape of diameter given by Eq. (1). The spinning speed is then reduced abruptly: as per Eq. (1), the drop must now retract to attain its final equilibrium diameter. If the final rotational speed is chosen to be sufficiently low (i.e., centrifugal forces are much smaller than interfacial forces), such retraction is essentially identical to that under quiescent conditions and the drop retracts into a sphere. Fig. 2 shows an example: a PPG-3500 drop was spun at 7620 rpm to achieve an aspect ratio of 12.6 (Fig. 2a). Upon reducing the rotational speed to 550 rpm, it retracted into a spherical shape within a few seconds (Fig. 2b). Note that the shape in Fig. 2b appears ellipsoidal due to the lensing mentioned in Section 2, however, correcting for the lensing confirms that the drop is almost exactly spherical. The same experiment was then repeated on the same system but with 0.1 wt% Pluronic L35 surfactant. The drop, initially being spun at 3750 rpm had an aspect ratio of about 11 (Fig. 3a). Fig. 3b shows the first unexpected observation: upon reducing the rotational rate to 540 rpm, the drop retracted partially within a few seconds, and then stopped retracting, retaining its irregular shape. This irregular shape persisted with no change for several hours as long as spinning was continued at 540 rpm. Figs. 4a–4d show the final irregular shapes of drops as the initial aspect ratio is varied. All four images refer to the same drop; the only difference is that the initial rotation speed was increased successively so that retraction started from cylinders of increasing aspect ratio. It is clear that at fixed surfactant concentration, the greater the initial aspect ratio of the cylinder, the more severely irregular is the final shape. Similarly, Figs. 4e–4g show that at fixed aspect ratio, increasing surfactant concentration also results in more irregular shapes. Heating the samples was found to erase this nonretraction behavior. Specifically, the irregular drop shapes (still spinning at a low rate) were heated, with the set point of the oven being
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Fig. 3. Retraction of a PPG drop suspended in PEG with Pluronic L35 (0.1 percent by weight) at the interface. (a) 3750, (b) 540 rpm.
Fig. 4. Final shapes of PPG3500 drops in PEG400 with Pluronic L35. (a–d) Surfactant concentration is fixed at 0.4%; initial aspect ratios are listed alongside each picture. (e–h) Initial aspect ratio was fixed at 34 ± 2; surfactant concentrations are listed alongside each picture. In all cases, the final rotational speed was 400 ± 100 rpm.
changed in 5 ◦ C increments. Between 35 and 40 ◦ C, the irregular drop shapes retracted gradually into spherical shapes. Further heating up to at least 65 ◦ C caused no further changes in the sample. Upon subsequent cooling from 65 ◦ C back to room temperature, the nonretraction behavior vanished and the samples behaved “normally,” i.e., initially elongated drops would retract readily into spherical shape upon reducing rotational speed. Thus, we conclude that heating to 65 ◦ C irreversibly erases the nonretraction behavior (it may return after extended storage: we have not tested this). All of the above observations about nonretraction and heatinginduced reversion to normal behavior remain valid for the WSB125 diblock surfactant, as well as four other surfactants that were
671
tested: Pluronic L64, Pluronic P65, Pluronic-R 10R5, and PluronicR 25R4. Various PPGs of molecular weights ranging from 2000 to 3500 and PEGs of molecular weights ranging from 400 to 600 also showed the same behavior. Thus the nonretraction seems to be fairly general for the PPG/PEG/Pluronic system. Irregular shapes such as Figs. 3b and 4 cannot be explained by a balance between interfacial tension and centrifugal forces, and indicate complex interfacial behavior that cannot be captured by interfacial tension alone. Before discussing the possible mechanisms for nonretraction, we will first briefly discuss the mechanism of “normal” retraction of elongated drops under quiescent conditions. The tips of an elongated drop have a higher curvature, and hence a higher capillary pressure, than the mid-section. This pressure gradient drives the retraction of an elongated drop [1]. Retraction generally stops only when the capillary pressure is the same everywhere; for surfactant-free drops this implies an equal curvature everywhere, i.e., a spherical final shape. One possible reason why Pluronic-containing drops do not retract into spheres is that interfacial tension gradients may develop as retraction proceeds: since retraction occurs from the tips, the surfactant concentration is expected to become higher at the tips than that at the mid-section, giving the tips a lower interfacial tension. Since the capillary pressure is the product of the local interfacial tension and local curvature, this raises the possibility that the drop surface can have a uniform capillary pressure even while the drop remains nonspherical. In such cases, capillary pressure gradients would no longer drive drop retraction [12], thus possibly explaining the nonretraction phenomenon. Yet, while interfacial tension gradients may play a significant role in retraction of surfactant-laden drops we believe that they cannot by themselves explain the nonspherical shapes seen here. First, the final shapes in Figs. 3b and 4 are quite complex and it seems unlikely that they have the precise surfactant distribution required to produce exactly the same capillary pressure everywhere. Moreover, even in the absence of capillary pressure gradients, Marangoni stresses should be able to drive retraction [12], whereas the drops of Figs. 3b and 4 remained stable for hours. A second possible reason for nonretraction is that the Pluronic endows the PEG/PPG interface with an interfacial yield stress. In oil/water or air/water systems, surface-active species are wellknown to give interfaces mechanical properties such as a dilational or shear modulus, shear viscosity, and possibly a yield stress. Nonspherical drop and bubble shapes induced by interfaciallyadsorbed particles have been known for over a century [13–16]; in such cases, the particles form a shell with some yield stress. There is much less literature on unusual interfacial mechanical properties in polymeric systems, principally because the high bulk viscosity of polymers makes any interfacial effects difficult to detect or measure. Nevertheless, there is no reason why polymer/polymer interfaces with interfacially-adsorbed species cannot show similarlycomplex mechanical properties. We can estimate the minimum magnitude of the interfacial yield stress, τy , by equating the capillary force driving drop retraction with the interfacial force resisting it:
π R2
σ R
= 2π R τy .
(2)
Thus, the interfacial yield stress must at least be on the order of the interfacial tension (i.e., on the order of 1 mN/m for Pluronicladen PEG/PPG interfaces as per our measurements). Interfacial yield stresses on the order of 1 mN/M should be within the range of measurement of existing instruments [17,18], although sample preparation may prove challenging. What may cause such an interfacial yield stress? While the interfacially-adsorbed Pluronic may itself endow the interface with a yield stress, the fact that heating erases this behavior irreversibly
672
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
suggests that the surfactant adsorption per se is not the cause. We hypothesize that a small amount of crystallinity is responsible for this behavior. In particular, surfactants such as Pluronic L35 have a melting temperature slightly below room temperature. However due to polydispersity, at room temperature, there may be some fraction of the surfactant that is present in the form of small crystallites. If these crystallites adsorb at the PEO/PPO interface (many solids are known to adsorb at liquid/liquid interfaces [19]), they may endow the interface with a small yield stress and cause nonretraction. The fact that heating to slightly-above room temperature erases nonretraction supports this hypothesis. It is also possible that some impurity in the surfactant or some aggregated structure formed by the surfactant as the initial blend coalesces into a large drop is responsible for this behavior. Regardless of its cause, the interfacial yield stress may confound any experiments that use the PEG/PPG/Pluronic as a model system at room temperature. We recommend that experimenters control the thermal history of their samples carefully, and in particular, heat the samples to above 65 ◦ C prior to experiments. 3.2. Autoextension upon heating As mentioned in the Introduction, the motivation for our research was to study the effect of surfactant on interfacial tensiondriven drop dynamics. In order to avoid the complex interfacial behavior discussed above, it seems judicious to heat all drops to 65 ◦ C, cool back to room temperature, and only then conduct further experiments of interfacial tension-driven dynamics. During heating and subsequent cooling steps, the sample was spun at a constant, slow speed of about 300 rpm to maintain the drop approximately at the center of the tube. During such heating steps of drops laden with WSB125 surfactant, we made a second surprising observation: when heating a drop at a constant, low rpm, upon reaching about 60 ◦ C, the drop would often stretch strongly. We call this phenomenon “autoextension” because even though the rotational rate is fixed, the drop appears to extend rapidly “all by itself.” Most interestingly, after autoextension, the drop would not remain in its extended state: it would either retract back into a slightly deformed shape, or break into several fragments that would retract into nearly-spherical drops. In general, drops that autoextended only modestly would subsequently retract, whereas those that autoextended severely would subsequently breakup. Fig. 5 illustrates an example. A PPG/PEG/WSB125 blend (overall composition 0.1 wt% WSB125, 0.3 wt% PPG; this same composition applies to all experiments in this section) was allowed to coalesce into a large drop at 25 ◦ C. At ∼300 rpm, the drop was only slightly elongated (Fig. 5a). The oven set point temperature was then changed to 65 ◦ C, and it took the oven about 4 min to reach 65 ◦ C. When the oven temperature reached ∼60 ◦ C, the drop stretched strongly (Fig. 5b) and then broke up (Fig. 5c). (As mentioned in Section 2, sample temperature cannot be measured directly, but we presume that during heating, the sample temperature lags behind the oven temperature.) In the case of Fig. 5, the entire process of stretching and breakup took less than 5 min. Such behavior was evident only in drops with surfactant; drops without surfactant stretched only slightly upon increasing temperature (suggesting a slight decrease in interfacial tension with increasing temperature), but never showed a subsequent retraction or breakup. The entire process of autoextension and subsequent breakup or retraction occurs at constant rotational speed ω , and hence as per Eq. (1), these processes must be regarded as reflecting changes in interfacial tension σ . Specifically, during autoextension, the diameter of the drop reduces sharply, indicative of a sharp decrease in interfacial tension. The subsequent breakup (or retraction) are both indicative of an increase in interfacial tension. In summary,
Fig. 5. A drop being spun at 300 rpm (a) while heating first autoextended into a highly elongated shape (b) and then broke up (c). (d) Schematic of interfacial tension changes with time during autoextension and subsequent breakup.
the interfacial tension changes nonmonotonically with time during heating as illustrated in Fig. 5d. Experiments to characterize the autoextension phenomenon were difficult to reproduce: although drops would generally autoextend and then retract or breakup, the amount of autoextension varied significantly. In very few cases, autoextension would not occur at all. This led us to believe that the autoextension phenomenon is highly sensitive to the thermomechanical history of the drop, and this was investigated in greater detail. Specifically, the sequence of the two processes—cooling the drop, and retracting the drop by reducing the rotational speed—was found to be critical. This is best illustrated by a comparison between the following two experimental protocols. In the first Protocol A, dubbed “retraction before cooling,” a blend of PPG/PEG/WSB125 of the above-cited composition was charged to the SDT tube, heated to 65 ◦ C and spun at 1300 rpm to obtain the elongated drop of Fig. 6a. The rotational rate was reduced while the drop was still at 65 ◦ C to allow retraction (Fig. 6b), and then the drop was cooled to room temperature (Fig. 6c). Upon reheating, the drop shape Fig. 6d was recorded. It is clear that there are only modest shape changes during heating, and in particular, autoextension did not occur. In the second Protocol B, dubbed “cooling before retraction,” the blend was first spun at 1300 rpm at 65 ◦ C (Fig. 6e) to coalesce into a large drop. It was then cooled to room temperature while still in its elongated shape at a high rotational rate (Fig. 6f). The rotational speed was then reduced while the drop was at room temperature to induce retraction into Fig. 6g. Upon reheating this drop, it elongated dramatically (Fig. 6h) and subsequently broke up. To summarize, although drops Figs. 6c and 6g are at the same rpm and temperature, upon heating they show dramatic differences in behavior. These differences are solely attributable to differences in their thermomechanical history. Specifically, allowing the drop to cool before retracting (Protocol B) seems to be responsible for autoextension. We propose the following mechanism to explain autoextension and subsequent retraction or breakup. The key idea is that the PEO block of the WSB125 surfactant is relatively long (MW of about
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Fig. 6. Effect of thermomechanical history on autoextension. Sequence (a–d) is Protocol A, “retraction before cooling:” (a) 1300 rpm, 65 ◦ C; (b) 340 rpm, 65 ◦ C; (c) 350 rpm, 25 ◦ C; (d) 350 rpm, 65 ◦ C. The drop retracts going from (a) to (b), and cools going from (b) to (c). Sequence (e–h) is is Protocol B, “cooling before retraction:” (e) 1300 rpm, 65 ◦ C; (f) 1350 rpm, 25 ◦ C cool; (g) 300 rpm, 25 ◦ C retract; (h) 290 rpm, 65 ◦ C. The drop cools going from (e) to (f), and retracts going from (f) to (g).
1000) and is therefore able to crystallize. Indeed, at room temperature, bulk WSB125 has a soft waxy consistency due to PEO crystallinity. Thus, when a PPG/PEG interface laden with WSB125 surfactant is cooled, we postulate that the interfacially-adsorbed surfactant crystallizes. Such crystallized surfactant is now trapped at the interface, and cannot desorb until the temperature is raised above a value that we denote the “interfacial melting temperature,” T mi , which lies between room temperature and 60 ◦ C. We will measure T mi more accurately later. The experiment of Fig. 6 can now be explained in more detail. Under the initial conditions (1300 rpm, 65 ◦ C), the surfactant has some interfacial concentration which is expected to be close to the equilibrium concentration. In Protocol A, the drop retracts while it is above T mi . Therefore as the drop retracts and the interfacial area reduces, the surfactant can desorb off the interface and maintain an interfacial concentration not too far above the equilibrium value. In Protocol B however, the drop retraction occurs at a temperature below T mi , at which the surfactant is immo-
673
bile and cannot desorb. Thus, as the interfacial area reduces, the surfactant concentration increases to, and remains, far above its equilibrium value. Upon reheating above T mi , two processes occur simultaneously (see schematic in Fig. 5d). First, the high interfacial concentration of surfactant immediately and sharply reduces the interfacial tension, thus inducing autoextension. Second, and more slowly, the surfactant starts desorbing off the interface so as to approach equilibrium. This gradually raises the interfacial tension, eventually inducing breakup or retraction. The SDT does not allow an exact determination of T mi since during a temperature ramp, the temperature inside the SDT tube lags behind the oven temperature by an unknown amount. However, T mi can be determined more accurately by an alternate experiment. The essence of the autoextension phenomenon is that an initial drop shape resulting from a balance between interfacial and centrifugal forces is altered because the interfacial forces reduce sharply when T exceeds T mi . Thus autoextension can be studied more conveniently outside the SDT by examining the shape of a sessile drop which is determined by a balance between gravitational and interfacial forces. Accordingly, we repeated the experiment of Fig. 6 in a glass vial. Temperature was controlled by immersing the vial in an oil bath; in this immersion setup, the vial temperature is nearly equal to the bath temperature and hence known precisely. A PPG/PEG/WSB125 blend of the same composition as above was poured into a vial. The vial was laid on its side in the oil bath at 65 ◦ C, and agitated with a magnetic stirrer to prepare an emulsion. The oil bath and the vial were cooled to 25 ◦ C with stirring, and then stirring was halted to allow the emulsion drops to coalesce into a large PPG sessile drop resting against the top curved wall of the vial. Similar to Protocol B above, the coalescence significantly reduced the surface area after cooling, thus we expect the surfactant to be trapped at the interface at a concentration far above equilibrium. The oil bath was then heated at the rate of approximately 1.5 ◦ C/min. Strong autoextension was observed when the oil bath reached 40 ◦ C; thus we conclude that T mi = 40 ◦ C. We also repeated the above experiment with coalescence occurring at 65 ◦ C, followed by cooling (thus, interfacial area decreased before cooling, analogous to Protocol A above). As expected in this case, reheating induced only small changes in drop shape upon heating. The value of T mi = 40 ◦ C is almost identical to the melting temperature of WSB125, thus providing support for the hypothesis that crystallization of interfacially-adsorbed surfactant is indeed the mechanism for interfacial trapping of the surfactant. We also note that the surfactant L35, which is liquid at room temperature, did not show autoextension behavior, further supporting surfactant crystallization at the interface as the trapping mechanism. Such interfacial trapping may be a general feature of crystallizable block copolymers adsorbed at interfaces between noncrystalline phases. 4. Summary and conclusions We studied drop retraction of Pluronic-laden PPG/PEG interfaces in the spinning drop tensiometer. While the motivation for this research was to study the effect of surface-active species on capillary instabilities of elongated drops under quiescent conditions, we instead observed two unusual phenomena that are of interest in their own right. The first is that under certain circumstances, Pluronic-laden elongated PPG drops in PEG do not retract, but instead maintain nonspherical shapes indefinitely. We believe that such nonretraction is attributable to an interfacial yield stress caused by addition of surfactant. The second is that under certain circumstances PPG drops in PEG elongate sharply upon heating. We show that such autoextension is directly related to a decrease in interfacial area at room temperature; the surfactant is then trapped at the interface (likely due to crystallization of the PEG
674
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
block) at concentrations far above equilibrium, which induces autoextension upon reheating. These unusual phenomena indicate that at room temperature, the PPG/PEG/Pluronic system is not a model system for two-phase flow experiments. Acknowledgments We are grateful to BASF for providing the PPG, PEG, and Pluronics for this research. This research was funded the ACS Petroleum Research Fund (Grant #39931-G9), and by a National Science Foundation CAREER grant (CBET-0448845). We are grateful to Dr. Steven Hudson, NIST, Gaithersburg for valuable suggestions. References [1] H.A. Stone, L.G. Leal, J. Fluid Mech. 198 (1989) 399. [2] A.J. Ramic, J.C. Stehlin, S.D. Hudson, A.M. Jamieson, I. Manas-Zloczower, Macromolecules 33 (2000) 371. [3] I. Welge, B.A. Wolf, Polymer 42 (2001) 3467.
[4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
M. Malmsten, B. Lindman, Macromolecules 26 (1993) 1282. H.M. Princen, I.Y.Z. Zia, S.G. Mason, J. Colloid Interface Sci. 23 (1967) 99. B. Vonnegut, Rev. Sci. Instrum. 13 (1942) 6. D.D. Joseph, M.S. Arney, G. Gillberg, H. Hu, D. Hultman, C. Verdier, T.M. Vinagre, J. Rheol. 36 (1992) 621. E.H.A. de Hoog, H.N.W. Lekkerkerker, J. Phys. Chem. B 105 (2001) 11636. A. Cohen, C.J. Carriere, Rheol. Acta 28 (1989) 223. M. Tjahjadi, J.M. Ottino, H.A. Stone, AIChE J. 40 (1994) 385. J. Martin, S. Velankar, Macromolecules 38 (2005) 10614. S. Velankar, H. Zhou, H.K. Jeon, C.W. Macosko, J. Colloid Interface Sci. 272 (2004) 172. W. Ramsden, Proc. R. Soc. London A 72 (1903) 156. A.B. Subramaniam, M. Abkarian, L. Mahadevan, H.A. Stone, Langmuir 22 (2006) 10204. S.U. Pickering, J. Chem. Soc. Abstracts 91, 92 (1907) 2001. A.B. Subramaniam, M. Abkarian, L. Mahadevan, H.A. Stone, Nature 438 (2005) 930. C.F. Brooks, G.G. Fuller, C.W. Frank, C.R. Robertson, Langmuir 15 (1999) 2450. P. Erni, P. Fischer, E.J. Windhab, V. Kusnezov, H. Stettin, J. Lauger, Rev. Sci. Instrum. 74 (2003) 4916. B.P. Binks, T.S. Horozov (Eds.), Colloidal Particles at Liquid Interfaces, Cambridge Univ. Press, Cambridge, 2006.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Note
Unusual behavior of PEG/PPG/Pluronic interfaces studied by a spinning drop tensiometer Jeffrey D. Martin, Sachin S. Velankar ∗ Department of Chemical Engineering, University of Pittsburgh, Pittsburgh, PA 15261, USA
a r t i c l e
i n f o
a b s t r a c t
Article history: Received 24 October 2007 Accepted 28 March 2008 Available online 8 April 2008
The effects of surfactants on the interfacial tension driven retraction of elongated drops were studied in a spinning drop tensiometer. Experiments were conducted on polypropylene glycol (PPG) drops suspended in polyethylene glycol (PEG), with Pluronic block copolymers as surfactants. Two unusual observations are reported here. In the first, initially-elongated drops generated at high rotational speed were allowed to retract by reducing the rotational speed. Pluronic-laden drops would not retract completely, but would instead maintain strongly nonspherical shapes indefinitely. We attribute such “nonretraction” to an interfacial yield stress induced by the Pluronic surfactant. In the second, drops being heated while spinning at a constant speed would elongate sharply at some temperature, and subsequently breakup. Such “autoextension” and breakup indicate complex nonmonotonic changes in interfacial tension with time during heating. We propose that autoextension occurs because at low temperature, interfacially-adsorbed surfactant is crystallized and hence trapped at the interface at a concentration far above equilibrium. © 2008 Elsevier Inc. All rights reserved.
Keywords: Interfacial tension Interfacial viscoelasticity Pluronic Interfacial crystallization Interfacial yield stress Nonspherical drops
1. Introduction A cylindrical liquid drop suspended in a quiescent fluid can either retract back into a spherical shape or can break up by a Rayleigh instability into a series of drops [1]. Generally, drops with modest aspect ratio (< ∼15) retract, whereas very long drops (aspect ratio > ∼15) breakup. We are presently experimentally investigating the effect of surfactant on both these interfacial tensiondriven processes. We chose for these studies an experimental system composed of polypropylene glycol (PPG) drops suspended in a oplyethylene glycol (PEG) matrix with PPO–PEO block copolymer surfactants (mostly from the Pluronic family) at the interface. This Note reports on some remarkable observations about the dynamics of this system made during the course of our research. Due to its easy and cheap availability, transparency, and chemical stability, the PPG/PEG/Pluronic system has been used as an experimental model to study ternary systems composed of two homopolymers and the corresponding block copolymer [2–4]. The observations described here can be a complicating factor if such model experiments are conducted at room temperature.
PPG, PEG, and PPO–PEO block copolymer surfactants were obtained from BASF and used as received. Most experiments used
Corresponding author. Fax: +1 412 624 9639. E-mail address: [email protected] (S.S. Velankar).
0021-9797/$ – see front matter doi:10.1016/j.jcis.2008.03.050
ρω2 a3
σ
2. Experimental
*
the four materials listed in Table 1. The Pluronic L35 is a triblock copolymer of PEO–PPO–PEO. The WSB125 is quoted as being a diblock copolymer of PEO and PPO, with a butanol endgroup. Limited experiments with additional surfactants or other molecular weights of PPG and PEG will be mentioned in specific sections of this paper. Most experiments were conducted in a spinning drop tensiometer (SDT), an instrument generally used to measure the interfacial tension between two fluids [5–7] by the following procedure: the SDT tube is filled with the two fluids of interest, with the lower density fluid being in a minority. The tube is then held horizontally and spun about its axis causing the lower density fluid to centrifuge to the center and form an elongated drop (Fig. 1a). At equilibrium, the shape of this drop is a balance between interfacial stresses (∼σ /a) and centrifugal stresses (∼ρω2 a2 ), where ρ is the density difference between the phases, and all other quantities are defined in Fig. 1a. If the aspect ratio L /a of the drop, exceeds about 4, the drop is nearly cylindrical [5] and the equilibrium relationship is given by
©
2008 Elsevier Inc. All rights reserved.
= 4 provided
L a
> 4.
(1)
This equation can be used to obtain the interfacial tension σ . If the equilibrium between interfacial and centrifugal stress is disturbed subsequently, the drop shape will change in response as illustrated in Figs. 1b–1d. For example, an increase in rotational speed or decrease in σ will cause the drop to reduce a, and hence increase L by stretching. In contrast, a decrease in rotational speed or an increase in interfacial tension will cause the
670
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Table 1 Materials used Material
MW (g/mol)
PPO (wt%)
PPG PEG Pluronic L35: PEO–PPO–PEO triblock Pluracol WSB125 PEO–PPO diblock
3500 400 1800 4000
100 0 50 50a
a
From 1 H NMR experiments. All other data was provided by BASF.
Fig. 2. Retraction of a PPG drop suspended in PEG in the SDT. Initial shape at (a) 7620, (b) 550 rpm.
of drop shape, and hence only uncorrected images will be presented. 3. Observations Experiments on the PPG/PEG/Pluronic system revealed two unusual observations as described in the following two sections. 3.1. Nonretraction Fig. 1. (a) Schematic of SDT experiment. (b–d) Possible evolutions of drop shapes in a SDT. (b) Increase in ω or decrease in σ causes a decrease in diameter, and hence stretching, (c) decrease in ω or increase in σ causes retraction if the initial drop has a low aspect ratio, (d) decrease in ω or increase in σ can induce capillary instability if the initial drop has a high aspect ratio.
drop to retract into a new equilibrium shape (if L /a is less than about 15) or breakup by a capillary instability (if L /a is much larger than 15) [8]. Interpreting such shape changes is crucial to this paper. The SDT used in this research has a sample tube with inner diameter of 12.7 mm and a length of about 12 cm. The tube can be rotated at speeds ranging from about 300 rpm to 12,000 rpm. Following the design of Joseph et al. [7], the tube spins in an electrically-heated oven allowing its temperature to be raised up to 100 ◦ C. The temperature is measured by thermocouples inserted into the oven. The sample temperature cannot be monitored directly. Experiments using temperature-sensitive pellets (Omega, Inc.) placed in the sample tube show that at steady state, the sample is at the same temperature as the oven. However, during heating, the sample temperature may lag behind the oven temperature. Such lag has not been characterized. For all experiments, the desired amount of PPG, typically 0.2 to 1% by weight (depending on the desired drop size), and the desired amount of surfactant were first dispersed as droplets in the PEG. This was to ensure that there were no diffusion limitations for surfactant adsorption onto the PEG/PPG interface. The blend was then poured into the SDT tube and the tube allowed to stand upright until the dispersed PPG droplets rose to the top and coalesced into a few large drops. The tube was then spun in the SDT where the few large drops centrifuged to the center and coalesced into a single elongated drop. In some cases, a diagonal stripe pattern was placed in the background to highlight the edges of the drops. This paper does not perform quantitative analysis of drop dimensions. However, we note that the drops appear distorted (e.g., spheres appear ellipsoidal) due to the lensing effect of the cylindrical sample tube. This can be corrected using a spherical plastic bead, imaged in the same fashion as a drop, for calibration. This paper is only concerned with qualitative aspects
A moderately elongated drop of fluid suspended in another fluid tends to retract due to interfacial tension into a spherical shape. Indeed, this is the basis of a popular method to measure interfacial tension between immiscible polymers [9–11]. Drop retraction can be induced in the SDT by the following procedure [11]: The drop is spun at high speed in order to draw it into an elongated cylindrical shape of diameter given by Eq. (1). The spinning speed is then reduced abruptly: as per Eq. (1), the drop must now retract to attain its final equilibrium diameter. If the final rotational speed is chosen to be sufficiently low (i.e., centrifugal forces are much smaller than interfacial forces), such retraction is essentially identical to that under quiescent conditions and the drop retracts into a sphere. Fig. 2 shows an example: a PPG-3500 drop was spun at 7620 rpm to achieve an aspect ratio of 12.6 (Fig. 2a). Upon reducing the rotational speed to 550 rpm, it retracted into a spherical shape within a few seconds (Fig. 2b). Note that the shape in Fig. 2b appears ellipsoidal due to the lensing mentioned in Section 2, however, correcting for the lensing confirms that the drop is almost exactly spherical. The same experiment was then repeated on the same system but with 0.1 wt% Pluronic L35 surfactant. The drop, initially being spun at 3750 rpm had an aspect ratio of about 11 (Fig. 3a). Fig. 3b shows the first unexpected observation: upon reducing the rotational rate to 540 rpm, the drop retracted partially within a few seconds, and then stopped retracting, retaining its irregular shape. This irregular shape persisted with no change for several hours as long as spinning was continued at 540 rpm. Figs. 4a–4d show the final irregular shapes of drops as the initial aspect ratio is varied. All four images refer to the same drop; the only difference is that the initial rotation speed was increased successively so that retraction started from cylinders of increasing aspect ratio. It is clear that at fixed surfactant concentration, the greater the initial aspect ratio of the cylinder, the more severely irregular is the final shape. Similarly, Figs. 4e–4g show that at fixed aspect ratio, increasing surfactant concentration also results in more irregular shapes. Heating the samples was found to erase this nonretraction behavior. Specifically, the irregular drop shapes (still spinning at a low rate) were heated, with the set point of the oven being
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Fig. 3. Retraction of a PPG drop suspended in PEG with Pluronic L35 (0.1 percent by weight) at the interface. (a) 3750, (b) 540 rpm.
Fig. 4. Final shapes of PPG3500 drops in PEG400 with Pluronic L35. (a–d) Surfactant concentration is fixed at 0.4%; initial aspect ratios are listed alongside each picture. (e–h) Initial aspect ratio was fixed at 34 ± 2; surfactant concentrations are listed alongside each picture. In all cases, the final rotational speed was 400 ± 100 rpm.
changed in 5 ◦ C increments. Between 35 and 40 ◦ C, the irregular drop shapes retracted gradually into spherical shapes. Further heating up to at least 65 ◦ C caused no further changes in the sample. Upon subsequent cooling from 65 ◦ C back to room temperature, the nonretraction behavior vanished and the samples behaved “normally,” i.e., initially elongated drops would retract readily into spherical shape upon reducing rotational speed. Thus, we conclude that heating to 65 ◦ C irreversibly erases the nonretraction behavior (it may return after extended storage: we have not tested this). All of the above observations about nonretraction and heatinginduced reversion to normal behavior remain valid for the WSB125 diblock surfactant, as well as four other surfactants that were
671
tested: Pluronic L64, Pluronic P65, Pluronic-R 10R5, and PluronicR 25R4. Various PPGs of molecular weights ranging from 2000 to 3500 and PEGs of molecular weights ranging from 400 to 600 also showed the same behavior. Thus the nonretraction seems to be fairly general for the PPG/PEG/Pluronic system. Irregular shapes such as Figs. 3b and 4 cannot be explained by a balance between interfacial tension and centrifugal forces, and indicate complex interfacial behavior that cannot be captured by interfacial tension alone. Before discussing the possible mechanisms for nonretraction, we will first briefly discuss the mechanism of “normal” retraction of elongated drops under quiescent conditions. The tips of an elongated drop have a higher curvature, and hence a higher capillary pressure, than the mid-section. This pressure gradient drives the retraction of an elongated drop [1]. Retraction generally stops only when the capillary pressure is the same everywhere; for surfactant-free drops this implies an equal curvature everywhere, i.e., a spherical final shape. One possible reason why Pluronic-containing drops do not retract into spheres is that interfacial tension gradients may develop as retraction proceeds: since retraction occurs from the tips, the surfactant concentration is expected to become higher at the tips than that at the mid-section, giving the tips a lower interfacial tension. Since the capillary pressure is the product of the local interfacial tension and local curvature, this raises the possibility that the drop surface can have a uniform capillary pressure even while the drop remains nonspherical. In such cases, capillary pressure gradients would no longer drive drop retraction [12], thus possibly explaining the nonretraction phenomenon. Yet, while interfacial tension gradients may play a significant role in retraction of surfactant-laden drops we believe that they cannot by themselves explain the nonspherical shapes seen here. First, the final shapes in Figs. 3b and 4 are quite complex and it seems unlikely that they have the precise surfactant distribution required to produce exactly the same capillary pressure everywhere. Moreover, even in the absence of capillary pressure gradients, Marangoni stresses should be able to drive retraction [12], whereas the drops of Figs. 3b and 4 remained stable for hours. A second possible reason for nonretraction is that the Pluronic endows the PEG/PPG interface with an interfacial yield stress. In oil/water or air/water systems, surface-active species are wellknown to give interfaces mechanical properties such as a dilational or shear modulus, shear viscosity, and possibly a yield stress. Nonspherical drop and bubble shapes induced by interfaciallyadsorbed particles have been known for over a century [13–16]; in such cases, the particles form a shell with some yield stress. There is much less literature on unusual interfacial mechanical properties in polymeric systems, principally because the high bulk viscosity of polymers makes any interfacial effects difficult to detect or measure. Nevertheless, there is no reason why polymer/polymer interfaces with interfacially-adsorbed species cannot show similarlycomplex mechanical properties. We can estimate the minimum magnitude of the interfacial yield stress, τy , by equating the capillary force driving drop retraction with the interfacial force resisting it:
π R2
σ R
= 2π R τy .
(2)
Thus, the interfacial yield stress must at least be on the order of the interfacial tension (i.e., on the order of 1 mN/m for Pluronicladen PEG/PPG interfaces as per our measurements). Interfacial yield stresses on the order of 1 mN/M should be within the range of measurement of existing instruments [17,18], although sample preparation may prove challenging. What may cause such an interfacial yield stress? While the interfacially-adsorbed Pluronic may itself endow the interface with a yield stress, the fact that heating erases this behavior irreversibly
672
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
suggests that the surfactant adsorption per se is not the cause. We hypothesize that a small amount of crystallinity is responsible for this behavior. In particular, surfactants such as Pluronic L35 have a melting temperature slightly below room temperature. However due to polydispersity, at room temperature, there may be some fraction of the surfactant that is present in the form of small crystallites. If these crystallites adsorb at the PEO/PPO interface (many solids are known to adsorb at liquid/liquid interfaces [19]), they may endow the interface with a small yield stress and cause nonretraction. The fact that heating to slightly-above room temperature erases nonretraction supports this hypothesis. It is also possible that some impurity in the surfactant or some aggregated structure formed by the surfactant as the initial blend coalesces into a large drop is responsible for this behavior. Regardless of its cause, the interfacial yield stress may confound any experiments that use the PEG/PPG/Pluronic as a model system at room temperature. We recommend that experimenters control the thermal history of their samples carefully, and in particular, heat the samples to above 65 ◦ C prior to experiments. 3.2. Autoextension upon heating As mentioned in the Introduction, the motivation for our research was to study the effect of surfactant on interfacial tensiondriven drop dynamics. In order to avoid the complex interfacial behavior discussed above, it seems judicious to heat all drops to 65 ◦ C, cool back to room temperature, and only then conduct further experiments of interfacial tension-driven dynamics. During heating and subsequent cooling steps, the sample was spun at a constant, slow speed of about 300 rpm to maintain the drop approximately at the center of the tube. During such heating steps of drops laden with WSB125 surfactant, we made a second surprising observation: when heating a drop at a constant, low rpm, upon reaching about 60 ◦ C, the drop would often stretch strongly. We call this phenomenon “autoextension” because even though the rotational rate is fixed, the drop appears to extend rapidly “all by itself.” Most interestingly, after autoextension, the drop would not remain in its extended state: it would either retract back into a slightly deformed shape, or break into several fragments that would retract into nearly-spherical drops. In general, drops that autoextended only modestly would subsequently retract, whereas those that autoextended severely would subsequently breakup. Fig. 5 illustrates an example. A PPG/PEG/WSB125 blend (overall composition 0.1 wt% WSB125, 0.3 wt% PPG; this same composition applies to all experiments in this section) was allowed to coalesce into a large drop at 25 ◦ C. At ∼300 rpm, the drop was only slightly elongated (Fig. 5a). The oven set point temperature was then changed to 65 ◦ C, and it took the oven about 4 min to reach 65 ◦ C. When the oven temperature reached ∼60 ◦ C, the drop stretched strongly (Fig. 5b) and then broke up (Fig. 5c). (As mentioned in Section 2, sample temperature cannot be measured directly, but we presume that during heating, the sample temperature lags behind the oven temperature.) In the case of Fig. 5, the entire process of stretching and breakup took less than 5 min. Such behavior was evident only in drops with surfactant; drops without surfactant stretched only slightly upon increasing temperature (suggesting a slight decrease in interfacial tension with increasing temperature), but never showed a subsequent retraction or breakup. The entire process of autoextension and subsequent breakup or retraction occurs at constant rotational speed ω , and hence as per Eq. (1), these processes must be regarded as reflecting changes in interfacial tension σ . Specifically, during autoextension, the diameter of the drop reduces sharply, indicative of a sharp decrease in interfacial tension. The subsequent breakup (or retraction) are both indicative of an increase in interfacial tension. In summary,
Fig. 5. A drop being spun at 300 rpm (a) while heating first autoextended into a highly elongated shape (b) and then broke up (c). (d) Schematic of interfacial tension changes with time during autoextension and subsequent breakup.
the interfacial tension changes nonmonotonically with time during heating as illustrated in Fig. 5d. Experiments to characterize the autoextension phenomenon were difficult to reproduce: although drops would generally autoextend and then retract or breakup, the amount of autoextension varied significantly. In very few cases, autoextension would not occur at all. This led us to believe that the autoextension phenomenon is highly sensitive to the thermomechanical history of the drop, and this was investigated in greater detail. Specifically, the sequence of the two processes—cooling the drop, and retracting the drop by reducing the rotational speed—was found to be critical. This is best illustrated by a comparison between the following two experimental protocols. In the first Protocol A, dubbed “retraction before cooling,” a blend of PPG/PEG/WSB125 of the above-cited composition was charged to the SDT tube, heated to 65 ◦ C and spun at 1300 rpm to obtain the elongated drop of Fig. 6a. The rotational rate was reduced while the drop was still at 65 ◦ C to allow retraction (Fig. 6b), and then the drop was cooled to room temperature (Fig. 6c). Upon reheating, the drop shape Fig. 6d was recorded. It is clear that there are only modest shape changes during heating, and in particular, autoextension did not occur. In the second Protocol B, dubbed “cooling before retraction,” the blend was first spun at 1300 rpm at 65 ◦ C (Fig. 6e) to coalesce into a large drop. It was then cooled to room temperature while still in its elongated shape at a high rotational rate (Fig. 6f). The rotational speed was then reduced while the drop was at room temperature to induce retraction into Fig. 6g. Upon reheating this drop, it elongated dramatically (Fig. 6h) and subsequently broke up. To summarize, although drops Figs. 6c and 6g are at the same rpm and temperature, upon heating they show dramatic differences in behavior. These differences are solely attributable to differences in their thermomechanical history. Specifically, allowing the drop to cool before retracting (Protocol B) seems to be responsible for autoextension. We propose the following mechanism to explain autoextension and subsequent retraction or breakup. The key idea is that the PEO block of the WSB125 surfactant is relatively long (MW of about
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
Fig. 6. Effect of thermomechanical history on autoextension. Sequence (a–d) is Protocol A, “retraction before cooling:” (a) 1300 rpm, 65 ◦ C; (b) 340 rpm, 65 ◦ C; (c) 350 rpm, 25 ◦ C; (d) 350 rpm, 65 ◦ C. The drop retracts going from (a) to (b), and cools going from (b) to (c). Sequence (e–h) is is Protocol B, “cooling before retraction:” (e) 1300 rpm, 65 ◦ C; (f) 1350 rpm, 25 ◦ C cool; (g) 300 rpm, 25 ◦ C retract; (h) 290 rpm, 65 ◦ C. The drop cools going from (e) to (f), and retracts going from (f) to (g).
1000) and is therefore able to crystallize. Indeed, at room temperature, bulk WSB125 has a soft waxy consistency due to PEO crystallinity. Thus, when a PPG/PEG interface laden with WSB125 surfactant is cooled, we postulate that the interfacially-adsorbed surfactant crystallizes. Such crystallized surfactant is now trapped at the interface, and cannot desorb until the temperature is raised above a value that we denote the “interfacial melting temperature,” T mi , which lies between room temperature and 60 ◦ C. We will measure T mi more accurately later. The experiment of Fig. 6 can now be explained in more detail. Under the initial conditions (1300 rpm, 65 ◦ C), the surfactant has some interfacial concentration which is expected to be close to the equilibrium concentration. In Protocol A, the drop retracts while it is above T mi . Therefore as the drop retracts and the interfacial area reduces, the surfactant can desorb off the interface and maintain an interfacial concentration not too far above the equilibrium value. In Protocol B however, the drop retraction occurs at a temperature below T mi , at which the surfactant is immo-
673
bile and cannot desorb. Thus, as the interfacial area reduces, the surfactant concentration increases to, and remains, far above its equilibrium value. Upon reheating above T mi , two processes occur simultaneously (see schematic in Fig. 5d). First, the high interfacial concentration of surfactant immediately and sharply reduces the interfacial tension, thus inducing autoextension. Second, and more slowly, the surfactant starts desorbing off the interface so as to approach equilibrium. This gradually raises the interfacial tension, eventually inducing breakup or retraction. The SDT does not allow an exact determination of T mi since during a temperature ramp, the temperature inside the SDT tube lags behind the oven temperature by an unknown amount. However, T mi can be determined more accurately by an alternate experiment. The essence of the autoextension phenomenon is that an initial drop shape resulting from a balance between interfacial and centrifugal forces is altered because the interfacial forces reduce sharply when T exceeds T mi . Thus autoextension can be studied more conveniently outside the SDT by examining the shape of a sessile drop which is determined by a balance between gravitational and interfacial forces. Accordingly, we repeated the experiment of Fig. 6 in a glass vial. Temperature was controlled by immersing the vial in an oil bath; in this immersion setup, the vial temperature is nearly equal to the bath temperature and hence known precisely. A PPG/PEG/WSB125 blend of the same composition as above was poured into a vial. The vial was laid on its side in the oil bath at 65 ◦ C, and agitated with a magnetic stirrer to prepare an emulsion. The oil bath and the vial were cooled to 25 ◦ C with stirring, and then stirring was halted to allow the emulsion drops to coalesce into a large PPG sessile drop resting against the top curved wall of the vial. Similar to Protocol B above, the coalescence significantly reduced the surface area after cooling, thus we expect the surfactant to be trapped at the interface at a concentration far above equilibrium. The oil bath was then heated at the rate of approximately 1.5 ◦ C/min. Strong autoextension was observed when the oil bath reached 40 ◦ C; thus we conclude that T mi = 40 ◦ C. We also repeated the above experiment with coalescence occurring at 65 ◦ C, followed by cooling (thus, interfacial area decreased before cooling, analogous to Protocol A above). As expected in this case, reheating induced only small changes in drop shape upon heating. The value of T mi = 40 ◦ C is almost identical to the melting temperature of WSB125, thus providing support for the hypothesis that crystallization of interfacially-adsorbed surfactant is indeed the mechanism for interfacial trapping of the surfactant. We also note that the surfactant L35, which is liquid at room temperature, did not show autoextension behavior, further supporting surfactant crystallization at the interface as the trapping mechanism. Such interfacial trapping may be a general feature of crystallizable block copolymers adsorbed at interfaces between noncrystalline phases. 4. Summary and conclusions We studied drop retraction of Pluronic-laden PPG/PEG interfaces in the spinning drop tensiometer. While the motivation for this research was to study the effect of surface-active species on capillary instabilities of elongated drops under quiescent conditions, we instead observed two unusual phenomena that are of interest in their own right. The first is that under certain circumstances, Pluronic-laden elongated PPG drops in PEG do not retract, but instead maintain nonspherical shapes indefinitely. We believe that such nonretraction is attributable to an interfacial yield stress caused by addition of surfactant. The second is that under certain circumstances PPG drops in PEG elongate sharply upon heating. We show that such autoextension is directly related to a decrease in interfacial area at room temperature; the surfactant is then trapped at the interface (likely due to crystallization of the PEG
674
J.D. Martin, S.S. Velankar / Journal of Colloid and Interface Science 322 (2008) 669–674
block) at concentrations far above equilibrium, which induces autoextension upon reheating. These unusual phenomena indicate that at room temperature, the PPG/PEG/Pluronic system is not a model system for two-phase flow experiments. Acknowledgments We are grateful to BASF for providing the PPG, PEG, and Pluronics for this research. This research was funded the ACS Petroleum Research Fund (Grant #39931-G9), and by a National Science Foundation CAREER grant (CBET-0448845). We are grateful to Dr. Steven Hudson, NIST, Gaithersburg for valuable suggestions. References [1] H.A. Stone, L.G. Leal, J. Fluid Mech. 198 (1989) 399. [2] A.J. Ramic, J.C. Stehlin, S.D. Hudson, A.M. Jamieson, I. Manas-Zloczower, Macromolecules 33 (2000) 371. [3] I. Welge, B.A. Wolf, Polymer 42 (2001) 3467.
[4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
M. Malmsten, B. Lindman, Macromolecules 26 (1993) 1282. H.M. Princen, I.Y.Z. Zia, S.G. Mason, J. Colloid Interface Sci. 23 (1967) 99. B. Vonnegut, Rev. Sci. Instrum. 13 (1942) 6. D.D. Joseph, M.S. Arney, G. Gillberg, H. Hu, D. Hultman, C. Verdier, T.M. Vinagre, J. Rheol. 36 (1992) 621. E.H.A. de Hoog, H.N.W. Lekkerkerker, J. Phys. Chem. B 105 (2001) 11636. A. Cohen, C.J. Carriere, Rheol. Acta 28 (1989) 223. M. Tjahjadi, J.M. Ottino, H.A. Stone, AIChE J. 40 (1994) 385. J. Martin, S. Velankar, Macromolecules 38 (2005) 10614. S. Velankar, H. Zhou, H.K. Jeon, C.W. Macosko, J. Colloid Interface Sci. 272 (2004) 172. W. Ramsden, Proc. R. Soc. London A 72 (1903) 156. A.B. Subramaniam, M. Abkarian, L. Mahadevan, H.A. Stone, Langmuir 22 (2006) 10204. S.U. Pickering, J. Chem. Soc. Abstracts 91, 92 (1907) 2001. A.B. Subramaniam, M. Abkarian, L. Mahadevan, H.A. Stone, Nature 438 (2005) 930. C.F. Brooks, G.G. Fuller, C.W. Frank, C.R. Robertson, Langmuir 15 (1999) 2450. P. Erni, P. Fischer, E.J. Windhab, V. Kusnezov, H. Stettin, J. Lauger, Rev. Sci. Instrum. 74 (2003) 4916. B.P. Binks, T.S. Horozov (Eds.), Colloidal Particles at Liquid Interfaces, Cambridge Univ. Press, Cambridge, 2006.