Transcription

1 This article was originally published in a journal published by Elsevier, and the attached copy is provided by Elsevier for the author s benefit and for the benefit of the author s institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues that you know, and providing a copy to your institution s administrator. All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier s permissions site at:

2 Journal of Colloid and Interface Science 310 (2007) Abstract Effect of interfacial rheology on model emulsion coalescence II. Emulsion coalescence H.W. Yarranton,P.Urrutia,D.M.Sztukowski Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, AB, T2N 1N4, Canada Received 20 October 2006; accepted 17 January 2007 Available online 14 February 2007 In Part I, surface pressure isotherms were measured for model interfaces between a dispersed water phase and a continuous phase of asphaltenes, toluene, and heptane. Here, the coalescence rate of model emulsions prepared from the same components is determined from measured drop size distributions at 23 C. A correlation is found between the initial coalescence rate and the interfacial compressibility. It is shown that the change in coalescence rate as the emulsion ages and coalesces can be predicted from surface pressure isotherm data also obtained at 23 C. The stability of the emulsions was further assessed in terms of free water resolved after a treatment of heating at 60 C and centrifugation. The emulsions were aged up to 24 h prior to treatment. The free water resolution appears to correlate to the capacity for coalescence of the interfacial film; that is, to the product of the initial film compressibility and (1-CR), where CR is the film ratio at which the film crumples Elsevier Inc. All rights reserved. Keywords: Asphaltenes; Emulsions; Interfacial rheology; Interfacial films; Emulsion stability 1. Introduction As discussed in Part I of this work, oilfield water-in-oil emulsions are believed to be stabilized, at least in part, by a film of adsorbed asphaltenes. However, the mechanism by which the asphaltene film stabilizes the emulsion is not well understood. Without this understanding, it has proven challenging to optimize the treatment of these emulsions. The objective of this work is to identify this mechanism and the relevant physical properties of the interfacial film. In Part I of this work, the interfacial elastic and viscous moduli and surface pressure isotherms were measured for asphaltene films. It was shown the total modulus of the interface correlated to the reciprocal of the initial compressibility from the surface pressure isotherms. Hence, the surface pressure isotherms provided similar data as the modulus measurements on the properties of an uncompressed interface. The surface pressure isotherms also showed that the asphaltene films underwent an apparent phase change as the interface was com- * Corresponding author. address: hyarrant@ucalgary.ca (H.W. Yarranton). pressed, as shown in Fig. 1. The compressibility of the second phase was typically five times lower than the first phase. Upon further compression, the interface crumpled; that is, the compressibility was reduced to zero. The decrease in compressibility implies that at least some of the asphaltenes are irreversibly adsorbed in the interface, consistent with previous observations [1,2]. Now consider the coalescence of an emulsion with an irreversibly adsorbed asphaltene film on the interface. As coalescence occurs, the total interfacial area decreases and the interfacial film is compressed. Since the adsorbed material cannot desorb, the asphaltene film will reach the low compressibility second phase and, eventually, the crumpling point. As the interface becomes less compressible, there will likely be a greater resistance to coalescence. For example, a lower compressibility interface will likely inhibit droplet deformation and the potential for water bridging between droplets. Hence, the coalescence rate is expected to decrease and eventually approach zero. In Part II of this work, we aim to determine if the emulsion coalescence rate and emulsion stability can be predicted from the surface pressure isotherm data collected in Part I. The coalescence and emulsion stability experiments are performed at 23 and 60 C, respectively, on model systems of water, as /$ see front matter 2007 Elsevier Inc. All rights reserved. doi: /j.jcis

3 254 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) Fig. 1. Surface pressure isotherm for 20 kg/m 3 asphaltenes in 25/75 heptol versus water at 23 C. phaltenes, toluene, and heptane with interfaces aged for up to 24 h. The experimental coalescence rate at 23 C is determined from the change in the drop size distribution during coalescence. A methodology is developed to predict the coalescence rate from film compressibilities measured at 23 C. A correlation for estimating emulsion stability at higher temperature treatments is also discussed. 2. Experimental methods 2.1. Materials Athabasca coker-feed bitumen was obtained from Syncrude Canada Ltd. and had been treated to remove most of the sand and water in preparation for upgrading. Reagant-grade n-heptane and toluene were purchased from Van Waters & Rogers Ltd. Reverse osmosis water was supplied by the University of Calgary water plant. Preparation of the asphaltene samples and the asphaltene content of the Bitumen 1 and Bitumen 2 samples are reported in Part I of this work. Many of the experiments use mixtures of toluene and heptane and, for convenience, a solution of X vol% heptane and Y vol% toluene is denoted as X/Y heptol Emulsion preparation, coalescence rate, and stability tests The procedures for emulsion preparation and testing were developed previously [3] and are summarized below. Model emulsions were prepared with asphaltenes, toluene, heptane, and water. A known mass of asphaltene was dissolved in toluene, after which heptane was added to make up the desired heptane to toluene by volume mixture. The heptane fraction in heptol was varied from 0 to 50%. The mixture was sonicated for 20 min prior to and after the addition of heptane to ensure complete asphaltene dissolution and mixture homogeneity. Water in a 40 vol% ratio was added dropwise to the hydrocarbon phase while the mixture was homogenized at 18,000 rpm with a CAT-520D homogenizer for 5 min. The emulsion was allowed to settle for anywhere between 1.5 and 24 h. In all cases, a continuous phase consisting solely of asphaltenes dissolved in heptol and a concentrated emulsion phase had separated after the minimum aging time, i.e., 1.5 h. To measure the drop size distribution, the concavity of a hanging drop slide was filled with a sample of the continuous phase. A small drop of the emulsion phase was added and examined with a Carl Zeiss Axiovert S100 inverted microscope equipped with video camera and Image Pro image analysis software. In this work, approximately drops were used. The drop size distribution did not change during the first 1.5 h and no free water was observed during settling in any experiment. The coalescence rate was assessed from the change in mean diameter of the emulsion over time, as will be discussed later. The mean diameter was found from drop size distributions of samples taken from a settled emulsion after 1.5, 4, 8, 16, or 24 h of aging. The drop size distributions were measured with a Carl Zeiss Axiovert S100 inverted microscope equipped with a video camera and image analysis software. Approximately drops were used, giving an expected error of 5 10% in the Sauter mean diameter [4]. The mean radius was more sensitive to the accuracy of the count of droplets less than 5 to 10 microns in diameter and repeatability was sometimes as low as ±30%. More details on the drop size distributions of water-in-asphaltene/heptol emulsions are provided in the supplemental information and elsewhere [3]. The stability of a given emulsion was assessed by measuring the water separated from the emulsion as a function of time. Samples of the settled emulsion phase (after aging for a given time at room temperature) were transferred into 12 cm 3 graduated centrifuge tubes. The tubes were closed with caps to prevent solvent evaporation and centrifuged for 5 min at 4000 rpm (1640 RCF) and placed in a constant temperature water bath at 60 C. After 2 h the emulsions were centrifuged for 5 min and the amount of resolved water was measured. The centrifuge tubes were returned to the heating bath for another 2 h and then centrifuged again before measuring the amount of resolved water. This procedure was repeated for a total treatment time of 8 h. The amount of resolved water was reported as the percent of total water volume contained in the given emulsion sample. The relative stability of all the emulsions was assessed by comparing the amount of resolved water at a given destabilization time Mass of asphaltenes on the interface The mass of asphaltenes on the emulsion interface was determined before and after coalescence to confirm the irreversibility of asphaltene adsorption. A brief description of the method follows but details are provided elsewhere [5]. The asphaltene mass surface coverage, Γ A, is given by ) (1 Ceq A (1) CA 0, where m t is the total mass of asphaltenes in the emulsion, d 32 is Γ A = m td 32 6V w the Sauter mean diameter of the emulsion droplets, V w the total volume of the water phase, C eq A the asphaltene equilibrium

4 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) (a) Fig. 2. Effect of aging time on (a) the inverse square of the mean drop diameter and (b) the calculated rupture rate for emulsions prepared from water and solutions of 5, 10, and 20 kg/m 3 asphaltenes in toluene at 23 C. concentration, and CA 0 the initial asphaltene concentration. The initial asphaltene concentration and the total volume of water are directly controlled variables while the Sauter mean diameter is calculated from the measured drop size distributions. The asphaltene equilibrium concentration is calculated from the mass of asphaltenes in the decanted continuous phase and the volume of that phase. The mass was determined after evaporating the solvent in air at 23 C and was repeatable to ±0.3%. 3. Results and discussion If an emulsion is stabilized with an irreversibly adsorbed film, the film will be compressed as the emulsion coalesces. As the molecules become more tightly packed, the compressibility of the film is reduced, as was shown in Part I of this work. The compressibility of the film at any given film ratio is also reduced as the film ages. Our hypothesis is that the coalescence rate of this type of emulsion is related to the compressibility of the film. Hence, to predict coalescence, the coalescence rate must be correlated to film compressibility and the change in film compressibility must be tracked as the film ages and as it is compressed during coalescence. To find a correlation between coalescence and film compressibility, the coalescence rate of the model emulsions is first determined from the drop size data at 23 C. Then, the initial coalescence rate is correlated to the initial phase 1 and phase 2 compressibilities. The time dependence of the surface pressure isotherm data is determined; that is, the trends of phase 1 and phase 2 compressibility, the phase change film ratio, and the crumpling ratio are determined. Finally, the change over time in compressibility and drop size of the model emulsions at 23 C are predicted Coalescence rate of model emulsions For a concentrated emulsion, such as the settled emulsion phase in these experiments, coalescence depends on the rupture (b) frequency of the interface. Deminiere et al. [6] showed that, for a monodisperse system, the mean diameter changes with time according to 1 (2) R 2 = 1 R0 2 8π 3 ωt, where R is the average drop radius, R 0 is the initial drop radius, ω is the rupture frequency, and t is time. They also observed that thesamescalingof1/r 2 versus t is expected for polydisperse systems. In Deminiere et al. s work, there were no aging effects and the rupture rate was constant over time. With the asphaltene systems, the interface becomes more rigid with time and the rupture rate is expected to decrease. Therefore, 1/R 2 was plotted versus time for emulsions prepared from 5, 10, and 20 kg/m 3 asphaltenes in toluene, 25/75 heptol, and 50/50 heptol at 23 C, as shown in Supplementary material and Fig. 2a. The data were fitted and the rupture rate at any time was determined from the slope, as shown in Fig. 2b. In all cases, the calculated rupture rates decreased exponentially to near zero values after approximately 4 to 8 h. Note that all of the coalescence experiments were conducted prior to any heating and centrifugation and no free water was observed during the experiments Reversibility of asphaltene adsorption during coalescence Both the occurrence of crumpling and the dramatic reduction in rupture rate suggests that the asphaltenes are irreversibly adsorbed on the interface. Simply, if the asphaltenes were free to leave the interface, the rupture rate would be constant and crumpling would not occur. To confirm the irreversibility of adsorption, the mass of asphaltenes on the emulsion interface was determined for emulsions prepared with 5, 10, and 20 kg/m 3 asphaltenes in toluene, 25/75 heptol, and 50/50 heptol at 23 C. Table 1 shows that, in most cases, the mass on the interface was constant over time suggesting that the asphaltenes were

5 256 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) Table 1 Effect of coalescence on the mass of asphaltenes on emulsion interfaces Aging time (h) Mass over initial mass (%) Mass per area (mg/m 2 ) 5kg/m 3 10 kg/m 3 20 kg/m 3 5kg/m 3 10 kg/m 3 20 kg/m 3 Toluene /75 heptol /50 heptol Fig. 3. Correlation between initial rupture rate (1.5 h of aging) and initial interfacial compressibility (60 min of aging) for 5, 10, and 20 kg/m 3 asphaltenes in toluene, 25/75 and 50/50 heptol at 23 C. irreversibly adsorbed on the interface. The exceptions were 10 kg/m 3 asphaltenes in toluene and 20 kg/m 3 asphaltenes in toluene and 25/75 heptol. In these systems, 10 to 40% of the asphaltenes were expelled from the interface during 24 h of coalescence. Not surprisingly, asphaltene adsorption is stronger and less reversible in more paraffinic solvents. It is not known why the adsorption is partially reversible at the higher concentrations Correlation of coalescence rate and interfacial compressibility Recall our hypothesis that, for an emulsion with an irreversibly adsorbed interfacial film, the rupture rate of the settled emulsion depends on the interfacial compressibility. Fig. 3 shows the initial rupture rate (measured after 1.5 h of settling) versus the initial interfacial phase 1 and phase 2 compressibilities (measured at 60 minutes). The drop size distributions from Fig. 4. Effect of aging time on the phase transition and crumpling film ratios for 20 kg/m 3 asphaltenes in 25/75 heptol at 23 C. which the rupture rates were determined were measured in a previous project and the earliest measurements were taken at 1.5 h. The compressibility data with the nearest aging time were measured at 60 min. The data of Fig. 3 were then fitted with the following equation: ω = ln{c I } , where c I is the interfacial compressibility. The correlation is merely adequate (±30%) but sufficient to test the hypothesis Prediction of emulsion coalescence If we assume that Eq. (3) holds at all film ratios and aging times, then the coalescence rate can be predicted with reasonable accuracy as long as the interfacial compressibility can be predicted. To predict the interfacial compressibility of a coalescing emulsion, it is necessary to account for the age of the interface and the film ratio of the interface. To facilitate the calculations, the phase 1 and phase 2 compressibilities as well as the film ratios at the phase transition (PR) and crumpling point (CR) were determined from the surface pressure isotherms at different ages (see data in supporting information for Part I of this work). The data were plotted versus time and the trends were fitted, as shown in Figs. 4 and 5. Now, since compressibility depends on the film ratio, which in turn depends on the coalescence rate, a simultaneous solution of compressibility and coalescence rate is required. The prediction method is illustrated conceptually in Supplementary material and the procedure is outlined below. Step 1: For any given time, t (n), the phase of the interfacial film is identified using Fig. 4 and the film ratio at that time, (A/A 0 ) (n). Step 2: Once the phase is known, the film compressibility at that time, c I(n), is found from Fig. 5. Step 3: The coalescence rate at that time, ω (n), is then determined from Eq. (3). (3)

6 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) Fig. 5. Effect of aging time on the phase 1 and phase 2 interfacial compressibilities for 20 kg/m 3 asphaltenes in 25/75 heptol at 23 C. Step 4: The mean drop radius is updated with a rearrangement of Eq. (2): [ ] R (n) = R (n 1) + (4) 1 8π 3 ω. (n) t R 2 (n 1) Step 5: The film ratio is updated. Recall that the total area of the interface of monodisperse droplets is given by A = 3V R where A is the interfacial area, V is the volume of the dispersed phase, and R is the radius. Hence, the film ratio after some coalescence is given by ( ) A A 0 (n) = R 0 R (n). In reality, the droplets are polydisperse and the Sauter mean diameter should be substituted for the mean diameter. However, the droplets were assumed to be monodisperse in the model. Also note that the initial mean radius microns could not be measured directly but a value of 5 microns was found to provide the best fit of the early time data. Step 6: The time is updated using a fixed increment, t, typically 0.2 to 0.5 h. Return to Step 1. Fig. 6 shows the predicted interfacial compressibility and mean drop radius for an emulsion prepared from water and 20 kg/m 3 asphaltenes in 25/75 heptol. The predicted mean radius is within the experimental error of the measured radius (approximately ±2 microns). The change in mean radius was predicted for emulsions prepared from the different solvents and at asphaltene concentrations of 20, 10, and 5 kg/m 3,as shown on Fig. 7 and in the supporting information. In all cases, the predictions matched the measured values within experimental error. Note that the film properties were not measured at 5kg/m 3. The properties measured at 1 and 10 kg/m 3 were averaged to model the coalescence rate at 5 kg/m 3. The successful predictions support the hypothesis that the coalescence of asphaltene stabilized emulsions depends on the compressibility of the asphaltene film Stability of model emulsions Only limited coalescence and no free water resolution were observed for emulsions settled under normal gravity at 23 C. To assess the stability of the emulsions, they were heated to 60 C and centrifuged for 5 min every 2 h. The amount of free water resolved is an indicator of the relative stability of the emulsion to coalescence. Fig. 8 shows the effect of asphaltene concentration and heptane-to-toluene ratio on the free water resolution after 1.5 h of aging. The emulsions are relatively unstable at asphaltene concentrations below 2 to 5 kg/m 3. At these low concentrations, there are insufficient asphaltenes in the solution to saturate the large interfacial area created by the homogenizer (a) (b) Fig. 6. Predicted change in interfacial compressibility (a) and mean droplet diameter (b) of a coalescing emulsion prepared from water and a solution of 20 kg/m 3 asphaltenes in 25/75 heptol at 23 C.

7 258 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) (a) Fig. 7. Predicted change in interfacial compressibility (a) and mean droplet diameter (b) of a coalescing emulsion prepared from water and solutions of 20 kg/m 3 asphaltenes in toluene, 25/75 and 50/50 heptol at 23 C. Fig. 8. Stability after 8 h of treatment at 60 C of emulsions prepared from water and solutions of 5, 10, and 20 kg/m 3 asphaltenes in toluene, 25/75 and 50/50 heptol at 23 C. and the emulsion rapidly coalesces until the interface is saturated [3]. After this initial rapid coalescence, the drop sizes are relatively large (see Supplementary material) and the emulsion is relatively unstable. In 25/75 and 50/50 heptol, the maximum stability of the emulsions occurs when there are just sufficient asphaltenes to saturate the initial interface (an initial concentration of approximately 2 kg/m 3 ). At higher asphaltene concentrations, emulsion stability decreases. Emulsion stability also increases as the interface ages, as shown in Fig. 9 for the 25/75 heptol systems. Aging also increased the stability of emulsions prepared from toluene although to a lesser extent. The previous results indicate that the most extensive coalescence occurs when both the initial film compressibility is high and the ultimate film ratio at which crumpling occurs is low. Therefore, we attempted to correlate the free water resolution from an emulsion stability test with the parameter c I0 (1 CR), (b) Fig. 9. Effect of aging at 23 C prior to treatment on the free water resolved from emulsions prepared from 25/75 heptol and treated (after aging) for 1.5 h at 60 C. where c I0 is the initial compressibility and CR is the crumpling film ratio. The initial compressibility correlates to the initial rupture rate. The expression (1 CR) represents the capacity for the interface to compress. Hence, the total expression is an approximate indicator of the capacity of the film to coalesce. Fig. 10 is a plot of the free water resolution from Figs. 8 and 9 versus c I0 (1 CR) determined from Table 2 in Part I of this work. Data for all three solvents and asphaltene concentrations are included. The film properties used in Fig. 10 for 5kg/m 3 are again an average of the measured properties at 1 and 10 kg/m 3. Data for coalescence capacity at 1 kg/m 3 could not be included because the free water resolution data at this concentration was affected by the high initial drop size. There appears to be a reasonably good correlation between free water resolution and the coalescence capacity. There also appears to be a threshold at c I0 (1 CR) = 0.2 m/mn below which

8 H.W. Yarranton et al. / Journal of Colloid and Interface Science 310 (2007) Fig. 10. Relationship between free water resolved from emulsions at 60 Cand c I0 (1 CR) determined at 23 C (5, 10, and 20 kg/m 3 asphaltenes in toluene, 25/75 and 50/50 heptol). Open symbols are based on an extrapolated CR. little free water can be resolved with the given treatment. Of course, this particular threshold only applies to the emulsions and treatments considered here. However, the coalescence capacity may be useful as a relative indicator for the stability of different emulsions stabilized by irreversibly adsorbed films. For example, the effectiveness of demulsifiers designed to promote the coalescence of asphaltene-stabilized emulsions could potentially be screened based in their effect on the coalescence capacity. To assess this potential, surface pressure isotherm and coalescence rate data are required for model and crude oil systems with demulsifiers. 4. Conclusions The coalescence rates of settled model emulsions consisting of water and solutions of asphaltenes, toluene, and heptane were successfully predicted based on the compressibility of asphaltene interfacial films. A single correlation of rupture rate to interfacial compressibility was found for the different solvents and over a range of asphaltene concentrations from 5 to 20 kg/m 3. It was necessary to account for the change in compressibility resulting from both aging and film compression. A coalescence capacity for an interfacial film was defined as the product of the initial interfacial compressibility (a measure of the initial coalescence rate) and 1 CR (a measure of how much the film can be compressed). The stability of model emulsions was measured in terms of free water resolution after a given treatment of heating and centrifugation. The free water resolution was found to correlate with the coalescence capacity. If the correlation holds for other emulsions stabilized by irreversibly adsorbed films, then the coalescence capacity may be an indicator of relative emulsion stability. In more general terms, this work demonstrates that film compressibility is a key property for the class of emulsions stabilized by irreversibly adsorbed interfacial films. Surface pressure isotherms were shown to be a suitable technique to determine film compressibility. It was demonstrated that emulsion coalescence could be predicted from film compressibility data. Such data can potentially provide a relatively simple means to compare the stability of different emulsions and perhaps to assess demulsifiers. The coalescence capacity considered here is offered as a potential comparator. Acknowledgments The authors thank Syncrude Canada Ltd. for providing samples and financial support. We are also grateful to AERI and NSERC for financial support. We thank Dr. Kevin Moran at Syncrude and Dr. Shawn Taylor at Oilphase-DBR (Schlumberger) for helpful discussions. Supplementary material The online version of this article contains additional supplementary material. Please visit DOI: /j.jcis References [1] E.M. Freer, C.J. Radke, J. Adhes. 80 (2004) 481. [2] T. Yeung, T. Dabros, J. Czarnecki, J. Masliyah, Proc. R. Soc. London Ser. A 455 (1999) [3] O.V. Gafonova, H.W. Yarranton, J. Colloid Interface Sci. 241 (2001) 469. [4] W.J. Dixon, F.J. Massey, Introduction to Statistical Analysis, McGraw Hill, New York, [5] D.M. Sztukowski, M. Jafari, H.W. Yarranton, J. Colloid Interface Sci. 265 (2003) 179. [6] B. Deminiere, A. Colin, F.L. Calderon, J. Bibette, in: B.P. Binks (Ed.), Modern Aspects of Emulsion Science, The Royal Society of Chemistry, Cambridge, UK, 1998, p. 261.