The stability of water-in-crude and model oil emulsions.

Sullivan, Andrew P. ; Zaki, Nael N. ; Sjoblom, Johan 等

INTRODUCTION

The existence of emulsions in both the production and refining of crude oil is one of the most persistent problems facing the petroleum industry. At the production end, emulsions are generated when water is co-produced with petroleum or pumped into the well to aid in petroleum recovery and this mixture passes through the wellhead, pipe bends, and choke valves. At the refinery, water is added to generate a large oil-water interfacial area to aid in the extraction of salts from the crude oil. The emulsions produced from these processes do not easily resolve into neat crude and water phases and a certain volume of the emulsion remains (Obah, 1988; Schramm 1992; Schubert and Armbruster, 1992; Sjoblom et al., 1997b; Tissot and Welte, 1984). The emulsified water can corrode refinery equipment such as distillation columns and dissolved salts in the water can poison catalysts (Obah, 1988). Some emulsions are very viscous and will foul machinery if allowed to continue through the refining processes. To minimize these problems, emulsions undergo separation procedures such as electrostatic coalescence and demulsifier addition. There is a great deal of ongoing research attempting to correlate crude oil and demulsifier properties with the effectiveness of these separation methods (Bhardwaj and Hartland, 1993; Breen et al., 2003; Hirato et al., 1991; Kim et al., 1996; Krawczyk et al., 1991; Malhotra and Wasan, 1986; Mohammed et al., 1994; Mukherjee and Kushnik, 1988; Singh, 1994; Sjoblom et al., 1990; Wasan, 1992; Zaki et al., 1996a, b).

Surface-Active Components in Petroleum

The mechanism of water-in-crude oil emulsion stabilization is not fully understood. The primary means, however, by which water droplets are stabilized, appears to be the formation of a viscoelastic, mechanically strong film at the droplet interface composed of asphaltenes (Cratin, 1969; Dodd, 1960; Graham et al., 1983; Mackay et al., 1973; Strassner, 1968; Taylor, 1992; Siffert et al., 1984). Asphaltenes are the portion of petroleum insoluble in an n-alkane solvent (usually n-pentane or n-heptane). Asphaltenes are polydisperse with respect to molecular weight and chemical functionality. The common structural theme among asphaltenes is a planar, polyaromatic, fused ring core imbedded with polar functionality, surrounded by aliphatic side chains and naphthenic rings (Figure 1a) (Bestougeff and Byramjec, 1994; Cimino et al., 1995; Dickie and Yen, 1967; Pelet et al., 1985; Speight, 1986, 1989; Yen, 1992). In petroleum, asphaltene molecules aggregate to minimize the interactions between their polar cores and the non-polar solvent and to increase the extent of [pi]-bond overlap (Moschopedis et al., 1976; Speight and Moschopedis, 1979; Yen, 1992). Asphaltenes are surface-active, having hydrophobic and hydrophilic portions, however, unlike typical surfactants, asphaltenes and their aggregates likely orient parallel to the interface, exposing their cores to the water phase. Scanning tunnelling microscopy has been utilized to image the surface of pyrolytic graphite with adsorbed asphaltenes (Watson and Barteau, 1994). The surface was highly ordered and flat with periodic features extending from the surface ca. 2 [Angstrom]. The results indicated asphaltene sheets oriented parallel to the surface with the terminal carbons of the aliphatic side-chains extending from the surface. The orientation of asphaltenes upon exposure to graphite may be similar to that with water interfaces, both surfaces preferentially interact with asphaltenic cores.

[FIGURE 1 OMITTED]

Resins, another class of surface-active molecules in petroleum, have their hydrophobic and hydrophilic portions on opposite ends, encouraging interfacial adsorption. Resins solvate asphaltenic aggregates by adsorbing to the faces of these aggregates, a process sometimes referred to as peptization (Figure 1a) (Kawanaka et al., 1989; Sheu et al., 1991; Speight and Moschopedis, 1979). The interaction of asphaltenes and resins and the subsequent adsorption to the water-oil interface forms a mechanically strong, viscoelastic, cross-linked network preventing droplet coalescence (Figure 1b). Mackay and coworkers (Mackay et al., 1973) observed this structure when they formed a water droplet in an asphaltene-containing oil phase, allowed time for adsorption to the interface, and withdrew the water. This process yielded a sac-like structure that folded as the water was removed but did not rupture.

Emulsions produced via asphaltene-resin aggregation and adsorption to the interface are a major concern to the petroleum industry. As oil is produced, the lighter material with lower asphaltene and water content is encountered first. As wells near completion, the light end supplies are exhausted, and the heavier, higher asphaltene content petroleum is produced. As heavier crudes are produced, the industry will have to develop ways to better handle and understand asphaltene physical chemistry and the emulsion problem.

Methods of Measuring Emulsion Stability

There have been many attempts at quantifying and correlating emulsion stability. So-called bottle tests and droplet size observations are common (Aveyard et al., 1990; Bhardwaj and Hartland, 1993, 1994; Breen et al., 2003; Ese et al., 1997; Gelot et al., 1984; Lawrence and Killner, 1948; McLean and Kilpatrick, 1997a, b; Menon and Wasan, 1988; Papirer et al., 1982; Shetty et al., 1992; Wasan, 1992; Yan and Masliyah, 1995a, b). In bottle tests, an emulsion is created and the volume of the disperse phase resolved over time is observed. To produce emulsions that will show resolution in a relatively short period of time, the level of mixing energy must be low, or the composition of the oil system must be one that produces weak emulsions. Both strong and weak emulsions resulting from a range of mixing energies and crude compositions are encountered in actual production and refinery conditions. Another possibility is to produce emulsions with a high mixing energy and propensity to emulsify and measure the water resolved with the application of a centrifugal field. This will produce a range of resolved water volumes, however this technique is very dependent upon protocol and only provides relative stabilities. The measure of the volume of water resolved due to coalescence of water droplets in a high centrifugal field may not be a true measure of real emulsion strength (Carroll, 1976). In refinery situations, centrifugation on a large scale for all emulsions is very expensive and not widely performed. Therefore, any predictive correlation with this technique as a basis is of limited value.

Droplet size is commonly measured by means of optical microscopy or light scattering techniques (Isaacs et al., 1990; Menon and Wasan, 1984). Optical microscopic observations of emulsion samples are difficult to perform without disturbing the emulsion system. In order to generate an accurate count and droplet size distribution, the emulsion droplets must be confined between a glass slide and cover slip to provide an unobstructed view of one droplet layer. This distorts droplet size and may result in some droplet coalescence, producing an inaccurate droplet count and size distribution.

[FIGURE 2 OMITTED]

The use of light scattering techniques is troublesome as well (Boyd et al., 1972; Carroll, 1976; Mikula, 1992; Parkinson and Sherman, 1972; Stalss et al., 1991). Petroleum is opaque so a very thin sample of emulsion must be observed. The observable path length may be very small for high asphaltene content systems, complicating sample preparation and observation. The conversion of light scattering data into droplet size information requires the assumption of a droplet size distribution. Lastly, light scattering techniques can not distinguish between coagulated and individual water droplets. In stable petroleum emulsion systems, many droplet collisions do not result in coalescence and coagulated droplets are common. With all of these difficulties, light scattering has somewhat limited value for this application.

The primary mechanism of emulsion stabilization is the formation and stability of the interfacial film discussed above. Therefore, a useful, predictive correlation would be one that probes this property directly. In this study, emulsion stability was quantified with the determination of the electric field strength required for emulsion breakdown in w/o emulsions. The technique is based on work by Sjoblom and co-workers (Fordedal et al., 1995; Fordedal et al., 1996; Sjoblom et al., 1997a; Skodvin et al., 1994; Skodvin and Sjoblom, 1996) with modifications in sample cell design, emulsion sampling procedures, and electronic components. To measure stability, an emulsion sample is placed between two electrodes and the voltage is steadily increased (Figure 2). At low voltages, a low level of current is observed due to conduction through the continuous oil phase. As the voltage is increased, water droplets begin to coagulate in small chains parallel to the electric field due to polarization of the water droplets from the movement of electrolytes in the water phase (Bailes and Larkai, 1982; Chen et al., 1994). As the voltage is increased further, the small chains form bridges spanning the entire gap between the electrodes. Finally, at a certain electric field strength, the electromotive force exerted on the electrolyte ions is sufficient to rupture the interfaces separating the water droplets. This phenomenon is observed as a large change in slope of the current versus voltage curve. The electric field (in kV/cm) at which this occurs is termed the critical electric field (cef).

The cef is a direct measure of the strength of the asphaltene-resin interfacial film and is simple to perform. Emulsion droplets encounter similar forces at the refinery in ac electrostatic coalescers. In these units, droplets coalesce due to increased collision and thinning of interfaces as droplets are deformed in an alternating electric field (Urdahl et al., 1996; Wang and Yan, 1994). In the cef technique, a dc power supply is used and coalescence does not occur by the same mechanism, but the important feature, thinning or disruption of the interface in the presence of an electric field, is identical. Therefore, no artificial artifacts of the experimental procedure are introduced in the measure of emulsion stability and it can easily be adapted to field work and testing. The technique produces a unique number for a given emulsion and is independent of the choice of experimental conditions which do not directly affect emulsion stability (gap width, sample size, etc.), unlike other techniques which are dependent on sample handling and testing procedures. The cef only depends on variables, such as temperature, water volume fraction, droplet size distribution, and asphaltene content, which control emulsion stability. Moreover, we have discovered that as long as the droplet size distribution is not bimodal with large droplets in the range of 50-100 ?m or more, then the measurement does not appear to depend on the fine features of droplet size in the range of 10 [micro]m or less. The measurable range of emulsion stability is large (0.01-5 kV/cm) and the test is rapid and reproducible. However, the cef technique does not provide an answer as to the extent of the water phase that separates upon demulsification. Neither do we get any information about the stability of the residual emulsion droplets remaining after field-induced coalescence.

In this study we have used the cef technique to develop a water-in-petroleum emulsion stability correlation with properties of the petroleum using information gathered from an investigation of cef in model oil systems. These tests have been performed at two different water/oil ratios producing different results. The development of this correlation has revealed important insight into the driving mechanisms for asphaltene aggregation and emulsion stabilization. Observations of cef versus time for many petroleum emulsions have led to the development of a simple kinetic model for the build-up of the interfacial film due to asphaltene adsorption that has revealed information on the effects of asphaltene chemistry and solvency. The similarities of petroleum results with those of model oils show that the observed effects with petroleum truly are due to asphaltenic films.

EXPERIMENTAL

Materials

A variety of crude oils were used for this study. Properties important for emulsion stability are listed in Table 1 for the crude oils investigated as well as others present in our laboratory. One important set of properties is the asphaltene and resin content. For the crude oils studied, these values range from 0.79% to 14.8% (w/w) for asphaltenes and 3.24 to 20.5% (w/ w) for resins, yielding a relatively narrow range of resin to asphaltene ratios (R/A = 0.90-4.25, except Alba (AL), R/A = 6.17). The asphaltene content was measured by precipitation in n-heptane and the resin content was obtained by sequential elution chromatography of the deasphalted crude adsorbed on silica gel (McLean and Kilpatrick, 1997a).

Another important set of properties for the crude oils is the hydrogen to carbon atomic ratio (H/C). This number provides a quantitative gauge of the aromatic vs. aliphatic nature of a chemical species (e.g. [H/C.sub.heptane]= 2.29, [H/C.sub.toluene] = 1.14). The ratios reported in the Table correspond to the whole crude oil, the asphaltene fraction, the resin fraction, the crude after removing the asphaltenes (DeA crude), and the difference in H/C of the DeA crude and asphaltenes ([DELTA] H/C). The hydrogen and carbon contents were measured by elemental analysis performed by Galbraith Laboratories. The H/C ratios of the DeA crudes were calculated from hydrogen and carbon contents of the whole crudes, asphaltenes, and the concentration of asphaltenes in the whole crude. Elemental analyses of asphaltenes were performed with a Perkin Elmer 2400 Series II CHNS/O Analyzer.

Other properties reported include density, which varies from 0.833 to 1.01 g/mL, and viscosity, which ranges from 3.14 cP to 2310 cP (3.14 to 2310 mPa-s) at 100[degrees]F (49[degrees]C) (except for SCS, a waxy crude with an unmeasurable viscosity below its pour point). These properties control the shear field in the oil sample during emulsification. Droplets formed during emulsification immediately begin to coalesce as they collide with other droplets if not stabilized by an asphaltenic film. When emulsification is completed, new droplets are no longer formed but coalescence continues between unstable droplets. Asphaltenic films build up over time and resist coalescence to a greater extent at longer times. The balance between the rates of asphaltene adsorption and droplet-droplet collisions determines the ultimate droplet size distribution and stability of the emulsion. As a result, factors that govern asphaltene and water droplet mobilities in the oil phase may be important to consider for the development of an emulsion stability correlation. For the crude oils in this study, density was measured pycnometrically, the viscosity was measured with a Rheometrics couette rheometer, and the kinematic viscosity was calculated from these two numbers.

Deionized water with 1% (w/w) NaCl added was used for creating the emulsions. The salt was necessary for the measurement of critical electric field. After NaCl addition, the pH of the water was adjusted to 6 with dilute HCl and NaOH.

Model oils used in this investigation consisted of n-heptane and toluene with asphaltenes fractionated from Arab Heavy (AH) [also known as Safaniya] crude oils. All solvents used were HPLC grade and supplied by Fisher Scientific.

Emulsion Preparation

Prior to emulsification, petroleum was loaded into a metal cylinder that was sealed and placed in a 100[degrees]C oven for 2 h in order to erase the thermal history and ensure the complete melting of wax crystals. The cylinder was then placed in a 60[degrees]C water bath for 30 min, after which a sample was transferred to a polyethylene jar with the appropriate amount of water (also at 60[degrees]C). Emulsions were prepared at two different oil/water ratios: 4/6 and 7/3 (v/v). The crude oil and water were mixed for 5 min (3 min at the level of the original bulk oil/water interface and 2 min at the bottom of the jar) at 15 000 rpm with a Virtis VirtiShear Cyclone IQ Homogenizer, using a 6 mm diameter internal shaft rotor/stator assembly with a gap width of 0.127 mm. The mixing was performed in a water bath to maintain the system at 60[degrees]C. After emulsification, the sample was placed in a 60[degrees]C oven until stability testing was performed.

[FIGURE 3 OMITTED]

For model oil experiments, AH asphaltenes were dissolved in toluene for 2 h. After this period, heptane was added to adjust the solvent to the appropriate degree of aromaticity. The model oils (referred to as heptol) varied from 10-60% toluene (v/v) in heptane. After 2 h of mixing, the appropriate amount of water was added and the mixture was emulsified for 3 min (2 min at the bulk oil/water interface and 1 min at the bottom of the jar). The temperature was not controlled for the model oil experiments because no wax was present as in the crude oil.

Critical Electric Field Measurement

The stability of the water-in-crude oil emulsions was gauged with critical electric field (cef) measurements. To measure cef, an emulsion sample was placed in the sample cell (in a 60[degrees]C oven) consisting of two, 1.0 cm diameter, gold plated, copper electrodes, separated with Mylar spacers and held in an aluminum casing. Figure 3 shows a side view of the cell. The cell was designed so the gap width could be varied but for all of the experiments it was 0.250 mm. Two holes were drilled through the top for sample introduction. A syringe was used to withdraw a sample from the middle of the emulsion and inject it through one of the holes in the cell. The cell was connected to a HP6634B power supply (0-100 V dc source), controlled by a PC through the use of a HP82350A interface card. Using this card, the power supply was controlled with a Visual Basic program.

After loading the sample in the cell, the voltage between the electrodes was increased in increments of 0.25 V every 5 s and the current was measured 2 s after every step change (to avoid current spikes). All of the emulsions for which we have reported cef showed no water resolved upon visual inspection after a 24 h period. The presence of resolved water would have indicated an emulsion containing large, unstable droplets. All of the emulsions tested were relatively stable and had droplet diameters in the range of 0.5-20 [micro]m. We used the Sauter mean droplet diameter ([d.sub.32]), defined in Equation (1), as the measure for droplet size.

[d.sub.32] = [summation] [n.sub.i][d.sub.3.sub.i]/[summation] [n.sub.i][d.sup.2.sub.i] (1)

In Equation (1), [d.sub.i] is an individual droplet diameter in a micrograph, measured by optical microscopy and image analysis, and [n.sub.i] is the number of droplets of that particular size. Typically, drop diameters of a very narrow diameter range, a so-called bin size, are grouped together as having a common diameter. Droplet size appears to be an important parameter in the stability of emulsions gauged by cef as we will show from the model emulsion results. Because of this sensitivity to droplet size, the emulsions were always sampled from the middle of the emulsion. Microscopic observations were performed in parallel with cef measurements to verify valid sampling procedures. All of the crude oil systems appeared to have droplet size ranges that were close enough to ensure valid comparisons from sample to sample. Model oil emulsion studies were conducted in the same manner except all steps were performed at room temperature.

Solubility Tests

The solubility of asphaltenes in various solvents was tested by preparing a 1.5 mL model oil sample with 1% asphaltenes (w/w) and injecting it through a glass microfibre filter disk attached to the end of a syringe (1.6 [micro]m pore size). After the sample was filtered through the disk, a 1.5 mL aliquot of the model oil solvent was pushed through to rinse any soluble material that may have adsorbed to the microfibres. The rinse step was performed quickly in order to prevent solubilization of any previously precipitated material. Both the filtrate and rinse were collected in a small vial. The asphaltenes in the syringe tip and vial cap that were not dissolved in the solvent were rinsed with methylene chloride into the original vial and labelled as precipitate. The filtrate and precipitate vials were dried in a nitrogen flushed vacuum oven at 70[degrees]C for 2 d and weighed.

RESULTS AND DISCUSSION

Asphaltene-Stabilized Model Oil Emulsions

60% Water

Initial model experiments were performed with organic solutions of varying heptane and toluene content, into which asphaltenes from Arab Heavy crude oil were dissolved. These solutions of varying toluene % and AH asphaltene content were emulsified with water such that the water content was 60% (w/o) and then the cef was measured after aging for 24 h. The results, along with the solubility of AH asphaltenes, are shown in Figure 4. The cef is observed to achieve local maxima with 2 and 3% asphaltene solutions at a toluene concentration of ca. 40%, very close to the solubility limit. This maximum in stability shifts to lower toluene concentrations at an asphaltene concentration of 5%. With the lower (2-3%) asphaltene concentrations, the maximum stability occurs at the limit of solubility, the point at which the asphaltenes are most surface-active and labile due to the fine aggregate size. Beyond the solubility limit at these asphaltene concentrations, sufficient inventory of asphaltenes precipitate, producing flocs which are much more weakly surface-active and insufficiently labile to self-assemble at the interface producing a strong interfacial film. At higher asphaltene concentrations (5-7%), the stability achieves a local maximum slightly beyond the solubility limit because the greater inventory of material enables the droplets to be better coated with asphaltenic film, despite the fact that some material precipitates beyond the solubility limit. This phenomenon is observed with high water content emulsions (e.g. 60%) in which there is a dearth of asphaltenic material until sufficiently high asphaltene concentrations are reached (as we will show below).

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

30% Water

Results of cef for model emulsions produced with AH asphaltene solutions in heptol and 30% water are shown in Figure 5. Values of cef for these emulsions were higher than those with 60% water. The total interfacial area was about the same as that found with 60% water, however the average droplet size was smaller. Larger droplets may facilitate droplet chaining between electrodes or droplet coalescence in the electric field. Similar cef values are observed for 0.5-3% AH asphaltenes at toluene concentrations above the solubility point, but are much lower at 0.25% AH asphaltenes. This suggests that there is some critical concentration of asphaltenes above which interfacial thickness is large enough to provide a strong barrier for droplet coalescence and increased asphaltene adsorption above this has little effect on emulsion stability. As for the 60% water emulsions, we see the maximal cef occurs at lower toluene concentrations as the asphaltene concentration is increased. The peak locations are about the same as those observed for AH asphaltenes with 60% water but in this case more solvent compositions were probed and more detailed shifts can be observed.

Sjoblom and co-workers (Sjoblom et al., 1997a) obtained slightly higher cef values with similar types of experiments. For model emulsions composed of 2% asphaltenes in decane-toluene mixtures, they found cef values ranging from 2.9 kV/cm at 20% toluene to 0.55 kV/cm at 80% toluene. Emulsions were unstable at 100% toluene. Details were not provided of the asphaltene type or water content of the emulsions, but the magnitude of the cef's and trend with solvent aromaticity they observed support the results we have obtained here.

Interfacial Film Thicknesses

Droplet sizes generally decreased, for all solvency conditions, with increasing asphaltene concentration (Figure 6). Generally speaking, one would expect increasing the toluene concentration to decrease the interfacial tension and thus reduce the droplet size of water-in-oil emulsions, all other things being equal. However, as toluene concentration is increased, the surface activity of the asphaltenes decreases and this has a greater impact than interfacial tension. All droplet size data on Figure 6 are for the soluble regime of AH asphaltenes. As expected, the droplet sizes decreased as asphaltene concentrations were increased for all of the solvents shown. For cef testing and droplet size observations, emulsions were sampled from the middle so the results may not be representative of the overall emulsion. For all other asphaltene concentrations, the droplet sizes decreased with concentration up until a critical concentration.

The total water-oil interfacial area, [A.sub.w/o] was calculated assuming a collection of monodisperse spherical droplets:

[A.sub.w/o] = [N.sub.d][A.sub.d] = ([[PHI].sub.d][V.sub.em]/[V.sub.d]) [A.sub.d] = 3[[PHI].sub.d][V.sub.em]/[R.sub.d] (2)

where [N.sub.d] is the number of water droplets, [A.sub.d] is the interfacial area per droplet, [[PHI].sub.d] is the volume fraction of droplets in the emulsion (0.3), [V.sub.em] is the total volume of emulsion (10 [cm.sup.3]), [V.sub.d] is the volume per droplet, and [R.sub.d] is the droplet radius. Film masses were calculated assuming 10% adsorption of asphaltenes based on interfacial film studies performed in our lab at conditions of maximum interfacial activity (Spiecker, 2001). This number is probably dependent on the asphaltene solubility conditions, but the range of solvent compositions was small enough that this is a good approximation.

Figure 7 displays the results of interfacial mass/area calculations for 45-55% toluene with AH asphaltenes (30% water) and 50% toluene with HO asphaltenes (60% water). Interfacial masses increased with asphaltene concentration. The slopes of the lines are similar for all cases for asphaltenes in the soluble regime. For HO asphaltenes in 50% toluene, cef did not increase past 2% asphaltenes due to solubility limitations. All of the trends with AH asphaltenes are linear up to 3%. These results are consistent with the solvency of the asphaltenes observed in previous experiments (Figures 4 and 5). Figure 8 displays cef versus interfacial mass/area for all of the solvency conditions in Figure 7. The cef increased with interfacial mass/ area up to about 1.50 mg/[m.sup.2], above which the cef was a weak function of interfacial mass/area. To calculate a rough estimate of the extent of interfacial coverage in these emulsions, we assumed an asphaltene molecular weight of 1000 g/mole and asphaltene molecular dimensions of 0.5 nm thick with a diameter of 2.0 nm. The molecular weight of asphaltenes has been an extensively researched subject (Acevedo et al., 1985; Acevedo et al., 1992; Ali and Saleem 1988; Ali et al., 1990; Al Jarrah and Al-Dujaili, 1989; Calemma et al., 1995; Cyr et al., 1987; McKay et al., 1978; Storm et al., 1990). Many techniques including, vapour pressure osmometry, gel permeation chromatography, viscometry, and mass spectrometry have been utilized to measure molecular weights with widely varying results. Numbers from 900 to 18 000 g/mole, have been recorded, with the lowest values found with the best solvents. An asphaltene molecular weight of 1000 g/mole was used in our analyses because it is at the low end of observed weights and probably corresponds to individual asphaltene molecules. The asphaltene molecular dimensions used were taken from x-ray diffraction results of Yen (Yen, 1992). We assumed an imperfect stacking of asphaltenes so the asphaltene stack diameter (25 [Angstrom]) is 25% larger than that of the individual molecule (20 [Angstrom]). With these assumptions, the number of molecules corresponding to the interfacial thickness, [n.sub.A,stack], was calculated:

[[GAMMA].sub.A] = [m.sub.A,ads]/[A.sub.w/o] = [[omega].sub.A,ads][m.sub.A,oil]/[A.sub.w/o] = [[omega].sub.A,ads]([w.sub.A,oil][[rho].sub.oil][V.sub.oil])/[A.sub.w/o] (3)

[n.sub.A,stack] = [[GAMMA].sub.A][N.sub.A]/[MW.sub.A] [A.sub.stack] = [[GAMMA].sub.A][N.sub.A]/[MW.sub.A] ([pi][R.sup.2.sub.stack]) (4)

where [[GAMMA].sub.A] is the asphaltene surface concentration, [m.sub.A,ads] is the adsorbed mass of asphaltenes, [m.sub.A,oil] is the mass of asphaltenes in the oil phase, [V.sub.oil] is the volume of the oil phase (7 [cm.sup.3]), [[rho].sub.oil] is the mass density of the oil phase (0.9 g/[cm.sup.3]), [[omega].sub.A,ads] is the fraction of the bulk asphaltene mass that reports to the interface (assumed to be 0.1), [w.sub.A,oil] is the weight fraction of asphaltenes in the oil phase, [N.sub.A] is Avogadro's number, [MW.sub.A] is the molecular weight of the asphaltenes (taken to be 1000 g/mol here) and [R.sub.stack] is the stack radius as estimated above (12.5 [Angstrom]). Evaluating [n.sub.A,stack] using these values, we obtain 4.4 molecules per stack, which seems reasonable given previous estimations that asphaltene aggregates consist of about 5 molecules in the best solvent conditions (Yen, 1992), so the critical interfacial concentration appears to be roughly a monolayer of asphaltene aggregates. The calculated interfacial thicknesses in terms of number of molecules are displayed in Figure 8.

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

Kinetic Model for Interfacial Film Formation

All of the cef measurements reported up until now were performed 24 h after emulsification. Greater differences among the various asphaltene concentrations were shown at shorter times in which the asphaltenic film did not have time to reach the critical thickness. Figure 9 shows cef results for model emulsions at various times from approximately 15 s to 24 h after emulsification. The cef for emulsions with 1% asphaltenes at 40% toluene changed very rapidly during the first half hour and slowly at longer times. The cef was 1.04 kV/cm after 30 min and reached 1.24 kV/cm after 24 h. For 0.5% asphaltenes and 40% toluene, the cef value did not change in the first 30 min, but then the change in cef with time was very similar to that observed for 1% asphaltenes. At short times, the concentration of adsorbed asphaltenes was too low to stabilize droplets and asphaltenes adsorbed to the interface but did not appreciably affect emulsion stability. After 30 min, the coverage reached a high enough concentration that increased adsorption led to increased stability at the same rate as the 1% asphaltene, 40% toluene condition. The results obtained for 1% asphaltenes and 50% toluene were very different from either of the 40% toluene results. In this case the cef increase was much more gradual over the whole 24 h period.

[FIGURE 9 OMITTED]

The results for the model systems are similar in nature to those obtained by Sjoblom and co-workers (Mouraille et al., 1998). Using a 50:50 mixture of "condensate F" and water with various combinations of asphaltenes and resins (0.9-5% A, 110% R), they observed cef over a period of one week. Measured cef values were found to generally increase over the whole aging period, with the majority of the increase occurring in the first day. The maximum electric field measured was 2.00 kV/cm, and is comparable to our findings.

The model for asphaltene adsorption and the stabilization of water-oil interfaces consists of two processes. First, the adsorption of asphaltene aggregates (A) to the interface, and second, the consolidation of the adsorbed asphaltenes into a rigid, cross-linked, interfacial structure (Jeribi et al., 2002). The adsorption mechanism is as follows:

[FORMULA EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

The terms in parentheses are the concentrations of asphaltenes at each condition. Bulk asphaltenes undergo Fickian diffusion to the interface where they immediately adsorb. The adsorbed asphaltenes then consolidate to form an interfacial film. Film consolidation involves the combination of an unconsolidated asphaltene molecule with the consolidated film. The adsorption and consolidation steps, when asphaltenes are adsorbing, are assumed to be irreversible and have rate constants of [k.sub.a] and [k.sub.c], respectively. This process is modelled with the following set of differential equations:

[d[GAMMA].sub.1]/dt = [k.sub.a][c.sup.s.sub.A]([[GAMMA].sub.MAX] - [[GAMMA].sub.1] - [[GAMMA].sub.2])- [k.sub.c][[GAMMA].sub.1][[GAMMA].sub.2] (6)

[d[GAMMA].sub.2]/dt = [k.sub.c][[GAMMA].sub.1][[GAMMA].sub.2] (7)

where [c.sup.s.sub.A] = [2c.sup.bulk.sub.A][(Dt/[pi]).sup.1/2] (8)

and D = kT/6[pi][eta]R (9)

[[GAMMA].sub.max] is the maximum concentration of asphaltenes that can adsorb to the interface, [c.sub.A.sup.bulk] is the concentration of asphaltenes in the bulk oil phase, D is the diffusivity of asphaltenes in the oil, k is Boltzmann's constant, ??is the viscosity of the oil, and R is the radius of asphaltene aggregates. The equation for concentration of asphaltenes near the interface was derived from Fick's law for diffusion assuming the bulk concentration remained constant. For the calculation of asphaltene diffusivity, the aggregate diameter was assumed to be 20 [Angstrom] and is based on observations of Yen in model oil systems. As asphaltenes adsorb, the driving force for further adsorption is reduced due to asphaltene-asphaltene repulsion and shielding of the water phase. This is accounted for with the saturation term, [[GAMMA].sub.max]. Under conditions of identical dispersed water content, we assume cef is proportional to the total amount of adsorbed asphaltenes. The data show that cef increases very rapidly at first followed by a more gradual increase with time, suggesting that adsorbed asphaltenes immediately provide emulsion stability and consolidation contributes to a lesser extent over a long time period. To model this behaviour we set cef equal to a linear combination of [[GAMMA].sub.1] and [[GAMMA].sub.2]:

cef = [[alpha].sub.1][[GAMMA].sub.1] + [[alpha].sub.2][[GAMMA].sub.2] (10)

[[alpha].sub.1] + [[alpha].sub.2] = 1 (11)

The cef value at the first point (3 min) was fit by assuming [[GAMMA].sub.1] and [[GAMMA].sub.2] were equal at that time. The values of [[GAMMA].sub.1] and [[GAMMA].sub.2] were then calculated:

cef(3 min) = ([[alpha].sub.1] min)[[GAMMA].sub.1] (3 min) + [[alpha].sub.2][[GAMMA].sub.2] (3 min) (12)

[[GAMMA].sub.1] (3 min) = cef(3 min)/[[alpha].sub.1]+[[alpha].sub.2] = [[GAMMA].sub.2](3 min) (13)

[[GAMMA].sub.1] (3 min) = [[GAMMA].sub.2](3 min) = cef(3 min) (14)

To calculate [[GAMMA].sub.max], the interface was assumed to be saturated with consolidated asphaltenes at 24 h.

cef(24 h) = [[alpha].sub.2][[GAMMA].sub.2](24 h) = [[alpha].sub.2] [[GAMMA].sub.max] (15)

[[GAMMA].sub.max] = cef(24 h)/[[alpha].sub.2] (16)

The assumption that the contributions from [[GAMMA].sub.1] and [[GAMMA].sub.2] to the magnitude of critical electric field at 3 min (the first measurement time) is arbitrary but comports with experimental observations and appears to provide a self consistent way of managing these unknown quantities. It also stands to reason that there should be much more "unconsolidated" asphaltenes, i.e., [[GAMMA].sub.1] at short time and because the weighting factor for it is much less than [[GAMMA].sub.2] that they might contribute approximately equally at short time.

The above set of equations could not be solved analytically. Runge-Kutta numerical methods were used to generate a series of parametric curves in which the parameters were systematically varied to find the best fit with the data. The best [k.sub.a]/[k.sub.c] combinations for several values of [a.sub.1] and [a.sub.2] were obtained by minimizing the sum of the squares of the differences between the experimentally measured cef and model prediction values for all times. For all model and crude oils, values of 0.2 and 0.8 for [[alpha].sub.1] and [[alpha].sub.2], respectively, provided a good fit to the data.

The kinetics of three different model systems were investigated: 0.5 and 1.0% AH asphaltenes in 40% toluene, and 1.0% AH asphaltenes in 50% toluene. The model was unable to predict a high enough rate of adsorption to account for the very rapid cef increase for 0.5 and 1.0% AH asphaltenes in 40% toluene. At these conditions, the driving force for asphaltene adsorption is very large due to the large chemical mismatch and high concentration of precipitated aggregates. The lyophobic forces associated with the asphaltene aggregates in these conditions may not be accounted for in the model. The fit of the model to the 1.0% AH asphaltene in 50% toluene results provided a [k.sub.a] of 0.8 [cm.sup.3][min.sup.-1][g.sup.-1] and a [k.sub.c] of 0.02 [min.sup.-1].

From model emulsion studies we have confirmed the importance of the cef technique and we have demonstrated the validity of the assumption that the magnitude of the cef depends directly on the interfacial film thickness (see Figures 7 and 8). We have also proposed a kinetic model for interfacial film formation that fits the data, and we have learned the importance of asphaltene concentration and solvency as gauged by [DELTA]H/C on emulsion stability. We will now use these findings to develop a correlation for petroleum emulsion stability and to discover the parameters that govern interfacial film development.

Petroleum Emulsions

Determination of Petroleum Emulsion Stability Correlation

Measurements of cef were performed for emulsions prepared with 30% water (v/v) for 12 different petroleums after aging for 24 h. All 12 petroleums produced emulsions that were at least stable to gravity over the 24 h aging period. The decreased water content led to greater distances between emulsified water droplets immediately after homogenization, resulting in lower droplet-droplet collision frequencies, allowing longer times for interfacial film development. Based on the model oil emulsion results, we know asphaltene solvency plays a large role in emulsion stability. Additional correlations in which resin content appeared with a negative exponent and the difference between the H/C ratio of the asphaltenes and petroleum solvent appeared with a positive exponent were attempted. By "petroleum solvent," what we mean here is all of the petroleum fluid excepting the "asphaltene solute," or more specifically, the so-called maltenes in the petroleum. The dependence of cef on the product of asphaltene concentration and [DELTA] H/C for petroleum emulsions, prepared with 30% water (v/v), after 24 h of aging, is displayed in Figure 10. A correlation coefficient of 0.88 was obtained when South China Sea crude was included. Much of the deviation from linearity is attributable to high value of cef for the emulsion obtained from this crude oil (South China Sea or SCS). This particular crude contains over 32% wax. The experiments were performed at 60[degrees]C, but a small percentage of the wax in SCS may be precipitated at this condition. Alternatively, the extent of asphaltene solvation, as measured by [DELTA] H/C, in SCS may not be comparable to that found in the other crudes. The effect of the long paraffinic wax molecules in SCS on asphaltene solubility differs significantly from the effect of a shorter, normal alkane which is not accounted for with the H/C ratio. For example H/ C's of n-decane and n-triacontane (C30) are very similar (2.20 vs. 2.07, respectively) despite the fact that they differ by twenty carbons and result in differences in asphaltene solubility (longer alkane chains solubilize asphaltenes better than shorter ones).

[FIGURE 10 OMITTED]

Therefore, SCS was removed from the correlation and the coefficient improved from 0.88 to 0.95. It should be noted that the data point corresponding to this crude oil, SCS, does not appear in Figure 10. The correlation does not pass through the origin, suggesting that, with crude oils, there are factors other than asphaltene content that contribute to emulsion stability and that these factors may predominate at very low asphaltene content. We estimate the extent of asphaltene interfacial coverage in these emulsions using:

[N.sub.AA] = [m.sub.A,ads][N.sub.A]/[n.sub.A,agg] (17)

[S.sub.AA] = [pi][R.sup.2.sub.AA] (18)

[S.sub.ads] = [S.sub.AA][N.sub.AA] (19)

[THETA] = [S.sub.ads]/[A.sub.W/O] (20)

where [N.sub.AA] and [S.sub.AA] are the number and cross-sectional area of asphaltene aggregates, respectively, [n.sub.A,agg] is the number of asphaltene molecules per aggregate, assuming columnar stacking of asphaltene molecules in groups of five (Yen, 1992), and [R.sub.AA] is estimated asphaltene aggregate radius (20 [Angstrom]). Here [S.sub.ads] is the total surface area covered by asphaltenes, [THETA] is the asphaltene fractional surface coverage, and [A.sub.w/o] is the total water-oil interfacial area calculated using Equation (2). The asphaltene concentration, [w.sub.asph,oil], varied among the crudes used in this study, from 0.79 to 14.8 wt%, corresponding to monolayer coverage of droplets with diameters of 9 to 0.5 [micro]m, respectively. Size distributions were troublesome to perform for petroleum emulsions and prone to significant errors because the emulsions were opaque. Droplet counts were not performed as for the model emulsions, but qualitative observations were made. All of the emulsions appeared to have average diameters a little greater but close to these values. The smallest droplets observed in our experiments were about a micron in diameter, implying that this was the predominant size immediately after emulsification before any coalescence occurred. At this point, coalescence was fast and droplet diameters increased rapidly, because the extent of interfacial coverage was low. As droplet diameters increased, the interfacial area decreased, yielding a higher fractional coverage, [THETA]. At long times the coalescence rate became very small as multi-layered interfacial films covered the entire interfacial area. Given the fact that asphaltene interfacial adsorption is not instantaneous, it is not surprising that the actual droplet size distribution is slightly larger than that predicted from monolayer coverage. All of the systems reported here are within the range of what we would term the "minimum drop size distribution" possible and are controlled by the extent of interfacial film formation rather than the extent of mixing.

The above correlation provides a remarkably good fit to the data without involving many petroleum characteristics considering the large variety of samples studied and possible correlating parameters. Resins play a large role in the solubilization of asphaltenes and would be expected to have a large impact on emulsion stability. All of the petroleums included in the correlation have R/A values within a relatively narrow range (0.924.25, except for Alba (AL) = 6.17) and the extent of resin solvation of asphaltene aggregates may be fairly uniform. For petroleum with very different R/A values, emulsion stability may be affected differently with variation of the asphaltene concentration or [DELTA] H/C.

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

Fit of Petroleum Emulsion Stability to Kinetic Model

The correlation is very encouraging and may serve the industry well in predicting long-time emulsion stability of water-in-crude oil systems at conditions in which wax issues are unimportant. In order to discover information on the kinetics of asphaltene adsorption, critical electric fields for all petroleums with 30% water were determined at various times from 3 min to 24 h after emulsification and the results are displayed in Figures 11 and 12. The data points represent the measured cef values while the lines are the predicted trends obtained from fitting the data to the kinetic model. The optimized fit parameters are listed in Table 2.

The emulsion stability data fit the model reasonably well. Some crudes showed a significant decrease in cef at short times followed by an eventual increase to the 24 h value. This was probably due to droplet settling rather than a weakening of the interfaces at short times. Immediately after emulsification, all droplets were small. As these small droplets coalesced, larger droplets formed and fell to the bottom of the emulsion. In crude oils with high viscosities or densities close to that of water, this settling occurred slowly, consequently, at short times larger droplets were used for cef testing when the sample was withdrawn from the middle of the emulsion. As a result, the water to oil ratio in the sample cell was higher and the measured cef was lower for these emulsions. This was also observed for some petroleums which were not as viscous and was probably due to sampling the emulsion from the wrong level. The rate constants from the model fits for B6 and CS were not used for further analysis because the experimental difficulties produced results that could not be fit well to the model. The data for AB show a lag time before a cef increase, similar to that observed for 0.5% AH asphaltenes in 40% toluene that may be due to its low asphaltene concentration (0.79%). The low bulk asphaltene concentration led to low concentration of adsorbed asphaltenes at short times and no increase in emulsion stability. After the first 30 min, the concentration of adsorbed asphaltenes was high enough to provide increased stability over the initial value, and cef increased at a rapid rate. At long times, the cef drops, probably due to sampling error. The 24 h cef was assumed, for model fitting purposes, to be 0.07 kV/cm greater than the 6 h value based on results for most of the other petroleums.

The dependence of [k.sub.a] on R/A of the crude oils is shown in Figure 13. The value of [k.sub.a] increases with R/A. Asphaltene aggregates are more solvated at higher values of R/A and consequently are smaller. Smaller aggregates are more mobile and adsorb to the water-oil interface faster because they are more influenced by interfacial forces. For the purpose of fitting the overall cef data, resin concentration is not as important, suggesting that the major mechanism by which resins control emulsion stability is the modification of the driving force for adsorption of near interfacial asphaltenes.

[FIGURE 13 OMITTED]

[FIGURE 14 OMITTED]

The qualities of the fits indicate this is an adequate model to describe the kinetics and is physically meaningful. A few important inferences can be drawn from these fits: (1) water droplets are stabilized very rapidly by adsorption of asphaltenes to yield an emulsion system which requires a significant electric field to induce coalescence; (2) a longer consolidation process occurs during which the asphaltenic film undergoes conformational changes which produce a stronger interfacial film. The length scale for asphaltene diffusion from the bulk to the oil-water interface in these well-mixed systems is small enough that the viscosity or density of the oil phase probably has a very small effect on the long time stability of the emulsion and the differences between petroleums is due to asphaltene interactions with the interface.

Sjoblom and co-workers (Mouraille et al., 1998) investigated the effect of aging on crude oil systems which were diluted with various amounts of solvents (0-50% added) including "condensate F"--a mixture of alkanes and aromatics, heptane, 50% heptane/50% toluene, and toluene. For all of the systems, they observed no effect on cef for emulsions with 20% water tested immediately after homogenization versus those aged one week. The stabilities ranged from 2.3 kV/cm for pure crude oil to about 0.5 kV/cm for crude oil diluted with 50% toluene. Both these results and the ones we have obtained confirm that the kinetics of asphaltene-film stabilization in crude oil systems is rapid.

Petroleum Blends: Emulsion Stability and Interfacial Film Formation Kinetics

The cef was measured for emulsions produced with AB/HO blends with 30% water. The results are shown in Figure 14 as cef vs. % A x [DELTA] H/C plot. The data do not fall on the whole crude stability correlation, but show increased stability over the whole range of compositions. The largest increase relative to the petroleum correlation is found for high AB content blends. We investigated the kinetics of interfacial film formation for each blend composition using the kinetic model in order to understand the discrepancies between the petroleum and blend data. The fitted kinetic parameters are listed in Table 3. The dependence of [k.sub.a] and [k.sub.c] on % A x [DELTA] H/C is shown in Figure 15. The values of [k.sub.a] and [k.sub.c] peak at 5% HO, the point at which emulsion stability deviates the most from the petroleum correlation. This result indicates that the positive deviations in emulsion stability were due to increased mobility and adsorption rate of the asphaltene aggregates at the water-oil interface as well as more rapid consolidation. When a small amount (1-5%) of HO was added to AB, the H/C value of the DeA crude did not substantially change from the value for AB, but the H/C value and concentration of the asphaltene fraction changed significantly due to the high asphaltene concentration in HO (14.8%) relative to AB (0.79%). HO asphaltenes were better dispersed in the blend solvent and the resulting asphaltene mobility was much higher than expected based on a simple ideal mixing assumption, producing deviations from the petroleum correlation.

The asphaltene adsorption rates in the model emulsions with AH asphaltenes in 40% toluene were relatively high for similar reasons as observed for those in blends with low concentrations of HO in AB. AH and AB asphaltenes are both highly aromatic asphaltenes (low H/C). In both the low HO blends and 40% toluene model oils, the H/C ratio of the solvent was higher than the original solvent the asphaltenes were dissolved in, and significant aggregation occurred providing a high driving force for interfacial adsorption.

[FIGURE 15 OMITTED]

CONCLUSIONS

Critical electric field measurement is a good technique to quantitatively gauge emulsion stability. We have applied the critical electric field technique to model emulsions and revealed the importance of asphaltene solvency to emulsion stability. We have shown conclusively with these tests that both the asphaltene concentration and solvent-asphaltene chemical mismatch are key parameters for determining emulsion stability. The correlation of interfacial mass with cef has revealed a critical extent of interfacial coverage, above which, emulsion stability is less affected by adsorbed asphaltenes. These studies have shown that cef truly is a direct probe of the interface. We have also developed a kinetic model for emulsion stabilization due to asphaltene adsorption. With this model, we have shown stabilization to be due to asphaltene adsorption followed by consolidation to form an interfacial film.

The ability to probe the strength of the interfacial film directly has resulted in clear and very understandable correlations for water-in-crude oil emulsion stability with characteristics of the system which control asphaltene interfacial adsorption. The correlations developed have a tremendous amount of physical and chemical research behind them. The results are physically appealing and remarkably simple. The kinetic model has revealed information on the stabilization of petroleum emulsions. Asphaltene aggregates that are more solvated by resins or the petroleum solvent adsorb more rapidly to the interface. The long time emulsion stability is dictated by overall crude oil properties, which are asphaltene-solvent chemical mismatch and asphaltene concentration. The similar behaviour of the crude and model oil systems with asphaltenes alone, verified that asphaltene adsorption to the water-oil interface is the primary mechanism responsible for crude oil emulsion stability. The differences when resins are added suggest some key property for emulsion stabilization is missing in model oils. This property is probably viscosity. In order to better investigate the kinetics of film formation a better model oil system needs to be found which better simulates petroleum behaviour.

It is very noteworthy that a simple correlation was found for petroleum emulsion stability considering the huge variety of petroleum properties and enormous number of potential parameters. This indicates that despite the complexity of petroleum, by conducting a careful, well thought out analysis, utilizing knowledge of petroleum chemistry and model oil emulsion results, the key parameters for petroleum emulsion stability can be identified. There are some issues that remain unaddressed. Is it possible to extend the cef technique to lower temperatures and for crudes in which wax solvency and precipitation plays a role? Can we extend the correlation to refinery emulsions (specifically desalters and API separators) in which inorganic solids (iron sulphide, iron oxide, calcium carbonate, etc.) can be present at levels of 1-5+% (w/w)? How should the correlation account for rapidly changing solvent conditions when crudes are blended? With crudes of higher R/A, such as SF, MI, and GM, blending with a very stable crude with a much lower R/A, such as HO, B-4, or B-6, will likely yield crude blends in which % resin will enter into the correlation. There are many possible directions to follow in this research effort. Based on the results of this investigation so far we are optimistic about future studies.

ACKNOWLEDGEMENTS

The authors would like to thank Dr. P. Matthew Spiecker for performing some of the solubility work and providing information on interfacial film masses that aided in the calculations of film thickness. We are also grateful to the Petroleum Environmental Research Forum for funding this work through grants 95-02 and 97-07 and to the National Science Foundation for funding this work through research grant CTS-9817127. JS would like to thank the Norwegian Research Council (NFR) and his oil consortium for financial support.

Manuscript received June 8, 2007; revised manuscript received September 28, 2007; accepted for publication August 8, 2007.

REFERENCES

Acevedo, S., G. Escobar, L. Gutierrez and H. Rivas, "Isolation and Characterization of Natural Surfactants from Extra Heavy Crude Oils, Asphaltenes, and Maltenes. Interpretation of Their Interfacial Tension-pH Behaviour in Terms of Ion Pair Formation," Fuel 71, 619-623 (1992).

Acevedo, S., B. Mendez, A. Rojas, I. Layrisse and H. Rivas, "Asphaltenes and Resins from the Orinoco Basin," Fuel 64, 1741-1747 (1985).

Al-Jarrah, M. M. H. and A. H. Al-Dujaili, "Characterization of Some Iraqi Asphalts II. New Findings on the Physical Nature of Asphaltenes," Fuel Science and Technology International 7, 69-88 (1989).

Ali, L. H., K. A. Al-Ghannam and J. M. Al-Rawi, "Chemical Structure of Asphaltenes in Heavy Crude Oils Investigated by NMR," Fuel 69, 519-521 (1990).

Ali, M. F. and M. Saleem, "Asphaltenes in Saudi Arabian Heavy Crude Oil Solubility and Molecular Weights in Hydrocarbon Solvents," Fuel Science and Technology 6, 541-556 (1988).

Aveyard, R., B. P. Binks, P. D. I. Fletcher and J. R. Lu, "The Resolution of Water-in-Crude Oil Emulsions by the Addition of Low Molar Mass Demulsifiers," Journal of Colloid and Interface Science 139, 128-138 (1990).

Bailes, P. J. and S. K. L. Larkai, "Liquid Phase Separation in Pulsed D.C. Fields," Transactions of the Institution of Chemical Engineers 60, 115-121 (1982).

Bestougeff, M. A. and R. J. Byramjee, "Chemical Constitution of Asphaltenes," T. F. Yen and G. V. Chiliingarian, Eds., Elsevier Science (1994), pp. 67-94.

Bhardwaj, A. and S. Hartland, "Dynamics of Emulsification and Demulsification of Water in Crude Oil Emulsions," Industrial and Engineering Chemistry Research 33, 1271-1279 (1994).

Bhardwaj, A. and S. Hartland, "Study of Demulsification of Water-in-Crude Oil Emulsion," Journal of Dispersion Science and Technology 14, 541-557 (1993).

Boyd, J., C. Parkinson and P. Sherman, "Factors Affecting Emulsion Stability, and the HLB Concept," Journal of Colloid and Interface Science 41, 359-370 (1972).

Breen, P. J., A. Yen and J. Tapp, "Demulsification of Asphaltene-Stabilized Emulsions-Correlations of Demulsifier Performance with Crude Oil Composition," Petroleum Science and Technology 21, 437-447 (2003).

Calemma, V., P. Iwanski, M. Nali, R. Scotti and L. Montanari, "Structural Characterization of Asphaltenes of Different Origins," Energy & Fuels 9, 225-230 (1995).

Carroll, B. J., "The Stability of Emulsions and Mechanisms of Emulsion Breakdown," Surface and Colloid Science 9, 1-65 (1976).

Chen, T. Y., R. A. Mohammed, A. I. Bailey, P. F. Luckham and S. E. Taylor, "Dewatering of Crude Oil Emulsions 4. Emulsion Resolution by the Application of an Electric Field," Colloids and Surfaces A: Physicochemical and Engineering Aspects 83, 273-284 (1994).

Cimino, R., S. Correra, A. D. Bianco and T. P. Lockhart, "Solubility and Phase Behavior of Asphaltenes in Hydrocarbon Media," E. Y. Sheu and O. C. Mullins, Eds., Plenum Press, New York (1995) pp. 97-130.

Cratin, P. D., "A Quantitative Characterization of pH-Dependent Systems," Industrial and Engineering Chemistry 61, 35-45 (1969).

Cyr, N., D. D. McIntyre, G. Toth and O. P. Strausz, "Hydrocarbon Structural Group Analysis of Athabasca Asphaltene and its G.P.C. Fractions by C-13 N.M.R.," Fuel 66, 1709-1714 (1987).

Dickie, J. P. and T. F. Yen, "Macrostructures of the Asphaltic Fractions by Various Instrumental Methods," Analytical Chemistry 39, 1847-1852 (1967).

Dodd, C. G., "The Rheological Properties of Films at Crude Petroleum-Water Interfaces," Journal of Physical Chemistry 64, 544-550 (1960).

Ese, M.-H., J. Sjoblom, H. Fordedal, O. Urdahl and H. P. Ronningsen, "Ageing of Interfacially Active Components and Its Effect on Emulsion Stability as Studied by Means of High Voltage Dielectric Spectroscopy Measurements," Colloids and Surfaces 123-124, 225-232 (1997).

Fordedal, H., Y. Schildberg, J. Sjoblom and J.-L. Volle, "Crude Oil Emulsions in High Electric Fields as Studied by Dielectric Spectroscopy. Influence of Interaction Between Commercial and Indigenous Surfactants," Colloids and Surfaces A: Physicochemical and Engineering Aspects 106, 33-47 (1996).

Fordedal, H., E. Nodland, J. Sjoblom and O. M. Kvalheim, "A Multivariate Analysis of W/O Emulsions in High External Electric Fields as Studied by Means of Dielectric Time Domain Spectroscopy," Journal of Colloid and Interface Science 173, 396-405 (1995).

Gelot, A., W. Friesen and H. A. Hamza, "Emulsification of Oil and Water in the Presence of Finely Divided Solids and Surface-Active Agents," Colloids and Surfaces 12, 271-303 (1984).

Graham, D. E., A. Stockwell and D. G. Thompson, "Chemical Demulsification of Produced Crude Oil Emulsions," P. H. Ogden, Ed., Royal Society of Chemistry: Sunbury-on-Thames, Middlesex, U. K. (1983) pp. 73-91.

Hirato, T., K. Koyama, T. Tanaka, Y. Awakura and H. Majima, "Demulsification of Water-in-Oil Emulsion by an Electrostatic Coalescence Method," Materials Transactions 52, 257-263 (1991).

Isaacs, E. E., H. Huang, A. J. Babchin and R. S. Chow, "Electroacoustic Method for Monitoring the Coalescence of Water-in-Oil Emulsions," Colloids and Surfaces 46, 177-192 (1990).

Jeribi, M., B. Almir-Assad, D. Langevin, I. Henaut and J. F. Argillier, "Adsorption Kinetics of Asphaltenes at Liquid Interfaces." J. Colloid and Interface Science 256, 268-272 (2002).

Kawanaka, S., K. J. Leontaritis, S. J. Park and G. A. Mansoori, "Thermodynamic and Colloidal Models of Asphaltene Flocculation," J. K. Borchardt and T. F. Yen, Eds., American Chemical Society: Washington, D. C. (1989) pp. 443-458.

Kim, Y.-H., A. D. Nikolov, D. T. Wasan, H. Diaz-Arauzo and C. S. Shetty, "Demulsification of Water-in-Crude Oil Emulsions: Effects of Film Tension, Elasticity, Diffusivity, and Interfacial Activity of Demulsifier Individual Components and their Blends," Journal of Dispersion Science and Technology 17, 33-53 (1996).

Krawczyk, M. A., D. T. Wasan and C. S. Shetty, "Chemical Demulsification of Petroleum Emulsions Using Oil-Soluble Demulsifiers," Industrial and Engineering Chemistry Research 30, 367-375 (1991).

Lawrence, A. S. C. and W. Killner, "Emulsions of Seawater in Admiralty Fuel Oil with Special Reference to their Demulsification," Journal of the Institute of Petroleum 34, 281 (1948).

Mackay, G. D. M., A. Y. McLean, O. J. Betancourt and B. D. Johnson, "The Formation of Water-in-Oil Emulsions Subsequent to an Oil Spill," Journal of the Institute of Petroleum 59, 164-172 (1973).

Malhotra, A. K. and D. T. Wasan, "Stability of Foam and Emulsion Films: Effects of the Drainage and Film Size on Critical Thickness of Rupture," Chemical Engineering Communications 48, 35-56 (1986).

McKay, J. F., P. J. Amend, T. E. Cogswell, P. M. Harnsberger, R. B. Erickson and D. R. Latham, "Petroleum Asphaltenes: Chemistry and Composition," P. C. Uden, S. Siggia and H. B. Jensen, Eds., American Chemical Society: Washington D.C. (1978), Vol. 170, pp 128-142.

McLean, J. D. and P. K. Kilpatrick, "Effects of Asphaltene Aggregation in Model Heptane-Toluene Mixtures on Stability of Water-in-Oil Emulsions," Journal of Colloid and Interface Science 196, 23-34 (1997a).

McLean, J. D. and P. K. Kilpatrick, "Effects of Asphaltene Solvency on Stability of Water-in-Crude Oil Emulsions," Journal of Colloid and Interface Science 189, 242-253 (1997b).

Menon, V. B. and D. T. Wasan, "A Review of the Factors Affecting the Stability of Solids-Stabilized Emulsions," Separation Science and Technology 23, 2131-2142 (1988).

Menon, V. B. and D. T. Wasan, "Coalescence of Water-in-Shale Oil Emulsions," Separation Science and Technology 19, 555-574 (1984).

Mikula, R. J., "Emulsion Characterization," L. L. Schramm, Ed., American Chemical Society: Washington D.C. (1992), pp. 79-130.

Mohammed, R. A., A. I. Bailey, P. F. Luckham and S. E. Taylor, "The Effect of Demulsifiers on the Interfacial Rheology and Emulsion Stability of Water-in-Crude Oil Emulsions," Colloids and Surfaces A: Physicochemical and Engineering Aspects 91, 129-139 (1994).

Moschopedis, S. E., J. F. Fryer and J. G. Speight, "Investigation of Asphaltene Molecular Weights," Fuel 55, 227-232 (1976).

Mouraille, O., T. Skodvin, J. Sjoblom and J.-L. Peytavy, "Stability of Water-in-Crude Oil Emulsions: Role Played by the State of Solvation of Asphaltenes and by Waxes," J. Dispersion Science and Technology 19, 339-367 (1998).

Mukherjee, S. and A. P. Kushnick, "Effect of Demulsifiers on Interfacial Properties Governing Crude Oil Demulsification," American Chemical Society: Toronto (1988), Vol. 33, pp 205-210.

Obah, B., "The Chemical Demulsification of Crude Oil Emulsion: Problem in a Niger Delta Oil Terminal," Erdohl, Kohle, Erdgas Petrochemie 41, 71-74 (1988).

Papirer, E., C. Bourgeois, B. Siffert and H. Balard, "Chemical Nature and Water/Oil Emulsifying Properties of Asphaltenes," Fuel 61, 732-734 (1982).

Parkinson, C. and P. Sherman, "Phase Inversion Temperature as an Accelerated Method for Evaluating Emulsion Stability," Journal of Colloid and Interface Science 41, 328-330 (1972).

Pelet, R., F. Behar and J. Monin, "Resins and Asphaltenes in the Generation and Migration of Petroleum," Organic Geochemistry 10, 481-498 (1985).

Schramm, L. L., "Petroleum Emulsions: Basic Principles," American Chemical Society, Washington, D.C. (1992), Vol. 231, pp. 1-49.

Schubert, H. and H. Armbruster, "Principles of Formation and Stability of Emulsions," International Chemical Engineering 32, 14-28 (1992).

Shetty, C. S., A. D. Nikolov and D. T. Wasan, "Demulsification of Water in Oil Emulsions Using Water Soluble Demulsifiers," Journal of Dispersion Science and Technology 13, 121-133 (1992).

Sheu, E. Y., D. A. Storm and M. M. D. Tar, "Asphaltenes in Polar Solvents," Journal of Non-Crystalline Solids 131-133, 341-347 (1991).

Siffert, B., C. Bourgeois and E. Papirer, "Structure and Water-Oil Emulsifying Properties of Asphaltenes," Fuel 63, 834-837 (1984).

Singh, B. P., "Performance of Demulsifiers: Prediction Based on Film Pressure-Area Isotherms and Solvent Properties," Energy Sources 16, 377-385 (1994).

Sjoblom, J., H. Fordedal, T. Jakobsen and T. Skodvin, "Dielectric Spectroscopic Characterization of Emulsions," K. S. Birdi, Ed., CRC Press, Boca Raton (1997), pp. 217-237.

Sjoblom, J., T. Skodvin, O. Holt and F. P. Nilsen, "Colloid Chemistry and Modern Instrumentation in Offshore Petroleum Production and Transport," Colloids and Surfaces A: Physicochemical and Engineering Aspects 123-124, 593-607 (1997b).

Sjoblom, J., H. Soderlund, S. Lindblad, E. J. Johansen and I. M. Skjarvo, "Water-in-Crude Oil Emulsions from the Norwegian Continental Shelf: Part II. Chemical Destabilization and Interfacial Tensions," Colloid Polym. Sci. 268, 389-398 (1990).

Skodvin, T. and J. Sjoblom, "Dielectric Spectroscopy on W/O Emulsions Under Influence of Shear Forces," Colloid and Polymer Science 274, 754-762 (1996).

Skodvin, T., J. Sjoblom, J. O. Saeten, O. Urdahl and B. Gestblom, "Water-in-Crude Oil Emulsions from the Norwegian Continental Shelf IX. A Dielectric Spectroscopic Characterization of Authentic as Well as Model Systems," Journal of Colloid and Interface Science 166, 43-50 (1994).

Speight, J. G., "Latest Thoughts on the Molecular Nature of Petroleum Asphaltenes," Preprints--ACS Division of Petroleum Chemistry 34, 321-328 (1989).

Speight, J. G., "Polynuclear Aromatic Systems in Petroleum," Preprints--ACS Division of Petroleum Chemistry 31, 818-825 (1986).

Speight, J. G. and S. E. Moschopedis, "Some Observations on the Molecular "Nature" of Petroleum Asphaltenes," Preprints--ACS Division of Petroleum Chemistry 24, 910-923 (1979).

Spiecker, P. M., "The Impact of Asphaltene Chemistry and Solvation on Emulsion and Interfacial Film Formation," Ph.D. Thesis, North Carolina State University, (2001).

Stalss, F., R. Bohm and R. Kupfer, "Improved Demulsifier Chemistry: A Novel Approach in the Dehydration of Crude Oil," Society of Petroleum Engineers--Production Engineering 334-338 (1991).

Storm, D. A., S. J. DeCanio, M. M. DeTar and V. P. Nero, "Upper Bound on Number Average Molecular Weight of Asphaltenes," Fuel 69, 735-738 (1990).

Strassner, J. E., "Effect of pH on Interfacial Films and Stability of Crude Oil-Water Emulsions," Journal of Petroleum Technology 20, 303-312 (1968).

Taylor, S. E., "Resolving Crude Oil Emulsions," Chemistry and Industry 20, 770-773 (1992).

Tissot, B. P. and D. H. Welte, "Petroleum Formation and Occurrence: A New Approach to Oil and Gas Exploration," Springer-Verlag, Berlin, 1984.

Urdahl, O., T. J. Williams, A. G. Bailey and M. T. Thew, "Electrostatic Destabilization of Water-in-Oil Emulsions under Conditions of Turbulent Flow," Transactions of the Institute of Chemical Engineers 74, 158-165 (1996).

Wang, L. and Z. Yan, "The Optimum Frequency in Electrically Enhanced Coalescence of W/O Type Emulsions," Journal of Dispersion Science and Technology 15, 35-58 (1994).

Wasan, D. T., "Destabilization of Water-in-Oil Emulsions," J. Sjoblom, Ed., Kluwer Academic Publishers, Netherlands (1992), pp. 283-295.

Watson, B. A. and M. A. Barteau, "Imaging of Petroleum Asphaltenes Using Scanning Tunneling Microscopy," Indusrial & Engineering Chemistry Research 33, 2358-2363 (1994).

Yan, N. and J. H. Masliyah, "Characteriztaion and Demulsification of Solids-Stabilized Oil-in-Water Emulsions Part 1. Partitioning of Clay Particles and Preparation of Emuksions," Colloids and Surfaces A: Physicochemical and Engineering Aspects 96, 229-242 (1995a).

Yan, N. and J. H. Masliyah, "Characterization and Demulsification of Solids-Stabilized Oil-in-Water Emuslions Part 2. Demulsification by the Addition of Fresh Oil," Colloids and Surfaces A: Physicochemical and Engineering Aspects 96, 243-252 (1995b).

Yen, T. F., "The Colloidal Aspect of a Macrostructure of Petroleum Asphalt," Fuel Science and Technology International 10, 723-733 (1992).

Zaki, N. N., M. E. Abdel-Raouf and A.-A. Abdel-Azim, "Propylene Oxide-Ethylene Oxide Block Copolymers as Demulsifiers for Water-in-Oil Emulsions, I. Effect of Molecular Weight and Hydrophilic-Lipophylic Balance on the Demulsification Efficiency," Monatshefte fur Chemie 127, 621-629 (1996a).

Zaki, N. N., M. E. Abdel-Raouf and A.-A. Abdel-Azim, "Polyoxyethylenated Bisphenol-A for Breaking Water-in-Oil Emulsions," Polymers for Advanced Technologies 7, 805-808 (1996b).

Andrew P. Sullivan (1), Nael N. Zaki (1), Johan Sjoblom (2) and Peter K. Kilpatrick (1) *

(1.) Department of Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, NC, U.S.A. 27695-7905

(2.) Ugelstad Laboratory, Department of Chemical Engineering, Norwegian University of Science and Technology, N-7491 Trondheim, Norway

* Author to whom correspondence may be addressed. E-mail address: peter_kilpatrick@ncsu.edu

Table 1. Summary of crude properties

Crude %A (a) %R (b) R/A

Arab Hvy II (AH) 6.68 7.46 1.12
Arab Berri (AS) 0.79 3.24 4.10
Alaska North Slope (ANS) 3.38 8.72 2.58
San Joaquin valley (SJV) 4.56 19.4 4.25
B-4 13.6 12.2 0.90
B-6 13.1 12 0.92
Alba (AL) 1.64 10.1 6.17
Malu Isan (MI) 0.18 4.86 27.0
Sour Maya (SM) 11.5 11 0.95
Canadon Seco (CS) 7.5 8.94 1.19
Statfjord (SF) 0.09 4.02 42.6
Gulf of Mexico (GM) 0.31 5.02 16.2
South China Sea (SCS) 3.46 6.05 1.75
Thums 1 (TH1) 5.09 18.7 3.67
Thums 2 (TH2) 3.31 12.5 3.77
Hondo (H0) 14.8 20.5 1.39

 H/C H/C H/C
Crude Whole (c) A (c) R (c)

Arab Hvy II (AH) 1.683 1.080 1.372
Arab Berri (AS) 1.814 1.020 1.349
Alaska North Slope (ANS) 1.710 1.057 1.408
San Joaquin valley (SJV) 1.518 1.170 1.38
B-4 1.593 (i) 1.222 1.514
B-6 1.553 (i) 1.224 1.536
Alba (AL) 1.651 1.144 1.433
Malu Isan (MI) 1.913 1.333 1.499
Sour Maya (SM) 1.615 1.087 1.410
Canadon Seco (CS) 1.680 1.028 1.376
Statfjord (SF) 1.844 1.289 1.412
Gulf of Mexico (GM) 1.782 1.117 1.374
South China Sea(SCS) 1.867 1.353 1.386
Thums 1 (TH1) 1.696 1.153 1.455
Thums 2 (TH2) 1.690 (i) 1.178 1.442
Hondo (H0) 1.667 1.248 1.508

 [DELTA] Density
 H/C H/C (g/mL)
Crude DeA (d) (DeA-A) @60F (e)

Arab Hvy II (AH) 1.725 0.645 0.946 (h)
Arab Berri (AS) 1.820 0.800 0.838
Alaska North Slope (ANS) 1.732 0.675 0.889
San Joaquin valley (SJV) 1.535 0.365 0.979
B-4 1.667 0.445 0.935
B-6 1.618 0.394 0.935
Alba (AL) 1.659 0.515 0.940
Malu Isan (MI) 1.914 0.581 0.845
Sour Maya (SM) 1.682 0.595 0.919
Canadon Seco (CS) 1.734 0.706 0.903
Statfjord (SF) 1.845 0.556 0.833
Gulf of Mexico (GM) 1.784 0.667 0.870
South China Sea (SCS) 1.886 0.533 0.858
Thums 1 (TH1) 1.726 0.573 0.952
Thums 2 (TH2) 1.718 0.50 1.01 (h)
Hondo (H0) 1.738 0.490 0.938

 Viscosity Kin Visc
 @100F @100F
Crude (CP) (f) (cst) (g)

Arab Hvy II (AH) 33.8 35.7
Arab Berri (AS) 4.39 5.24
Alaska North Slope (ANS) 12.8 14.4
San Joaquin valley (SJV) 1390 1420
B-4 2310 2470
B-6 2030 2170
Alba (AL) 136 145
Malu Isan (MI) 38.2 45.2
Sour Maya (SM) 75 81.6
Canadon Seco (CS) 70 77.5
Statfjord (SF) 3.14 3.77
Gulf of Mexico (GM) 7.11 8.17
South China Sea (SCS) -- --
Thums 1 (TH1) 152 160
Thums 2 (TH2) 656 650
Hondo (H0) 363 387

(a) = n-heptane precipitation

(b) = sequential elution chromatography

(c) = combustion and GC analysis (Galbraith Laboratories)

(d) = calculated from combustion and GC analyses of whole crude, A,
and R {H,C = whole(H,C)%[A.sup.*][A(H,C)])}

(e) = from assays (supplier of crude)

(f) = rheology (stress rheometer, couette geometry)

(g) = dynamic viscosity/density

(h) = pycnometry

(i) = corrected numbers based on assumption 100-%C-%H=%O from water
in crude sample (for high water content crudes)

Table 2. Model parameters for crude oil emulsions
([[alpha].sub.1] = 0.2, [[alpha].sub.2] = 0.8)

Crude oil [k.sub.a] ([cm.sup.3] [k.sub.c] (x[10.sup.3]
 [min.sup.-1][g.sup.-1]) [min.sup.-1])

AB 2.5 6
AH 0.1 3
AL 1.2 3
ANS 3.9 2
B4 1.2 4
B6 0.3 5
CS 4.8 9
HO 0.1 16
SCS 10 2
SJV 2.6 2
SM 0.8 4
TH1 0.3 4

Table 3. Model parameters for crude oil blend emulsions
([[alpha].sub.1] = 0.2, [[alpha].sub.2] = 0.8)

%HO in AB/HO [k.sub.a] ([cm.sup.3] [k.sub.c] (x[10.sup.3]
 [min.sup.-1][g.sup.-1]) [min.sup.-1])

0.0 2.5 6
1.0 4.8 7
5.0 14 67
11.4 0.2 20
18.0 0.6 23
25.0 1.0 4
50.0 1.2 6
75.0 1.9 4
100 0.1 16