Analysis of Formation of Water-in-Oil Emulsions - Energy & Fuels

3674

Energy & Fuels 2009, 23, 3674–3680

Analysis of Formation of Water-in-Oil Emulsions Clint P. Aichele,*,† Walter G. Chapman,† Lee D. Rhyne,‡ Hariprasad J. Subramani,‡ and Waylon V. House§ Department of Chemical and Biomolecular Engineering, Rice UniVersity, 6100 S. Main, MS 362, Houston, Texas 77005, CheVron Energy Technology Company, 1400 Smith Street, Houston, Texas 77002, and Department of Petroleum Engineering, Texas Tech UniVersity, 8th and Canton, Box 43111, Lubbock, Texas 79409-3111 ReceiVed March 5, 2009. ReVised Manuscript ReceiVed May 4, 2009

This work analyzes water-in-oil emulsion formation in a Taylor-Couette shear cell. ASTM certified brine is the dispersed phase, and an assortment of crude and model oils are used as continuous phases. This work employs the pulsed field gradient with diffusion editing (PFG-DE) nuclear magnetic resonance (NMR) technique to measure the drop size distributions. The PFG-DE technique is particularly useful because it is not constrained by the optical properties of the emulsions, investigates the entire emulsion, and does not assume a form of the drop size distribution. The effect of the imposed shear field on the drop size distributions is studied (a) experimentally by subjecting the emulsions to a wide range of shear fields in the Taylor-Couette cell and (b) computationally via numerical simulations in Fluent. Such quantitative information on the mechanisms of formation of water-in-crude-oil emulsions is invaluable because conventional optical measurement techniques encounter difficulties when attempting to characterize water-in-crude-oil emulsions.

Introduction The formation of emulsions occurs frequently in nature and in many industrial sectors, including the food,1 pharmaceutical,2 and energy industries.3-5 Because of the intense interest in emulsions, a large body of literature exists regarding single drop breakup and dilute emulsion formation in the absence of surfactants.6,7 However, a paucity of data exists for emulsion formation of concentrated emulsions, particularly for crude oil emulsions which contain an assortment of surface active material.8,9 These systems are particularly difficult to investigate with traditional methods such as microscopy and light scattering. However, nuclear magnetic resonance (NMR) is not constrained by the optical properties of opaque emulsions.3 Therefore, this work employs the pulsed field gradient with diffusion editing (PFG-DE) NMR technique to provide quantitative drop size information for complex, emulsified systems. G.I. Taylor pioneered the initial work on single drop breakup of a Newtonian drop in a Newtonian continuous phase.10,11 Taylor developed a deformation theory which predicted the deformation of a drop in simple shear flow. * To whom correspondence should be addressed. Telephone: +1 713 348 3581. Fax: 1 713 348 5478. E-mail: [email protected]. † Rice University. ‡ Chevron Energy Technology Company. § Texas Tech University. (1) Chappat, M. Colloids Surf. A: Physicochem. Eng. Aspects 1994, 91, 57–77. (2) Cao, A.; Hantz, E.; Taillandier, E. Colloids Surf. 1985, 14, 217– 229. (3) Pen˜a, A.; Hirasaki, G. J. In Emulsions and Emulsion Stability; Sjo¨blom, J., Ed.; CRC: Boca Raton, FL, 2006; pp 283-309. (4) Emulsions: Fundamentals and Applications in the Petroleum Industry; Schramm, L. L., Ed.; American Chemical Society: Washington, DC, 1992. (5) Sjo¨blom, J.; Aske, N.; Auflem, I. H.; Brandal, O.; Havre, T. E.; Saether, O.; Westvik, A.; Johnsen, E. E.; Kallevik, H. AdV. Colloid Interface Sci. 2003, 100, 399–473. (6) Rallison, J. M. Annu. ReV. Fluid Mech. 1984, 16, 45–66. (7) Stone, H. A. Annu. ReV. Fluid Mech. 1994, 26, 65–102. (8) Dalmazzone, C. Lubrication Sci. 2005, 17, 197–237. (9) Zhao, X. J. Rheol. 2007, 51, 367–392.

D)

a-b 19λ + 16 ) Ca a+b 16λ + 16

(1)

The length of the drop along the x-axis is given by a, while the height of the drop along the y-axis is given by b. The drop deformation, D, depends linearly on the Capillary number, Ca, as well as the viscosity ratio between the dispersed and continuous phases, λ. Ca )

γµCPr σ

(2)

In eq 2, γ is the shear rate applied to the drop and r is the radius of the undeformed drop. The capillary number is the ratio between the disruptive viscous shear stresses from the continuous fluid, γµCP, and the restoring interfacial forces of the drop, σ/r. Taylor’s deformation theory agrees with experimental data.12-15 Subsequently, Grace performed experiments to determine the critical Capillary number at drop breakup as a function of viscosity ratio in simple shear flow.12 Grace’s famous work established the map of drop breakup for single drops and dilute systems in simple shear in the absence of surfactant. Recently, Zhao directly visualized the transient breakup of dilute emulsions (0.1-0.2 wt %) in simple shear flow in the absence of surfactant which provided specific information about breakup mechanisms in different regions of viscosity ratio.9 Subsequent researchers attempted to provide theoretical validation to Grace’s experimental results. Hinch and Acrivos developed theoretical predictions for low viscosity ratios (10) Taylor, G. I. Proc. R. Soc. London 1932, 138, 41–48. (11) Taylor, G. I. Proc. R. Soc. London 1934, 146, 501–523. (12) Grace, H. P. Chem. Eng. Commun. 1982, 14, 225–277. (13) Guido, S.; Villone, M. J. Rheol. 1998, 42, 395–415. (14) Rumscheidt, F. D.; Mason, S. G. J. Colloid Sci. 1961, 16, 238– 261. (15) Torza, S.; Cox, R. G.; Mason, S. G. J. Colloid Interface Sci. 1972, 38, 395–411.

10.1021/ef900192v CCC: $40.75  2009 American Chemical Society Published on Web 05/21/2009

Analysis of Formation of Water-in-Oil Emulsions

Energy & Fuels, Vol. 23, 2009 3675

Figure 1. PFG-DE pulse sequence (adapted from Flaum, 2006).18

(λf0).16 This theory, referred to as the slender body theory, states that as the dispersed phase viscosity approaches zero the drop assumes a slender shape before breakup. For higher values of the viscosity ratio, Barthes-Biesel and Acrivos developed theoretical predictions for drop breakup.17 Of great industrial importance is the formation of concentrated (>10 vol %) water-in-oil emulsions in the presence of surfactants, particularly crude oil emulsions.8 This work presents experimental data regarding emulsion formation of concentrated emulsions in the presence of surfactants in welldefined flow fields. This work employs NMR to measure drop size distributions using the PFG-DE technique. The experimental data are coupled with computational fluid dynamics simulations to obtain quantitative insight about concentrated emulsion formation. Experimental Methods The PFG-DE technique was used to measure the drop size distributions. Figure 1 shows the NMR pulse sequence. Equation 3 is the governing magnetization equation for this technique.18

M(g, t) )

∫ ∫ f(r, T ) exp(- Tt )R

sp(∆, δ, g, r)

2

(

2

exp -

×

(

))

1 1 ∆+δ - 2δ drdT2 (3) T1 T2 T1

T1, the longitudinal relaxation time, was assumed to be equal to T2 in this work, which is a valid assumption for liquids at 2 MHz.19 The attenuation of water droplets as a result of restricted diffusion is given by Rsp, and it is described by the relations developed by Murday and Cotts.20

{

∞

Rsp ) exp -2γ2gg2

∑R

m)1

1 2 2 2 m(Rmr

- 2)

[

2δ R2mDDP

-

Ψ (R2mDDP)2

]}

(4)

Ψ ) 2 + exp(-R2mDDP(∆ - δ)) - 2 exp(-R2mDDPδ) 2 exp(-R2mDDP∆) + exp(-Rm2 DDP(∆ + δ))

(5)

The time between gradient pulses is ∆; δ is the gradient pulse duration; the gyromagnetic ratio is given by γg; the gradient strength (16) Hinch, E. J.; Acrivos, A. J. Fluid Mech. 1980, 98, 305–328. (17) Barthes-Biesel, D.; Acrivos, A. J. Fluid Mech. 1973, 61, 1–21. (18) Flaum, M. Fluid and rock characterization using new NMR diffusion editing pulse sequences and two dimensional diffusivity-T2 maps, Ph.D. Thesis, Rice University, Houston, TX, 2006. (19) Callaghan, P. T. Principles of Nuclear Magnetic Resonance Microscopy; Oxford University Press: New York, 1991. (20) Murday, J. S.; Cotts, R. M. J. Chem. Phys. 1968, 48, 4938–4945.

Figure 2. Example of the two-dimensional result obtained from the PFG-DE technique.

is g; the dispersed phase diffusivity is DDP; r is the radius of the emulsion droplet, and Rm is the mth positive root of eq 6.

1 J (Rr) ) J5/2(Rr) Rr 3/2

(6)

Jk is the Bessel function of the first kind with order k. The distribution of both drop size and transverse relaxation, f(r,T2), is determined using a two-dimensional inversion with regularization.18,21,22 Figure 2 shows an example of the twodimensional information provided by this measurement. Recently, this technique was shown to resolve drop size distributions of polydisperse emulsions because the form of the drop size distribution is not assumed, thereby making it a robust technique to investigate complex emulsified systems.23 This work used two crude oils and two model oils as the continuous phases. The dispersed phase of all the emulsions was ASTM certified brine. Figure 3 shows that each crude oil was matched with a model oil of similar viscosity. Table 1 contains the values of the brine viscosity divided by the oil viscosity at 303.2 K for the four systems. The viscosities of each fluid were measured using a Brookfield Programmable DV-III viscometer. The viscosities of the oils were independent of shear rate over the range of shear rates used in the viscosity measurements. For the crude/model oil A viscosity measurements, the shear rate range was 20-100 s-1. For the crude/model oil B viscosity measurements, the shear rate range was 1-20 s-1. The viscosities of the model oils were (21) Hu¨rlimann, M. D.; Venkataramanan, L. J. Magn. Reson. 2002, 157, 31–42. (22) Venkataramanan, L.; Song, Y. Q.; Hu¨rlimann, M. D. IEEE Trans. Signal Process. 2002, 50, 1017–1026. (23) Aichele, C. P.; Flaum, M.; Jiang, T.; Hirasaki, G. J.; Chapman, W. G. J. Colloid Interface Sci. 2007, 315, 607–619.

3676

Energy & Fuels, Vol. 23, 2009

Aichele et al.

Figure 3. Viscosity as a function of temperature for the crude and model oils.

Figure 4. SARA analysis for the four oils used in this work (provided by Dr. Jill Buckley’s laboratory at New Mexico Tech, 2008).24

Table 1. Viscosity Ratios for the Four Systems system

viscosity ratio

brine/crude oil A brine/model oil A brine/crude oil B brine/model oil B

0.06 0.05 0.005 0.004

Table 2. Effect of Surfactant on Model Oil Viscosity (298.2 K) fluid model model model model

oil oil oil oil

µ (cP)

A with no surfactant A with 4 vol % Span 80 B with no surfactant B with 4 vol % Span 80

31.0 34.4 405.0 412.4

Table 3. Densities of the Fluids fluid

density (g/mL)

ASTM brine Span80 crude oil A model oil A crude oil B model oil B

1.03 1.00 0.85 0.81 0.91 0.80

measured without surfactant. To quantify the effect of surfactant on the model oil viscosities, the viscosities of the model oils with 4 vol % Span80 were measured at 298.2 K, and the results are provided in Table 2. Table 3 provides the densities of the fluids, and they were measured by weighing a known volume of the oil on a Sartorius balance. The SARA analysis provides the amounts of saturates, aromatics, resins, and asphaltenes in each oil, as shown in Figure 4.24 The SARA analysis indicates that the model oils are essentially free of resins and asphaltenes. Naturally occurring surfactants stabilized the brine-in-crude-oil emulsions. The nonionic surfactant, Span80 (Sigma-Aldrich), stabilized the brine-in-model-oil emulsions. The specific surface area of Span80 is 3.7 × 10-7 µm2/molecule.25 Assuming a drop diameter equal to 10 µm, the total interfacial area of the dispersed phase is 7.2 × 1012 µm2, thereby requiring 0.0138 g Span80 for monolayer coverage. In all experiments, the concentration of the oil soluble surfactant was 10 times the needed concentration for monolayer coverage. With the brine fraction equal to 0.2, the concentration of Span80 was 0.356 wt % in terms of the mass of oil. The surfactant was dissolved in the model oils before the addition of brine. (24) Buckley, J.; Creek, J.; Wang, J.; Fan, T., personal communication, SARA analysis, 2008. (25) Peltonen, L.; Hirvonen, J.; Yliruusi, J. J. Colloid Interface Sci. 2001, 240, 272–276.

Figure 5. Interfacial tension between the two crude oils and brine.

Figure 6. Interfacial tension between the two model oils with surfactant and brine.

Interfacial properties provide critical information regarding emulsion formation.8 A KSV CAM 200 instrument performed the interfacial property measurements using the pendant drop technique. The drop profile and densities of the oil and brine enable the determination of the interfacial tension using the Young-LaPlace equation. Figure 5 shows the equilibrium interfacial tension of the two crude oils suspended in brine. The interfacial tension of crude oil A (22 mN/m) was greater than crude oil B (12 mN/m) indicating a difference in surface active material. Figure 6 shows the equilibrium interfacial tension of the model oils in the presence of surfactant. In the presence of surfactant, the suspended oil drop



(

µeff ) µCP 1 -

Figure 7. Taylor-Couette flow leads to the development of secondary flow patterns referred to as Taylor vortices.

quickly rose off the tip of the needle, indicating the ability of Span80 to rapidly decrease the interfacial tension. These data provide an order of magnitude estimate of 5 mN/m for the interfacial tension of both model oil systems. A Taylor-Couette flow device formed the emulsions in this work. The rotating, inner cylinder was composed of Torlon with radius equal to 19.1 mm. The stationary, outer cylinder was composed of glass with radius equal to 21.6 mm. Both simple and secondary flows can exist in this geometry as shown in Figure 7. The dimensionless Reynolds and Taylor numbers describe the flow in this geometry.26

Re )

Ta )

Friω(ro - ri) µ

F2riω2(ro - ri)3 µ2

(7)

(8)

G.I. Taylor showed that instabilities in the flow field can arise, and these instabilities are primarily governed by the critical Taylor number, Tac.27 Tac )

(

π4 1 +

(

0.0571 1 - 0.652

)

(ro - ri) 2ri

) (

(ro - ri) (ro - ri) + 0.00056 1 - 0.652 ri ri

)

-1

(9)

For the geometry used in this work, Tac ) 1960. Before the fluids were subjected to shear in the Taylor-Couette device, the inner cylinder was moved up and down 20 times to promote interaction between the brine and oil phases. This action dispersed large brine drops in the oil. For all experiments, the fluids were sheared with the Taylor-Couette device for 10 min. After shearing, the emulsions were placed in the NMR instrument, and the drop size distributions were measured under quiescent conditions. A typical drop size distribution measurement required 6 h. For all emulsions, the volume of brine was 12 mL and the volume of oil was 48 mL, thereby yielding a dispersed phase fraction equal to 0.2. The sample height for all samples was 4 cm.

φ φmax

)

-2.5φmax

(10)

The maximum packing fraction for a face-centered cubic lattice is represented by φmax.29 A multiphase simulation was also performed to provide comparisons to the single-phase simulations. The eulerianeulerian multiphase model was used to perform the simulations using a pressure-based solver. The Schiller-Neumann drag coefficient was incorporated, and the no slip boundary condition was used at the walls. Figure 8 shows that the single-phase and multiphase simulations produced similar results for the axial velocity profile at the same inner cylinder rotational speed (3000 rpm). In addition, the multiphase simulation showed that the water volume fraction throughout the fluid domain was within 1% of the initial concentration, thereby signifying that no significant depletion or enrichment of water drops occurred throughout the fluid domain. Therefore, single-phase simulations are appropriate to simulate the flow fields in this work. The convergence criterion for the transient simulations was 0.001. The two-dimensional mesh consisted of 72 721 nodes, and the solver implemented the Green-Gauss node based algorithm. Grid refinement comparisons were performed to ensure that the results were grid independent. For all cases, the flow was modeled as laminar because k-ε turbulent calculations predicted no turbulence for the range of inner cylinder rotational speeds used in this work. Simulations performed in three dimensions showed that there were no deviations between twoand three-dimensional calculations. Results and Discussion Crude/Model Oil A Systems. Figure 9 shows the results of the simulations for the swirl and axial velocities for the crude/ model oil A systems at the minimum inner cylinder rotational speed used in the experiments (1600 rpm). The effective viscosity for the crude/model oil A systems at 303.2 K was 36 cP which was 1.8 times the pure fluid viscosity. The swirl velocity is the component of the velocity coming out of the plane, while the axial velocity is the component of the velocity that exists along the vertical axis. Secondary flows in the form

Computational Methods This work implemented the computational fluid dynamics package Fluent 6.4.11 (as marketed by Ansys Inc.) to simulate the flow fields. Single-phase simulations were performed, and the effect of the dispersed phase was incorporated in the fluid viscosity using the Krieger-Dougherty relation for effective viscosity.28 (26) Kataoka, K. Encyclopedia of Fluid Mechanics; Gulf Publishing Company: Houston, TX, 1986; Vol. 1. (27) Taylor, G. I. Phil. Trans. R. Soc. London 1923, A223–289. (28) Krieger, I. M.; Dougherty, T. J. J. Rheol. 1959, 3, 137–152.

Figure 8. Comparison between axial velocity profiles obtained from single-phase and multiphase simulations (3000 rpm).

Figure 9. Swirl and axial velocity profiles for the crude/model oil A systems (µKD ) 36 cP, 1600 rpm).

3678


Aichele et al.

Figure 10. Swirl and axial velocity profiles for the crude/model oil A systems (µKD ) 36 cP, 3000 rpm).

Figure 12. Drop size distributions of the brine-in-model-oil-A emulsions.

Figure 11. Drop size distributions of the brine-in-crude-oil-A emulsions.

of Taylor vortices begin to emerge. For these conditions, Re ) 190 and Ta ) 4712 which exceeds Tac. Using the same Krieger-Dougherty viscosity at the maximum inner cylinder rotational speed (3000 rpm), the presence of Taylor vortices intensifies as shown in Figure 10. For these conditions, Re ) 356 and Ta ) 16565 which greatly exceeds Tac. Using the same velocity scale for comparison, Figures 9 and 10 illustrate the emergence of secondary flows in the form of Taylor vortices for the crude/model oil A systems as the inner cylinder rotational speed increases. Figure 11 shows the brine-in-crude-oil-A drop size distributions as measured with NMR. At least two emulsions were measured at each rotational speed. Figure 12 shows the brinein-model-oil-A drop size distributions. The increase in multimodality of both the crude oil A and model oil A drop size distributions is attributed to the emergence of Taylor vortices in the Taylor-Couette flow device as shown by the CFD simulations. Three-dimensional simulations using an inner cylinder rotational speed equal to 3000 rpm showed the existence of a multimodal distribution of strain rates throughout the fluid domain. Figure 13 shows the maximum strain distribution as determined by a droplet tracking investigation in the threedimensional CFD simulation. This analysis consisted of injecting 18 000 water droplets with diameter equal to 1 µm across the periodic face of the three-dimensional geometry. The droplets were tracked for 5000 time steps, and the maximum strain rate experienced by each droplet is shown in Figure 13. The dominant flow direction in the fluid domain is circumferential as driven by the rotational speed of the inner cylinder. However, not all flow paths experience the same shear. This analysis shows (29) Becher, P. Emulsions: Theory and Practice, 3rd ed.; Oxford University Press: Washington, D.C., 2001.

Figure 13. Maximum strain rate distribution as obtained by a particle tracking analysis in the three-dimensional CFD simulation.

that droplets can experience different shear fields which could enhance multimodality of the drop size distribution. The experimental data were compared to two predictive models. The first model, referred to as the Grace model, is based on Grace’s experimental work showing critical capillary number as a function of viscosity ratio.12

( )

dmax ) 2Cac(λ)

µCPγ σ

-1

(11)

Grace’s model predicts the maximum stable drop diameter of a single drop in simple shear in the absence of surfactant. The second model, proposed by JMH Janssen, accounts for partially mobile water/oil interfaces, thereby relaying coalescence information during breakup for a multidrop system.30,31 (30) Grizzuti, N.; Bifulco, O. Rheol. Acta 1997, 36, 406–415. (31) Janssen, J. M. H. Dynamics of liquid-liquid mixing, Ph.D. Thesis, University of Technology, Eindhoven, The Netherlands, 1993.


dcoal ) 2λ-2/5

( )( ) 4 hc √3

2/5

µCPγ σ


-3/5

(12)

The critical thickness of the film between two approaching drops is given by hc.32 hc ∼

Hr ) ( 8πσ

1/3

(13)

H is the Hamaker constant (∼10-20 J); r is the initial drop radius; and σ is the interfacial tension. Equation 14 shows the effective shear rate that was used in this work in eqs 11 and 12. γ)

ωri (ro - ri)

Figure 16. Swirl and axial velocity profiles for the crude/model oil B systems (µKD ) 400 cP, 1600 rpm).

(14)

The angular velocity of the inner cylinder is ω, and the inner and outer cylinder radii are ri and ro, respectively. Figure 14 shows the comparison between the experimental mean and maximum diameters and the two predictive models

Figure 17. Drop size distributions of the brine-in-crude-oil-B emulsions.

for the brine-in-crude-oil-A system. The mean diameters were calculated based on eq 15, while the maximum diameters were calculated based on 99% of the cumulative volume of the drop size distribution. dv ) Figure 14. Comparison of experimental data to the Grace and Janssen models for the brine-in-crude-oil-A emulsions.

Figure 15. Comparison of experimental data to the Grace and Janssen models for the brine-in-model-oil-A emulsions.

Σ(fidi) Σfi

(15)

For the ith bin of the drop size distribution, fi is the amplitude from NMR and di is the corresponding diameter. The drop sizes of the brine-in-model-oil-A emulsions were compared to the predictive models as shown in Figure 15. Similar to the brine-in-crude-oil-A emulsions, the drop sizes of the brine-in-model-oil-A emulsions are consistent with the Grace model because the final drop size is smaller than the maximum stable drop diameter predicted by the Grace model. The Janssen model qualitatively predicted both the size and trend of the maximum diameter of the model oil A drop size distributions. Crude/Model Oil B Systems. With the dispersed phase fraction equal to 0.2, the effective viscosity for the crude/model oil B systems at 303.2 K was 400 cP. Figure 16 shows that even at the maximum inner cylinder rotational speed used experimentally the CFD simulations predict that no secondary flows develop. For these conditions, Re ) 18 and Ta ) 43 which is well below Tac. Simple shear flow dominates throughout the cell with a linear gradient in swirl velocity. Figure 17 shows the drop size distributions for the brine-incrude-oil-B emulsions as obtained by NMR. Figure 18 shows the drop size distributions for the brine-in-model-oil-B emulsions. At the same inner cylinder rotational speeds, the brine(32) Chesters, A. K. Trans. Inst. Chem. Eng. 1991, 69, 259–270.

3680


Figure 18. Drop size distributions of the brine-in-model-oil-B emulsions.

Aichele et al.

Figure 20. Comparison of experimental data to the Grace and Janssen models for the brine-in-model-oil-B emulsions.

The Grace and Janssen model comparisons are shown in Figures 19 and 20. Both the brine-in-crude-oil-B and brine-inmodel-oil-B data are consistent with the Grace model. Conclusions

Figure 19. Comparison of experimental data to the Grace and Janssen models for the brine-in-crude-oil-B emulsions.

in-crude-oil-B emulsions exhibited multimodality while the brine-in-model-oil-B emulsions showed broad, unimodal drop size distributions. Given the similarity in viscosity, the difference in emulsion morphology between the two systems likely arises from different interfacial properties and availability of surfactant during emulsification. Interfacial tension measurements indicate that with the addition of 0.356 wt % Span80 to the model oil B the surfactant migrates to the interface quickly. The crude oil B, however, contains a multitude of surfactants that take a period of time to migrate to the interface. These surfactant heterogeneities within crude oil B could possibly explain the distinct populations of drops characteristic of the emulsions formed with this crude oil. With a homogeneous surfactant concentration in the model oil B, the emulsion drops exist with unimodal drop size distributions in simple shear flow.

This work provides quantitative drop size distribution data for complex, concentrated emulsions. The PFG-DE technique effectively quantifies the morphology of these systems by not assuming a form of the drop size distribution and by considering the entire emulsion. By coupling the experimental data with computational fluid dynamics simulations, this work provides important quantitative insight about concentrated emulsion formation. The crude/model oil A drop size distributions reflect the emergence of secondary flows as verified by CFD simulations. As the inner cylinder rotational speed increases in these systems, the drop size distributions become more multimodal. CFD simulations show that the secondary flows can lead to multimodal strain rate distributions, thereby leading to multimodality in the drop size distributions. The experimental mean diameters reflect a power law dependence on ((µCPγ)/σ) as expected by theories presented in the literature. The crude/model oil B emulsions display distinctly different morphologies. The brine-in-crude-oil-B emulsions display multimodality at each inner cylinder rotational speed, while the brine-in-model-oil-B emulsions exhibit broad, unimodal drop size distributions. This distinct difference in morphology of the emulsions likely arises from differences in interfacial properties between the two oils. Acknowledgment. We thank Chevron Energy Technology Company for financial support of this project. In addition, we thank George J. Hirasaki, Mark Flaum, Dick Chronister, Alberto Montesi, and Jeff Creek for insightful discussions. Details regarding the PFGDE technique can be found in Mark Flaum’s thesis.18 EF900192V

Analysis of Formation of Water-in-Oil Emulsions - Energy & Fuels

Recommend Documents