Characterization of water-in-crude-oil emulsions in a complex shear

?eld

Clint P.Aichele a,?,Walter G.Chapman b,Lee D.Rhyne c,Hariprasad J.Subramani c

a School of Chemical Engineering,Oklahoma State University,423Engineering North,Stillwater,OK74078,United States

b Department of Chemical and Biomolecular Engineering,Rice University,6100S.Main St.,MS362,Houston,TX77005,United States

c Chevron Energy Technology Company,1400Smith St.,Houston,TX77002,Unite

d States

a r t i c l e i n f o

Article history:

Received12August2013

Received in revised form26November2013 Accepted5December2013

Available online16December2013

Keywords:

Emulsions

Crude oil

NMR

Drop size distribution

Complex shear

Multi-phase?ow a b s t r a c t

Understanding the formation and stability characteristics of concentrated emulsions generated in com-plex?ow?elds is of importance to the energy industry.Knowledge about the transient stability of water-in-crude-oil emulsions aids in formulating?ow assurance strategies to problems such as water separation and gas hydrate formation.During the production of crude oil,the produced?uids can expe-rience a wide range of?ow conditions,such as?ow through pumps,reductions and expansions,and per-forations which affect both the formation and stability of the emulsions.This work focuses on complex water-in-crude-oil emulsion formation and stability.The emulsions are formed with a rotating turbine using synthetic seawater as the dispersed phase and two crude oils as the continuous phase.This work employs the Carr–Purcell–Meiboom–Gill

(NMR)technique to mea-

sure the transient drop size distributions.The mixing Reynolds number,

energy input,and mixing time on emulsion The effect of mild

shear after emulsi?cation was also evaluated.

1.Introduction

Formation and stability characteristics of emulsions created in

complex?ow?elds are particularly applicable to the energy

industry[1].The transient stability of water-in-crude-oil emul-

sions affects both water separation and methane hydrate forma-

tion in?owing systems.The shear history of an emulsion has a

profound impact on the overall behavior of the emulsion through-

out its life in a process system.In addition,the presence of contam-

inants and natural surfactants in crude oil systems affects the

stability of crude oil emulsions.Fundamental understanding and

quanti?cation of concentrated emulsion formation and stability

mechanisms in complex crude oil systems is still lacking.

In the laboratory,emulsions are generated using a variety of

techniques,one of which is the use of a turbine mixer such as a

Rushton turbine[2].Rushton turbines have been widely used in

industry and academia to investigate mixing phenomena and drop

breakup[2,3].Rushton turbines facilitate adequate mixing of the

immiscible?uids and provide suf?cient shear to disperse water

drops in oil.In addition,published correlations facilitate the calcu-

lation of power and energy applied to the emulsion systems using

Rushton turbine mixers.

Hinze pioneered drop breakup in turbulent?ow by showing

that the maximum equilibrium drop size(d max)was a function of

the energy dissipation rate[4].

d max?C eà0:4

0:6

e1T

In this expression,C is a correlation constant,e is the energy dissi-

pation rate,r is the interfacial tension,and q CP is the density of the

continuous phase.Work by Hinze showed that drop breakup in a

stirred tank environment primarily occurs in the impeller region

due to turbulent pressure?uctuations along the drop’s surface

[4].Subsequently,Chen and Middleman used similar arguments

to empirically derive an expression for the Sauter diameter for di-

lute emulsions with an inviscid dispersed phase produced by Rush-

ton turbines[5].

d32

?0:053Weà3=5e2T

The Sauter mean diameter,d32,is the size of a drop which has

the same volume-to-surface area ratio as the entire population of

drops[6].The impeller diameter is represented by the symbol L

in Eq.(2).The Sauter diameter gives important information about

the available interfacial area of the drops in the distribution.The

Sauter diameter is commonly used in the literature to describe

https://www.360docs.net/doc/8c14117658.html,/10.1016/j.exptherm?usci.2013.12.005

Corresponding author.Tel.:+14057449110.

E-mail address:clint.aichele@https://www.360docs.net/doc/8c14117658.html,(C.P.Aichele).

drop breakup in turbulent?ow that incorporates turbine mixers [7].The Weber number,We,is de?ned according to the following equation:

We?q

N2L3

re3T

The density of the continuous phase is q CP,the rotational speed of the impeller is N,the diameter of the impeller is L,and the interfacial tension is r.Calabrese accounted for the viscosity of the dispersed phase and adjusted the correlation accordingly[8].

d32 L ?0:053Weà3=51t0:97V0:79

h i3=5

e4TThe viscosity group that accounts for the impact of the dis-

persed phase viscosity is given by V i.

V i?

1=2

re5T

The dispersed phase viscosity is represented by l DP.

To account for increased dispersed phase fractions,Calabrese developed an empirical formulation for dispersed phase fractions less than or equal to0.2[8].

d32 L ?0:054e1t3uTWeà3=51t4:42e1à2:5uTV i

d32

1=3

"#3=5

e6T

The dispersed phase volume fraction is represented by u.Eq.(6)ac-counts for the effect of dispersed phase volume fraction on disrup-tive energy.

A lack of data exists for concentrated emulsion formation in complex?ows in the presence of surfactants,and there is presently a need for the direct measurement of drop size distributions of opaque emulsions such as crude oil emulsions.The NMR technique employed in this work?lls that gap by directly measuring the emulsion droplet size distributions and transient behavior of con-centrated crude oil emulsions.The direct measurements of real crude oil systems illustrate the complexities that arise in these systems.

2.Experimental methods

For all experiments presented in this paper,the dispersed phase was ASTM D1141synthetic seawater(also referred to as brine in this paper)purchased from Ricca Chemical.The volume fraction was equal to0.2by combining10mL brine with40mL oil.Two crude oils were used in this work,and the densities of all?uids used are contained in Table1.The viscosities of the two oils were measured using a Brook?eld rheometer for a range of temperatures as shown in Fig.1.Fig.2shows the interfacial tensions for both crude oils in the presence of brine that were measured using the pendant drop technique.A KSV CAM200was used to perform the pendant drop measurements.

A six bladed turbine,also referred to as a Rushton turbine,was used to form the emulsions.This emulsi?cation system was se-lected based on its ability to generate?ow conditions in the turbu-lent regime[9].A digital tachometer was used to measure the rotational speed of the impeller.The diameter of the impeller was30mm.The emulsions were prepared in a glass vessel with the inside diameter equal to43mm and the length equal to 210mm.Fig.3shows the dimensions of the turbine relative to the glass vessel.The length of the rotor shaft was198mm,as shown in Fig.4.The turbine blades were approximately5mm thick.For all experiments,the emulsions were mixed and mea-sured in the same vessel.Therefore,the entire emulsion that was prepared was also used for the measurement.Because of the high viscosity of the crude oil B,the brine-in-crude-oil-

B emulsions were formed in a circulating water bath with temperature equal to303.2K.A two sample Wheaton bench top roller was used to ap-ply mild shear to select samples after emulsi?cation.For all exper-iments for which mild shear was applied after emulsi?cation,the rotational speed was set to4rpm.Table2summarizes the exper-imental conditions evaluated in this work.The Reynolds numbers reported in Table2were calculated according to the following equation[2,3].

Re?

L2N q CP

e7T

Based on correlations by Rushton et al.[2,3],the power number, N P,was determined and the power input to the system was calculated.

N P

N L

aàlogeReT

àá

N3L5

e8T

Nomenclature

a,b,C empirical constants

d i diameter of th

e i th bin(l m)

d max max.stabl

e diameter(l m)

d v volum

e weighted mean diameter(l m) d32Sauter diameter(l m)

E energy input(kJ)

f i amplitude of the i th bin(a.u.)

g gravitational constant(m2/s)

L impeller diameter(mm)

N rotational speed of impeller(rpm)

N P power number

P power input(W)

Re Reynolds number

T2transverse relaxation(ms)

T2,bulk bulk transverse relaxation(ms)T2,DP,i transverse relaxation of the dispersed phase of the i th bin(ms)

T2,lm log mean of transverse relaxation distribution(ms)

V i dimensionless viscosity group

We Weber number

e energy dissipation rate(J/s)

r interfacial tension(mN/m)

SR surface relaxivity(l m/s)

q CP continuous phase density(g/mL)

q DP dispersed phase density(g/mL)

l CP continuous phase viscosity(Ns/m2)

l DP dispersed phase viscosity(Ns/m2)

u dispersed phase volume fraction

s time between p/2and p pulses(l s)

Table1

Densities of the?uids.

Density(g/mL) Synthetic seawater 1.03

Crude oil A0.85

Crude oil B0.91

C.P.Aichele et al./Experimental Thermal and Fluid Science53(2014)190–196191

A2MHz NMR spectrometer was used to measure the trans-verse relaxation distributions of the water-in-crude-oil-emulsions using the Carr–Purcell–Meiboom–Gill(CPMG)[10,11]technique.The NMR pulse sequence for the CPMG technique is given in Fig.5.The decay of the magnetization vector in the transverse plane,M xy,is governed by a multi-exponential function[6].

M xye2n sT

M xye0T

X m

i?1

f i expà

2n s

T2;i

e9T

The time between p/2and p pulses is given by s and the num-ber of echoes collected is n.To obtain a T2distribution,the multi-exponential function is?t to the experimental data.Based on the designated list of T2values,a corresponding list of f i values is ob-tained.The f i values represent the fraction of the1H nuclei which have a relaxation time of T2,i.Multiple sets of f i satisfy the?tting [12],so a regularization method developed by Huang provides the best?t to the multi-exponential function[13,14].Examples of T2distributions are provided in Fig.6which contains the T2dis-tributions for crude oil A,crude oil B,and brine at303.2K.Both crude oil T2distributions are broad indicating the presence of a multitude of components.By contrast,the pure brine consists of a narrow peak with log mean value equal to2550ms(ms).

The T2distribution of the dispersed phase of an emulsion(T2,DP) is a function of the bulk T2value of the dispersed phase(T2,bulk),the surface relaxivity(SR),and the surface:volume ratio[6].Therefore, the drop size distribution of the emulsion can be obtained using the following equation:

d i?6SR

T2;DP;i

T2;bulk

à1

e10T

Pena demonstrated the utility of using this NMR technique for emulsion characterization with con?rmation by optical micros-copy[6].

The surface relaxivity of a water/oil system can be determined by performing a pulsed?eld gradient with diffusion editing (PFG-DE)measurement and thereby measuring both the mean drop size and mean T2[15–18].

SR?

d v

T2;lm

T2;bulk

e11T

The volume weighted mean diameter is denoted by d v and the log mean of the T2dispersed phase T2distribution is T2,lm.The sur-face relaxivity between the brine and crude oil A was measured to be0.4l m/s.The surface relaxivity of the brine-in-crude-oil-B sys-tem was measured to be0.9l m/s.

3.Results and discussion

3.1.Re$3000with no applied shear after emulsi?cation

The Reynolds number,Re,provides important information about the conditions that lead to emulsion formation and transient stability.This section contains results for high mixing Reynolds number conditions(Re$3000)with no applied shear after emulsi-?cation.Due to the limitations of the emulsi?cation equipment, only crude oil A emulsions are considered in this section.The Sau-ter diameter was calculated for each distribution[6].

d32?

P M

i?1

f i

P M

i?1

f i

e12T

The drop size distributions were obtained at several times after preparation(0min,30min,1h,2h,5h,and24h).Fig.7 shows the transient mean Sauter diameters for10min and 1min of mixing.The error bars shown in Fig.7were calculated based on the sample standard deviation of multiple points at each time.The broad error bars with1min of mixing are attributed to the heterogeneity of the emulsion at short mixing times.With

192 C.P.Aichele et al./Experimental Thermal and Fluid Science53(2014)190–196

10min of mixing,the emulsion has more time to experience mix-ing;therefore,the error bars are narrower compared to 1min of mixing.Fig.8shows that with one minute of mixing,the trans-verse relaxation of the dispersed phase overlapped the bulk brine T 2value 24h after emulsi?cation,thereby introducing an increase in width of the drop size distribution.Eq.(10)shows that if T 2,DP overlaps T 2,bulk ,ambiguous results arise in the drop size distribu-tion.Therefore,the CPMG technique was unable to yield the drop

size distribution 24h after emulsi?cation.These results indicate that the mixing time impacted both emulsion formation and tran-sient stability of these emulsions.

3.2.Re $3000with applied shear after emulsi?cation

This section presents results for brine-in-crude-oil-A emulsions formed with high mixing Reynolds numbers and applied shear after emulsi?cation.Fig.9shows that the emulsions resisted coalescence

198 mm

turbine that was used to make the emulsions.The mixer was connected to a shaft with length equal Table 2

Summary of experimental conditions.Crude oil type Mixing time (min)Re P (W)E (kJ)Shear after emulsi?cation A 1030657.0 4.2No A 12963 6.50.4No A 102758 5.4 3.3Yes time progresses,the magnetization vector in the transverse plane decays according to the transverse

1520

2530d 32 (μm )

C.P.Aichele et al./Experimental Thermal and Fluid Science 53(2014)190–196193

in the presence of a mild shear,as was expected based on work in the literature for dilute emulsions in the presence of surfactants[19]. Nandi observed a reduction in the coalescence frequency of1vol.% emulsions in the presence of surfactant by applying mild shear. The authors attributed this observation to the ability of surfactants to stabilize the thin?lms between drops,and the applied shear rate minimized the contact time during droplet collisions,thereby reduc-ing the coalescence rate[19].Fig.9also shows that the mean diam-eter experiences a small decrease over time indicating that an equilibrium size has not been achieved24h after preparation in the presence of a mild shear.The results for both10min and1min of mixing with applied shear after emulsi?cation indicate that the emulsions resisted coalescence throughout the duration of the experiments.In contrast to1min of mixing with no applied shear after emulsi?cation,1min of mixing with applied shear after emul-si?cation produced emulsions with drop size distributions that could be measured24h after formation.

3.3.Re$400with no applied shear after emulsi?cation

This section contains results for emulsions formed using low mixing Reynolds numbers($400).Both crude oil A and crude oil B emulsions were considered in this mixing?ow regime.Fig.10 illustrates that for all times after emulsi?cation,the dispersed T2 distributions overlapped the bulk brine T2value for the crude oil A emulsions.Therefore,no drop size distributions were obtained at any time after emulsi?cation.

Brine-in-crude-oil-B emulsions were also formed using low mixing Re conditions.Though the Reynolds numbers for the crude oil B and crude oil A emulsions were similar,the power required for making emulsions with the crude oil B(91.5W)was larger than the power required to make emulsions with the crude oil A (0.04W).Fig.11shows that the mean Sauter diameters for both mixing times did not change signi?cantly over24h for the crude oil B emulsions.For the same mixing Reynolds number as the brine-in-crude-oil-A emulsions,the drops in the brine-in-crude-oil-B emulsions remained dispersed throughout the duration of the experiments,while the brine-in-crude-oil-A emulsions re-sulted in two phase separation.

3.4.Re$400with applied shear after emulsi?cation

The effects of mild shear after low mixing Reynolds numbers (Re$400)was investigated in both the crude oil A and crude oil B systems.Fig.12shows that for the brine-in-crude-oil-A emul-sions,the dispersed brine T2distributions overlapped the bulk brine T2value.Drop size distributions were not obtained at any time due to the overlap of the dispersed phase T2distributions and the bulk brine T2value.

Fig.13shows the mean Sauter diameters as a function of time after emulsi?cation for the crude oil B https://www.360docs.net/doc/8c14117658.html,ing low mixing Reynolds numbers for the brine-in-crude-oil-B emulsions with ap-

194 C.P.Aichele et al./Experimental Thermal and Fluid Science53(2014)190–196

plied shear after emulsi?cation produced emulsions that resisted coalescence throughout the duration of the experiments.

https://www.360docs.net/doc/8c14117658.html,parison to published correlations

Correlations exist for drop size as a function of?ow conditions for emulsions formed with Rushton turbines in turbulent?ow [5,7,8,20].Fig.14illustrates the comparison of crude oil A drop size data to these correlations.The correlations by Chen and Middle-man and Wang were developed for dilute systems in the absence of surfactant in turbulent?ow formed by Rushton turbines in baf-?ed tanks.The correlation developed by Wang accounts for the vis-cosity of the dispersed phase.The correlation developed by Calabrese accounts for the dispersed phase fraction,and it was developed for turbines in baf?ed tanks.However,it does not account for the presence of surfactant.The deviation of the exper-imental data from the correlations at high Weber numbers could be caused by the inability of the correlations to account for dynamic interfacial tension effects that could lead to smaller drop sizes than predicted.The dynamic nature of the interfacial behav-ior of complex surfactants in crude oil systems must be accounted for in order to develop a robust model to predict this behavior. Such required data includes the dynamic interfacial properties such as the dilatational modulus and interfacial elasticity.With this dynamic interfacial data,existing equilibrium models could be tailored to account for the dynamic interfacial behavior,or ad-vanced predictive models could be formulated to account for the heterogeneities in concentrated crude oil emulsions.

4.Conclusions

The drop size distributions of brine droplets in brine-in-crude-oil emulsions are sensitive to an assortment of parameters includ-ing energy input,crude oil viscosity,and interfacial properties. These parameters affect both emulsion formation and transient stability of the emulsions.In this study,the mixing Reynolds num-ber,the mixing time,power input,energy input,and applied shear after emulsi?cation were evaluated in order to observe their effects on the transient drop size distributions in brine-in-crude-oil emulsions.

The brine-in-crude-oil-A emulsions remained dispersed at high mixing Reynolds numbers($3000)and ten minutes of mixing. With one minute of mixing,the drop size distributions of the emul-sions could not be measured24h after formation.Applied shear after emulsi?cation slightly reduced the drop size and reduced coalescence of the emulsions formed with1min of mixing.

At low values of the mixing Reynolds number($400),the brine-in-crude-oil-A emulsions displayed coalescence immediately after preparation,and applied shear after emulsi?cation did not prevent coalescence.In the same range of the mixing Reynolds number,the brine-in-crude-oil-B emulsions produced measurable drop sizes and resisted coalescence for the duration of the experiments. Application of shear after emulsi?cation to the brine-in-crude-oil-B emulsions had little effect on the drop size distributions.

Acknowledgements

The authors thank Chevron ETC for?nancial support of this pro-ject.In addition,the authors thank George J.Hirasaki,Mark Flaum, Alberto Montesi,and Jeff Creek for insightful discussions. References

[1]C.W.Angle,H.A.Hamza,T.Dabros,Size distributions and stability of toluene

diluted heavy oil emulsions,AIChE J.52(3)(2006)1257–1266.

[2]J.H.Rushton, E.W.Costich,H.J.Everett,Power characteristics of mixing

impellers:Part I,Chem.Eng.Prog.46(8)(1950)395.

[3]J.H.Rushton, E.W.Costich,H.J.Everett,Power characteristics of mixing

impellers:Part II,Chem.Eng.Prog.46(9)(1950)467.

C.P.Aichele et al./Experimental Thermal and Fluid Science53(2014)190–196195

[4]J.O.Hinze,Fundamentals of the hydrodynamic mechanism of splitting in

dispersion processes,AIChE J.1(3)(1955)289–295.

[5]H.T.Chen,S.Middleman,Drop size distribution in agitated liquid–liquid

systems,AIChE J.13(5)(1967)989–995.

[6]A.Pe?a,G.J.Hirasaki,NMR Characterization of Emulsions,in:J.Sjoblom(Ed.),

Emulsions and Emulsion Stability,CRC,Boca Raton,FL,2006,pp.283–309. [7]R.V.Calabrese,T.P.K.Chang,P.T.Dang,Drop breakup in turbulent stirred-tank

contactors.Part I:Effect of dispersed-phase viscosity,AIChE J.32(4)(1986) 657–666.

[8]R.V.Calabrese,C.Y.Wang,N.P.Bryner,Drop breakup in turbulent stirred-tank

contactors Part III:Correlations for mean size and drop size distribution,AIChE J.32(4)(1986)677–681.

[9]J.A.Boxall,C.A.Koh,E.D.Sloan,A.K.Sum,D.T.Wu,Droplet size scaling of

water-in-oil emulsions under turbulent?ow,Langmuir28(2011)104–110.

[10]H.Y.Carr, E.M.Purcell,Effects of diffusion on free precession in nuclear

magnetic resonance experiments,Phys.Rev.94(1954)630–638.

[11]S.Meiboom, D.Gill,Modi?ed spin-echo method for measuring nuclear

relaxation times,Rev.Sci.Instrum.29(1958)688–691.

[12]A.N.Tikhonov,V.Y.Arsenin,Solution of Ill-Posed Problems,Winston&Sons,

Washington,DC,1977.

[13]C.C.Huang,Estimation of Rock Properties by NMR Relaxation Methods,Rice

University,Houston,1997.[14]K.J.Dunn,https://www.360docs.net/doc/8c14117658.html,Torraca,J.L.Warner,D.J.Bergman,On the Calculation and

Interpretation of NMR Relaxation Times Distribution.in:69th Annual Technical Conference and Exhibition,New Orleans,LA,1994,pp.45–54. [15]M.D.Hurlimann,L.Venkataramanan,Quantitative measurement of two

dimensional distribution functions of diffusion and relaxation in grossly inhomogenous?elds,J.Magn.Reson.157(2002)31–42.

[16]L.Venkataramanan,Y.Q.Song,M.D.Hurlimann,IEEE Trans.Signal Process.50

(2002)1017–1026.

[17]C.P.Aichele,M.Flaum,T.Jiang,G.J.Hirasaki,W.G.Chapman,Water in oil

emulsion droplet size characterization using a pulsed?eld gradient with diffusion editing(PFG-DE)NMR technique,J.Colloid Interface Sci.315(2007) 607–619.

[18]M.Flaum,Fluid and rock characterization using new NMR diffusion-editing

pulse sequences and two dimensional diffusivity-T2maps,Rice University, Houston,2006.

[19]A.Nandi,A.Mehra,D.V.Khakhar,Suppression of coalescence in surfactant

stabilized emulsions in shear?ow,Phys.Rev.Lett.83(12)(1999)2461–2464.

[20]C.Y.Wang,R.V.Calabrese,Drop breakup in turbulent stirred-tank contantors.

Part II:Relative in?uence of viscosity and interfacial tension,AIChE J.32(4) (1986)667–676.

196 C.P.Aichele et al./Experimental Thermal and Fluid Science53(2014)190–196