OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary

OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary OIL SPILL SCIENCE chapter 15 – oil spill dispersants a technical summary

Oil Spill Dispersants: A Technical Summary

Recently, the National Academy of Sciences released its study of the use ofchemical dispersants in the United States.1This report is particularly instructiveand provides some useful assessments of the situation Their assessments andrecommendations will be summarized in the applicable sections of this chapter.The prime motivation for using dispersants has been stated to be reduction

of the impact of oil on shorelines To accomplish this reduction, the dispersantapplication must be highly successful and effectiveness high As some oilwould still come ashore following treatment, there is much discussion on whateffectiveness is required to significantly reduce the shoreline impact.2A majorissue that remains is the actual effectiveness during spills so that these valuesOil Spill Science and Technology DOI: 10.1016/B978-1-85617-943-0.10015-2

can be used in estimates and models in the future A significant physical factmust also be considereddthat is, the lifetime of the dispersion Because not alldispersions are stable and will degrade to surface slick and some residualdispersion, the utility of dispersants in any case should consider this fact.The second motivation for using dispersants is to reduce the impact on birdsand mammals on the water surface As the National Academy of Sciences (NAS)committee on dispersants states, little or no research on this matter has beencarried out anytime since the 1980s In their report (p 274) they note the following:

Of additional concern is the effect of dispersed oil and dispersants on the waterproofproperties of feathers and their role as thermal insulators One of the recommendations

of the NRC (1989) report was that studies be undertaken to “assess the ability of fur andfeathers to maintain the water-repellency critical for thermal insulation under dispersedoil exposure conditions comparable to those expected in the field.” This recommendation

is reaffirmed because of the importance of this assumption in evaluating the mental trade-offs associated with the use of oil dispersants in nearshore and estuarinesystems because it has not been adequately addressed.1

environ-The third motivation for using dispersants is to “promote the biodegradation

of oil in the water column.” The effect of dispersants on biodegradation is still

a matter of dispute A number of papers state that dispersants do not promotebiodegradation, whereas others indicate that dispersants suppress biodegrada-tion The most recent papers, however, confirm that promotion or suppression is

a matter of the surfactant in the dispersant itself and the factors of mental conditions More details of recent findings will appear in the subsequentdiscussion What is very clear at this time is that the surfactants in some of thecurrent dispersant formulations can either suppress or have no effect onbiodegradation Further, there are issues about the biodegradability of thesurfactants themselves, and this fact can confound many tests of dispersed oilbiodegradation Several questions remain unanswered, however An importantissue that never comes up is that it is known that oil-degrading bacteria largelylive on the water surface, where they would feed on natural hydrocarbons in theabsence of spills Would not putting oil in the water column then remove it fromthese bacteria? However, in the case of oil seeps or oil-contaminated sediments,there are microbial colonies associated at depth Another serious question isthat of timescale Biodegradation takes place over weeks, months, and years.Dispersion half-lives are 12 to 24 hours

environ-This author prepared a review of dispersants in 2002 and covered the period

to 1997.2Another review covered the period from 2002 to mid-2008.3The latterreview was combined with the earlier review to provide coverage from 1997 to

2008.4The comprehensive review contained over 450 references Literature notcovered in this present summary is covered in the reviews.5Another biblio-graphic search was published during this time as well but did not contain

a review.6Previous reviews covered the various oil spill dispersant topics.7Allthese are reviews of the literature, and most cover similar topics as this review

15.1.1 What Are Dispersants?

Many surfactant mixtures for treating oil spills have been promoted in thepast four decades to overcome the extensive problems and costs of physicalrecovery Of particular interest in this section are dispersants; these areformulations containing surfactants as active ingredients

Surfactants have varying solubilities in water and varying actions toward oiland water.8The parameter used to characterize surfactants is the hydrophilic-lipophilic balance (HLB) HLB is determined using theoretical equations thatrelate the length of the water-soluble portion of the surfactant to the oil-solubleportion of the surfactant A surfactant with an HLB between 1 and 8 promotesthe formation of water-in-oil emulsions and one with an HLB between 12 and

20 promotes the formation of oil-in-water emulsions A surfactant with an HLBbetween 8 and 12 may promote either type of emulsion, but generally promotesoil-in-water emulsions Dispersants have an HLB in this range

Dispersants are formulated to “disperse” oil slicks into the sea or anotherwater body Surface-washing agents, or beach cleaners as they are sometimescalled, are surfactant formulations designed to remove oil from solid surfacessuch as beaches Emulsion breakers and inhibitors are intended to break water-in-oil emulsions or to prevent their formation

Although many of these treating agents have been promoted, few are stillbeing produced More than 100 dispersants have been tested for toxicity andeffectiveness by Environment Canada, but only 2 remain on the department’s list

of accepted products.9The compendium of oil spill treating agents prepared bythe American Petroleum Institute in 1972 lists 69 dispersants and 43 surface-washing agents, most of which are also listed as dispersants.10Only two of theseare commercially available today, each being produced in a different formulation.More than 300 surface-washing agents have been sold in the North Americanmarket, but only about 36 of these are still commercially available There were 26surface-washing agents on the U.S National Contingency Plan List in 2010 It isestimated that approximately 600 dispersants have been sold worldwide, of whichonly about 200 were ever tested in the lab or field, even in a limited way Theabundance of products makes it difficult for potential buyers and environmen-talists to discriminate between effective products and those that are ineffective orcould actually cause more damage than if the oil were left without intervention.15.2 THE BASIC PHYSICS AND CHEMISTRY OF DISPERSANTS 15.2.1 Formulations

Dispersants are oil spill treating agents formulated to disperse oil into water inthe form of fine droplets Typically, the HLB of dispersants ranges from 9 to 11.Ionic surfactants can be rated using an expanded scale and have HLBs rangingfrom 25 to 40 Ionic surfactants are strong water-in-oil emulsifiers, very soluble

in water, and relatively insoluble in oil, which generally work from the water

onto any oil present Such products disappear rapidly in the water column andare not effective on oil Because they are readily available at a reasonable price,however, many ionic surfactants are proposed for use as dispersants Theseagents are better classified as surface-washing agents Some dispersants containionic surfactants in small proportions, yielding an average HLB more toward

15 than 10 Studies on the specific effect of this mixing on effectiveness ormode of action have not been done

A typical dispersant formulation consists of a pair of nonionic surfactants inproportions to yield an average HLB of 10 and some proportion of ionicsurfactants Studies have been done on this mixture, one of which usedstatistical procedures in an attempt to determine the best mixture of the threeingredients.11 An improvement in performance was claimed by adjusting thethree ingredients Several patents are held on dispersants.12-14 The typicalingredients, from patents, are listed inTable 15.1 Some dispersants listed foruse in Canada, the United States, and Europe are listed inTable 15.2.15.2.2 Nature of Surfactant Interaction with Oil

Surfactants interact with oil and oil droplets to yield a temporary lower-energystatedgiven many conditions and circumstances 15-18 The disperse state isoften called an emulsion, and in the oil spill trade it is known as a dis-persiondto distinguish these oil-in-water emulsions from water-in-oil emul-sions (called emulsions and sometimes mousse) Some surfactants will alignalong slick and droplet interfaces and thus promote the temporary stabilization

of droplets in water This droplet stabilization is enhanced by the presence ofsurfactants at the interface

TABLE 15.1 Contents of Dispersants (patent information)

Hydrocarbon-based -1 Sorbitan monooleate

Ethoxylated monooleate

Na dioctyl sulfosuccinate Solvent - hydrocarbon and butyl cellosolve Hydrocarbon-based-2 Sorbitan monooleate

Ethoxylated sorbitan monooleate Ethoxylated sorbitan trioleate

Na tridecyl sulfosuccinate Solvent - hydrocarbon and butanols Hydrocarbon-based-3 Mixtures of polyethylene glycol monoleate

Solvent- hydrocarbon Aqueous-based-1 Tall oil esters (35%), ethyl dioxitol (47%)

Sorbitan monolaurate (7%), water (10%) Calcium Sulfonate (1%)

TABLE 15.2 Listed Dispersants in Various Countries (lists may not be

complete due to changes with time)

Product Manufacturer/Origin Canada

United States Britain France Corexit 9500

Enersperse xx BP, Britain (old stocks) U

Saf-Ron Gold Sus Env Tech., Mesa, AZ U

Sea Brat #4 Alabaster, Pasadena, TX U

Agma DR 379,

OSD 569

15.3 THE BASIC NATURE OF DISPERSIONS OR OIL-IN-WATER EMULSIONS

It is well known that most emulsions are not stable and will break down intotheir constituent parts This effect is due to a large number of forces, as will bedescribed in this section, but also to the fact that the stabilizers or surfactants actusing weak forces It is also known that chemically dispersed oil destabilizesafter the initial dispersion There is an extensive body of literature on surfac-tants and interfacial chemistry, which includes an abundance of experimentaldata on the topic as well as many theoretical approaches to it This report willsummarize both the data and the theory The phenomenon of resurfacing oil isthe result of two separate processes: destabilization of an oil-in-water emulsionand desorption of surfactant from the oil-water interface

Almost every paper on the topic of the stability of emulsions notes thatemulsions are not stable.15-20There are also many books on the topic.21-24It isnoted that emulsions are not thermodynamically stable, but may be kineticallystable depending on the timescale considered In the case of kinetic stability,emulsions are not stable in terms of years, and the scale of time consideredtypically relates to consumer products and may be months In terms of oil spilldispersions, half-life may be only a matter of hours The destabilization of

TABLE 15.2 Listed Dispersants in Various Countries (lists may not becomplete due to changes with time)dcont’d

Product Manufacturer/Origin Canada

United States Britain France Super-dispersant 25 Oil Slick Dispersants, UK U

Veclean Oil Dispersant Westchem B.V., NL U

oil-in-water emulsions such as chemical oil dispersions is a consequence of thefact that emulsions are not thermodynamically stable Natural forces move theemulsions to a stable state, which consists of separated oil and water What isimportant is the rate at which this occurs An emulsion that stays sufficientlystable until long past its practical use consideration may be said to be kineti-cally stable.

There are several forces and processes that result in the destabilization andresurfacing of oil-in-water emulsions such as chemically dispersed oils Theseinclude gravitational forces, surfactant interchange with water and subsequentloss of surfactant to the water column, creaming, coalescence, flocculation,Ostwald ripening, and sedimentation

15.3.1 Forces of Destabilization

15.3.1.1 Droplet Separation

The most important force in resurfacing oil droplets from an oil-in-wateremulsion is gravitational separation.25 Droplets in an emulsion tend to moveupward when their density is lower than that of water This is true for almost allcrude oil and petroleum dispersions as they usually have droplets with a densitylower than that of the surrounding water Dense or heavy oils are poorly, if at all,dispersible The rate at which oil droplets will rise due to gravitational forces isdependent on the difference in density of the oil droplet and the water, the size ofthe droplets (Stokes’s Law, as will be described in Section 15.3.2), and therheological properties of the continuous phase The rise rate is also influenced

by the hydrodynamical and colloidal interactions between droplets, the physicalstate of the droplets, the rheological properties of the dispersed phased, theelectrical charge on the droplets, and the nature of the interfacial film

Creaming is a process that is simply described by the appearance of thestarting dispersed phase at the surface.25Creaming is the process that might bedescribed in the oil spill world as resurfacing Robins describes creaming atlength, noting that it is a very important phenomenon in the food-processingbusiness.19As much as 40% of the cost of developing a new food emulsioninvolves the long-term testing related to creaming Examples of this includeyogurt, whipped cream, jam, and many other types of food Sedimentation is thereverse of creaming and occurs when the dispersed phase is denser than water.Coalescence is the joining of two or more droplets to form a larger droplet.Coalescence is an important destabilization process in oil spills Changes indroplet size resulting in coalescence have been monitored as an emulsiondestabilizes

Ostwald ripening may be an understated mechanism in the destabilization

of oil-in-water emulsions.25 Basically, Ostwald ripening is the growth oflarger emulsion droplets by absorption of soluble components from the watercolumn The effect is to remove soluble material from the water column andsmaller droplets, resulting in an increased growth of the larger droplets The

phenomenon occurs because the soluble components of the dispersed phase aremore soluble in the larger droplets than in the water and the smaller droplets.Although the Ostwald ripening phenomenon has not been investigated with oil-in-water emulsions to the same extent as other phenomena, it is believed to bequite important Studies of undecane, hexadecane, benzene, and octane-in-water emulsions have shown that Ostwald ripening is an important factor indestabilization.25

Flocculation is another process that occurs when two particles cometogether to form an agglomerate of particles, but the particles do not coalesce

15.3.1.2 Surfactant Separation

It is well known that there is an exchange of surfactants between the target dropletand the surrounding water.18 This promotes destabilization of the emulsion.When the water is in a large ratio to the droplet concentration, surfactant is largelylost and destabilization is relatively rapid In laboratory tests, a small ratio ofoil-to-water then becomes important in simulating the conditions at sea.Surfactants will distribute between the bulk phase (water) and the interface

to achieve equilibrium between the two phases This equilibrium depends onthe watereoil solubility characteristics of the surfactant In a closed system,this equilibrium is achieved rapidly with little loss of surfactant In an opensystem, however, equilibrium is never achieved, the surfactant leaches into thewater, and over a period of hours, little surfactant is left in the oil droplets.The Marangoni effect is an important phenomenon in terms of surfactantstability and dynamics.26 This effect is due to the tendency of surfactantconcentrations to quickly distribute over an interface If there is a deficit insurfactant concentration on one side of a droplet, the surfactant quickly moves

to restore the equilibrium concentration over the droplet The restoration ofequilibrium is known as Marangoni stabilization

Marangoni instability arises as a result of this surfactant flow because theflow continues and results in areas of greater and lesser surfactant concentrationover the droplet interface Some researchers have noted that Marangoniinstability was periodic and was about on the order of 1000 seconds for oneparticular system.27 It was noted that convective instability periodicallyswitched between a slow and a more rapid transport regime During

a convective stage, fast absorption of surfactant occurred with rapid inflow ofsurfactant to the interface During a diffusive stage, desorption occurred andgradients built up until the system became unstable again

Several studies have shown, both experimentally and theoretically, thatsmall surfactants will displace larger surfactants or polymers at the interface.Although studies have shown that mixed surfactant systems yield more stableemulsions as a rule, the size difference between surfactants is critical to this Amixed surfactant system with large and small surfactants will essentially bemore stable than one stabilized by the small surfactants alone, as these smallsurfactants will displace the larger surfactants at the interface This condition is

predicated on the fact that the concentration of surfactants at the interface isgreat and there actually is interference between surfactants.

Several additional forces have been described and are summarized in theliterature.25These are depicted graphically inFigure 15.1

15.3.2 The Science of Stabilization

The basis for much of the physics and chemistry surrounding emulsion stability

is that emulsions are not thermodynamically stable.15One view of this is thattwo immiscible liquids are combined or one immiscible liquid is dispersed intoanother immiscible liquid Since the interfacial tension between these two

Capillary force

Film strength Film thinning Film (Gibbs) elasticity Film thickness Film viscosity

Hydrodynamic forces

Flocculation Depletion flocculation

Surface deformation force Van der Waals force

Radius

Sub-layer viscosity

Velocity

of approach

Oscillatory structure Interactions

Brownian

movement

Marangoni circulation

Surfactant concentration Surfactant diffusion Absorption Desorption Surfactant type Surfactant chain length Steric stabilization Surfactant precipitation Surface tension

FIGURE 15.1 The forces and influences on two droplets approaching each other.

liquids will always be greater than zero despite the amount or type of tants, there is a force or energy leading toward destabilization Furthermore, theinterfacial energy is vastly increased by increasing the area between the twoliquids through the process of increasing the number of droplets This results in

surfac-an energy imbalsurfac-ance that will tend to force the two media to separate.Kinetic stability is another consideration when describing an emulsion Anemulsion is said to be kinetically stable when significant separation, usuallyconsidered to be half or 50% of the dispersed phase, occurs outside of theusable time Therefore, if the time of use is one day, an emulsion with a half-life

of more than one day may be considered to be usable In food emulsions, thisstability would be well past the stated shelf life It should be noted, however,that food emulsions are poor examples for crude oil-in-water emulsionsbecause their stability can be controlled in closed systems by adding enoughsurfactants and gelling the water media, thereby negating coalescence andsuppressing surfactant loss

The function of any emulsifying agent or surfactant is to stabilize (somewhat)

an otherwise unstable system The emulsifying agent does so by absorption at theliquideliquid interface as an oriented interfacial film This oriented film performstwo functions: (1) it reduces the interfacial tension between the two liquids andconsequently the thermodynamic instability of the system resulting from theincrease in the interfacial area between the two phases, and (2) it decreasesthe rate of coalescence of the dispersed liquid droplets by forming mechanical,steric, and/or electrical barriers around them The steric and electrical barriersinhibit the close approach of one droplet to another The mechanical barrierincreases the resistance of the dispersed particles to mechanical shock and inhibitsthem from coalescing when they do collide When emulsions form, the emulsi-fying agents or surfactants reduce the amount of work required for formation.Stability can be defined as the resistance of the droplets to coalescence.15Creaming, or standard gravity separation, was not considered to be destabili-zation in classical terms because it occurs with or without emulsifier stabili-zation The classic destabilization processes were considered to be coalescence,flocculation, and phase inversion The rate of coalescence was stated to be theonly quantitative measure of emulsion stability These factors have nowchanged to encompass broader areas, as shown in the present review

It has been found that the rate at which the droplets of a macroemulsioncoalesce from larger droplets depends on a number of factors: the physicalnature of the interfacial film, the existence of an electrical or steric barrier onthe droplets, the viscosity of the continuous phase, the size distribution of thedroplets, the phase volume ratio, and the temperature These factors are dealtwith in greater detail in this section

The physical nature of the interfacial film is important The droplets ofdispersed liquid in an emulsion are in constant motion and frequently collide Ifthe interfacial film surrounding the two colliding droplets in an emulsionruptures, the droplets will coalesce to form a larger droplet, and eventually the

emulsion will separate and break The strength of the film then becomes animportant factor in emulsion stability Highly purified surfactants tend to formweak interfacial films, whereas mixtures of different types of surfactants tend toform stronger films, although there is evidence that smaller surfactants displacelarger ones at the interface Mixtures of a water-soluble and an oil-solublesurfactant are often used in oil spill dispersants, for example, a mixture of Spanand Tween surfactants As these surfactants may be of different sizes, there may

be a problem with destabilization

The presence of an electrical charge on the dispersed droplets can create anelectrical barrier, preventing two particles from closely approaching each other.While ionic surfactants are sometimes used in oil spill dispersants for thispurpose, these surfactants will rapidly partition to the water phase in dilutesystems such as at sea Nonionic surfactants, such as those that typicallyconstitute the bulk of oil spill dispersants, have a lesser charge and pose a weakelectrical barrier to coalescence

The viscosity of the continuous phase is an important factor in dispersionstability An increase in the viscosity of the continuous phase reduces thediffusion of the droplets and thus the frequency of collisions This is given bythe classic equation:25

where D is the diffusion rate,

k is the Boltzmann constant,

T is the absolute temperature,

h is the viscosity of the liquid continuous phase, and

a is the radius of the droplets

It is obvious from this equation that the diffusion occurs inversely to theviscosity of the continuous phase This is very important for oil spill dispersion

as the viscosity of the continuous phase is that of water and is, in fact, very low.Therefore, there is a high diffusion rate, high collision rate, and potential forcoalescence This contrasts with many food emulsions in which the continuousphase is deliberately rendered viscous to reduce coalescence and thus increasestability Equation(1)predicts that oil spill dispersions will not be as stable asmany other emulsions

Another factor influencing the rate of coalescence of the droplets is the sizedistribution of the droplets The smaller the range of sizes, the more stable theemulsion The larger particles have less surface area for the volume and thus aremore thermodynamically stable and tend to grow at the expense of the smallerdroplets As this process continues, the emulsion destabilizes An emulsionwith a fairly uniform size distribution is more stable than one with a widedistribution of particle sizes This factor is also significant when discussingoil spill dispersions Oil spill dispersions have wide distribution of dropletsizes because of the nature of crude oil and the wide distribution of compounds

in it.1Oil spill dispersions are therefore less stable by nature than many othertypes of emulsions.

Temperature is an important factor in emulsion stability, for a change intemperature causes changes in interfacial tension between the two phases.Temperature can also cause differential changes in other factors such as therelative solubility of the surfactant in the two phases and in the diffusion in thesystem Emulsifying agents are usually most effective when near the point ofminimum solubility in the solvent in which they are dissolved because at thispoint they are most surface-active Since the solubility of the emulsifying agentusually changes with temperature, the stability of the emulsion also changeswith temperature The classic equation to describe this is by Smoluchowski:15

n is the number of particles per cm3

Combining this equation with diffusion equations yields an expression forthe rate of coalescence of particles and thus for the stability of the emulsion:

V is the volume of the dispersed phase, for example, volume per unit volume,

k is the Boltzmann constant,

T is the absolute temperature,

E is the energy barrier to coalescence,

h is the viscosity of the liquid continuous phase, and

A is the collision factor as defined by the left portion of the equation.This is the most important equation in describing the stability of oil-in-water emulsions, for it shows that volume and the viscosity of the continuousphase (e.g., water) are critical parameters in describing stability or increasedcoalescence In other words, oil spill dispersions will always have low stability.Further, this shows that the effect of temperature is exponentialdthat is, collisions are increased with temperaturedor emulsions are actuallymore stable at lower temperaturedthat is, from the collision point of view only.The forces between particles or droplets are an important physicalconsideration in describing stability.15The stronger the force between particles,

if opposite sign, the greater the stability of an emulsion These forces might beconsidered to be of four types: soft or electrostatic forces, hard sphere, van derWaals, and steric forces

The soft or electrostatic forces and van der Waals interparticle forces aredescribed in the well-established theory of the stability of dispersions byDerajaguin and Landau (in 1941) and Verwey and Overbeek (in 1948); thus this

between the repulsive and attractive potential energies of interaction betweenthe dispersed particles or droplets Repulsive interactions are due to either thesimilarly charged electrical double layers surrounding the particles or tosolvent-particle interactions Attractive interactions are believed to be mainlydue to the van der Waals forces between particles For dispersion to occur, therepulsive forces must be larger than the attractive forces

Stokes’s rising rate e The classic Stokes’s equation is:

s ¼ 2Drga2

where s is the rise rate,

Dr is the density difference between the disperse and droplet phases,

g is the gravitational constant,

a is the droplet radius,

F(F0) is a volume dependent correction factor and is 1 for dilute solutions,and

Dh is the difference between the viscosity of the disperse and droplet phases.This equation is very important in terms of understanding the resurfacing ofoil spill dispersions It shows that for the smallest droplets at 1 m below theslick, the rise rate would be about a year (or forever) and for the largest dropletsimmediately below the slick, rise rate is a few seconds Several researchershave shown that surfactants do not affect the base rise rate, but others questionwhether the Stokes’s rate is far too slow.25,27Many researchers have shown thatthe rise rate predicted by the Stokes’s equation is far too slow compared toexperimental measurements These might be explained by the destabilizationprocesses described in this report, namely, coalescence, flocculation, andOstwald ripening All of these processes serve to increase particle diameter andthus significantly increase the rise rate A doubling of a droplet radius results in

a quadrupling of the rise rate for that particular droplet

15.3.3 Oil Spill Dispersions

There are some measurements of the half-lives of oil and hydrocarbon sions in the literature.25Some of these papers presented data from which thehalf-life of the particular emulsion could be calculated.25This is shown inTable15.3 The half-life data for crude oil emulsions are all very similar with anaverage half-life of about 12 hours Resurfacing has been noted during severallarge tank tests as well.28

emul-Sterling et al studied the coalescence of Arabian crude oil emulsions with thedispersant Corexit 9500.29For the range of pH from 4 to 10 and salinity, 10%,

Summary Literature Data

Oil Type

Dispersant/

Surfactant

Average Half-Life (hours) Nominal

Lower Range

Upper

Other Factors

polyacrylamide

30%, and 40%, thez potentials range from e3 to e10 mV This potential wouldnot be sufficient to produce significant resistance to coalescence Coalescencekinetics of the premixed crude oil and dispersant were determined with a range ofshear rates and salinity It was found that increasing shear rate increases coa-lescence as predicted by the extensive body of literature on the topic Sterling andcoworkers found that the dispersed oil fraction decreased with increasing coa-lescence and especially with time The half-life extrapolated from the dataimplies that the half-life of the Arabian crude oil emulsion with Corexit 9500dispersant was about 5 hours for a shear rate of 5 s1, about 4.5 hours for 10 s1,about 4 hours for 15 s1, and about 3.5 hours for a shear rate of 20 s1.

It is important to emphasize that, although increased turbulence enhancescoalescence, increased turbulence at the sea surface may also result in redis-persion The two processes do not appear to balance off because in experi-ments, it was noted that maintaining energy only decreased the resurfacing rate

by about 10% over time periods of 8 to 96 hours.25

15.3.4 Significance of Emulsion Stability

Crude oil-in-water dispersions are similar to many other types of emulsions inthat they are stable under some conditions for a period of hours During thistime, destabilization processes are underway that result in oil resurfacing.Because of movement of the slick, from which the dispersion occurred andbecause the water column may have differential movement from the slick,resurfacing oil will likely appear in areas outside the residual slick As resur-facing is a slow process and goes on for many hours, most of the oil will not bevisible on the surface unless processes such as Langmuir cells were to recon-centrate the oil into slicks or unless there was no relative movement betweenthe surface slick and the water column

There is a vast body of information and experimentation and a broadconsensus on the stability of such emulsions The stability and resurfacing ofcrude oil emulsions are influenced by the following forces

1 Natural stabilization/destabilization forces

The most important force is gravity As most oils are less dense than water,their emulsion droplets are also less dense than water and will rise Thereappearance of oil on the surface is known as creaming

There are many destabilizing forces to emulsions such as chemicallydispersed oil, including coalescence, flocculation, Ostwald ripening, andphase inversion It is known that coalescence of droplets is the most importantdestabilization process for emulsions similar to dispersed crude oil emulsions

2 Standard tendency of emulsions to instability

There are many repulsive forces and attractive forces between droplets Thenet result of these forces is to destabilize the droplet after some period oftime

3 Instability of interfacial film with surfactants

The interfacial films stabilized by surfactants are subject to a number ofdestabilization processes, including Marangoni circulation, hydrodynamicdestabilization, oscillatory forces, pulsing, thermal instabilities, surfactantdesorption, and others These forces weaken the interfacial film and conse-quently destabilize the emulsion

4 Loss of surfactant

In addition to the mechanisms of interfacial stability reduction noted inpoint 3, there is a net loss of surfactant in dilute emulsions such as oil spilldispersions This net loss is caused by the tendency of the surfactants toequilibrate between the water bulk phase and the oil droplet interface Ascrude oil emulsions are continually being diluted, surfactant movementfrom the interface to achieve equilibrium constitutes a loss of surfactant

to the system This loss of surfactant accelerates the destabilization of theemulsion

5 The heterogeneous mixture of compounds in oil

Oil consists of dozens of major constituents, most of which are verydifferent in size and properties This results in the formation of verydifferent droplet sizes In addition, the effect of surfactant is quite different

on the various fractions of the oil

6 Wide distribution of droplet sizes

Because crude oil dispersions have a wide distribution of droplet sizes withmuch of the volume in the micron-sized area, the emulsions have a lowerstability It has been demonstrated that emulsions of micron-sized dropletsare less stable It has also been shown that the presence of even a few largerdroplets will destabilize an emulsion, as this triggers destabilizationprocesses such as Ostwald ripening

7 Low viscosity of water

Because the viscosity of water is low, destabilization processes are moreprevalent in water than in other bulk fluids The low viscosity of waterincreases coalescence and the diffusion of surfactants away from oil drop-lets in oil-in-water emulsions

8 Increasing dilution of the emulsion

For dilute emulsions, surfactant desorption becomes surfactant loss Furthersurfactant absorption would not occur As crude oil dispersions or emul-sions are dilute and become increasingly dilute with time, they destabilizethrough surfactant loss and through many of the other processes noted inthis report Crude oil dispersions would be considered less stable thanmost other emulsions typically studied

9 Effect of sea energy or turbulence

Increased sea energy or turbulent energy increases the amount of cence that occurs, resulting in greater resurfacing However, increasedturbulence also causes redispersion, thus offsetting the effect of the recoa-lescence somewhat

coales-Thus in summary, dispersions are at best a transient phenomenon Theeffectiveness value is a changing value at a given point in time To illustratethis, a model was developed, based on the stability equations above, to look atthe implications of these physics for the effectiveness and half-lives of emul-sions.30Figure 15.2illustrates the typical effectiveness change over a period oftime.Figure 15.3shows the change in half-life with increase depth of mixing.

as indicated by the wave height

15.4 EFFECTIVENESS

Effectiveness remains a major issue with oil spill dispersants It is important torecognize that many factors influence dispersant effectiveness, including oilcomposition, sea energy, state of oil weathering, type of dispersant used andamount applied, temperature, and salinity of the water The most important ofthese factors is the composition of the oil, followed closely by sea energy andthe amount of dispersant applied.1It is equally important to recognize that theonly thing that really counts is effectiveness on real spills at sea Moreemphasis should be put on monitoring this real effectiveness so that there is realinformation for assessment and modeling

Effectiveness issues are confounded by the simple fact that many tests,regardless of scale, show highly different results depending on how they areconstructed and operated Detailed scientific examination of most of these tests

Typical re-dispersion

2%/hour additional redispersion

FIGURE 15.2 The results of a typical series of runs of dispersion model with a 60% tiveness as a start and a 50% redispersion effectiveness This figure shows the typical decline in oil

effec-in the water column as time proceeds The two curves shows the difference effec-in assumptions, the first with typical redispersion and the second with an additional 2%/hour decline in effectiveness.

shows major deficiencies in procedure or analytical methods Further, testersshould recognize the fact that effectiveness changes with time, and this factorshould be built into testing procedures More emphasis is needed on looking atthe real results from real spills.

Another major issue is that of the toxicity of dispersants and dispersed oil,

an issue that will be discussed later Another issue to keep in mind is that oflong-term effects The long-term effects of chemically dispersed oil have notbeen well studied and therefore remain largely a topic for speculation On

a community level, there have been very few studies such as the TROPICSstudy; however, no molecular-level studies were undertaken on any of thesestudies.1,31These issues will be discussed later and form a “cluster” of majorconcerns on oil spill dispersants

15.4.1 Introduction to Effectiveness

Dispersant effectiveness is typically defined as the amount of oil that thedispersant puts into the water column compared to the amount of oil thatremains on the surface Many factors influence dispersant effectiveness,including oil composition, sea energy, state of oil weathering, type of dispersantused and the amount applied, temperature, and salinity of the water The mostimportant of these is the composition of the oil, followed closely by sea energyand the amount of dispersant applied

Change in half-life reduced redispersion (2%/hr.)

FIGURE 15.3 Plot of the half-lives of the dispersions with the two models of redispersion It can

be seen that half-lives are not affected greatly by which model of redispersion is chosen, because redispersion is much more an influence after about 120 minutes and model consideration here is about 48 hours.

One of the major confusions that persist is the relationship of effectiveness

to viscosity There is a certain belief that a “viscosity cut-off” of effectivenessfor dispersants exists.4 In fact, certain components of oil, such as resins,asphaltenes, and larger aromatics or waxes, are barely dispersible, if at all Oilsthat are made up primarily of these components will disperse poorly whendispersants are applied On the other hand, oils that contain mostly saturates,such as diesel fuel, will readily disperse both naturally and when dispersants areadded The additional amount of diesel dispersed when dispersants are usedcompared to the amount that would disperse naturally depends primarily on theamount of sea energy present In general, less sea energy implies that a higherdose of dispersant is needed to yield the same degree of dispersion as when thesea energy is high This should not be attributed to viscosity alone, but primarily

to oil composition Oils that typically contain a larger amount of resins,asphaltenes, and other heavier components are typically more viscous and lessdispersible Viscosity, however, does not track composition very well and thus isonly an indicator of dispersibility A “viscosity cut-off” does not exist

While it is easier to measure the effectiveness of dispersants in the ratory than in the field, laboratory tests may not be representative of actualconditions Important factors that influence effectiveness, such as sea energyand salinity, may not be accurately reflected in laboratory tests Resultsobtained from laboratory testing should therefore be viewed as representativeonly and not necessarily reflecting what would take place in actual conditions.When testing dispersant effectiveness in the field, it is very difficult tomeasure the concentration of oil in the water column over large areas and atfrequent enough time periods It is also difficult to determine how much oil isleft on the water surface as there are no methods available for measuring thethickness of an oil slick and the oil at the subsurface often moves differentlythan an oil slick on the surface Any field measurement at this time is bestviewed as an estimate

labo-The NAS committee on dispersants reviewed effectiveness testing.1It notedthat as the physical scale of the effectiveness increases, the cost and realismincrease, but the degree to which factors that affect dispersion can be controlledand the ability to quantitatively measure effectiveness decrease The committeealso states that when modeling or prediction is carried out, viscosity is aninsufficient predictor of dispersion efficiency The chemical composition of oil

is important, and several factors of composition have been shown to correlatewell with dispersant effectiveness Two other factors relating to dispersanteffectiveness are the dispersant-to-oil ratio and the oil-to-water ratio, but themost important factor may be the energy applied, energy dissipation rate, ormixing energy In reviewing testing, the NAS committee notes that severalimportant principles of experimental design are often ignored, includingsystematic errors that affect the outcome in one direction and random errors.Common systematic errors in dispersant effectiveness measurement includedignoring the evaporation of volatile compounds, poor analytical methods, and

incomplete recovery of floating oil These three errors, as an example given inthe NAS report, introduce a positive bias in the estimates of dispersanteffectiveness.

15.4.2 Field Trials

Previous workers and reviews have put forward discussion on field trials.1,32Field tests can provide opportunities to test and train on full-scale applicationequipment as well as to develop and test full-scale monitoring equipment and toverify oil fate and transport models Field tests, however, are subject to highcosts, and legal issues may impede the carrying out of these tests A majorlimitation on field trials is the limited data set that can be obtained from onegiven trial The experimental design of field trials is an issue, and a primaryobjective should be to obtain an unbiased estimate of the variation that existsbetween two experimental slicks Another major limitation of field trials is theinability to measure remaining oil slick thickness NAS or most scientists donot believe sorbent testing to be an accurate method Measurement of oil in thewater column is also fraught with difficulties; use of fluorometers, it has beenobserved, only gives a relative measurement The output of fluorometers alsochanges with time, aromatic composition, and so on Visual observation hasbeen used, but a suggestion to improve visual results is to use ‘blind’ observerswho are not aware of the particular treatment applied Visual observation issubject to many variables including position of the sun, cloud cover, andviewing angle The NAS committee notes that results from field trials aregenerally lower than those obtained in the laboratory, suggesting that the energyregimes in the laboratory are higher than those encountered in field trials.1Massbalances should be attempted on field trials In conclusion, the complexitiesand costs of carrying out meaningful field trials suggest that more effort beplaced on improving bench-scale and mesocosm research projects As arecommendation, the NAS committee stated that future field-scale work should

be based on systematic and coordinated bench-scale and wave-tank testing.Many field trials have been conducted in the past to assess the effectiveness

of dispersants Several papers have assessed the techniques used to measureeffectiveness in these tests.33There is no general consensus that effectivenessand other parameters can actually be measured in the field using some of thecurrent methodologies

The effectiveness determined during field trials varies significantly Recentresults, which may be more reliable, claim that dispersants removed about 10 to40% of the oil to the subsurface.1,33

The purpose of these tests were:

1 To quantify the effectiveness of dispersants on a given oil in a given cation situation

techniques

3 To measure concentrations of oil in the subsurface as a result of dispersantuse

4 To determine dispersibility conditions and relationships between factors

5 To quantify application factors, such as effect of application rate and dropletsize

Table 15.4shows the tests conducted in the past.35-62Several papers haveassessed the techniques used to measure effectiveness in these tests.33,63Theeffectiveness determined during these trials varies significantly The validity ofthe older results is even more questionable because of both the analyticalmethodology, which is now known to be incorrect, and data treatmentmethods.1,33,34,63,64

Most tests relied heavily on developing a mass balance of oil in the watercolumn and that left on the surface.33In early tests, samples from under the oilplume were analyzed in a laboratory using colorimetric methods Colorimetricmethods are not valid for this type of analysis and are no longer used Fluor-ometry has recently been used, but this method is also unreliable because itmeasures only a small and varying portion of the oil (middle aromatics) anddoes not discriminate between dissolved components and oil that actuallydispersed Furthermore, it is impossible to calibrate fluorometers for wholeoil dispersions in the laboratory without using accurate techniques such asextraction and gas chromatographic analysis It is known that the aromatic ratio

of the oil changes as a result of the dispersion process.33

In early tests, it was not recognized that the plume of dispersed oil formsnear the thicker oil in the tail of the slick and that this plume often moves off in

a separate trajectory from the slick.63 Many researchers tried to measure thehydrocarbon concentrations beneath the slick and then integrated this over thewhole slick area As the area of the plume is always far less than this area,the amount of hydrocarbons in the water column was greatly exaggerated Sincethe colorimetric techniques used at the time always yield some value of hydro-carbons, the effectiveness values were significantly increased When effective-ness values from past tests were recalculated using only the area where the plumewas known to be, those values decreased by factors as much as 2 to 5.33,63Although no applications of dispersants on freshwater spill have beenfound, one field test was carried out in fresh water.65 Effectiveness was notmeasured The ASTM standards on the use of dispersants in fresh water such aslakes and rivers suggest that they not be used in fresh water primarily becausemost lakes and rivers are used as sources of drinking water.66-68

In summary, testing in the field is difficult because effectiveness valuesdepend on establishing a mass balance between oil in the water column and onthe surface Because this mass balance is difficult to achieve, results arequestionable All tests relied heavily on developing a mass balance between oil

in the water column and that left on the surface In early tests, samples fromunder the oil plume were analyzed in a laboratory using colorimetric methods,

Rate, D: 0

Sea State

Effectiveness Claimed

% unless noted otherwise North Sea -

Great Britain

35 1997 136 Forties Crude - Weathered 50 Corexit 9500 airplane 1:19 3-4 good

135 Forties Crude - Weathered 50 Slickgone NS airplane 1:19 3-4 good

134 Alaska North Slope

-Weathered

North Sea- Norway 36 1995 133 Troll - Weathered 15 Corexit 9500 helicopter 1:20 1-2 good

132 Troll - Weathered 15 Corexit 9500 ship 1:20 1-2 good

131 Troll - Weathered 15 Corexit 9500 control control 1-2 North Sea -

123 Forties Crude continuous

North Sea -Norway 38 1994 117 Sture Blend Crude 20 Corexit 9500 helicopter 1:12 4-5 good

116 Sture Blend Crude 20 control then-Corexit

39 1993 115 MF/GO (Medium Fuel

Oil/ Gas Oil)

20 Dasic Slickgone NS airplane 1:10 2-3 good

North Sea - Great

Britain

40 1993 113 MF/GO (Medium Fuel

Oil/ Gas Oil)

continuous 50L/min

Rate, D: 0

Sea State

Effectiveness Claimed

50L/min

North Sea - Great

Britain

43 1992 109 Forties Crude 12.3 LA 1834 then Dasic

LTSW

airplane 1:100, 1:28 4-5 good

108 Forties Crude 12.3 LA 1834 airplane 1:100 4-5

Beaufort Sea - Canada 44 1986 106 Topped Federated Crude 2.5 control

105 Topped Federated Crude 2.5 Corexit CRX-8 helicopter 1:1 2-3 poor

104 Topped Federated Crude 2.5 BP MA700 helicopter 1:1 2-3 poor

103 Topped Federated Crude 2.5 BP MA700 helicopter 1:10 2-3

North Sea

-Haltenbanken

45 1985 101 Topped Statfjord Crude 12.5 Alcopol premixed 250 ppm 1-2

99 Topped Statfjord Crude 12.5 Finasol premixed, 3m below

surface

Brest, Protecmar VI 46 1985 97 Fuel Oil part of below Dispolene 355 ship-aerosol 1:9 1

94 Fuel Oil 5 control control 1

89 Statfjord 10 Corexit 9527 airplane, Islander 1:75 1

Rate, D: 0

Sea State

Effectiveness Claimed

Mediterranean

Protecmar V

68 Light Fuel 5 Dispolene 325 airplane, CL215 1:2.8 2

66 Light Fuel 5 Dispolene 325 airplane, CL215 1:2.4 3

North Sea, Britain 49 1982 64 Arabian 20 Corexit 9527 airplane, Islander 1:4 1

58 Statfjord 0.2 control control 2-3 2.6

Mediterranean,

Protecmar III

50 Light Fuel 6.5 Dispolene 325 airplane, CL215 1:3 1-2 50 Mediterranean,

Protecmar II

42 Light Fuel 1-5.5 BPIIOOX ship, helicopter, 1-3 Mediterranean,

Protecmar I

47 1979 28-41 Light Fuel 3 each BPIIOOX ship, helicopter, 1-3

(Continued )

Rate, D: 0

Sea State

Effectiveness Claimed

Long Beach, USA 58 1979 27 Prudhoe Bay 1.6 2% Corexit 9527 ship 1:11 2-3 62

Victoria, BC, Canada 59 1978 21 Prudhoe Bay 1.6 2% Corexit 9527 ship 1:67 2-3 5

Southern California,

USA

15 North Slope 0.6 Several,

demonstration

Several, demonstration 1-2

13 North Slope 0.8 Corexit 9527 ship >1:5 1-2

12 North Slope 3.2 Corexit 9527 airplane, Cessna > 1:5 1-2

10 North Slope 0.8 BPI IOOWD ship, WSL > 1:5 0-1

9 North Slope 1.7 Recovery þ helicopter > 1:5 0-1

8 North Slope 3.2 Corexit 9527 airplane, Cessna > 1:5 0-1

7 North Slope 1.7 Control later Corexit

9527

control then helicopter > 1:5 0-1

Wallops Island, USA 60 1978 6 La Rosa 1.7 Corexit 9527 helicopter 1:11 1 50

which are notoriously inaccurate Fluorometry has recently been used, but thismethod is also unreliable because it measures only a small and varying portion

of the oil (middle aromatics) and does not discriminate between dissolvedcomponents and oil that actually dispersed There is further discussion onanalytical techniques in a later section of this report The points raised inSection 15.4: Tank Tests are valid for field tests as well

In summary, testing in the field is fraught with measurement difficulties;however most of the past tests showed poor effectiveness, and the overallaverage of those that assigned values was 16%

15.4.3 Laboratory Tests

Many different types of procedures and apparatus for testing dispersants aredescribed in the literature Fifty different tests or procedures are described in onepaper.69Only a handful of these are now used, however, including the Labofina,Warren Springs, or rotating flask test; the swirling flask test; and the baffled flasktest Most of these procedures are used only on occasion for special studies.Some are used for regulatory purposes to screen dispersants for effectivenessprior to national approval Some common tests are listed inTable 15.5.Several investigators have reported results of apparatus comparison testsconducted in early years.70-76 In the several papers reviewed, all authorsconcluded that the results of the different tests do not correlate well, but someconclude that some of the rankings are preserved in different tests Generally,the more different types of oil tested, the less the results correlate It has beenshown that laboratory tests can be designed to give a comparable value of oildispersion if the parameters of turbulent energy, oil-to-water ratio, and settlingtime are set at similar valuesdbut most importantly if correct analyticalprocedures are applied.73 In the literature, different protocols are sometimesdescribed for the same apparatus The testing protocol used can sometimeschange the data more than the actual physical test

Fingas measured, calculated, or estimated energy and work in severallaboratory vessels and compared to estimates of energy/work at sea.77-79Somemeasurements completed by particle image velocimetry (PIV) and anemometrywere compared to these calculated values The initial measurements and esti-mates indicate that the energy in several laboratory vessels is similar and that itmay be equivalent to those encountered at sea under moderate wind and waveconditions Two techniques have been initiated to measure energy Themeasurement technique chosen to do this is PIV In this method, seed parti-clesdwhich could be oil dropletsdare put into the fluid and the fluid is illu-minated with a laser The movement of a particle in a given cell is measured as

a function of time This can occur as fast as 30 to 200 Hz, depending on theapparatus Turbulent energy can be calculated at each point in the image frame.The other method used is hot wire anemometry Although this method can yielddata similar to PIV, it requires the intrusion of a probe into the area The

Test Name Alternate Names (s) Energy Source Water Volume (L) Prime Use Where Used

general

ASTM, Canada USA, others

LABOFINA Warren Springs

at present Cascading Weir Flume fall over weir constant flow

Wave-Plate Tank South African

methods are compared in several laboratory vessels under several energyconditions.

Kaku and Boufadel have conducted similar measurements in some ratory apparatuses.80,81The laboratory apparatuses compared are the swirlingflask and the baffled flask

labo-Some of these laboratory data were compared to the field data by Lunel andcoworkers, and the results are shown in Table 15.6.1,75,76,82While the datacorrelate somewhat to the field data, with a wide spread in effectivenessnumbers and the few data points, this correlation should not be overstated.Another interesting point is that the effectiveness values obtained in the fieldare lower than the data obtained in the laboratory, indicating that the energylevels may be much higher in laboratory tests than those in the field conditionsdescribed here Furthermore, all of the laboratory tests yield effectivenessvalues much higher than field values This is contrary to what was thought inprevious years

The results of a number of dispersant effectiveness tests taken from lished laboratory results were compared in one study.8The correlation amongtests varies from high to low This may be due to errors associated with themeasurement, such as errors in measurement of volumes and variances inenergy of the apparatus It was also noted that the ranking of effectiveness isgenerally consistent; that is, those oils and dispersants that show the highest orlowest effectiveness do so in all tests

pub-A lot of work has been done recently on the new Environmental ProtectionAgency (EPA) test entitled the baffled flask 83-86 This apparatus has been

TABLE 15.6 Intercomparison of Laboratory and Field Effectiveness Results

Effectiveness Results in Percent

Oil Type Dispersant

Field Test [16e17]

SF GC

SF

CA IFP

WSL Lab 1

WSL Lab 2 Exdet

Medium Fuel Oil Corexit 9527 26 54 50 91 42 42 67 Medium Fuel Oil Slickgone NS 17 49 46 94 29 23 50 Medium Fuel Oil LA 1834/Sur 4 2 2 50 16 11 38 Forties Crude Slickgone NS 16 47 65 95 28 25 60 Forties Crude LA 1834/Sur 5 2 2 61 15 12 53 Correlation with field test (R 2 ) 0.89 0.7 0.54 0.87 0.94 0.41 Ratio Lab test/field test 0.4 0.35 0.19 0.56 0.62 0.27 Legend: SF ¼ Swirling Flask, GC ¼ analysis by Gas Chromatography, CA ¼ Colorimetric Analysis, IFP ¼ French Institute for Petroleum test, WSL ¼ Warren Springs Laboratory Test, EXDET e an Exxon test.

studied extensively including energy studies, variation with temperature,salinity, and operational parameters This test is a high-energy test and uses anobsolescent colorimetric analytical method More description of this will begiven in an analytical section below.

Reviews of testing have noted several conclusions on laboratory testing.1,8Bench-scale testing is widely used to evaluate the performance of dispersantsand the physical and chemical mechanisms of oil dispersion A major disad-vantage is, of course, that it is difficult to scale the results of these tests topredict performance in the field Several factors that are difficult to extrapolateinclude energy regimes, dilution due to horizontal and vertical advection, andturbulent diffusion Bench-scale tests are very useful for determining theeffectiveness of various dispersanteoil combinations, salinity, temperatureeffects, effects of oil composition, and effects of oil weathering It has also beennoted that the conditions for some tests are not realistic and result in unreal-istically high effectiveness values Many operators of these tests are not usingvalid analytical techniques

15.4.4 Tank Tests

There has been high interest in tank tests recently The U.S National Academy

of Sciences focused much attention on tank testing in its recent report.1It notesthat the physical characteristics of wave tanks imply that the encounter prob-ability of the dispersant with the oil slick will be higher than can be achievedduring a real spill response Thus, wave-tank tests provide upper limits onoperational effectiveness There is concern that wave-tank tests may also notcount for the skinning of oil that often occurs with weathering Skinning occurswhen heavier components of the oil, typically resins, float to the top and coatthe rest of the oil so that penetration from either side is slowed

Another concern is that the dispersant application system should simulatethe droplet-size distributions and impact velocities in real application systems.The wave energies used in tanks should be scalable to actual sea states It is alsonoted that coalescence and resurfacing of dispersed oil droplets occur and thatwave-tank experiments should include investigation of these phenomena Insummary, the advantage of wave tanks is to investigate operational effective-ness components and observe diffusion of droplets more like those at sea Thedispersant droplet size generation in tanks may be an important factor Thecommittee feels that the measurement of effectiveness should also includethe measurement of dispersed oil droplet size Further, factors such as massbalance and analytical methods require careful consideration.1,87

The EPA and the Canadian Department of Fisheries constructed a new testtank at Bedford Nova Scotia Extensive calibration, wave, and energymeasurement were carried out at this facility.88-92 In recent times, moreextensive testing has been carried out, and quantitative relationships havedeveloped between energy dissipation rate and dispersion.90 Recent testing

showed effectiveness amounts ranging from 53 to 95%, depending on tions and dispersants In a related test, effectiveness was found to be 21 to 36%under regular waves and 42 to 62% under breaking waves.91This is very similar

condi-to the findings at other well-controlled facilities such as at the Texas A&Mfacility and the Imperial Oil facility The energy factor under breaking waves inthe Bedford Tank created much smaller droplet sizes as would be expected In

a related study, a model was created for droplet sizes produced under breakingwaves based on tank test results.92

OHMSETT, a government-run test tank in New Jersey, has continued work

on dispersant tank testing.93-100 The facility is still working on the mendations for analytical methods

recom-Tank testing began with the Imperial test tank facility in Calgary, Alberta.101Workers at this facility were pioneers in measuring mass balance, energy, andother essential parameters to tank dispersant testing The range of effectivenessachieved was generally from 20 to 60% in a few hours and 10 to 30% over oneday Texas A&M followed up with a new test facility near Corpus Christi.102This facility advanced the art of measuring mass balance, energy, and other tanktesting parameters The effectiveness they measured was similar to that of theImperial tank crew

The author and coworkers prepared extensive studies on tank testing.103-106The following presents 17 critical factors that need to be considered andincluded in any test for measuring the effectiveness of dispersants in a tank or inthe field in order for that test to be valid

1 Mass balance

The measurement of effectiveness should include the determination of

accounted-for mass balances typically vary from 50 to 75% It is mended that mass balance should be calculated in all wave-tank studies

that the strips could be later removed and extracted for oil A water-surfaceoil slick quantification protocol was developed, using solid-phase extractiondisks Initially the group was able to account for only 10 to 33% of the oiloriginally placed in the tank After considerable effort, the mass balance wasimproved to about 50 to 75% This illustrates the problems of attaining

a mass balance

Mass balance is very difficult to achieve in large test tanks, especially infull-scale field tests Brown et al reported on tank tests of dispersant effec-tiveness.101,110Effectiveness was measured in two ways: by accumulatingthe concentrations of oil in the water column by fluorometric measurementsand by removing and weighing oil on the surface The results of these twomeasurements, the amount of oil unaccounted for, and the differencebetween the two measurements, are shown inTable 15.7

These data show that between 0 and 68% of the oil in the tank can be counted for Furthermore, in two cases (2 and 3 in the table), the amount ofoil was overcalculated This shows the difficulty in attaining a mass balance,even in a confined test tank Brown et al observed that the problem wasaccentuated by the heterogeneities in oil concentration in the tank.101Some of the unaccounted oil may have been in regions where the concen-trations of oil were higher than average It should also be noted that surfaceremoval exaggerated the amount of oil dispersed from a factor of 1 to 8, with

3 Hour

Surface Removal

Percentage Unaccounted For

Percent Difference Between Methods

numbers are meaningless The above examples show that mass balance ses, even in the more controlled tank tests, can vary from a few percent andhigher If the measurement made does not account for the discrepancies inmass balance, then very high errors result A typical example of this is usingonly the oil remaining on the surface as an indicator of dispersant effective-ness In a very highly controlled test series, this number can be from 0 to67% greater than the oil actually dispersed (factors of 1 to 8 times theamount recovered).

los-The experiences of Bonner et al show that there are major losses of oil inthree areas that historically have not been considered in performing a massbalance.107-109 These are adhesion to walls, adhesion to sediments, andformation of invisible slicks These three losses can account for over 50%

of the oil loss in certain cases Methods for the measurement of each of theseoil losses were developed and applied The adhesion to the walls wasmeasured by placing strips of wall material into the test tank and laterremoving and quantifying the oil on these strips It was noted that the ageand conditions of these strips were important as the more weathered tankwould hold more oil than the newer, unweathered strips

Bonner and coworkers noted that the sediment (even that from the ently clean tank bottom) must be collected and oil content measured.107-

appar-109

Oil in thin slicks was measured using a solid-phase extraction diskheld by vacuum to retain both the disk and oil adhered to the disk

A question that must be dealt with is, as in the title of the Brown et al paper,

“where has all the oil gone?”110In summary, the mass balance problemsrevolve around analytical problems; loss of oil through thin, invisiblesheens; calculation difficulties; inability to recover surface oil after disper-sant applied; and loss to tank walls and also in the presence of large hetero-geneities in oil concentrations in the water column

2 Proper controls

A proper control experiment is needed in order to accurately assess

a dispersant tank test or a field trial The control slick must be treatedequally to the test slick in every respect except for the application of disper-sant The measurement of any factor of the dispersant slick should then becompared to the same measured factor of the control

3 Analytical methods

Few analytical methods can be used in field situations even in a test tank.Very early in the tank testing program, fluorometers, particularly Turnerfluorometers, were used In early years before Global Positioning Systems(GPS), it was difficult to assess the position at which samples were taken ifthe sampling probes were not fixed to the tank Now accurate GPS datacoupled directly to fluorometer data can provide reasonable positionaldata for the fluorometric readings

Some of the earlier trials used grab samples that were subsequently takenfor analysis by Ultraviolet (UV) or Infrared (IR) absorption.63,103These

methods are notoriously inaccurate and scientifically incorrect and havelong since been replaced by gas chromatography methods A furtherproblem is that of sample preservation Samples must be chilled immedi-ately and treated to prevent bacteria growth and hydrocarbon loss Standardprocedures are available, but in early trials these were not applied.

The use of fluorometry for oil measurement has been examined indetail.94-97 These studies show that fluorometry is a sensitive, but notnecessarily accurate, means of oil determination A fluorometer uses

UV or near UV to activate aromatic species in the oil The UV activationenergy is more sensitive to the naphthalenes and phenanthrenes, whereasthe near UV is more sensitive to large species such as fluorenes Thecomposition of the oil changes with respect to aromatic content as itweathers and is dispersed, with the concentration of aromatics increasing.Thus, the apparent fluorescent quantity increases in this process

Studies then showed that because the amount and distribution of PAHs,the target compound for fluorometers, change with time during the course

of a chemical dispersion event, a fluorometer can never be truly brated” for a particular oil and dispersant combination.111-114The compo-sition of the oil changes with respect to aromatic content as it weathersand is dispersed, with the concentration of aromatics increasing A fluo-rometer reading will always remain a relative value and even with carefulcalibration can only give indications that are as much an order of magni-tude from the true value Efforts continue on fluorescent measures, butthere needs to be more recognition that this method will always be relative

samples directly from the output of a fluorometer, there are significantdifferences because the target compounds of the PAHs have differentdistributions and concentrations with time, conditions, and weathering

of the oil.Figure 15.4shows the lack of relation between the total leum hydrocarbon (TPH) measured by Gas Chromatography-Flame Ioni-zation Detection (GC-FID) and the reading of the fluorometer Samplesfor analysis were taken directly from the output of the fluorometer, andreadings were recorded for comparison.Figure 15.4shows that the corre-lation between readings is poor

petro-The calibration of fluorometric readings is critical, but one should bear inmind that an exact reading is impossible, as pointed out earlier.113-116Themost important factor is how the oil is introduced to the fluorometer andthe subsequent readings made The physical factors that influence howmuch of the oil the fluorometer sees are the solubility and dispersibility

of the particular oil and the subsequent evaporation or volatilization ofthe oil A typical procedure is to add oil and dispersant to a container(e.g., a bucket) and then pump this through a flow-through fluorometer.Most often, that amount of oil added is taken as the amount of oil read

by the fluorometer The problem with this method is that most of the

oil is not dispersed into the water column and that some large amounts ofsoluble species are present, which would not be the case in the sea Tests

of these types of methods show that the fluorometer calibration curve isgenerally between 5 and 10 times greater than is the actual case Thus,

a reading of 15 ppm in the field is actually a reading of somewherebetween 1.5 and 3 ppm As this was generally the case in a few test-tank trials, the actual ppm readings provided are far too high and cannotsimply be converted into actual values

A somewhat better method of calibrating a fluorometer is to use weatheredoil (to about the percentage expected in the field) and introduce this to

a closed container After about 15 minutes of pumping, take a sample andanalyze it by a standard gas chromatography (GC) method.114Then the addi-tion continues, one increment at a time, with sampling and analysis at eachincrement After the numbers are collected, this will form a relatively goodcalibration curve But because of the differences in chemical composition,this calibration curve could also give results as high as twice that of actualconcentration

The most reliable method of calibrating a fluorometer is to perform the abovecalibration procedure, but repeat it throughout the actual experiment Almostsimultaneous samples are relatively easy to collect from the fluorometer asthe flow from the output of the fluorometer can be captured and preservedfor later analysis This is generally done when the fluorometer reading is

Fluorometer Reading (relative units)

relatively stable to ensure correspondence between the sample and the rometric value The actual values and the previously prepared calibrationcurve can be compared to examine the differences in composition It should

fluo-be noted that this method was followed in some of the past test-tank work.This is the case for the data illustrated inFigure 15.4

One must remember that, in fact, the aromatic concentrations in thedispersed plume are a composite of dispersed and dissolved aromatics andthis total concentration is changing with time

The effects of running probes into the water column have not been fullyexamined Several devices have been created in the past to examine thesubsurface water column; however the standby usually ends up beingweighted hose Tests show that there is significant retention on Tygon tubingand that pumping for up to one hour may be required to clear this line to thepoint of background measurements Teflon tubing appears to show a smallereffect, although less testing has been conducted on this There may be

a serious effect on measurements, depending on how the tubes or samplingdevices are deployed Some test-tank work used fixed probes with Teflontubing

Another complication to sampling is the retention of surface oil on thesampling tubes, weights, and pumps that are lowered into the water Asthe equipment goes through the surface slick, which is present over greaterareas than the in-water plume, some of the surface oil will be retained onthe sampling equipment and will be read as oil concentration at that depth.Although some experimenters have dragged the submerged sampling train

to the next sample point to avoid this problem, this action may also dragoil on the outside of the sampling gear

In summary, fluorometry is a technique sometimes used for measuringrelative concentrations of oil in the water column at test tanks It must benoted that fluorometers cannot truly be calibrated for the oil as there aremany variables, as explained earlier The errors encountered all increasethe apparent value of the oil concentration in the water column Incorrectcalibration procedures can distort concentration values up to 10 times theiractual value, or even more Correct analytical methods involve performingaccurate GC measurements both in the laboratory and in the field duringthe actual experiment Furthermore, water sampling gear must be deployed

in such a way as to avoid disturbing the underwater plume or carrying oilfrom one level or area to another Fixed probes and sampling after a period

of time may assist in minimizing this disturbance problem

4 Time lag and length of time plume is followed

There are certain time characteristics of the dispersion process that must beunderstood First, the time to visible action after the dispersant applicationvaries from 15 to 90 minutes Fast action is herding and not dispersion Thevisible action is generally taken as the appearance of a yellow-to-coffee-colored plume in the water The second item of timing to note is that the

dispersant may continue to act for up to an hour after application Third, themovement and dispersion of the plume are generally slow, although theplume is generally visible for about 3 hours and is never visible after about

8 hours Finally, the oil in the plume will resurface slowly over the nextseveral days

It is important to take measurements as long as possible The Beaufort Seaexperiment is a good example Three slicks were laid and two left ascontrols.44 Two days later, three slicks were found at sea, and each hadthe same orientation and geometry as one on the first day of the experiment.The largest slick, in area, was the dispersed slick, although the oil contentwas not known The interpretation of the results would have been quitedifferent if the slick had not been followed for 2 days

Brown et al noted that they had to measure their test tank after 24 hours toyield a reasonable result Measurements before about 6 hours were found

to be of little value.110

In summary, as the oil concentration in the water is constantly changing,any value should be expressed as a function of time and place Further,since the dynamics of the dispersion process change rapidly in the firstfew hours, values should only be used after about 6 hours

5 Mathematics of calculation and integration

Several examples of the effects of integrating and averaging incorrectly aregiven in a past paper.63This effect is exacerbated if nonzero oil concentra-tion values are measured in areas outside of the plume These errors areillustrated inFigure 15.5 Methods of integration and handling backgroundvalues can easily change the values by as much as threefold, even for theidentical case.33

Another concern regarding the mathematics is that related to the use offixed water column concentration values to determine effectiveness.Although this method has not been used in recent years, it was thought to

be a reliable means of estimating effectiveness The assumption that ismade is that the slick is evenly distributed in 1, 2, or 3 meters Then oncethe concentration is measured, an effectiveness is assigned.Table 15.8illus-trates the variances in using this type of scheme This shows that oneconcentration could yield a wide range of effectiveness values depending

on what assumption one makes Because one cannot make oil thicknessmeasurements and because the depth of mixing is not simply a fixed depth,this type of procedure is not a valid method for determining effectiveness

6 Lower and upper limits of analytical methods

The lower and upper limits of the analytical methods applied are anotherimportant factor, especially in field situations If the measurement is lessthan the lower limit, use of these values can result in serious errors, asshown above The lower analytical limit should be taken as twice the stan-dard deviation, or about 0.3 ppm for an older fluorometer or about 0.1 ppmfor a newer unit Use of double the standard deviation is standard