United States Environmental Protection Agency Risk Reduction Engineering Laboratory Cincinnati, OH 45268 Research and Development EPA/600/S-92/065 October 1992 EPA Project Summary Chemical Oil Spill Dispersants: Update State-of-the-Art on Mechanism of Action and Laboratory Testing for Performance John R. Clayton, Jr., James R. Payne, Siu-Fai Tsang, Victoria Frank, Paul Marsden, and John Harrington Chemical dispersants are formula- tions designed to facilitate dispersion of an oil slick into small droplets that disperse to non-problematic concen- trations in an underlying water column. This project had two primary objectives: (1) update information on mechanisms of action of dispersants and factors affecting their performance and (2) evaluate selected testing procedures in the laboratory for estimating perfor- mance of different dispersant agents. The first objective resulted in a report updating information on chemical dis- persants, their mode of action, variables affecting dispersant performance in the field as well as the laboratory, and a discussion of a number of laboratory and rapid-screen field tests for esti- mating performance, information de- rived in the course of preparing this report was used to select three labora- tory testing procedures for evaluation of performance characteristics: the Revised Standard EPA test, the Swirling Flask test, and the IFP-Dilution test, in the laboratory, these three procedures were evaluated for their precision of results in estimating dispersant perfor- mance, costs associated with conduct- ing a given procedure, and the ease of conducting that procedure (e.g., num- ber of tests performed in 8 hr, skill level required of an operator, and overall complexity of the procedure). The precision of results for dispersion performance for all procedures was approximately the same (standard de- viation of 7% to 9% in dispersion per- formance values). Costs to perform a procedure and ease of conducting that procedure favored the Swirling Flask test. This Project Summary was developed by EPA's Risk Reduction Engineering Laboratory, Cincinnati, OH, to announce key findings of the research project that is fully documented in the two re- ports listed at the end of this summary (see Project Report ordering informa- tion at back). Introduction All tasks performed for this work assign- ment were elements of the Oil Spills Re- search Program that was initiated in re- sponse to the Oil Pollution Act (OPA) of 1990. The work supports the EPA workgroup concerned with revision of sub-part J (dis- persant effectiveness and toxicity) of the National Contingency Plan (NCP) as re- quired by the OPA. Primary deliverables from SAIC to EPA's Releases Control Branch/Risk Reduction Engineering Labo- ratory (RCB/RREL) include (1) a State-of- the-Art (SOTA) report on mechanisms of action and factors influencing dispersant performance and (2) a laboratory evalua- tion of candidate National Contingency Plan protocols for testing dispersant perfor- mance of candidate agents. The RGB may use information derived from the SOTA report as well as the laboratory studies as part of the work assignment to assist in evaluation of candidate tests for estimat- ing performance of dispersant agents as Printed on Recycled Paper ------- well as planning follow-on studies with chemical dispersants. State-of-the-Art Report on Chemical Dispersants The SOTA report for chemical disper- sants includes discussions of the follow- ing topics: (1) the mechanism of action of chemical dispersants for oil spills, (2) fac- tors affecting the performance of disper- sants, (3) some common laboratory methods used to measure dispersant per- formance, (4) aspects of the analytical measurement of dispersant performance in the laboratory, (5) a brief summary of dispersant applications and their perfor- mance in field trials and spills-of-opportu- nity, and (6) recommendations for future laboratory studies. The discussion of the laboratory methods includes detailed in- formation for a number of the more com- monly used tests, as well as similarities and differences among testing procedures. Differences among tests are particularly important because they may be respon- sible for not only significant differences in results between laboratory testing methods but also poor correlations between labo- ratory results and data from field tests. Four general types of laboratory testing methods are considered: (a) tank tests, (b) shake/flask tests, (c) interfacial surface tension tests, and (d) flume tests. Infor- mation is presented for general ap- proaches used in laboratory studies, limi- tations inherent to the laboratory mea- surements, and the relevance of laboratory results to field studies or situations. Brief descriptions also are presented for a number of rapid field tests for estimating dispersant performance. Limitations in- herent to measurements obtained with the latter tests are discussed. General Mechanism of Action of Chemical Dispersants Chemical dispersants are designed to promote the break-up or dispersion of an oil slick into small droplets that distribute into a water column. The small oil droplets should not recombine or coalesce to reform surface slicks. Ideally, dispersed oil drop- lets will be subject to not only dilution to non-problematic concentrations in the water column but also enhanced microbial degradation as the oil-water interface in- creases. A typical commercial chemical disper- sant is a mixture of three types of chemi- cals: surface active agents (i.e., surfac- tants), solvents, and additives. Solvents are primarily present to promote the dis- solution of surfactants and additives into a homogeneous dispersant mixture. Addi- tives may be present for a number of purposes such as increasing the biode- gradability of dispersed oil mixtures, im- proving the dissolution of the dispersant into an oil slick, and increasing the long- term stability of the dispersion. For the actual dispersion process, however, the most important components in the disper- sant mixture are the surfactant molecules. These are compounds containing Irath oil- compatible (i.e., lippphilic or hydrophobic) and water-compatible (i.e., hydrophilic) groups. Because of this amphiphatic nature (i.e., opposing solubility tendencies), a surfactant molecule will reside at oil-water interfaces with the hydrophobic and hy- drophilic groups positioned toward the oil and water phases, respectively. As such, the surfactant will reduce the oil-water in- terfacial surface tension. The lowering of oil-water interfacial surface tension will promote dispersion of oil droplets into the underlying water with minimal mixing en- ergy. The oil droplets will remain dispersed in a water column if they are small enough to allow for natural water currents or Brownian motion to prevent rising to reform surface slicks. Factors Affecting Chemical Dispersion of Oil and Its Measurement A variety of factors have major influ- ences on the ability of chemical agents to disperse oil into water in both laboratory tests as well as actual field situations. These factors can include physical and chemical properties of an oil, the compo- sition of a dispersant formulation, the method of applying the dispersant to an oil slick, the mixing energy available for dispersing treated oil into a water column, the dispersant-to-oil ratio, the oil-to-water ratio, temperature, and salinity. Trie sam- pling and analysis methods for evaluating dispersion performance also can influence measurement results. Crude and refined petroleum products are complex mixtures of hydrocarbon compounds that can contain compounds in five broad categories: lower molecular weight (1) aliphatics and (2) arcimatics, and higher molecular weight (3) asphaltenes, (4) resins, and (5) waxes. Interactions between the aliphatics, aro- matics, asphaltenes, resins, and waxes allow for all of the compounds to be maintained in a liquid-oil state. That is, the lower molecular weight components (i.e., the aliphatics and aromatics) act as solvents for the less soluble, higher mo- lecular weight components (i.e., the asphaltenes, resins, and waxes). In addi- tion to inherent differences in chemical compositions among different parent oils, oil that is released onto a water surface will undergo a variety of rapid, dynamic changes in both chemical composition and physical properties. Such changes are known as weathering and result from se- lective dissolution and evaporation losses of lower molecular weight components as well as photooxidation and microbial deg- radation of selective compounds. Com- plex crude oil mixtures remain as relatively stable liquid phases as long as the sol- vency interactions occurring in the bulk oil are maintained and thermodynamic con- ditions remain constant. If this equilibrium state is changed (e.g., due to weathering processes), the solvency strength of the oil may become insufficient to keep higher molecular weight components in solution and lead to their precipitation as solid particles. Accompanying changes in the physical state and chemical properties of the oil can affect the way chemical dis- persants interact with the oil that has un- dergone such changes. Despite the pre- ceding complexities associated with dif- ferent oils, dispersant formulations are frequently designed with the intent to deal with a relatively broad range of oil types and properties. The dispersant application method can be one of the most critical elements de- termining whether a particular dispersant will produce dispersion of oil or not. In field situations, dispersant is normally ap- plied from aircraft (airplanes or helicopters) or surface vessels (boats). The dispersant is applied as relatively small droplets that descend onto a slick in a manner providing broad spatial coverage. The size of the applied dispersant droplets is important to successful application. Droplets that are too large may penetrate through an oil slick without interaction. Droplets that are too small may not reach the slick because of air or wind transport between the appli- cation source and the slick. Following application of a chemical dis- persant to an oil slick on water, dispersion of the oil requires input of mixing energy that results in injection of the oil as droplets into the underlying water column. The mixing energy is generally supplied by ambient wave action in field situations or mechanical agitation of test solutions in laboratory systems. Dispersion of the droplets into the water column is countered by the buoyancy of the oil droplets, which depends on the density and size of the droplets, their rise velocities as described by Stokes' Law, and natural advective processes that result in horizontal and vertical transport and dilution of the oil. In addition to mixing energy, other factors ------- that can affect dispersion performance of an oil include the ratios of dispersant-to- oil and oil-to-water as well as ambient temperature and salinity in the water. Evaluation of dispersant performance in the laboratory (as well as the field) must incorporate appropriate sampling and analysis methods into a testing procedure. In the numerous studies that compare dis- persion-performance values among test- ing procedures, agreement for results is generally poor. At least a portion of this variability is attributable to variations in the sampling approaches. For example, different laboratory testing procedures will collect test samples from reaction vessels at various times after agitation in the re- action vessel ends. A settling time (i.e., collection of samples only after agitation has stopped for some predefined period of time) may be incorporated into a testing procedure to allow larger, less stable dis- persed oil droplets to return to a surface slick before smaller, more stable dispersed oil droplets are recovered in a subsurface water sample. In addition to the preceding issues re- lated to sampling methodology, the ana- lytical methods chosen to quantify amounts of dispersed oil in samples also are im- portant for dispersion measurements. The most widely used methods for quantifying amounts of dispersed oil in laboratory test samples involve extraction of a water sample with a suitable solvent and quan- titation by UV-visible spectrophotometric or (less frequently) gas chromatographic methods. However, selection of the ana- lytical wavelength(s) for spectrophotomet- ric measurements can be important. Mea- surements in different laboratory testing procedures have been made at wave- lengths from 340 to 620 nmeters, with wavelengths selected in part on the optical (or color) characteristics of particular oils and dispersants being tested as well as the optical characteristics of the available spectrophotometric system. Laboratory Tests for Dispersant Performance A variety of laboratory testing methods have been used to evaluate dispersant performance. In general, laboratory tests can be placed into four categories: (1) tank tests with water volumes ranging from 1 to 150 L, (2) shake/flask tests that are conducted on a relatively smaller scale and require less sophisticated laboratory equipment, (3) interfacial surface tension tests that measure properties of the treated oil instead of djspersant performance di- rectly, and (4) flume tests using flowing water systems with the capacity for break- ing/nonbreaking waves to generate en- ergy regimes that can more closely simu- late real-world conditions in large water bodies (e.g., oceans and coastal bays). Each type of test uses a general approach of (1) establishing an oil slick on water, (2) applying dispersant to the slick, (3) apply- ing energy to the oil-dispersant-water sys- tem, and (4) measuring the amount of oil dispersed into the water. Significant differences are inherent to the various methods. For example, differ- ent methods for adding the dispersant to oil include premixing of dispersant with oil, slowly pouring dispersant onto the oil, spraying the oil surface with a fine mist of either neat dispersant or dispersant premixed with seawater, or pouring the dispersant into the water before adding the oil. Test-specific variations in the ratio of the oil-to-water volumes can affect not only the relative performance of dispersant surfactants (e.g., hydrophilic versus li- pophilic) but also the magnitude of wall- effects in test containers. A variety of ap- proaches have been used to provide mix- ing energy to test systems, such as circu- lating pumps and spray hose systems, high velocity air streams that produce small waves on the water's surface, raising and lowering of a metal hoop in the water, rotating or shaking separatory funnels, shaking flasks on a shaker table, and vertically flowing water in a test cylinder. Another variation concerns the time after mixing ends that water samples are with- drawn from the test solutions in the differ- ent procedures. In summary, the wide va- riety of test conditions can make compari- son of results among different methods quite problematic. Detailed descriptions of the following procedures for evaluating performance of chemical dispersants are presented in the SOTA report. 1) Tank tests: Mackay/Nadeau/Steelman (MNS) test, Revised Standard EPA test, oscillating hoop test, IFP-Dilu- tion test, and flowing cylinder test. 2) Shake/Flask tests: rotating flask test (Labofina/Warren Spring Laboratory), swirling flask test, and Exxon disper- sant effectiveness test (EXDET). 3) Interfacial Surface Tension tests: drop- weight test. 4) Flume tests: cascading weir test and Delft Hydraulics flume test. Table 1 summarizes features and es- sential procedural components in these testing methods. Strengths and limitations associated with each testing method are presented in the SOTA report. In addition to the preceding methods, descriptions also are presented in the SOTA report for five rapid-screen field tests for evaluating dispersant performance (the EPA field dispersant effectiveness test, the API field dispersant effectiveness test, the Mackay simple field test, the Pelletier screen test, and the Fina spill test kit*). The detailed laboratory test methods identified in Table 1 do not readily lend themselves to onsite applications in the field. In contrast, the rapid-screen field tests have been developed to provide fast, qualitative information regarding an oil's dispersibility in field situations. These rapid- screen tests are, however, inherently lim- ited in the scope of information that they can provide because of their necessary simplicity for use in field situations. Evaluation of Laboratory Test Procedures to Assess Dispersant Performance Primary objectives in evaluations of laboratory testing procedures were to ob- tain estimates of the repeatability of mea- surements for dispersion performance with different testing procedures, evaluate comparability of results obtained with the procedures for selected dispersant agents and oils, and summarize the qualitative ease of conducting each testing procedure (i.e., how many individual test runs can be performed in a given period of time, the complexity of a testing procedure in relation to the required training time and skill level of an operator, and associated costs for both equipment and conduct of tests). All of these objectives have relevance to the suitability of a testing procedure for use as a routine laboratory testing method. Additional advantage could derive from identifying one or more testing procedures that could be performed in a more rapid and efficient manner than the current Re- vised Standard EPA protocol and that could provide results giving dispersion- performance rankings for different disper- sant agents equivalent to those obtained with testing procedures used by agencies in other countries (e.g., Canada and countries of Western Europe). Selection and Experimental Design for Test Procedures Tests selected for evaluation in the labo- ratory included the currently accepted Re- vised Standard EPA test, Environment Canada's Swirling Flask test (including three versions: premixed, 1-drop, and 2- drop dispersant additions), and the IFP- * Mention of trade names or commercial products does not constitute endorsement or recommendation for use. ------- Table 1. Summary of Features of Laboratory Methods to Test Dispersant Performance Water Energy Energy Volume Test ID Source Rating1 (mL) OWRt MNS Revised Std. EPA Oscillating hoop IFP-dlliition\ Flowing cylinder Labotlna rotating flask Swirling flask EXDET Drop-weight Cascading wolf Delft flume High velocity 3 air stream Pump 3 Oscillating 3 hoop Oscillating 1-2 hoop Vertical flow 1 ofwator Rotating 3 vessel Shaker table 1-2 Wrist-action 1-2 shaker None 0 Water 2-3 passing over weirs Wave paddle 2-4 6000 1:600 130,000 1:1300 35,000 1:175 4000-5000 1:1000, then decrease 1000 1:1200, than decrease 250 1:50 120 1:1200 250 Variable (NA)» (NA) 150,000 Variable 4,500,000 Variable Dispersant Application Method Dropwise/ premix Dropwise Dropwise/ premix Dropwise Pramtx Dropwlso Premix/ dropwise Premix/ dropwise Water-oil interaction Spray Spray DOR* Variable 3:100 to 1:4 Variable Variable 1:25 1:25 1:10 to 1:25 Variable (NA) Variable Variable Settling Time (min) None None None None 10 1 10 None (NA) None None Complexity Rating"* 3 3 3 2 2 1 1 1 2 4 4 'Energy Rating: 0=none; 4=highest. tOWR* oil-to-water ratio (v:v). * DOR s dispersant-to-oil ratio (v:v). "Complexity Rating: 1* lowest; 4 = highest. H (NA) * not applicable. Dilution test (Centre de Documentation de RecherchS et d'Experimentations sur les Pollutions Accidentelles des Eaux, Plouzane, France). Test oils used in some or all portions of the laboratory study in- cluded Prudhoe Bay crude, South Louisi- ana crude, Alberta Sweet Mixed Blend (ASMS), Arabian crude, Bunker C, and No. 2 fuel oil. Test dispersants used in all portions of the study included Corexit 9527, Coraxit CRX-8, and Enersperse 700. Com- mon elements through tests with all of the testing procedures included the following. • oil types: Prudhoe Bay and South Louisiana crudes • dispersant types: Corexit 9527, Corexit CRX-8, Enersperse 700, and "no dispersant" controls • test types: EPA-10 min, EPA-2 hr, premixed Swirling Flask, and IFP-Di- lution • analytical wavelengths: 340, 370, and 400 nmeter absorbance • duplicate measurements for particu- lar groups • water temperature (not a specified variable of interest for these studies, but one that did exhibit slight varia- tions) Results of Laboratory Tests Information for the primary objectives of the laboratory study are summarized in Table 2. Estimates of precision or repeat- ability for dispersion-performance values (i.e., standard deviations about means) were approximately 7% to 9% for all of the testing procedures. These values should be viewed as preliminary estimates, however, because they are generated with only a limited number of oils and disper- sant agents. Furthermore, final estimates for precision associated with a given test- ing procedure should incorporate measure- ments from multiple laboratories. The nonquantitative criteria in Table 2 (i.e., number of tests that can be performed in 8 hrs; costs associated with equipment acquisition, conduct of tests, and waste disposal; and qualitative items such as necessary skill level of an operator and overall complexity of a testing procedure) favor the Swirling Flask procedure for conducting multiple tests in a relatively short period of time for the least amount of cost. General trends for the results of disper- sion performance in the laboratory tests are illustrated in Figure 1 for the two com- mon test oils (Prudhoe Bay and South Louisiana crudes) with the four primary testing procedures (Revised Standard EPA-10 min, Revised Standard EPA-2 hr, ------- premixed Swirling Flask, and IFP-Dilution) and the three dispersants (Corexit 9527, Corexit CRX-8, and Enersperse 700). Dis- persion for- all test oils in all procedures was near zero in the absence of chemical dispersant agents. With addition of the dispersants, performance values were con- sistently highest with the EPA test. Simi- larities in relative trends for dispersion per- formance with the different dispersants were observed in the Revised Standard EPA and Swirling Flask procedures for Prudhoe Bay crude. Trends for dispersion performance with the different dispersants and South Louisiana crude were less com- parable In the EPA and Swirling Flask procedures. Relative performance trends among dispersants In the IFP-Dllutlon pro- cedure did not appear to be comparable to either the EPA or Swirling Flask tests. Statistical analyses of results showed that the major portions of differences in disper- sion-performance values for the three pri- mary testing procedures (Revised Stan- dard EPA, premixed Swirling Flask, and IFP-Dilution) could be accounted for by -differences in the oils and dispersants used in tests (i.e., at least 85% of the total variance for results with each testing pro- cedure). Dispersion-performance values for measurements at the three analytical wavelengths (340, 370, and 400 nmeters) were negligible (i.e., <0.5% of the total variance). Recommendations for Laboratory Studies Chemical dispersion of oil into water in laboratory studies involves complex inter- actions between many variables including the chemical and physical properties of oils and dispersants, the method of appli- cation (and its effectiveness) and mixing of a dispersant with an oil, the source and magnitude of mixing energy available to the system, the dispersant-to-oil ratio, the oil-to-water ratio, temperature, and the sa- linity of the aqueous medium. Extrapola- tion of results from dispersion performance studies in a laboratory to field situations must take into account additional compli- cating variables including rapid changes that occur in properties of the oil with time (i.e., natural weathering), field application methods and logistics, ambient weather and meteorological conditions, and local sea-state or oceanographic conditions (e.g., wave heights, currents, turbulent mix- ing regimes, etc.). The breadth of these variables makes It unlikely that any single laboratory test will ever be completely suit- able to quantify performance of chemical dispersant agents for all possible environ- mental scenarios. More realistically, many laboratory test results should be used to apply relative rankings to performance by various dispersant agents, including pos- sible assignment of "pass/fail" status to individual dispersants. Scientific criteria to assign "pass/fail" status continue to be subjects for future discussion and study. Much has been learned about perfor- mance of dispersants and their mecha- nisms of action from studies conducted with the variety of laboratory testing methods shown in Table 1. Test results are, however, frequently contradictory for reasons that are probably related to test- specific characteristics. Adoption of stan- dard experimental protocols (e.g., selection of specific reference oils and dispersants, consideration of the weathered state of test oils, use of specific oil-to-water ratios, selection or not of designated settling-times to be used in the conduct of experiments, and consideration of the natural dispersibilities of oils in a given test) might lead to closer agreement in performance results among testing methods. Further advances in testing methodologies, how- ever, remain to be developed and refined, particularly as they relate to the environ- mental relevance and performance of dis- persant agents. For example, approaches used to generate environmentally relevant mixing energies in laboratory studies could be improved. Continued investigation and analysis of dispersed oil droplet sizes might explain differences in energy levels and estimates of dispersion performance in dif- ferent laboratory testing systems, which could lead to improved, standardized test designs. In general, laboratory experiments also have not been designed to evaluate the effects of herding of oil on dispersion results. Current testing methodologies are Inadequate to Investigate dispersion In thin versus thick slicks, which Is Important for dispersant applications at sea because slicks are usually nonuniform In thickness and distribution on the water's surface. At the same time, it Is highly desirable that a laboratory testing method be simple, re- quire equipment that is relatively easy to acquire and fabricate, require a minimum of operator training and sophistication, and allow for the conduct of a reasonably large number of tests yielding quantifiable results in an acceptably short period of time. From the standpoint of using chemical dispersants for mitigating effects of oil spills in real-world situations, development and refinement of application techniques and protocols for applying dispersants in the field remain as critical needs. Successful application of chemical dispersants in field situations continues to be extremely prob- lematic. Further studies in areas of appli- cation technologies are definitely war- ranted. All reports for the work assignment were submitted in fulfillment of Contract No. 68- C8-0062 by Science Applications Interna- tional Corporation under the sponsorship of the U.S. Environmental Protection Agency. Table 2. Results of Test Procedures Used to Evaluate Performance of Chemical Dispersant Agents Test Procedure Revised Standard EPA - 10 min Revised Standard EPA -2hr Swirling Flask (Premixed) - (2 oils) Swirling Flask {Premixed) - (4 oils) IFP-Dilution -^ Standard Deviation for Performance 8.8%* 7.2% 7.8% 8.1% 7.2% No. Tests/8 hr 2 2 24-36 24-36 4-5 Equip. Cost ($) 2,280 2,280 1,225 1,225 3,160 Cost Run ($) 600 600 22 22 202 Complexity of Procedure High High Low Low Moderate Operator Skill Level Moderate Moderate Moderate Moderate Moderate * Bold values for standard deviations are estimates because variances among groups are heterogeneous by Bartlett's test for homogeneity. ------- Prudhoe Bay Crude Mean DIspersant Performance (%) EPA - 10 mln EPA - 2 hr Sw. Flask-premix IFP-dilution Test ID South Louisiana Crude Mean DIspersant Performance (%) C9527 EN700 EPA - 10 mln EPA - 2 hr Sw. Flask-premix IFF'-dilution Test ID Figure 1. DIspersant performance for four testing procedures with two oils and three dispersants. Values are means from replicate measurements. •U.S. Government Printing Office: 1992— 648-080/60139 ------- ------- John R. Clayton, Jr., Siu-Fai Tsang, Victoria Frank, Paul Marsden, and John Harrington are with Science Applications International Corporation, San Diego, CA 92121. James R. Payne is with Sound Environmental Services, Inc., Carlsbad, CA 92008 ChoudhrySarwaris the EPA Project Officer (see below). Completed reports produced in the project are the following: (1) "Chemical Oil Spill Dispersants: Update State-of-the-Art on Mechanisms of Action and Factors Influencing Performance with Emphasis on Laboratory Studies. Final Report," (OrderNo. PB92-222 207/AS; Cost: $19.00, subject to change) (2) " Chemical Oil Spill Dispersants: Evaluation of Three Laboratory Procedures for Estimating Performance. Final Report," (Order No. PB92-222 041/AS; Cost: $26.00, subject to change) Both reports will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Risk Reduction Engineering Laboratory U.S. Environmental Protection Agency Edison, NJ 08837-3679 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati, OH 45268 Official Business Penalty for Private Use $300 BULK RATE POSTAGE & FEES PAID EPA PERMIT No. G-35 EPA/600/S-92/065 ------- |