USEPA Dispersant Toxicity Testing
June 30,2010
Comparative Toxicity of Eight Oil Dispersant Products on Two Gulf of
Mexico Aquatic Test Species
U.S. Environmental Protection Agency
Office of Research and Development
U.S.EPA/ORD Contributors
National Health and Environmental Effects Research Laboratory
Michael J. Hemmer, Mace G. Barron and Richard M. Greene
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approvedfor publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
1

-------
USEPA Dispersant Toxicity Testing
June 30,2010
1.	Introduction
Large quantities of Louisiana sweet crude oil have been released into the Gulf of Mexico
since the explosion of the Deepwater Horizon oil exploration platform on April 20, 2010. As
part of the integrated response effort to mitigate the impact of the oil in the environment, the
decision was made to use dispersants listed on the U.S. Environmental Protection Agency's
(EPA) National Contingency Plan (NCP) Product Schedule (EPA 2010a). Dispersants are being
applied offshore on the surface as well as underwater at the source of the leak. The EPA
conducted independent studies to assess the relative acute toxicity of eight dispersants on the
NCP Product Schedule.
This report summarizes results of the first phase of testing obtained from acute toxicity
tests conducted with eight oil dispersants using two Gulf of Mexico aquatic species: (1) the
mysid shrimp, Americamysis bahia, an aquatic invertebrate, and (2) the inland silverside,
Menidia beryllina, a small estuarine fish. These species are standard test organisms used in a
variety of EPA toxicity test methods. The eight dispersants tested were Corexit 9500A, Dispersit
SPC1000, JD-2000, Nokomis 3-AA, Nokomis 3-F4, Saf-Ron Gold, Sea Brat #4 and ZI-400. The
tests were conducted using an established contract testing laboratory and in compliance with the
Good Laboratory Practice regulations as provided in EPA 40CFR160 (USEPA, 40CFR Part
160). The approach described herein utilized consistent test methodologies within a single
laboratory which provided a means to assess acute toxicity estimates across dispersants and
independently evaluate the NCP Product schedule toxicity information. The next phases of this
study will examine the acute toxicity of Louisiana sweet crude oil and dispersant-sweet crude oil
mixtures on mysids and Menidia - the results will be reported separately at a later date.
2.	Test Methods
The acute toxicity test methods followed, with slight modification, the requirements
specified in U.S. Environmental Protection Agency's 62 FR 15576, Appendix C of Part 300 -
Swirling Flask Dispersant Effectiveness Test, Revised Standard Dispersant Toxicity Test and
Bioremediation Agent Effectiveness Test (USEPA, 1997) and the EPA Test Method 821-R-02-
012, Methods for Measuring the Acute Toxicity of Effluents and Receiving Waters to Freshwater
andMarine Organisms (USEPA, 2002). Specific modifications are shown in Appendix A.
2

-------
USEPA Dispersant Toxicity Testing
June 30,2010
The exposure concentration range for each dispersant was chosen to bracket the estimated
median lethal concentration (LC50) values reported in the NCP Product Schedule. The LC50 is
defined as the concentration of a substance causing mortality in 50% of test organisms for a
specified time interval, in this case, 48-hours for the mysid test and 96-hours for the silverside
test. The commercially available statistical software package, CETIS® was used for the
calculation of LC50 values using an automated decision tree adapted from EPA for selection of
the appropriate statistical method (CETIS, 2009; USEPA, 1994). Point estimate procedures used
to calculate LC50 values included linear regression methods, the non-parametric Spearman-
Karber method and the binomial method. A qualitative comparison was made between LC50
values for the eight dispersants tested as well as with those available in the NCP Product
Schedule. Note that the reproducibility of static acute tests among laboratories using the same
species/toxicant combination has been reported to generally fall within a factor of 3.5 among
laboratories when using nominal concentrations (unmeasured treatment concentrations) for both
freshwater and marine species (USEPA, 1981). Given the use of whole organisms in these tests,
some variation in response attributable to differences in parameters such as culture and
acclimation conditions, stock populations or variable water quality is expected and acceptable.
3. Results - Mysid Toxicity Tests
3.1	Mysid Testing Schedule
Following the first round of eight acute toxicity tests, dispersant LC50s were greater than
the highest concentration tested for four of the eight dispersants. Definitive acute toxicity tests
were repeated using higher test concentrations for JD-2000, Saf-Ron Gold, Sea Brat #4 and ZI-
400.
3.2	Mysid Test Acceptability
Control performance (without dispersant) met all criteria for an acceptable exposure in
each test (>90% survival). All water quality parameters were within ranges specified in the
protocol with the exception of dissolved oxygen for the high test concentration (56 ppm) in the
Nokomis 3-AA exposure at 24 hours, which was measured at 56% of saturation. As dissolved
oxygen levels were >60% at other time points in the test and the toxicity was clearly dose
3

-------
USEPA Dispersant Toxicity Testing
June 30,2010
related, the departure observed in the 56 ppm concentration at 24 hours was not considered to
have had a negative impact on the exposure with Nokomis 3-AA.
3.3 Mysid Toxicity Results
In the first series of acute toxicity tests, LC50 values and 95% confidence intervals were
successfully determined for Corexit 9500A, Dispersit SPC 1000, Nokomis 3-AA and Nokomis
3-F4 and in the second series of acute tests, LC50s were calculated for JD-2000, Saf-Ron Gold,
Sea Brat #4 and ZI400. Test results are summarized in Table 1.
The LC50 values for dispersant acute tests with mysids ranged from 12 ppm for Dispersit
SPC1000 to 788 ppm for JD-200 (Table 1). EPA uses a five-step scale of toxicity categories to
classify pesticides based on their acute toxicity to aquatic organisms: LC50 values of >100 ppm
are considered practically nontoxic; >10 to 100 ppm as slightly toxic; > 1 to 10 ppm as
moderately toxic; LC50s of 0.1 to 1 ppm as highly toxic and LC50s <0.1 ppm as very highly
toxic (USEPA, 2010b). Using this toxicity classification, Corexit 9500A, Dispersit SPC1000,
Nokomis-3 AA, Nokomis 3-F4, Sea Brat #4 and ZI-400 would be classified as slightly toxic
whereas JD-2000 and Saf-Ron Gold would be classified as practically non-toxic to mysids
(Table 1).
Based on comparison of LC50 values and 95% confidence intervals across the eight
dispersants tested in the present study, the rank order toxicity (most to least toxic) of the
dispersants to mysids was: (1) Dispersit SPC1000, (2) Nokomis 3-AA, (3) Nokomis 3-F4,
Corexit 9500A, (4) ZI-400, Sea Brat #4, (5) Saf-Ron Gold, and (6) JD-2000.
Factor ratios were used to compare LC50s derived for the same species/dispersant
combination from different laboratories. The factor ratios between LC50 values determined in
this study and NCP reported LC50 values were calculated as a ratio by dividing the higher of the
two LC50 values by the lower LC50 value for each of the eight dispersants, respectively (Table
1). As an example, using information from Table 1, the factor ratio for Corexit 9500A was
determined as 42/32.2 = 1.3. The factor ratios calculated for Corexit 9500A, Dispersit SPC1000,
Nokomis 3-AA, Nokomis 3-F4, Saf-Ron Gold and ZI-400 were less than or equal to 2.6 which
4

-------
USEPA Dispersant Toxicity Testing
June 30,2010
was considered within normal inter-laboratory variability (USEPA, 1981). Results for JD-2000
and Sea Brat #4 showed lower toxicities (i.e., higher LC50s) with factor ratios of 8.7 and 4.6,
respectively, compared to their reported NCP LC50 values.
3.4 Mysid Reference Toxicant Test
A 48-hr acute toxicity test was conducted with the standard reference toxicant, sodium
dodecyl sulfate (SDS), to evaluate the relative sensitivity of the mysids used in the series of
dispersant toxicity tests. The mysids tested with SDS were from the same population and age
range used for dispersant testing. The 48-hr LC50 and 95% confidence interval calculated for
SDS was 23 ppm [19-26 ppm] which was consistent with the reported NCP LC50 values for
SDS.
4. Results - Mettidia Toxicity Tests
4.1	Menidia Testing Schedule
Following the first round of acute toxicity tests, dispersant LC50s were determined to be
greater than the highest concentration tested for two of the eight dispersants. Definitive acute
toxicity tests were repeated using higher test concentrations for Corexit 9500A and JD-2000.
4.2	Menidia Test Acceptability
Control performance met all criteria for an acceptable exposure in each of the eight
dispersant tests conducted (> 90%). All water quality parameters were within ranges specified in
the test protocol for Menidia beryllina.
4.3	Menidia Toxicity Results
In the first series of acute tests, LC50 values and 95% confidence intervals were
successfully determined for Dispersit SPC 1000, Nokomis 3-AA, Nokomis 3-F4, Saf-Ron Gold,
Sea Brat #4 and ZI-400. In the second series of repeat acute tests, an LC50 was calculated for
Corexit 9500A but not for the dispersant JD-2000. These data are summarized in Table 2. In the
case of JD-2000, 20% mortality was observed in the highest concentration tested of 5,600 ppm,
followed by no mortality observed in the next two highest exposure concentrations which
indicated an LC50 > 5,500 ppm. At the highest concentration, solid material was observed at the
5

-------
USEPA Dispersant Toxicity Testing
June 30,2010
bottom of the replicate test vessels suggesting saturation of the dispersant may have been
achieved.
The LC50 values for dispersant acute toxicity tests with Menidia ranged from 2.9 ppm
for Dispersit SPC1000 to 130 ppm for Corexit 9500A; the LC50 for JD 2000 exceeded the
highest test concentration of 5,600 ppm. Using the EPA toxicity classification, Dispersit
SPC1000 would be considered moderately toxic whereas Nokomis-3AA, Nokomis 3-F4, Saf-
Ron Gold, Sea Brat #4 and ZI-400 would be classified as slightly toxic, and Corexit 9500A and
JD-2000 as practically non-toxic to inland silversides.
Based on comparison of LC50 values and 95% confidence intervals, the rank order
toxicity (most to least toxic) of the dispersants to Menidia were: (1) Disersit SPC1000, (2)
Nokomis 3-F4, Nokomis 3-AA, ZI-400, (3) Saf-Ron Gold, (4) Sea Brat #4, (5) Corexit 9500A,
and (6) JD-2000.
The factor ratios calculated for Dispersit SPC1000, Nokomis 3-AA, Nokomis 3-F4, Saf-
Ron Gold, Sea Brat #4 and ZI-400 were less than or equal to 1.83 which was considered within
normal inter-laboratory variability. The factor ratios of 5.2 and 13.8 for Corexit 9500A and JD-
2000 indicate that the LC50 values reported for Corexit 9500A and JD-200 in the NCP Product
Schedules would be considered different (i.e., lower) from the LC50 values determined in the
present study.
Possible explanations for the 13.8 fold difference between the reported NCP LC50 for
JD-2000 and the highest exposure concentration tested in the present study may be attributable to
batch-to-batch variability in the manufacturing process, instability of the stored product over
time, or a change in the product formulation.
A A Menidia Reference Toxicant Test
A 96-hr acute toxicity test was conducted with the reference toxicant SDS to evaluate the
relative sensitivity of the Menidia used in the series of dispersant toxicity tests. The Menidia
tested with SDS were from the same population and age range used for dispersant testing. The
6

-------
USEPA Dispersant Toxicity Testing
June 30,2010
96-hr LC50 and 95% confidence interval calculated for SDS was 9.5 ppm [8.7-10 ppm] which
was consistent with the reported NCP LC50 values for SDS. It should be noted that during the
last 24 hours of the test, the temperature dropped to 22°C, which was 2 degrees below the
acceptable criteria and thus invalidated the test. However, there was no difference in mortality
counts between the 72-hour and the 96-hour observations suggesting the temperature change had
no negative impact on the test or the final calculated LC50.
5.0 Conclusions
The present study provided an independent, quantitative assessment of acute toxicities of
eight dispersants to two aquatic species inhabiting Gulf of Mexico waters. Toxicity was
determined as the LC50 derived from standard short term acute tests using standard test species,
specifically the Gulf mysid, Americamysis bahia, and the inland silverside, Menidia beryllina. In
general, the toxicity values (i.e., LC50s) for mysids ranged over nearly two orders of magnitude
and for Menidia over three orders of magnitude. Given the expected range of inter-laboratory
variability, the results of the present study were consistent with test results reported in the NCP
Product Schedule, with the exception of two dispersants for each test species which yielded
higher LC50s (i.e., lower toxicity) than reported in the NCP. The rank order toxicity of the eight
dispersants was generally similar to the information provided in the NCP Product Schedule. For
both test species, Dispersit SPC1000 was the most toxic and JD-2000 the least toxic. The other
six dispersants varied in relative toxicity to mysids and Menidia, with LC50 values ranging from
20 to 130 ppm. Overall, the dispersants were classified as being slightly toxic to practically non-
toxic to both test species, with the exception that Dispersit SPC1000 would be considered
moderately toxic to Menidia. Corexit 9500A, the dispersant currently applied offshore at the
surface and underwater, falls into the slightly toxic category for mysids and the practically non
toxic category for Menidia.
Short-term acute toxicity tests using consistent methodologies and test organisms provide
important and fundamental information on oil spill dispersants and other toxicants. The next
phase of testing will examine the acute toxicity of Louisiana sweet crude oil and dispersant-
sweet crude oil mixtures on mysids and Menidia. The comparative toxicity analysis of
dispersants, sweet crude oil and dispersant-sweet crude oil mixtures on standard aquatic test
7

-------
USEPA Dispersant Toxicity Testing
June 30,2010
species will provide improved understanding of potential toxicological effects associated with
this oil spill.
6.0 References
CETIS. 2009. Comprehensive Environmental Toxicity Information System: Users Manual.
Tidepool Scientific Software, McKinleyville, CA.
USEPA. 40 CFR, Part 160. Federal Insecticide, Fungicide, and Rodenticide Act. Good
Laboratory Practices Standards; Final Rule. Office of the Federal Register,
National Archives and Records Administration. U.S. Government Printing Office,
Washington, D.C.
USEPA. 1981, Results: Interlaboratory Comparison- Acute Toxicity Tests Using Estuarine
Animals. February 1981. EPA-600/4-81-003. U.S. Environmental Protection
Agency, Gulf Breeze, Florida.
USEPA. 1994. Short-Term Methods for Estimating the Chronic Toxicity of Effluents and
Receiving Waters to Marine and Estuarine organisms, 2nd ed. Environmental Monitoring
Systems Laboratory, Office of Research and Development, U.S. Environmental
Protection Agency, Cincinnati, OH. EPA/600/4-51/003.
USEPA. 1997. Swirling Flask Dispersant Effectiveness Test, Revised Standard
Dispersant Toxicity Test, and Bioremediation Agent Effectiveness Test. Part 300,
Appendix C. EPA 62 FR 15576. U.S. Environmental Protection Agency, Washington,
D.C. April 1, 1997.
USEPA. 2002. Methods for Measuring the Acute Toxicity of Effluent and Receiving Waters
to Freshwater and Marine Organisms. 5th Edition, October 2002. EPA-821-R-02-012.
U.S. Environmental Protection Agency, Washington, D.C.
USEPA. 2010a. http://www.epa.gov/emergencies/content/ncp/product schedule.htm
USEPA. 2010b. http://www.epa.gov/oppefedl/ecorisk ders/toera analysis eco.htm#Ecotox.
Accessed June 23, 2010.
8

-------
USEPA Dispersant Toxicity Testing
June 30,2010
Table 1. Results of mysid 48-hr static acute toxicity tests with eight dispersants. LC50 values
(ppm), 95% confidence intervals [in brackets] and the toxicity classification of dispersant LC50s
derived in the present study. NCP Product Schedule listing of dispersant LC50s and 95%
confidence intervals [in brackets] shown in right column for comparison.
Dispersant
This Study
LC50 (ppm)
[95% CI]
Toxicity
Category1
NCP Product
Schedule
LC50 (ppm)
[95% CI]d
Dispersit SPC 1000
12 [10-14]a
Slightly Toxic
16.6 [14.1-19.6]
Nokomis 3-AA
30 [27-34]b
Slightly Toxic
20.2 [17.4-22.8]
Corexit 9500A
42 [38-47]°
Slightly Toxic
32.2 [26.5-39.2]
Nokomis 3-F4
42 [38-47]°
Slightly Toxic
32.2 [28.4-36.5]
ZI -400
55 [50-61]b
Slightly Toxic
21.0 [17.9-24.5]
Sea Brat #4
65 [57-74]a
Slightly Toxic
14.0 [±10.4]
Saf-Ron Gold
118 [104-133]b
Practically Non-Toxic
63.0e [52.9-75.1]
JD-2000
788 [627-946]a
Practically Non-Toxic
90.5e [76.1-108]
toxicity classification per USEPA 2010 applied to results of present study
aEstimated by linear regression method
bEstimated by Speannan-Karber method
Estimated by binomial method
dValues as reported in NCP Product Schedule documentation by manufacturer
eClassified as slightly toxic according to values provided in NCP Product Schedule
9

-------
USEPA Dispersant Toxicity Testing
June 30,2010
Table 2. Results of Menidia 96-hr static acute toxicity tests with eight dispersants. LC50 values
(ppm), 95% confidence intervals [in brackets] and the toxicity classification of dispersant LC50s
derived in the present study. NCP Product Schedule listing of dispersant LC50s and 95%
confidence intervals [in brackets] shown in right column for comparison.
Dispersant
This Study
LC50 (ppm)
[95% CI]
Toxicity
Category1
NCP Product
Schedule
LC50 (ppm)
[95% CI]d
Dispersit SPC 1000
2.9 [2.5-3,2]b
Moderately Toxic
3.5 [3.1-4.0]
Nokomis 3-F4
19 [16-21]b
Slightly Toxic
29.8 [24.0-35.4]
Nokomis 3-AA
19 [17-21]b
Slightly Toxic
34.2 [29.2-37.95]
ZI -400
21 [18-23]b
Slightly Toxic
31.8 [28.7-35.1]
Saf-Ron Gold
44 [41-47]b
Slightly Toxic
29.4 [25.2-34.3]
Sea Brat #4
55 [49-62]b
Slightly Toxic
30.0 [±16.2]
Corexit 9500A
130 [122-138]b
Practically Non-Toxic
25.2e [13.6-46.6]
JD-2000
>5,600
Practically Non-Toxic
407 [330-501]
toxicity classification per USEPA 2010 applied to results of present study
bEstimated by Speannan-Karber method
dValues as reported in NCP Product Schedule documentation by manufacturer
eClassified as slightly toxic according to values provided in NCP Product Schedule
10

-------
USEPA Dispersant Toxicity Testing
June 30,2010
Appendix A
Test parameter
Specified in SubPart J
Appendix C (USEPA 1997)
Method used in present study, and
specified in USEPA 2002
Photoperiod and light
intensity
*24 hr light
*higher intensity light
* 16 hr light/8 hr dark
*Moderate intensity light
Glassware cleaning
*Hexane immersion
*Acetone rinse
Reference toxicant test
*Two species simultaneously
* Staggered tests
Rangefinder tests
*Prior to definitive test
*Use NCP data to define test concentrations
Mysid age
*5-7 day old larvae
* 1 to 6 day old; all within 24 hr same age
Toxicant stock
solution preparation
for mysid test
*Blender 10,000 rpm
*gas tight syringes
*Top stirring at 70% vortex
*graduated glass pipettes
Mysid test solution
mixing
*no specification
* short term gentle mixing following stock
addition
Mysid additions to test
chambers
*no specification
impartial, two at a time
Menidia age
*7 day old larvae
*9-14 day old, all within 24 hr same age
Menidia test solution
mixing
*test jars on shaker platform
*same procedure as for mysids
Dilution Water
*Natural Seawater Prefered
* Salinity adjusted, 20 [j,m filtered natural
seawater
11

-------