Toxicity of Binary Chemical Munition Destruction Products: Methylphosphonic Acid,Methylphosphinic Acid,2-Diisopropylaminoethanol,DF Neutralent,and QL Neutralent

Abstract

This paper reports the toxicity and environmental impact of neutralents produced from the hydrolysis of binary chemical agent precursor chemicals DF (methylphosphonic difluoride) and QL (2-[bis(1-methylethyl)amino]ethyl ethyl methylphosphonite). Following a literature review of the neutralent mixtures and constituents, basic toxicity tests were conducted to fill data gaps, including acute oral and dermal median lethal dose assays, the Ames mutagenicity test, and ecotoxicity tests. For methylphosphonic acid (MPA), a major constituent of DF neutralent, the acute oral LD₅₀ in the Sprague-Dawley rat was measured at 1888 mg/kg, and the Ames test using typical tester strains of Salmonella typhimurium and Escherichia coli was negative. The 48-h LC₅₀ values for pH-adjusted DF neutralent with Daphnia magna and Cyprinodon variegatus were >2500 mg/L and 1593 mg/L, respectively. The acute oral LD₅₀ values in the rat for QL neutralent constituents methylphosphinic acid (MP) and 2-diisopropylaminoethanol (KB) were both determined to be 940 mg/kg, and the Ames test was negative for both. Good Laboratory Practice (GLP)-compliant ecotoxicity tests for MP and KB gave 48-h D. magna EC₅₀ values of 6.8 mg/L and 83 mg/L, respectively. GLP-compliant 96-h C. variegatus assays on MP and KB gave LC₅₀ values of 73 and 252 mg/L, respectively, and NOEC values of 22 and 108 mg/L. QL neutralent LD₅₀ values for acute oral and dermal toxicity tests were both > 5000 mg/kg, and the 48-h LD₅₀ values for D. magna and C. variegatus were 249 and 2500 mg/L, respectively. Using these data, the overall toxicity of the neutralents was assessed.

Chemical warfare agents are extremely toxic substances. The United States is obligated to destroy these munitions in order to comply with the Chemical Weapons Convention (formally the 1993 Paris Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and on Their Destruction). The U.S. Army has developed procedures to destroy these weapons. As a result of growing public concerns about and opposition to incineration, many chemical weapons are being destroyed by neutralization with water (National Research Council [NRC] 1996a). Little toxicity data have been published for many products of these neutralization reactions. We have tested the end products of two such neutralization reactions (“neutralents”) of binary chemical munition components for acute toxicity in mammals and several aquatic species, as well as mutagenic potential.

Although there is considerable data available for the chemical warfare agents GB and VX (NRC 2003), many risk assessors cannot gain direct access to some of the original data, and significant data gaps remain for many products of neutralization. The data reported herein will be helpful to the risk assessment community, because this report allows all concerned to see the impact of the substances on various biota and to draw conclusions about the risks of such substances based on original data rather than on estimates. The present experiments developed acute toxicity values because the neutralents were generated over a relatively short period of time and shipped to a commercial hazardous waste disposal facility, where they will be destroyed. An injury from these neutralents is likely to be the result of a spill or other accidental, acute exposure.

Binary chemical munitions contain two chemicals that, although less toxic than a chemical warfare agent, combine after the weapon is fired to form agent while in flight. In the fall of 2006, the U.S. Army completed the destruction of the binary agent components DF (methylphosphonic difluoride; CAS reg. no. 676-99-3) and QL (the 2-[bis(1-methylethyl)amino]ethyl ethyl ester of methylphosphonous acid; CAS reg. no. 57856-118), which are precursors of the chemical agents GB and VX, respectively.

The neutralization process destroys DF and QL through hydrolysis. DF neutralent has a pH between 1 and 2.4 and is primarily composed of 71% water, 21% methylphosphonic acid (MPA; CAS reg. no. 993-13-5), and 8% hydrogen fluoride (HF; CAS reg. no. 7664-39-3). HF and MPA are the main products on DF hydrolysis; see Figure 1A . Minor constituents of DF neutralent include 0.3% sodium fluoride (NaF; CAS reg. no. 7681-49-4) and 0.015% methylphosphonic fluoride (MF; CAS reg. no. 1511-67-7); DF is not detected in the neutralent, with a detection limit of 0.0098%. QL neutralent has a pH between 7 and 8 and is primarily composed of 82% water, 9.5% 2-[bis(1-methylethyl)amino]ethanol (designated as KB; CAS reg. no. 96-80-0), 5.2% methylphosphinic acid (MP; CAS reg. no. 4206-94-4), and 3% ethanol (CAS reg. no. 64-17-5). MP, KB, and ethanol are the main products on QL hydrolysis; see Figure 1B . QL is not detected in the neutralent, with a detection limit of 0.5%.

A search of the open literature was conducted to identify known information concerning the toxicity of the major neutralent constituents (Watson et al. 2004). We found that HF, NaF, and ethanol have already been the subjects of extensive research, whereas other neutralent constituents had poorly understood toxicological properties. This paper reports the results of experiments designed to characterize the acute mammalian toxicity, genotoxicity and ecologic toxicity of these substances.

There is limited toxicological information available for MPA (a component of DF neutralent) and for MP and KB (components of QL neutralents). Data available for MPA indicate that this compound is of minimal toxicity to animals and aquatic organisms. MPA LD₅₀ values are reported to equal or exceed 5000 mg/kg in the mouse and the rat (as referenced in Williams et al. 1987), although the experimental details of this study could not be found. Additionally, recent studies have shown that the oral approximate lethal dose (ALD) of MPA in the rat is 2300 mg/kg (Finlay 2004). Although these values are consistent with the idea that MPA is not likely to be acutely toxic to humans unless large amounts are ingested, an additional acute oral lethality test was conducted in order to verify this and to provide published data for use in future risk assessments where ingestion must be considered, e.g., to assess MPA concentrations in aqueous discharges. Ecological testing of MPA indicated low toxicity to freshwater protozoan communities and selected fish species (including bluegill fish and fathead minnows) (Williams et al. 1987). MPA is relatively persistent in the environment; the primary fate appears to be binding to soil and very slow biodegradation (Small 1984). Bioaccumulation of MPA in organisms is expected to be minimal based on a low estimated log K _ow value of −2.28 (Small 1984). Thus, additional ecotoxicity test on MPA were not conducted.

Less is known about the environmental fate of MP and KB. MP in solution appears to oxidize slowly to MPA upon storage in the presence of oxygen (NSCMP 2005). Analogs of KB are water soluble and readily biodegraded (Rothkopf and Bartha 1984), suggesting that biodegradation of KB should be its environmental fate. Additionally, a range-finding LD₅₀ of 1070 mg/kg for KB in Carworth-Wistar rats was reported (Smyth et al. 1954).

Additionally, data on the toxicity of DF and QL neutralents, as they exist as mixtures, are lacking. Testing of the mixtures was also conducted in order to determine whether or not synergistic interactions occur between components of the neutralents, which might lead to otherwise unpredicted toxicological outcomes.

Potential human exposure routes for DF and QL neutralents are primarily via dermal contact and inhalation of volatile components following spills; oral exposure was determined to be extremely unlikely. The following tests were performed to fill the identified data gaps:

Acute oral toxicity tests in Sprague Dawley rats to measure the median lethal dose (LD₅₀) for MPA, MP, KB, and QL neutralent

Toxicity tests to measure the acute dermal LD₅₀ for QL neutralent

Good Laboratory Practice (GLP)-compliant acute eco-toxicity tests for MP and KB using guidelines from the Organisation for Economic Cooperation and Development (OECD) and/or guidelines supplied by the U.S. EPA Office of Prevention, Pesticides, and Toxic Substances (OPPTS)

Acute ecotoxicity tests for DF neutralent and QL neutralent; these were performed with the same procedures as the GLP study but without full GLP documentation. The less expensive test was conducted because DF and QL neutralents are unlikely to be generated in the future, whereas the component substances could be of greater interest in the future.

Ames mutagenicity tests for KB, MPA, and MP

Acute dermal and oral LD₅₀ studies were not performed on DF neutralent because its low pH and the hydrofluoric acid concentration make it strongly corrosive. Acute oral toxicity tests were conducted on MP, KB, MPA, and QL neutralent. The rat was the experimental model because this is the species for which the most oral LD₅₀ data are available, allowing us to assess the toxicity of these components in relation to other chemicals.

Inhalational testing was not considered because MPA and KB are both very soluble in water and relatively non-volatile substances (MPA melts at 105°C, and the manufacturer indicates that KB boils at 187°C to 192°C). MP is also relatively non-volatile, as it is substantially ionized at the pH of QL neutralent.

EXPERIMENTAL

Test Articles

DF and QL neutralents were provided by the U.S. Army. KB was purchased from Sigma-Aldrich (purity ≥ 99%). MPA and MP were prepared by Prime Organics, Inc. (Woburn, MA). The MP was characterized by both ¹H [(D₂O) δ 8.15 (s, 1H) 2.89 (s, 1H) 1.15 (d, 3H)] and ³¹P NMR [(D₂O) δ 36.37]. The first batch of MP was approximately 84% pure and the second batch of MP was approximately 80% pure by integration of the peaks in the ³¹P NMR. The major impurity was tentatively identified as the methyl ester of MP based on ¹H and ³¹P NMR [³¹P (D₂O) δ 35.03 for the methyl ester].

During an attempted large-scale preparation of MP, a potential safety hazard in the synthetic procedure of Boyd and Regan (1994; depicted in Figure 2) was uncovered. A brief fire occurred within the vessel while the dichloromethane/methanol solvent was being removed during an attempt to prepare several kg of MP. The entire batch did not ignite, most likely because there was still a significant amount of nonflammable dichloromethane remaining. We believe that the pyrophoric nature of the intermediate silyl-protected phosphine (I in Figure 2) is roughly comparable to that of tert-butyl lithium. This hazard had not been noted previously, presumably because of the smaller scales involved. We were able to prepare a limited quantity (220 g) of MP by synthesizing two smaller batches avoiding any further dangerous incidents during the synthesis. Anyone attempting this scale up should modify the reaction conditions by using excess methyl iodide to ensure that the Arbuzov reaction goes to completion, thus depleting the highly reactive P(III) species, which is presumably pyrophoric, rather than the P(V) product. Furthermore, the reaction mixture should not be allowed to reach dryness before methanol is added to hydrolyze the silyl groups and the vacuum should be released to inert gas rather than the atmosphere.

Median Lethal Dose (LD₅₀) Assays

GLP-compliant acute oral toxicity assays using Sprague Dawley rats were performed on MP, KB, MPA, and QL neutralent at Illinois Institute of Technology Research Institute (IITRI, Chicago, IL). The tests were performed in accordance to U.S. Environmental Protection Agency (EPA) Office of Prevention, Pesticides, and Toxic Substances (OPPTS) Guideline 870.1100 (U.S. EPA 1998a), and were performed in compliance with Department of Defense (DoD) Directive 3216.1 requiring DoD funded animal studies to adherence to the National Research Council Guide for Laboratory Animal Facilities and Care (NRC 1996b).

Briefly, male and female Sprague-Dawley derived rats (Crl:CD IGS BR) from Charles River Laboratories (Wilmington, MA), were approximately 7 weeks of age and weighed approximately 150 to 225 g on arrival. Animals were individually housed in steel wire cages, fed food and water ad libitum in rooms maintained on 12-h light/12-h dark cycles. Room temperature was kept between 18°C and 26°C, and humidity ranged from 30% to 70%. For each treatment group, five males and five females were used. Animals were randomly assigned to treatment or control groups and the body weight variation of the experimental rats did not exceed ± 20% of the mean weight of each sex. Animals were administered the test articles or vehicle (water) control via oral gavage following an overnight fasting period in order to minimize animal-to-animal drug absorption differences. The first dose was 2000 mg/kg. If no lethality was demonstrated at this dose, then an additional group of animals (5/sex) was similarly dosed at 5000 mg/kg. If no lethality was demonstrated at that dose, then no further testing was required and the LD₅₀ was determined to be >5000 mg/kg. If compound-related mortality was induced at 2000 mg/kg, then two additional groups of animals (5/sex/group) were dosed at levels less than 2000 mg/kg. If compound-related mortality occurs at 5000 mg/kg, then additional groups of animals (5/sex/group) were dosed at levels less than 5000 mg/kg, but greater than 2000 mg/kg. The acute oral median lethal dose LD₅₀ and the associated 95% confidence interval was then determined using the method of Miller and Tainter (1994).

The acute dermal toxicity assay was also conducted at IITRI, in accordance with OPPTS guideline 870.1200 (U.S. EPA 1998b) and OECD Guideline 402 (OECD 1987). Male and female New Zealand white rabbits from Kuiper Rabbit Ranch (Gary, IL) weighing between 2.0 and 3.0 kg and approximately 3 to 4 months of age were used in this study. Animals were randomly assigned to treatment or control groups and the body weight variation of the experimental rabbits did not exceed ± 20% of the mean weight of each sex. The test substance or vehicle (water) control was held in contact with the skin with a gauze semiocclusive patch secured with nonirritating tape to shaved dorsal skin. Gauze was removed 24 h following application. Treated groups were exposed to 5000 mg/kg of the test article. If no mortality was seen at this concentration, the acute dermal LD₅₀ was reported to be >5000 mg/kg, otherwise additional testing would be conducted at lower doses. LD₅₀ values are calculated in the same manner as the oral LD₅₀ values described above.

Ames Mutagenicity Assay

The Ames Mutagenicity assay was conducted at BioReliance (Rockville, MD) on MP, KB, and MPA in compliance with OECD Guideline 471 (OECD 1997) and with the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH 1996, 1997). Studies were conducted with and without Aroclor 1254-induced rat liver S9 metabolic enzymes prepared from male Sprague-Dawley rats induced with a single 500 mg/kg intraperitoneal injection of Arochlor-1254, 5 days prior to sacrifice.

Selection of dose levels for the mutagenicity assay was based upon the toxicity and precipitation profile of the test article assessed in a preliminary toxicity assay. The highest dose used in this preliminary assay was the highest workable concentration in deionized water, up to a maximum of 5 mg/plate.

In the Ames assay, four strains of Salmonella bacteria (TA98, TA100, TA1535, and TA1537), and one strain of E. coli (WP2 uvrA) bacteria were exposed to negative vehicle controls (water) and to at least eight concentrations of the test article, one plate per level, in both the presence and absence of S9 activation. The test system was exposed to the test article via the plate incorporation methodology originally described by Ames, McCann, and Yamasaki (1975) and updated by Maron and Ames (1983). Positive controls as described in OECD 402 (OECD 1987) were plated concurrently with the assay. 2-Aminoanthracene (1 μ g/plate) was the positive control for all S9-activated strains. The following positive controls were used in the absence of S9 activation: 2-nitrofluorene at 1 μ g/plate for TA98, sodium azide at 1 μ g/plate for TA100 and TA1535, 9-aminoacridine at 75 μ g/plate for TA1537, and methyl methanesulfonate at 0.5 mg/plate for WP2. Deionized water negative controls were also plated for each tester strain with and without S9 activation. Experimental bacterial plates were incubated for approximately 48 to 72 h at 37°C ± 2°C. The number of revertant colonies per plate was scored to determine if the test article increased the rate of mutation.

Ecotoxicity Assays

Basic ecotoxicity tests were conducted for QL and DF neutralent mixtures on Daphnia magna (water flea) and Cyprinodon variegatus (Sheepshead minnow) at Biological Monitoring (Blacksburg, VA) for a 48-h period in accordance to the acute EPA method defined in EPA 821 R02012 (2002). Due to the acidic pH of DF neutralent (pH ∼ 2.0), neutralization to a pH of 7.1 to 8.0 using sodium hydroxide was performed so that the testing would address toxicological attributes of DF neutralent other than the acidic pH. This neutralization converts HF to NaF and MPA to its monosodium salt, the same process that would occur in most spills because of the buffering capacity of many natural water bodies. If a large amount of DF neutralent were released into a small enough body of water, the buffer capacity would be exhausted and the acidity of the neutralent would be expected to dominate the impact.

First, range-finding assays were conducted with D. magna and juvenile C. variegatus to determine the ideal nominal concentrations of test articles to use in the definitive assays to maximize the probability of obtaining a definitive endpoint. The maximum nominal concentration in these assays was 1000 or 2500 mg/L, depending on the solubility of the test article so if these substances did not contribute to toxic outcomes at concentrations below this, the median lethal concentration (LC₅₀) might be expressed as >1000 or >2500 mg/L. In the definitive test, 2 chambers of 10 test subjects each were exposed to 5 concentrations of each test article and then observed for signs of toxicity (death, immobility). Statistical analyses for the basic ecotoxicity tests were performed using a computer program supplied by the U.S. EPA that utilizes binomial, moving average, probit, and/or Spearman-Karber statistical methods for the determination of LC₅₀ values with 95% confidence intervals and no adverse effect concentration (NOAEC) values (Finney 1971; Thompson 1947; Stephan 1977; Cornell 1965). The most conservative LC₅₀ value provided by a statistical test appropriate for the distribution of the data from the ecotoxicity test was used. The LC₅₀ determined from D. magna testing in QL was calculated using the Probit method, and the LC₅₀ determined from D. magna testing in pH-adjusted DF was calculated using the Spearman-Karber method.

Ecotoxicity tests for MP and KB were conducted on D. magna and C. variegatus at Wildlife International Ltd. (Easton, MD) in compliance with Good Laboratory Practices (GLP). The D. magna assay was conducted according to OECD Guideline 202 (OECD 1984), as well as OPPTS Guideline 850.1010 (U.S. EPA 1985) and ASTM Standard E-729-96 (ASTM 1996). Median effective concentrations (EC₅₀) were calculated when possible, based on the number of dead Daphnids observed in each test concentration after each 24 hr interval or exposure. The C. variegatus studies on MP and KB used procedures in accordance with OPPTS Guideline 850.1075 (U.S. EPA 1996) and ASTM Standard E729-96 (ASTM 1996). Observations of mortality and other clinical signs were made throughout the 96 hr test period, and these data were analyzed using an EPA computer package by C. E. Stephan (1978). Cumulative percent immobility and mortality observed in the treatment groups was used to calculate EC₅₀ and LC₅₀ values in the same manner these values were calculated for the basic ecotoxicity tests with the exception that the Spearman method was not used. The no observed effect concentration (NOEC) was determined by visually interpreting the clinical observation data. For methylphosphinic acid, values for the D. magna and C. variegatus studies were determined using the binomial method. For KB, values for the D. magna study were determined using the binomial method, and values for the C. variegat U.S. study were determined using Probit analysis.

Fluoride Content

In order to estimate the DF:water ratio in the DF neutralent sample provided to us, the fluoride content of DF neutralent was measured by ProChem Analytical Inc., Roanoke, VA using chromatographic techniques described in U.S. Environmental Protection Agency Method 300.0 (U.S. EPA 1993).

RESULTS

Acute Oral Toxicity Assays

Data for the acute oral toxicity test on MP are presented in Table 1. At 2000 mg/kg, all 10 animals died on day 1. Adverse clinical observations were seen in all animals dosed with 2000 and 1000 mg/kg. Adverse signs included coldness to touch, labored breathing, hypoactivity, hunched posture, redness around the nose, tremors, and lacrimation. At necropsy all animals that died as a result of dosing with 2000 and 1000 mg/kg had dilated, fluid filled stomachs with streaked mucosa and dilated small and large intestines. Pale streaks on the liver were also commonly observed in these animals. Of the three surviving animals in 1000 mg/kg, two had no gross lesions and one had urinary bladder calculi (not thought to be treatment related). At 500 mg/kg, all animals were cold to the touch, and in all cases this subsided on day 4 of the study; no adverse effects seen upon necropsy. The acute oral LD₅₀ value for MP was calculated to be 940 mg/kg (combined sexes), with a 95% confidence interval of 476 to 1858 mg/kg).

Data for the acute oral toxicity test on KB are presented in Table 2. At doses of 1250, 1750, 2400, 3000, and 5000 mg/kg KB, all deaths occurred on day 1 of the study except for one male rat dying on day 2 in each of the 1250, 1750, and 3000 mg/kg dose groups. At 5000 mg/kg, all animals died before substantial clinical observations could be obtained. At 3000 mg/kg, one male and one female died prior to be able to assess clinical observations. Of the remainder of the animals that died after clinical observations were taken, all exhibited adverse clinical signs including hypoactivity, coldness to the touch, tremors, and labored breathing. At 2400 mg/kg, all five males and four of the females died before clinical observations were taken. The remaining female rat was hypoactive, cold to touch, with slight tremors and labored breathing before being found dead. At 1750 mg/kg, two males and four females were observed to be prostrate prior to death, and all the males were found to be hypoactive before death. At 1250 mg/kg, hypoactivity and tremors were seen in all the males before death as well as in three of the females. At 1100 mg/kg, all males and two females were observed to be hypoactive with tremors. At 500 mg/kg, all animals had adverse clinical findings including hypoactivity, slight tremors, rough hair coat, ataxia, an emaciated appearance, and/or redness around the mouth and nose. In all cases, these adverse effects subsided by study day 4. Upon necropsy, all animals dosed at levels above 500 mg/kg had test article–related gross lesions. Animals dying quickly after dosing exhibited fluid filled stomachs with pigmentation and dilatation. Additional necropsy observations included thin wall of the stomach, stomach adhered to the liver, spleen adhered to the stomach or liver with dark pigmented parenchyma, dilated urinary bladder with calculi, and/or red testes or ovaries. At 500 mg/kg, no gross lesions were observed in two males and two females. In three males and three females, the liver was adhered to the stomach. The acute oral LD₅₀ value for KB was calculated to be 940 mg/kg (combined sexes), with a 95% confidence interval of 615 to 1430 mg/kg.

Data for the acute oral toxicity test on MPA are presented in Table 3. At 2000 mg/kg, all deaths occurred on day 1 of the study. At 1750 mg/kg, one female died on day 3 and one male on day 2. At 2000 mg/kg, four males and five females were hypoactive and cold to the touch with labored breathing. In addition, three males and two females were observed with convulsions and one male was observed with tremors and redness around the nose/eyes. No adverse clinical observations were seen in any other dose groups. Upon necropsy, all animals treated with 2000 mg/kg had at least one gross lesion noted, including a dilated stomach (four males and two females), a dilated small and large intestine (three males and three females), kidney foci (one male and one female), and enlarged stomach (three females). At 1750 mg/kg, four males and two females had no gross lesions. One male had a dilated stomach, and dilated intestine filled with yellow fluid. Two females had pigmented ovaries, and one female had a dilated, thin-walled stomach filled with fluid and mucosa, as well as a tan liver focus. Upon necropsy, ovary pigmentation was observed in two females that had not died, and in the male that died, the stomach was dilated and fluid-filled, with a thin wall, mucous covering and black pigment. The small intestine was dilated and filled with yellow fluid and the liver exhibited a tan lateral lobe focus. At 1500 mg/kg, one female had bilateral red ovary pigmentation. At 1000 mg/kg, no gross lesions were observed. The acute oral LD₅₀ value for MPA was calculated to be 1888 mg/kg (combined sexes), with a 95% confidence interval of 1462 to 2438 mg/kg.

In the QL neutralent acute oral assay, no mortalities occurred as a result of treatment with doses up to 5000 mg/kg. All animals exhibited adverse clinical signs after dosing. At 2000 mg/kg animals were cold to the touch after dosing, but this subsided within an hour. At 5000 mg/kg, all animals were cold to the touch and the females were also hypoactive with labored breathing, convulsions, and slight tremors. All effects seen at 5000 mg/kg subsided within 3 h after dosing. Upon necropsy there were no treatment-related gross lesions found. The acute oral LD₅₀ value for QL neutralent was taken to be 5000 mg/kg (both sexes).

Acute Dermal Toxicity Assays

In the QL neutralent acute dermal assay, no deaths were observed up to the highest dose tested (5000 mg/kg), nor were any gross lesions observed upon necropsy. Soft stool was observed in one female on days 1 and 3 of the study, and diarrhea was observed in one male on day 9 of the study. It is unclear whether these effects were definitively associated with test article administration. The acute dermal LD₅₀ value for QL neutralent was taken to be 5000 mg/kg (both sexes).

Ecotoxicity Assays

Data from ecotoxicity tests using QL neutralent and DF neutralent with D. magna are presented in Table 4. The 48-h. LC₅₀ value for DF neutralent adjusted to neutral pH was calculated to be 1600 mg/L, with a 95% confidence interval of 1420 to 1790 mg/L. The 48-h LC₅₀ value for QL neutralent was calculated to be 250 mg/L, with a 95% confidence interval of 180-320 mg/L. All C. variegatus survived both in control experiments and at concentrations of 156, 313, 625, 1250, or 2500 mg/L with QL neutralent and DF neutralent adjusted to neutral pH. Thus, the 48-h LC₅₀ in C. variegatus for both DF neutralent adjusted to neutral pH and QL neutralent is taken to be >2500 mg/kg.

Data from ecotoxicity tests on KB and MP are presented in Tables 5 and 6, respectively. The 48-h LC₅₀ value for MP in D. magna was calculated to be 6.8 mg/L, with a 95% confidence interval of 4.3 to 12 mg/L. The 96-h LC₅₀ value for MP in C. variegatus was calculated to be 70 mg/L, with a 95% confidence interval of 60 to 100 mg/L. The 48-h LC₅₀ value for KB in D. magna was calculated to be 80 mg/L, with a 95% confidence interval of 50 to 100 mg/L. The 96-h LC₅₀ value for KB in C. variegatus was calculated to be 250 mg/L, with a 95% confidence interval of 220 to 280 mg/L. Finally, the 96-h NOEC in C. variegatus was found to be 22 mg/L for MP and 108 mg/L for KB.

Mutagenicity Assays

Mutagenicity tests performed on MP, MPA, and KB indicated that the mutation frequency in treated bacteria was never statistically higher than the controls concentrations tested, with and without S9 metabolic enzymes, as presented in Tables 7 to 9. Thus, acute exposure to these substances is unlikely to cause genotoxicity.

DF Neutralent Fluoride Measurement

The fluoride concentration of DF neutralent was measured as 166,000 mg/L with a method detection limit of 10,000 mg/L due to sample dilution.

DISCUSSION

DF Neutralent and Its Components

MPA

Toxicological data for MPA were incomplete, although based on the available data, it appeared unlikely that MPA would have a significant adverse impact on the toxicity of DF neutralent. The results of the tests on MPA support our initial conclusion; the acute oral LD₅₀ for MPA is 1888 mg/kg. This value is lower than the 5000 mg/kg value referenced by Williams et al. (1987) from an unpublished study, and similar to the 2300 mg/L acute lethal dose (ALD) value in Sprague-Dawley rats reported by Finlay et al. (2004). Differences from the Williams et al. value could be attributed to differences in the exact strain of rat used, and small procedural differences. Additionally, MPA was found not to be mutagenic. Previously existing data indicate that the ecological risk of MPA is low (Williams et al. 1987). The acute oral toxicity and mutagenicity values of MPA reported here will also be useful in evaluating the risks to human health of the hydrolysates obtained from chemical agents GB and VX, which also contain MPA as constituents.

DF Neutralent Mixture

The ecotoxicity values for the neutralized DF neutralent indicated that DF neutralent will not be markedly toxic to aquatic organisms if spilled into an environment in which the pH would be buffered. Furthermore, the fluoride result indicates that the DF neutralent supplied for this test was made up at a 2.7:1 water:DF ratio rather than the 5:1 ratio that is being used at the Binary Destruction Facility (BDF) and produces the composition described in the Introduction. Thus, we believe that the reported values likely overestimate the toxicity of DF neutralent as produced at the BDF.

Our current results support the previous conclusion that the largest risk from DF neutralent was likely from the HF content, which can have severe effects via the inhalational (NRC 2004) and dermal pathways. Two studies address the effects of dermal exposure to aqueous HF at concentrations close to that found in DF neutralent. In one study, Japanese workers exposed to bleach containing 9.5% HF experienced severe dermal irritation, but adverse systemic effects associated with exposure to larger amounts of HF such as cardiac arrhythmias, skeletal fluorosis, and hypocalcemia were not observed (Fujimoto et al. 2002). In a second study, a retrospective investigation of 237 workers exposed to rust removers containing 6% to 11% HF (El Saadi et al. 1989) indicated that patients reported significant pain, dermal irritation, redness, discoloration, and blistering, but no signs of systemic toxicity were noted in any of the patients. The largest body surface area burns in these two studies were 3% to 4% and 6% to 11% of total body skin surface area in a pair of children who were under the mistaken impression that rust remover was suntan lotion (El Saadi et al. 1989). These children were tested for signs of systemic toxicity associated with HF exposure, and none were observed. It should, however, be noted that in the case of the neutralization of DF, it is possible that workers would come into contact with volumes of HF larger than those experienced by those using the HF-containing cleaning solutions. Therefore, in case of dermal exposure to DF neutralents, personnel should be monitored closely for signs of systemic toxicity.

Large spills of DF neutralent are capable of generating airborne concentrations of HF vapor at which the general population, including susceptible individuals, could experience life-threatening adverse health effects or death for single, nonrepetitive exposure. The airborne HF concentrations would depend on numerous incident-specific parameters such as the amount of DF neutralent spilled and the meteorological conditions at the time of the spill. Therefore, utmost care should be taken to avoid exposure to vapor from DF neutralent in the case of an accidental spill.

QL Neutralent and Its Components

KB

Our acute oral toxicity and mutagenicity results for KB indicate low toxicity to humans unless a significant amount is ingested. The LD₅₀ value of 940 mg/kg is similar to a single dose range–finding LD of 1070 mg/kg (Smyth et al. 1954). Results of the GLP/OPPTS-compliant ecotoxicity tests suggest that releases of KB into aquatic environments can result in harm to aquatic life, although at the amounts present in QL neutralent, KB appears less harmful than MP.

MP

The acute oral toxicity and mutagenicity results for MP indicate that MP is of minimal toxicity to humans. Results of the GLP/OPPTS-compliant ecotoxicity tests indicate that MP can harm aquatic life at relatively low concentrations; QL neutralent would have to be diluted by a factor of 7600 in order for the MP concentration to drop below the D. magna EC₅₀. Testing of the neutralent mixture revealed that the mixture is very minimally toxic; both oral and dermal acute toxicity tests resulted in LD₅₀ values greater than 5000 mg/kg (the largest dose tested).

QL Neutralent Mixture

The test results for QL neutralent are consistent with the test results for the individual components of the neutralent. QL neutralent would be considered at most slightly toxic via ingestion (which is not a significant exposure pathway for a waste material) and dermal exposure. A 5000-gallon spill of QL neutralent would contaminate 64 acre-feet of water to the D. magna LC₅₀; this corresponds to an entire tanker truckload spilled into a moderate to large farm pond.

SUMMARY

The results of this study have allowed for a more complete toxicological evaluation of acute exposure to DF and QL neutralents and have contributed to a much improved understanding of the toxicological characteristics of DF and QL neutralents and their components. The risks from DF neutralent stem primarily from its HF content. QL neutralent is unlikely to cause toxic effects in humans under dermal or inhalational exposure scenarios resulting from hazardous waste handling and transportation, and only large releases into aquatic environments, e.g., catastrophic rupture of a tanker, are expected to results in possible harm to aquatic life.