National Advisory Committee on Acute Exposure Guideline Levels for Hazardous Substances [NAC]

This interim report presents the committee’s conclusions and recommendations for improving NAC’s AEGL [acute exposure guideline level] documents for 25 chemicals: allyl alcohol, bis-chloromethyl ether, chloromethyl methyl ether, bromine pentafluoride, bromine trifluoride, chlorine pentafluoride, carbon tetrachloride, chloroform, chlorosilanes (26 selected compounds), epichlorohydrin, formaldehyde, hydrogen bromide, hydrogen iodide, methyl bromide, methyl chloride, nitric acid, nitric oxide, nitrogen dioxide, nitrogen tetroxide, piperidine, titanium tetrachloride, toluene, trimethylbenzenes (1,2,4-; 1,2,5-;and 1,3,5-TMB), vinyl acetate monomer, and vinyl chloride.

At its meeting held on June 15-18, 2010, the committee reviewed the TSDs [technical support documents] on chlorine pentafluoride, bromine pentafluoride, and bromine trifluoride. Presentations on the TSDs were made by Heather Carlson-Lynch, of Syracuse Research Cooperation.

The committee observed that the AEGL [acute exposure guideline level] values for the three halogen fluorides are linked with each other and three other compounds, hydrogen fluoride, chlorine trifluoride, and chlorine dioxide, chemicals for which TSDs have already been published (NRC 2004, 2007). Thus, on the basis of the review of the TSDs at the meeting (see details below), and excerpted analyses below on related compounds, the committee strongly recommends publication of the halogen fluorides as a single document with chlorine trifluoride, chlorine dioxide, and hydrogen fluoride as appendixes or possibly republishing chlorine trifluoride and chlorine dioxide from Volume 5 of Acute Exposure Guideline Levels for Selected Airborne Chemicals and hydrogen fluoride from Volume 4 as chapters, as well as chapters on chlorine pentafluoride, bromine pentafluoride, and bromine trifluoride. Alternatively, it should be ensured that references are made throughout the document to hydrogen fluoride, chlorine trifluoride, and chlorine dioxide.

Regardless of which approach is chosen, an expansion of the analysis below showing the dissociation paths of the different agents to explain the relative toxicities is important to understand the toxicities of these agents and should be provided in whatever document or documents are developed. This information belongs in Section 4 of the TSD, Special Considerations, and should also be included in the TSD’s Executive Summary.

The following are excerpts from the TSD on chlorine trifluoride (NRC 2007) and are provided as the basis for the discussion of the dissociation paths and relative toxicities below:

• Chlorine trifluoride (ClF3) is unstable in air and rapidly hydrolyzes to hydrogen fluoride (HF) and a number of chlorine-containing compounds, including chlorine dioxide (ClO2). The toxic effects of ClF3 are due, at least in part, to the actions of both HF and ClO2.

• In the moist respiratory tract, ClF3 is predicted to hydrolyze to ClOF, which further degrades to ClO2F and ClF (Dost et al. 1974). ClO2F rapidly hydrolyzes to ClO2, HF, and ClOx anions; the first two
products predominate and are thought to be responsible for ClF3 toxicity, as the ClOx anions are relatively nontoxic.

• The chemical reactivity of the halogenated fluorine compounds in order of decreasing reactivity is chlorine pentafluoride (ClF5) > ClF3 > bromine pentafluoride (BrF5) > iodine heptafluoride (IF7) > chlorine monofluoride (ClF) > bromine trifluoride (BrF3) > bromine monofluoride (BrF) (Bailey and Woytek 1994).

• In the monkey, ClF3 is slightly less toxic than ClF5 but 7 times more toxic than HF. (In all three species for which data are available, ClF5 is almost exactly 10 times more toxic than HF.) In the rat and
mouse, ClF3 is approximately 4 times more toxic than HF.

On the basis of these observations, the committee recommends that the following discussion, suitably modified and expanded as appropriate, be included in any document or documents developed for the halogen fluorides. ClF5, ClF3, BrF5, BrF3, HF, and ClO2 are toxicologically related, and all produce the toxic effect at the point of absorption which is primarily related to the agent’s physical form (vapor, mist, and aerosol). The relative toxicities of these agents are

ClO2 > ClF5 > ClF3 > BrF5 > BrF3 > HF.

These toxicities could be expressed in terms of HF equivalents. ClF3 is approximately 7 times more toxic than HF, and ClF5 is approximately 10 times more toxic than HF. The relative toxicities indicate that ClO2, an intermediate in the dissociation of ClFx, plays a role in the toxicity of these agents. (In the moist respiratory tract, ClF3 is predicted to hydrolyze to ClOF, which further degrades to ClO2F and ClF [Dost et al. 1974]. ClO2F rapidly hydrolyzes to ClO2, HF, and ClOx anions; the first two products predominate and are thought to be responsible for ClF3 toxicity, as the ClOx anions are relatively nontoxic.) If a similar path exists for bromine to form BrO2, it is expected to be less toxic than ClO2, as BrO2 is less reactive
than ClO2.

Because the toxicity data for the individual chemicals are sparse, each chemical is compared with ClF3, and the AEGL values are derived via or supported by the comparison, the Summary and Sections 2, 3, and 4 and much of Sections 5, 6, 7, and 8.3 should be straightforward to develop. Descriptions of the toxicity of ClF3 should be reduced to cross-references to the relevant sections in the appendixes. Other redundancies could likewise be reduced. This consolidation would also strengthen the material in Section 8.3., as the larger data set generated by including all the halogen fluorides provides greater confidence. The other sections and the appendixes could be structurally awkward, so consolidation will be needed.

References to the ClF3, ClO2, and HF documents will need to be rechecked as the three halogen fluoride documents are combined into one document. The summary table of AEGL values could either be a table for each compound or a table for each AEGL, the rows being the separate compounds.

As noted below in the section Comments Pertaining to All TSDs, better justification is needed for reducing the intraspecies factor to 3 for direct-acting irritants.

Below are comments on the specific halogen fluorides discussed at the June meeting.

Chlorine Pentafluoride

The following is excerpted from the Executive Summary of the TSD:

Chlorine pentafluoride (ClF5) is a strong oxidizer that was once considered for use as a missile propellant. No human data were available for development of AEGL values. . . . The AEGL-1 is based on empirical data as well as analogy with hydrogen fluoride (HF) and chlorine trifluoride (ClF3). The empirical data point is a no-observed effect level for the endpoint of irritation of 3 ppm for 10 minutes in the rat. . . . The sensory irritation and reversible mild lung congestion observed in monkeys, rats, and mice following exposure to 30 ppm for 10 minutes, 20 ppm for 30 minutes, or 10 ppm for 60 minutes or following exposure of dogs to 30 ppm for 10 minutes meets the definition of the AEGL-2. . . . The AEGL-3 values are based on a lethality study with rats.

Specific Comments

AEGL-1

The TSD for ClF5 proposes an AEGL-1 value of 0.30 ppm for exposure durations of 10 min, 30 min, and 1 h, but it does not apply the value to exposure durations of 4 h and 8 h. The rationale provided is that the value at those durations is similar to the corresponding AEGL-2 values. The committee recommends not setting any AEGL-1 values, as the sensory warnings are too close to AEGL-2 effects.

AEGL-2

The observation of “severe irritation” and “lung congestion” (on page 21, lines 27-34) are AEGL-2 effects. The observation of “irritation without pathology” (on page 22, lines 9-11) indicates changes
below the definition of the AEGL-2 and, therefore, is suitable as a POD for AEGL-2 values. Using this POD will result in a 1-h AEGL-2 value being similar to the 1-h AEGL-1 value, which reinforces the
recommendation above to not set AEGL-1 values.

AEGL-3

The committee approved the derivation of the AEGL-3 values for ClF5.

Other Comments

Page 17, lines 29-31: “Although most review sources indicate that the reaction with water is violent, both Smith (1963) and Dost and Wang (1970) reported that the reaction with water is slow. (Slow
reaction indicates poor scrubbing in the upper respiratory tract.).” The discrepancy noted in this sentence would benefit from further discussion. Some of the pathology reported in Section 3 indicates, in
accordance with the parenthetical statement, that ClF5 does indeed penetrate to the alveoli, and this information was used in the AEGL-2 derivation. The discrepancy might be resolved in Section 4 of a consolidated TSD that incorporates information on ClF3, HF, and ClO2.

Page 17, lines 42-44: The committee recommends retaining the text that states, “The authors stated that the toxicity of ClF3 is comparable to that of ClO2 on a chlorine equivalent basis and is
comparable to that of HF on a fluorine equivalent basis.” When taken in context of the relative toxicities of ClF3, ClF5, and HF, it adds to the discussion and was reported by the authors. This information (and the citation) belongs in Section 4 with the discussion on relative toxicities and mechanisms described above.

Page 18, lines 4-6: “These observations suggest that the effects of ClF5 exposure may be more likely to be due to the direct irritation of the respiratory tract than to fluoride poisoning.” This is a weak statement. The entire document is based upon direct action at the point of absorption. Can it not be stated that the effects are due to direct irritation of the respiratory tract and not due to fluoride poisoning? See page 24, lines 14-15. See also similar comments on BrF5.

Page 19, lines 36-47, and page 20, lines 1-4: The revised section still does not provide a clear basis for the statement that concentration is more important than duration of exposure for effects other than irritation. The committee recommends rewriting the section to state that

The data from the MacEwen and Vernot studies indicate that, at least for the direct irritant responses to ClF5, concentration may be more important than exposure duration. However, for the other effects observed, the role of exposure duration versus concentration is difficult to interpret because these studies provided few qualitative and quantitative details of the pathology findings. Discordant findings could be due to the dissociation to other agents or to a metabolic pathway.

Page 19, Section 4.4.2: The two paragraphs on susceptibility are found in each of the three halogen fluoride documents. The committee recommends keeping both paragraphs (some of the
documents have one or both paragraphs deleted), as the information is relevant to all three compounds.

Page 20: The deletion of section Concurrent Exposure Issues would indicate that no relevant data are available (see Standing Operating Procedures [NRC 2001]).

Page 24, lines 13-17: The discussion of the relative toxicities of ClF3, ClF5, and HF should be moved to Section 4 (Special Considerations).

Page 24, line 21-23, and Table 13: In this section, the AEGL values for ClF5 are compared with those for ClF3 and HF. The AEGLs reported for ClF3 in Table 13 are inaccurate and should be updated
with the final published values (NRC 2007). The accompanying paragraph should be revised accordingly. (Specifically, the paragraph should note that the AEGL values for ClF3 are lower than those for HF. There should also be discussion about the reason for the two compounds being more similar for longer-duration AEGLs than for shorter-duration AEGLs, including the fact that the relative toxicities of the compounds should be the same if tissue destruction is the end point, unless the saturation point has been reached and toxicokinetics become the driving factor.) The committee also recommends that a table of the AEGL values for ClO2 be added to the TSD for completeness. ClO2 is a breakdown product of ClF5 and ClF3 and probably accounts for why the two halogen chlorides are much more toxic than HF in terms of HF equivalents. The revised paragraphs comparing AEGL values for ClF5 with those for ClF3, HF, and ClO2 should be moved to Section 4 (Special Considerations) rather than appear in Section 8 (Comparison with Other Standards and Guidelines), which should only consider values for ClF5.

Page 25, Section 8.3: This section on data adequacy and research needs should be rewritten according to guidance in the Standing Operating Procedures (NRC 2001).

Bromine Pentafluoride

The following is excerpted from the Executive Summary of the TSD on BrF5:

Bromine pentafluoride (BrF5) is a strong oxidizing chemical that is used as a fluorinating agent and as an oxidizer in rocket propellant fuels. No data on human exposures were available. A single study provided information on lethal and non-lethal values for the rat. . . . In the absence of empirical data, no AEGL-1 values were developed. . . . In the absence of data relevant to derivation of AEGL-2 values for BrF5, data for the structurally-related chemical, chlorine pentafluoride (ClF5), were used. . . . The AEGL-2 values for ClF5 are based on a series of exposures with four species. . . . The AEGL-3 values for BrF5 are based on the highest non-lethal value in the rat study of Dost et al. (1970), 500 ppm for 40 min.

Page 6, lines 35-37: Time-scaling for BrF5 is based on a revised ClF5 time-scaling factor. Footnote b of Table 3 (on page 7) should acknowledge that by noting that the 4-h and 8-h values were time-scaled from the 60-min value.

Page 9, lines 35-39, and page 10, line 2: Is this statement attributable to Darmer (1971) or to the NAC? If the NAC, the statement should be removed, as it is speculation.

Page 10, lines 4-8 and 12-13: The sentence on lines 4-8 should indicate the compound to which the rats were exposed, and the sentence on lines 12-13 should specify the concentration of BrF5. Unlike ClF5, in this study of BrF5, Dost et al. (1968) reported fluoride in the bones and other organs. The TSD should build a case for no systemic effects from fluoride.

Page 10, lines 12-13: The statement that systemic effects are unlikely is appropriate, but a citation is needed to support it, especially since the preceding paragraph discusses the systemic
distribution of fluoride as a result of BrF5 exposure.

Page 10, lines 20-22: The discussion of relative chemical reactivity of halogenated fluorine compounds should be expanded when the TSDs on ClF5, BrF5, and BrF3 are combined. The Bailey and
Woytek (1994) study should be reviewed for information on specific relative toxicities.

Page 12, Section 4.4.2: The two paragraphs on susceptibility are found in each of the three halogen fluoride documents. The committee recommends keeping both paragraphs (some of the
documents have one or both paragraphs deleted), as the information is relevant to all three compounds.

Page 16, Section 8.3: This section includes descriptive statements of the data used without assessment of data adequacy or of what, if any, additional research would be useful to improve the AEGLs. See Standing Operating Procedures (NRC 2001, page 53-57) for requirements.

Bromine Trifluoride

The following is excerpted from the Executive Summary of the TSD on BrF3:

Bromine trifluoride (BrF3) is an extremely reactive and corrosive oxidizing agent used in nuclear reactor fuel processing; as a fluorinating agent; and, potentially, in rocket and missile fuels. . . . In the absence of empirical information on BrF3, AEGL values were based on the chemical analogue, chlorine trifluoride (ClF3). . . . The AEGL-1 values for ClF3 are based on slight irritation as evidenced by rhinorrhea (nasal discharge) observed in two of two dogs during the first 3 hours of a 6-hour exposure to an average concentration of 1.17 ppm. . . . The AEGL-2 values for ClF3 were based on signs of irritation (salivation, lacrimation, rhinorrhea, and blinking of the eyes) in two of two dogs exposed to a concentration of 5.15 ppm for 6 hours. . . . Lethality data for ClF3 (1-hour LC50 values [concentrations of a substance that is lethal to 50% of test organisms in a given time]) were available for the monkey, rat, and mouse. . . . The AEGL-3 values were based on the highest 1-hour concentration that resulted in no deaths in monkeys.

No appendixes were included in this TSD.

Page 6, line 21: A study reporting “obvious” lacrimation in dogs, which was used to derive AEGL-1 values, should not be characterized as mild and transient. The somewhat late onset of the
obvious lacrimation might have been due to a mechanism-based delay (e.g., the main responsible chemical species might have been a metabolite or dissociation product and not BrF3 itself) or to an
oversight of the onset at an earlier time point. Regardless, it was “obvious” and not mild when it was observed. Obvious lacrimation should be considered an AEGL-1 effect and not as a no-observed-adverseeffect level (NOAEL) for AEGL-1 (see Standing Operating Procedures [NRC 2001, page 41]).

Page 9, lines 13-14 and 24-25: The statement that systemic effects are unlikely is appropriate, but a citation is needed to support it.

Page 11, Section 4.4.2: These two paragraphs are found in each of the three halogen fluoride documents. The committee recommends keeping both paragraphs (some of the documents have one or both paragraphs deleted), as the information is relevant.

Page 13, line 37: The committee recommends using “lesser toxicity” rather than “lower toxicity”when comparing BrF3 and ClF3, as the latter description might be misinterpreted.

Page 17, Section 8.3: The Section states that there were no BrF3 data. The inference is that structure-activity relationships are adequate to derive AEGL values using data from ClF3 and other
halogen fluorides and HF and that no further research is needed. If that is the case, then an explicit statement to that effect should be made in this section. See Standing Operating Procedures (NRC 2001, pages 53-57) for requirements.

Comment References

Bailey, W.I., and A.J. Woytek. 1994. Fluorine compounds, inorganic (halogens). Pp. 342-355 in Kirk-Othmer Encyclopedia of Chemical Technology, Vol. 11, 4th Ed., Vol. 11. New York: John Wiley & Sons.

Darmer, K.I. 1971. The acute toxicity of chlorine pentafluoride. Pp. 291-300 in Proceedings of the 2nd Annual Conference on Environmental Toxicology, 31 August, 1 and 2 September 1971. AMRL-TR-71-120. Aerospace Medical Research Laboratory, Wright-Patterson Air Force Base, OH.

Dost, F.N., and C.H. Wang. 1970. Studies on Environmental Pollution by Missile Propellants. AMRL-TR-69-116. Aerospace Medical Research Laboratory, Wright-Patterson Air Force Base, OH. January 1970.

Dost, F.N., D.J. Reed, A. Finch, and C.H. Wang. 1968. Metabolism and Pharmacology of Inorganic and Fluorine Containing Compounds. AMRL-TR-67-224, AD 681 161. Aerospace Medical Research Laboratory, Wright-Patterson Air Force Base, OH. August 1968.

Dost, F.N., D.J. Reed, T.D. Cooper, and C.H. Wang. 1970. Fluorine distribution in rats following acute intoxication with nitrogen and halogen fluorides and with sodium fluoride. Toxicol. Appl. Pharmacol. 17(3):573-584.

Dost, F.N., D.J. Reed, V.N. Smith, and C.H. Wang. 1974. Toxic properties of chlorine trifluoride. Toxicol. Appl. Pharmacol. 27(3):527-536.

NRC (National Research Council). 2001. Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals. Washington, DC: National Academy Press.

NRC (National Research Council). 2004. Acute Exposure Guideline Levels for Selected Airborne Chemicals, Vol. 4. Washington, DC: National Academies Press.

NRC (National Research Council). 2007. Acute Exposure Guideline Levels for Selected Airborne Chemicals, Vol. 5. Washington, DC: National Academies Press.

Smith, D.F. 1963. Chlorine pentafluoride. Science 141(3585):1039-1040.