FLUOROCARBON TOXICITY 337 compounds. Accordingly, to evaluate the practical hazard, the first toxicity designs were predicated on the brief exposure to a relatively high concentration that a householder, a refrigeration repairman, or a fire fighter might encounter. The Underwriters' Laboratories designed an inhalation study in which groups of 12 guinea pigs were exposed to a graded series of concentrations of the test compound. At each sampling time (five minutes, 30 minutes, one hour, and two hours) three guinea pigs, if surviving the exposure, were removed from the chamber for ob- servation or pathologic examination of vital organs within a two to ten day period after exposure. In addition to the Underwriters' work on the then new fluorocarbon refrigerants, they also included refrigerants in use at the time, e.g., ammonia and sulfur dioxide. The comparison pro- vided by this inclusion was invaluable in comparing novel refrigerants with materials already in use. Studies on several refrigerants in the decade from 1931 to 1941 led to a classification system which has been used to grade the safety of new compounds. In this system, class 1 com- pounds are the most toxic, causing death or serious injury when inhaled continuously by guinea pigs at 0.5 to 1.0% (by vol.) for five minutes, e.g., sulfur dioxide. Class 2 is comprised of materials judged hazardous (on the same basis) when concentrations of 0.5 to 1.0% (by vol.) are in- haled for thirty minutes, e.g., ammonia and methyl bromide. Classes 3 and 4 are distinguished by an exposure which proves toxic at 2.0 to 2.5% for one hour and two hours, respectively, e.g., carbon tetrachloride and chloroform (class 3), dichloroethylene and methyl chloride (class 4). Class 6 includes materials which during a two-hour exposure at 20% pro- duce no injury, e.g., dichlorodifluoromethane, dichlorotetrafluoroethane, and monobromotrifluoromethane. Classes 4-5, 5a, and 5b are not pre- cisely defined by the above parameters they encompass compounds more toxic than those in class 6 but less toxic than class 4 materials, e.g., trichlorotrifluoroethane and methylene chloride (class 4-5) trichloro- fluoromethane, chlorodifluoromethane, and CO2 (class 5a) ethane, pro- pane, and butane (class 5b). The primary purpose of this system was to gauge the acute inhalation hazard, and it was evident that the fluoroalkanes possess low toxicity on the basis of the above grouping. Subsequent work has confirmed this judgment, and experience in the production and marketing of fluoro- alkanes attests to a low order of toxicity for man. Toxicological data, Underwriters' Class, and Threshold Limit Values (TLV) of the American Conference of Governmental Industrial Hygienists for a number of fluoroalkanes are summarized in Table I.

388 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS As far as threshold limit values (TLV) are concerned, the first indus- trial hygienic standards for airborne contaminants proposed for some fluoroalkanes by Cook (4) in 1945 ranged from 5000 ppm for dichloro- fluoromethane to 100,000 ppm for dichlorodifluoromethane. The values were subsequently lowered to 1000 ppm not because of toxicity but on the premise that good engineering practice demands that no environ- mental contaminant should be allowed to exceed 1000 ppm (on an eight- hour average), an engineering standard rather than an indicator of tox- Table I Inhalation Toxicity of Fluoromethanes Exposure Conc. Fatal U.L. • Group Structure (%) Hours (-k or --) Class TLV b A CHCla 2.0 2 d- 3 50 CHC12F 10.0 1 d- 4-5 1000 CHC1F2 20.0 2 - 5a (1000) CHF • 20.0 2 -- 6 c (1000) B CC14 2.0 2 d- 3 10 CCI•F 10.0 2 -- 5a 1000 CCI•F• 9,0.0 2 -- 6 1000 CC1F,• 20.0 2 -- 6 c (1000) CF4 9,0.0 9, - 6 •' (1000) C CHIC1 2.0 2 q- 4 100 CH•CI• 5.0 2 d- 4-5 500 CHC12F 10 1 q- 4-5 1000 CClaFa 20 2 - 6 1000 • U. L. signifies Underwriters' classification. The higher the value the lower the toxicity. •' TLV = Threshold Limit Value published by the American Conference of Governmental Industrial Hygienists, 1966 Values. Figures in parentheses indicate provisional values. c Based on data from Haskell Laboratory. icity or health hazard. Only carbon dioxide has a TLV greater than 1000 ppm, viz, 5000 ppm. Thus, while the TLV of 1000 ppm is commensurate with the low toxicity of the fluoromethanes, it is primarily a good house- keeping standard. Industrial experience has established its validity and utility. Much of the early toxicity work on the fluoroalkanes was purely taxonomic, i.e., the assignment of individual compounds to categories. Toxicology admits of no foolproof deductive method by which the tox- icity of individual compounds can be determined from chemical struc- tures or properties. This applies especially to the fluorocarbons. In spite of the strength of the C-F bond and its intramolecular stabilizing