Document 4QvpYrvZKMY0zd07axxz5ja1V

Odor as an Aid to Chemical Safety: Odor Thresholds Compared with Threshold Limit Values and Volatilities for 214 Industrial Chemicals in Air and Water Dilution John E. Amooref Olfacto-Labs, PO Box 757, El Cerrito, California 94530, USA Earl Hautala Western Regional Research Center, US Department of Agriculture, Agricultural Research Service, Berkeley, California 94710, USA Key words: odor threshold; threshold limit value; volatility; solubility; distribution ratio; chemical safety. The body of information in this paper is directed to specialists in industrial health and safety, and air and water pollution, who need quantitative data on the odor thresholds of potentially hazardous chemical vapors and gases. The literature, largely unorganized, has been reviewed for 214 compounds and condensed into tables based on consistent units. Data on the volatility, solubility, ionization and water-air distribution ratio at 25 C are included. From the currently recommended threshold limit value (TLV), a safe dilution factor and an odor safety factor are calculated for each compound. The equivalent data are presented for both air and water dilutions of the chemicals. Available data are summarized on the variability of odor sensitivities in the population, and the increased odor concentrations that are required to elicit responses from persons whose attention is distracted, or who are sleeping. This information is reduced to calibration charts that may be used to estimate the relative detectability, warning potential and rousing capacity of the odorous vapors. Each compound has been assigned a letter classification, from A to E, to indicate the margin of safety, if any, that may be afforded by the odor of the compound as a warning that its threshold limit value is being exceeded. INTRODUCTION The human sense of smell, although not as acute as that of some other mammals and certain insects, can be a valuable source of information about chemicals in the environment. The nose is exceedingly sensitive to certain repulsive smelling compounds, produced in trace amounts by patho genic or putrifying bacteria and molds, such as methyl mercaptan, trimethylamine, 1-pyrroline and isovaleric acid. Although these chemicals themselves are generally harmless to man in the concentrations occurring naturally in air, water or food, heightened odor sensitivities to them may have developed from the protection offered against dangerous or fatal infection or food poisoning. With the advent of the industrial revolution, persons have been exposed to diverse chemicals, many of which are commonly found in workplace settings at concentrations much higher than occur naturally. Some of these pose an inherent risk to health at certain concentrations. In recognition of this potential hazard, the American Con ference of Governmental Industrial Hygienists (ACGIH) publishes an annual listing of Threshold Limit Values (TLV).1 (TLV is a registered trademark of ACGIH, whom we thank for permission to use the TLV designation in this paper.) The TLV used in this paper is the time-weighted average value. Based on the best available industrial health data, it is defined as the time-weighted average concentra tion for a normal 8-h work-day and a 40-h work-week, to f Author to whom correspondence should be addressed. which nearly all workers may be repeatedly exposed, day after day, without adverse effect. The actual concentrations of specific chemicals in the working environment can be sampled and analyzed by various chemical and instrumental means, to determine whether the TLV is being exceeded. The necessary equip ment, however, is often expensive, cumbersome and slow, and requires professional skills to operate and interpret. Nevertheless, there is a little-considered alternative, the human nose, that could serve as a first-line warning system for hazardous concentrations of many chemical vapors. The nose is perfectly placed to sample the inspired air, monitors rapidly and continuously, and may even exceed the sensi tivity of the best instruments. It is, however, at best only semi-quantitative, and it requires calibration to determine its sensitivity to those chemicals that are of importance in industrial practice. In this regard, it is necessary to evaluate the increased concentration of a compound that may be required to alert the average person to the presence of an odor, while engaged in another activity which requires attention. The typical variability of the population for odor sensitivity and responsiveness should also be taken into consideration. METHODS Literature search for basic data A search was conducted for the olfactory and physiochemical characteristics of all volatile compounds and gases listed CCC-02 60-4 3 7X/8 3/0003-02 7 2 509.50 272 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 Wiley Heyden Ltd, 1983 ODOR AS AN AID TO CHEMICAL SAFETY in the Threshold Limit Values1 for 1982. The first objective was to find literature values for the odor-detection thresholds, measured by dilution in either air or water. Dilution of odorants in air can be achieved either dynami cally, by adding a calibrated flow of odorant vapor to an air-stream, or statically, by dispersing a known amount of odorant in a vessel or chamber. In the water-dilution procedure, the odorant is prepared as a series of aqueous dilutions in closed, partially filled vessels from which the head-space vapors can be sniffed. Previous reviews include those of Laffort,2 Patte et al,,3 van Gemert and Nettenbreijer,4 van Gemert,s Fazzalari6 and the ACGIH Documen tation of Threshold Limit Values.1 In practically every case, we consulted the original articles, so as to minimize errors of transcription, calcula tion or duplication. Nearly all of the odor thresholds and references are available in the recent comprehensive com pilations by van Gemert.4'5 If an author gave only a recognition threshold, this was accepted, because recognition of an odor requires on average only about three times the detection threshold concentration.8 If, for any compound, an odor threshold could be located, then a further search was conducted for relevant physical data. The molecular weights, liquid densities and ionization constants (of acids and bases) for these common compounds can be found in laboratory handbooks. The vapor pressures at 25 C were usually interpolated by linear regression computations from the tables of Stull.9 Solu bilities in water at 25 C were often interpolated graphically from data collected by Seidell and co-workers.10,11 More current information is given in Verschueren's handbook.12 Certain missing data on vapor pressures, solubilities, ioniza tion constants, and also occasionally data on the air-water partition coefficient, were found in Beilstein's Handbuch13 and its four supplements. A few solubilities were estimated by extrapolation of homologous series or by comparison with isomers. The air-water partition coefficient describes the relative distribution of a chemical in this two-phase system. Quanti tatively, it is the ratio of the concentrations of the chemical in air and water (both expressed as g l'1) at equilibrium. For compounds of finite water solubility, the coefficient Table 1, Literature odor thresholds for n-butyl alcohol3 Water-dilution threshold Air-dilution threshold Original data 9 I'1 0,005% (v/v) 1 mg/I 1.00 ppm (w/v) 2.5 ppm (v/v) 0.50 ppm (v/v) 4.03 X10'3 1.00X10--' i .oo x i o-5 2.01 X10'3 4.02 XIO'4 2.0 mg/ kg 3.6X10_,,M/I 2.77 ppm (w/v) 2.00 XI0-3 2.67 X10_I 2.77X10 3 6.5 X 10~J g/1 6.50 X10"1 Original data 1 Mg/I 0.565X10'* mol I-1 0.000223 mg/I ActJS = 6 X10"6 Act3T = 7.OX 10"* 1 5 ppm (v/v) Act 25 = 5X10`S 33 mg/m3 1.10 X1013 mol/cc 1.2 mg/m3 0.013 mg/I -- |og10 M/I = 7,91 0.30 ppm (v/v) 3.16 ppm (v/v) 62 ppm (v/v) 0.0231 mmHg 0.390 ppm (v/v) 2.8X 10'1 ppm (v/v) 3.5 ppm (v/v) log2 ppb = 10.42 g r1 looxio-* 4.18X10-' 2.23 X10" 1.61 XI0"' 4.09X10-* 1.45X10-* 4.56 XI 0~5 3.60X10"' 3.60 XI0"' 7.24X10-' 1.40X10-* 3.30 XIO'* 1.45X10-' 1.34X10-" 1.20 X10' ` 1.30X10'* 9.12 X 10"' 9.11 X10"' 9.60 X 10"* 1.88X10-" 7.20 X 10"' 9.61 X10-* 9.97X10-' 9.23X10'* 1.18 X 10"* 8.50X10-' 2.34 XIO-4 1.06X10'* 4.15X10-* First reference Passy, 1892 Backman, 1917 Jung, 1936 Gavaudan, 1948 Mullins, 1955 Moncrieff, 1957 Scherberger, 1958 Nazarenko, 1962 Rosen, 1962 Baker, 1963 Gavaudan, 1966 May, 1966 Fiath, 1967 Dravnieks, 1968 Khachaturyan, 1969 Corbitt, 1971 Laffort, 1973 Heilman, 1974 Moskowitz, 1974 Moskowitz, 1974 de Grunt, 1975 Hertz, 1975 Lillard, 1975 Piggott, 1975 Dravnieks, 1976 Williams, 1977 Amoore, 1978 Laing, 1978 Punter, 1980 Geometric mean, air-dilution threshold = 2.54 X 10'* g I'1 (W = 29} = 2.54 mg m"3 = 0.835 ppm (v/v) Standard deviation -- x/- 7.14; Standard error = x/-h 1.44 aMW=74.1g; 0,s 0.806 g mi"1; VP,, = 6.99 mmHg; S,, = 73.0 g I'1; air-water partition coefficient at 25 C = 3.6 X 10`* lexpt.}, 3.61 X10'* (calc.). JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. G, 1983 273 J. E. AMOORE AND E. HAUTALA Table 2(a) Air-dilution odor threshold data on 214 industrial chemicals. The numerical data are mostly rounded off to two significant figures. Note that ppm on this half of Table 2 are in v/v units QtI 1_1) for the gaseous chemical in air dilution. See Methods for further explanation of each column. TLVs are reproduced from Ref. 1 (1982) with permission from ACGIH Substance Acetaldehyde Acetic acid Acetic anhydride Acetone Acetonitrile Acetylene Acrolein Acrylic acid Acrylonitrile Allyi alcohol Ally 1 chloride Ammonia /7-Amyl acetate sec-Amyl acetate Aniline ' Arsine Benzene Benzyl chloride Biphenyl Bromine 0 t) Bromoform 1,3-Butadiene Butane 2-Butoxyethanol n-Butyl acetate n-Butyl acrylate n-Butyl alcohol sec-Butyl alcohol ferr-Buty! alcohol n-Butylamine n-Butyl lactate n-butyl mercaptan p-fert-Butyltoluene Camphor Carbon dioxide Carbon disulfide Carbon monoxide Carbon tetrachloride Chlorine Chlorine dioxide a-Ch loroacetophenone Chlorobenzene Chlorobromomethane Chloroform Chloropicrin 0-Chloroprene o-Chlorotoluene m-Cresol rrans-Crotonaldehyde Cumene 1 Threshold limit value (ppm; v/v) 100 to 5 750 40 140 000` 0.1 10 2 2 1 25 100 125 2 . >. 0.05 Oarr10 1 0.2 0.1 0.5 1000 800 25 150 10 50 100 100 5 5 0.5 10 2 5000 10 50 5 1 0.1 0.05 75 200 10 0.1 10 50 5 2 50 2 Volatility at 25C (ppm; v/v) g 20 000 6700 290 000 120 000 9 360 000 5800 140 000 33 000 480 000 g 5200 9200 630 g 120 000 1600 11 270 000 8000 9 9 1300 16 000 7100 9200 23 000 55 000 93 000 590 -49 000 850 450 9 470 000. 3' 140 000 9 9 9.9 15 000 190 000 250 000 34 000 290 000 4700 180 -- 41 OOO 5900 3 Air odor threshold (ppm; v/v) 4 Standard error 0.050 0.48 0.13 13 170 1.7 1.5 ' 1.1 1.6 2.8 620 0.16 0.094 17 i.i 2.8 1.5 2.4 1.3 1.2 5.2 0.054 0.0020 1.1 2.5 2.0 2.1 1.6 0.50 12<- 0.044 0.00083 0.051 _ 1.6 4^ 1.1 2.2 1.3 1.6 2700 0.10 0.39 2.3 2.5 1.4 2.5 0.035 0.83 2.6 47 1.8 5.3 1.4 2.0 2.6 2.5 7.0 0.00097 5.0 0.27 74 000 - 1.4 1.9 1.5 0.11 100 000 96 0.31 9.4 ' 1.9 10 1.8 1.8 1.6 0.035 0.68 400 85 0.78 1.1 1.6 1.7 1.4 15 0.32 0.00028 0.12 0.088 7.9 1.5 2.4 1.1 2.9 5 Safe dilution factor 6 Odor safety factor 7 Odor safety class 10 000 2000 1300 390 3000 2000 21 39 57 0.23 A C B B D 7 3 600 000 580 72 000 16 000 480 000 40 000 52 74 310 230 0.61 110 0.12 * 18 0.84 4.8 1800 61 000 1.9 B D B E c D C A A c 20 000 000 12 000 1600 56 2 700 000 ,--------- ,. j.3*0 0.10 0.85^ 23 240 2.0 16 000 1000 1300 52 no 0.39 640 0.29 250 390 E ^=r.-0 __ B c D A D B B 720 180 230 550 19 000 290 60 38 2.1 2.7 B B B c c 120 97 000 85 230 200 0.71 510 2.0 7.3 0.067 D B C c E 47 000 20 000 29 000 1 000 000 10 000 000 92 0,00050 0.052 3.2 0.011 B E E C E 200 200 940 25 000 340 000 1.4 110 0.50 0.12 0.13 C B D E E 29 000 94 36 20 000 120 0.68 150 17 000 17 570 D B A C A 274 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 8. 1983 ODOR AS AN AID TO CHEMICAL SAFETY ical in air 182) with 7 Odor safety class A c B B D 8 D B E c D C A A C c B C O A D R B B B c C D c c E E E D A c A Table 2(b) Water-dilution odor threshold data on the same 214 chemicals. Note that ppm on this half of Table 2 are in w/' units (mg 1 ) for the chemical in aqueous solution. The numerical values in Table 2 are almost invariably com piled, averaged, re-calculated or extrapolated from the literature, and are not new experimental determinations Substance Acetaldehyde Acetic acid (A/4.7) Acetic anhydride Acetone Acetonitrile Acetylene Acrolein Acrylic acid (A/4.31 Acrylonitrile Ally 1 alcohol i Ally! chloride 1 Ammonia (B/9.2) /r-Amyl acetate sec-Amyl acetate Aniline (B/4.6) Arsine Benzene Benzyl chloride Biphenyl Bromine 89 Water TLV equivalent (ppm; w/v) 67 2000 d 1100 70 Solubility at 25 "C (ppm; w/v) 00 os d oo (150) 0.066 1.1 26 1000 200 000 oo 73 000 09 (0.0075) 7.1 68 110 120 3600TM 280 000 1800, ,, 1700 37 000 (0.000035) (0.15) 0.28 0.12 0.012 670 1800 460lo 6.7 33 000 10 Water odor threshold (ppm; w/v) 0.034 97 d 20 300 (0.67) 0.11 9.1 14 (0.0089) 1.5 0.037 0.0017 65 (0.00035) (0.17) 0.012 0.00050 0.0063 11 Molecular weight (g) 12 Density at 20-25C (g ml"') 13 Water-air distribution ratio (w/v) 14 Number of thresholds performed air water 44 0.79,. 370 63 60 1.05 82 000 14 4 102 1.08 d 2- 58 0.79 620 20 8 41 0.78 1000 3- 26 9 56 0.84 72 1.05 53 0.80 58 0.85 1.0 290 240 5600 2- 71 ~ 22 4" 76 17 130 130 93 78 78 127 154 160 0.94 9 0.88 0.87 1.02 9 0.88 1.10 s 3.12 2.4 400 130 160 16 000 0.22 4.6 55 95 19 211 2 54 -1 91 1- 19 4 2- -1 4-- , Bromoform 1,3-Butadiene Butane 2-Butoxyethanol n-Butyl acetate 0.20 (0.88)[0.051) 65 3100 850 61 oo 6800 0.51 (0.0014) (0.17) 0.17 253 54 53 118 116 2.89 9 9 0.90 0.88 38 0.40 0.027 91 41 6- 4- 93 /r-Butyl acrylate n-Butyl alcohol sec-Butyl alcohol fert-Butyl alcohol n-Butylamjne (8/1 0.6) n-Butyl lactate n-Butyl mercaptan (A/10.8) p-fe/T-Butyltoluene Camphor Carbon dioxide (A/6.4) Carbon disulfide Carbon monoxide Carbon tetrachloride Chlorine Chlorine dioxide (A) ar-Chloroacetophenone Chlorobenzene Chlorobromomethane Chloroform Chloropicrin /3-Chloroprene o-Chiorotoluene , m-Cresol (A/10.1) rrans-Crotonaldehyde Cumene 2.2 420 730 620 17 1BOO,0 73 000 200 000 oo oo 370 (0.0061) (0.064) 7.5 (7.51 (0.036) (0.0013) (0.027) (0.0065) 0.0071 42 000 600,0 ~5.5 1700,, 1400 1700 . 26 . 770 6300 87 000 ,, dd 5.5 1100 17 -16 000 (0.28) 7100 (0.0048) 1600 (0.016) (1.1) 640 7.2 (0.45) 480 100w 23 000 150 000,,, 53 0.0078 7.1 19 290 6.2 128 74 74 74 73 520 (0.000012) (0.032) 1.0 (110) 146 90 148 152 44 (0.00039) (2.7) (0.52) (0.0020) 0,67 * 76 28 154 71 67 d 0.050 34 (2.4) (0.0371 155 113 129 119 164 (0.024) (0.0069) 0.037 0.42 (0.00080) 88 127 108 70 120 0.90 0.81 0.81 0.78 0.73 0.98 0.84 0.86 s g 1.26 9 1.59 9 9 s 1.10 1.93 1.48 1.65 0.96 1.08 1.03 0,85 0.86 43 2800 2400 2000 1100 21 20 9 51 41 32 12 000 3.3 1.1 600 0.83 - 6- 93 2-- 1.2 6 - 0.023 2- 0.85 10 1 2.2 7 - 26 1 d 16 16 5.7 7.1 2- 62 -- 14 1 1 0.45 4.1 29 000 1200 1.8 21 11 33 11 61 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 275 Table 2(a)--Continued Substance Cyclohexane Cyclohexanol Cyclohexanone Cyclohexene Cyclohexytamine Cyclopentadiene Decaborane Diacetone alcohol Diborane o-Dichlorobenzene p-Dichlorobenzene frans-1,2-Dichloroethylene f3,0'-Dich|oroethyl ether Dicyclopentadiene Diethanolamine Diethylamine Diethylaminoethanol Diethyl ketone Diisobutyl ketone Diisopropylamine /V-Dimethylacetamide Dimethylamine /V-Dimethylaniline A/'Dimethylformam ide 1,1 -Dimethylhydrazine 1,4-Dioxane Epichlorhydrin Ethane Ethanolamine 2-Ethoxyethanol 2-Ethoxyethyl acetate Ethyl acetate Ethyl acrylate Ethyl alcohol Ethylamine Ethyl o-amyl ketone Ethyl benzene Ethyl bromide Ethyl chloride Ethylene Ethylenediamine Ethylene dichloride Ethylene oxide Ethylenimine Ethyl ether Ethyl formate Ethylidene norbornene Ethyl mercaptan A/-E thy 1 morpholine Ethyl silicate Fluorine Formaldehyde Formic acid Furfural Furfury) alcohol i Threshold limit value (ppm; v/v) 2 Volatility at 25C (ppm; v/v) 300 50 25 300 10 130 000 2000 6000 99 000 15 000 75 0.05 50 0.1 50 -- 560 000 - 110 1600 9 1800 75 200 5 5 3 1200 420 000 1500 3600 78 10 10 200 25 5 310 000 2900 22 000 3300 110 000 10 10 5 10 0.5 2600 9 1000 3100 210 000 25 2 140 0001 3 5n 52 000 21 000 g 780 7100 5n 400 5 1000 10 2700 1 20 000 50 000 75 000 g 25 100 200 1000 140 000' 10 10 1n 0.5 400 3600 13 000 610 000 9 9 16 000 110 000 9 260 000 700 000 t 100 5 0.5 5 10 320 000 710000 11 000 3000 1g r9 5 57 000 2 2100 10 810 3 Air odor threshold (ppm; v/v) 4 Standard error [xM 25 0.15 0.88 0.18 2.6 1.9 0.060 0.28 2.5 0.30 2.8 2.1 2.2 -- - _ _ _ 4.2 0.18 17 0.049 0.0057 0.27 4.1 16 - 1.9 - 0.13 0.011 2.0 0.11 1,8 2.9 -- 2.1 -- 3,9 47 0.34 0.013 2.2 1.7 _ 3.1 3.8 46 5.5 24 0.93 120 000 2.6 2.7 2.4 12 5.9 9.0 0.056 3.9 0.0012 84 0.95 _ 1.8 4.1 1.8 2.6 6.0 2.3 3.1 4.2 290 1.0 88 430 1.5 8.9 _ 2.7 - - 2,6 _ 2.1 1.6 1.3 3.3 31 1.6 0.014 1.4 0.00076 2.0 1.4 18 17 4.9 0.14 0.83 49 0.078 B.O 2.3 1.9 1.7 - 5 Safe dilution factor 430 39 240 330 1500 7500 2300 33 10 000 000 37 17 2100 290 720 26 31 000 290 110 130 21 000 260 100 000 200 370 410 000 1000 11 000 7 260 1400 530 300 10 000 75 100 000 140 . 130 3100 1000 7 1600 11 OOO 1 000 000 520 000 1800 3200 1 400 000 2100 300 1 000 000 1 000 000 11 000 1000 81 . Table 2* 6 Odor safety factor 12 340 28 1600 3.8 40 0.83 180 0.040 160 420 12 100 870 1 ' 77 910 97 230 2.7 0.27 29 400 4.6 0.30 1.1 2.1 1.2 1.2 1.8 89 100 4000 12 11 4.2 44 64 240 490 10 0.11 0.0023 0.32 45 3.3 350 650 3.5 0.57 7.3 1.2 0.10 25 1.2 7 Odor safetyY class c B B A C B D B E 8 B C B A C B A B C D B B C D C c c c c B B A C C C B B B B C E E D B C B A C D C C E C c ; , Substan Cyclohe Cyclohe Cyclohe Cyclohe Cyclohe Cyclope Decabor Diacetoi Diboran 0-Dichlt p-Dichit vans-1 jjjJ'-Dic 1 Dicycloj Dlethan Diethyl; Diethyl; Diethyl Diisobu Diisoprr N-Dime Dimetfv W-Djme W-Dime 1,1 -Dirr 1,4-Dio Epichlo Ethane Ethano 2-Etho> 2-Etho; Ethyl a< Ethy 1 a Ethyl a Ethylar Ethyl n Ethyl b Ethyl t CtitnFyi.. l| c. Cc tinh.yi liaer. Ethylei Ethylei Ethylei Ethylei ctrvyi p ' 1 Ethyl 1 Ethyii< Ethylr W-Eth> Ethyl t 1 Fluorii Formr Fofmi Furfui Furfur 276 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 19*13 7 -lass 3 3 3 3 J ' * ' ; ` : . . . Table 2{b)--Continued 1 i 1 | Substance j Cyclohexane j Cyclohexanol . Cyclohexanone , CycJohexene j Cyclohexylamine (B/10.6) 8 Water TLV equivalent (ppm; w/v| (0.13) 940 240 (0.65) 94 9 Solubility at 25 C (ppm; w/v) 55 36 OOO -54 000 210 j Cyciopentadiene Decaborane | Diacetone alcohol , Diborane 1 0-Dichlorobenzene (0.24) d 3.9 | p-D ic hfo robenzene 1 rrans-1,2-Dichloroethylene ' 0j3'-Dichloroethy( ether 1 Dicyclopentadiene j Diethanolamine (B/8.9) ! j Diethylamine (B/11.0) 1 Dietbylaminoethano) (8/8.8) 1 Diethyl ketone Diisobutyf ketone | Diisopropylamine (B/11.0) 4.7 (3.0) 36 240 000 36 450 3.3 3.5 i A/'Dimethyiacetamide Dimethylamine (B/10.7) A/-DimethyJaniline {B/5.2) A/-Dimethy Ifo rmam id e 1,1-Dimethylhydrazine (B/7.2) 8.6 9.9 -1800 d 140 79 6300 11 000 oo 48 000 430 oo 550 000 2000 oo 1,4-Dioxane Epichlorhydrin Ethane Ethanolamine (8/9.5) 2-Ethoxyethanol 240 6.4 (8.8) 23 000 oo 65 000 60 oo oo 2-Ethoxyethyl acetate Ethyl acetate Ethyl acrylate Ethyl alcohol Ethylamine (B/10.7) 450 270 1.5 9000 45 200 000!o 73 000 15 000 oo I6 Ethyl /j-amyl ketone Ethyl benzene Ethyl bromide Ethyl chloride Ethylene Ethylenediamine (B/10.0) Ethylene dichloride Ethylene oxide Ethylenimine (B/8.0) Ethyl ether ' Ethyl formate Ethylidene norbornene Ethyl mercaptan (A/10.5) /V-Ethylmorpholine (8/ ) Ethyl silicate 10 (1.3) (2.9) (4.7) (19) 0.80 0.33 d .34 - 1500 160 9000 4700 130 oo 8600 270 00O,o cod 56 000 35 100 000 (0.0049) 7000 oo dd Fluorine Formaldehyde Formic acid (A/3.7) Furfural Furfury 1 alcohol d 0.73 170 89 d d 550 000 oo 86 000 od 10 11 12 13 Water odor threshold (ppm; w/v) (0.011) 2.8 8.3 (0.00039) 25 Molecular weight (g) Density at 20-25 0 (9 ml"1) 84 0.78 100 0.95 98 0,95 82 0.81 99 0.87 Water-air distribution ratio (w/v) 0.12 4600 2400 0.64 2300 (0.0060) 64 d 0.024 66 122 116 28 147 0.80 5 0.94 g 1.30 1.2 d 13 0.011 (0.26) 0.36 22 000 147 97 143 132 105 s 1.26 1.21 s 1.10 10 3.8 1200 19 000 000 0.47 4.7 0.014 1.3 73 117 86 142 101 0.71 0.88 0.81 0.81 0.72 1200 640 23 -170 0.29 0.025 50 87 45 121 73 60 0.94 g 0.96 0.94 0.79 460 400 230 3.0 (7.5) 20 000 190 88 1.03 2700 92 1.18 840 30 9 0.051 61 1.02 3 100 000 90 0.93 5.0 2.6 0.00038 760 4.3 132 88 100 46 45 0.97 0.90 0.92 0.79 0.69,5 16 000 180 74 4800 2400 2.5 (0.029) (0.046) (0.019) (0.039) 128 106 109 64 28 0.83 0.87 1.43 9 g 80 2.9 3.3 1.8 0.12 16 000 SO 0.90 7.0 99 1.26 4 140 44 g 170d 43 0.83 0.75 74 0.71 20 180 d 28 11 74 120 (0.0000075) 62 115 d 208 d 0.60 1700 3.5 d 38 30 46 96 98 0.92 0.83 0.90 0.93 9 9 1.22 1.16 1.13 120 3.9 d d 590 18 000 11 000 d 14 Number of thresholds performed air water 6- 32 82 1- 1 1- 112 132 23 2"1 2-- 1-- 61 1-- 3- 121 1- 62 3- 21 2-- 71 2- 2- 121 1- 84 21 13 5 33 1- 23 1 141 11 82 221 7" 11 2-- 12 1 2-- 2 1- 94 45 23 1-- JOURNAL 0 F APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 277 Table 2(a)--Continued Substance Halothane Heptane Hexachlorocyclopentadiene Hexach loroethane Hexane Hexylene glycol Hydrazine Hydrogen bromide Hydrogen chloride Hydrogen cyanide Hydrogen fluoride Hydrogen seienide Hydrogen sulfide Indene Iodoform Iscamyl acetate isoamyl alcohol Isobuty! acetate isobutyl alcohol Isophorone Isopropyl acetate Isopropyl alcohol Isopropy lamine Isopropyl ether Maleic anhydride Mesityl oxide 2-Methoxyethanol Methyl acetate Methyl acrylate Methyl acrylonitrile Methyl alcohol Methylamine Methyl n-amyl ketone A/-Methylaniline Methyl/7-butyl ketone Methyl chloroform Methyl 2-cyanoacrylate Methyicyclobexane c/s-3'Methylcyc)ohexanol Methylene chloride Methyl ethy! ketone Methyl formate Methyl hydrazine Methyl isoamyl ketone Methyl isobutyl carbinol Methyl isobutyl ketone Methyl isocyanate Methyl isopropyl ketone Methyl mercaptan Methyl methacrylate Methyl /7-propyl ketone a-Methyl styrene Morpholine Naphthalene Nickel carbonyl 1 Threshold limit value Ippm; v/v) 2 Volatility at 25 C {ppm; v/v) 50n 400 0.01 10 50 390 000 60 000 78 770 200 000 25 0.1 3 5 10 100 18 000 9 9 970 000 3 0.05 10 10 0.6 9 9 9 2200 -49 100 7100 100 4300 150 26 000 50 16 000 5 450 250 400 5 250 0.25 79 000 57 000 740 000 210 000 -170 15 5n 200 10 1 13 000 16 000 270 000 110 000 88 000 200 10 50 0.5 5 160 000 g 2000 640 5000 350 2 400 50 100 160 000 -530 61 000 710 550 000 200 100 0.2 50 25 1 30 000 760 000 65 000 4800 7800 50 0.02 200 0.5 100 9500 630 000 39 000 9 52 000 200 50 20 10 0.05 21 000 3800 13 000 120 520 000 3 Air odor Threshold (ppm; v/v) 4 Standard error (xM 5 Safe dilution factor 33 150 0.030 0.15 130 _ 1.7 5.1 2.0 . 7900 150 7800 77 4000 50 3.7 2.0 0.77 0.58 - 1.1 -- 2.2 1.9 4.0 180 000 330 000 200 000 97 000 0.042 0.30 0.0081 0.015 0.0050 1.2 1.5 3.9 1.8 330 000 20 000 000 100 000 220 81 0.025 0.042 0.64 1.6 0.20 1.6 1.3 1.8 2.0 - 71 43 170 330 89 2.7 22 1.2 0.017 0.32 2.9 1.8 2.8 - - 320 140 150 000 850 670 0.45 2.3 4.6 0.0048 7.0 26 26 3.5 -- 850 3200 1400 11 000 88 000 100 3.2 0.35 1.7 0.076 2.0 4.6 2,1 -- 800 100 000 40 1300 1000 120 2.2 630 500 250 2.8 - - 1.2 470 260 150 14 5500 . 5.4 600 1.7 0.012 0.070 1.9 2.9 - - 660 7600 330 000 96 310 0.68 2.1 1.9 0.0016 0.083 2,3 - 2.3 2.0 1.9 190 32 000 000 200 2 000 000 520 11 0.29 0.01 0.084 0.30 2.2 4.0 - 1.9 3.3 110 76 670 12 10 000 ooo 6 Odor safety factor 1.5 2.7 0.34 64 0.37 7 Odor safety class C C D B D 0.50 0.027 1.5 6.5 17 71 0.17 1200 690 120 3900 2300 . 230 30 25 D E C c c B E A A 8 A A 8 8 C 93 18 4.1 15 000 0.77 e c c A D 33 2.1 44 2100 0.14 B C B A E 2.0 3.1 140 0.29 66 C c B D B 2.8 0.91 0.63 0.10 0.40 37 0.17 0.12 4200 360 73 0.0094 100 300 1200 C D D E D B E E A B B E B B A 18 170 2000 120 0.17 C B A B E Table : | Substai | Halath; Heptan I Hexact I Hexach I Hexane ^ Hexylei I Hydraz ' Hydrog | Hydrog j Hydrog | Hydrog I Hydrog 1 Hydrog j Indene I lodofor I Isoamyi | Isoamyl I Isobuty j Isobuty ' Isophor I Isoprop | Isoprop Isoprop Isoprop Maleic < J Mesityl 1 2-Metht | Methyl Methyl 1 Methyl j Methyl i Methyl* ^ Methyl | W-Meth- I Methyl j Methyl ( Methyl I Methylc I c/s-3-Mf j Methyli | Methyl j Methyl ; Methyl . Methyl 1 Methyl I - Methyl Methyl Methyl j Methyl j Methyli i Methyl a-Met h Morph* I Naphth j Nickel 278 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. G. 1983 i 7 Odor safety class cc D B D D C C C B E A A B A A B B C B C C A D B C B A E C C B D B C D D E 0 E E A B 3 = 3 3 3 a, 3 Table 2(b)--Continued a 9 Substance Water TLV equivalent (ppm; w/v) Solubility at 25C (ppm; w/v) Haiothane Heptane Hexachlorocyciopentadiene Hexachloroethane Hexane (0.44) (0.020) 0.0026 (0.65) (0.0024) 3400 2.9 20 50 9.5 Hexylene glycol Hydrazine (B/8,5) Hydrogen bromide (A) Hydrogen chloride (A) Hydrogen cyanide (A/9.2) Hydrogen fluoride (A/3.2) Hydrogen selenide (A/3.9) Hydrogen sulfide (A/7.0) Indene Iodoform Isoamyl acetate Isoamyl alcohol Isobutyl acetate Isobutyl alcohol Isophorone Isopropyl acetate Isopropyl alcohol Isopropylamine (B/10.5) Isopropyl ether Maleic anhydride Mesityl oxide 2-Methoxy ethanol Methyl acetate Methyl acrylate Methyl acrylonitrile - d d 3.0 1 200 000 500 000 90 d 19 (0.00035) 6800 (0.036) 3500 (0.18) -40 1.3 110 66 630 34 310 140 1400 26 000 5900 89 000 12 000 97 3000 20 12 d 30 000 OO ocr 10 000 d 35 130 4.5 0.29 29 000 CO 220 000 49 000 25 000 Methyl alcohol Methylamine (B/10.6) Methyl /)-amyl ketone A/-Methy laniline (B/4.8) Methyl /)-butyl ketone Methyl cMoteform Methyl 2-cyanoacrylate Methylcyclohexane c/r-3-Methylcyclohexanol Methylene chloride Methyl ethyl ketone Methyl formate Methyl hydrazine (B/7.9) Methyl isoamyl ketone Methyl isobutyl carbinol Methyl isobutyl ketone Methyl isocyanate Methyl isopropyl ketone Methyl mercaptan (A/10.7) Methyl methacrylate Methyl n-propyl ketone a-Methyl styrene Morpholine (B/8.7) , Naphthalene Nickel carbonyl 1500 7,4 40 5.3 17 550 000 4300 670030 16 000 (2.8) 1300 (0.092) 660 3.6 14 9300 19 000 310 25 56 - 53 210 000 170 000 CO 5400 16000 94 18 000 dd 320 (0.00751 60 000 ~ 14 000 30 15 000 270 7.4 54 000 560 " 2.5 (0.000012) 30 130 10 1 1 Water odor threshold (ppm; w/v) (0.29) (0.0073) 0.0077 (0.010) (0.0064) 160 d d 0.17 Molecular weight (9) 197 100 273 237 86 118 32 81 3627 d 20 (0.0021) 81 (0.000029) 34 (0.00026) 116 0.011 394 0.017 0.27 0.15 10 5.4 130 88 116 74 138 1.0 160 4.9 0,00080 d 1.0 3.0 0.0021 2.1 740 2.4 0.28 18 0.25 102 60 59 102 98 98 76 74 86 67 32 31 114 107 100 (0.97) (0.15) 6600 9.1 8.4 150 0.013 0.15 133 111 98 114 85 72 60 46 114 102 1.3 100 d 57 3.1 86 (0.000024) 48 0.025 100 15 0.043 0.021 (0.000072) 86 118 87 128 171 12 Density at 20-25C (g mr`l 1.87 0.68 1.70 s 0.66 0.92 1.01 g g 0.70 0.96 g 9 1.01 s 0.87 0.80 0.87 0.80 0.92 0.87 0.78 0.69 0.73 s 0.85 0.97 0.93 0.95 0.80 0.79 g 0.81 0.99 0.81 1.34 1.11 0.77 0.91 1.34 0.80 0.97 0.87 0.81 0.81 0.80 0.96 0.80 g 0.94 0.81 0.91 1.00 s 1.32 13 Water-air distribution ratio (w/v) 14 Number of thresholds performed air wate: 1.1 0.012 23 6.7 0.014 1 4 1 2 - 1 1 -- d d 270 1_ 21 1623 d 2.1 2.6 3.7 130 2- 125 1 11 3-- 120 1700 48 2100 4800 92 3000 -1700 11 d 83 53 31 7c 1-- 412 4 21 1_ 1 570 210 130 110 2- 25_ 1-- 1 5600 580 170 2400 800 13 4 23 22 11-- 1.4 0.057 2800 10 311141 530 100 240 510 81 31-- 11 460 d 460 7.6 73 51-- 11 82 41 380 31 47 0.035 21 31 1-- 64 3-- JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6,1983 279 gar r m e" Table 2(a)--Continued Substance Nitrobenzene Nitroethane Nitrogen dioxide Nitromethane 1-Nitropropane 2-Nitropropane m-Nitrotoluene Nonane Octane Osmium tetroxide 1 Threshold limit value (ppm; v/v) 2 Volatility at 25C Ippm; v/v) i 360 100 27 000 39 100 47 000 25 13 000 10" 2 200 300 0.0002 22 000 -280 6000 18 000 12 OOO Oxygen djfluoride Ozone Pentaborane Pentane Perchloroethylene 0.05 0.1 0.005 600 50 9 9 270 000 670 000 25 000 Phenol Phenyl ether Phenyl mercaptan Phosgene Phosphine 5 460 1 29 0.5 2000 0.1 9 0.3 S Phthalic anhydride Propane Propionic acid n-Propyl acetate n-Propyl alcohol 1 140 ooo' 10 200 200 0.67 9 5400 43 000 26 000 Propylene Propylene dichloride Propylene glycol 1-methyl ether Propylene oxide n-Propyl nitrate 140 000' 75 100 20 25 9 69 000 16000 700 000 30 000 Pyridine Quinone Styrene Sulfur dioxide 1.1.2.2-Tetrachioroethane 5 0.1 50 2 5 27 000 130 9600 9 8400 T etrahydrofuran Toluene T oiuene*2,4-disocyanate o-Tolukhne 1.2.4-T rchloroben2ene 200 100 0,005" 2 5 230 000 37 000 -21 330 570 Trichloroethylene Trichlorofluoromethane 1.1.2- Trichloro-1,2,2- trifluoroethane Triethylamine Trimethylamine 50 1000 1000 10" 10" 99 000 9 430 000 93 000 9 1.3.5-T rimethylbenzene Trimethyl phosphite n-Valeraldehyde Vinyl acetate Vinyl chloride 25 3600 2 34 000 50 21 000 10 140 000 59 Vinyiidene chloride Vinyl toluene m-X ylene 2,4-Xylidine 5" 50 100 2 790 000 2400 11 000 190 3 Air odor threshold (ppm; v/v) 4 Standard error (x/*) 0.018 2.1 0.39 3.5 11 1,7 2.6 4.2 70 0.045 47 48 0.0019 2.2 4.1 3.2 -` 0.10 0.045 0.96 400 27 - 1.9 1.9 1.8 0.040 0.0012 0.00094 0.90 0.51 1.5 3.7 4.4 1.7 2.5 0.053 16 000 0.16 0.67 2.6 -- 1.3 1.8 4.1 1.7 76 0.25 10 44 50 0.17 0.084 0.32 1.1 1.5 3.0 4.5 - 1.4 3.0 2.0 1.3 2.1 2.0 2.9 0.17 0.25 1.4 5.4 1.6 2.9 4.1 2.1 28 5.0 45 1.7 - 0.48 2.1 0.00044 1.4 0.55 0.00010 0.028 0.50 3000 1.9 2.5 1.6 3.7 190 10 1.1 0.056 3.7 2.1 - 5 Safe dilution factor 360 270 330 000 470 520 2200 140 30 61 61 000 000 20 000 000 10 000 000 54 000 OOO 1100 490 92 29 4100 10000 000 3 300 000 0.7 7 540 220 130 7 920 160 35 000 1200 5300 1300 190 500 000 1700 1100 370 4200 170 110 2000 1000 430 9300 100 000 150 17 000 420 14 000 200 000 160 000 48 110 97 fable 2(1 6 Odor safety factor 56 46 7.8 29 2.3 0.14 45 4.3 6.3 0.10 0.50 2.2 0.0052 1.5 1.8 130 800 530 0.11 0.58 7 Odor safety class 6 8 C c 6 cB C E D c E C c B BA E D 19 8.8 61 300 78 c c B B B 1800 300 10 0.45 0.50 A B C D D 30 1.2 160 1.7 3.4 B C B c c 99 34 0.030 8.0 3.6 8 B E C C 1.8 200 22 c B C 21 23 OOO c A 45 20 000 1800 20 0.0017 B A A C E 0.027 5.0 92 36 E 0 B B Substanc Krtroben 1 Nitroctht 1 Nitrogen Nitromet 1 1-Nitropi 1 2-Nitropr ^7-Nitroti i1 Nonane Octane | Osmium | 1 Oxygen d i 1 Ozone Pentabort ( Pentane Perchloro ! 1 Phenol ( Phenyl et' ii Phenyl m* Phosgene : ' Phosphim ! Phthalic a Propane ! Propionic rt-Propyl ? rt-Propyl i Propylene Propylene Propylene Propylene /j-Propyl r Pyridine (' Quinone Styrene Sulfur dio 1,1,2,2-Te Tetrahydr Toluene Toluene-2 o-Toluidir 1.2.4-Triel Trichloroe Trich lorof 1,1,2-Tricl trifluor Triethylan Trimethyl. 1.3.5-Trinr Trimethyl n-Valerald Vinyl acet Vinyl chlo VinylidenVinyl tolu m-Xylene 2,4-Xylid. 280 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 Table 2(b)--Continued Substance 8 Water TLV equ ivalent (ppm; w/v) 9 10 Solubility at 25C (ppm; w/v) Water odor threshold (ppm; w/v) 11 Molecular weight (g) 12 Density at 20-25C (g ml'1] 13 Water-air distribution ratio (w/v) 14 Number of thresholds performed air water Nitrobenzene Nirroethane (A/8.4) Nitrogen dioxide (A) Njtromethane (A/10.2) 1-Nitropropane (A/~8) ' 6.0 100 d 260 29 2100 27 000 d 110 000 15 000 0.11 2.2 d 9.1 12 123 75 46 61 89 1.20 1.05 g 1.13 1.00 1200 330 d 1000 310 13 2 -1 6-1 21 2-Nitropropane (A/7.7] /n-Nitrotoluene Nonane Octane Osmium tetroxide (A/12.0) 7.6 3.6 (0.0056) (0.011) 0.0012 16 000 500JO -0.17 0.66 69 000 53 0.080 (0.0013) (0.0017) 0.012 89 137 128 114 254 ' OxVQen difluoride Ozone Pentaborane Pentane Perch loroethylene (0.0000054|d 10030d (O.OOOOII)d 54 (0.00064) 6100 (0.00028) 48 d dd 63 (0.033) 38 10.022) 72 (0.31) 150 (0.17) 166 0.98 1.16 0.72 0.70 s g g 0.63 0.62 1.61 210 320 0.0054 0.0077 580 1 1 2 2 1 0.049d 1 3.2 6 d1 0.019 3 0.90 3 1 -- -- -- - _ -- - 1 Phenol (A/10.0) Phenyl ether Phenyl mercaptan (A/6.5) Phosgene Phosphine Phthalic anhydride Propane Propionic acid (A/4.9) n-Propyl acetate n-Propyl alcohol 1000 150 0.15 d (0.00011) 85 OOO 4300 610 d 370,, d (9.0) 1700 92 1800 d 62 19 O0O,o CO 7.9 94 0.18 170 0.00028 110 d 99 (0.00020) 34 d (1.01 28 0.31 23 148 44 74 102 60 s 1.07 1.08 g 9 s g 1.00 0.89 0.80 52 000 21 000 66 d 0.27 16 6 23 22 66- d 1-- 0.036 2- 56 000 11 2 110 4- 3600 12 5 Propylene Propylene dichloride Propylene glycol 1-methyl ether Propylene oxide n-Propyl nitrate (50) (3.0) 14 7.4 350 2800 370 000 880O,0 (0.028) (0.010) 31 15 42 113 90 58 105 g 1.16 0.92 0.83 1.05 0.21 8.8 300 69 31 1121- Pyridine (B/5.2) Quinone Styrene Sulfur dioxide (A/1.9) 1,1,2,2-Tetrachloroethane 28 11 (1.7) 0.19 1.7 oo 14 000 320 88 000 2900 0.95 9.3 (0.011) 0.11 0.50 79 108 104 64 168 0.98 s 0.90 g 1.60 1700 25 000 7.8 37 50 15 10 21 10 3 13 " 31 Tetrahydrofu ran Toluene Toluene-2,4-diisocyanate o-Toluidine (B/4.4) 1,2,4-T richlorobenzene (1.41 d 91 (0.23) oo 540 d 15000 -26 (0.042) d 11 (0.064) 72 92 174 107 181 0.89 0.86 1.22 1.00 1.45 3.8 d 10 000 6.1 3-- 18 2 431 11 Trichloroethylene Trichlorofluoromethane 1,1,2-Trichloro-l ,2,2- trifluoroethane Triethylamine (8/10.9) Trimethylamine (B/9.7) (0.55) 1100 8.8 71 000 4.5 410 OOO,, (0.31) 0.42 0.00020 131 137 187 101 59 1.46 1.49 1.56 0.73 9 2.1 210 190 71 11- 41 31 1,3,5-Trimethylbenzene Trimethyl phosphite n-Valeraldehyde Vinyl acetate Vinyl chloride Vjnylidene chloride Vinyl toluene m-Xylene 2,4-Xylidine (B/4,9) (0.67) 97 dd 29 12 000 1.8 25 000,,, (0.0057) 1100 (0.041) (2.11 (1.6) 66 6400 - 100 170 6400 (0.015) d 0.017 0.088 (3.4) (1.5) (0.42) (0.017) 1.8 120 124 86 86 62 97 118 106 121 0.86 1.05 0.81 0.93 g 1.22 0.90 0.86 0.97 5.4 d 170 50 0.44 2.0 8.7 3.7 6600 63 1" 13 41 3-- 2182 1- JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. G, 1983 281 J. E. AMOORE AND E. HAUTALA at 25 C can be calculated14 from the vapor pressure and the solubility at 25 C. The coefficients for some of the com pounds that are infinitely soluble in water at 25 C were calculated from tabulated activity data15 or measured experimentally.14 The results for n-butyl alcohol, which has provided the most plentiful odor-threshold data, are given in Table 1 as a demonstration of data reduction. The original threshold data, in a variety of concentration units, were converted2 into common units of g l"1. Any water dilution thresholds were further converted to the equivalent air dilution threshold, through multiplication by the air-water partition coefficient.14 The relationship between odor-intensity sensation and odorant concentration is exponential.16 Therefore, in order to preserve the normal distributions of olfactory-threshold measurements, all chemical concen trations of odorants were calculated on a logarithmic scale. Hence the geometric mean of all 29 odor thresholds, expressed in air dilution, was computed (by converting to the logarithms, finding their arithmetic mean, and taking its antilogarithm).2 The mean air dilution threshold, in g T1, was finally converted to mg m'3, and to ppm by volume. Explanation of Table 2 (odor thresholds) Column 1. Threshold limit values (TLV) adopted by ACGIH, 1982.1 The superscript n indicates that the TLV used is the value proposed in the 1982 Notice of Intended Changes. The superscript i indicates an inert gas (simple asphyxiant) for which no TLV is assigned by ACGIH, merely a requirement that the oxygen content of the air not be reduced below 18%. This would be expected to occur if the asphyxiant reaches 14%, or I40 000pprn, which is in effect the TLV for inert gases. Column 2. The volatility in ppm (v/v) is given by the literature vapor pressure (in mmHg at 25 C) multiplied by 1316 (1 000 000 ppm per 760 mmHg). ~ indicates approxi mate value obtained by extrapolating the linear regression from vapor pressures recorded at substantially higher temperatures, g, gaseous at 25 C. Column 3. Air-dilution odor thresholds are geometric averages of all available literature data, omitting extreme points and duplicate quotations. Odor thresholds originally measured in water dilution were converted to the equivalent air dilution, as illustrated in Table 1 for n-butyl alcohol. Column 4. When two or more acceptable literature thresholds were located, the standard error of their mean was calculated. The standard error is the standard deviation divided by the square root of the number of literature thresholds. This factor is applicable to the data in columns 3, 6 and 10. The smaller the standard error, the greater the confidence that may be placed in the accuracy of the mean threshold value. (It should be borne in mind, however, that a small standard error, based on only two thresholds, could itself be the result of a fairly probable coincidence.) Column 5. Safe dilution factor, for the saturated vapor at 25 C, is the volatility divided by the threshold limit value (column 2 divided by column 1). For substances that are less than infinitely soluble in water, the same safe dilution factor applies to the saturated solution at 25 "C (column 9). Column 6. Odor safety factor is the threshold limit value divided by the odor threshold (column 1 divided by column 3). This factor may be interpreted quantitatively by reference to Fig. 2, in terms of what percentage of attentive persons can detect the TLV concentration, and what percentage of distracted persons will perceive a warning of the TLV concentration. 1 Colum j measui or air c I 1 PI I 1P Column 7. The scale of odor safety classes is explained in Table 3. Class A substances provide the strongest odorous warning of their presence at threshold limit value concen- trations, whereas class E substances are practically odorless at the TLV concentration. Table 3. Odor safety classification Odor safety Class factor Interpretation ] g, gase I Colum i of the I values case, t 2 and estinia dividir 2), tl ii moleci 26-550 1-26 warning of TLV concentration in the air 50-90% of distracted persons perceive warning of TLV Less than 50% of distracted persons perceive warning of TLV Colurr literati averag colunt thresh 0.18-1 10-50% of attentive persons can detect TLV dilutic. v----- concentration-in the.air-__________ _____ ___ <0.18 Less than 10% of attentive persons can detect Ion the TLV ___--------- in Tal symbi dissoc in col Column 8. Water TLV equivalent is the concentration of | in wh the substance in water, which will generate the air TLV , poten concentration in the headspace of a stoppered flask or be les other closed system. It is calculated from column 1 by | or shi multiplying by the distribution ratio in column 13, then for a dividing by 24400 (volume in nil of one gram molecule of [ these vapor at 25 C) and multiplying by the molecular weight. | tratioi Solutions with values in parentheses lack enough per Hasse' sistence for reference purposes, due to an unfavorably low | will si water-air distribution ratio (< 10) in column 13; d, decom- | and tl poses in water. 1 Th propc Column 9. Solubility in ppm (w/v) is the literature could solubility (expressed as g L1 of saturated solution at 25C) made multiplied by 1000. '"indicates uncertain or extrapolated water values. Temperatures other than 25 C are indicated by estini subscripts. odor missii Columns 10. Water-dilution odor threshold is the concen estim tration of the substance in water which will generate the air ful fc odor threshold concentration in the headspace of a comp stoppered flask. It is calculated from column 3 by multi plying by the distribution ratio in column 13, then dividing by 24 400 and multiplying by the molecular weight. Values Varia in parentheses have the same meaning as in column 8. Wher Column 11. The molecular weight (MW, rounded off to the nearest whole number expressed in grams) can be used to convert the air concentrations in ppm (v/v) (columns 1, 2 and 3) into mg m~3. Multiply by MW and divide by 24.4 (volume in liters of one gram molecule of vapor at 25 C). given popu Gaus a lo; as a inter 282 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6,1983 i ODOR AS AN AID TO CHEMICAL SAFETY nit value / column ively by attentive nd what warning lained in odorous concenodorless Column 12. The density (D, at 20-25 C) is needed when measuring out liquid odorants by volume to prepare water or air dilutions: 1 ppm (w/v) = 1 mg [or (lID) (A) per liter of water MW / MW \ 1 ppm (v/v) =----- mg [ or----------- ul 1 per cubic meter 24.4 \ 24.4 xD 1 0f air g, gaseous at 20 C; s, solid at 20C. Column 13. The water-air distribution ratio is the reciprocal of the air-water partition coefficient. Where experimental values are unavailable in the literature, which is usually the case, the ratio has been calculated from data in columns 9, 2 and 11, or from other approaches mentioned earlier. An estimate of the water-air distribution ratio is given by dividing the solubility (column 9) by the volatility (column 2), then multiplying by 24 400 and dividing by the molecular weight (column 11). s perceive he air e warning ; perceive tect TLV Column 14. The numbers indicate how many original literature odor thresholds were included in calculating the average threshold in column 3 and the standard error in column 4. On the left is the number of air-dilution thresholds, and on the right the number measured in water dilution. :an detect ition of iir TLV Task or n 1 by 3, then rcule of weight. gh perbly low decom- | | j , i | I . erature 125C) polated ted by | | toncenthe air of a multilivid ing Values lonizable odorants (weak acids and bases) are indicated in Table 2(b) by appending to the compound name the symbol A for acid and B for base, followed by the acid dissociation constant pWa. Data given for such compounds in columns 8, 9, 10 and 13 are accurate only for solutions in which the odorant is practically un-ionized and hence potentially volatile. That is, the pH of the solution should be less than two pH units lower than the pK% for an acid, or should be more than two pH units higher than the p/fa for a base. The odorant volatilities at pH values outside of these limits can be estimated by calculating the concen tration of the un-ionized species using the HendersonHasselbalch equation.16 For demonstration purposes, it will suffice to make solutions of the acids in 0.01 N H2SC>4, and the bases in 0.01 N NaOH. The data in Table 2 are incomplete for some physical properties of 25 compounds, because no literature values could be located, and no justifiable estimates could be made. The missing data are mostly water solubilities or water-air distribution ratios, which in turn preclude estimates of TLV equivalents in water and water-dilution odor thresholds. If the reader is aware of values for the missing data, or knows of more accurate measurements or estimates of the recorded data, the authors would be grate ful for the information. Odor threshold data on TLV-listed compounds not included in Table 2 would also be welcome. Variance of human responsiveness to odors off to re used i mns 1, ide by lpor at When the individual olfactory detection thresholds for a given compound are determined on a sample of the human population, the data typically generate a (log)normal or Gaussian distribution.17 For this result, it is necessary to use a logarithmic scale for the odorant concentration, such as a binary or decadic dilution series. The quantitative interpretation of a Gaussian curve is facilitated by re plotting the data on probability graph paper. The resulting probit approximates a straight line if the distribution of sensitivities in the population is in fact normal. Literature data on the percentages of persons responding to odorants when they were attentive,18 distracted,19 or asleep20 were replotted as probits in Figs 1,2 and 3. Ethyl mercaptan Figure 1. Tests of responsiveness of persons to fuel gas odorants. The data were taken from the report by Whisman et a!.,''' Figs 12 and 13, and Table 28, then re-plotted on log/probit coordinates. In the misdirected tests, the attention of the subjects was deliberately channeled to other matters. Note that the concentration units in this Figure are ppb lv/v). Some chemicals, but not all, besides having a true odor, also cause immediate irritation in the nose, eyes or throat. The sensation of stinging, prickling or burning, conveyed by the trigeminal or 5th cranial nerve, is quite distinct from the smell sensation carried by the olfactory or 1st cranial nerve.21 Irritation usually requires a higher chemical concentration than odor, and trained normal subjects can readily report the distinct irritation threshold.22 Another approach is to use subjects who have suffered a chronic loss of their olfactory nerve function, but still retain an active trigeminal nerve sensitivity.23 Explanation of Table 4 (irritant thresholds) Column 1. In this Table, each odor threshold was derived from the same source which reported the irritation threshold; hence the odor threshold in Table 4 may differ from that given for the same compound in Table 2(a), column 3, which may be an average of several literature values. Columns 2 and 3. Irritation thresholds are the Lowest con centrations that cause immediate stinging or burning sensations in the nose, or stinging or lacrimation of the eye. JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 283 J. E. AMOORE AND E. HAUTALA -4 -2 0 Qrnory steps (!og2) 2 463 10 12 Awakening Odor safely factor (multiple of threshold! Figure 2. A practical guide to the quantitative interpretation of odor safety factors. The coordinates are log/probit, so care is required in interpolating between marked intervals. The sloping lines indicate the percentages of the population expected to respond to various fractions or multiples of the mean detection threshold concentration {1.0 on the x axis). The detection line represents the performance of fully attentive persons under good laboratory con ditions. The warning line shows what may be expected for distracted persons under factory or field conditions. The warning line is based on the results of Whisman et ai 19 for the gas odorants ethyl mercaptan and thiophane. In four compounds, designated by superscript a, they are the lowest concentrations that could be distinguished from pure air by a general anosmic, i.e. by a person who has no olfactory nerve sensation, but whose trigeminal nerve sensitivity is intact. Column 4. The lower of the nose and eye threshoids (if both are available) was used for calculating this ratio of irritation and odor thresholds. Column 5. The irritation hazard factor is obtained by dividing the nose or eye irritation threshold (whichever is lower, columns 2 or 3) by the threshold limit value from Odor safety factor (multiple of threshold) Figure 3. An illustration of the efficacy of certain vapors in awakening sleeping persons. The data were taken from the work of Fieldner er a/,,30 Tables 8, 12 and 14, then plotted on log/probit coordinates. The irritants were ally! alcohol on the left, and crotonaldehyde on the right. The odorants were ethyl mercaptan () phenyl ether (*} and isoamyl acetate (). The concentrations in this Figure are stated as multiples of the odor thresholds reported by Fieldner er ai, Table 2(a), column 1. This datum indicates by what multiple the TLV is exceeded, if eye or nose irritation can be detected. Column 6. References in italics indicate that thresholds were obtained using water dilutions. RESULTS AND DISCUSSION Literature search for odor thresholds The ACGIH compilation includes approximately 350 appreciably volatile compounds for which time-weighted average threshold limit values have been adopted or pro- Table 4. Irritant threshold concentrations of ten industrial chemicals. See Methods for further explanation of each column Substance 1 Odor threshold (ppm; v/v) 23 Irritation thresholds Nose (ppm; v/v} Eye ' (ppm;v/v) 4 Ratio of irritation and odor thresholds 5 Irritation hazard factor 6 Reference Acetaldehyde Acetic acid Acrolein Ally! alcohol Benzyl chloride cr-Ch loroacetophenone frarw-Crotonaldehyde Formic acid Propionic acid Pyridine 0.066 0.16 1.8 1.4 0.040 0.040 0.11 130 0.24 0.71 2200 160a 11 30 35 0.034 14 11003 3703 700a 11 000 12 59 8.0 0.022 19 33 000 1000 6.1 21 200 0.55 130 8.5 1500 990 22 16 110 15 8.0 0.44 7.0 220 37 140 22 23 22 22 22 22 22 23 23 27 3 Detection threshold for a general anosmic. I posed.1 A> I we were a ' olfactory < I aii or wat literature, I units for I reported t gathered f j used 18 di. . ing their d; l 29 thresho- | mental met The lacl I the inconsi ] variability i ! wide range I literature fi ! of Table 1 I 0.835 ppm. I and calculat * to two sigi | concentrate i statistical d ' should be i j differences standard di I thresholds \ i yields a fai '. literature vr standard en factor of x mately a 6 j threshold fo 0.58 ppin ai I probability i I ppm and (! could, if nC' 1 and smaller factors betw I by redeterm j procedure werror. j In the litt total of 105I to be rejectei j without cons: partition coc I ability of Rr | were discardt from the nea 1 same compor two or mori mean threshc | pound. The ; t for all these I factor of x/ ; yielded only could be cah | column 4 oi olfactory thn of whether tl in the literati i assume the s: I for these singl 284 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 ODOR AS AN AID TO CHEMICAL SAFETY posed.1 Among these, there are 214 compounds for which we were able to locate at least one literature value for the olfactory detection or recognition threshold, measured in air or water dilution. The data are widely scattered in the literature, and there is little conformity in the choice of units for expressing the results. For example, the 29 reported thresholds for n-butyl alcohol (Table 1) were gathered from the works of 26 principal authors, who used 18 different systems of concentration units in publish ing their data, in 24 journals. Furthermore, no two of these 29 thresholds were measured by precisely the same experi mental method. The lack of standardization, taken in conjunction with the inconsistent purity of the chemical samples and the variability of human sensitivity, is responsible for the rather wide range of threshold concentrations usually found in the literature for a given compound. As indicated at the foot of Table 1, the mean threshold for n-butyl alcohol is 0.835 ppm. (In this compilation, the data were collected and calculated to three significant figures, then rounded off to two significant figures for the Tables.) The threshold concentrations having been calculated as logarithms, statistical deviations and errors from the geometric mean should be stated in the form of factors (rather than the differences used with ordinary arithmetic means). The standard deviation of the logarithms of the observed thresholds was 0.854 log10 units, for which the antilog yields a factor of x/7.14. Taking into account all 29 literature values (i.e. dividing by V29), this reduces to a standard error of 0.159 log10 units, corresponding to a factor of x/t-1.44. This indicates that there is approxi mately a 68% probability ( let or SD) that the true threshold for n-butyl alcohol lies between (0.835/1.44) = 0.58 ppm and (Q.835 x 1.44)= 1.20 ppm. There is a 96% probability (2o) that it lies between (0,58/1.44) = 0.40 ppm and (1.20 x 1.44) = 1.73 ppm. Olfactory thresholds could, if necessary, be obtained with greater consistency and smaller standard errors, by determining conversion factors between different experimental methods,2,3'14 or by redetermining the thresholds by using a standardized procedure with careful minimization of known sources of error. In the literature, we found for these 214 compounds a total of 1054 acceptable thresholds. Some thresholds had to be rejected on the grounds that they had been measured without consideration of substantial ionization, unfavorable partition coefficients, likely impurities or the inapplic ability of Raoult's law. A few remaining extreme points were discarded because they diverged more than 100-fold from the nearest of two or more other thresholds for the same compound.24 For 152 of the compounds, we found' two or more acceptable thresholds. We calculated the mean threshold and its standard deviation for each com pound. The average of the individual standard deviations for all these 152 multiple threshold compounds was a factor of x/7.0. The remaining 62 compounds each yielded only one usable threshold, so no standard error could be calculated, which accounts for the dashes in column 4 of Table 2(a). The uncertainty in a given olfactory threshold measurement should be independent of whether the compound has been reported several times in the literature, or only once. As a rough guide, we may assume the same average standard error factor of x/=7.0 for these single-threshold compounds. Safe dilution factors for saturated vapors The procedure of expressing threshold limit values, volatilities and odor thresholds all in the same units (ppm; v/v) brings to light certain relationships that are not apparent when miscellaneous units are used. Nearly ail of the compounds in Table 2(a) have volatilities at 25 C which exceed, sometimes by an enormous factor, their threshold limit values. Accordingly, a sniff, from the head space of a bottle or drum, or from a confined space con taining a spill, of almost any of these substances, inevitably exceeds the TLV. The safe dilution factor in column 5 indicates the minimum number of volumes of uncon taminated air that would be required to dilute, to the safe level, one volume of air that has been saturated by exposure to the named compound (assuming perfect mixing). Plant location, layout, ventilation, chimneys and emergency procedures should be designed with the realization of the safe dilution factor in mind, at least for compounds for which dilution ventilation is an allowable method of control. Any increase in temperature of the chemical above 25 C increases the required safe dilution factor, in pro portion to the vapor pressure. A majority of these compounds are not completely miscible with water. Nevertheless, a saturated solution of any volatile compound is theoretically capable of saturating the headspace to the same concentration as the pure com pound could achieve. Whether or not it will do so in a finite time depends upon the water-air distribution ratio, the relative volumes of air and water, and the degree of agitation. To err on the safe side, it would be prudent to use the same safe dilution factor in calculating the number of volumes of clean water which would be needed to dilute one volume of a saturated aqueous solution of the compound before discharge to a sewer, lagoon or river, where this is permitted. Odor safety factors as chemical safeguards When the threshold limit value is substantially higher than the odor threshold, the intrinsic odor of the compound usually, but not invariably, provides an indication of its presence, at a concentration level low enough that no harm is likely to the human observer. Conversely, if the odor threshold is much higher than the TLV, then anybody detecting the odor of the compound has a warning that a safe vapor concentration has already been exceeded. The exposed worker would be well advised to request a pro fessional evaluation and perhaps instrumental assessment of the situation. It should be determined whether the applicable TLV criterion (time-weighted average, short term exposure limit or ceiling value)1 is likely to be exceeded in the particular working regime, and if so, what the health significance may be. The potential warning power of a given chemical is conveniently expressed by the odor safety factor (column 6 of Table 2(a)), which is simply the TLV divided by the odor threshold. Any chemical with an odor safety factor less than 1.0 carries the risk that hazardous concentrations will not be detected by odor. Conversely, an odor safety factor greater than 1.0 bears the promise that a hazardous concentration could be perceived by smell. Nevertheless, the question of whether or not a hazardous concentra- J0URNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 285 1. E. AMOORE AND E. HAUTALA Lion will actually be smelt, is quite complex, and depends upon a variety of circumstances. (A very few people, roughly 1 in 500, have no true sense of smell at all;25 the existence of anosmic persons, while of some practical importance, is omitted from our discussion.) The average odor threshold has not been sufficiently rigorously evaluated for all these compounds, many of which possess measured or implied standard errors as large as seven-fold. This is not, in principle, an insurmountable problem, because 63 compounds in Table 2(a), column 4, already have thresholds evaluated with standard errors less than two-fold. Equal, or better, accuracy could readily be attained by new experimental measurements on the deficient compounds. The ability of members of the population to detect a given odor is strongly influenced by the innate variability of different persons' olfactory powers, their prior experience with that odor, and by the degree of attention they accord to the matter. The thresholds listed in column 3 of Table 2(a) represent the most favorable conditions for testing. The subjects were well aware that these were tests of their sense of smell, they were attentive and they were trying their best to detect the presence of the odor. Even so, the odor-detecting ability of different people varies over quite a wide range. The compilation of individual sensitivities to a given compound typically yields a Gaussian or bell-shaped curve," provided that a logarith mic concentration scale is employed. For this normal distribution, the standard deviation is a measure of the spread of odor sensitivity in the population. We have evaluated this standard deviation with seven odorants: isobutyl isobutyrate, isovaleric acid, 1-pyrroline, trimethylamine, isobutyraldehyde, androst-16-en-3-one and pentadecalactone, each tested with 18-443 normal observers. The average standard deviation was 1.97 binary steps, which may be rounded off at two binary steps.15 The standard deviation indicates that 68% of people tested, on the average, will have a personal threshold that lies within the range from one-fourth of the mean, to four times the mean, threshold of the population. The effect of distracted attention In connection with testing the efficacy of certain odorants as warning agents for fuel gas. Whisman et al.i9 conducted a thorough study of the influence of various degrees of distraction on the responsiveness of people to these wellknown warning odors. Their `directed' test corresponds with usual laboratory conditions, in which the attention of the subject is purposely focused on the sole objective of detecting an odor. In the `semi-directed' test, the subjects were asked to report on visual, tactile, aural and nasal stimuli in the test room. In the. `undirected' test, the subjects were given no indication of the object of the exercise. In the `misdirected' test, the attention of the participants was deliberately distracted by asking each to try to read some print in a dim light and to judge the temperature of the room. All except the directed tests were performed with inexperienced subjects recruited by a mobile laboratory arriving unannounced at shopping centers, and each volunteer was used for one test only at one odorant concentration. Whisman et al. found that the responsiveness of the subjects to a given concentration of odorant was sub stantially decreased in the semidirected, undirected and misdirected tests, compared with their performance in the directed test mode. The misdirected test was probably the most difficult set of conditions imposed upon the subjects In our opinion, the misdirected test is the most appropriate of the available models for evaluating the effects of con ditions encountered in industrial practice. A factory worker would not be familiar with odor-threshold testing tech niques, but would hopefully be aware that chemical vapors may be hazardous, and might know that a distinct smell indicates the presence of appreciable vapor in the air. On the other hand, the worker is likely to be concentrating on following instructions, reading charts, controlling equip, ment and generally trying to get the work done. Such a degree of mental distraction, as Whisman et al. showed is ample to divert attention away from any but the most obvious of odors. In Fig. 1, the results of Whisman et al.19 for their directed and misdirected test modes are presented in log2/ probit coordinates, which have the advantage of exhibiting an approximately linear relationship between olfactory stimulus and response. Each data point in the directed tests was obtained from 22 subjects, and in the misdirected tests from over 100 subjects. The data points were fitted by a logarithmic transformation linear regression, from which the slope and 50% response intercept were obtained. The directed test threshold for ethyl mercaptan, at which 50% of the subjects would respond, was found by extrapolation to be 0.17 ppb. In the misdirected test situation, however, the 50% response threshold was at 4.8 ppb, or 28 times higher. Furthermore, the slope of the regression line is shallower, so that disproportionately higher concentrations are required to elicit a response from 90% of the partici pants. The results for thiophane (tetrahydrothiophen) are virtually superimposable upon those for ethyl mercaptan, except that about double the concentration of odorant is needed to achieve a given level of response. That is, 0.35 ppb for detection threshold and 8.7 ppb for mis directed threshold, or 24 times higher. The good agreement between the results for ethyl mercaptan and for thiophane encourages us to generalize the data, so as to provide a practical guide for interpreting threshold ratios and odor safety factors (Fig. 2). This graph is set in log/probit coordinates. Since neither the logarith mic nor the probit scales go to zero, the origin of the graph is considered to be the intersection of threshold multiple 1.0 on the x axis, with 50% persons responding on the y axis. This, by definition, is the average detection threshold, measured under laboratory conditions, i.e. a directed test. The logarithmic binary step concentration scale and the standard deviation intervals are also entered in Fig. 2. It was previously demonstrated15 that the sensitivities of people to various odorants exhibit standard deviations close to 2.0 binary steps. Hence, the detection line in Fig. 2 is based on this generalization, and constructed by drawing a line with a slope of 2.0 binary steps per standard deviation unit, through the origin of the graph. The detec tion line is shown as a broken line above 95% response, because there are some indications that a small percentage of the population has specific anosmias to one or more ol the sulfurous odorants." Such persons, while they may perceive most other odors normally, are found to have an innate lower sensitivity or `odor blindness' to the typical gas odorants. I The war j misdirectei It was cor I eaptan and of the 50' I directed te | Their geon for the th | drawn to L I multiple va ' likewise d | regression averaged w I standard de I Therefoj ! subjects, th j to 26 time 50% of attc I illustration j Whisman t 1 between th| level that w j The availat I warning lint j Odor safety ! Figure 2 r available da soundly er chemicals. I pronounced | among men I to flatten : intercept ol | two, quite i i feel that tl i represent a | public awar and bottled I gas' is an in I may have a < of a distrae j mentally as.1 Until mt j relationship | classificatioi safety indie I adopting th. , guides. Acci detection tl the odor, a of people \ traded. Sec | tions at whi and the ol ] get a warn indicated b> Our tent Table 3. / class A con persons. To I be at least compound 286 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 i d and in the ly the ejects, priate ` conrnrker tech^apors smell ir. On ing on iauipuch a red, is most their 'g 2/ biting ictory i tests 1 tests 1 by a which I. The i 50% lation vever, times ine is ations articiphen) ethyl ion of . That r mis- ethyl -ralize reting graph 'arithgraph jltiple the y shold, 1 test, id the 2. It ,es of ations Fig. 2 awing ndard detec- lonse, :ntage ore of ' may rve an ypical ODOR AS AN AID TO CHEMICAL SAFETY The warning line in Fig. 2 is based on the average of the misdirected data for both ethyl mercaptan and thiophane. It was constructed as follows. The results for ethyl meTcaptan and for thiophane (Fig. 1) showed that the ratios of the 50%-detection thresholds in the misdirected and directed test protocols were 28.3 and 24.5, respectively. Their geometric mean is 26.3, which was rounded off to 26 for the threshold1 multiple. In Fig. 2, the warning line is drawn to intersect the 50% response level at the threshold multiple value of 26-fold. The slope of the warning line was likewise determined by averaging the slopes of the regression lines for the misdirected tests in Fig. 1. The averaged warning line has a slope of 3.5 binary steps per standard deviation unit. Therefore, in order to be perceived by 50% of distracted subjects, the concentration of gas odorant had to be raised to 26 times the concentration that could be detected by 50% of attentive subjects in laboratory test conditions. This illustration lends emphasis to the compelling conclusion of Whisman et al.19 that there is a substantial difference between the level of odorant that can be detected, and the level that will be detected, in a given set of circumstances. The available data do not permit extrapolation of the warning line in Fig. 2 below the 50% response level. Odor safety classification of chemicals Figure 2 represents a provisional synthesis of the best available data. The slope of the detection line appears quite soundly established, and to be applicable to many chemicals. For those uncommon chemicals that exhibit a pronounced and frequently occurring specific anosmia among members of the population,26 the curve is expected to flatten at higher response percentages. The slope and intercept of the warning line, however, are based on only two, quite closely related, fuel gas odorants. Intuitively, we feel that tire results for ethyl mercaptan and thiophane represent a relatively favorable case, because, thanks to the public awareness developed by the suppliers of household and bottled gas, it is a widely known fact that the `smell of gas' is an indication of danger. In other words, gas odorants may have a better chance of penetrating the consciousness of a distracted person than many other odors that are not mentally associated with harmful consequences. Until more data become available, we propose that the relationships in Fig. 2 can be used to set up a provisional classification of the 214 chemicals, according to the level of safety indicated by their odors. For this purpose, we are adopting the 10%, 50% and 90% response levels as practical guides. According to Fig. 2, the obvious benchmarks are the detection threshold at which 50% of people can perceive the odor, and the higher warning threshold at which 50% of people will notice the odor even when they are dis tracted. Secondary criteria are provided by the concentra tions at which 10% of attentive people can detect the odor, and the other extreme where 90% of distracted people get a warning of the odor. These four borderlines are indicated by vertical lines in Fig. 2. Our tentative odor safety classification is presented in Table 3. At their threshold limit value concentration, class A compounds will be perceived by 90% of distracted persons. To achieve this rating, the odor safety factor must be at least 550; i.e. the threshold limit value for the compound is more than 550 times higher than its odor threshold. At the other extreme, class E compounds at their TLV concentration can be detected by less than 10% of attentive persons. In this category, the odor safety factor is below 0.18. The quantitative ranges for the intermediate B, C and D classifications are as indicated in Table 3. The zones of odor safety factor for the five classes are also labeled on Fig. 2. The odor safety class of each of the 214 compounds, for which adequate data are available, are entered in column 7 of Table 2(a). Class A compounds provide the strongest odorous warning of their presence at the TLV level, whereas class E compounds aTe practically undetectable by odor at their TLV concentration. The effect of sleeping Although it is not considered relevant to most workplace situations, the power of an odorant to waken a sleeping person is significant where industrial products can escape into a residential area. This is an obvious risk with house hold gas, and the question was included in a study by Fieldner et al20 Their data for several odorants are displayed in log10/probit coordinates in Fig. 3. They tested three compounds (ethyl mercaptan, phenyl ether and isoamyl acetate) which can be regarded as more or less purely olfactory stimulants, i.e. they have little or no irritating power for the trigeminal nerve. Each data point in Fig. 3 was calculated from the results of tests with three to eight sleepers. The points were then fitted by linear regression. The performances of these three odorants seem fairly concordant, and imply that an odorant concentration about 20 000 times the normal detection threshold is required to awaken 50% of soundly sleeping persons. That is more than 700 times stronger a stimulus than suffices to serve as a warning for wakeful, but misdirected, observers (Fig. 2). If this result were applicable to all odorants, it would mean that virtually none of the 214 compounds examined in Table 2(a) would awaken the average person, without exceeding the TLV. There is, however, a complicating factor. Some odorants, besides stimulating the olfactory nerve, also irritate the trigeminal nerve. Two examples are included on the left side of Fig. 3. These substances were far more effective in waking the sleepers. A 50% response was obtained at 27 times the odor threshold of crotonaldehyde, and at only three times the odor threshold of ally1 alcohol. From the comments of those that woke up, it is obvious that the irritation was the determining factor. It is an interesting observation that the trigeminal nerve has some sort of a `hot line' directly into the subconscious, that is denied to the olfactory nerve. Some data on irritant thresholds Trained normal observers can report distinct concentration levels at which a vapor produces nasal or eye irritation, quite apart from its odor. Katz and Talbert22 tabulated considerable data, from which we have selected those compounds that are on the ACGIH list (Table 4). We have also added a few compounds from our own work, in which nasal irritation thresholds were obtained from an anosmic person lacking the ability to perceive true odors as opposed to irritants. The ratio of the irritation and odor JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 287 J. E. AMOORE AND E. HAUTALA thresholds for these compounds ranges from 33000 for acetaldehyde, to less than unity for a-chloroacetophenone. Where this ratio is relatively small, it seems very likely that irritation would become an important factor in determining the intercept and slope of the warning line in Fig. 2. If irritation of the trigeminal nerve can wake a sleeping person so effectively, it seems very likely also to be able to preempt the attention of a distracted person. No quanti tative treatment of this factor is possible at present, because irritant thresholds are available for so few of the compounds on the TLV list, and no tests have been reported on perception of irritants by distracted persons. It may, however, be worth noting the irritation hazard factor in column 5 of Table 4. These figures indicate the degree to which the TLV is being exceeded, if there is appreciable eye or nose irritation for art attentive subject. Threshold in water dilution Many of the odor thresholds found in our literature survey had been measured by sniff-tests from the head-space above aqueous dilutions. Theoretically, the air-dilution threshold and the water-dilution threshold are simply related by the air-water partition coefficient of the odorant, provided the concentrations are measured in equivalent weight per volume units. This expectation has been borne out in comparisons made for n-butyl alcohol, pyridine and isovaleric acid,14 and has been further supported by the data for many compounds listed in Table 2(b). For example, the data for n-butyl alcohol in Table 1 exhibit, for the reported olfactory thresholds, more than a 1000-fold range, yet the group means of the 20 air thresholds and the nine water thresholds differ by a factor of only about three-fold, and this is not considered signifi cant (/>>0.1). Odor thresholds measured in air and water dilutions are generally concordant, unless the water-air distribution ratio is less than approximately ten. In that case, the reported water-dilution threshold concentration is liable to be too high, due to substantial evaporative loss of odorant from the solution during the course of conducting the odor threshold tests. The air-dilution thresholds in column 3 of Table 2(a) are based on a pool of all available data from both air- and water-dilution measurements, omitting water thresholds for compounds with unfavorable water-air distribution ratios. The water dilution thresholds in column 10 of Table 2(b) were generally calculated from the data in column 3 of Table 2(a), by applying the water-air distribution ratio. In this way, we have been able to calculate water-dilution thresholds for many compounds for which only air-dilution threshold data were previously available. By applying the same distribution ratio, the water equivalent concentrations were also calculated for the TLV, and are listed in column 8 of Table 2(b). With odorants that are ionizable (acids and bases), these calculations are strictly valid only within specified pH limits, as explained in the Methods section. We felt that it would be informative to provide the theoretical water threshold and TLV data, even for com pounds with distribution ratios of less than ten. The equi librium air concentration can develop and persist in conditions of high liquid-vapor volume ratio and low vapor loss, such as a closed vessel or a sewer. TLV and threshold data for odorants with distribution ratios less than ten are in parentheses in Table 2(b). This is to indicate that those solutions lack enough persistence to serve as reliable standards in setting up water dilution sniff-tests for training or testing personnel. CONCLUSION The interpretation of these data in any particular safety or pollution problem will depend markedly on the individual circumstances. The threshold data in the Tables and Figures are based on averages for samples of the population, pre sumably in good health. Individuals can differ quite markedly from the population average in their smell sensi tivity, due to any of a variety of innate, chronic or acute physiological conditions.13'28'29 likewise, the timeweighted average threshold limit values are for workers, who by the mere fact of being able to work evidently represent a generally healthy segment of the population. Continuing exposure to an odor usually results in a gradual diminution or even disappearance of the smell sensation. This phenomenon is known as olfactory adapta tion or smell fatigue.30 If the adaptation has not been too severe or too prolonged, sensitivity can often be restored by stepping aside for a few moments to an uncontaminated atmosphere, if available. Unfortunately, workers chronically exposed to a strong odor can develop a desensitization which persists up to two weeks or more after their de parture from the contaminated atmosphere. In such cases, it should be the responsibility of supervisors and inspectors to note the odor and take appropriate action. Hydrogen sulfide and perhaps other dangerous gases can very quickly lose their characteristic odor at high concentrations. At levels of H2S above 100 ppm (over 10 000 times the average detection threshold), the sense of smell is rapidly abolished, so that potentially lethal concen trations may not be detected by odor at all.31 Certain commercial diffusible odor masking or suppressing agents may reduce the perceptibility of odors, without removing the chemical source. The use of such agents might interfere with the capability of the nose to provide a warning at the expected concentration level. There are many potential applications of these data in chemical safety and in air- and water-pollution control, some of which have been mentioned previously. In addition, we believe that the data might find some less apparent uses: Table 2 is also a guide to what data are in the literature on odor thresholds, on TLV-listed substances, is unavailable, unconfirmed or erratic. Readily prepared water dilutions could be used to test the individual smell thresholds of workers to the chemicals they handle. A water TLV dilution of an odorant could be prepared to demonstrate quickly to workers the practical experience of its TLV concentration. The general experimental pro cedures for preparing and testing aqueous solutions of odorants have been described.32 These concepts could improve the reliability of odor breakthough as an indication of when to change the organic vapor cartridge in a respirator. The feasibility might be considered of using class A or B compounds as warning odorants to be added to class D or E substances, or to pesticides. The water-air distribution ratios could also be a guide to the possible success of water-scrubbing as a means of removing vapors from effluent gases. | The 1 i those rei > The valu I revision, 1 The US I Administ have est; While thi j recommei ACG1H T I certain cc | lines witl exposure I the TLV , ] adjust the > ratio, and | Values in Every i ! a property | maintainin * recognized ) tion and ii I margin, if ; conditions condoning sistent stc obtained tJ The fir; something j 1. Thresh o. Agents i I of Gove* 2. P. Laffo pour 19: ' 3. F, Patte, j values o Chem. Sf | 4. L. J. va 1 Odour T, k Nutritior * 5. L. J. van j Air, Sup/ Research | 6. F. A, F; Values L | Philadetp 7. Doeumen I printing. Hygienist 8. T. M. Ht propertieJ. Air Pol 9. O. R. Si ' Chem. 39 \ 10. A. Seidell ! D. Van N , 11. A. Seidel Organic < New Yod 12. K, Versch Chemical. 13. Beilstetn'i Suppleme 14. J. E. Ait i cornparati (1978). 288 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. G, 1983 I ODOR AS AN AID TO CHEMICAL SAFETY ible ling ' or ual ires irelite nsiute Tie- in, itly 1a lell ita- `.00 red ted illy ion deses, ors :ses igh ver : of enain nts ing ere the i in rol, In less the i, is iter lell A to ! of >roof uld ion ia ling 1 to -air ible rors I The TLVs used in Table 2 and discussed in this paper are I those recommended by the ACGIH in its 1982 listing.1 ' The values are re-published annually, and are subject to I revision, usually with two years notice of intended changes. The US Government Occupational Safety and Health I Administration (OHSA) and many State Administrations I have established their own lists of permitted exposures. While the values adopted are often based on the ACGIH . recommendations, they may not coincide with current ACGIH TLVs, and quite different standards may be set for I certain compounds. Some foreign governments issue guide- j lines with independently derived limits. If the applicable exposure limit for a particular compound is different from I the TLV cited in Table 2, column 1, it will be necessary to adjust the values in columns 5, 6 and 8 by the appropriate ratio, and perhaps reassign the odor safety class (column 7). 1 Values in Table 4, column 5 may also have to be altered. Every chemical that can be detected by smell exhibits 1 a property that can be turned to advantage as an aid in | maintaining safe operating conditions. It must be recognized that background odors, odor fatigue, preocupa| tion and individual insensitivity may combine to reduce the I margin, if any, between odor detection and safe operating i conditions. No odor safety factor is large enough to justify condoning the presence of a fleeting odor, let alone a per sistent stench, unless professional assurance has been obtained that the working conditions are safe. The fust detectable odor should be a sure signal that something abnormal has happenedsomewhere.lt maybe the last warning. During chemical operations, when an odor is detected, the source should be located and the concen tration determined. Then effective steps can be taken to prevent the escape of vapor, and restore a neutral and healthful odor background. Even in the unnatural environ ment of the industrial workplace, our sense of smell has much to offer as a natural safety warning system. Acknowledgements We are very grateful to Dr R. G. Buttery for measuring the air-water partition coefficients of some infinitely soluble compounds by gas chromatography, and to Mr C. J. Thompson for an advance copy of his manuscript with Whisman et al. on the responsiveness of people to gas odorants.1' We thank Mr W. D. Kelly, Executive Secretary of the American Conference of Governmental Industrial Hygienists, Inc., for permission to use the TLV data from Ref. 1 in Table 2(a). The preparation of this paper was supported in part with funding provided under Service Order No. 34 016, from the Hazard Evalua tion System and Information Service, Department of Health Services-Depaitment of Industrial Relations, State of California. This report has been reviewed by the staff of the Hazard Evalua tion System and Information Service Section, Department of Health Services-Department of Industrial Relations, State of California, and approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the Hazard Evaluation System and Information Service Section, nor does mention of trade names or commercial products constitute endorse ment or recommendation for use. Reference to a company and/or product in this publication is only for purposes of information and does not imply approval or recommendation for the product by the US Department of Agri culture to the exclusion of others which may also be suitable. I I -- ...... ........ --------------------------- ... REFERENCES ---------- 1 .............. 1. Threshold Limit Values lor Chemical Substances and Physical Agents in the Workroom Environment, American Conference of Governmental Industrial Hygienists, Cincinnati (1982). 2. P. Laffort, Essai de standardisation des seuils olfactifs humains pour 192 corps purs. Arch. $ci. Physiol. 17, 75-105 (1963), 3. F. Patte, M. Etcheto and P. Laffort, Selected and standardized values of suprathreshold odor intensities for 110 substances. Chem. Sens. Flavor 1, 283-305 (1975). 4. L. J. van Gemert and A. H. Nettenbreijer, Compilation of Odour Threshold Values in Air and Water, Central Institute for Nutrition and Food Research, Zeist (1977). 5. L. J. van Gemert, Compilation of Odour Threshold Values in Air, Supplement ///, Central Institute for Nutrition and Food Research, Zeist (1980). 6. F. A. Fazzalari, Compilation of Odor and Taste Threshold Values Data, American Society for Testing and Materials, Philadelphia (1978). 7. Documentation of the Threshold Limit Values, 3rd edn, 4th printing, American Conference of Governmental Industrial Hygienists, Cincinnati (1977). 8. T. M. Heilman and F. H. Small, Characterization of the odor properties of 101 petrochemicals using sensory methods. J. Air Poll. Control. Assoc. 24, 979-982 (1974). 9. D. R. Stull, Vapor pressure of pure substances, ind. Eng. Chem. 39, 517-550 (1947). 10. A. Seidell, Solubilities of Organic Compounds, 3rd edn, Vol.2, D. Van Nostrand, New York (1941}. 11. A. Seidell and W. F. Linke, Solubilities of Inorganic and Organic Compounds, Suppl. to 3rd edn, D. Van Nostrand, New York (1952). 12. K. Verschueren, Handbook of Environmental Data on Organic Chemicals, Van Nostrand Reinhold, New York (1977), 13. BeUstein's Handbuch der organischen Chemie, 4th edn and Supplements 1-4, Julius Springer, Berlin (1918 ff.). 14. J. E. Amoore, and R. G. Buttery, Partition coefficients and comparative olfactometry. Chem. Sens. Flavor 3, 57-71 (1978). 15. G. J. Pierotti, C. H. Deal and E. L. Derr, Activity coefficients and molecular structure. Ind. Eng. Chem. 51,95-102 (1959). 16. J. E, Amoore, Odor theory and odor classification. In Fragrance Chemistry, ed. by E. T. Thejmer, pp. 27-76. Academic Press, New York (1982). 17. J. E. Amoore, Odor blindness as a problem in odorization. Am. Gas Assoc. Oper. Sect. Proc., Distribution Coni. pp. 242-247 (1968). 18. J. E. Amoore, P. Pelosi and L. J, Forrester, Specific anosmias to 5or-androst-16-en-3>one and oi-pentadeca lactone: the urinous and musky primary odors. Chem. Sens. Flavor 2, 401-425 (1977). 19. M. L. Whisman, J. W. Goetzinger, F. O. Cotton, D. W. Brink man and C. J. Thompson, A New Look at Odorization Levels for Propane Gas, Bartlesville Energy Research Center, Bartels- ville, OK (1977). 20. A. C. Fieldner, R. R. Sayers, W. P. Yant, S. H. Katz, J. B. Shohan and R. D. Leitch, Warning Agents for Fuel Gases, U.S. Dept, of Commerce, Bureau of Mines, Monograph 4 11931). 21. R. I. Henkin, The definition of primary and accessory areas of olfaction as the basis for a classification of decreased olfactory acuity. In Olfaction and Taste it, ed. by T. Hayashi, pp. 235 252. Pergamon Press, Oxford (1967). 22. S. H. Katz and E. J. Talbert, Intensities of odors and irritating effects of warning agents for inflammable and poisonous gases. U.S. Dept, of Commerce, Bureau of Mines, Technical Paper 480 (1930). 23. J. E. Amoore, D. Venstrom and A. R. Davis, Measurement of specific anosmia. Percept Motor Skills 26,143-164 (1968). 24. V. C. Allison and S. H. Kat2, An investigation of stenches and odors for industrial purposes. Ind. Eng. Chem. 11, 336-338 (1919). . 25. P. M. Patterson and B. A. Lauder, The incidence and probable inheritance of smell-blindness. J. Heredity 39, 295-297 (1948). 26. J. E. Amoore, Specific anosmia and the concept of primary odors. Chem. Sens. Flavor 2, 267-281 (1977). JOURNALOF APPLIED TOXICOLOGY,V0L.3,NO.G,1983 289 }. E. AMOORE AND E. HAUTALA 27. A. H. Sherman, J. E. Amoore and V. Weigel, The pyridine scale for clinical measurement of olfactory threshold: a quantitative reevaluation. Otolaryngol. Head Neck Surg. 87. 717-733 (1979). 28. R. A. Schneider, The sense of smeli in man -- its physiologic basis. New Engl. J. Med. 277, 299-303 (1967). 29. R. L. Doty, A review of olfactory dysfunctions in man. Am. J. Otolaryngol. 1, 57-79 (1979). 30. G. T. Pryor, G, Steinmetz and H. Stone, Changes in absolute detection threshold and in subjective intensity of supra- threshold stimuli during olfactory adaptation and recovery Percept Psychophys. 8, 331-335 (1970). ' 31. Occupational Exposure to Hydrogen Sulfide: Criteria for a Recommended Standard, National institute for Occupational Safety and Health, Washington DC (1977). 32. J. E. Amoore, Directions for preparing aqueous solutions of primary odorants to diagnose eight types of specific anosmia Chem. Sens. Flavor A. 153-161 (1979). ' Received 17 August 1982; accepted (revised) 15 February 1983 j I . ; .. ' 290 JOURNAL OF APPLIED TOXICOLOGY, VOL. 3, NO. 6, 1983 INTRC t There , coneeri biologi. j three c active I extend[ uraniui ' means j proved of a : body b | reduce: iodide ^ ment < elemen aminep Desj which * of mix j proced Rec ; amidol nuclide 1 sodiun ' perchli can be antido J f Auth Wile;