7, in the Chemical laboratory Edited by NORMAN V. STEERE, 140 Melbourne Ave., S.E. Minneapolis, Minn. 5541 4
XCIX.
Estimation of Flash Point Temperature* Richard W. Prugh, Senior Engineer. Safety & Fire Protection. E. I.du Pont de Nemours Co., inc., Wilmington, Del. 19898
A simple technique for determining the approximate flash point temperature of a compound. knowing only its atomic composition and boiling point, has been developed from experimental data and some theoretical considerations.
It is generally agreed that before a chemical is first used or produced, appropriate precautions should be taken against its known or suspected explosive, combustible, flammable, toxic, or radioactive nature. A problem which frequently arises in research and development laboratories is how to estimate the degree of combustibility or flammability. For some non-toxic, "on-explosive, and "on-radioactive materials, a simple test can be conducted to determine if a match flame or spark will cause ignition of vapors *This paper was presented a t the 1971 meeting of the Research and Development Section of the National Safety Council, published in the 1971 National Safety Congress Transactions, and is reprinted by permission.
from the material a t room temperature. If ignition is obtained the material is flammable, and there would be a fire or explosion hazard unless adequate precautions are taken. If there is no ignition, the test must be considered inconclusive, since the material might evolve vapors a t some higher temperature which might he easily ignited. Thus, there would he a fire or explosion hazard during handling or use of the material a t or above the higher temperature. The distinction between combustible and flammable is intended to differentiate between degrees of hazard. For example, Webster's dictionary defines combustible as "capable of burning," whereas flammable is "capable of being easily ignited." Generally, the term "flammable" is assigned to materials whose vapors can be
feature
ignited by a spark or flame a t temperatures which might normally be encountered during storage or handling. Accidental ignition of such vapors could result in a flash fire or explosion, depending upon the degree of confinement. The upper limit to "normal" temperatures which should he anticipated differs with regulatory or advisory authorities. The National Fire Protection Association
garding precautions against ignition by static electricity. Factory Mutual (21 makes a distinction between liquids having flash points above or below 110°F (43.3-C) regarding flame arresters in vessel vents, and a flash point of 80PF(26.7'C) regarding storage in basements. The U.S. Department of Transportation has used a flash point temperature of 80°F to determine whether or not a container of flammahle liquid must bear a r e d warning label (3). Thus, to determine the precautions which should be taken to assure safe handling or use of a chemical, a vapor ignition
surf&. (Continued on page A86)
Volume 50. Number 2. February 1973
/
A85
Safety
...
test should he conducted at 140°F (6OSC), using the most conservative guidelines. It is often not convenient or economically practical t o conduct such a test prior to preliminary experimental work, hecause of the toxicity, radioactivity, or other hazardous property. Therefore a method of estimating a material's flash point would he useful. Several methods for estimating flash point have been proposed. For hydroearbons, it has been observed that there is a close relationship between flash point (FP) and boiling point (BPI. This relationship (2, 4) can he described by the equation
the flash point) could be easily determined from the boiling point. Unfortunately, there is no single convergence point, hut there are two areas of convergence (one for alcohols, and another for many other materials) near the infinite-temperature ordinate. For the purpose of the present flashpoint estimation method, two points of convergence on the infinite-temperature ordinate are assumed. This assumption results in some loss of accuracy in estimated flash point. The concentration of vapor in air a t the flash point i.e., the approximate lower flammable limit, LFL, in volume-per cent can be estimated from the vapor pressure (VP, in mmHg) a t the flash-paint temperature, using Dalton's and Amagat's laws: LFL = ( VP/760) (1001
F P = 0.683 B P - 119 [FPand BPin "F] ( F P = 0.683 B P - 66)[in0C] and is applicable t o hydrocarbons having initial boiling points between 200°F and 700°F (93.3"C and 37%). Another method of estimating flash point, using molecular structure, has been proposed (51. An intermediate calculation to determine autoignition temperature1 is required and is based on the total number of carbon atoms in the molecule and the number of carbon atoms in branches. The flash point is then calculated from an equation involving the autoignition temperature and the number of hydrogen atoms in radicals. During the 1967 National Safety Congress, the relationships between lower explosive or flammable limit, vapor pressure, and flash point were explained and demonstrated (6). The statement was made that the vapor concentration in air above a liquid a t its flash point should he equivalent to the lower flammable limit. In general, the vapor concentration corresponding to the lower flammable limit occurs a t a temperature slightly below the flash point, when calculated from a vapor pressure curve. This discrepancy may result from experimental difficulties in attaining equilibrium in the testing apparatus and the fact that the flash point test is essentially a downward-propagation test, whereas lower flammable limits conservatively state upward-propagation vapor or gas concentrations. The theoretical and and flash paint would allow calculation of one characteristic (such as flash point) if the other two were known. However, far many novel chemicals, these data are not available. The observation that vapor pressure curves of many hydrocarbons converge to a point was demonstrated hy Cox (71. Thus the locus of this point and the narma1 boiling point would be sufficient to describe the vapor pressure characteristics of any particular hydrocarbon. If there were only one convergence point for all organic and inorganic materials, the vapor pressure a t any given temperature (e.g., ~
I Autoienition t e m ~ e r a t u r eis the mini-
A86
/ Journal of Chemical Education
A relationship between lower flammable limit (LFL, in volume-per cent) and stoichiametric concentration (CST, in volume-per cent) for paraffin hydrocarbons was documented by Jones and Lloyd (8) and can be described by the equation: LFL
+ 0.55 CST
Other organic and inorganic materials behave similarly, and use of the 0.55 coefficient generally does not result in a relative error greater than 20 per cent in the lower flammable limit, as compared with an experimentally determined lower limit. Calculation of the staiehiometric eoncentration of a vapor in air is relatively straightforward, if it can he presumed that: (a) all carbon (C) is axidized to COX; (b) all sulfur (S) is oxidized to SOz; (c) all halogens (X) react with hydrogen to form hydrogen halides; (d) any remaining hydrogen ( H ) is oxidized t o water; and (el all oxygen ( 0 ) in the molecule is available for 'oxidation of other constituents. An equation relatine the stoichiometric eoncentratmn ('.
