Computationally evaluate self-reactivity hazards

Chemical Engineering Progress, Feb 2003 by Murphy, Michelle R, Singh, Surendra K, Shanley, Edward S

Here is a method that rates the tendency of new compounds to self-explode, without the need for performing extensive laboratory tests.

OVER THE YEARS ORGANIC CHEMISTS HAVE synthesized millions of compounds. Several thousand new compounds have been prepared and studied in sufficient detail to be included in the CAS Registry (www.cas.org) during each recent year. The original synthesis is frequently carried out on a very small scale, often less than a gram. On this scale, self-reactivity is usually not a serious concern. When the compound proves to have interesting properties it may be prepared in larger quantities. Self-reactivity hazards as well as toxicity, flammability and other hazards must then be assessed with increasing rigor as scale-up continues.

The ultimate assessment of a self- reactivity hazard often requires physical testing such as adiabatic calorimetry, dropweight impact tests, explosive-initiation tests, self-heating tests and others. Such examination, requiring specialized skills and facilities, can be both tedious and expensive and may require considerable quantities of material.

Computational assessment of reactivity hazard can provide substantial guidance with respect to hazard management. It is fast, inexpensive, risk-free, and applicable in advance of actual synthesis. It is not to be regarded as a substitute for physical testing.

This article provides a comparative means for assigning a hazard index to a particular compound. Utilizing two common properties of these compounds, heat of reaction to form most stable products, (delta)H^sub r^, and calculated adiabatic reaction temperature to form the most stable products (CART) comparisons are made with known hazardous materials to provide a basis for predicting the hazardous nature of lesser-known materials. Five classes of compounds are investigated - hydrocarbons, carbon-hydrogen-oxygen compounds, nitro compounds and nitrates, nitrogen compounds other than nitro compounds and nitrates, and organic peroxides. The evaluative process presented works particularly well for four of the five classes, but not for the organic peroxides. Care should be exercised when working with these compounds.

The hazard assessment described herein is based on the potential for rapid release of stored chemical energy, as exemplified in the case of standard explosives. (The hazard posed by relatively slow pressurization of closed vessels is not considered.) The self-reactivity hazards of concern here are, therefore, kinetic in nature. Available computational methods do not permit ready estimation of reaction rates for fast or explosive reactions. Instead, computational means for reactivity hazard assessment are based upon thermochemistry.

The driving force for any chemical reaction, explosive or otherwise, is measured by the free energy change, (delta)G^sub r^, associated with the reaction. Nevertheless, thermochemical hazard prediction methods in common use are based upon the heat of reaction to form most stable products, (delta)H^sub r^. Rationalization for this practice is based upon the fact that heat of reaction data, (delta)H^sub r^, are more readily available than (delta)G^sub r^ data, and that these two parameters frequently provide similar hazard rankings.

Experience has shown that (delta)H^sub r^ is, in fact, a useful indicator of reactivity hazard potential. Our associates and we have recently demonstrated that the CART is also useful in hazard evaluation (1). Estimation of (delta)H^sub r^ and CART depend upon the availability of values for the standard heat of formation of the compound's composition and its decomposition products. The most stable decomposition products are usually simple compounds with well-known thermochemical properties. The problem of estimating heat of reaction is, therefore, reduced to estimating the standard heat of formation, (delta)H^sub f^^sup 0^, for the composition in question. Means for estimating (delta)H^sub f^^sup 0^, (delta)H^sub r^, and CART and for their interpretation in terms of hazard potential are discussed in what follows.

Estimating (delta)H^sub f^^sup 0^

Calorimetric measurements - Calorimetric estimation of (delta)H^sub f^^sup 0^ for a new compound is typically carried out by measuring the amount of heat released during exhaustive oxidation of the compound and then subtracting the already-known heats of formation of the products resulting from complete oxidation. A tedious and demanding enterprise, precision calorimetry is not a practical means for evaluating the thermochemical properties of the vast number of new compounds synthesized each year.

Literature values - In spite of the difficulties noted above, precision calorimetric data are available for several thousand organic compounds. These data are available in various handbooks and in critical collections such as that of Pedley (2), and of the U.S. Dept. of Commerce, National Institute of Standards and Technology (NIST) databases (3). The carefully reviewed and internally consistent data in such collections constitute the "gold standard" of thermochemistry. Throughout this article, cited (delta)H^sub f^^sup 0^ values are taken from Pedley unless otherwise identified.