From this plot we may determine (for example) that for a typical atmospheric lifetime of toluene of 10 hours the concentration of p-cresol (which reacts six times faster than toluene with OH) should reach a maximum in 3.5 hours. If the yield of o-cresol is 5 percent (see below) and the ambient toluene concentration is .020 ppm, this maximum concentration calculated from eq. (1) is about 0.2 ppb.

Assuming pseudo first order toluene loss the variation of a product with toluene concentration is given by

For the simple case where a product is totally unreactive, R, = 0 and a plot of P, vs. T will give a straight line with slope = α. For the

case where the product does react further, the second term on the left hand side of eq. (4) corrects for this loss of product. The variation of Q-cresol and of benzaldehyde for one of our experiments are shown in figures 3 and 4. This experiment was carried out by irradiating toluene and NO2 each at about 4 ppm in a 250 L evacuable glass vessel with a mixture of fluorescent black lights and sun lamps.

The yield of each product may be determined from the slope of these plots. For benzaldehyde we obtain a 2.5 percent yield and for o-cresol a 5 percent yield. These yields are much lower than those measured by Hendry in his low pressure flow system.

We have been initially skeptical of our low yields, especially for o-cresol since it is about ten times lower than the yield reported by Hendry. To double check this result we have carried out experiments in which we start with a mixture of toluene and o-cresol (4 ppm and 1 ppm) respectively). The decay of o-cresol is then modified by formation of -cresol from toluene. Equation (3) holds for any initial product concentration so we have plotted the data for this experiment in the required form in figure 5. The yield of p-cresol is found from the slope to be 5 percent, in agreement with the other experiments. This experiment has the advantage of generating a large

i

The values of a; may be obtained from the individual rate measurements of Atkinson and others. However, considering the errors present in each measurement a separate measurement of the ratio itself may be preferable. We have made such measurements by irradiating a mixture of P. and toluene at about a 10 to 1 ratio (P;/T). A' plot of In P, vs. In T gives the value of R as the slope. This analysis is not sufficient if the product photolyzes to any appreciable extent. For the case of o-cresol the photolytic lifetime in our reaction vessel is 10 minutes and is probably negligible.

For the more general case of non first order toluene loss (variable OH concentration) we may still derive an expression to analyze production formation and loss. For the same mechanism given above (reactions a and b) it can be shown that

1

Slope = .05

Importance

Study of the atmospheric reaction processes of aromatic hydrocarbons is in its early stages. Our current knowledge about these compounds is rather primitive compared to the alkanes and alkenes. However, aromatics are major components of urban atmospheres and elucidation of their reaction pathways is essential for the following reasons.

1) Oxidant formation. As major urban hydrocarbons which react relatively rapidly with hydroxyl radical, aromatics will contribute directly to ozone formation and buildup and they seem to generate appreciable quantities of PAN, itself a harmful oxidant.

2) Direct health effects. Oxidant products of aromatic hydrocarbons are poorly characterized and present a potential health hazard of undetermined magnitude.

3) Aerosol formation. Gas phase mass balances for smog chamber experiments with aromatics are very poor and may indicate appreciable aerosol formation. If so, the aerosol so formed may contribute to a heterogeneous component of tropospheric chemistry which is currently unrecognized; this heterogeneous component may well impact other areas in particular NO, conversion to nitric acid or free radical loss processes.

Current Status

1) Rate constants. Considerable work has been done to determine the reaction rate constants for hydroxyl radicals with the chief aromatic constituents of the atmosphere. Agreement between various groups is quite good so this question is resolved. Ozone and other free radicals (H02, NO3) are known to react slowly with aromatics and are therefore, at present of minor importance. Relative rates of ring addition versus side chain abstraction, while less certain than the overall rates, are also fairly well settled.

Rate constants for reaction of OH with some of the more important reaction products of aromatic hydrocarbons (cresols, benzaldehyde, etc.) have

also been measured.

2) Product identification. Major products of the reaction of OH with toluene which have been identified are the following: cresols, nitrotoluenes, benzaldehyde, benzyl alcohol, benzyl nitrate, peroxybenzoyl nitrate, peroxyacetyl nitrate, and carbon monoxide. Yields of these compounds have been determined at low pressure and are becoming available at atmospheric pressure as well. Currently the low pressure yields are considerably higher on an absolute basis than those at high pressure, but on a relative basis are in good agreement. Products have been identified for the reactions of some other aromatics as well.

Products of the subsequent reactions of the primary products are known in a few cases.

The hydroxyl radical is a key component in controlling loss rates of the primary products but other processes, such as direct photolysis or reaction with 03, RO2, RO, NO3, etc. may be important as well.

3) Ozone formation. Only a limited amount of work has been done on modeling the ozone profiles in aromatic/NO or mixed hydrocarbon/NO systems which include aromatics, because of the general lack of knowledge about the detailed reactions involved. However, the limited work done to date indicates that ozone profiles are different than those in nonaromatic systems and in some cases are difficult to model unless unique radical-radical reactions are invoked.

4) Analytical techniques. Current studies of aromatic hydrocarbon systems are severely hampered by a lack of versatile techniques for analyzing the high molecular weight products involved. Techniques which have been employed include gas chromatography, gas chromatography-mass spectrometry and to a limited extent Fourier transform infrared spectrometry. These techniques are difficult to employ when they are successful, and are often unsuccessful. Much time has been spent in adapting these techniques for use in the study of aromatic hydrocarbons, but they still suffer from some inherent problems.

Recommendations

1) Absolute yields of the major known primary products of aromatic-OH reactions should be determined at atmospheric pressure. These aromatics would include as a minimal set benzene, toluene, the xylenes, trimethyl benzene and some alkyl benzenes such as ethyl benzene.

2) Rate constants for the various processes these products undergo should be determined.

Although a large number of compounds are involved, competitive kinetic studies employing several compounds simultaneously may suffice in some cases. This would reduce the total number of necessary experiments.

3) A carbon mass balance for the gaseous products including CO and CO2 should be obtained for the major aromatics. The mass balance should include the carbon content of any aerosol formed.

4) New analytical techniques should be investigated for application to the study of aromatics. These techniques would be doubly useful because they would be equally applicable to the study of higher moleculas weight alkanes and alkenes. Techniques which might be investigated include improvement of gc sampling techniques and separation efficiency on the column, direct mass spectral analysis employing non-framentation ionization, liquid chromatography, and field desorption mass spectrometry.

In all these techniques every attempt should be made to work at realistic reactant concentrations and total pressures and to induce minimal sample alteration. However, some low pressure techniques may have to be employed (e.g., direct ms sampling) because of the lack of any other viable alternatives for direct analysis of intermediates. The current advancement of knowledge in this area is now limited by analytical methodology. Advancement of knowledge in the alkane and alkene systems will soon suffer the same fate, as the chemistry of the low molecular weight compounds becomes worked out, and higher members of the series are studied.

5) Heterogeneous processes may be of great importance in the aromatic hydrocarbon systems. The impact of these processes may well extend beyond the purely aromatic systems and influence the chemistry of NO and of free radicals generated from other classes of compounds. An attempt should be made to assess the significance of these processes on the overall chemistry of the troposphere.

6) Compouter modeling of the aromatic hydrocarbon system should be continued in order to assess the ozone forming potential of these hydrocarbons. It is expected that these modeling efforts will become more meaningful as more basic rate and product data become available.

Recommendations 1 to 3 may be expected to be completed with current funding in the next year or two. Recommendations 4 and 5 are much more ambitious and will require a long term committment and considerable additional funds for instrument development.