166 GEHRS 



the effluent, contributed 42% of the total toxicity. The acid fraction, 

 which was almost 29% of the effluent, contributed about 52% of the 

 toxicity. Preliminary characterization work showed naphthalenes, 

 phenanthrene, and anthracene in the neutral fraction and phenols 

 and cresols in the acid fraction. Such an approach provides useful 

 information (identification of active components) to the design 

 engineer and control technologist through a process of selective 

 elimination. 



Parkhurst, Gehrs, and Rubin (1978) also compared the toxicity 

 of an unfractionated effluent wdth a reconstituted effluent (various 

 fractions mixed together in same concentration as initial effluent). 

 That they found no significant difference in results suggests the 

 validity of fractionation and toxicity testing as an initial step in 

 evaluating complex chemical mixtures. 



Both approaches have several weaknesses, however. The identities 

 and concentrations of the specific compounds being delivered are 

 unknown. Interaction* are masked. Hence, developing predictive 

 information from these types of studies is not possible. Using 

 chemical fractionation rather than boiling point to separate complex 

 mixtures appears to have several advantages, the most important 

 being that separation is into similar groups (phenolics, aromatics, 

 etc.) rather than across groups by boiling point. By combining the 

 fractionation— subfractionation scheme with subsequent molecular 

 weight profiles, we can obtain a rough estimate of the amount of the 

 various classes, as well as the distribution (by weight) within the 

 various classes. Using the data developed by Herbes, Southworth, and 

 Gehrs (1977) (discussed in the following section) enables initial 

 prediction of both potential effects and causative agents. 



Neither of these screening approaches provides the information 

 needed to understand mechanisms of effect, however, and, hence, 

 neither permits predictive capabilities. Predictions can be made only 

 by using specific compounds in studies of transport, fate, and effects. 



LIQUID EFFLUENT COMPOSITION 



No data are available on the exact composition of effluents that 

 might be released from coal-conversion facilities. Herbes, 

 Southworth, and Gehrs (1977) used data from the coking industry 

 as a surrogate to estimate potential effluent components. They 

 Eirranged the compounds into five groups on the basis of chemical 

 structure and used available data on anticipated concentrations in 

 liquid effluents, plus efficacy of waste-water treatment methods, to 

 calculate expected concentrations released to the environment. 



