and sample collection, shipping, storage, and analysis by any of t h e three techniques.
Table V. Values and Precision of Trihalomethane (THM) Concentration in Tap Water Samples as Determined by the Solvent Extraction Technique and Using the Electron Capture Detector THM concn ( p g / L ) and re1 SD ( % ) CHC13 CHBrC12
sample
pg/L
Yo
A
31 60 98 168
1
B
C D
2 3 3
%
2.2 4.9 8.4 11.1
1
1 12 3
Acknowledgments Review of this manuscript by Guy L. LeBel a n d Valerie M. Douglas is gratefully acknowledged.
CHBr2CI
pglL
Yo
Literature Cited
0.1 0.5 1.0 2.6
3 2 6 2
(1)
Rook, J. J . , J . Water Treat. Exam., 23, 234-43 (1974). ( 2 ) ‘Tardiff, R. G., J . Am. Water Works Assoc., 69, 658-61 (1977). ( 3 ) Symons, J . M., Bellar, T. A,, Carswell, .J. K., DeMarco, J . , Kropp, K. L., Roheck, G . G., Seeger, D. R., Slocum, D. J., Smith, B. L., Stevens, A. A,, J . Am. Water Works Assoc., 67, 643-7 (1975). (4) Keith, L. H., “Identification & Analysis of Organic Pollutants in A‘ater”, Ann Arbor Science Publishers, Inc., Ann Arbor, Mich.,
sion. T h e HS technique showed relatively poor precision a n d inferior sensitivity. T h e sensitivity of t h e SE a n d HS techniques can be improved by increasing t h e volume of t h e aliquot injected into the gas chromatograph, but overloading of t h e E C detector by organohalides m u s t be avoided. T h e relatively poor precision for calibration solutions as compared to the results for t a p water is probably due to lack of precision in t h e spiking procedure. Both t h e HS and SE techniques allowed processing of 6 s a m p l e d h , whereas only 2.5 samples/h could be analyzed by t h e G S technique. Also, t h e G S technique required some specialized equipment a n d minor modification of t h e gas chromatograph. In view of its precision, relative accuracy, simplicity, and speed of analysis t h e SE technique was judged t h e most suitable for monitoring T H M s in water. Culture tubes equipped with screw caps and Teflon-coated silicone disks are suitable for calibration solution preparation
1976. ( 5 ) ”National Survey for Halomethanes in Drinking Water”, Health and U’elfare Canada. 77-EHD-9, 1977. (6) Bush, €3.. Narang, R. S., Syrotynski, S., Bull. Ent>iron.Contam. T O X i c O l . , 13, 436-41 (1977). (71 Morris. R. L.. .Johnson. L. G.. J . Am. Watc3r Works A ~ s o c 68. .. ,
I
492-4 (19761. (8) Richard, .J. ,J., dunk, C;. A , , J . Am. Water Works Assoc., 69, 62-4 (1977). (9) Mieure, J . P., J . Am. Water Works Assoc., 69,60-2 (1977). (10) Smillie, R. D., Nicholson, A. A., hleresz, O., Duholke, W. K., Rees, G. A. V., Roberts, IC. Fung, C., “Organics in Ontario Drinking Raters, Part 11. A Survey of Selected Water Treatment Plants”, Ontario Ministry of the Environment, April 1977. (11) Bellar. T. A , , Lichtenberg, J . .J., J . Am. Water Works A.>,\oc.,66, 7:19-44 (1974). (12) Nicholson, A. A,, Meresz, O., Lemyk, B., Aria/. Chem., 49,814-9 (1977). Keceiced f o r r e v i r u September 26, 1978. Accrpted April 9, 1979
Selective Adsorption of Organic Homologues onto Activated Carbon from Dilute Aqueous Solutions. Solvophobic Interaction Approach and Correlations of Molar Adsorptivity with Physicochemical Parameters Georges Belfort Department of Chemical and Environmental Engineering, Rensselaer Polytechnic Institute, Troy, N.Y. 12181
T h e solvophobic interaction theory with simplifying assumptions is used as a basis for correlating t h e molar a d sorption capacity of different single organic solutes such as aliphatic alcohols, aldehydes, ketones, organic acids, a n d substituted aromatic compounds with physicochemical (polar a n d steric) structural parameters. T h e well-known effects of molecular weight, polarity, solubility, branching, and type of aromatic substituents on adsorption a r e discussed in terms of these smooth correlations. From preliminary evidence, these correlations also appear to be useful in predicting t h e competitive adsorption selectivity of homologues from two-, three-, a n d five-solute containing admixtures. Althoughactivated carbon adsorption has been widely used for odor and color removal in the water industry, only recently has it been seriously considered for dissolved organic removal from water supplies ( I ) . T h i s is in major p a r t due to t h e growing concern for potential carcinogenic, mutagenic, a n d teratogenic compounds found in drinking waters ( 2 , 3 ) . Interest has recently intensified because of t h e Environmental Protection Agency’s (EI’A) proposal for a maximum containment level (MCL) of 100 parts per billion ( p p b ) for total trihalomethanes (TTHM) in drinking water ( 4 ) . Fur0013-936X/79/0913-0939$01.00/0
thermore, it has been proposed that granular activated carbon (GAC) treatment be used t o remove T T H M from t h e water so as to comply with t h e standard. These proposals have illicited t h e anticipated vociferous response from water utility managers, State Environmental Protection Agencies ( 5 ) ,and t h e American Water Works Association ( 6 ) ,and it is not clear whether they will be promulgated into law or not. What is clear, however, is that the 309 volatile organic compounds and 55 pesticides that so far have been identified in drinking water will be closely monitored, a n d t h a t those compounds which are found to be a serious health hazard will have to be removed or reduced in concentration ( 7 ) .T h e fact t h a t E P A has proposed GAC as t h e probable removal method indicates t h e cardinal role it will probably play in the future. G r a n u l a r Activated Carbon Whether used in t h e powdered (PAC) or granular form (GAC), activated carbon is particularly effective in removing m a n y of t h e microorganic contaminants found in drinking water. T h e efficiency of GAC for removing biologically derived odors of industrial origin, sulfide odors, pesticides. hydrocarbons, and haloforms is reviewed in detail by McCreary and Snoeyink ( 1 ) . Chow a n d David present results of a study of organic compounds in municipal wastewater that are resistant
@ 1979 American Chemical Society
Volume 13, Number 8, August 1979 939
t o carbon-adsorption treatment (8).T h e y suggest t h a t t h e most resistant organic compounds are small molecules (four primary carbons) with weak branching, Q; is reduced relative to the nonbranched alcohol, but since the molecule is still a flexible rodlike molecule, it behaves according to the linear alcohol Q, vs. u* correlation (Le., solute no.’s 1 through 6) in Figure 1. Examples of this behavior are shown in Figure 1 for 3methyl-1-pentanol and 5-methyl-1-heptanol (solute no.’s 10 and 11, respectively). These solutes are named 2-ethylbutanol and 2-ethylhexanol, respectively, by Giusti e t al. (22).Thus,
I
18 16 14 12 IO
I
I
I
I
I
I
I
15
-
8 -
ALOEH YOES \ ,
6 -
i
4 2 -
o - - w -020
I
-010
00
0 IO
040
0.50
-10
1
-08
-06
U*
Figure 2. Effect of polar parameter on molar adsorption on activated carbon. Solute numbers for ketones and aldehydes same as in Table I. Adsorption data from Giusti et al. (22)
I
00 t o 2
-02
-04
XES Figure 4. Effect of steric parameter on molar adsorption on activated carbon. Solute numbers for ketones and aldehydes are the same as in Table I. Adsorption data from Giusti et al. (22)
34 Q
-
-020
-010
00
+O 10 “ + 0 40
+O 50
U*
0
2
6
4
8
29
IO
E
Figure 3. Effect of polar parameter on molar adsorption on activated carbon. Solute numbers for organic acids are the same as in Table I. Adsorption data from Giusti et al. (22)
Figure 5. Effect of normalized composite relative IR shift parameter on molar adsorption on activated carbon. Solute numbers for aromatic organics are the same as in Table II. Adsorption data from AI-Bahrani and Martin (23)
with slight branching 3-methyl-1-pentanol (solute no. 10) has a more reduced rnolar adsorbability Q:, than the linear n-amyl alcohol (solute no. 5). Both, however, are on t h e same correlation line. (ii) Ketones a n d Aldehydes. An asymptotic trend similar t o t h a t for alcoh.ols is shown in Figure 2 for ketones and aldehydes. Both curves tend to a maximum Qill as g X tends to a value of about -0.130. As the substituted group increases in carbon number, so Q; increases. For t h e aldehydes, Q”, reaches a minimum for acetaldehyde (solute no. 18) of about 5 x 10W g-mol/g of carbon. As in Figure 1for t h e alcohols, branching of the substituent (solute no. 15) results in its Q, vs. a* coordinate being slightly off t h e primary correlation (Le., for solutes 12, 13, and 14). Also, stabilization (i.e., decreased motional degrees of freed o m ) of t h e two substituents on t h e ketone by cyclization (solute no. 16) results in a decrease in adsorhahility. Conversion from aldehydes t o ketones by t h e replacement of a hydrogen by a methyl group for t h e related pairs acetaldehyde/acetone (solute no.’s 18/12], propionaldehyde/methyl ethyl ketone (solute no.’s 19/13), and hutyraldehyde/methyl propyl ketone (solute no.’s 20/14) results in increased adsorhabilities (AQ,) of2.41 X 3.23 X 10-1,and 1.44 X 10-4 g-mol/g of carbon, respectively. (iii) Organic Acids. Once again a similar relationship to that observed above is depicted in Figure 3 for organic acids. As the polarity of the solute o* is decreased below a* = 0.0, a sudden rise in t h e adsorbability is observed. Once again maximum Q, is reached as n* approaches a value of about -0.130. ( b )Steric Effects. Ketones and Aldehydes. Since polarity of functional groups is thought to follow t h e order (70’, 76)
COOH > O H > C=O > C-0-C, dominant steric effects are most likely to occur for aldehydes, ketones, a n d ethers. T h i s is shown as a function of adsorption in Figure 4 for ketone and aldehyde functional groups for a series of substituents [CH:3, CHsCH: