Envlron. Scl. Technol. 1988, 22, 470-472
with H,. When sulfidation was carried out in the absence of H2, as in period C6 (Figure l),H2S consumption was much higher and SO2was formed in larger amounts. The constant level of SO2 observed before breakthrough is an indication that reduction of V205is performed by means of reaction 8. An important aspect in relation to the desulfurization of a coal-derived fuel gas is the effect of H20 on the interaction of H2Swith V205or the reduced vanadium oxides present. In sulfidation period C3 in the above batch of experiments, 7 mol% H 2 0was added to the reaction gas. The resulting H2S retention was found to be much lower (Figure 1)and the breakthrough sharper, suggesting that H20 and H2S compete for the same adsorption sites. Because of its low sulfur capacity in the presence of water vapor, vanadium oxide is not a practical sorbent for desulfurization of coal-derived fuel gas, which usually contains 10-30% water. I t could be of some utility for desulfurization of fuel gas of very low water content. Registry No. H2S, 7783-06-4; VzOs, 1314-62-1; V203,7s, 12165-50-3.
Literature Cited (1) M E R C Hot Gas Cleanup Task Force MERC/SP-78/2;1978.
(2) Westmoreland, P. R.; Harrison, D. P. Environ. Sci. Technol. 1976, 10, 659. (3) Westmoreland, P. R.; Gibson, J. B.; Harrison, D. P. Enuiron. Sci. Technol. 1977, 11, 488. (4) JANAF Thermochemical Tables;Stull, D. R., Prophet, H., et al., Eds.; U S . Government Printing Office: Washington, DC, 1971; NSRDS-NBS 37; Supplement, J.Phys. Chem. Ref. Data 1974,4, 1. (5) Jalan, V.; Desai, M.; Frost, D.; Wu, D. “Final Report to DOE on Contract 31-109-38-5804”;Giner Inc.: Waltham, MA, 1981. ( 6 ) NBS Technical Note (U.S.) 1971, No. 270-5. ( 7 ) Mills, K. C. Thermodynamic Data f o r Inorganic Sulfides, Selenides and Tellurides; Butterworth: London, 1974. (8) Jalan, V.; Desai, M.; Brooks, C.; Waterhouse, R. Proceedings of the Third Annual DOE Contractors’ Meeting on Contaminant Control in Hot Coal Derived Gas Streams; DOE/METC: Morgantown, WV, 1983; DE64000216. (9) Matsuda, S.; Kamo, T.; Imahashi, J.; Nakajima, F. Ind. Eng. Chem. Fundam. 1982,21, 18. (10) Tamhankar, S. S.; Bagajewicz, M. J.; Gavalas, G. R.; Sharma, P. K.; Flytzani-Stephanopoulos, M. Ind. Eng. Chem. Process Des. Dev. 1986, 25, 429. (11) Fukuda, K.; Dokiya, M.; Kameyama, T.; Kotera, Y. Ind. Eng. Chem. Fundam. 1978,17, 4.
Received for review February 4,1987. Accepted October 10,1987.
Adsorptive Displacement Analysis of Many-Component Priority Pollutants on Activated Carbon. 2. Extension to Low Parts per Million (Based on Carbon) Sharad Thakkar and Milton Manes” Chemistry Department, Kent State University, Kent, Ohio 44242
An earlier paper described the analysis of multiple trace contaminants on activated carbons that are extracted by adsorptive displacement, i.e., equilibration in a solvent (e.g., dichloromethane) containing a large excess of a strongly adsorbing solute/ displacer (e.g., benz [a ]anthracene-7,12-dione). The method was previously applied to the simultaneous determination of 25 base-neutral priority pollutants, including such strongly adsorbed pollutants as benz[a]anthracene, at loadings less than 0.1 mg/g. The lower limits for detection and analysis have been extended downward to the low-ppm range based on carbon. The method should be applicable to the analysis of strongly adsorbed organic impurities in water at levels that are too low for solvent extraction methods and to the monitoring of carbon beds in water purification plants.
Introduction A preceding paper (1) described the simultaneous determination of multiple trace contaminants on activated carbon by the application of adsorptive displacement, a process in which the carbon sample is equilibrated in a good solvent (e.g., dichloromethane) with a high concentration of a strongly adsorbed compound (“displacer”),e.g., benz[a]anthracene-7,12-dione.Under these circumstances, many of the trace contaminants go completely into solution; most of the others exhibit isotherms that are linear, with slopes that are mutually independent (I, 2) and with zero intercept. The principal exceptions are phenolic compounds, which exhibit nonlinear isotherms at low capacities (fractions of a milligram per gram); these com-
* Address correspondence to this author at.5 Bayard Road, No. 412, Pittsburgh, PA 15213. 470
Environ. Sci. Technol., Vol. 22, No. 4, 1988
pounds also exhibit linear isotherms with zero intercepts when equilibrated with phenolic displacers. For contaminants that are not completely extracted into the solvent, the amount of each unextracted compound is readily calculated from the (predetermined) slope of its isotherm (2).
The previous paper (1) described the application of adsorptive displacement to the simultaneous determination of some 25 base-neutral priority pollutants at loadings down to about 0.1 mg/g carbon, even for such strongly adsorbed (“refractory”)compounds as anthracene, phenanthrene, and benz[a]anthracene. Although these levels were considered sufficiently low to suggest the potential power of the method for the analyses of pollutants in drinking waters at extremely low concentrations, it seemed likely that these already low limits could be extended downward by at least an order of magnitude by the application of relatively simple techniques. This paper describes the improvements in sensitivity that were attained.
Experimental Section Except as noted, the materials and experimental conditions were essentially the same as in the previous paper (I), except that the solvent here was dichloromethane (without methanol addition). Attempts were first made to improve the sensitivity of the analysis by concentrating the filtered equilibrium SOlutions by solvent evaporation under ambient conditions. These attempts were unsatisfactory, partly because of loss of some of the more volatile components, even at only tenfold concentration, but also because of the appearance of new impurity peaks with increasing concentration. The desired improvement came about from adjustments in the gas chromatographic technique that made it possible to
0013-936X/88/0922-0470$01.50/0
0 1988 Amerlcan Chemlcal Society
I
CONC (MG/L)
Flgure 1. Adsorption Isotherms of preloaded adsorbates showing nonzero Henry coeff lcients with benz [a ]anthracene7,12dione as the displacer at 25 O C : (0) 2-nltrophend, (A)benz[b]fluoranthene, and (0) dlbenz[a ,h]anthracene.
analyze for components at 0.1 ng (i.e., at 0.1 ppm in a 1 pL sample of equilibrium solution). The average standard deviation for repeat runs was up to &lo%. As in the earlier work, all displacers were individually equilibrated with similarly loaded carbon samples. The equilibration was carried out at 25 OC in 125-mL screwcapped Erlenmeyer flasks with Teflon cap liners, which were shaken in a shaker bath at 25 OC. Preloaded carbon samples were used throughout; they were prepared (as in earlier work) by adding the carbon to a solution of the multiple adsorbates in methylene chloride, followed by equilibration for at least 48 h and slow evaporation of solvent to dryness. This preloaded carbon was then shaken with a solution of displacer. Benzanthracene-7,12-dione and anthraquinone were used as displacers in saturated solution. For pentachlorophenol and 4-chloro-3-methylphenol (used to displace phenols) their higher solubility in methylene chloride made the use of saturated solution awkward; in these cases, 1 g of displacer was used in 25 mL of solvent. The equilibration was continued for at least 7 days, after which time the carbon was allowed to settle. The drained supernatant liquid was cleared by centrifuging in a screw-capped centrifuge tube. A measured sample of the cleared solution was immediately spiked with a measured volume of internal standard (acenaphthene) solution and stored for analysis. The clear spiked solutions were analyzed in a Hewlett-Packard HP-5890 gas chromatograph equipped with a flame ionization detector. The data reported here were obtained with a 60-m Supelco SPB-5 wide-bore (0.75 mm) glass capillary column with a methyl silicone-phenyl silicone coating. Quantitative calculations were carried out by an HP-3392 integrator in peak height mode. The results were used to plot adsorption isotherms as mg of adsorbatelg of carbon vs equilibrium concentration in mg/L.
Results and Discussion Earlier results (1)were for the preloaded carbon in the loading range of 0.1-2 mg/g of carbon. The present results are for the lower loading range of 0.001-0.07 mg/g of carbon. With benzanthracene-7,12-dioneas a displacer, the following compounds showed essentially complete displacement from the carbon: 1,4-dichlorobenzene,hexachloroethane, 1,2-dichlorobenzene, isophorone, nitrobenzene, naphthalene, 1,2,4-trichlorobenzene, hexachlorobutadiene, 2,6-dinitrotoluene, diethyl phthalate, phenanthrene, 4-bromophenyl phenyl ether, fluoranthene, anthracene, and benzyl butyl phthalate. Large errors in quantitation of dioctyl phthalate at these low concentrations resulted from peak overlap with the large displacer peak. Figures 1and 2 show the linear adsorption isotherms of the adsorbates with nonzero Henry coefficients. (Neither phenol nor 4-chloro-3-methylphenol showed any detectable displacement at these low loadings.)
CONC. (MG/L)
Flgure 2. Adsorption Isotherms of preloaded adsorbates showing nonzero Henry coefficients wlth benz[a]anthracene7,12 dione as the displacer at 25 OC: (0) fluorene and (A) 2-chloronaphthalene.
m-
u
5
002
a Oo0
020406081012
020406081012
CONC.(MG/L) Flgure 3. Adsorption isotherms of preloaded refractory adsorbates with anthraquinone as the displacer at 25 OC: (0)benzanthracene and (A) benz[b]fluoranthene. Dlbenz[a ,h]anthracene showed no displacement.
CONC (MG/L)
Flgure 4. Adsorption isotherm of preloaded 2-nitrophenol with pentachlorophenol as the dispiacer at 25 'C. Phenol and 4-chloro-3methylphenol showed no displacement.
Figure 3 shows the results with anthraquinone as a displacer for the more refractory benz[a]anthracene and benz[ blfluoranthene. Both show linear adsorption isotherms with Henry coefficients of 0.45 and 0.55 L/g respectively. Dibenz[a,h]anthracene showed essentially no displacement by anthraquinone. Figure 4 shows results with pentachlorophenol as a displacer for phenolic compounds. 2-Nitrophenol showed linear adsorption behavior with a Henry coefficient of 0.056 L/g. Both phenol and 4-chloro-3-methylphenol showed essentially no displacement by pentachlorophenol. Finally, Figure 5 shows the results with 4-chloro-3-methylphenol as a displacer for phenolic compounds. In this case, both phenol and 2-nitrophenol show linear adsorption behavior with Henry coefficients of 0.025 and 0.008 L/g, respectively. We first consider the results with benzanthracene7,Wdione as a displacer. Fifteen adsorbates that had been completely displaced (at higher loadings) in earlier work (I) were also completely displaced here. On the other hand, phenol and 4-chloro-3-methylphenol, which had earlier exhibited nonlinearity at higher loadings, were not detectably displaced at all. The other adsorbates, as shown in Figures 1 and 2, exhibit the now familiar linear isoEnviron. Sci. Technol., Vol. 22, No. 4, 1988
471
0.06
T
3 002 n
a
Oo0
0 2 0 4 0 6 0 8
1 12 14 16 18 (MG/L)
2
22
CONC
Flgure 5. Adsorptlon isotherm of preloaded phenolic adsorbates with 4-chloro-3-methylphenol as the displacer at 25 O C : (0)2-nitrophenol and (0)phenol.
therms with zero intercepts. Differences between the slopes and those previously observed may be ascribed to the methanol in the solvent in the previous work, which would be expected to affect some solubilities and therefore the slopes. The importance of a proper displacer for the more refractory components is shown by a comparison of the data in Figure 3 (with anthraquinone as the displacer) and the data for the same adsorbates (benz[a]anthracene, benz[blbifluoroanthene, and dibenz[a,h]anthracene) in Figure 1. Not only are the two plotted adsorption isotherms considerably steeper with anthraquinone, but dibenz[a,hlanthracene, which is well recovered in Figure 1,was not detectably recovered with anthraquinone. Turning now to the phenolic compounds (Figures 4 and 5 ) , we find that although both pentachlorophenol and 4-chloro~3-methylphenalare phenolic derivatives, pentachlorophenol showed poor displacing power; neither phenol nor 4-chloro-3-methylphenolwere detectably displaced by pentachlorophenol at these low levels (Figure 4). On the
472
Environ. Scl. Technol., Vol. 22, No. 4, 1988
other hand, 4-chloro-&methylphenol is a strong displacer for both phenol and 2-nitrophenol, which exhibited linear displacement behavior with low Henry coefficients (Figure 5). One may speculate that the specific adsorption of the phenol is via the donation of a pair of electrons from the oxygen atom, in which case electron-withdrawing groups such as chloro and nitro groups would be expected to reduce their reactivity; this could explain the relatively weak displacing power of pentachlorophenol observed here, and of 2-nitrophenol observed earlier ( I ) , and would also account for the invariably lower Henry coefficients of 2nitrophenol as compared to phenol. Although the adsorption of phenolic compounds is specific, present results indicate it to be reversible. In summary, the power of the method is illustrated by the detection of dibenz[a,h]anthracene at loadings below 10 pg/g of carbon and with no detectable irreversible adsorption. Applicability and limitations are discussed in the previous paper ( I ) . Registry No. C, 7440-44-0; HzO, 7732-18-5; 2-O2NC6H,OH, 88-75-5; C&&OH, 108-95-2; CBCl6OH,87-86-5; benz[ blfluoranthene, 205-99-2; dibenz[a,h]anthracene, 53-70-3; fluorene, 86-73-7; 2-chloronaphthalene, 91-58-7; benzanthracene, 56-55-3; benz[a]anthracene-7,12-dione, 2498-66-0; 4-chloro-3-rnethylpheno1, 59-50-7; anthraquinone, 84-65-1.
Literature Cited (1) Thakkar, S.; Manes, M. Environ. Sci. Technol. 1987, 21, 546-549. (2) Gu, T.; Manes, M. Enuiron. Sci. Technol. 1984,18, 55-57.
Received for review March 30,1987. Revised manuscript received September 14,1987. Accepted November 11, 1987. This work was supported by a grant ( U S . EPA/R811314-01-0) from the Environmental Protection Agency.
