Environ. Sci. Technol. 1991, 25, 722-729 Cunningham, R. P.; Asahara, H.; Bank, J. F.; Scholes, C. P.; Salerno, C. P.; Surerus, K.; Munck, E.; McCracken, J.; Peisach, J.; Emptage, M. Biochemistry 1989,28,445C-4455. Kennedy, M. C.; Werst, M.; Telser, J.; Emptage, M. H.; Beinert, H.; Hoffman, B. M. Proc. Natl. Acad. Sci. U.S.A. 1987,84, 8854-8858. Wolfe, R. S. Trends Biochem. Sci. 1985, 10, 396-399. Pfaltz, A.; Juan, B.; Fassler, A.; Eschenmoser, A.; Jaenchen, R.; Gilles, H. H.; Diekert, G.; Thauer, R. Helv. Chim. Acta 1982, 65, 828-854. Walsh, C. T.; Orme-Johnson,W. H. Biochemistry 1987,26, 4901-4906. Juan, B.; Pfaltz, A. J . Chem. SOC.,Chem. Commun. 1988, 293-294. Klecka, G. M.; Gonsior, S. J. Chemosphere 1984, 13, 391-402. Muller, 0.; Muller, G. Biochem. Z. 1963, 337, 179-189. Gossett, J. M. Environ. Sci. Technol. 1987, 21, 202-208. Ho, T.-L.; Wong, C. M. Synth. Commun. 1973,3,237-239. Friedrich, W. Vitamin B12und Verwandte Corrinoide; Georg Thieme Verlag: Stuttgart, 1975; pp 47-55. Johnson, A. W.; Mervyn, L.; Shaw, N.; Smith, E. L. J. Chem. SOC.1963, 4146-4156. Mays, M. J.; Wilkinson, G. Nature 1964, 203, 1167. March, J. Aduanced Organic Chemistry; John Wiley & Sons: New York, 1985; pp 295-304.
(39) Vogel, T. M.; Criddle, C. S.; McCarty, P. L. Enuiron. Sci. Technol. 1987,21, 722-736. (40) Freedman, D. L.; Gossett, J. M. Appl. Enuiron. Microbiol. 1989, 55, 2144-2151. (41) Chapra, S . C.; Canale, R. P. Numerical Methods for Engineers; McGraw-Hill: New York, 1988. (42) Maltoni, C.; LeFemine, G. Enuiron. Res. 1974, 7, 387-396. (43) Smith, E. L.; Mervyn, L.; Muggleton, P. W.; Johnson, A. W.; Shaw, N. Ann. N . Y. Acad. Sci. 1964, 112, 565-574. (44) Mousa, M. A.; Rogers, J. E. Abstracts, Annual Meeting of the American Society of Microbiologists, Anaheim, CA, 1990; Q45, p 296. (45) Fathepure, B. Z.; Vogel, T. M. Abstracts, Annual Meeting of the American Society of Microbiologists, Anaheim, CA, 1990; $38, p 294. (46) Mikesell, M. D.; Boyd, S. A. Appl. Environ. Microbiol. 1986, 52, 861-854. (47) Gibson, S. A.; Suflita, J. M. Appl. Enuiron. Microbiol. 1986, 52, 681-688. (48) Gibson, S. A.; Suflita, J. M. Appl. Environ. Microbiol. 1990, 56. 1825-1832.
Received f o r review July 24, 1990. Revised manuscript received November 14,1990. Accepted November 15,1990. This research was supported by National Institutes of Health Grant GM41235 (to L.P.W.).
Nonequilibrium Sorption and Transport of Neutral and Ionized Chlorophenols Linda S. Lee,” P. Suresh C. Rao, and Mark L. Brusseaut Soil Science Department, University of Florida, Gainesville, Florida 3261 1-0 15 1
w For a series of chlorophenols sorption nonequilibrium was assessed by fitting a bicontinuum sorption model to breakthrough curves measured by miscible displacement techniques. A single log-log inverse relationship was observed between the desorption rate coefficient &, h-l) and the equilibrium sorption constant ( K mL/g) for all chlorophenols as well as for a series o!Lhlorobenzenes. This suggests that approach to sorption equilibrium for neutral and ionized chlorophenols is constrained in a manner similar to that for nonpolar hydrophobic organic chemicals. For neutral pentachlorophenol (PCP), K , values decreased log-linearly with increasing volume fraction methanol (f,),with successful extrapolation of these data to aqueous systems. A log-linear relationship was observed between k, and f, in good agreement with independent estimates. For a series of ionized chlorophenols, the fraction of instantaneous sorption domains (F)increased with increasing solute hydrophobicity, while F decreased with increasing f, for neutral PCP. Introduction
Chlorinated phenols are of environmental concern due to anthropogenic inputs from industrial wastes, degradation of chlorinated pesticides, and the use of pentachlorophenol as a wood preservative. An understanding of the behavior of chlorophenols in the environment necessitates an assessment of the processes influencing their fate and transport in soils and groundwater. In the pH range relevant to most environmental scenarios, some of the more chlorinated phenols can partially or totally ionize. This can drastically modify their aqueous solubility, sorption, transport, and bioavailability. Present address: Soil and Water Science Department, University of Arizona, Tuscon, AZ 85721. 722
Environ. Sci. Technol., Vol. 25, No. 4, 1991
Much of the current research on sorption of chlorophenols is based on equilibrium measurements (1,2). On the basis of an analysis of a large data set for pentachlorophenol (PCP) sorption from aqueous solutions by several sorbents over a broad pH range, Lee et al. (2) showed that equilibrium sorption could be described with a knowledge of pH, organic carbon (OC) content of the soil, and the acid dissociation constant (pK,) for PCP. The influence of sorption nonequilibrium on organic contaminant transport has recently been recognized to be of importance (3-7). Brusseau and Rao (8)have compiled and analyzed an extensive sorption kinetics data base, revealing the existence of an inverse log-log relationship between desorption rate coefficients (kz,h-l) and corresponding equilibrium sorption constants (Kp,mL/g). They noted two distinct relationships, based on the chemical nature of the sorbates (i.e., for nonpolar and polarizable hydrophobic organic compounds). Following Brusseau and Rao (a), sorption nonequilibrium to be considered here is postulated to involve diffusion of sorbate molecules into the interior of the sorbent, which is visualized as a polymeric matrix of organic matter (intraorganic matter diffusion, IOMD). Several authors (9-13) have proposed that soil organic matter may be viewed as a three-dimensional network of polymer chains, exhibiting a relatively open, flexible structure perforated with “voids”. In this study, miscible displacement techniques were used to measure desorption rate coefficients for a series of chlorophenols and to assess the importance of IOMD for both the neutral and ionized forms. The solution matrix was altered in three ways: (i) The solution p H was adjusted to change the degree of ionization of the chlorophenols. (ii) Methanol was added (from 35 to 80 volume %) as a cosolvent. Organic cosolvents increase the solubility of
0013-936X/91/0925-0722$02.50/0
0 1991 American Chemical Society
Figure I.Schematic diagram (not to
scale) of the miscible displacement apparatus.
hydrophobic organic chemicals, thus decreasing their sorption and increasing their mobility in soils (4,14-19). Organic cosolvents also cause swelling of the organic matter matrix (20).thus increasing the accessibility of interior sorptive domains to sorbates. Such solvent effects on the organic matter matrix are reflected generally through an increase in the observed sorption rate coefficients (4). Only pentachlorophenol (PCP) was used in the present study to investigate the effects of cosolvent characteristics on retardation and sorption nonequilibrium. Of the chlorophenols, PCP was considered an ideal candidate, since it is the most acidic (pK, = 4.74) and also the most hydrophobic in its neutral form (log KOw= 5.01). (iii) The electrolyte matrix was varied to examine the effects of charge and concentration of the electrolyte on chlorophenol sorption; formation of neutral ion pairs hetween chlorophenolate ions and inorganic cations ( I , 2,21) is influenced by ionic strength ( p ) and electrolyte type. Investigating the effects of the various electrolyte matrices should help to assess mechanisms involved in the sorption and transport of chlorophenols.
Materials and Methods Sorbents. The sorbent used for this study was a sample of Eustis tine sand (Psammentic Paleudult) collected from the 0-30-cm depth increment at a field site near Gainesville, FL. The particle-size distribution for Eustis soil is as follows: 96.470, 1.8%, and 1.8% by weight for sand, silt, and clay, respectively. The cation exchange capacity (CEC) of the soil was 3.37 mequiv/100 g. The organic carbon content, determined hy the Walkley-Black method (22),was 0.39%. The soil was air-dried and passed through a 2-mm sieve prior to use. Chemicals. The chlorophenols and pentafluorobenzoic acid (PFBA), all >98% pure, were purchased from Aldrich. Tritiated water (3H20) was purchased from New England Nuclear and was used a t an activity of 0.19 Bq/mL. The sodium hydroxide, calcium chloride, calcium hydroxide, hydrochloric acid, potassium chloride, and potassium hydroxide were of reagent-grade purity, purchased from Fisher Scientific. Scinti-Verse I1 scintillation cocktail obtained from Fisher Scientific was used in the liquid scintillation counting (LSC) assay. Concentrations of the chlorophenols in the influent solutions for the column studies were hetween 3 and 12 Fg/mL.
Table I. Summary of Column Parameters column
ID B1
B2 83 D5
L, em
g/em3
mL/cm3
mL
P
5.5 5.5 6.9 5.5
1.75 1.73 1.64 1.69
0.383
10.33
100 (81. 1%) ,.~,
0.379 0.405 0.396
10132
123 (118,135) 166 (108,224) 190 (183,205)
p.
~~
8.
~
~~~
VO,
13.72 9.64
~~~
‘95% confidence limits in parentheses.
Experimental Procedures. Miscible displacement techniques, similar to those described by Brusseau et al. (231, were used to study the transport of two nonreactive tracers (3H20 and PFBA) and a series of chlorophenols. A schematic diagram of the experimental setup is presented in Figure 1. Equilibrium sorption isotherms were measured by using a hatch slurry method described by Lee et al. (21, and all isotherms were linear. A total of four soil columns were used for this study. They will he designated as B1, B2, B3, and D5. The columns were packed with air-dried soil in small increments to establish uniform hulk density. Distinct layering was avoided by mixing the bed surface between increments. Prior to packing, the column apparatus was sterilized by rinsing with methanol, and the soil for column B1 was sterilized with 2450-MHz, microwave radiation (24). A 0.01 N electrolyte solution was pumped through the columns until steady-state, water-saturated conditions were established. Column weights were monitored periodically throughout the miscible displacement studies with variations in column weights of no more than 0.24%. These weights were used to estimate column pore volume, with adjustments made to account for any dead volume attributed to the end plates and the tubing connections. Note that the accuracy of the column void volume is affected by the dead volume estimates, which incur a maximum error of approximately 5%. All solutions were filter-sterilized (0.45 pm) and degassed with helium before use. The length (L, cm), volumetric water content (8, mL/cm3), hulk density ( p , g/cm3), and pore volume (V,,, mL) of the soil columns are given in Table I, along with column identification numbers used in further discussion. The column and solution matrix used for the displacement of chlorophenols varied. In summary, column B1 was used with 0.01 N KCI as the solution matrix, columns B2 Environ. Sci. Technol.. VoI.
25, No. 4. 1991 723
1
*
0.8
0
2 C-
0.0
C 0
6 P)
I
3
C
8
2,3,4,5-Tetrachlorophenol (0.01 NKCI, pHk10)
0.4
-d
,1
I
m
0.2
0 Flow-throughUV detectlon
a
Fractlon
collectlon-HPLC/UVanalysls
0 0
20
40
60
80
100
120
140
160
Data Analysis. Miscible Displacement of Nonretarded Solutes. A pulse of electrolyte solution containing either tritiated water or pentafluorobenzoic acid was displaced through each of the columns as a measure of flow of nonretarded solutes. PFBA has a pK, of 1.49; therefore, in most soil-solutions PFBA is completely ionized (pH >> pK,). Bowman (25)has shown that PFBA can be used as a nonadsorbed, reference solute. The convective-dispersive solute transport equation (26) was fitted to the measured BTCs to estimate values of both the Peclet number ( P ) and the total retardation factor ( R T ) . This model can be stated as follows: RT (dC*/dp) = (1/P)(d2C*/dX2) - (dC*/dX) (1) P = vL/D
Pore Volume, p
Figure 2. Comparison of measured breakthrough curves for 2,3,4,5tetrachlorophenol obtained by flow-through monitoring of column effluent with a UV detector and by collecting effluent fractions followed by HPLC-UV analysis.
and D5 were used for cosolvent studies, and column B3 was used for ionic strength studies with CaCl,. A porewater velocity ( u ) of -90 cm/h was used for all columns. Solute concentrations in the column effluent were monitored continuously and/or by collecting effluent fractions. Continuous monitoring was done using a flowthrough variable-wavelength UV detector (Gilson Holochrome, Waters 450 or LDC UV) connected to a linear chart recorder (Fisher Recordall 5000); this technique has been described by Brusseau et al. (23). Effluent fractions were collected with an Isco fraction collector and then analyzed by HPLC-UV (Waters WISP 710B automated sampler; Eldex 9800 ternary solvent pump; Waters radial compression column module with a C-18 column; Gilson 115 UV detector; HP3392A integrator) or by liquid scintillation counting (Searle Delta 300). To ensure that the effluent monitoring technique did not influence the data output, breakthrough curves (BTC) were measured for partially ionized 2,3,4,5-tetrachlorophenol(TeCP) by both the continuous-monitoring and fraction-collecting methods. Measured breakthrough curves from the two methods yielded comparable results (Figure 2). Similar agreement was observed by Brusseau et al. (23) for the herbicide atrazine and for trichloroethene. As a quality control, influent solutions for each experiment were analyzed by HPLC-UV before and after measurement of the BTC to confirm that solute concentrations in the reservoirs remained constant for the duration of an experiment. Also, in all experiments samples were taken a t the point of maximum effluent concentration and compared with influent concentrations. Column effluent pH was also measured continuously or by collecting fractions a t periodic intervals followed immediately by measurement. A Corning 130 or a Brinkman 686 pH meter was used with a combination glass electrode. Flow-through monitoring of pH was facilitated by the insertion of an Ingold microelectrode (Catalog No. 6030-#3) into a specially bored three-way Teflon tee valve (cell volume ca. 750 pL) connected in-line to the column effluent (see Figure 1). A Hamilton three-way valve was placed between the detector and the pH device in order to standardize the electrode under the same conditions as used for monitoring effluent pH. Since all influent solutions were degassed prior to use, the flow-through p H technique avoids introducing air into the effluent, which can alter the pH. Overall changes in pH during a run were within 0.1 pH unit. Any shifts that occurred were simply due to small differences in the pH values of the tracer and nontracer solutions. 724
Environ. Sci. Technol., Vol. 25, No. 4, 1991
RT = (1
+ pKp/O)
(2)
x
= x/L p = ut/L C* = C/C, (3) where C* is the solute concentration in the column effluent (C, pg/mL) expressed as a fraction of the concentration in the column influent (C,, pg/mL), x is the distance (cm); t is time (h), u is pore-water velocity (cm/h), D is the hydrodynamic dispersion coefficient (cm2/h), and other parameters are as previously described. For breakthrough curves where column effluent concentrations reached the input concentration (i.e, C* = l), retardation factors ( R T )were obtained by calculating the area above the BTC (27). For BTCs that did not reach C/C, = 1, RT values were estimated by moment analysis; i.e., using the first, normalized, absolute temporal moments corrected for a non-Dirac pulse (23, 28). A bicontinuum sorption model (29-31) was used to characterize sorption nonequilibrium. This model assumes that sorption occurs in two types of domains: an instantaneous-equilibrium type and a rate-limited type, with sorption in the latter domain characterized by a first-order rate coefficient, h2 (h-l). Sorption in both domains was assumed to follow a linear sorption isotherm, with Kp and F representing the equilibrium sorption constant (mL/g) and the fraction of equilibrium-type sorptive domains, respectively. Parameters for the bicontinuum model can be grouped into four dimensionless terms (32): P = uL/D (4) RT = (1 + pKp/S) (5)
P
=
[(e + PKfl/(O + PKp)l
(6)
0 = [k2(1 - P)RTLI/U (7) where w is the Damkholer number representing the degree of sorption nonequilibrium, P is the fraction of total retardation attributed to sorption in the equilibrium domain, h2 is the first-order desorption rate coefficient (8,23),and other parameters are as described previously. Parameters for the bicontinuum model were estimated from the measured breakthrough curves by using a curve-fitting program CFITIM (32), based on nonlinear least-squares optimization techniques. The parameters that can be estimated by CFITIM are P , RT, P, w , and the pulse size in units of pore volumes (J). Of these five parameters the pulse size (J)was determined experimentally, while P and RT were determined as previously described. The unknown values of and w for the measured breakthrough curves of chlorophenols were then estimated by using the curve-fitting program, while P, RT, and J values remained fixed.
Results and Discussion
Miscible Displacement of 3H20. Effluent breakthrough curves for both 3H20and PFBA from all columns
Table 11. Summary of Sorption Nonequilibrium Parametersn
solute
pK,
9.37b p-chlorophenol 2,4-dichlorophenol 7.85c 6.94" 2,4,5-trichlorophenol 2,3,4,5-tetrachlorophenol 6.35" 2,4-dichlorophenol 7.85 4.74' pentachlorophenol 4.74 pentachlorophenol 4.74 pentachlorophenol 4.74 pentachlorophenol 4.74 pentachlorophenol 4.74 pentachlorophenol 4.74 pentachlorophenol 6.94 2,4,5-trichlorophenol 2,3,4,5-tetrachlorophenol 6.35
ID B1 B1 B1 B1 D5 D5 B2 B2 B2 B2 B3 B3 B3 B3
column
f,
electrolyte
0 0 0 0 0 0.75 0.35 0.5 0.6 0.75 0 0 0 0
KCl KC1 KCl KC1 CaC1, CaC1, CaC1, CaC1, CaC1, CaC1, CaCl, CaCl, CaCl, CaCl,
ji
0.01 0.01 0.01 0.01
0.015 0.015 0.015 0.015 0.015 0.015 0.0038 0.0155 0.0155 0.0155
pH
% ionized
K,, mL/g
k,, h-l
0 2 8 47 0 0 0 0 0 0 100 100 100 100
0.19 0.70 1.85 4.19 0.62 0.05 4.77 1.17 0.57 0.12 0.45 0.81 0.14 0.26
46.5 (41.1, 51.9) 10.6 (10.0, 11.2) 7.16 (5.99, 8.33) 2.07 (1.48, 2.64) 8.96 (8.27, 9.65) I125 (120, 130) 1.49 (1.05, 1.93) 4.38 (4.09, 4.67) 6.91 (6.57, 7.26) 31.5 (30.7, 32.3) 14.2 (13.3, 15.1) 10.2 (9.8, 10.6) 25.0 (24.7, 25.4) 22.7 (22.5, 22.9)
6.8 6.1 5.9 6.3 5.5