Environ. Sci. Technol. lQQ2,26, 2483-2489
1st ed.; Cambridge University Press: Cambridge, England, 1986; pp 114-115,289-293. (21) Scraton, R. E. Basic Numerical Methods: An Introduction to Numerical Mathematics on a Microcomputer, 1st ed.; Edward Arnold London, 1984, pp 61-63. (22) Buffle, J.; Altmann, R. S. In Aquatic Surface Chemistry,
l e t ed.; Stumm, W., Ed.; Wiley-Interscience: New York, 1987; Chapter 13, pp 351-383. Received for review March 17,1992.Revised manuscript received July 21, 1992. Accepted August IO, 1992.
Dissolution of Monoaromatic Hydrocarbons into Groundwater from Gasoline-Oxygenate Mixtures Mette Poulsen, Lloyd Lemon,+ and James F. Barker’
Waterloo Centre for Groundwater Research, University of Waterloo, Waterloo, Ontario, Canada N2L 301
rn The effects of the “oxygenate” additives methanol and methyl tert-butyl ether (MTBE) on the aqueous solubility of benzene, toluene, ethylbenzene, and xylenes (BTEX) from gasoline were evaluated through equilibrium batch experiments. For a gaso1ine:water ratio of 1 : l O (v/v), up to 15% MTBE or up to 85% methanol in gasoline produced no enhanced BTEX solubility. However, at higher gaso1ine:water ratios, aqueous methanol concentrations above 10% enhanced BTEX solubility. The initial methanol content of the gasoline and the equilibrating gasoline- to water-phase ratio controlled the aqueous methanol concentration. Partitioning theory and the experimental results were used to calculate aqueous benzene and methanol concentrations in successive batches of fresh groundwater equilibrating with the fuel and subsequent residuals. These successive batches simulated formation of a plume of contaminated groundwater. The front of the plume generated from high-methanol gasoline equilibrating with groundwater a t a gaso1ine:water ratio of more than 1 had high methanol content and elevated BTEX concentrations. Thus, releases of high-methanol fuels could have a more serious, initial impact on groundwater than do releases of methanol-free gasoline. Introduction Monoaromatic hydrocarbons such as benzene, toluene, ethylbenzene, and xylenes (termed BTEX as a group) are the most water-soluble and mobile and among the most potentially harmful hydrocarbons found in gasoline. Their concentrations in groundwater in contact with gasoline can be estimated, within a factor of 2 (1-3),from the aqueous solubility of the pure component, ita mole fraction in the gasoline, in accordance with Raoult’s law (assuming gasoline is an ideal solution whose components have unit activity coefficients):
ciw= sixi
g
(1)
where Ciwis the equilibrium concentration of component i in the water phase, si is the solubility of pure component i in water, and xig is the mole fraction of component i in the gasoline. Benzene has a “solubility” (si)of -1800 mg/L, but since it makes up less than 5% (v/v; wLg) of most gasolines, the benzene concentration in waters affected by gasoline should be less than 90 mg/L. This is still 4 orders of magnitude above regulatory action levels, which are typically near 5 pg/L. This solubility could be dramatically increased if a water-soluble cosolvent is present in the gasoline (4, 5). Present address: Jacques Whitford Environmental Ltd., P.O. Box 1116, Fredericton, NB, Canada E3B 5C2. 0013-936X/92/0926-2483$03.0010
Oxygen-containing organic compounds (“oxygenates”), such as ethers and alcohols, are common additives to gasoline (6) and are potential cosolvents. Fuels with high oxygenate content will contain less BTEX than nonoxygenated fuels. Applying eq 1,a decrease in the BTEX content of the gasoline would decrease the BTEX concentration in impacted groundwater. However, the welldocumented cosolvency effect (5, 7-13) would increase BTEX solubility by increasing the term si. For unexpectedly high BTEX concentrations to appear in groundwater, it is not sufficient that the solubility of BTEX simply be higher than in systems with no cosolvent present. Enough BTEX must be available to partition into the groundwater a t high concentrations along with the cosolvent. Here, this potential enhancement of BTEX solubility is considered for methanol and MTBE in gasoline. Theory Partitioning Theory. This paper treats the dissolution of BTEX and oxygenates in gasoline-water systems as attaining an equilibrium partitioning between aqueous and organic phases of specified volumes or volume ratios, an approach outlined in refs 14 and 15. This approach is well-suited for spill situations where the gasoline- to aqueous-phase ratio may be site-specific. A mass balance expression for the equilibrium partitioning of each component (e.g., benzene) in a two-phase system of gasoline and water is
where ybgis the volume fraction of benzene in the gasoline, V , is the initial volume of the gasoline (m3),pb is the density of benzene (g/m3), cbgand cbware the equilibrium concentrations of benzene in the gasoline and in the water phase, respectively (g/m3), and V, and V, are the volumes of the gasoline phase and the water phase, respectively, at equilibrium (m3). The left-hand side of eq 2 represents the total mass of benzene, and the right-hand side expresses the distribution of this mass between the gasoline and the water phases. The distribution of benzene between the two phases can be described in terms of a gasoline-water partitioning coefficient, Kbw,where
Kbw = cbg/cbw
(3)
We will assume that (1)applies: then Cbw
= b.$J
(4)
where xbg is the mole fraction of benzene in gasoline and
0 1992 American Chemical Society
Environ. Sci. Technol., Vol. 26, No. 12, 1992
2483
Table I. Volume Percent and Estimated Mole Fraction of BTEX in PS-6 Gasoline" benzene
toluene
1 2 3 4 5
run
2.117 1.905 2.141 2.087 2.159
3.740 3.301 3.691 3.437 3.425
av
2.082 5.06 0.02969
3.519 5.23 0.04392
% RSDb
MFc a
ethylbenzene 1.580 1.235 1.503 1.763 1.768 1.570 14.52 0.01700
volume % p-xylene
m-xylene
o-xylene
2.030 1.826 1.899 1.614 1.674
4.222 3.940 4.169 3.935 4.094
1.970 1.969 2.071 2.185 2.243
15.659 14.176 15.474 15.021 15.363
1.809 9.83 0.02263
4.072 3.5 0.04406
2.088 6.31 0.01959
15.139 4.04 0.17689
total BTEX
The BTEX volume percent measurements for runs 1-3 were made by vapor injection and selected-ion monitoring on a Hewlett-Packard
GC-MS. The injection comprised 100 r L of vapor from the equilibration of 3 WL of gasoline in a l-L bottle (external standard technique). Runs 4 and 5 were performed by split solvent injectioq of gasoline diluted in hexane with an m-fluorotoluene internal standard onto a GC with an FID detector. RSD, relative standard deviation. e MF, mole fraction.
sbis the solubility of pure benzene in water (g/m3). Since benzene solubility is low, the change in benzene concentration in the gasoline phase is assumed to be negligible. The gasolinewater partitioning coefficient (K",) can then be expressed as
Kbgw =
bb,Pb)/(XbgSb)
(5)
The mole fractions of BTEX in PS-6 gasoline were approximated from our analyses of BTEX (Table I) and analysis of the other, major organic compounds in PS-6 gasoline (17)and are included in Table I. This expression for Kb, (eq 5) can be substituted into eq 2 to obtain a value for the benzene concentration in the aqueous phase. cbw= yb,(Vg/Vg)pb/(Kbgw + V,/V,)
(6)
For the case of pure gasoline, the relative volumes of the gasoline and water phases did not visibly change during equilibration; hence V, is equal to Vr This treatment permits calculation of the aqueous BTEX concentrations by considering both the phase volume ratio (u,/u,) and the partitioning between the gasoline and aqueous phase (Kb,). When the experimentally observed aqueous BTEX concentrations significantly exceed concentrations predicted using the equilibrium partitioning model, the discrepency is attributable to the "cosolvency effect"-the higher solubility of BTEX in a water-methanol mixture than in water. Oxygenates in gasoline will partition between the gasoline phase and the aqueous phase so that phase volumes, densities, and molar compositions of the two phases change. Methanol [ K , = 0.0051 (5)]partitions preferentially to the aqueous phase while the less hydrophilic MTBE [ K , = 15.7 (5)]partitions preferentially to the gasoline. For calculating the equilibrium aqueous BTEX concentrations in oxygenate-gasoline-water mixtures at high water to gasoline ratios, all methanol was assumed to be in the aqueous phase and all MTBE in the gasoline phase. At lower water to gasoline ratios, aqueous methanol concentrations, where measured, were used in the calculation of aqueous BTEX concentrations. If no experimental determination of aqueous methanol concentration was available, the aqueous methanol concentration and the phase volumes at equilibrium were estimated using partitioning theory as follows. The mass balance expression for methanol, derived from eq 6, is cmw = Y",VlgPrn/(KmgwVg + V,)
(7)
Assuming that methanol and water form ideal mixtures, 2484
Environ. Sci. Technol., Vol. 26, No. 12, 1992
methanol will partition into the aqueous phase, resulting in significant changes to the phase volumes according to
v, = v', + (C",Vw)/Prn
(8)
v, = (V,P9 / (P" - Cmw) v, = v,+ v', - v,
(9)
and rearranging (10)
The following relation is obtained by substituting eqs 9 and 10 into eq 7: cmw = Y m g V g P m / ( P g w ( V+g V,) + (1 - P,)V,)
(11)
The aqueous concentration of methanol and the volume of the aqueous phase can be calculated from eqs 8 and 11 by iteration. Equation 10 can be used to calculate V, if the initial volumes of water and gasoline are known. Calculated values were used to predict the contaminant plume resulting from dissolution of gasoline into groundwater. Cosolvency Model. Good agreement was reported (11) between the log-linear and UNIFAC models in some solvent systems, so where the cosolvent effect becomes significant, the simple log-linear model of Yalkowsky and Rosemann (13)is used here. From ref 13 (12) log S m = log S w + P f c where S, and S, are the solubilities in water and in the water-cosolvent mixture, f , is the volume fraction of cosolvent in the aqueous phase, and 6 is a measure of the relative ability of the cosolvent to solubilize hydrophobic organic compounds (cosolvency power) as expressed by P = 1% (SC/S,) (13) where S, is the solubility in pure cosolvent. Also = a log (KO,)+ b (14) where KO,is the odanol-water partitioning coefficient and a and b are constants applying to a group of compounds (18). Substituting the expression for from eq 13 into eq 12 gives (15) log s, = f , log s, + (1- f,) log s, Equation 14 does not appear to be valid for low aqueous concentrations of cosolvent. For cosolvent contents less than 20% by volume, Banerjee and Yalkowski (19) observed a linear relationship: s, = fcVHs'c + (1- fcvH)sw (16) where VHis the ratio of the hydration shell volume to cosolvent volume and Sicis the solubility within the hy-
Table 11. Average Aqueous BTEX Concentrations for Various GasolineOxygenate Mixtures in a 10:l (v/v) Water to Gasoline System init oxygenate content (7%) 0 (PS-6)
methanol 5 10 15 50 85 MTBE 5 10 15
oxygenate concn (mg/L)
benzene
toluene
ethylbenzene
p-xylene
m-xylene
o-xylene
total BTEX 122.8
65.5
33.1
3.9
3.9
10.2
6.2
4110 8000 11300 43000 61500
63.4 67.0 64.9 60.6 55.2
33.5 33.0 32.4 31.6 35.4
4.0 4.1 3.9 4.0 4.7
4.0 4.1 3.9 3.9 4.7
10.5 10.5 10.1 9.9 11.7
6.3 6.5 6.2 6.2 7.4
121.7 125.2 121.5 116.2 119.0
1760 3650 5140
60.1 60.5 57.2
31.7 30.5 28.7
3.8 3.7 3.5
3.8 3.7 3.5
9.9 9.6 9.0
6.0 5.9 5.5
115.4 114.0 107.4
0.0
dration shell of the cosolvent. For toluene solubility in methanol-water mixtures with less than 20% methanol, Banerjee and Yalkowski (19) estimated V , = 6.7 and Si, = 1.4Swfrom experimental results. With cosolvent contents greater than 20% by volume, the conventional exponential behavior was observed. Since the log-linear relationship starts at the breakpoint [f’,, s,(f’J], rather than at the point of water solubility (0, sw), eq 12 should be modified to log S m = log Sm(f’J + P ( f c - f’J (17) Experimental Section Methanol and MTBE were glass-distilled (BDH, Ltd.; Omnisolv, 99.9% and 99.0% purity, respectively); PS-6 gasoline was provided by API and ita composition is given in ref 17. The BTEX composition of the particular batch of PS-6 gasoline supplied was determined by two different methods. The results are presented in Table I. Groundwater from the Canadian Forces Base Borden, ON, Canada, research site (20) was used in all experiments. This water had total dissolved solids of -230 mg/L. All laboratory experiments used the shake-flask batch contacting equilibration procedures (17). All experiments were conducted at 10 OC,the groundwater temperature. Triplicate solutions were prepared in 60-mL Hypovials filled without headspace and then rotated at 40 rpm in a 10 OC refrigerator for 4 h. Initial experiments determined that no change in concentrations occurred after 1h of rotation. The Hypovials were placed upside down in a GSA rotor head and centrifuged for 15 min a t 2000 rpm inside at 10 “C Sorvall centrifuge to separate the gasoline and water phases. The separated water phase was removed by glass syringe and dispensed into 18mL glass Hypovials containing 0.2 mL of 10% sodium azide bacteriacide. Dissolved BTEX concentrations were measured by hexane extraction followed by gas chromatographic determination using a 6 m long, 0.32 mm inside diameter fused-silica column with a 0.25-flm bonded Carbowax 20M stationary phase operated isothermally at 90 “C and a flame ionization detector (FID). An internal standard, m-fluorotoluene, was added in the hexane. Generally the BTEX method detection limits were 1-2 pg/L, the relative standard deviations for replicates were f5% or less, and the accuracy with respect to standard solutions, prepared gravimetrically, was virtually 100%. Oxygenate concentrations were determined by direct aqueous injection onto a Hewlett-Packard 5840A gas chromatograph with a 3 m long, 0.3 cm inside diameter column packed with 3% SP1500 on Carbopack B (80-100 mesh) and a FID. Method detection limits were 100 and 250 pg/L for methanol and MTBE, respectively, with relative standard deviations typically less than 4% and bias less than 4 % .
-
BENZENE mg/L
70 1
. .
40t 30 0
40
20
80
60
100
METHANOL CONTENT IN GASOLINE (%) 70
50
BENZENE, mg/L
t
40
0
io
20
30
MTBE CONTENT IN GASOLINE (%)
Figure 1. Effect of (a) methanol content and (b) MTBE content In gasoline on aqueous benzene concentration. Initial aqueous- to gasdlnephase ratk 1 0 1 (v/v). Calculated trend shown as dashed line and normalized trend shown as solid line. See text for details.
Results and Discussion Aqueous BTEX concentrations were essentially constant after equilibration of fixed volumes of PS-6 gasoline with varying volumes of groundwater as long as the aqueousphase volume (V,) to gasoline-phase volume (V,) was less than -2O:l. Thus, Vw/V of 1O:l was used here to investigate the aqueous solu%ilityof BTEX from gasoline containing variable concentrations of oxygenates. The aqueous BTEX concentrations are about the same for equilibrium with gasolinemethanol mixtures as with pure gasoline (Table I1 and Figure la). Slightly lower aqueous BTEX concentrations were observed for equilibrium with MTBE (Table I1 and Figure lb). Three additional equilibrium experiments were done to investigate the influence of high methanol contents in water equilibrated with gasoline and with pure benzene: Environ. Sci. Technoi., Vol. 26, No. 12, 1992
2485
250
BENZENE mg/L
BENZENE g/L
6
200 150 -
0' 0
20 30 40 AQUEOUS METHANOL CONTENT (%) 10
I 50
10 20 30 AQUEOUS METHANOL CONTENT (%I
0
BENZENE mg/L
10000
-
10
0
20 40 60 80 AQUEOUS METHANOL CONTENT (%) Figure 2. Effect of aqueous methanol concentration on aqueous benzene concentration for the gasoline-methanol-water system. Linear and logarithmic scales used to Illustrate the calculated trends (solid lines) based on the linear and log-linear models, respectively.
(a) aqueous methanol contents from 0 to 90% (v/v) for gasoline- to aqueous-phase ratios of 1:lO (v/v), (b) aqueous methanol contents from 0 to 90% (v/v) for benzene- to aqueous-phase ratios of 1:lO and 1:l (v/v), and (c) 50% aqueous methanol content (v/v) for gasoline- to aqueous-phase ratios between 1:l and 1:lOOO (v/v). Methanol and groundwater were mixed prior to the addition of various proportions of the gasoline or benzene phase. The results of experiments a-c for benzene are summarized in Figures 2-4, respectively. Figure 2 shows that, for an aqueous- to gasoline-phase ratio of lO:l, aqueous benzene concentrations increase dramatically when the aqueous methanol content exceeds -20%. A similar increase in aqueous benzene solubility from a pure benzene phase is observed in Figure 3. Complete dissolution of the benzene phase was noted for systems with initial aqueous methanol contents greater than 75%. Figure 4 illustrates the decrease in the aqueous concentrations as the benzene and p-xylene pools are depleted and thus as V,/V, increases beyond 1O:l. The equilibrium partitioning model developed above is used to evaluate the experimental data in order to identify enhanced BTEX solubility due to cosolvent effects. Normalizing Predicted Trends. For each experiment, the aqueous BTEX concentrations were calculated using the equations developed in the Theory section. The calculated trends were compared with the trends of experimental data as illustrated in Figure la. In general, the calculated values (lower line on Figure la) underestimate the experimental values; however, similar trends are apparent. The discrepancy probably reflects uncertainties in the volume or molar composition of the gasoline. 2486
Envlron. Scl. Technol., Vol. 26, No. 12, 1992
1
CALCULATED VALUE
1
8
40
300
vw/
BENZENE mg/L
1
10
100
1000
vw/vg
P-XYLENE ma/L
I
I 407
20t
I 0' 1
10
100
1000
v w I vg
Flgure 4. Effect of aqueous- to gasollne-phase ratios (VJV,) on benzene and p-xylene concentrationsfor an aqueous phase with 50% (vlv) methanol. The normalized predicted trend line Is shown.
benzene in the gasoline phase. The aqueous concentrations of other BTEX compounds remained relatively constant for methanol contents up to 90%. The averaged experimental data are generally well-represented by the predicted trends (Figure la). This suggests that the potential for enhanced BTEX solubility due to the presence of methanol in the gasoline is minimal for the situation that produces low aqueous methanol contents. For the calculation involving MTBE, it is assumed that MTBE remains within the gasoline phase a t equilibrium. The experimental results suggest that approximately 46% of the available MTBE partitioned into the aqueous phase. As all of the oxygenate is assumed to remain within the gasoline phase, the value of the equilibrium constant, Kbm, remains constant as the oxygenate content of the gasoline is changed. We conclude from the match of experimental and predicted trends (Figure lb) that no significant cosolventderived enhancement of BTEX solubility occurred. This is consistent with the findings of Cline et al. (5). Mihelcic (9) suggested that aqueous MTBE concentrations of 1% could produce ethanol BTEX solubility. However, this represents -20% of the pure-phase MTBE solubility and could not result from fuels containing the proportions of MTBE used in our experiments (15% and less) and present in current gasoline fuels (5). Effect of High Aqueous Methanol Concentrations on BTEX Solubilities. The calculated and experimentally determined benzene concentrations for aqueous phases with various methanol contents in equilibrium with gasoline and with pure benzene are presented in Figures 2 and 3, respectively. Essentially all of the benzene was dissolved a t equilibrium for systems with methanol con-
tents greater than approximately 75% and vi-,= lo. Consequently, the experiments were repeated with Vw/Vb = 1 with the results included in Figure 3. A breakpoint between linear and logarithmic behavior appears at an aqueous methanol content between 25 and 35%. Aqueous concentrations of benzene for low methanol contents were calculated using the linear relationship (eq 16) suggested in ref 19. The values of VH and Sic/Sw estimated by in ref 19 were used in solving eq 16. A linear relationship between /3 and log Kow(values from ref 21) was found, confirming the applicability of eq 14 and establishing the values for a and b as 2.57 and -3.01, respectively. For aqueous methanol contents greater than 25%, the aqueous BTEX concentrations were calculated using the log-linear relationship (eq 17) followed by eqs 5 and 6. The mole fractions, volume fractions, and relative phase volumes change as the system approaches equilibrium. These changes have little effect on the aqueous solute concentrations as long as the equilibrium methanol content is used in the calculations. If experimentally determined aqueous methanol contents were not available, the methanol content was calculated from the initial methanol content partitioned between water and gasoline according to the octanol-water partitioning coefficient. In Figure 3, breakpoints were evident at aqueous methanol contents below 20 and 35%. The linear relationship (eq 16 with VH and Sictaken as toluene from ref 19) appears valid at methanol contents below 20%,while the log-linear relationship is better above 35% (see Figure 3). At methanol contents above 55% the experimental values are higher than the calculated values. The trend of the calculated curve (Vw/V,= 10) diverges from the log-linear trend at higher methanol contents due to complete dissolution of the benzene phase (Le., no more benzene is available and only one phase is present). The predicted and experimental aqueous benzene and p-xylene concentrations for an initial 1:l water-methanol mixture in equilibrium with gasoline at various gasolineto aqueous-phase volume ratios are shown in Figure 4. For this calculation the aqueous methanol content at equilibrium was assumed to be 43.6%,the observed content when Vw/V, was 10. Again, agreement between the prediction and experimental results is satisfactory, confirming the reliability of the calculation. Dissolved BTEX Plumes Resulting from Spills of Methanol-Gasoline Mixtures. In a real spill situation fresh groundwater will continually flow past a fmed volume of spilled gasoline. Because the composition of the gasoline phase will change due to dissolution of the soluble components (BTEX), the aqueous concentrations of successive groundwater volumes will also change. This can be simulated as a series of successive batch equilibrations of quantities of water and fuel, in which the water is removed after each equilibration and replaced with water which, in turn, equilibrates with the residual fuel. The equations developed in the Theory section were used to calculate the aqueous- and gasoline-phase compositions sequentially. Gasolines with 0 and 85% methanol (here, termed M85) were considered, with initial gasoline to water ratios (v/v) of 1:l and, for 85% methanol, also 1O:l. The aqueous methanol and benzene concentrations calculated for each successive batch of water are given in Table 111. An octanol-water partitioning coefficient of methanol, Pow = 0.178, was used in approximation of the gasolinewater partition coefficient. Cline et al. (5) more recently reported Km, = 0.0051, derived from experiments with a water to gasoline ratio of 201 and initial methanol Environ. Scl. Technol., Vol. 26,
No. 12, 1992 2487
Table 111. Aqueous Methanol and Benzene Concentrations in Successive Batches of Water Exposed to Gasoline Pools with Varying Methanol Content and with Varying Initial Gasoline to Water Ratios batch methanol content
concn (ma/L) benzene methanol benzene methanol benzene
0.85
0.0
0.85
0.85
0.85
(V/d
init gasoline: water ratio (v/v)
1:l
1
52.70 360000 52.70 52.70 10200 52.70 52.70 266 52.70 52.69 6.9 52.69 0.2 52.69 0.2 52.69 0.0 0.0 52.69 52.69 0.0 52.69 0.0 52.69
1" 2 2" 3 3" 4 5 6 7 8 9 10 10"
1:l
"Values calculated with K,"
1:l
101
793.36 705000 829.08 56.60 149000 50.79 50.82 31300 50.54 50.58 6600 50.48 1390 50.38 293 50.26 61.8 50.14 13.0 50.00 2.7 49.84 0.6 49.61
1O:l
2571.56 2564.60 86.04 53.04 58.54 51.95 53.41 52.35 52.11 52.05 52.03 52.02 52.00 51.83
= 0.0051.
content lower than lo%, quite different from our case. Calculations with the two values of P, are included for selected batches in Table 111. The differences in these values do not significantly change the implications discussed below. The initial batch of gasoline-equilibrated water will form the leading edge of the groundwater contamination plume. Later batches will represent the portion of the plume nearer the source. The methanol is leached, quickly, resulting in a discrete, short methanol pulse (the initial 5-10 batches in Table 111),which will advance at about the groundwater velocity. The cosolvency effect of methanol produces higher benzene concentrations from M85 than from methanol-free gasoline. A high initial M85 to water ratio produces higher benzene concentrations in the initial batches (Table 111). These high benzene concentrations will likely be near the leading edge of the plume generated from M85. The more hydrophobic aromatics will likely be retarded relative to the benzene (20). Note that the later batches of M85-equilibrated water have benzene concentrations only slightly lower than batches equilibrated with methanol-free gasoline. Thus, unlike the methanol plume, the BTEX plume will extend back to the source for an extended time. This has recently been confirmed in preliminary dynamic experiments conducted with M85 in sand columns (22), where the effluent concentrations of methanol and BTEX followed, generally, the trend predicted by the sequential batch calculation (Table 111). Eventually, but not quickly, the BTEX will be depleted in the gasoline and the aqueous concentration will decrease. There is less BTEX in the M85 than in normal gasoline, so this decrease in aqueous concentrations will occur earlier with M85. Of course, the plume of BTEX will be somewhat modified by dispersion, by sorptive retardation, and potentially by biotransformation (20). This calculation does, however, illustrate that a downgradient well can expect to see very high BTEX and methanol concentrations initially from a spill of M85, especially if V,/V, is