Monod Kinetics for Aerobic Biodegradation of Petroleum

Mar 6, 2007 - ... data that document the aerobic biodegradation of hydrocarbon vapors in .... A new model for the spectral induced polarization signat...
8 downloads 0 Views 165KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 2343-2349

Monod Kinetics for Aerobic Biodegradation of Petroleum Hydrocarbons in Unsaturated Soil Microcosms DAVID W. OSTENDORF,* THEODORE H. SCHOENBERG, ERICH S. HINLEIN, AND SHARON C. LONG Civil and Environmental Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003

We use Monod kinetics to calibrate previously published data that document the aerobic biodegradation of hydrocarbon vapors in soil microcosms from a weathered petroleum spill site. Monod kinetics offer insight into biodegradation mechanics because they address biomass growth as well as substrate depletion. A blend of five aromatics and five alkanes dose the microcosm sets at four strengths, and a finite difference model describes the response superimposed across the constituent substrates. An observed initial biomass XO of 125 g biomass/m3 soil moisture and an endogendous decay rate b of 0.102 day-1 calibrate all four dosages and agree with heterotrophic plate counts. Common maximum specific growth rates µMJ and half saturation constants KSJ calibrate each constituent across the four dosages. The biodegradable alkanes exhibit µMJ values ranging from 0.0190 to 0.0996 day-1, while the aromatic rates vary from 0.0946 to 0.322 day-1. One of the alkanes (2,2,4trimethylpentane) is recalcitrant. The half saturation constants range from 0.000083 to 0.000355 g substrate/m3 soil moisture for the biodegradable alkanes, which imply zeroorder kinetics. The aromatic KSJ values vary from 5.02 to 14.3 g substrate/m3 soil moisture, and suggest first-order kinetics. The yield YJ increases with dosage concentration for all the biodegradable constituents, varying from 0.0533 to 1.58 g biomass/g substrate.

documents this natural aerobic attenuation of petroleum hydrocarbon vapors at various scales. Richards et al. (1), Aelion and Bradley (2), and Schirmer et al. (3) study the degradation of aviation gasoline, jet fuel constituents, and m-xylene in soil microcosms, respectively. Moving up in scale, Hohener et al. (4) assess gasoline constituent vapor fate and transport in stainless steel bioreactors and columns, while Lahvis et al. (5) use soil gas tubing clusters to quantify the biodegradation of hydrocarbon vapors at field spill sites in situ. The literature offers biological and transport insights into hydrocarbon biodegradation. Aliphatic and aromatic hydrocarbons at ppm concentrations can serve as growth substrates for microorganisms after initial oxidation. Carboxylated aliphatics can enter the beta-oxidation pathway, while carboxylated aromatics proceed through catechol metabolic pathways (6, 7). Thus, it can be hypothesized on biochemical grounds that Monod growth kinetics could appropriately be applied to this research. Robinson and Tiedje (8) suggest that substrate utilization in soil microcosms calibrates these kinetics in the absence of endogenous decay. From a transport perspective, Monod kinetics couple transient biomass concentration to the soil gas studies and provide a second calibrating data set that strengthens the substrate analyses and data. This coupling is well represented theoretically in groundwater models (9-12), whose sophistication requires a large number of parameter estimates. By contrast, many vadose zone investigators assume steady biomass concentration and apply zero order (4), first order (13), variable order (14), or vertically integrated (5) kinetics to assess hydrocarbon, oxygen, or carbon dioxide concentrations in soil gas. We improve this well-established literature by applying Monod kinetics with endogenous decay to a blend of substrate constituents, then using a finite difference model to recalibrate a thorough, existing soil microcosm data set (14). Our calibrated specific growth rates, endogenous decay rate, half saturation constants, and yields for 10 petroleum hydrocarbon constituents reduce the number of assumed kinetic values input to the detailed, dissolved phase codes. Since Monod kinetics are fundamental and the partitioning is measured, we can combine calibrated parameters to estimate first order, zero order, or Michaelis-Menten kinetics for simpler, volatile models of aerobic transport in the unsaturated zone.

Theory Introduction Semivolatile petroleum hydrocarbons pose a continuing threat to the subsurface environment due to their widespread use as gasoline, diesel fuel, jet fuel, and heating oil, and the relatively slow transport times of the unsaturated and saturated zones. These uses require distributed storage, transport, and handling facilities, with potential for release through leakage or accidents. Hydrocarbon vapors are an important phase of petroleum hydrocarbon contamination in shallow unconfined aquifers because the nonaqueous phase liquid is semivolatile and less dense than water. Thus the separate phase spreads out across the water table and evaporates into the overlying unsaturated zone. The hydrocarbon vapors and oxygen in the pore space are electron donors and acceptors for aerobic biodegradation of the contaminant, respectively, provided microorganisms and nutrients exist in the vadose zone. A substantial literature * Corresponding author e-mail: [email protected]. 10.1021/es062313l CCC: $37.00 Published on Web 03/06/2007

 2007 American Chemical Society

Biomass and Substrate Conservation Equations. We follow (4) and base our analysis on dissolved hydrocarbons. The conservation of biomass X (per unit volume of soil moisture) balances storage, Monod kinetics, and endogenous decay b in a soil microcosm, which includes a blend of N petroleum hydrocarbon vapor constituents

VW

dX dt

) V WX

(

N

µMJCJ

∑K J)1

SJ

+ CJ

-b

)

(1a)

X ) XO (t ) 0)

(1b)

with time t and initial biomass concentration XO. Since the reactions are assumed to occur in water, the half saturation constant KSJ of the Jth hydrocarbon constituent and dissolved substrate concentration CJ are expressed as mass of substrate per volume VW of soil moisture. The Jth maximum specific growth rate µMJ has the units of 1/time (biomass growth rate VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2343

per unit biomass), and quantifies the ability of an individual microorganism to grow upon the dissolved substrate constituent. Nutrients and oxygen are assumed abundant and do not limit the biomass growth. We superimpose the constituent contributions to biomass: each microorganism grows upon each dissolved hydrocarbon. The conservation of dissolved substrate concentration balances storage, equilibrium partitioning, and reaction for each constituent in the soil microcosm

(VW + VGHJ + MSKDJ)

dCJ µMJVWXCJ )dt YJ(KSJ + CJ)

(2a) (2b)

GJ ) HJCJ

(2c)

SJ ) KDJCJ

(2d)

with Jth initial substrate concentration COJ and yield coefficient (g biomass created/g substrate consumed) YJ. The latter reflects the partial assimilation of carbon into cellular materials, with the remainder respired by the biomass. Partitioning is important. The hydrocarbons sorb to the dry soil mass MS (15), deposit on the air-water interface in the soil pores (16), and evaporate into the soil gas volume VG, which includes the headspace of the microcosm (14). Thus the substrate disappears from the microcosm slower than its soil moisture defined degradation rate, due to retardation by the remaining partitions, including the headspace. By the same token, the sorbed, deposited, and volatile substrate partitions contribute energy and mass to the microorganisms in the soil moisture. The Jth Henry constant HJ relates volatile (GJ) to dissolved concentrations on a dimensionless basis (17), while the distribution coefficient KDJ reflects both sorbed and deposited hydrocarbon constituent mass (SJ) per mass of dry soil. The Henry constant rests on chemical properties of the hydrocarbons (18), while abiotic control data (3, 14) define KDJ. We eliminate the reactions between eqs 1a and 2a and find N

∑YR

J DJCJ)

J)1

) -bX

(3a)

VGHJ MSKDJ + VW VW

(3b)

dt RDJ ) 1 +

with Jth retardation factor RDJ for the dissolved phase in the microcosm. Equation 3b implies that volatile, insoluble constituents are significantly retarded in soil microcosms with relatively large headspaces when the analysis proceeds on a dissolved basis. Insoluble, sorbed hydrocarbons are also retarded in soil microcosms with relatively large soil to moisture ratios. Robinson and Tiedje (8) combine eqs 2a and 3 for a single substrate in the absence of endogenous decay and solve analytically with a nonlinear, implicit solution, predicated on the constancy of the sum of the biomass and substrate fractions used to generate biomass. The variables do not separate for multiple constituent substrates with endogeneous decay, however, and we approximate the solution instead by finite difference approximation. The Jth substrate concentration after the Ith time increment CJI rests on known prior conditions, and eq 2a becomes 2344

9

]

µMJXI-1∆t

(4)

YJRDJ(KSJ + CJI-1)

with time increment ∆t. The current substrate concentration and eq 3a then may be used to approximate the biomass concentration XI after the Ith time increment N

XI ≈ XI-1(1 - b∆t) +

∑R

DYYJ(CJI-1

- CJI)

(5)

J)1

CJ ) COJ (t ) 0)

d(X +

[

CJI ≈ CJI-1 1 -

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007

The Monod kinetics, endogenous decay, and initial biomass calibrate a given constituent data set for a known retardation factor and initial substrate dose. The calibration is difficult because all the constituents couple through the biomass term. Stoichiometry and Yield Coefficient. The fraction FJ of substrate used for growth and the assumed stoichiometry of aerobic biodegradation determine YJ and the relationship between mass based concentrations of the reactants and products in the dissolved phase (19). The stoichiometry rests in part on an assumed biomass formula of C5H7O2N (20)

Q + (1 - F )(M + )]O + [ (Q4 - 3M 20 ) 4

CMHQ + FJ

J

2

FJM FJM Nf C5H7O2N + (1 - FJ)MCO2 + 5 5 Q Q 7M FJ + (1 - FJ) H2O (6) 2 10 2

[(

)

]

with carbon number M and hydrocarbon number Q in the petroleum hydrocarbon. The yield coefficient is simply the ratio of the biomass and substrate mass concentrations, and accordingly depends in part on FJ

YJ ) FJYMJ YMJ )

M(molar‚biomass) 5(substrate‚molar‚mass)

(7a) (7b)

The maximum yield coefficient YMJ describes the complete use of substrate for growth.

Materials and Methods Ostendorf et al. (14) prepared sets of aerobic soil microcosms from a petroleum spill site in eastern Massachusetts in June 1998. The spill, which consisted of diesel fuel and automobile gasoline, had weathered under a paved surface for over 20 years at the time of sampling. The soil was a nonuniform silty sand with 53 g moisture/kg dry soil, 1.2 g organic C/kg dry soil, 1.9 g total N/kg dry soil, and 0.3 g P/kg dry soil. Eighteen kg of separate phase petroleum/m2 horizontal area was distributed across the capillary fringe, which extended 2-3 m below the pavement. Soil microcosms were obtained over a depth interval of 0.6-0.8 m using U.S. Environmental Protection Agency aseptic drilling protocol (21), which featured steam cleaned core barrels driven ahead of hollow stem augers. Autoclaved 10 mL syringe barrels were used to obtain 20 g soil samples from the core barrels as they were extruded in a nitrogen filled glovebox. The gaseous soil microcosms consisted of 12.3 mL autoclaved glass vials, equipped with Mininert valves and water seals (1). The latter reduce leakage by the semivolatile headspace vapors, which can be appreciable in soil microcosms (4). A 5.0 g dry soil mass (MS) and a soil moisture volume of 0.46 mL (VW) were placed in each vial under a laminar flow hood, resulting in a headspace and total gas volume (VG) of 10 mL. Heterotrophic plate counts were conducted by Ostendorf et al. (14) before dosage, resulting in an initial population of 11.5 colony forming units/µg dry soil.

TABLE 1. Hydrocarbon Partitioning in Soil Microcosms constituent (formula)

YMJ

KDJ (mL/g)

HJ

2-methylhexane (C7H16) 2,2,4-trimethylpentane (C8H18) 3-methylheptane (C8H18) ethylbenzene (C8H10) m-xylene (C8H10) o-xylene (C8H10) n-nonane (C9H20) cumene (C9H12) 1,2,4-trimethylbenzene (C9H12) n-decane (C10H22)

1.58 1.59 1.59 1.71 1.71 1.71 1.59 1.70 1.70 1.59

7.68 3.14 9.00 0.661 0.658 0.527 238 1.46 1.47 558

128 114 40.9 0.323 0.264 0.194 216 0.532 0.221 181

The sets were dosed with a blend of ten hydrocarbon vapors, subject to empty bottle and sodium azide controls (14). All experiments were conducted at 20 deg C. Table 1 lists the five alkanes and five aromatics used, 7-10 carbon compounds (M) with molar masses varying from 100 to 142 g/mol. The maximum yield (eq 7b), also listed in the Table, varies from 1.58 to 1.71. The headspace of replicate microcosms was dosed at four strengths: weakest (76.8 g hydrocarbons/m3 soil moisture), weak (115 g/m3), strong (154 g/m3), and strongest (192 g/m3). These substrate doses were much less than the nutrients (3,300 g P/m3 soil moisture, 21 000 g total N/m3 soil moisture) and oxygen (5500 g O2/m3 soil moisture) in the microcosms, so that the aerobic reactions were substrate limited only. The dosage composition was uniformly distributed across the constituents, although nonuniform partitioning resulted in varying vapor concentrations COJ at the start of the experiments (Figure 1, Table 2, and Table 3). The dissolved aromatic concentrations were 2 orders of magnitude higher than the alkane concentrations due to solubility effects. This contrasts markedly to the original vapor phase presentation of the experiment (14), which featured GJ values of order 0.3 g/m3-soil gas for all constituents. The empty bottle controls quantified leakage and other abiotic losses, which were minor over the 200 h duration of the microcosm study. The sodium azide controls calibrated initial concentrations and partitioning. Typical live microcosm data, sketched as symbols in Figure 1, were adjusted for these losses and partitioning. At least 20 observations documented degradation of each dosage and constituent. All constituents degraded significantly within 200 h of exposure, except for 2,2,4-trimethylpentane, which was recalcitrant. Ethylbenzene, n-nonane, and n-decane degraded most rapidly.

Results Partitioning. Ostendorf et al. (14) followed Schirmer et al. (3) by using the vapor concentrations in the sodium azide controls to calibrate the retardation factors, distribution coefficients, and initial concentrations. Tables 2 and 3 partition the results on the dissolved basis used in the present analysis. Soil microcosms exhibit strong retardation when hydrocarbon concentrations are based on the dissolved partition. Retardation factors for the alkanes range from 990 to 10 000, reflecting their insolubility. The aromatic retardation varies from 10.9 to 28.4, due to appreciable dissolution. Equation 3b distributes the hydrocarbon mass among dissolved, gaseous, and sorbed partitions, with the results cited in Table 1. The lightest alkanes are primarily volatile, while the heavier alkanes exhibit coequal volatility and sorption. Dissolution accounts for 3-10% of the aromatic hydrocarbons, with the remainder distributed appreciably between the soil gas and the soil mass. Calibrated Monod Kinetics. A nested Fibonacci search (22) calibrates b, µMJ, KSJ, and YJ by minimizing the root-

vapor (% mass)

sorbed (% mass)

aqueous (% mass)

97.1 98.6 90.0 46.2 41.3 38.5 64.5 40.7 22.1 39.3

2.91 1.36 9.90 47.2 51.5 52.3 35.5 55.8 73.4 60.7

0.035 0.040 0.101 6.58 7.20 9.14 0.014 3.52 4.59 0.0010

mean-square δ of the error defined for each constituent and dosage by

error )

measurement - calibration COJ

(8)

Tables 2 and 3 and Figure 1 summarize the calibration. We optimize the endogenous decay rate for all dosages and constituents with the following result

XO ) 125

g × biomass m × soil × moisture

(9a)

3

b ) 0.102 day-1

(9b)

The adopted initial biomass value rests on the observed initial population of 11.5 cells/µg dry soil and an individual microbial biomass mB of 10-6 µg, with the latter reflecting a 1 µm3 microorganism volume (23) and a density of 1000 kg/ m3. Equation 9b is at the upper end of the b estimate of 0.01-0.1 day-1 adopted by Essaid et al. (12) and higher than the 0.02 day-1 value proposed by Molz et al. (11). The overall substrate model error of 12% indicates reasonable accuracy for a common parameter value across all the experiments. We also measured heterotrophic plate counts at the end of the four experiments with a range of 116-187 g biomass/m3 soil moisture (10.9-17.2 cells/µg dry soil). These data are 12-52% higher than the calibrated values, with a root-meansquare error of 33%. This order of magnitude biomass agreement further endorses the independent substrate calibration, which rests on far more data. The half saturation constant and µMJ vary for each constituent, but remain constant for the four dosages. Figure 1 displays observed and calibrated constituent concentrations for the strong dosage, which typifies all the experiments and calibration. All the hydrocarbons degraded except 2,2,4trimethylpentane. The latter recalcitrance is also noted by Solano-Serena et al. (24), who attribute the relative persistence to the quaternary carbon group in the compound structure. Tables 2 and 3 list the calibrated Monod kinetics. The maximum specific growth rate for the four biodegradable alkanes varies from 0.0190 to 0.0996 day-1, increasing with molar mass. The KSJ values range from 0.000083 to 0.000355 g substrate/m3 soil moisture. Since COJ exceeds KSJ, the microorganisms degrade the alkanes on a zero-order basis for the range of tested substrate concentrations, and the calibrations plot as straight lines in Figure 1. The aromatic µMJ varies from 0.0946 to 0.322 day-1, with KSJ ranging from 5.02 to 14.3 g substrate/m3 soil moisture. The aromatic kinetics tend toward first order and plot as curves in Figure 1. Constituent errors range from 7.9 to 20%, suggesting reasonable model accuracy, particularly since the growth rate and half saturation constants remain fixed across the four dosages. VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2345

FIGURE 1. Observed (symbols) and calibrated (curves) dissolved hydrocarbon concentrations for strong dose (154 g hydrocarbons/m3 soil moisture) experiment. Data are adjusted for abiotic leakage, which is minor.

We calibrate the yield fraction for each dose and constituent, however, resulting in the values cited in Tables 2 and 3. The optimal YJ (and FJ) rise with rising dosage for every degradable constituent. 2346

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007

Discussion Biomass Concentration. The initial biomass concentration falls within the 1-103 microorganisms/µg dry soil range cited

TABLE 2. Calibrated Monod Kinetics for Aerobic Degradation of Lighter Hydrocarbon Vapor Constituents, Based on Dissolved Partition constituent (J) 2-methylhexane (1)

COJ (g/m3) 0.00314 0.00453 0.00595 0.00766

YJ (FJ)

RDJ

µMJ (day-1)

KSJ (g/m3)

δ (%)

0.909(0.575) 2,870 0.0276 0.000175 7.9 1.21(0.765) 1.53(0.968) 1.58(1.00)

2,2,4-trimethyl- 0.00396 0.293 pentane (2) 0.00587 0.374 0.00766 1.56 0.00974 0.601

2,510 0.00110 0.000038 10

3-methyl0.00934 0.458(0.288) 990 heptane (3) 0.0133 0.614(0.386) 0.0177 0.765(0.481) 0.0224 0.859(0.540)

0.0190 0.000083 8.3

ethyl0.458 benzene (4) 0.663 1.01 1.24

0.0692(0.040) 15.2 0.257 0.171(0.100) 0.246(0.144) 0.483(0.283)

9.51

8.0

m-xylene (5)

0.163(0.095) 13.9 0.252 0.416(0.243) 0.622(0.364) 1.09(0.635)

10.8

9.4

0.477 0.701 1.03 1.34

TABLE 3. Calibrated Monod Kinetics for Aerobic Degradation of Heavier Hydrocarbon Vapor Constituents, Based on Dissolved Partition constituent (J)

COJ (g/m3)

YJ (FJ)

RDJ

µMJ (day-1)

KSJ (g/m3)

δ (%)

o-xylene (6)

0.572 0.861 1.16 1.57

0.192(0.112) 10.9 0.491(0.287) 0.791(0.463) 1.30(0.758)

n-nonane (7)

0.000949 0.00121 0.00186 0.00230

0.234(0.147) 7,280 0.0242 0.000355 9.6 0.395(0.248) 0.348(0.219) 0.424(0.266)

cumene (8)

0.235 0.314 0.442 0.571

0.0742(0.044) 28.4 0.179(0.105) 0.251(0.148) 0.403(0.237)

0.322 5.02

8.8

1,2,40.385 trimethyl- 0.611 benzene (9) 0.783 1.04

0.0533(0.031) 21.8 0.103(0.061) 0.184(0.108) 0.301(0.177)

0.0946 8.99

20

n-decane (10) 0.000812 0.000989 0.00156 0.00192

0.661(0.415) 10,000 0.0996 0.000163 16 1.39(0.871) 1.24(0.781) 1.58(0.996)

0.185 14.3

12

by Atlas and Bartha (25) and is the same order of magnitude as the eight microorganisms/µg dry soil value put forth by Alexander (23) for a typical surface (3-8 cm depth) soil horizon. The calibrated XO is near the bottom of the 20-300 colony forming units/µg of dry soil range found by Song and Bartha (26) in loamy surface soil, and the same order of magnitude as the 64 cells/µg of dry soil calibration observed by Mu and Scow (27) in uncontaminated silty loam surface soil. Our observed biomass is an order of magnitude less than the 300 cells/ µg of dry soil found by Hohener et al. (4) in unsaturated silty sand. Schirmer et al. (3) estimate biomass created from aerobic degradation of m-xylene in groundwater microcosms, and find populations that range from 3 to 23 g microorganisms/m3 soil moisture, an order of magnitude less than eq 9a. We note that groundwater lies below the unsaturated zone, and the decrease of microbial population with increasing depth below the ground surface is well

documented in the literature (23). Chen et al. (11) adopt an XO of one cell/µg dry soil in their groundwater code, an order of magnitude less than our unsaturated zone calibration. Molz et al. (10) assume a value of six cells/µg dry soil. Correlation of Maximum Specific Growth Rate and Half Saturation Constant. Figure 2 displays the correlation between and µM and KS for three constituents of the strong dosage experiment. Equation 9 specifies the initial biomass and endogenous decay rate, while the yield varies to minimize δ for specified µM and KS values. This exercise yields contours of root-mean-square errors on the µM and KS plot. The 3-methylheptane minimum error contour of 8.2% ranges between a maximum specific growth rate of 0.01 and 0.02 days-1 and is bounded by the µM axis. Thus, a unique µM value emerges from the calibration, while KS is small and unimportant. This corresponds to zero-order kinetics. The 8.0% minimum error contour for ethylbenzene occupies a slotted region in Figure 2 that indicates a direct correlation between the half saturation constant and maximum specific growth rate. This is a consequence of the first order behavior exhibited by the aromatics, which implies a constant µM/KS ratio. Only the heavier alkanes degrade with Monod kinetics: the minimum error contour of 9.7% for n-nonane encloses a unique pair of µM and KS values. We thus conclude that Tables 2 and 3 cite unique µM values for biodegradable alkanes and unique µM/KS values for aromatics. Only the heavier alkanes exhibit unique half saturation constants in Table 3. Yield Coefficients and Limiting Respiration. The yield coefficients and FJ values increase with dosage concentration for all the biodegradable hydrocarbon vapors, though the maximum specific growth rate remains fixed. All the calibrated YJ values fall within their stoichiometric limits (YMJ) cited in Table 1. The alkane yields generally exceed their aromatic counterparts and vary from 0.23 to 1.58 g biomass/g substrate, which corresponds to an FJ range of 0.147-0.996. The aromatic yields range from 0.053 to 1.30 g biomass/g substrate, so that FJ varies from 0.031 to 0.758. The calibrated Y5 range of 0.163-1.09 includes the average value of 0.52 g biomass/g m-xylene found in aerobic groundwater microcosms by Schirmer et al. (3). The calibrated yield range brackets values commonly adopted for hydrocarbon biodegradation in groundwater; Borden and Bedient (9), Hohener et al. (4), and Chen et al. (11) adopt a yield coefficient of 0.50 g biomass/g substrate, while Molz et al. (10) assume 0.28 g biomass/g substrate and Essaid et al. (12) assume a yield of 0.25 g biomass/g substrate. The wide range of calibrated FJ values is typical for soil based microorganisms: Alexander (28) cites growth fractions varying from 0.2 to 0.9 in soil. Substrate Utilization Rate Comparisons. We may use the separately calibrated Monod parameters of eq 9, Table 2, and Table 3 to compare our results with other calibrations of the aerobic utilization rate of petroleum hydrocarbon substrates. These rates vary widely in form, phase, and partition, and should be computed carefully as a consequence. Although the other studies calibrate petroleum hydrocarbon data, they reflect varying source composition, nutrient concentrations, soil characteristics, microbial populations, soil moisture, and soil temperature. They also describe substrate utilization rather than growth kinetics, and so compare with a range of our values that rest on the yield coefficients of Tables 2 and 3. In view of all the above sources of variability, the comparisons are order of magnitude at best. Simulations (29) and spill site investigations are still needed to scale hydrocarbon remediation kinetics from the microcosm up to the field, as noted by Madsen (30). The published dissolved phase Monod kinetics for one of a blend of constituents in unsaturated soil are easiest to discuss because they compare most directly to our data and VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2347

FIGURE 2. Sensitivity of root-mean-square error δ (contours shown in percent) to KS and µM for different constituents, with all other parameters set at their calibrated values. Zero order alkane (3-methylheptane) is insensitive to KS, while first order aromatic (ethylbenzene) has constant µM/KS ratio, indicated by slotted minimum error region. The n-nonane exhibits true Monod behavior, shown by defined, closed minimum error contour. separate partitioning from the kinetics. Hohener et al. (4) report a maximum specific substrate utilization rate of 0.96 g dissolved m-xylene/g biomass-day in an unsaturated column, which lies in our µ5/Y5 calibration range of 0.23-1.6 g m-xylene/g biomass-day. Chen et al. (11) calibrate a value of 8.3 g benzene/g biomass-day in a blend of aromatics, somewhat larger than our µ4/Y4 range of 0.53-3.7 g ethylbenzene/g biomass-day. Hohener et al. (4) measure maximum specific utilization rates of 2.5 g dissolved n-octane/g biomass-day and 0.21 g dissolved n-hexane/g biomass-day. Table 3 gives ranges of 0.057 < µ7/Y7 < 0.103 day-1 for n-nonane and 0.063 < µ10/Y10 < 0.15 day-1 for n-decane; our n-decane agrees with the n-hexane of ref 4, but our n-nonane and the n-octane rate found by ref 4 differ markedly. Maximum specific total source substrate utilization rates from the groundwater literature may be compared to a superposition (µM/Y) T of our biodegradable constituent rates

() µM Y

T

≈

N

µMJ

J)1

YJ

∑

(excluding 2)

(10)

Schirmer et al. (3) calibrate an aerobic rate of 4.1 g dissolved m-xylene/g biomass-day in groundwater microcosms, with m-xylene as the single source compound. Total maximum specific substrate utilization rates of hydrocarbon mixtures may also be found in the groundwater literature: Borden and Bedient (9) specify 1.7 g hydrocarbons/g biomass-day, and Essaid et al. (12) calibrate a value of 0.5 g hydrocarbons/g biomass-day. Table 2, Table 3, and eq 10 specify a range of 2.1 < (µM/Y)T < 13 dissolved substrate/g biomass-day that is comparable to these literature values. Most of our total is aromatic, due to the combination of relatively high µMJ and low YJ for these constituents. This trend is also noted by the 2348

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 7, 2007

modelers: the high value of ref 3 represents aromatic degradation, while refs 9 and 12 include the slower degrading alkanes in the creosote and crude oil spills. First-order dissolved substrate utilization kinetics also require biomass and KSJ estimates, and so do not compare as directly as the maximum specific rates. Hers et al. (13) estimate a first order, dissolved phase biodegradation rate of 12 day-1 for m- and p-xylenes in unsaturated, vapor contaminated soil below the slab of a petrochemical plant. This estimate lies within the 2.7-18 day-1 m-xylene range of µM5XO/(Y5KS5) from eq 9a and Table 2. Nielsen et al. (31) observe a first order, sorbed, dissolved phase biodegradation rate varying from 0.04 to 0.1 day-1 for o-xylene utilization in in situ groundwater microcosms. This is less than the 0.21.3 day-1 range for µM6XO/[(1 + MSKD6/VW)Y6KS6] from Table 3, adjusted for sorption and deposition (eq 3b and Table 1). These estimates incorporate biomass, which is likely to be higher in the unsaturated zone, leading to our overprediction of the groundwater rate found by ref 31. Sorption and deposition are important to the ref 31 comparison since they reduce the true first-order rate to an apparent (retarded) value an order of magnitude lower.

Acknowledgments The Massachusetts Highway Department funded this research under Interagency Service Agreement 38721. The views, opinions, and findings contained in this paper are the authors’, and do not necessarily reflect the official view or policies of MassHighway. This paper does not constitute a standard, specification, or regulation.

Literature Cited (1) Richards, R. J.; Ostendorf, D. W.; Switzenbaum, M. S. Aerobic soil microqcosms for long-term biodegradation of hydrocarbon vapors. Hazard. Waste Hazard. Mater. 1992, 9, 397-410. (2) Aelion, C. M.; Bradley, P. M. Aerobic biodegradation potential of subsurface microorganisms from a jet fuel-contaminated aquifer. Appl. Environ. Microbiol. 1991, 57, 57-63. (3) Schirmer, M.; Butler, B. J.; Roy, J. W.; Frind, E. O.; Barker, J. F. A relative-least-squares technique to determine unique Monod kinetic parameters of BTEX compounds using batch experiments. J. Contam. Hydrol. 1999, 37, 69-86. (4) Hohener, P.; Duwig, C.; Pasteris, G; Kaufmann, K.; Dakhel, N.; Harms, H. Biodegradation of petroleum hydrocarbon vapors: laboratory studies on rates and kinetics in unsaturated alluvial sand. J. Contam. Hydrol. 2003, 66, 93-115. (5) Lahvis, M. A.; Baehr, A. L.; Baker, R. J. Quantification of aerobic biodegradation and volatilization rates of gasoline hydrocarbons near the water table under natural attenuation conditions. Water Resour. Res. 1999, 35, 753-765. (6) Atlas, R. M. Microbial degradation of petroleum-hydrocarbons - an environmental perspective. Microbiol. Rev. 1981, 45, 180209. (7) Bossert, I.; Bartha, R. In The Fate of Petroleum in Soil Ecosystems; Atlas, R. M. Ed.; MacMillan: New York, NY, 1984; pp 453-474. (8) Robinson, J. A.; Tiedje, J. M. Non-linear estimation of Monod growth kinetic-parameters from a single substrate depletion curve. Appl. Environ. Microbiol. 1983, 45, 1453-1458. (9) Borden, R. C.; Bedient, P. B. Transport of dissolved hydrocarbons influenced by oxygen-limited biodegradation. 1. Theoretical development. Water Resour. Res. 1986, 22, 1973-1982. (10) Molz, F. J.; Widdowson, M. A.; Benefield, L. D. Simulation of microbial-growth dynamics coupled to nutrient and oxygentransport in porous media. Water Resour. Res. 1986, 22, 12071216. (11) Chen, Y. M.; Abriola, L. M.; Alvarez, P. J. J.; Anid, P. J.; Vogel, T. M. Modeling transport and biodegradation of benzene and toluene in sandy aquifer material - comparisons with experimental measurements. Water Resour. Res. 1992, 28, 1833-1847. (12) Essaid, H. I.; Bekins, B. A.; Godsy, E. M.; Warren, E.; Baedecker, M. J.; Cozzarelli, I. M. Simulation of aerobic and anaerobic biodegradation processes at a crude oil spill site. Water Resour. Res. 1995, 31, 3309-3327. (13) Hers, I.; Atwater, J.; Li, L.; Zapf-Gilje, R. Evaluation of vadose zone biodegradation of BTX vapours. J. Contam. Hydrol. 2000, 46, 233-264. (14) Ostendorf, D. W.; Hinlein, E. S.; Schoenberg, T. H. In Bioremediation of Contaminated Soils; Wise, D. L., Trantolo, D. J., Cichon, E. J., Inyang, H. I., Stottmeister, U., Eds.; Marcel Dekker: New York, NY, 2000; pp 607-632. (15) Chiou, C. T.; Porter, P. E.; Schmedding, D. W. Partition equilibria of non-ionic organic-compounds between soil organic-matter and water. Environ. Sci. Technol. 1983, 17, 227-231.

(16) Hoff, J. T.; Gillham, R.; Mackay, D.; Shiu, W. Y. Sorption of organic vapors at the air-water-interface in a sandy aquifer material. Environ. Sci. Technol. 1993, 27, 2789-2794. (17) Baehr, A. L. Selective transport of hydrocarbons in the unsaturated zone due to aqueous and vapor-phase partitioning. Water Resour. Res. 1987, 23, 1926-1938. (18) Mackay, D.; Shiu, W. Y.; Ma, K. C. Physical-Chemical Properties and Environmental Fate; Chapman & Hall/CRCnetBase: New York, 1999. (19) Freijer, J. I. Mineralization of hydrocarbons in soils under decreasing oxygen availability. J. Environ. Qual. 1996, 25, 296304. (20) Grady, C. P. L.; Lim, H. C. Biological Wastewater Treatment; Marcel Dekker: New York, NY, 1980. (21) Leach, L. E.; Beck, F. P.; Wilson, J. T.; Kampbell, D. H. Proceedings of the 2nd National Outdoor Action Conference on Aquifer Restoration, Groundwater Monitoring, and Geophysical Methods; NGWA: Las Vegas, NV, 1988; pp 31-51. (22) Beveridge, G. S. G.; Schechter, R. S. Optimization: Theory and Practice; McGraw-Hill: New York, NY, 1970. (23) Alexander, M. Introduction to Soil Microbiology; Krieger: Malabar, FL, 1991. (24) Solano-Serena, F; Marchal, R; Lebeault, J. M.; Vendercasteele, J. P. Distribution in the environment of degradative capacities for gasoline attenuation. Biodegradation 2000, 11, 29-35. (25) Atlas, R. M.; Bartha, R. Microbial Ecology; Addison-Wesley: Reading, MA, 1998. (26) Song, H. G.; Bartha, R. Effects of jet fuel spills on the microbial community of soil. Appl. Environ. Microbiol. 1990, 56, 646651. (27) Mu, D. Y.; Skow, K. M. Effect of trichloroethylene and toluene concentrations on TCE and toluene biodegradation and the population-density of TCE and toluene degraders in soil. Appl. Environ. Microbiol. 1994, 60, 2661-2665. (28) Alexander, M. Biodegradation and Bioremediation; Academic: New York, NY, 1994. (29) Abreu, L. D. V.; Johnson, P. C. Simulating the effect of aerobic biodegradation on soil vapor intrusion into buildings: Influence of degradation rate, source concentration, and depth. Environ. Sci. Technol. 2006, 40, 2304-2315. (30) Madsen, E. L. Determining in situ biodegradationsfacts and challenges. Environ. Sci. Technol. 1991, 25, 1663-1673. (31) Nielsen, P. H.; Bjerg, P. L.; Nielsen, P.; Smith, P., Christensen, T. H. In situ and laboratory determined first-order degradation rate constants of specific organic compounds in an aerobic aquifer. Environ. Sci. Technol. 1996, 30, 31-37.

Received for review September 27, 2006. Revised manuscript received December 25, 2006. Accepted February 8, 2007. ES062313L

VOL. 41, NO. 7, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2349