Novel Chamber to Measure Equilibrium Soil–Air Partitioning

Environ. Sci. Technol. 2008, 42, 4870–4876

Novel Chamber to Measure Equilibrium Soil–Air Partitioning Coefficients of Low-Volatility Organic Chemicals under Conditions of Varying Temperature and Soil Moisture ´ W O L T E R S , * ,† ANDRE VOLKER LINNEMANN,‡ KILIAN E. C. SMITH,§ EVA KLINGELMANN,| BYUNG-JUN PARK,⊥ AND HARRY VEREECKEN Forschungszentrum Ju ¨ lich GmbH, Institute of Chemistry and Dynamics of the Geosphere IV: Agrosphere, D-52425 Ju ¨ lich, Germany

Received February 7, 2008. Revised manuscript received April 3, 2008. Accepted April 3, 2008.

The need to determine soil–air partitioning coefficients (KSA) of low-volatility organic chemicals as a measure of their distribution in the soil surface after release into the environment resulted in the development of a novel chamber system, which has been filed for patent. A major advantage of this pseudo-static system is that sufficient time can be factored into the experiment to ensure that the system has achieved equilibrium. In a highly precise method, the air is collected in adsorption tubes and subsequently liberated in a thermodesorption system for the quantitation of the adsorbed compound. The precision of the method is great enough that even the effects of temperature and soil moisture on the soil–air partitioning of very low-volatility compounds can be quantified. Because of analytical detection limits, quantitation of these influences has not been possible to date. Functionality of the setup was illustrated by measurements on the fungicide fenpropimorph. KSA values of fenpropimorph displayed a negative relationship with temperature and soil moisture. The type of application (spraying or incorporation) and the use of formulated compounds was shown to have a major impact on the measured KSA values. Comparison with calculations using an estimation method revealed that the use of experimentally determined KSA values will facilitate a more adequate consideration of volatilization in recent model approaches. * Corresponding author phone: +49-621-718858-0; fax: +49-621718858-10; e-mail: [email protected]. † Current address: Dr. Knoell Consult GmbH, Dynamostrasse 19, D-68165 Mannheim, Germany. ‡ Current address: North Rhine Westphalia State Agency for Nature, Environment and Consumer Protection, FB 64, Leibnizstrasse 10, D-45659 Recklinghausen, Germany. § Current address: Department of Environmental Chemistry and Microbiology, National Environmental Research Institute, Frederiksborgvej 399, P.O. Box 358, DK-4000 Roskilde, Denmark. | Current address: Technische Universita¨t Berlin, Institute for Ecology, Salzufer 12, D-10587 Berlin, Germany. ⊥ Current address: National Institute of Agricultural Science and Technology, RDA, Suwon 441-707, Republic of Korea. 4870

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 13, 2008

Introduction Some low-volatility organic chemicals (e.g., pesticides) are intentionally released into the environment at large during their application. Others (e.g., polychlorinated biphenyls (PCBs)), although produced for use in a number of industrial applications, were never intended to be directly released into the environment. Nevertheless, they have found their way into environments far away from their original site of use or disposal (1). In the terrestrial environment, it is the soil compartment that contains the largest mass fraction of these chemicals. To assess the risk related to the presence of these chemicals in soils, there is an urgent requirement to understand the behavior of such chemicals in the soil. One particular area of concern is the ability of such soil sequestered chemicals to partition back to the atmosphere. Here, they will once more become available for long-range transport to pristine regions and for entry into natural and wildlife food chains. Therefore, the soil–air distribution has been shown to have a major impact on the environmental fate of many organic chemicals. Even though promising approaches have been developed to predict the partitioning of such organic chemicals to soil, particularly under dry conditions (2), measurements of soil–air distribution are still at an early stage of development for very low-volatility compounds. For determining soil–air distributions, dynamic systems (flow-through systems) have been used to date. Here, a purified air stream is passed through a column filled with soil so that equilibrium is reached between the soil and the air phases (3, 4). Sampling of the head space in the chambers into which the air stream enters after traversing a chamber containing soil has up to now been used exclusively for determining the kinetics of volatilization but would also allow for the calculation of soil–air partitioning coefficients by use of the measured equilibrium concentrations in the soil and air (5, 6). A potential problem in the use of dynamic methods lies in the attainment of equilibrium between soil and air phases. This might not necessarily be reached for very involatile compounds, even at very low flow rates often used. In addition, this difficulty is compounded by soil processes since for the determination of the partitioning coefficient, it is necessary for an equilibrium between the air and the upper soil layer (which is in direct contact with the air) to be established. A precondition for equilibrium is, however, that the exchange with air does not lead to an emptying of the upper soil layer. Substance losses at the soil surface must be compensated for by diffusion from the deeper soil layers. This kinetic problem could result in a disequilibrium during measurement, leading to partitioning coefficients that are not constant over time. To date, it has not been possible to measure soil–air partitioning coefficients of low-volatility compounds. Furthermore, analytical detection limits, and sampling reproducibility in particular, pose a problem in characterizing temperature and soil moisture effects such that quantification of these influences has not been possible with sufficient precision. It was thus the goal of this work to develop a device for determining the phase partitioning coefficients of lowvolatility substances in the soil–air system. The initial objective was the optimization and adjustment of the chamber features to ensure the reliability of the measurement procedure and to obtain reproducible results. In the second stage, experiments were performed investigating the soil–air partitioning of the fungicide fenpropimorph under conditions 10.1021/es800372m CCC: $40.75

 2008 American Chemical Society

Published on Web 05/29/2008

FIGURE 1. Construction and operation of soil-air partitioning chamber. Nine sealable sampling ports are available for air sampling, temperature, and humidity sensors (not shown, three in each cap and three along the length of the chamber). For air sampling, air is drawn using a programmable pump through thermal desorption tubes filled with a solid adsorbent (Tenax). Analysis was performed by thermal desorption GC-MSD.

FIGURE 2. Air supply and air sampling unit used in the soil-air partitioning chamber. T1-3 are thermosensors. of varying temperature and soil moisture as compared to predictions of an estimation method.

Theoretical Considerations Soil–Air Partitioning. An organic chemical in the soil will partition between the soil solids, the interstitial soil solution, and the gas filled soil pores (4). For a sorbed chemical to volatilize from the surface of the soil, it must first desorb from the soil solids into the soil interstitial solution, from where it can partition into the soil air. Once in the soil air at the surface, there is the potential for transportation across the boundary layer and into the bulk atmosphere (see Figure S1, Supporting Information). Detailed considerations on the partitioning between the soil solids and the soil gas phase can be taken from the Supporting Information. Estimation Method by Smit et al. (7). The estimation method by Smit et al. (7) allows pesticide volatilization from bare soil using a regression equation to be predicted. This method correlates cumulative volatilization reported in the literature to the calculated fraction of the pesticide in the gas phase (8). Data used as input parameters for the calculations on fenpropimorph can be taken from Table S1, Supporting Information. A comparison of the calculated fraction in the gas phase with the measured fraction enables an evaluation of the predictive use of estimation approaches based on the underlying equation (see Supporting Information).

Experimental Procedures Chamber Design and Construction. In the following discussion, the main features of the novel chamber system are summarized. A detailed description of the setup can be taken from the underlying patent applications (9). The basic element of the phase partitioning chamber is a double-walled glass tube (1.00 m in length, i.d. of 0.15 m) equipped with several sampling ports situated at regular intervals along the

chamber body and glass threads used for connecting the chamber with a cooling system and for fixing the measuring instruments (Figure 1). The total chamber volume is 17.7 L. To minimize sorption artifacts, the whole chamber is constructed out of glass. The use of sealing rings and quickrelease caps on both sides of the glass tube enables gastight sealing of the whole chamber. The glass axle is provided with propellers for mixing the air and adjusted in the chamber by ball-and-socket joints installed in both caps (Figure 3A). On one side of the apparatus, the axle is connected to a stirrer. The air sampling unit (Figure 2) is attached to the chamber via an additional sampling port in the cap on the opposite side of the air inflow. The chamber can be run using a cooling-heating system at temperatures between 5 and 40 °C. The chamber body and end caps were insulated using Armaflex insulating tape to ensure an even temperature throughout. To minimize photodegradation of susceptible compounds, the chamber was provided with a reflecting foil layer under the insulating material to block out light. For experiments, a metal tray (0.90 m × 0.10 m with a wall height of 0.04 m) filled with a very thin soil layer (layer thickness of ∼5 mm), which was adjusted to a defined water content and a defined pesticide concentration, was put into the glass tube. Purified air was fed into the glass chamber. The problem of gastight fittings at each end seizing up when the chamber is run over prolonged periods was resolved by the application of a thin layer of silicone grease to the fittings, taking care that this did not cover the part of the glass shaft in contact with the chamber air. Air Supply and Air Sampling Unit. A suction pump (GS 301, Desaga GmbH, Wiesloch, Germany) allowed for adjustment of defined air exchange rates ranging from 0.1 to 1.5 L min-1. The incoming air was purified by activated charcoal and passed through two wash bottles filled with water that VOL. 42, NO. 13, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4871

FIGURE 3. Glass axles to be used for studies on soil-air partitioning and Henry’s law constants. (A) Glass axle to be used for studies on soil-air partitioning (total length: 184 cm, 1.4 cm i.d.); 12a-d: ingress of air and 13a-d: propellers for mixing air. (B) Sandblasted glass cylinder to be used for studies on Henry’s law constants (total length: 181 cm, 9.5 cm o.d.); 10a,b: part of cylinder that sticks out through the central opening (9). was maintained at the same temperature as the chamber cooling-heating system to bring the air humidity of the incoming air up to near saturation. The temperature gradient along the entire chamber body was less than 1 °C. After leaving the chamber, the compounds contained in the exhaust air were trapped on Tenax tubes and quantified by subsequent thermal desorption. Studies on Soil–Air Partitioning. For experiments on the temperature dependence of the soil–air partitioning of fenpropimorph, 500 g of air-dried and 2 mm sieved soil (gleyic cambisol; CORG ) 0.931% and soil bulk density ) 1250 kg m-3) was used. A 50 g subsample was taken and added to a mortar. One milliliter of a methanolic spike solution of fenpropimorph (purity of 95% active substance (a.s.), supplied by Dr. Ehrenstorfer, Augsburg, Germany) at a concentration of 1 µg mL-1 was added to the surface of the soil, taking care that the sides of the mortar were not contaminated. The soil was allowed to become air-dry and subsequently homogenized. The spiked soil inoculum was added to the remaining 450 g of soil to a solvent rinsed jar, resulting in a soil concentration of 2 µg kg-1. The jar was sealed and wellmixed on a rotary shaker for at least 2 h. Immediately prior to use, the spiked soil was adjusted to the required soil moisture of 40% MWC (maximum water holding capacity) by adding the appropriate volume of distilled water. Half of the water volume was added to the metal tray, the soil was transferred to the tray, and finally the remaining water was 4872

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 13, 2008

sprayed on the soil surface. The tray was introduced into the chamber, and the system was sealed. Air was drawn continuously at a flow rate of 0.2 L min-1 through the chamber and the downstream Tenax tubes. This corresponds to a turnover time of the air in the chamber of approximately 1.5 h. A sampling duration of 5 h per Tenax tube was chosen, and about 10 Tenax tubes were sampled per experiment. The constant moisture content of the soil was monitored by initial and final weighings of the soil tray. For experiments on the effect of soil moisture on the partitioning of fenpropimorph applied as an EC (emulsion concentrate) formulation Corbel (supplied by BASF AG, Limburgerhof, Germany), a modified application procedure was used. One milliliter of an aquatic Corbel solution at a fenpropimorph concentration of 100 µg mL-1 was added to the second half of the water used for adjustment to the required soil moisture, and this was sprayed on the soil surface. This resulted in a final soil concentration of 200 µg kg-1 fenpropimorph. Soil Extraction. For the determination of soil concentrations, a cleanup method for the spiked soil used in the chamber was developed. Twenty grams of soil was weighed into cellulose thimbles and Soxhlet extracted for 12 h using 100 mL of acetone. The Soxhlet extract was filled in a separating funnel, and 400 mL of water, 50 mL of saturated sodium chloride solution, and 50 mL of dichloromethane (DCM) were added. The mixture was shaken for 2 min and allowed to separate, and the bottom DCM layer was collected in a solvent cleaned round-bottomed flask. Another 50 mL of DCM was added, and the mixture was shaken for a further 2 min. The DCM layer was collected with the previous fraction, rotary evaporated to dryness, and taken up in 500 µL of hexane. The sample was added to a silica column (silica slurry packed in hexane into a column with a tap until first mark; sodium sulfate (baked overnight at 250 °C) was added until the second mark) and eluted with 10 mL of hexane, 5 mL of hexane/DCM (1:1), and 10 mL of DCM. An injection standard (20 µg of 1-methylnaphthalene) was added to the various fractions that were then all rotary evaporated to approximately 500 µL final volume and quantitated by a gas chromatography/mass selective detector (GC/MSD). In preliminary studies, the losses during the cleanup procedure of soil were quantified by comparing the levels in spiked and unspiked soil samples. Three replicates each of 20 g of soil were spiked with fenpropimorph (2 µg; equivalent to a soil loading of 100 µg kg-1). In addition, three soil samples remained unspiked. The replicates were Soxhlet extracted and cleaned as described previously. Recoveries were calculated by subtracting the blank levels from the levels found in the spiked samples. Thermodesorption System (TDS). A combination of adsorption and subsequent thermal desorption was used for air analysis. Here, the components were initially trapped on a selective adsorbent (Tenax tubes) and after achievement of the required enrichment were subsequently thermally desorbed and transferred onto a GC column. The system consisted of a TDS (TDS-2, Gerstel GmbH, Mu ¨lheim a.d. Ruhr, Germany), a temperature programmable cooled injection system (CIS-4), and a GC instrument equipped with massselective detector. TDS tubes were spiked with 10 µL of internal standard, 1-methylpyrene, prior to analysis. The adsorbed compounds were thermally desorbed (TDS-2), cryofocused in the CIS-4 (solvent vent mode; initial temperature of -50 °C, programmed at 12 °C/min to 260 °C, 1 min hold), and transferred onto the GC column. GC-MSD. Sample analysis was conducted using a HewlettPackard 6890-5973 GC-MSD (Hewlett-Packard, Ratingen, Germany) equipped with a 25 m fused silica DB-5 capillary column with a 0.25 mm i.d. and 0.32 µm film thickness. The

TABLE 1. Temperature Dependence of Soil–Air Partitioning of Active Substance (a.s.) Fenpropimorpha temp (°C)

CAIR a.s. (pg L-1)

CSOIL a.s. (µg kg-1)

KSA (L g-1)

measured fraction in gas phase

10

1.16

2

1.731 × 103

2.045 × 10-5

40

1.40c (1.22 and 1.58)

2

1.450 × 103c (1.634 × 103 and 1.265 × 103)

2.483 × 10-5c (2.166 × 10-5 and 2.799 × 10-5)

a

KSA is soil–air partitioning coefficient. b According to Smit et al. (7). Tenax tubes). Single values are given in parentheses.

MSD source (held at 230 °C) was operated in positive electron ionization mode, while the quadrupole mass filter (held at 150 °C) was operated in selective ion monitoring (SIM) mode (the MS fragments (m/z) monitored for fenpropimorph were 128 (100%), 173 (3.2%), and 303 (M+, 4.3%)). The injector and GC-MSD transfer line were operated at 240 and 290 °C, respectively. An HP 6890 series autoinjector was used to inject 1 µL of sample in splitless mode. The oven temperature started at 90 °C (3 min hold) and was then programmed at 8 °C/min to 280 °C and held for 1 min. Throughout the run, the carrier gas (helium) was maintained at 1.0 mL min-1. Pesticide concentrations were calculated by comparing the ratio of peak area response of the target analyte to that of the internal standard (1-methylpyrene) measured in samples to those of calibration standards covering the appropriate concentration range.

Results and Discussion Validation and Preliminary Studies. The development of a cleanup method for the spiked soil samples used in the partitioning chamber was necessary to allow for the determination of the soil concentrations for use in the soil–air distribution calculations. Recoveries of the cleanup procedure were calculated by subtracting the blank soil levels from the levels found in the spiked soil samples. Recoveries of fenpropimorph covered an acceptable range of values between 80 and 115%. In addition, the homogeneity of the soil concentrations was documented by low standard deviations up to 8%. The soil–air partitioning coefficient pertains to the gaseous concentration in the air. Possibly, ultrafine particles could have been suspended from the soil surface during the running of the experiment. Thus, inclusion of a particle fraction in the air concentration measurements would lead to incorrect estimation of the soil–air partitioning coefficient and in particular to its relationship with environmental variables such as temperature or relative humidity. Although soil resuspension should not have been an issue, given the very low mixing of the air in the chamber, this artifact was tested for by employing a glass fiber filter prior to the gaseous adsorbent. The levels of fenpropimorph measured on the glass fiber filters were negligible, and thus, this potential artifact was discounted. Preliminary studies and calculations allowed for an estimation of the flow rate and running time required for sampling sufficient amounts of pesticide to exceed analytical detection limits. The running time was expected to have consequences for the practicality of the chamber to perform the required measurements, the significance of potential artifacts, and the logistics of the study as a whole. With reference to the case of application of fenpropimorph, a running time of 5 h and a flow rate of approximately 0.2 L min-1 was appropriate to ensure a sufficient mass on the Tenax tubes and thus a reliable detection of the Tenax trapped compound. The volumes of soil used in the chamber studies should contain a sufficient mass of compounds to avoid depletion

c

calcd fraction in gas phaseb 2.518 × 10-7 (1.2% of measured fraction) 3.624 × 10-6 (14.6% of measured fraction)

Mean value of two experiments (each with 10

during the running of the experiment. Here, a further confounding factor to consider is the nature of the compound reservoir in the soil involved in the equilibration process. There exists evidence from a number of studies of a rapidly desorbing fraction of pesticide in soil and a more inaccessible slow desorbing fraction (12). For the purpose of this exercise, a conservative estimate of the fast responding surface compartment was preferable. This was set at 10% of the total mass of contaminant in the soil. For fenpropimorph, a soil concentration of 2 µg kg-1 applied in the chamber was sufficient to avoid depletion. At this soil concentration, the expected final total mass of the compound in the air was sufficiently small that the soil reservoir was negligibly depleted. The fundamental assumption in measuring the soil–air partitioning coefficient is that the system attains equilibrium. The major advantage of the pseudo-static system used here is that the small air flow of 0.2 L min-1 maximized the contact time between air and soil. Conditions close to equilibrium were verified by sequential measurements of the air concentrations. As the air and soil approached equilibrium, the air concentrations reached a plateau. Equilibrium conditions were attained more rapidly by stirring the air in the chamber using a glass axle provided with propellers (Figure 3A) to reduce any boundary layer above the soil surface. This results in a large surface area/volume ratio of soil available for exchange with air. Temperature Dependence of Soil–Air Partitioning of Fenpropimorph. Studies on the soil–air partitioning of fenpropimorph were performed at temperatures of 10 and 40 °C. The fraction in the gas phase was calculated as the ratio of the total pesticide mass in the gaseous phase of the chamber to the amount remaining in the soil (Table 1). An increase of approximately 20% of the fraction in the gas phase was determined when the temperature was increased from 10 to 40 °C (corresponding to a decrease of KSA (soil–air partitioning coefficient) from 1.731 × 103 to 1.450 × 103 L g-1), thus indicating that increases in temperature are accompanied by an increasing tendency toward partitioning into the gas phase. Temperature influences partitioning mainly through its effect on vapor pressure and through its effect on the movement of the chemical to the surface by diffusion or mass flow in evaporating water. Because of these effects, increases in temperature are usually associated with increases in volatilization rate. However, this may not always be the case since an increase in temperature is also associated with an increase in the drying rate of the soil surface, thereby possibly decreasing vapor density and resulting in less volatilization than at lower temperatures (13). To achieve water-saturated air and thus to prevent drying of the soil surface, the air stream passes through two wash bottles (Figure 2), so the decrease of KSA values can be exclusively attributed to the increase in temperature. Measurements were compared to the calculated fraction in the gas phase using the estimation method by Smit et al. (7). The formulation of the postapplication pesticide distribution over the gas, liquid, and solid phases resulted in a VOL. 42, NO. 13, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4873

TABLE 2. Soil Moisture Dependence of Soil–Air Partitioning of Corbela temp (°C)

soil moisture (% MWC)d

CAIR a.s. (pg L-1)

CSOIL a.s. (µg kg-1)

KSA (L g-1)

measured fraction in gas phase

25

40

1.13

200

177 × 103

1.996 × 10-7

25

70

3.62

200

55.2 × 103

6.414 × 10-7

25

glass surface applicationc

8.50

calcd fraction in gas phaseb 9.806 × 10-7 (491% of measured fraction) 7.199 × 10-7 (112% of measured fraction)

a

KSA is soil–air partitioning coefficient; concentrations and fractions in the gas phase refer to the a.s. fenpropimorph. According to Smit et al. (7). c Application of 1 mL of aquatic Corbel solution (100 mg L-1 a.s.) to glass plate. d MWC is maximum water holding capacity.

b

calculated pesticide fraction in the gas phase of 2.518 × 10-7 at 10 °C and 3.624 × 10-6 at 40 °C, thus underestimating the measured value by up to 2 orders of magnitude. Deviations between measurements and calculations might be attributed to the underlying assumptions used in the calculations (e.g., linear adsorption isotherms), which are supposed to be idealized. For several pesticides, a deviation from linearity was observed with a gradual decrease in the soil–water partitioning with increasing apparent pesticide equilibrium concentration, resulting in a nonlinear isotherm with a negative curvature (13, 14). In general, nonlinearity is observed, especially with pesticides that are not extremely hydrophobic and therefore not limited by solubility to extremely low concentrations (15). With regards to volatilization and its description in models, the use of overestimated soil sorption partitioning coefficients will generally result in underestimated volatilization rates, regardless as to whether a simple screening approach or a more sophisticated model is used. Consequently, the introduction of experimentally determined soil–air partitioning data instead of calculated values is the only way to take into consideration this effect and will contribute to a clear improvement of model approaches. Soil Moisture Dependence of Soil–Air Partitioning of Corbel. Investigations to elucidate the soil moisture effect on soil–air partitioning were performed after spraying of Corbel (EC formulation of fenpropimorph) directly onto the soil surface. The soil-surface application of this widely used formulation was performed at an increased application rate of 200 µg kg-1 fenpropimorph. This adapted the application conditions close to those of agricultural practice (10). An increase of the soil moisture from 40 to 70% MWC while maintaining a constant temperature of 25 °C resulted in an increase of CAIR from 1.13 to 3.62 pg L-1, corresponding to KSA values of 177 × 103 and 55.2 × 103 L g-1, respectively, thus clearly revealing an increased tendency toward transfer into the gas phase under moist conditions (Table 2). This tendency is mainly due to increasing soil adsorption under dry conditions, which causes a reduction of the effective vapor pressure and results in a lower volatilization tendency (11). The highest concentration in the air was measured after application to the glass surface (CAIR ) 8.50 pg L-1). Obviously, the glass surface did not permit any significant adsorption of fenpropimorph, as generally has been shown for pesticides (24). The effective vapor pressure of fenpropimorph applied to soil is likely to be lower than the pure vapor pressure of fenpropimorph applied to glass due to interactions of the pesticide with the soil matrix, consequently resulting in a reduced partitioning tendency of soil-applied fenpropimorph. KSA values of fenpropimorph applied as EC formulated Corbel were markedly higher than those of fenpropimorph applied as a pure compound. This finding might be attributed to the effects of adjuvants, including surfactants used in the pesticide formulation. Solubilizing effects leading to apparently increased water solubility and reduced volatilization of 4874

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 13, 2008

a pesticide are commonly known through the preparation of formulations (25). The calculated fractions in the gas phase using the estimation method by Smit et al. (7) agreed reasonably well with the measured value for the experiment at 70% MWC. For the experiment at 40% MWC, the calculated fraction in the gas phase was overestimated by a factor of about 5. It is striking to note that the model predicted decreasing volatilization with increasing soil moisture, as illustrated by a decrease of the calculated fractions in the gas phase from 9.806 × 10-7 (40% MWC) to 7.199 × 10-7 (70% MWC). This is a general limitation of currently available volatilization approaches: existing models explicitly assume partitioning coefficients as being independent of the water content and consequently calculate lower pesticide vapor pressures at the surface under moist conditions (16, 17). Consequently, these model predictions are not in line with the well-known tendency of pesticides toward enhanced volatilization at high soil moisture contents (18, 19). Even the calculations of model approaches used in the registration procedures of pesticides, such as PELMO (20), PEARL (21), or PRZM (22), are affected by the same weaknesses. Recently, some of the underlying assumptions used in these models were improved, particularly by including increased sorption of pesticides in dry soils (23). However, a description of the water content in the top soil layer is still subject to uncertainty, and consequently, the extent of the predicted volatilization increase after irrigation markedly differs from measured values. Thus, future model improvement will be very closely connected with advanced experimentation to obtain more detailed information on soil–air partitioning, finally resulting in a calibration or adjustment of model approaches. The novel chamber enables a reliable quantitation of the phase distribution at the soil–air interface and will facilitate a more adequate description of the soil moisture dependence of volatilization as part of PEC (predicted environmental concentration) models. Regarding the behavior of pesticides in the soil, a number of important questions remain unresolved. It is not clear as to whether the total amount measured in the soil using conventional extraction methodologies is in fact available for exchange with the atmosphere. The fraction that is available for soil–air partitioning may decrease over time as the pesticide becomes more tightly bound to the soil matrix. The question arises as to whether this association is reversible or not (e.g., when the organic matter component of the soil that acts as the main sorbing compartment undergoes degradation). A detailed understanding will eventually form an important part of the risk assessment process for such chemicals. In addition, more detailed investigations will address the question as to whether KSA is a linear function of compound physicochemical properties. Should KSA be related to compound physicochemical properties such as KOA (octanol–air partitioning coefficient), the experimental determination of this relationship would allow for the

prediction of the behavior of other similar chemicals for which no experimental data are available. Outlook for Measurement of Water–Air Partitioning. Until now, the broad application spectrum of the chamber has not been fully utilized, especially for reliable measurements of Henry’s law constants that are expected to deliver substantial progress in elucidating water–air partitioning. This would require reconfiguring the chamber by using a water reservoir and a sandblasted glass cylinder (Figure 3B) instead of a glass axle (Figure 3A). For this purpose, the rotary roller dips into the sample holder and picks up a water film during its rotation, which then comes into contact with the air in the chamber. The surface of the rotary roller is roughened so that the wetting of the surface with water is facilitated. Through use of the rotary roller, the surface area of the liquid is increased. To determine Henry’s law constant, air is fed into the housing and comes into contact with the water film covering the surface of the rotary roller. The air stream introduced through the connecting fittings can be discharged at the opposite side of the housing from the device and subsequently subjected to a quantitative chemical analysis. Even with state-of-the-art methods and devices for measurements of Henry’s law constants (26, 27), to date it has not been possible to measure distribution coefficients of low-volatility compounds such pesticides with dimensionless Henry’s law constants lying in the range of 10-7 to 10-9. The precision of the novel chamber system should be great enough that even the Henry’s law constants of these very low-volatility compounds can be determined (9). The use of the new setup to measure water–air partitioning of such low-volatility compounds may contribute to an improvement of model approaches, such as the rice-water quality model (RICEWQ), which calculates the mass volatilized using an empiric volatilization coefficient (28).

Acknowledgments Mention of brand names is for information only and does not imply endorsement or exclusion of other products that might also be suitable. For technical assistance, the authors thank M. Krause, S. Ko¨ppchen, and R. Niehaus. J. H. Smelt (Alterra Green World Research, Wageningen, The Netherlands) and G. D’Orsaneo are specially acknowledged for their support with the setup and optimization of the chamber.

Supporting Information Available Theoretical considerations of soil–air partitioning; images to illustrate diffusion steps involved in exchange of pesticides between soil and atmosphere (Figure S1) and exchange of gaseous chemicals between air and soil (Figure S2); and illustration of estimation method by Smit et al. (7) and summary of data used in this estimation method for calculations on fenpropimorph (Table S1). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Baker, J. E.; Eisenreich, S. J. Concentrations and fluxes of polycyclic aromatic hydrocarbons and polychlorinated biphenyls across the air-water interface of Lake Superior. Environ. Sci. Technol. 1990, 24, 342–352. (2) Goss, K.-U.; Buschmann, J.; Schwarzenbach, R. P. Adsorption of organic vapors to air-dry soils: Model predictions and experimental validation. Environ. Sci. Technol. 2004, 38, 3667– 3673. (3) Hippelein, M.; McLachlan, M. S. Soil-air partitioning of semivolatile organic compounds. 1. Method development and influence of physical-chemical properties. Environ. Sci. Technol. 1998, 32, 310–316. (4) Cousins, I. T.; McLachlan, M. S.; Jones, K. C. Lack of an aging effect on the soil-air partitioning of polychlorinated biphenyls. Environ. Sci. Technol. 1998, 32, 2734–2740. (5) Ayris, S.; Harrad, S. The fate and persistence of polychlorinated biphenyls in soil. J. Environ. Monit. 1999, 1, 395–401.

(6) Morrissey, F. A.; Grismer, M. E. Kinetics of volatile organic compound sorption/desorption on clay minerals. J. Contam. Hydrol. 1999, 36, 291–312. (7) Smit, A. A. M. F. R.; Van den Berg, F.; Leistra, M. Estimation Method for Volatilization of Pesticides from Fallow Soil; DLO Winand Staring Centre, Environmental Planning Bureau Series 2: Wageningen, The Netherlands, 1997. (8) Wolters, A.; Kromer, T.; Linnemann, V.; Ophoff, H.; Stork, A.; Vereecken, H. Volatilization of [14C]fluoranthene and [14C]diflufenican after soil surface application under field-like conditions: Measurement and comparison with different model approaches. Agronomie 2002, 22, 337–350. (9) D’Orsaneo, G.; Linnemann, V.; Smelt, J. H.; Smith, K. E. C.; Wolters, A. Method for determining phase distribution coefficients, in addition to a corresponding device. International Patent Application No. PCT/DE2002/004550, filed on Dec. 12, 2002, published on July 3, 2003; European Patent Application No. EP20020794993, September 22, 2004; U.S. Patent Application No. 2005/0074366, April 7, 2005; and Japanese Patent Application No. 2003-555186, May 12, 2005. (10) Klingelmann, E. Degradation Studies in Microcosms and Distribution Studies in a Novel Chamber System on the Fungicide Fenpropimorph in Soil. M.S. Thesis (in German), University of Bielefeld, Bielefeld, Germany, 2003. (11) Taylor, A. W.; Spencer, W. F. Volatilization and vapor transport processes. In Pesticides in the Soil Environment: Processes, Impacts, and Modeling; Cheng, H. H., Ed.; SSSA: Madison, WI, 1990; pp 213-269. (12) Cornelissen, G.; Van Noort, P. C. M.; Parsons, J. R.; Govers, H. A. J. Temperature dependence of slow adsorption and desorption kinetics of organic compounds in sediments. Environ. Sci. Technol. 1997, 31, 454–460. (13) Wolters, A. Pesticide Volatilization from Soil and Plant Surfaces: Measurements at Different Scales versus Model Predictions. Ph.D. Thesis, University of Technology, RWTH, Aachen, Germany, 2003; http://darwin.bth.rwth-aachen.de/opus3/volltexte/ 2003/600/pdf/03_130.pdf. (14) Wauchope, D. R.; Yeh, S.; Linders, J. B. H. J.; Kloskowski, R.; Tanaka, K.; Rubin, B.; Katayama, A.; Ko¨rdel, W.; Gerstl, Z.; Lane, M.; Unsworth, J. B. Pesticide soil sorption parameters: Theory, measurement, uses, limitations, and reliability. Pestic. Manage. Sci. 2002, 58, 419–445. (15) Chiou, C. T.; Porter, P. E.; Schmedding, D. W. Partition equilibria of non-ionic compounds between soil organic matter and water. Environ. Sci. Technol. 1983, 17, 227–231. (16) Baker, J. M.; Koskinen, W. C.; Dowdy, R. H. Volatilization of EPTC: Simulation and measurement. J. Environ. Qual. 1996, 25, 169–177. (17) Wolters, A.; Linnemann, V.; Herbst, M.; Klein, M.; Scha¨ffer, A.; Vereecken, H. Pesticide volatilization from soil: Lysimeter measurements versus predictions of European registration models. J. Environ. Qual. 2003, 32, 1183–1193. (18) Spencer, W. F.; Farmer, W. J.; Cliath, M. M. Pesticide volatilization as related to water loss from soil. J. Environ. Qual. 1973, 2, 284–289. (19) Wolters, A.; Kromer, T.; Linnemann, V.; Scha¨ffer, A.; Vereecken, H. A new tool for laboratory studies on volatilization: Extension of applicability of the photovolatility chamber. Environ. Toxicol. Chem. 2003, 22, 791–797. (20) Klein, M. PELMO: Pesticide Leaching Model, User’s Manual Version 2.01; Fraunhofer-Institute for Environmental Chemistry and Ecotoxicology: Schmallenberg, Germany, 1995. (21) Leistra, M.; Van der Linden, A. M. A.; Boesten, J. J. T. I.; Tiktak, A.; Van den Berg, F. PEARL Model for Pesticide Behavior and Emissions in Soil-Plant Systems; Descriptions of Pprocesses in FOCUS PEARL v. 1.1.1.; Alterra-Rapport 013, Alterra Green World Research: Wageningen, The Netherlands, 2001. (22) Carsel, R. F.; Imhoff, J. C.; Hummel, P. R.; Cheplick, J. M.; Donigian, A. S., Jr. PRZM-3, a Model for Predicting Pesticide and Nitrogen Fate in the Crop Root and Unsaturated Soil Zones: User’s Manual for Release 3.0; National Exposure Research Laboratory, Office of Research and Development, U.S. EPA: Athens, GA, 1997. (23) Wolters, A.; Klein, M.; Vereecken, H. An improved description of pesticide volatilization: Refinement of the Pesticide Leaching Model (PELMO). J. Environ. Qual. 2004, 33, 1629–1637. (24) Kromer, T.; Ophoff, H.; Stork, A.; Fu ¨ hr, F. Photodegradation and volatility of pesticides: Chamber experiments. Environ. Sci. Pollut. Res. 2004, 11, 107–120. (25) Katagi, T. Surfactant effects on environmental behavior of pesticides. Rev. Environ. Contam. Toxicol. 2008, 194, 71–177. VOL. 42, NO. 13, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4875

(26) Fendinger, N. J.; Glotfelty, D. E. A laboratory method for the experimental determination of air-water Henry’s law constants for several pesticides. Environ. Sci. Technol. 1988, 22, 1289– 1293. (27) Mackay, D.; Shiu, W. Y.; Sutherland, R. P. Determination of air-water Henry’s law constants for hydrophobic pollutants. Environ. Sci. Technol. 1979, 13, 333–337.

4876

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 13, 2008

(28) Ferrari, F.; Karpouzas, D. G.; Trevisan, M.; Capri, E. Measuring and predicting environmental concentrations of pesticides in air after application to paddy water systems. Environ. Sci. Technol. 2005, 39, 2968–2975.

ES800372M