Organic Carbon Partition Coefficient Research Articles

Sorption of humic acids and α‐endosulfan by clay minerals

Sorption of alpha-endosulfan by kaolinite and montmorillonite alone and in the presence of sorbed and dissolved humic acid (HA) was investigated (pH 8 and 25 degrees C). Three types of HA, Elliot soil HA (EHA), Peat HA (PHA), and Summit Hill HA (SHHA), were used to represent typical humic substances found in soils. For sorption of HA by either mineral, Freundlich sorption coefficient (K(f)) values appeared to decrease in the order of EHA > PHA > SHHA, which followed increasing polarity (expressed as the O/C atomic ratio) and decreasing percent-carbon content. For both clays, sorption of alpha-endosulfan by the HA mineral complex was greater than for sorption by the clay alone. Sorption of alpha-endosulfan by the HA mineral complexes followed the same order as the K(f) of the HAs (EHA > PHA > SHHA). Based on the amount of HA adsorbed by each mineral, organic carbon partition coefficients (K(oc)) were determined for sorption of alpha-endosulfan by two of the HA mineral complexes. The value of K(oc) for alpha-endosulfan sorption was greater for kaolinite EHA than kaolinite SHHA. However, the opposite trend was found with the montmorillonite HA complexes. Montmorillonite appeared to sorb alpha-endosulfan and/or HA with higher affinity than kaolinite, which likely is due to its 2:1 layer structure and higher surface area. Sorption of endosulfan diol, a hydrolysis product, by the minerals was much less than the parent pesticide.

Sensitive Parameters In Predicting Exposure Contaminants Concentration In A Risk Assessment Process

A sensitivity analysis (SA) was conducted on the analytical models considered in the risk-based corrective-action (RBCA) methodology of risk analysis, as developed by the American Society for Testing of Materials (ASTM), to predict a contaminant's concentration in the affected medium at the point of human exposure. These models are of interest because evaluations regarding the best approach to contaminated site remediation are shifting toward increased use of risk-based decision, and the ASTM RBCA methodology represents the most effective and internationally widely used standardized guide for risk assessment process. This paper identifies key physical and chemical parameters that need additional precision and accuracy consideration in order to reduce uncertainty in models prediction, thereby saving time, money and engineering effort in the data collection process. SA was performed applying a variance-based method to organic contaminants migration models with reference to soil-to-groundwater leaching ingestion exposure scenario. Results indicate that model output strongly depends on the organic-carbon partition coefficient, organic-carbon content, net infiltration, Darcy velocity, source-receptor distance, and first-order decay constant.

Carbaryl, 2,4-D, and Triclopyr Adsorption in Thatch-Soil Ecosystems

Thatch development in intensively managed turf sites may cause environmental concerns for greater sorption or leaching of applied chemicals in terrestrial ecosystems. To determine the adsorption potential of Carbaryl (1-Napthyl N-methylcarbamate), 2,4-D (2,4-dichloro-phenoxyacetic acid), and Triclopyr (3,5,6-trichloro-2-pyridinyloxyacetic acid) in turf ecosystems, composite thatch and underlying soil samples from three- and six-year-old stands of cool-season Southshore creeping bentgrass (Agrostis palustris Huds.) and warm-season Meyer zoysiagrass (Zoysia japonica Steud.) were collected. The samples were processed and analyzed for total organic carbon (COrg); extractable (CExt), humic (CHA) and fulvic acid (CFA); anthrone reactive nonhumic carbon (ARC) fractions; and CHA and CFA associated iron (Fe) contents. Pesticide adsorption capacity (K f ) and intensity (1/n), organic carbon partition coefficient (K OC ) and Gibbs free energy change (Δ G) were calculated for thatch materials and the underlying soils using a modified batch/flow technique. Both bentgrass (BT) and zoysiagrass thatch (ZT) contained a greater concentration of CExt, CFA, CHA, and ARC than the respective soils (BS and ZS). The CExt, CFA, CHA, and ARC concentration was higher in BT compared with ZT. The BT contained a greater concentration of bound Fe in both CFA and CHA fractions than in BS, whereas ZT had more bound Fe in CHA fraction than in ZS. On average, the BT had a greater concentration of bound Fe in CExt, CFA, and CHA fractions than in the ZT. Among the pesticides, Carbaryl had higher K f and 1/n values than 2,4-D and Triclopyr for both thatch and soil. Although the K OC and Δ G values of Carbaryl were higher in both BT and ZT than in the underlying soils, the K OC and Δ G values of 2,4-D were significantly higher in BS and ZS than in the overlying thatch materials. The 2,4-D and Triclopyr had higher leaching indices (LI) than Carbaryl for both BT and ZT materials than the respective soils. The Carbaryl, however, had a higher LI for soils than for thatch materials. Averaged across thatch materials and soils, COrg accounted for 96, 85, and 84% variations in Carbaryl, 2,4-D, and Triclopyr adsorption, respectively. Among the COrg fractions, lignin followed by CFA and CHA accounted for greater adsorption of pesticides, especially Carbaryl. The concentration of CHA and CFA bound Fe did not correlate with K f and 1/n values of pesticides.

Field study of pesticide leaching in a Himatangi sand (Manawatu) and a Kiripaka bouldery clay loam (Northland). 2. Simulation using LEACHM, HYDRUS-1D, GLEAMS, and SPASMO models

Abstract We applied 5 pesticides, along with bromide as a conservative tracer, at 2 field sites (Northland and Manawatu, North Island of New Zealand), and their distributions in the soil profile were monitored through soil core and suction cup sampling for 18 months. Field-measured values were simulated using LEACHM, HYDRUS-1D, GLEAMS, and SPASMO models and compared against the measured datasets at both these sites. All 4 models predicted the measured dataset for each of the compounds with reasonable degree of accuracy; however, some discrepancies were observed. Prediction by GLEAMS was by far the best, as indicated by the goodness of fit parameters. Pesticide leaching parameters, organic carbon partition coefficient (Koc), and degradation half-life (T1/2) were optimised using the PEST optimisation package. For diazinon and terbuthylazine the parameters obtained for each of the models were generally close to the literature range. However, Koc and T1/2 values for atrazine, hexazinone, and procymidone were found to be slightly lower than the available literature range data.

Herbicide sorption to concrete and asphalt

In order to accurately predict the fate and behaviour of herbicides following application to hard surfaces, an understanding of the processes involved is required. Previous studies have demonstrated that herbicides with a low soil organic carbon partition coefficient (K(oc)) are preferentially removed from concrete and asphalt surfaces. It is possible, therefore, that sorption may play a role in retaining herbicides to hard surfaces. This study provides an indication of the extent to which herbicides may sorb to concrete and asphalt, and examines the results in the light of other research to identify the significance of sorption in describing herbicide losses after application. There was measurable sorption of herbicides to concrete and asphalt although this was limited compared with that to soils, especially for concrete. An exponential relationship between experimentally derived values of K(p) and literature values of K(oc) for asphalt was established (r(2) = 0.931); there was a weaker relationship for concrete (r(2) = 0.606). The experimentally derived K(p) values could be used to enhance the prediction of herbicide loss to receiving waters following application. It is probable that the fate of herbicides applied to concrete and asphalt surfaces depends more heavily on physical processes than is the case in soils.

Mobility and Persistence of Petroleum Hydrocarbons in Peat Soils of Southeastern Mexico

Spilled crude petroleum from oil wells contains numerous hydrocarbons, some of which are toxic and threaten life. We have studied the mobility and persistence of hydrocarbons in waterlogged soils that contain large proportions of fermented organic matter (Histosols) and large concentrations of dissolved organic carbon (DOC) in the State of Tabasco, Mexico. We sampled soil and phreatic water at sites polluted by oil spills for several decades, as well as at sites that had only recently (few weeks) been polluted, and compared their hydrocarbon contents with those of unaffected sites in the same area. Samples were analyzed for 16 non-alkylated polyaromatic hydrocarbons (PAHs) and n-alkanes from nC9 to nC34. The spilled hydrocarbons had remained predominantly in the organic surface horizons of the soil where spillage occurred; there was little evidence of movement within the soil. The fraction of low molecular weight compounds was larger at sites of recent spills than where spills happened several decades ago. Nevertheless, sites of old spills still contained large concentrations of hydrocarbons, among which those of low molecular weight represented from 30 to 49% of total PAHs and from 50 to 84% of total n-alkanes, indicating that volatilization or microbial degradation is slow in these soils. In the peat horizons the measured organic carbon partition coefficients (K oc ) for the higher molecular weight PAHs were consistently smaller than those estimated by empirical equations by up to two orders of magnitude. The dissolved organic carbon of these peat soils seems to influence this behavior. At sites of old spills, partition coefficients for the PAHs were larger than at sites of recent spills.

Simulations of Fungicide Runoff Following Applications for Turfgrass Disease Control.

Computer simulations of fungicide loading in surface water runoff were conducted with fungicides commonly used in golf course fairways and lawns in Kentucky. For all fungicides, values for degradation half-life and organic carbon partition coefficient were obtained from published sources; other input parameters were selected to simulate conditions typical in local swards. Spray programs were tested using a 21-year period of weather data for Lexington, KY. Predicted amounts of fungicide in runoff were determined, and predicted fungicide concentrations in runoff (mg/liter) were compared with 50% lethal concentration (LC50) values for rainbow trout and Daphnia magna. All simulated chlorothalonil applications produced runoff with concentrations that greatly exceeded the LC50 values for both indicator species. For some applications, concentrations of azoxystrobin, iprodione, and pentachloronitrobenzene exceeded LC50 values of at least one indicator species. Under the conditions simulated, runoff concentrations of metalaxyl, propiconazole, thiophanate methyl, and triadimefon were well below LC50 values of the indicator species. Although actual amounts of fungicide loaded into runoff were relatively low, these simulations suggest that turfgrass applications of fungicides with high intrinsic toxicity to indicator species could pose a risk to populations of primary and secondary consumers in aquatic ecosystems.

Prediction of soil organic carbon partition coefficients by soil column liquid chromatography

To avoid the limitation of the widely used prediction methods of soil organic carbon partition coefficients ( K OC) from hydrophobic parameters, e.g., the n-octanol/water partition coefficients ( K OW) and the reversed phase high performance liquid chromatographic (RP-HPLC) retention factors, the soil column liquid chromatographic (SCLC) method was developed for K OC prediction. The real soils were used as the packing materials of RP-HPLC columns, and the correlations between the retention factors of organic compounds on soil columns ( k soil) and K OC measured by batch equilibrium method were studied. Good correlations were achieved between k soil and K OC for three types of soils with different properties. All the square of the correlation coefficients ( R 2) of the linear regression between log k soi and log K OC were higher than 0.89 with standard deviations of less than 0.21. In addition, the prediction of K OC from K OW and the RP-HPLC retention factors on cyanopropyl (CN) stationary phase ( k CN) was comparatively evaluated for the three types of soils. The results show that the prediction of K OC from k CN and K OW is only applicable to some specific types of soils. The results obtained in the present study proved that the SCLC method is appropriate for the K OC prediction for different types of soils, however the applicability of using hydrophobic parameters to predict K OC largely depends on the properties of soil concerned.

Ranking and classification of non-ionic organic pesticides for environmental distribution: a qsar approach

The partitioning of pesticides into different environmental compartments depends mainly on the physico-chemical properties of the studied chemical. To rank pesticides according to their distribution tendency in various media, we propose a combination of two multivariate approaches: Principal Component Analysis and Hierarchical Cluster Analysis. In such explorative methods we take into account physico-chemical properties like the organic carbon partition coefficient (K oc), the n-octanol/water partition coefficient (K ow), water solubility (S w), vapour pressure (VP) and Henry's law constant (H), being the more relevant to the determination of environmental partitioning. The 54 studied pesticides of various chemical categories are thus ranked in 4 a priori classes according to their environmental behaviour (sorbed, soluble, volatile and non-volatile/medium class) and finally assigned to the defined four classes by different classification methods (CART, K-NN, RDA) using theoretical molecular descriptors. This QSPR approach allows a rapid indication of the environmental distribution of pesticides starting only from their molecular structure.

Transport of Atrazine and Metribuzin in Three Soils of the Humid Pampas of Argentina1

Widespread groundwater contamination has prompted studies on the fate and transport of solute through soil. Large quantities of atrazine and metribuzin are applied annually in the Humid Pampas of Argentina, creating the need to study the fate of these herbicides in soils of the region. The objective of this work was to study the vertical transport of atrazine and metribuzin in packed soil columns for three loam soils representative of the Humid Pampas of Argentina. Bromide was used as a nonreactive tracer. The convection dispersion equation was fitted to chemical breakthrough data to obtain a parameter characterizing chemical transport. Bromide breakthrough curves (BTCs) were similar among soils. BTCs for atrazine and metribuzin revealed significant interaction among soils and herbicides. The average values for the organic carbon (OC) partition coefficients (Koc) derived from column flow experiments were 119 and 48 ml/g for atrazine and metribuzin, respectively. Metribuzin in leachate was 97.3% of the tot...

Sorption of the Herbicides Isoxaben and Oryzalin to Soils and Container Media

Abstract The herbicides isoxaben (Gallery) and oryzalin (Surflan) are commonly used to prevent weed emergence around ornamental plants in nurseries and landscapes. Saturated batch experiments were conducted in the laboratory to generate equilibrium sorption isotherms for these herbicides in soils and container media. Soils consisted of [A] a sandy loam containing 1.24% organic matter (by wt) after sieving, and [B] a silt loam containing 1.36% organic matter (by wt) after sieving. Container media obtained from nurseries consisted of [C] a mixture of 50% peat and 50% sand (by vol) containing 18.5% organic matter (by wt) after sieving, and [D] a mixture of 70% bark, 25% peat, and 5% fine stone (by vol) containing 34.9% organic matter (by wt) after sieving. Sorption isotherms were generated from these data. Based on sorption coefficients (Kd, Kf), container mix D sorbed both herbicides to a much greater extent than did the other substrates. Oryzalin sorption was greater than isoxaben sorption in each soil and container mix. For oryzalin, organic carbon partition coefficient (Koc) values in the soils were greater than those in the container mixes. Compositional differences in organic matter between soils and container media may account for differences in herbicide sorption in these media.

Modeling Volatile Chemical Transport, Biodecay, and Emission to Indoor Air

AbstractA model is presented for estimating vapor concentrations in buildings because of volatilization from soil contaminated by non‐ aqueous phase liquids (NAPL) or from dissolved contaminants in ground water. The model considers source depletion, diffusive‐ dispersive transport of the contaminant of concern (COC) and of oxygen and oxygen‐limited COC biodecay. Diffusive‐advective transport through foundations and vapor losses caused by foundation cross‐flow are considered. Competitive oxygen use by various species is assumed to be proportional to the product of the average dissolved‐phase species concentration and a biopreference factor. Laboratory and field data indicate the biopreference factor to be proportional to the organic carbon partition coefficient for the fuel hydrocarbons studied. Predicted indoor air concentrations were sensitive to soil type and subbase permeability. Lower concentrations were predicted for buildings with shallow foundations caused by flushing of contaminants by cross‐flow. NAPL source depletion had a large impact on average exposure concentration. Barometric pumping had a minor effect on indoor air emissions for the conditions studied. Risk‐based soil cleanup levels were much lower when biodecay was considered because of the existence of a threshold source concentration below which no emissions occur. Computed cleanup levels at NAPL‐contaminated sites were strongly dependent on total petroleum hydrocarbon (TPH) content and COC soil concentration. The model was applied to two field sites with gasoline‐contaminated ground water. Confidence limits of predicted indoor air concentrations spanned approximately two orders of magnitude considering uncertainty in model parameters. Measured contaminant concentrations in indoor air were within model‐predicted confidence limits.

Influence of rhamnolipids and triton X-100 on the desorption of pesticides from soils.

A rhamnolipid biosurfactant mixture produced by P. aeruginosa UG2 and the surfactant Triton X-100 were tested for their effectiveness of enhancing the desorption of trifluralin, atrazine, and coumaphos from soils. Sorption of both surfactants by the soils was significant and adequately described by the Langmuir-type isotherm. Values of maximum sorption capacity (Qmax) and Langmuir constant (Klang) did not correlate with the amount of soil organic matter. Our results indicate that clay surfaces play an important role in the sorption of surfactants. When surfactant dosages were high enough to reach soil saturation and maintain an aqueous micellar phase, pesticide desorption was only enhanced. At dosages below soil saturation, surfactants sorbed onto soil, increasing its hydrophobicity and enhancing the sorption of the pesticides by a factor of 2. Similar values of water-soil partition coefficients (Ksol*) for aged and fresh added pesticides to soils indicate that the aging process used did not significantly after the capability of either surfactant to desorb the pesticides. A model able to estimate equilibrium distributions of organic compounds in soil-aqueous-micellar systems was tested against experimental results. The determined organic carbon partition coefficients, Koc values, indicate that, on a carbon normalized basis, sorbed Rh-mix is a much better sorbent of pesticides than TX-100 or soil organic matter. These results have significant implications on determining the effectiveness of surfactants to aid soil remediation technologies.

Effect of flood-induced chemical load on filtrate quality at bank filtration sites

Riparian municipal wells, that are located on riverbanks, are specifically designed to capture a portion of the river water through induced infiltration. Runoff from agricultural watersheds is found to carry enormous amounts of pesticides and nitrate. While the risk of contamination for a vast majority of sites with small-capacity vertical wells is low, potential exists for medium to large capacity collector wells to capture a fraction of the surface water contaminants during flood. Prior monitoring and current modeling results indicate that a small-capacity (peak pumpage 0.0315 m 3/s) vertical bank filtration well may not be affected by river water nitrate and atrazine even during flood periods. For a medium capacity (0.0875–0.175 m 3/s) hypothetical collector well at the same site, potential exists for a portion of the river water nitrate and atrazine to enter the well during flood periods. Various combinations of hydraulic conductivity of the riverbed or bank material were used. For nitrate, it was assumed either no denitrification occurred during the period of simulation or a half-life of 2 years. Equilibrium controlled sorption (organic carbon partition coefficient of 52 ml/g) and a half-life of between 7.5 and 15 weeks were considered for atrazine. Combinations of these parameters were used in various simulations. Peak concentrations of atrazine or nitrate in pumped water could vary from less than 1% to as high as 90% of that in the river. It was found that a combination of river stage, pumping rates, hydraulic properties of the riverbed and bank, and soil/pesticide properties could affect contaminant entry from river water to any of these wells. If the hydraulic conductivity of the bed and bank material were low, atrazine would not reach the pumping well with or without sorption and degradation. However, for moderately low permeable bank and bed materials, some atrazine from river water could enter a hypothetical collector well while pumping at 0.0875 m 3/s. It was interesting to note that doubling the pumpage of this collector well would bring in more ground water from the aquifer (with no atrazine) and thus have a lower concentration of atrazine in the filtrate. For highly conductive banks, it is possible to find some atrazine at a vertical well for a sustained pumpage rate of 0.0125 m 3/s if the effect of sorption is neglected. However, with equilibrium sorption, the concentration would be below the detection limit. On the other hand, if a collector well of capacity 0.0875 m 3/s is used at the place of the vertical well with highly conductive banks, atrazine concentration in the filtrate would be about 80% of river water even assuming equilibrium sorption and a half-life of 7.5 weeks. Remediation of river water contamination of the aquifer using ‘scavenger’ wells between the river and the pumping well(s) was not a feasible option due to the contact of the aquifer with a highly conductive bank at the site. However, moving the existing pumping well(s) 100 m upstream would have negligible impact from the bank-stored water.

A subregional-scale method to assess aquifer vulnerability to pesticides.

A method to predict aquifer vulnerability to pesticide contamination at the subregional scale was developed. The assessment method was designed to incorporate relevant hydrologic and pesticide-transport information and to use generally available data. The method assumes steady-state advection of pesticides in the vadose zone, including sorption and biological decay. The solution is presented as a vulnerability index (VI) that increases as the aquifer vulnerability increases. The hydrologic input data for the VI model are obtained from the soil survey geographic database. Pesticides were grouped into three leachability classes using a leachability ratio (half-life divided by organic carbon partition coefficient). Pesticide transformation is assumed to occur in the surface layer. The influence of vertical transport in the remainder of the vadose zone has been incorporated by applying a multiplying factor to the VI that varies with depth to ground water. Hydraulic conductivity is used as a surrogate for soil-water velocity for practical purposes. The performance of the VI model is evaluated using ground water data from Weld County, Colorado. The model is demonstrated to be successful at predicting relative vulnerability, defined as the magnitude of pesticide concentration and percent of wells in a unit that exhibit a pesticide detection. Areas of low, medium, and high vulnerability are assigned. Results indicate that the vulnerability classifications are most dependent on the leachability ratio, hydraulic conductivity, and organic carbon content.

Carbendazim sorption–desorption in Vietnamese soils

Four Vietnamese soils (denoted AG, CT, ST and TG) which differed with respect to pH (pH 2.9–5.4), clay (17–50%) and organic matter (0.3–9.8%) content, were selected for sorption and desorption studies of carbendazim using the batch equilibration technique. Sorption increased with increasing organic carbon (OC) and clay content. K d values for carbendazim sorption on AG, CT, ST, TG soils at initial concentration of 20 μg/g were 12.5, 127, 8.1 and 9.6 ml/g, respectively. The OC partition coefficients ( K oc) for AG, CT, ST and TG were 1140, 1300, 2700 and 960 ml/g, respectively. Carbendazim was strongly sorbed and the binding was less reversible in the acid sulfate soil (CT), than in the other soils. The CT soil had both the highest OC content (9.8%) and the highest clay content (49.8%). The influence of pH on carbendazim sorption was studied in the ST and CT soils. Sorption of carbendazim by the sandy ST soil (OC 0.3%; clay content 26.3%) increased as the pH decreased, while sorption of carbendazim by the CT soil decreased as pH decreased.

Screening of pesticides for environmental partitioning tendency

The partitioning tendency of chemicals, in this study pesticides in particular, into different environmental compartments depends mainly on the concurrent relevance of the physico-chemical properties of the chemical itself. To rank the pesticides according to their distribution tendencies in the different environmental compartments we propose a multivariate approach: the combination, by principal component analysis, of those physico-chemical properties like organic carbon partition coefficient ( K oc), n-octanol/water partition coefficient ( K ow), water solubility ( S w), vapour pressure and Henry's law constant ( H) that are more relevant to the determination of environmental partitioning. The resultant macrovariables, the PC1 and PC2 scores here named leaching index (LIN) and volatality index (VIN), are proposed as preliminary environmental partitioning indexes in different media. These two indexes are modeled by theoretical molecular descriptors with satisfactory predictive power. Such an approach allows a rapid pre-determination and screening of the environmental distribution of pesticides starting only from the molecular structure of the pesticide, without any a priori knowledge of the physico-chemical properties.

A study of atrazine transport through a soil profile on the Gnangara Mound, Western Australia, using LEACHP and Monte Carlo techniques

A preliminary risk assessment of atrazine leaching at a strawberry farm on the Gnangara Mound, Western Australia, was performed using LEACHP in conjunction with Monte Carlo techniques. Monte Carlo techniques were also used to implement a sensitivity analysis of LEACHP to determine influential parameters. As an extension to the sensitivity analysis, a model comparison was performed with similar models, in order to quantify differences arising from the use of different models. The Monte Carlo risk assessment involved simultaneous variation of LEACHP's soil parameters and selected chemical parameters to assess the risk of atrazine leaching past the root-zone at the experimental site. This Monte Carlo risk assessment predicted that of the 100 mg/m2 of atrazine applied at the experimental site, the mean cumulative mass per unit area of atrazine to leach below the root-zone after 263 days was 49 mg/m2, with a standard deviation of 4.1 and a range from 37 to 62 mg/m2. The Monte Carlo sensitivity analysis of LEACHP tested selected soil and chemical parameters, one at a time, to determine the influence of each parameter on the models predictions of cumulative mass per unit area of atrazine to leach below the root-zone. The chemical parameters organic carbon partition coefficient (Koc) and degradation rate constants were found to be more influential than the soil parameters, with Koc being the most influential parameter. The most influential of the soil parameters was organic carbon content, followed by bulk density. The model comparison involved 4 separate simulations: LEACHP using Richards' equation (LEACHP (Richards)), LEACHP using Addiscott's model of water and solute movement (LEACHP (Addiscott), MACRODB, and GLEAMS. The results from this exercise showed that LEACHP (Richards) and MACRODB made similar predictions for the cumulative mass per unit area of atrazine to leach below the root-zone (49 and 50 mg/m2, respectively), while GLEAMS and LEACHP (Addiscott) made significantly lower predictions (35 and 38 mg/m2, respectively). The results from this study indicate there is a significant risk of atrazine leaching at the experimental site; however, more detailed field measurements and further modelling would enable more accurate predictions.

Estimating Dissolved Organic Carbon Partition Coefficients for Nonionic Organic Chemicals

A literature search was performed for dissolved organic carbon/water partition coefficients for nonionic organic chemicals (KDOC), and KDOC data were taken from more than 70 references. The KDOC data were evaluated as a function of the 1-octanol/water partition coefficients (KOW). A predictive relationship of KDOC = 0.08KOW with 95% confidence limits of a factor of 20 in either direction was developed using KDOC data based upon naturally occurring dissolved organic carbon. Inclusion of KDOC data for Aldrich humic acid, a reagent-grade organic carbon, resulted in a slightly different predictive relationship of KDOC = 0.11KOW with 95% confidence limits of a factor of 14 in either direction. The large uncertainties in these relationships are, in part, caused by the variability in structure and composition of dissolved organic carbon (DOC) in sediments, soils, and surface waters. This variability is not accounted for by the hydrophobicity parameter. For individual chemicals, ranges in KDOC values approaching 2 orders of magnitude were observed among investigations using Aldrich humic acid as the DOC. These large ranges of KDOC values suggest that measurement techniques are also, in part, responsible for the large uncertainties in these relationships.

Estimating the Organic Carbon Partition Coefficient and Its Variability for Hydrophobic Chemicals

Numerous correlations have been developed between the organic carbon/water partition coefficient KOC and various molecular properties and descriptors, but most notably the octanol/water partition coefficient KOW and water solubility. From an analysis of the theory underlying in this partitioning and an examination of the existing database, it is suggested that the preferred approach is to correlate the quantity log(KOC/KOW), which is essentially the ratio of activity coefficients in octanol and organic carbon, with a molecular property or with log KOW rather than log KOC with log KOW. Such an approach is presented for estimating KOC for hydrophobic chemicals, including an expression of uncertainty limits, based on a correlation derived between log(KOC/KOW) and log KOW. In its simplest form the correlation is that KOC = 0.35KOW subject to variation by a factor of 2.5 in either direction. It is suggested that new experimental data be assessed for consistency and achievement of true equilibrium by comparison...