Report No. SRC-TR-99-020

Evaluating Potential POP/PBT Compounds

for Environmental Persistence

Prepared by:

Dallas Aronson

Philip H. Howard

Environmental Science Center

Syracuse Research Corporation

6225 Running Ridge Road

North Syracuse, NY 13212-2509

November 1, 1999

TABLE OF CONTENTS

1. Introduction

2. Sources of Information

3. Persistence in Soil, Water and Sediment

3.1. Biodegradation

3.1.1. Definition of Biodegradation

3.1.2. Kinetics of Biodegradation

3.1.3. Process for Evaluation of Experimental Biodegradation Data Versus the Criteria

3.1.3.1. Overview of the Evaluation Procedure

3.1.3.2. Realistic Biodegradation Studies

3.1.3.2.1. Field Studies

3.1.3.2.2. Grab Sample Studies

3.1.3.3. Screening Studies

3.1.3.4. Pure Culture Studies

3.1.4. Prediction of Biodegradation Using Estimation Programs

3.1.4.1. Review of Biodegradation Estimation Programs

3.1.4.2. Review of Biodegradation Kinetics from Estimation Programs

3.1.4.3. Recommended Use of Biodegradation Estimates Versus POP/PBT Criteria

3.2. Abiotic Degradation Processes

3.2.1. Hydrolysis

3.2.1.1. Hydrolysis Kinetics

3.2.1.2. Interconversion of Hydrolysis Rates Between Different Media

3.2.1.3. Hydrolysis Estimation Programs

3.2.2. Photolysis

3.2.2.1. Direct Photolysis

3.2.2.2. Indirect Photolysis

3.2.2.3. Photolysis in Aquatic Environments

3.2.2.4. Soil Photolysis

4. Persistence in the Atmosphere

4.1. Direct Photolysis

4.2. Indirect Photolysis/Photooxidation

5. Summary

6. References

1. Introduction

Both international and national level efforts are currently being applied to identify environmental contaminants where persistence needs to be determined. These types of chemicals are referred to as Persistent Organic Pollutants (POP) or Persistent, Bioaccumulative, and Toxic (PBT or PTB) compounds. Compounds with a slow degradation rate in the environment, i.e. are resistant to biodegradation, hydrolysis, and photolysis processes, are classified as persistent and have often been considered as potential environmental problems. A more exacting approach recognizes that a compound released to the environment has a tendency to accumulate in one medium more than others and that partitioning, transport, and transformation rates of any particular compound will differ in each medium. Degradation processes in the dominant medium (where the compound is preferentially accumulated) are expected to have more effect on overall persistence of the measured compound than degradation processes in the other media (Webster et al., 1998). Optimally, to describe persistence of a compound in the environment, information on the amount released to each medium, how it partitions among the environmental media, its potential for transport, and a degradation half-life in each medium are required.

Persistence half-life criteria have been proposed by several official organizations/ governments worldwide and have been reviewed by Rodan et al. (1999). These are summarized as follows:

• atmospheric half-lives of 2 to 5 days
 
• aquatic half-lives of 2 to 6 months
 
• soil half-lives ranging from 6 to 12 months
 
• sediment half-lives ranging from 6 to 12 months
 

Due to the inherent variability in experimentally-measured half-lives, these persistence criteria should not be considered as specific cutoff values but as guidelines for making scientifically-valid judgements regarding the ability of a compound to persist or not in an environmental medium (Klecka, 1999).

In the following pages, procedures are discussed for determining if a chemical should be classified as persistent using the above criteria. Persistence in air is considered separately from persistence in other environmental media (water, soil, sediment). However, the basic approach is similar for each environmental medium and is outlined in Figure 1. Initially, the available experimental information is collected. In soil, water and sediment, this includes kinetic information on both biotic (biodegradation) and abiotic (photolysis and hydrolysis) processes as summarized in the following equation:

Figure 1. Overview of recommended procedure for determining persistence.

2. Sources of Information

Sources of environmental fate information can be accessed via electronic databases as well as by on-line searches as summarized in Table 1. Many of the sources listed below contain redundant information; Hazardous Substances Databank (HSDB), Environmental Fate Data Base (EFDB) and Chemical Abstracts, in combination, are suggested as the most generally productive and convenient sources of information. If environmental fate information is not available for a particular chemical, substructure searches may be necessary to identify structurally-similar chemicals. Once the available data is collected, compounds can be separated into those with experimental degradation (hydrolysis, photolysis and/or biodegradation) data and those without.

Table 1. Databases available for environmental fate information

3. Persistence in Soil, Water and Sediment

Degradation in soil, water, and sediment environments is mainly due to biodegradation, hydrolysis, and photolysis. Compounds which are not susceptible to these processes will be persistent in the environment. Information is presented in this section to assist in evaluating the experimental data for inclusion and methods are described to determine half-lives/first-order rate constants from available data for those compounds with available experimental data. Methods for estimating the degradation potential of a compound which does not have experimental data are given in section 3.1.4. The emphasis will be to obtain first-order rate constants or half-lives for each degradation process from the available data for each specified medium. Half-life and first-order rate constant values are interconvertible using the following equation:

3.1. Biodegradation

Biodegradation is expected to be the major mechanism of loss for most chemicals released into the environment. Common terms and biodegradation kinetics are defined, the different types of biodegradation studies outlined and procedures for the evaluation of data in order to obtain half-life values in each medium are given below.

3.1.1. Definition of Biodegradation

Biodegradation is defined as the biologically-mediated structural alteration of an organic chemical, the parent compound, resulting in the formation of metabolites. Both the extent of degradation and the rate of this process are important. These are dependent on numerous factors including the chemical structure itself, the concentration of chemical (high versus low concentration), the density of the microbial population, the concentration of nutrients present, temperature, oxygen content, whether other compounds are present (can either enhance or decrease the rate of reaction), etc.

A single alteration of the parent structure is sufficient to label a compound as "having undergone biodegradation". This is called primary biodegradation and the remaining parent chemical is most often measured by GC or HPLC. The alteration may or may not result in the loss of toxicity, persistence, or bioaccumulative characteristics, e.g. metabolites may have POP/PBT characteristics as well. Other studies may measure biodegradation as mineralization which is the complete degradation of a compound, in most cases, to carbon dioxide (both aerobic and anaerobic conditions), methane (anaerobic conditions) and water. This process is also referred to as ultimate biodegradation. Ultimate biodegradation studies often use a radiolabeled compound and measure O₂ consumption, dissolved organic carbon (DOC) loss, and/or CO₂ evolution from the study system. As expected, mineralization half-lives can be considerably longer than those reported for the primary degradation of the parent compound. For example, a recent report collected published rate constants for primary and ultimate biodegradation of benzene (median value 0.096 and 0.0013 day^-1, respectively) and toluene (median value of 0.2 and 0.009 day^-1, respectively) (Aronson et al., 1999). As currently written, the persistence criteria do not specify whether the measured endpoint is primary degradation or mineralization but it is assumed that primary degradation is intended. Primary degradation of most POP/PBT compounds is expected to be sufficient to reduce their ability to be persistent, bioaccumulative and/or toxic.

3.1.2. Kinetics of Biodegradation

A biodegradation curve showing a lag phase, a growth phase and a stationary phase is given in Figure 2.

Figure 2. Representative biodegradation curve.

When a compound is initially released to the environment, an acclimation period, also called the lag phase, is often observed where biodegradation does not proceed or proceeds at a rate which is considerably less than that measured during the following growth phase. The presence of a lag phase may be due to several factors: the development of new enzyme pathways, the time required for the induction of enzyme(s) in a reaction pathway or to significantly increase the microbial population capable of degrading the added compound, or the concentration of the compound may be toxic to microorganisms in the environment and, until it is partially diluted, biodegradation does not proceed. The actual length of acclimation required for a particular compound in a chosen environment is not predictable as it can vary with location, concentration of the compound, oxygen concentration, temperature, microbial populations, etc. (Alexander, 1994). After the initial acclimation period, a lag phase is not seen again as long as further additions of the compound are made within a reasonable time (from weeks to over a year) (Alexander, 1994).

Once biodegradation begins it proceeds at a particular rate and the kinetics of this process can be mathematically defined. Although other rate equations [most commonly: zero-order, three-half-order, Monod kinetics (Scow et al., 1986)] may more suitably define the biodegradation of a compound under some circumstances, first-order kinetics are commonly used as an approximation because:

• they are easy to calculate even when limited data are available,
• they allow for standardization with previously published data and
• many environmental model programs require the input of first-order rate constants.

First-order rate kinetics are most appropriate for use when the substrate concentration is low (too low to support microbial growth, estimated as about 1 ppm) as is frequently the case with a compound released to the environment and biomass concentrations do not change significantly.

The first-order biodegradation rate constant, k_Bio, is mathematically defined by the differential:

where S is substrate concentration and t is time. If a rate constant is not specifically stated, it is still often possible to extract both time, concentration, and possibly biomass data suitable for the rough estimation of a first-order rate constant using equation 3, above. While first-order kinetics are widely accepted as a simplification of more complex biodegradation kinetics for the reasons given above, first-order kinetics do not account for increased or decreased degradation rates with decreasing concentration which are commonly seen in biodegradation studies where compounds are released into the environment. In addition, a lag phase, if reported, is not included in the calculation; thus it is possible for a compound with a long acclimation period to be considered biodegradable even if a considerable lag phase is present.

When standardization and prediction are important, division of the rate constant by the population (or concentration) of microbes present in the measured environment has been recommended in order to normalize for differences in biomass between locations (surface soil versus subsurface soil, oligotrophic lake versus eutrophic lake) (Larson et al., 1999; Paris et al. 1981). In principle, this should be done only if the microbial biomass which is actively degrading the added compound is measured in the study; however, in practice, biomass values are commonly based solely on direct bacterial counts or ATP concentrations which may overestimate the responsible bacterial population. Biodegradation studies which do not include this information would need to be standardized using a generic biomass concentration. As active biomass concentrations for a single environmental medium can vary with location, nutrient status and time, prediction of this value(s) could introduce even further variability into the biodegradation kinetics of a given compound.

Environmental conditions may significantly affect the rate of biodegradation reported for a compound in the environment. In most cases, organic compounds are more readily degraded in the presence of oxygen (aerobic conditions) than when oxygen is not present (anaerobic or anoxic conditions). The presence of other compounds which can be more readily utilized than the studied compound in the same environment or may be toxic to the microbial population may reduce the rate of degradation of the studied compound. The rate of biodegradation of a water-insoluble compound released to a soil with a high organic content or a eutrophic lake may be affected by the bioavailability of the released compound. Over time, a compound can become sequestered in organic material or soil aggregates (Alexander, 1994) and may be only slowly available to microorganisms for biodegradation through the processes of desorption from the particle surface or diffusion (Shelton et al., 1998). If permanent, covalent binding to the soil occurs, the compound may not be available at all for microbial degradation. The importance of strong adsorption of compounds to soil, sediment or particulates can be indicated from physical/chemical properties, e.g. high log K_ow, low water solubility.

3.1.3. Process for Evaluation of Experimental Biodegradation Data Versus the Criteria

This section reviews the different biodegradation study types available in the literature, examines their strengths and weaknesses and provides methods for determining half-lives from each. These half-lives can then be compared to the POP/PBT criteria. Biodegradation tests include pure culture studies, enrichment studies using a preacclimated inoculum, screening studies, grab sample studies and field studies (Howard and Banerjee, 1984; Howard et al., 1987). Non-standardized experimental methods and natural variation due to different samples and sample types are mainly responsible for the considerable variation in the rates of reaction reported in the literature. Thus the choice of a single "best" value is often difficult for compounds with large quantities of biodegradation data; other options in this case include providing the data as a range, giving the median or mean value or providing a frequency distribution of the available data. At the other extreme are compounds for which there are currently no available data. In this case, estimation of biodegradation using QSBR or SBR, or use of structurally-similar (analogous) compounds with experimental data may be possible.

3.1.3.1. Overview of the Evaluation Procedure

The more environmentally realistic tests are given higher priority during the evaluation phase. Field studies are evaluated first as they are potentially the most environmentally-relevant data available. Grab sample studies are given second priority, ready biodegradability screening tests 3^rd priority, and inherent biodegradability screening tests 4^th priority. As there is still debate as to whether exceeding the published criteria in only one medium (soil, water, or sediment) is sufficient for designating a compound as persistent in the environment, whenever possible first-order rate constants should be located or predicted for each environmental medium. If possible, studies measuring mineralization versus primary biodegradation should be separated and evaluated within each study type and medium separately.

If sufficient data are available to give a range of first-order rate constants for each environmental medium, then this should be used along with a measure of central tendency such as the mean or median value. For example, aerobic studies were collected for bis(2-ethylhexyl)phthalate in soil/sediment (Aronson et al., 1999). Half-life values for primary degradation ranged from 8 to 630 days (n=26; median value=73 days). A frequency histogram of this data is presented in Figure 3. If the data span the persistence criteria, as they do in this example, it may be necessary to look more closely at the available data and choose values for that compound which are most representative of the conditions found in the environment of interest.

Figure 3. Frequency histogram for the published primary biodegradation rate constant values of bis(2-ethylhexyl)phthalate in soil and sediment.

3.1.3.2. Realistic Biodegradation Studies

Both field and grab sample studies provide data which is environmentally relevant. As they can give the best indication of biodegradation in the natural environment, these studies are given greater weighting than screening tests during the evaluation process.

3.1.3.2.1. Field Studies

As field studies directly measure the loss of a compound in the actual environment [usually as a dissipation half-life (DT₅₀)], they are the most realistic source of information for the persistence of a particular compound. However, mechanisms of loss include both degradation and transport processes and distinguishing between these can sometimes be difficult. Transport mechanisms include leaching and volatilization from soil, adsorption to suspended solids/sediment, and dilution and volatilization from water. These transport processes do not affect the structure of the original compound.

Techniques which are used in field studies to account for non-degradation loss include mass balance modeling and the use of tracers to account for dispersion, sorption and volatilization processes. Mass balance modeling requires that the system is at equilibrium in the measured environment which may not be the case, particularly for persistent compounds which are often water insoluble and adsorptive. This method has been used in groundwater studies (Borden et al., 1997); in this case, rows of monitoring wells were set up perpendicular to the flow and used to estimate the mass flux of contaminant crossing each row. First-order rate constants were estimated for the contaminant based on changes in mass flux versus distance. Studies using tracers are also reported. Considerations for selecting an appropriate tracer are based on the biodegradability (should be non-degradable under the study conditions) and sorption/volatility of the tracer compound (should be similar to that of the studied compound) in that system. A study by Kim et al. (1995) measured the loss of toluene in a stream in situ using tracers to measure transport processes. In this study, the importance of transport versus degradation as a loss mechanism changed with the season. Biodegradation was the dominant process during the warmer months; however, when temperatures dropped in winter, volatilization became a significant loss mechanism (Kim et al., 1995). Field studies in soil often measure compound residues at different depths or use soil cores to measure losses due to leaching (Lichtenstein and Schulz, 1959; Stewart and Chisholm, 1971). Physical properties such as the Henry's Law constant, log K_ow, and water solubility can be used to estimate K_oc and volatilization effects, when this information is not available, in order to predict whether loss due to transport could be significant.

Recommended Analysis of Field Data Versus POP/PBT Criteria

A diagram outlining the decision process for using field studies to obtain overall persistence half-lives for comparison with POP/PBT criteria is given in Figure 4. If field data are available with adequate controls for transport effects (leaching, volatilization, dilution) then this information is given the highest priority. First-order rate constants for loss in a field study can be directly converted to overall persistence half-lives using equation 2, after correcting for transport loss, as degradation is due to both biotic and abiotic processes.

Figure 4. Use of field studies to obtain overall persistence half-lives.

Differences in environmental parameters such as soil type, pH, organic carbon content, temperature, nutrient status, time of year and moisture between different locations can result in considerable variation among degradation rates derived from field studies. A measured DT₅₀ in the field longer than 6 months in soil or 2 months in water (POP/PBT criteria) strongly indicates a compound which fulfills the persistence criteria for POP/PBT compounds. However, a DT₅₀ under 2 or 6 months reported in a field study must be carefully examined for loss due to transport processes.

3.1.3.2.2. Grab Sample Studies

If field data are not available or only available in some environmental media or adequate controls were not run, then grab sample data should be used. Grab sample tests (also called "batch tests") are used to measure the loss of a compound in an environmental sample of water, soil, or sediment, or some combination of these (e.g. a water/sediment microcosm) in the laboratory. The most commonly encountered tests are river die-away and soil metabolism tests. In these tests, the environmental medium is collected and treated with a compound and the loss of the added compound or formation of metabolites is monitored. Greater control of environmental parameters (e.g. substrate concentration, temperature, moisture content, presence of other compounds) is possible in these tests. Because biodegradation can be considered separately from other degradation and transport processes in grab sample tests, a more accurate measurement of the biodegradation of the monitored compound within that system is possible in comparison to that measured in a field test. In addition, metabolites can be identified by common laboratory procedures definitively showing the transformation of the studied compound by a microbial population.

Grab sample tests where surfactants are added to increase bioavailability or a specialized microbial population or enrichment culture is added to increase rates of degradation should be evaluated carefully. In some cases, the surfactant can be used as a carbon source inhibiting biodegradation rates of the insoluble compound or it may be toxic to the microbial population. Acclimated cultures can give data which might be relevant to biodegradation under conditions where a compound is continuously released, but cannot be applied to situations where the compound is intermittently released. Many of these types of studies are more appropriately used to evaluate the bioremediation potential of a compound rather than to determine its intrinsic biodegradability.

Intermedia Conversion Factors for Grab Sample Studies

Conversion factors are available to predict biodegradation in different environmental media. Federle et al. (1997) measured the aerobic mineralization rates in river water amended with sewage and sludge-amended soil for nine compounds, mainly long-chain surfactants, and obtained a mean scaling factor of 1.1; scaling factors ranged from 0.14 to 3.24. In addition to the variation in the scaling factors, the correlation between the different rates in soil and water was not strong (r=0.53). The scaling factor reported by Federle et al. (1997) is very similar to that derived by Boethling et al. (1995). These authors measured the mean ratio of aerobic soil half-lives to aerobic water half-lives (looking at primary biodegradation and mineralization separately) which had been published for 20 chemicals (low K_ow range of -0.38 to 7.6) and obtained a scaling factor of roughly one, again indicating that half-lives reported in soil are reasonably similar to those in water. Scaling factors ranged from 0.13 to 5.22 for the primary biodegradation data and from 0.049 to 5.66 for the mineralization data.

The published POP/PBT criteria do not distinguish between aerobic and anaerobic conditions although it is assumed that aerobic conditions are implied. Aerobic conditions are expected to predominate in drained soil, surficial sediments and fresh and marine waters, while anaerobic conditions may be found in sediment, the hypolimnia of stratified lakes, waterlogged wetland soils, groundwater or waterlogged soils. Studies reporting the degradation of a compound under anaerobic conditions may be used to estimate degradation rates under aerobic conditions (Boethling et al., 1995). Published half-lives in surface soil were compared to half-lives in flooded soils for 25 different compounds. Rates of degradation in aerobic soil were determined to be approximately 3 to 4 times faster than that found in flooded soil (presumed anaerobic) whether primary degradation or mineralization was the measured endpoint. However, judgement should be exercised when using this scaling factor as compounds containing either multiple halogen atoms or nitroaromatics can degrade much more rapidly under anaerobic conditions when compared to aerobic conditions.

Boethling et al. (1995) compared biodegradation half-lives in marine water versus fresh water for 11 compounds. While it is generally considered that rates of degradation are slower in salt water (Bartholomew and Pfaender, 1983), these authors reported that a scaling factor can not be predicted based on this alone.

Recommended Analysis of Grab Sample Data Versus POP/PBT Criteria

If field studies are not available and grab sample data are found with adequate controls for volatilization (if compound is volatile) then grab sample data can be used to calculate first-order rate constants for use in the determination of overall persistence. If appropriate controls are present (sample incubated in dark to prevent photolytic degradation, sterile controls present), the degradation mechanisms can be separated and a first-order biodegradation rate constant calculated. First-order biodegradation rate constants in a grab sample study are summed with first-order hydrolysis and photolysis rate constants (see sections 3.2.1. and 3.2.2. for further information) to obtain an overall persistence rate constant. If the studied compound is susceptible to hydrolysis and the hydrolysis rate constant was determined for a specific pH value then the overall persistence rate constant is relevant only at that specific pH. If possible, several pH values should be used giving a range of k_Hyd values. Once an overall rate constant or rate constant range (using the high and low value for each process) for degradation in each medium is calculated, the persistence half-life or half-life range can be determined using equation 2 for comparison to POP/PBT criteria. If separation of the different degradation mechanisms is not possible based on the controls used, but non-degradation loss can be accounted for, an overall persistence first-order rate constant can be calculated after correcting for any non-degradation loss. This value can then be converted to a half-life and compared to the POP/PBT criteria. A diagram outlining the decision process for using grab sample studies to obtain overall persistence half-lives for comparison with POP/PBT criteria is given in Figure 5.

Rates of transformation obtained in the laboratory are generally believed to be more rapid than those reported in the field (Aronson and Howard, 1997; Aronson et al., 1999), at least for compounds which are not readily degraded (Barker et al., 1987). Thus, rates reported from grab sample studies may be higher than the actual transformation rate which would be seen in the environment. This may be an oversimplification as a recent paper by Di et al. (1998), studying the biodegradation of 8 pesticides in both laboratory and field conditions, reported that a consistent relationship between degradation rates in the two study types was not seen. For many compounds though, laboratory results may represent a best case scenario providing conditions suitable for obtaining maximum rates of degradation in the environment. Therefore, a compound which exceeds the persistence criteria due to biodegradation alone in a grab sample study is likely to be "persistent" according to the POP/PBT criteria unless abiotic degradation is significant.

 

3.1.3.3. Screening Studies

When data from field or grab sample studies are not available, screening studies can be conditionally used to estimate the potential for biodegradation in the environment. Screening studies measure the biodegradation of a compound using a small quantity of inoculum (sewage, activated sludge, surface water, soil, or sediment) added to a defined, usually mineral salts, medium. These studies were developed to identify compounds which would rapidly undergo ultimate biodegradation in the environment and are the most commonly reported biodegradation tests found in the literature. However, these tests do not mimic conditions in the environment for the following reasons:

• the test chemical concentration is very high as is necessitated by the analytical method (unless ¹⁴C is used),
• the microbial population can be quite high compared to some environmental media in order to simulate biodegradation during sewage treatment and
• other carbon sources are limited to facilitate analysis of biodegradation of the tested chemical
• inorganic nutrients are provided so that, theoretically, none is in short supply.
 

Thus, the calculation of half-lives from this data in order to predict biodegradation in the environment, although it is done, is questionable.

The OECD (Organization of Economic Cooperation and Development) divides screening test methods into ready and inherent biodegradability tests. Ready biodegradability tests (RBT) [modified OECD screening test, CO₂ evolution test, manometric respirometry test, DOC die-away test, closed bottle test, and the MITI(I) test] measure mineralization over a 28-day period using a low concentration of a sewage or activated sludge as an inoculum and a high concentration of the test compound. The use of an acclimated inoculum or a longer incubation period results in the classification of the test as an inherent biodegradability test (IBT, see below) (Strujis and van den Berg, 1995). A compound is considered to be readily biodegradable under RBT conditions if:

• 60-70% biodegradation has occurred in 28 days (including the lag phase) and
• the time between the start of mineralization (defined as 10% of the theoretical) and the plateau does not exceed 10 days (the "10-day window") (OECD, 1981).
 

Inherent biodegradability tests [Zahn-Wellens test, SCAS test, MITI(II) test] are run with high microbial population densities, again using a sewage or activated sludge inoculum. As with an RBT, mineralization is the measured endpoint and similar criteria for degradability are used. A compound which "fails" the inherent biodegradability test is considered to be non-biodegradable; this, however, does not necessarily mean that they are so slowly degraded that they would meet POP/PBT criteria.

The stringent conditions applied in a screening test (i.e., mineralization as the measured endpoint, incorporation of the lag phase into the biodegradation time frame) indicate that the results can be useful, if the chemical passes, in predicting whether biodegradation will or will not occur in the environment. The high concentration of compound added to many of these tests, however, means that results from the test can be affected by the solubility or toxicity of the added compound often resulting in false negative results (Federle et al., 1997). This is less of a concern in grab sample and field studies which use lower substrate concentrations.

Published Extrapolations of Screening Test Results to Grab Sample Tests

Kinetic data from screening tests have been used to predict biodegradation rates in the environment (Federle et al., 1997; Struijs and van den Berg, 1995). Federle et al. (1997) measured mineralization rates in the modified Sturm test (Ready CO₂ test) and in die-away tests using river water and sludge-amended soil for nine diverse compounds. Mineralization rates in the environmental media were not correlated to rates reported in the screening test (r=0.04, 0.12 in river water and soil, respectively), despite the addition of a microbial inoculum from the screening test to "standardize" the microbial population in the other environmental media. The rate of biodegradation in the screening test was, however, strongly correlated to the solubility and the log K_ow of the added compound. The authors concluded that while prediction of biodegradation kinetics in environmental media was not possible from screening test results, the ready tests are a very useful indicator of the ability of a compound to undergo complete biodegradation.

Struijs and van den Berg, 1995, attempted to use results reported in standardized OECD biodegradation tests to assign first-order biodegradation rate constants for degradation in water and soil. This was done by referencing the first-order rate constant to the expected population density in the studied medium. A compound which passes the RBT criteria (including the 10-day window) has a lower limit of 0.14 day^-1. First-order rate constants for water and soil are assigned as follows:

• surface waters - a rate constant of 0.14 day^-1 is given to compounds passing the RBT criteria. This is based on the assumption that the microbial density in surface water is similar to that found in the RBT tests.


• soils - a rate constant of 14 day^-1 (a half-life of less than one day) is used to describe biodegradation in the soil pore water for compounds which pass the RBT criteria. This is based on the assumption that the microbial density in soil is 100 times greater than that found in the RBT test and that the compound is found completely in the soil pore water. If a compound is not expected to reside completely in the pore water; i.e. adsorption to the soil occurs, a correction can be made to the rate constant as described by the authors, using the soil partition coefficient (k_P); in this method, no biodegradation is assumed when a compound is adsorbed to soil/sediment.
 

These rate constants were validated using only a small sample of compounds (14 for surface water; 4 for soil) and the authors report that the extrapolation of these rate constants to more persistent compounds appears to result in an overestimation of biodegradation in the environment. A study by Federle et al. (1997) comparing biodegradation rates for nine compounds in soil and water showed that for these compounds, rates of biodegradation were not strongly related to microbial biomass and the assumption that it is, made by Struijs and van den Berg (1995) above, may be flawed.

The EU Technical Guidance Document (TGD) (EC, 1994), assigns rate constants based on ready and inherent test data; these are not measured rate constants but are "guesstimates" based on expert judgement. Extrapolation of screening test results to surface water is given as follows:

• a rate constant of 0.047 day^-1 (half-life of 15 days) is given for a readily biodegradable compound which meets the criteria of 60-70% biodegradation in 28 days as well as the 10-day window,
• 0.014 day^-1 (half-life of 50 days) for a compound which is readily biodegradable but does not meet the 10-day window,
• 0.0047 day^-1 (half-life of 150 days) for a compound which does not pass the ready biodegradability test but does meet the requirements of degradability for an inherent test (70%).
• any other result (fails ready test and no inherent test data or fails inherent test and no ready test data available) gives a value of 0 day^-1 (half-life of infinity).
 

Because sorption to soil may decrease the bioavailability of some compounds (mainly those with high soil/water partition coefficients) soil half-lives were assigned by the TGD based on both screening test results and sorption characteristics (EC, 1994). A compound with a K_p (soil partition coefficient) of 100 l/kg, is given half-lives of 30, 90, and 300 days for passing the ready test including the 10-day window, passing the ready test but not the 10-day window, and for passing the inherent test, respectively. A compound with a K_p value >100 l/kg but 1000 l/kg has assigned half-lives of 300, 900, and 3000 days for the same test results listed above. Thus, a compound with a K_p value >100 l/kg will exceed the POP/PBT criteria in soil even if the compound is considered biodegradable in screening level tests. K_p can be calculated from the K_oc (the soil adsorption coefficient) value if the organic carbon content for an average soil is assumed.

Some objections to the selection of these rate constants have been made. Assigning a rate constant of 0 to a compound which may have failed the ready test, even if only just, has been criticized as being excessively stringent. Another, more liberal, scheme proposes first-order rate constants based on screening test results (Boethling, personal communication):

• a rate constant of 0.14 day^-1 (half-life of 5 days in soil and water) is given for a readily biodegradable compound which passes the RBT,
• 0.069 day^-1 (half-life of 10 days in soil and water) for a compound that does not pass the RBT but shows 40% mineralization in 28 days,
• 0.023 day^-1 (half-life of 30 days in soil and water) for a compound which shows 20% but <40% mineralization in 28 days in a RBT or passes an IBT(70% mineralization in 28 days),
• 0.0069 day^-1 (half-life of 100 days in soil and water) for a compound which shows 20% but < 70% mineralization in 28 days in an IBT,
• <20% mineralization in the RBT or the IBT results in a first-order rate constant of 0 day^-1.
 

Other qualifications for compounds which have low water solubility or are toxic and might fail screening tests based on reasons not related to biodegradability may also need to be considered.

Recommended Analysis of Screening Test Data Versus POP/PBT Criteria

Screening biodegradation data have been used to calculate or estimate biodegradation rate constants when neither field or grab sample studies are available (see above), but this does not seem scientifically defensible. These screening biodegradation tests are not good simulations of environmental soils and surface waters. They use mixed culture inocula and mineral salts media and are run for relatively short periods of time (< 1 month). The inoculum is often from sources such as sewage that are quite different than surface waters and soil and may also be preacclimated to the studied compound. These tests are often run at high concentrations of the test chemical which may affect the degradation kinetics as well as result in toxicity effects to the microorganisms; concentrations of carbon sources, other than the test chemical, are very low. A better approach would be to use the screening study data as a qualitative indication of whether the chemical is persistent or not. Also, a weight-of-evidence approach should be used as was done in the development of the BIODEG file for the Environmental Fate Data Base (Howard et al., 1987) since screening biodegradation studies are well known for having considerable variability (Howard et al., 1987) and a procedure is needed if more than one test result is available. The following paragraph discusses a reasonable approach which is summarized in Table 2.

If a chemical consistently reaches a pass level in one or several RBT or IBT methods, the chemical should be considered non-persistent. However, when there is more than one test there is often some conflicting data, especially for chemicals that are less biodegradable. If there are conflicting data, but the chemical reaches a pass level in one of the RBT, the chemical should be considered non-persistent. If the chemical passes an IBT method and there are no other data and even if it does not pass any RBT methods, the chemical should still be considered non-persistent, especially if the criteria are a half-life of six months, although the rate in the environment may be somewhat slow (half-life of months).

If the chemical does not reach a pass level in any of the screening tests, the actual results should be examined. If no biodegradation is observed in any of the tests, the chemical should be considered persistent. If some biodegradation (>20% degradation) occurs in any of the RBT methods, the chemical is probably not very persistent and should pass the 6 to 12-month soil/sediment half-life criteria and may pass the 2-month aquatic criteria if tested in a grab sample test. Therefore chemicals in this class should be considered non-persistent. Similarly, if a chemical exhibits some biodegradation (>20% degradation) in an IBT, but not enough to pass, the chemical may not be persistent, but would require grab sample testing to confirm this result; until grab sample testing is conducted, the chemical probably should be considered persistent.

If complex mixtures (e.g., creosote, PCBs, toxaphene) are being tested, screening test results will need to be more carefully assessed. There may be portions of the mixture that are readily biodegraded (e.g., mono- and dichlorobiphenyls within the PCB mixture) which may allow small percentages of degradation to occur while the majority of the material may be persistent.

Table 2. Criteria for assessing screening biodegradation tests

Type of Testa,b	Criteria	Conclusion	Comment
RBT	Passes at least one	Non-persistent	Rate in environment should be rapid
IBT	Passes at least one	Non-persistent	Rate in environment may be slow
RBT	Does not pass, but degradation is >20%	Non-persistent	Most likely non-persistent; would be desirable to conduct testing
IBT	Does not pass, but degradation is >20%	Persistent	May be non-persistent, but requires further testing
RBT	Does not pass and degradation is <20%	Persistent	May be non-persistent, but requires further testing
IBT	Does not pass and degradation is <20%	Persistent	Probably very persistent

^aRBT - Ready Biodegradation Test; ^bIBT - Inherent Biodegradation Test

Back to Table of Contents

3.1.3.4. Pure Culture Studies

Pure culture studies, where a compound is added to a mineral salts medium and inoculated with a pure culture of bacteria or fungi, can not be used to determine the biodegradation kinetics of a compound in the environment. They are useful mainly in identifying metabolites and thus possible pathways of biodegradation (Howard et al., 1987). These studies are generally characterized by high substrate and nutrient concentrations. The extent of breakdown of the chemical shown in a pure culture study also may not be representative of that found in the environment. Often biodegradation in the environment proceeds through the action of several different microbial populations. The result of no biodegradation in a pure culture study may be from toxicity effects due to the high substrate concentrations used or may simply reflect the inability of that particular bacterial strain to biodegrade the compound in question.

Back to Table of Contents

3.1.4. Prediction of Biodegradation Using Estimation Programs

While experimental data are considered to be the best indication of degradation in the environment, many compounds which will be screened for POP/PBT persistence characteristics will not have any available degradation information. Since completion of either laboratory or field tests to measure the degradation potential of a compound requires an investment in both time and resources, estimation programs can be used in the initial stages of POP/PBT identification to predict the degradation potential of a compound without experimental data and place it into categories such as "little potential concern" (model predicts rapid degradation), "possible concern" (intermediate probability of degradation) and "persistent" (model predicts resistence to degradation). Laboratory or field testing should be required for compounds placed into the last two categories if it is shown that they either exceed POP/PBT criteria for bioaccumulation or toxicity or if no data are available for any of the other POP/PBT criteria.

Characteristics intrinsic to the compound itself, such as high molecular weight, low water solubility, the presence of three or more aromatic rings, branching in the molecule, and the presence of halogen atoms (aerobic conditions only) have all been implicated as properties indicating that a compound will have increased resistance to biodegradation (Alexander, 1994; Howard, 1999). However, the biodegradation kinetics of a particular compound are also dependent on environmental conditions, the presence of other compounds, and the availability and concentration of the studied compound. Because of this, it is difficult to predict the rate of biodegradation of a compound based solely on its structure and nearly all of the available modeling programs provide only qualitative or semi-quantitative data.

Back to Table of Contents

3.1.4.1. Review of Biodegradation Estimation Programs

Many biodegradation estimation methods have been published; however, most of these are suitable only for a particular structural type or limited structural types and are not appropriate for screening the biodegradation potential of large numbers of compounds which may have many different structures and structural fragments. The programs listed below were developed to predict aerobic biodegradation and have been trained on a fairly large number of compounds.

Niemi et al. (1987) used data from BOD₅ screening studies to highlight structural fragments which inhibit (16 different fragments) or assist (12 different fragments) biodegradation. Compounds which have a non-degradable fragment present are considered to be resistant to biodegradation; compounds without a non-degradable fragment(s) but with a degradable fragment(s) are considered to be degradable. However, compounds which have both non-degradable and degradable fragments cannot be classified by this model. Training was completed on 284 compounds and 95.5% were correctly predicted to biodegrade slowly. A validation exercise on this method using MITI test results for 756 compounds correctly classified 68.7% of the non-degradable compounds (Degner et al., 1993).

Boethling et al. (1994) and Howard et al. (1992) used experimental data from the biodegradation database, BIODEG, (Howard et al., 1986) to calculate coefficients by both linear and non-linear regression equations for 36 different structural fragments [Biodegradability Probability Program (BPP), BIOWIN]. Molecular weight is also considered. In this model, a structural fragment which is known to increase the biodegradability of a compound usually gets a positive coefficient and a fragment known to decrease biodegradability usually gets a negative coefficient. Estimations can be made solely on the basis of the compound's molecular weight but the reliability of the estimation is probably quite low. Validation of the linear (BPP1) and non-linear (BPP2) BIOWIN models of primary degradation was performed using 295 compounds; prediction of rapidly-degrading compounds by both the non-linear and linear regression equations was 97% correct. The non-linear regression equation was better able to predict compounds which degrade slowly, (76% versus 86% linear vs. non-linear, respectively). However, Langenberg et al. (1996), using MITI test data for 488 compounds, reported that only 44 and 54% were correctly classified as slowly degraded by the linear and non-linear regression equations in BIOWIN, respectively. 93 and 86% of 385 rapidly-degrading compounds with MITI test data were correctly predicted using the linear and non-linear regression equations in BIOWIN, respectively.

A second estimation within the BIOWIN program predicts both primary (BPP4)and ultimate (BPP3) biodegradation based on a survey of 200 compounds by 17 biodegradation experts resulting in semi-quantitative estimations. 85 (primary biodegradation) to 94% (ultimate biodegradation) of the rapidly-degrading compounds (defined as a biodegradability score 3.5 for primary degradation and a score of >2.5 for ultimate degradation) and 79 (primary biodegradation) and 72% (ultimate biodegradation) of the slowly-degrading compounds (defined as a biodegradability score <3.5 for primary degradation and a score of 2.5 for ultimate degradation) were correctly predicted (Boethling et al., 1994).

A recent addition to the BIOWIN program is based on MITI test data. Several of the original fragments were revised and new fragments were added to the original program to account for toxicity effects. Validation of the new MITI method was performed on MITI results for 589 compounds; 335 of these were classified as "biodegrades slowly" based on experimental results. The non-linear regression equation was better able to predict compounds which degrade slowly, 84% versus 86% (linear vs. non-linear, respectively), than the original BIOWIN program. Overall, approximately 83 to 84% of the total compounds were predicted correctly.

Loonen et al. (1996) developed an estimation method to predict both ready and non-ready biodegradability using data from the MITI I test. One hundred and eleven structural fragments are used as model descriptors and calculation is by Partial Least Squares discriminant analysis. The model was generated using data from 600 compounds; validation of the method, using 198 different compounds with MITI I data, showed that 88% and 86% of the not ready and the ready test results, respectively, were correctly predicted using this model when borderline results were excluded.

Back to Table of Contents

3.1.4.2. Review of Biodegradation Kinetics from Estimation Programs

Currently, extrapolation of biodegradation estimations to actual default rate constant values is not possible. Some authors have used qualitative estimation data, however, to classify compounds as either persistent or readily degraded. According to the SAR/MPD joint project, an estimate in BIOWIN of weeks/months or longer for ultimate biodegradation (BPP3<3) suggests that the compound will not pass a RBT but should be tested for inherent biodegradability. A predicted biodegradability of days or days/weeks (BPP3>3) was considered to be a good indication of a readily biodegradable compound (U.S. EPA, 1994). Tyle and Niemela (1999) in a screening exercise used BPP1<0.15 and BPP3<2.2 (months or greater) to predict compounds which may be persistent in the environment. bkh Consulting Engineers has used a BPP1<0.1 and a value of 2 or less in the ultimate survey model to screen for persistent compounds (bkh Consulting Engineers, 1998).

Back to Table of Contents

3.1.4.3. Recommended Use of Biodegradation Estimates Versus POP/PBT Criteria

Until a more detailed comparison of known persistent compounds with the selected biodegradation estimation method is completed, it is not possible to scientifically select exact probability values as kinetic predictors. Optimally, with further experience, values could be selected and used to classify compounds as readily biodegraded, persistent, and as having intermediate biodegradability. As qualitative estimates can not be used to determine an overall persistence half-life and unless experimental data from hydrolysis or photolysis studies meet the proposed criteria, estimations of biodegradation should only be used to rank compounds with a similar lack of data.

Realizing that data from the MITI test are best used to predict whether a compound is degradable (a result of non-degradable could be due to toxicity, low water solubility effects, or a poorly acclimated microbial population), an initial screening using the soon-to-be-released MITI BIOWIN program to remove "persistent or potentially persistent" compounds which have a probability of <0.5 is proposed as a conservative method. This method has been shown to predict resistant compounds in the MITI test with 85% accuracy and should focus attention on potentially persistent compounds. The original BIOWIN program could then be used to screen the remaining compounds for those with a BPP1 >0.5 (readily degraded compounds) which would be classified as readily biodegraded, BPP1 of 0.2 to 0.5 as potentially persistent and a BPP1 of <0.2 as persistent. This estimation method order is suggested to maximize the number of correctly classified persistent compounds. The cutoffs at 0.2 and 0.5 are arbitrary and need review by comparison with experimental data. In addition, all estimation methods are likely to provide false negatives and positives. For example, a branched peroxide was reported to be resistant to biodegradation by the BIOWIN program suite although peroxides are very reactive and should rapidly degrade in the environment. In this case, the estimation method adds correction factors to impart resistance to biodegradation due to the branching of the molecule; however, as peroxide structures were not used in the development of this method, factors which could correct for the degradable peroxide structure are not available. Therefore, once chemicals are sorted into various classes of persistence, the chemicals need to be reviewed by experts familiar with the models to identify obvious chemicals that have been incorrectly estimated.

It should be reiterated however, that biodegradation estimation programs are not a substitute for experimental data and should be used with caution. They can be used to readily screen a large number of compounds, and may be useful in the initial classification of compounds without experimental data. However, a compound which is shown to have intermediate persistence or to be persistent based on current estimation methods, needs further laboratory or field testing before being definitively identified as a persistent compound according to POP/PBT criteria.

If current modeling programs are not appropriate for the chemical structure studied (contains only structural fragments which are not included in biodegradation models), then experimental degradation data for structurally-analogous compounds should be located and judgement used to obtain a scientifically-valid estimation of degradation for the unknown compound. In some cases, experimental data will not be available, the modeling programs will not be suitable for the compound in question and data on analogous compounds will not be found. For these compounds, it may be reasonable to determine other characteristics of PBT/POPness such as bioaccumulation and toxicity; if criteria for bioaccumulation and toxicity are exceeded, then experimental degradation data should be required.

Back to Table of Contents

3.2. Abiotic Degradation Processes

Certain chemical structures are susceptible to abiotic degradation. The major abiotic reactions considered in this report are hydrolysis and photolysis. This section provides information on calculating rate constants for these processes; biodegradation rate constants calculated from grab sample tests should be summed with rate constants from abiotic degradation processes, resulting in a value for overall persistence in the environment (see section 3.1.3.2.2.).

In comparison to microbial degradation, abiotic reactions are often slow or are specific to a particular pH range or light condition which may not occur where the compound is released to the environment. In addition, abiotic degradation generally results in small alterations of the compound and rarely results in mineralization. It should be noted though that abiotic processes can, in some cases, degrade a biologically-recalcitrant molecule into intermediates which are more readily biodegraded.

Back to Table of Contents

3.2.1. Hydrolysis

In aquatic environments, moist soils and groundwater, hydrolysis can be an important degradation mechanism for susceptible compounds. Hydrolysis reactions can be divided into acid- and base-catalyzed hydrolysis and neutral hydrolysis; as a compound may be susceptible to any or all of these mechanisms, a summation of the rate constants as shown below (see equation 8) should be made if possible. Esters in particular can undergo neutral, base-catalyzed, and acid-catalyzed hydrolysis simultaneously. Compounds in the following structural groups may be susceptible to hydrolysis: alkyl halides, allyl and benzyl halides, polyhalogenated methanes, epoxides, aliphatic acid esters, aromatic acid esters, amides, carbamates, phosphoric acid esters, thiophosphoric acid esters, phosphoric acid halides, dialkylphosphonohalidates, and dialkylphosphorohalides (Lyman, 1990). However, aromatic acid esters and amides only hydrolyze at a significant rate at very high pH values. Compilations of hydrolysis rate constants have been published in papers by Mabey and Mill (1978), Ellington et al. (1987) and Ellington (1989).

Back to Table of Contents

3.2.1.1. Hydrolysis Kinetics

Neutral hydrolysis is the reaction of an organic compound, theoretically described as RX, with water. In this process, the parent carbon-X bond is broken and a carbon-oxygen bond is formed at the same time (Lyman, 1990). The rate of neutral hydrolysis is pseudo first-order as water is present in excess and the rate changes only with the concentration of substrate present in the reaction mixture:

where k_n is the rate constant for neutral hydrolysis, S is the substrate concentration and t is time. In both acid- and base-catalyzed hydrolysis, hydrogen and hydroxide ion concentrations, respectively, as well as substrate concentration control the rate of reaction and thus the rate kinetics are second-order as follows:

where k_b is the rate constant for base-catalyzed hydrolysis and OH^- is the hydroxide ion concentration and,

where k_a is the rate constant for acid-catalyzed hydrolysis and H⁺ is the hydrogen ion concentration. A pseudo first-order rate constant can be calculated from the acid- and base-catalyzed second-order rate equations (valid only at the selected pH) by multiplying the second-order rate constant with the hydrogen or hydroxide ion concentration at the measured pH value.

The rate of hydrolysis can be affected by pH, ionic strength, temperature, and the presence of solvents, humic acids and metal ions. Most fresh waters in the environment are characterized by low ionic strength and a pH value generally ranging from 4 to 9 (Hemond and Fechner, 1994). When acid- or base-catalyzed rate kinetics are studied, the pH is varied, preferably over this range. Data from laboratory studies with measured hydrolysis rates within the 4 to 9 pH range, and using fresh water, distilled water, or a minimally-buffered aqueous system should be suitable for the measurement of environmentally-relevant rate constants. The buffer is used to keep the pH stable as some hydrolysis reaction products can change the pH of the reaction medium. Other authors conduct studies using low concentrations of substrate in order to avoid pH changes.

In some cases, hydrolysis is measured at elevated temperatures so that observable rates can be measured. Hydrolysis is temperature dependent with the rate constant increasing as the temperature of the study increases. Studies run at non-environmental temperatures [an average non-winter temperature of 20 C was determined for 111 streams and rivers making up over 95% of the total water volume in the U.S. (USGS, 1972-1973)] can be corrected using the Arrhenius equation:

where E_A is the Arrhenius activation energy (kcal/mol), R is the gas constant (1.987 cal/deg-mol), and T is the temperature (in Kelvin).

If a compound is insoluble in water, hydrolysis may be measured in a mixed water-organic medium. Standard protocols (e.g. EPA - OPPTS 835.2130) use a 1% acetonitrile solution (1% ethanol is used in place of acetonitrile if the experiment is carried out at temperatures above 80 C) and this can be assumed to give hydrolysis rates similar to that in water alone.

Each hydrolysis rate constant should be corrected separately for temperature, ionic strength, and solvent effects, if necessary or possible, before summing to form an overall hydrolysis rate constant. Assuming first-order in substrate concentration, the summation equation becomes:

where k_Hyd is the overall hydrolysis rate constant. At a fixed pH, the overall kinetics are pseudo first-order and the hydrolysis half-life is valid only at that pH value.

Back to Table of Contents

3.2.1.2. Interconversion of Hydrolysis Rates Between Different Media

The abiotic neutral hydrolysis of compounds in sediment-water systems has been shown to be comparable to that reported in water alone (Wolfe et al., 1989). Hydrophobic esters, however, exhibited a decreased rate of base-catalyzed hydrolysis in the presence of sediment (Macalady and Wolfe, 1985). As the effect of ionic strength on the hydrolysis reaction rate of any given compound cannot be predicted (Mabey and Mill, 1978), hydrolysis rates calculated in salt or brackish water can not be used without qualification to estimate fresh water hydrolysis, and vice versa. Perdue (1983) reports that, based on limited data, the rate of base-catalyzed hydrolysis is inhibited while the rate of acid-catalyzed hydrolysis is increased when susceptible compounds are incubated in the presence of humic acids. Humic acids are commonly found in natural waters.

Hydrolysis data obtained in aqueous systems must be used carefully to describe hydrolysis in moist soil; organic carbon and clay content can affect the sorption of a compound to the soil matrix, soil temperatures can reach higher values than that in an aquatic environment, and varying soil moisture can act to either increase or decrease the rate of hydrolysis in comparison to that measured in an aqueous system. The effect from these differing conditions is not currently predictable (Wolfe et al., 1989). Soil hydrolysis data can sometimes be inferred from soil grab sample studies where biodegradation loss is corrected for loss due to hydrolysis.

Back to Table of Contents

3.2.1.3. Hydrolysis Estimation Programs

A structure-based aqueous hydrolysis estimation program (HYDROWIN; Meylan, 1995) can be used to estimate rate constants for acid- and base-catalyzed hydrolysis (Mill et al., 1987). Neutral hydrolysis is not predicted by this estimation method. Chemical classes which are predicted by HYDROWIN are esters, carbamates, epoxides, halomethanes and selected alkyl halides. Other structures such as isocyanates, acyl halides, cyclic esters, ureas, amides, and carbonate esters are highlighted by the program (either as reacting extremely rapidly, e.g. isocyanates and acyl halides, or as reacting very slowly such as cyclic esters or amides) although a rate constant is not determined. Not all structures which can hydrolyze are included in this estimation program. For example, the hydrolysis of benzyl and allyl halides can not be predicted by HYDROWIN. ATHIAS (Abiotic Transformations of Halogenated hydrocarbons In Aqueous Solution) can be used to obtain base-catalyzed and neutral hydrolysis rates at various pH values and temperature conditions for 58 different halogenated aliphatic compounds (alkyl halides) (Ellenrieder and Reinhard, 1988).

Back to Table of Contents

3.2.2. Photolysis

Degradation of a compound in the environment is possible by direct and/or indirect photolysis. Both processes are reviewed below and the calculation of rate constants from this data examined.

Back to Table of Contents

3.2.2.1. Direct Photolysis

A compound that is directly photolyzed adsorbs light energy which then, in some cases, causes the chemical structure to be altered. Adsorption of light at environmental wavelengths (290 to 750 nm) however, is not sufficient evidence to indicate that a compound will undergo photodegradation. Many photoexcited compounds will release the adsorbed energy without undergoing chemical transformation. Compounds such as alkenes or aromatic rings, with unsaturated carbon/carbon bonds, as well as nitrosamines, benzidines, chlorinated organics and some metal complexes are susceptible to direct photodegradation (Zepp, 1980).

Direct photolysis of any compound can occur under the "right" experimental conditions; laboratory test results can show photolysis under conditions which are not environmentally relevant and should be carefully screened because of this. Artificial light sources and filters are used in laboratory screening experiments to approximate sunlight. Both mercury-vapor lamps and xenon arc lamps are commonly used; photolysis kinetics measured using a xenon arc lamp are the closest to that from sunlight (Zepp, 1980). Because the light emitted from these sources contains wavelengths less than 290 nm, the light must be filtered. Pyrex filters exclude wavelengths less than 290 nm; however, quartz filters exclude wavelengths less than 230 nm and thus are not appropriate for use in experiments estimating the potential of direct photolysis in the environment. As many of these light sources also generate heat, a cooling system should be used. If the compound is not very soluble in water, polar solvents such as acetonitrile and methanol or a mixture of water and the solvent can be used (Zepp, 1978).

Laboratory studies are used as a screening tool, showing the ability or inability of a compound to photolyze. Several researchers have attempted to extrapolate laboratory-measured rates to natural samples exposed to sunlight by measuring the intensity of the light and correcting it for that found in sunlight. This disregards, however, the spectral distribution of the light used in the photolysis study versus that in sunlight. This is affected in turn by the "solar zenith angle, ozone layer thickness and other environmental factors" (Zepp, 1980) such that the spectral distribution seen in winter sunlight can be significantly different from that seen in the summer. Because of this, many photolysis studies measure the photodegradation of a compound in sunlight, on a clear summer day at midday.

Kinetics of Direct Photolysis

The rate of direct photolysis is based on the light absorption of the measured compound, I_O, and the quantum yield, :

The quantum yield has been measured for only a few compounds; Mill (1989) suggests that it can be assumed to be 1 for screening purposes (assumes 100% efficiency). At low concentrations of compound, as is expected in the environment, the rate of light absorption is proportional to the substrate concentration. Thus, first-order kinetics can be assumed and direct photolysis can be described by first-order kinetics:
where k_dir is the direct photolysis rate constant (Mill, 1989). However, the quantum yield is rarely greater than 0.01 so the estimated rate from this procedure will be considerably higher than will actually occur. As this is not a conservative method, the assumption of a quantum yield value is not recommended.

If photolysis studies are not available, spectral data, usually generated in solvents other than water can be used to indicate that a compound may be susceptible to photodegradation. Data optimally should be corrected for solvent effects. A compound which adsorbs strongly above 290 nm and particularly in the visible (400 to 700 nm) range may potentially photolyze. If the compound does not adsorb above 290 nm then it is unlikely that it will photolyze at environmentally-relevant rates in the environment.

Back to Table of Contents

3.2.2.2. Indirect Photolysis

Compounds which may not undergo direct photodegradation may be photolyzed indirectly via sensitized photolysis or photooxidation. During sensitized photolysis, another compound present in the environment [the photosensitizer (e.g., quinones, humic acid, flavins, as well as trace metals, nitrate, nitrite, and hydrogen peroxide) (Zepp, 1991)] adsorbs light energy and then transfers it either indirectly via various reactive intermediates or directly to the compound of interest resulting in potential structural alteration. Measured degradation rates which are higher in natural waters when compared to distilled water or a large variation in the rates measured in different natural waters indicate that indirect photolysis may be occurring. Rates reported in laboratory studies where sensitizers (such as acetone, diethylamine) are added to the water may be affected by the concentration of sensitizer added.

Photooxidation in aquatic environments is possible for some chemical structures. Light can be directly absorbed by dissolved organic matter, naturally present in many fresh water sources, and used to generate oxidants which then react with the organic compound resulting in a structural change. Oxidants which are commonly found in aquatic environments include the peroxy radical (RO·), hydroxy radical (HO·), singlet oxygen (¹O₂), ozone (O₃), and triplet diradicals (Mill, 1989).

Kinetics of Indirect Photolysis

Calculation of indirect photolysis rates is more complicated than that for direct photolysis as molar absorptivities and concentrations of the photosensitizers present in natural waters are usually unknown. The reaction kinetics are second order, dependent on both the concentration of the sensitizer/oxidant and the compound present in that environment (Mill and Mabey, 1985; Mill, 1999) as follows:

where k_sen and k_ox are the second-order rate constants for photosensitization and photooxidation, respectively, and the concentrations of sensitizer and oxidant are given by [Sen] and [Ox], respectively. Average concentrations of the photosensitizers/oxidants present in natural waters have been published (Mill, 1999); however, concentrations can vary considerably both temporally and spatially. The kinetics can be reduced to first order if it is assumed that the concentration of oxidant/sensitizer does not change during the study period. A single rate constant should be calculated for both sensitization and for photooxidation by summing the rate constants for each reacting species.

The overall photolysis rate constant is the sum of both direct and indirect photolysis processes described as:

 

3.2.2.3. Photolysis in Aquatic Environments

In aquatic photolysis studies, photolysis rates in natural versus distilled water vary due to the presence in natural waters of particulates (may increase in summer months due to algal and bacterial growth) and dissolved substances (such as humic acids) which attenuate sunlight, result in other, indirect photoreactions, and may interact either chemically or physically to make the compound less available for reaction (Hwang et al., 1987; Zepp and Wolfe, 1987). Thus experiments which are conducted under realistic environmental conditions, are most likely to give the most environmentally-relevant rate constant for that body of water. Laboratory experiments conducted in water or water/solvent mixtures can suggest that a compound may photolyze in the environment; however, data from these experiments should be used carefully to calculate realistic rate constants. For example, data on 13 different PAHs show that these compounds photolyze rapidly with half-lives, due to direct photolysis alone, calculated at latitude 40 N at midday in midsummer ranging from 0.13 hours for 9-methylanthracene to 4.4 hours for chrysene to 21 hours for fluoranthene (Zepp and Schlotzhauer, 1979). However, several of these compounds can be quite persistent in the environment; they can partition to particulate material in the water column and end up in the sediment layer where photolysis is unlikely.

3.2.2.4. Soil Photolysis

Photolysis may occur on soil surfaces although judgement should be practiced if rate constant values from this data are used. For example, laboratory studies measuring the photolytic degradation of DDT in soil reported a half-life of 125 days (Vollner and Klotz, 1994); however, field studies in soil showed that this compound had substantially longer degradation half-life values, ranging from 2 to over 15 years (Lichtenstein and Schultz, 1959; Stewart and Chisholm, 1971). Some compounds may leach below the soil surface and become unavailable or, over time, compounds may become strongly adsorbed to the soil matrix resulting in stabilization to photolytic degradation. Photooxidation is also assumed to occur on soil surfaces;however, very little is published on this subject due to the complexity of the system (Mill, 1989).

Rate constants derived from aquatic systems can not be used to estimate the kinetics of photolysis on soil surfaces; therefore, a first-order photolysis rate constant should be calculated separately for both soil and water, if possible. Photolysis in sediment is not expected to be a major degradation process (Zepp and Schlotzhauer, 1979). Estimation programs which can estimate either direct or indirect photolysis rates in water or soil are not currently available for screening purposes (Mill, 1989; Mill, 1999).

4. Persistence in the Atmosphere

The procedure for screening compounds for persistence in the atmosphere is similar to that outlined above for determining persistence in soil, sediment and water. Experimental data are given the highest priority in the calculation of persistence half-lives in the atmosphere. If experimental data are not available then estimation methods can be used, within the constraints of the method, to estimate atmospheric half-lives. Unlike estimation methods for biodegradation, the estimation of the reaction of the hydroxyl radical with organic compounds is fairly well defined, a quantitative value is given and the available models are considered to be fairly accurate.

In the atmosphere, organic compounds are distributed into the gaseous, aqueous, and/or particulate phase depending mainly on the liquid phase-vapor pressure of a particular compound. An organic compound with a liquid-phase vapor pressure < 10^-8 mm Hg at the ambient air temperature is expected to exist mainly in the particle phase; a compound with a liquid-phase vapor pressure > 10^-4 mm Hg will be found almost completely in the gas phase (Eisenreich, 1981). Compounds with liquid phase-vapor pressures between 10^-4 and 10^-8 mm Hg will exist in both the gas and particle phases. Removal from the atmosphere can occur by both physical (wet and dry deposition) and chemical (degradation) processes. Chemical degradation processes include direct and indirect photolysis in the gaseous, aqueous and most likely the particulate phases as well as hydrolysis of susceptible compounds in the aqueous phase.

4.1. Direct Photolysis

Light in the troposphere is limited to wavelengths greater than or equal to 290 nm. Therefore, a compound must be able to adsorb light between 290 and 800 nm in order to potentially directly photolyze in the troposphere. As mentioned with photolysis in water, however, adsorption of light does not, by itself, ensure that a compound will degrade. Absorption spectra can only show the potential for a compound to directly photolyze in the troposphere; experimental data are necessary to prove that photolysis does occur and that it proceeds at a environmentally-significant rate. The rate constant for gas-phase photolysis in the troposphere can be calculated if experimental data are available using the following equation:

Compounds which are found completely or partially in the particle phase may be less susceptible to photodegradation. Coal stack ash acted to stabilize pyrene resulting in reduced photolytic degradation (Dunstan et al., 1989). A study comparing photolytic half-lives of 18 PAH adsorbed to 15 different coal fly ash samples reported that darker ash samples (higher carbon content) adsorbed the most light, reducing photolytic degradation of the adsorbed organic compound (Behymer and Hites, 1988). While the PAH compounds benzo[a]pyrene, pyrene, and anthracene are rapidly photolyzed in aqueous solution, they are resistant to photodegradation when adsorbed to coal fly ash (Korfmacher et al., 1980).

4.2. Indirect Photolysis/Photooxidation

Most organic compounds in the troposphere can be structurally altered by reaction with hydroxyl radicals (daytime reaction), and some compounds are susceptible to reaction with nitrate radicals (nighttime reaction), or ozone. As all of these oxidants are photolytically produced, this process is called atmospheric photooxidation. Of these reactions, gas-phase reaction with hydroxyl radicals is generally the most important and most common removal process in the troposphere (Atkinson, 1989); if reaction with hydroxyl radicals is slow or non-existent it is likely that the compound will not react significantly with either ozone or nitrate radicals (OECD, 1992). In general, compounds with unsaturated carbon/carbon bonds, furans, pyrrole, dimethyl sulfide (and possibly aliphatic amines and hydrazines) and phenols are potentially susceptible to nighttime reaction with nitrate radicals. Aliphatic compounds with internal unsaturated carbon/carbon bonds, acetylamines and hydrazines are susceptible to attack by ozone; reaction of aromatic and saturated hydrocarbons with ozone is not environmentally significant.

Measurement of rate constants for the reaction of hydroxyl or nitrate radicals with organic compounds is made using either the absolute or the relative rate method. The absolute method measures either 1) the loss of one species in the presence of a known excess of the second reactant (reaction then becomes pseudo first-order) or 2) the loss of both species (second-order rate kinetics). In the relative rate method, the rate of hydroxyl or nitrate radical reaction for an unknown organic compound is determined relative to the known reaction rate of another organic compound (which was measured by absolute rate constant methods). In both methods, discharge flow, flash (or laser) photolysis, steady-state photolysis, and pulse radiolysis are commonly used to generate the required radical concentration and a wide variety of detection systems have been used to measure the loss of the studied compound(s) (Atkinson, 1989).

Compilations of experimentally-derived rate constants are available for the gas-phase reaction of organic compounds in the atmosphere with inorganic species (Atkinson, 1989; Atkinson, 1994). If experimental rate constant values are not available, several modeling programs are capable of estimating reasonable values for the gas-phase reaction with the hydroxyl radical and with ozone (see section below). Experimental rate constants measured at elevated temperatures should be corrected to ambient temperatures using the Arrhenius equation (see equation 7) before a half-life is calculated. Rate constants are often given for temperatures of 20 to 25 C and while this is probably higher than temperatures found at many higher altitudes, the calculated half-lives are conservative values. Experimental or estimated reaction rates for compounds in the gas-phase may not be accurate for particle-phase compounds. The adsorption of a compound to particulate material is believed to reduce its ability to be photooxidized and may thus increase the lifetime of the compound in the troposphere.

Methods for estimating the reaction of gas-phase hydroxyl radicals and ozone with organic compounds in the troposphere if experimental data are missing are available. The Atmospheric Oxidation Program (AOP), based on an estimation method published by Atkinson (1985, 1987, 1988), estimates rate constants for the reaction between gas-phase hydroxyl radicals and organic compounds and between ozone and olefinic/acetylenic compounds based on structure-activity relationships (Meylan and Howard, 1993). Validation of this program compared the estimated versus experimental rate constants for 448 compounds; 90% of the estimations were within a factor of 2 of the experimental data and 95% were within a factor of 3. Gas-phase hydroxyl radical rate constants based on molecular orbital calculations can be estimated for selected structures (Klamt, 1993; Klamt, 1996). Currently, programs estimating the reaction with nitrate radicals are not available although compounds are screened for susceptible structures in the AOP program. As with any estimation program, results should be carefully screened for compounds with structures which were not used in the development of the estimation method. Results for these types of compounds are not reliable and thus should only be used with caution in the POP/PBT decision process.

Once a bimolecular rate constant is obtained for the required reaction, calculation of the half-life can be made using the following equations.

(1) For the hydroxyl radical reaction:

The concentration of hydroxyl radicals (or other oxidants) used in the calculation of these second-order rate constants has a large effect on the value of the atmospheric half-life. For example, nearly all commercial chemicals will exceed the current 2-day criterion if the average hydroxyl radical concentration for North America in the winter is used. Published average values of hydroxyl radical, nitrate radical, and ozone concentrations in the troposphere are available (Table 3). The POP/PBT atmospheric half-life criterion was based on the default hydroxyl radical concentration present in Atkinson's model (1.5x10⁺⁶ radicals/cm³), thus, use of this or similar concentrations could provide a consistent approach to determining atmospheric lifetimes. However, as atmospheric concentrations of reactive species can vary substantially (e.g. between pristine and polluted areas, time of year, latitude), it is recommended that a range or distribution of concentrations be used for each compound. Nitrate radicals are very susceptible to photolysis and thus this is essentially a nighttime reaction; the average concentration given in the table below is the nighttime concentration over land. Nitrate radical concentrations over marine areas are at least 10-fold less than over land (Noxon, 1983). Reaction of chemicals with hydroxyl radicals is a daytime reaction.

Table 3. Concentration of reactive inorganic species in the troposphere

An overall first-order rate constant for degradation in the atmosphere can be determined as follows:

5. Summary

Half-life criteria have been published in order to identify potential POP/PBT compounds based on their persistence in the environment. This report provides information on degradation processes in the environment and methods for collecting persistence data, evaluating this data for validity, and calculating half-life values for comparison to the POP/PBT persistence criteria. Although a compound released into the environment will have a tendency to accumulate in one medium over another, the focus of data collection initially should be on collecting the available degradation information and calculating persistence half-lives for each medium whenever possible.

In soil, water, or sediment environments, experimental data, when available, is used in the following order: field studies, grab sample studies, ready biodegradability screening tests and finally inherent biodegradability screening tests. Field and grab sample studies are given the highest priority for calculation of persistence half-life values as they represent the most environmentally-realistic study conditions. A summary of persistence half-life calculations from these tests is as follows:

• Field tests - overall persistence half-life is directly calculated from test data after correction for transport loss.
 
• Grab sample tests - when abiotic and biotic loss are not distinguished, an overall persistence half-life is directly calculated from test data after correction for non-degradation loss. If a first-order biodegradation rate constant from the grab sample test is given or can be calculated, this value can be added to first-order hydrolysis and photolysis rate constants to obtain an overall first-order persistence rate constant. This rate constant can then be readily converted into an overall persistence half-life.
 
• Screening tests - results are used to screen for ready and not ready biodegradability. These tests provide a good source of information about the potential for degradation; however, the determination of kinetics from these tests, although done, is not scientifically defensible. Screening test data incorporate loss due to both hydrolysis and biodegradation and thus, unless photolysis is a major mechanism of loss for the studied compound, further incorporation of abiotic effects into the final judgement of degradability is not necessary.

Experimental data may not be available for many compounds and in this case, estimation of the potential for biodegradation may be possible. Estimates of degradability should be used with caution and preferentially as a screening and classification tool; any final decision on the degradability or non-degradability of a compound should be based on scientific evidence alone. Estimation programs are best used to classify compounds into categories such as readily degraded, intermediate degradability, and persistent. If current estimation programs are not able to accurately determine the biodegradation potential of particular compounds, experimental data for structurally-similar compounds can be used in combination with scientific judgement to determine a reasonable estimation of degradability for the unknown compound. In some cases, experimental data will not be available, the modeling programs will not be suitable for the compound in question and data on analogous compounds will not be found. For these compounds, it may be reasonable to determine other characteristics of PBT/POPness such as bioaccumulation and toxicity; if criteria for bioaccumulation and toxicity are exceeded, then experimental degradation data should be required.

Atmospheric half-lives are more readily determined than those in soil, sediment, or water environments. In general, experimental data are limited to approximately 600 chemicals; but estimation methods are considered to be quite accurate. However, current estimation methods are for gas-phase reactions with major inorganic species in the troposphere and may thus underestimate atmospheric half-lives for those compounds (which include many POP/PBT compounds) which are found partially or mainly in the particulate phase.

Whether data are collected for air, soil, water, or sediment, variation in half-life values are expected for each studied compound. This will be due to spatial and temporal differences, differing concentrations of reactive species, variations in experimental protocols or monitoring methods, as well as numerous other factors. For this reason, whenever possible, a range of half-lives should be collected along with a measurement of central tendency. As the published POP/PBT criteria were meant as guidelines to identify persistent and non-persistent compounds they should be viewed as such and scientific judgement applied as necessary when results conflict.

6. References

Alexander, M. 1994. Biodegradation and Bioremediation. Academic Press: San Diego, CA.

Aronson, D. and Howard, PH. 1997. Anaerobic Biodegradation of Organic Chemicals in Groundwater: A Summary of Field and Laboratory Studies. Prepared for: American Petroleum Institute et al. http://esc.syrres.com/~esc/anabiodeg.htm. November 12, 1997.

Aronson, D., Citra, M., Shuler, K., Printup, H., and Howard, PH. 1999. Aerobic Biodegradation of Organic Chemicals in Environmental Media: A Summary of Field and Laboratory Studies. Prepared for: U.S. Environmental Protection Agency, Athens, GA.

Atkinson, R. and Carter, WPL. 1984. Kinetics and mechanisms of the gas-phase reactions of ozone with organic compounds under atmospheric conditions. Chem. Rev. 84: 437-470.

Atkinson, R. 1985. Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic compounds under atmospheric conditions. Chem. Rev. 85: 69-201.

Atkinson, R. 1987. A structure-activity relationship for the estimation of rate constants for the gas-phase reactions of OH radicals with organic compounds. Intern. J. Chem. Kinet. 19: 799-828.

Atkinson, R. 1988. Estimation of gas-phase hydroxyl radical rate constants for organic chemicals. Environ. Toxicol. Chem. 7: 435-442.

Atkinson, R. 1989. Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with organic compounds. J. Phys. Chem. Ref. Data. Monograph No.1.

Atkinson, R. 1991. Kinetics and mechanisms of the gas-phase reactions of the NO₃ radical with organic compounds. J. Phys. Chem. Ref. Data. 20: 459-507.

Atkinson, R. 1994. Gas-phase tropospheric chemistry of organic compounds. J. Phys. Chem. Ref. Data. Monograph 2.

Atkinson, R. 1996. Atmospheric chemistry of PCBs, PCDDs and PCDFs. Issues Environ. Sci. Technol. 6: 53-72.

Barker, JF., Patrick GC., and Major D. 1987. Natural attenuation of aromatic hydrocarbons in a shallow sand aquifer. Ground Water Monit Rev. 7: 64-72.

Bartholomew, GW. and Pfaender, FK. 1983. Influence of spatial and temporal variations on organic pollutant biodegradation rates in an estuarine environment. Appl. Environ. Microbiol. 45: 103-109.

Behymer, T.D. and Hites, R.A. 1988. Photolysis of polycyclic aromatic hydrocarbons adsorbed on fly ash. Environ. Sci. Technol. 22: 1311-1319.

Boethling, R.S., Howard, P.H., Meylan, W.M., Stiteler, W., Beauman, J. and Tirado, N. 1994. Group contribution method for predicting probability and rate of aerobic biodegradation. Environ. Sci. Technol. 28: 459-465. (BIOWIN Software available from Syracuse Research Corp., Environmental Science Center, North Syracuse, NY 13212)

Boethling, R.S., Howard, P.H., Beauman, J.A. and Larosche, M.E. 1995. Factors for intermedia extrapolation in biodegradability assessment. Chemosphere. 30: 741-752.

Borden, R.C., Daniel, R.A., LeBrun IV, L.E. and Davis, C.W. 1997. Intrinsic biodegradation of MTBE and BTEX in a gasoline-contaminated aquifer. Water Res. Research. 35: 1105-1115.

Chemicals Inspection and Testing Institute. 1992.