The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 1– 7 DOI 10.1007/b11460
Introduction Marian K. Stanley 1 · Kenneth A. Robillard 2 · Charles A. Staples 3 1 2 3
American Chemistry Council, 1300 Wilson Blvd., Arlington, VA 22209, USA E-mail:
[email protected] Eastman Chemical, Rochester, NY, USA Assessment Technologies, Inc. 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA
The esters of 1,2-benzene dicarboxylic acid, commonly called phthalate esters, are a diverse group of compounds that have broad use in a wide array of industrial applications. Regulatory oversight of the manufacture, transport, use and disposal of phthalate esters has produced a large amount of data regarding the properties, environmental fate, exposure, and toxicity of these compounds. Such data are critical for the development of safe and accepted production practices, effluent discharge limits, and human exposure limits. The following chapters present, in detail, the information that has been collected regarding these properties. Keywords. Phthalate, Manufacture, Use, Releases, Regulation
1
General Description
. . . . . . . . . . . . . . . . . . . . . . . . .
1
2
Manufacture of Phthalate Esters . . . . . . . . . . . . . . . . . . .
2
3
Nomenclature and Physical/Chemical Properties
. . . . . . . . .
3
4
Chemical Interactions with Vinyl . . . . . . . . . . . . . . . . . .
3
5
Uses of Phthalate Esters
. . . . . . . . . . . . . . . . . . . . . . .
3
6
Disposal and Releases into the Environment . . . . . . . . . . . .
5
7
Regulations and Phthalate Esters . . . . . . . . . . . . . . . . . .
6
8
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
1 General Description Phthalate esters are widely used industrial chemicals. Higher molecular weight phthalate esters act as an additive which imparts flexibility in vinyl resins; this is the highest volume use of phthalates. Both linear and branched phthalate esters are used in the manufacture of vinyl articles. The linear esters provide superior low temperature properties to the finished vinyl products and also © Springer-Verlag Berlin Heidelberg 2003
2
M.K. Stanley et al.
have lower volatility. The C8–C13 phthalate esters are the dominant vinyl plasticizers with di-2-ethylhexyl phthalate, diisononyl phthalate predominant and diisodecyl phthlatate. The lower molecular weight phthalates are used as plasticizers in some non-vinyl resins, including acrylics, urethanes and cellulosics. The various esters used in commerce have alkyl side chains containing from 1 to 13 carbon atoms. Table 1 contains a listing of the most common phthalate esters. Most of the high molecular weight phthalates esters are used in the manufacture of a wide variety vinyl goods, both commercial and consumer. The lower molecular weight phthalates, those with alkyl side chains from 1 to 4 carbon atoms, have a very broad use which includes consumer products and pharmaceuticals. This will be detailed in the use section.As plasticizers, phthalates are additives which improve the flexibility, processability and softness of vinyl. Phthalates with alkyl side chains lower than C6 are not often used alone as a plasticizer because of volatility concerns. In general, the factors that dictate selection of a phthalate or combination of phthalates for a particular application are functionality and economics of use [1]. Overall, phthalate esters are used because they combine qualities such as compatibility, permanence, efficiency and processability at reasonable cost. Compatibility problems with the vinyl resin preclude the use of phthalates esters of molecular weight higher than ditridecyl. In vinyl, dibutyl phthalate is only used in isolated cases in conjunction with higher molecular weight plasticizers to reduce volatility. Dipropyl and dipentyl phthalate, C3 and C5, are not available commercially in the United States.
2 Manufacture of Phthalate Esters The ortho phthalate esters are generally manufactured by the sequential addition of either branched or normal alcohols to phthalic anhydride in the presence of an acid catalyst. The alcohol manufacturing processes are stable, so although the phthalates produced from branched alcohols are complex substances, they are not variable. Phthalate esters are products of simple esterification reactions, which can be carried out readily in heated kettles with agitation and provision for water removal.While some plants produce phthalates by the batch method, newer, highly automated plants operate continuously, particularly if they emphasize a single product. The purity requirements for commercial plasticizers are very high and phthalate diesters are usually colorless and mostly odorless. The reaction usually requires an excess of alcohol, which is readily recycled. Diisononyl phthalate (DINP) is a complex substance assigned two different CAS numbers. CAS number 68515-48-0 is manufactured from polygas branched olefin that is converted to alcohol moieties consisting mainly of 3,4-, 4,6-, 3,6-, 3,5-, 4,5- and 5,6-dimethyl-1-heptanol. The CAS number 28553-12-0 is produced from dimerized n-butene that is converted primarily to methyl octanols and dimethyl heptanols. This CAS number also represents DINP which is produced from n-butene and isobutene that are converted to alcohols, with 60% consisting of methylethyl hexanols. The two types of DINP are considered commercially
Introduction
3
interchangeable. Other phthalates that are complex mixtures are diisodecyl phthalate (DIDP) and the linear phthalates D610P and D711P.
3 Nomenclature and Physical/Chemical Properties The physical and chemical properties of phthalate esters are well documented and will be discussed in detail in this handbook. A summary of some of their physical properties is contained in Table 1. The phthalate esters discussed here are liquids at room temperature. The diesters derived from the lower molecular weight alcohols such as DMP and DEP are colorless fluids of low viscosity, but phthalate esters become more viscous and oily as the size of the alkyl side-chain increases. They have low freezing points, many well below 0 °C (see Table 1). Generally, the water solubility of the alkyl phthalate ester varies inversely with the length of the alkyl side chain. DMP is the most hydrophilic and water soluble of the esters. The C10, C11 and C13 esters are the most hydrophobic and least water soluble (< 0.001 mg/l). Most of the dialkyl phthalates are soluble in common organic solvents such as benzene, toluene, xylene, diethyl ether, chloroform and petroleum ether [2]. In many cases alternate names are used in the literature and in commerce for the common phthalate esters. Table 1, while not intending to be exhaustive, lists these synonyms. Note the DOP (“dioctyl phthalate”) is used as a synonym for DEHP (di(2-ethylhexyl)phthlate).
4 Chemical Interactions with Vinyl Incorporating phthalate esters into a polymeric matrix reduces the glass transition temperature of the polymer [3]. Phthalate esters are not bound to the polymer with covalent chemical bonds and are therefore able to migrate to the surface of the polymer matrix where they may be lost by a variety of physical processes. Nevertheless, various chemical-physical attractive forces hold the phthalate ester tightly within the vinyl matrix, so that such migration occurs at a very low rate. Retention in the polymer matrix is one of the main factors in considering which phthalate ester to use. The ester must be sufficiently nonvolatile to remain in the compound during its mixing and formation stages
5 Uses of Phthalate Esters Uses of phthalate esters can be broadly split into three general categories – vinyl plasticizers, plasticizers for non-PVC polymers and other minor specialized applications. In the United States, DEHP, DINP and DIDP account for 52.2% of phthalates consumed. Linear phthalates account for 21.4% of the consumption and the 26.4% balance of US consumption includes all other phthalates esters [4].
Dimethyl Phthalate Diethyl Phthalate Di-n-Butyl Phthalate Diisobutyl Phthalate Butylbenzyl Phthalate Dihexyl Phthalate
Diisoheptyl Phthalate
Di-n-Octyl Phthalate Di (n-Hexyl, n-Octyl, n-Decyl) Phthalate Di(2-Ethylhexyl) Phthalate Diisononyl Phthalate
Diisodecyl Phthalate
Di(Heptyl, Nonyl, Undecyl) Phthalate
Diundecyl Phthalate Ditridecyl Phthalate
DMP DEP DnBP DIBP BBP DHP
DIHP
DnOP D610P
DIDP
D711P
DUP DTDP
DINP
DEHP
Phthalate Ester
Abbreviation
C30H50O4 C34H58O4
C26H42O4
C28H46O4
C26H42O4
C24H38O4
C24H38O4 C25H40O4
C22H34O4
C10H10O4 C12H14O4 C16H22O4 C16H22O4 C19H20O4 C20H30O4
Formula
Table 1. Physical properties of phthalate esters
28553-12-0 68515-48-0 26761-40-0 68515-49-1 3648-20-2 68515-44-6 68515-45-7 111381-89-6 111381-90-9 111381-91-0 3648-20-2 119-06-2 68515-47-9
131-11-3 84-66-2 84-74-2 84-69-5 85-68-7 84-75-3 68515-50-4 7188-89-6 6815-44-6 117-84-0 25724-58-7 68515-51-5 117-81-7
CAS No.
249-079-5 271-090-9 247-977-1 271-091-4 222-804-9 271-086-7 271-087-2 – – – 222-884-9 204-294-3 271-089-3
204-214-7 247-210-0 271-091-4 204-211-0
205-011-6 201-550-6 201-557-4 201-553-2 201-622-7 201-559-5 271-093-5 276-15-8
EINECS No.
447.7 {432.7–474.7} 530.8 {506.8–544.8}
418.6 {362.6–474.7}
446.7 {432.7–446.7}
418.6 {418.6–432.6}
390.6
390.6 404.6 {334–447}
363
194.2 222.2 278.4 278.4 312.4 334.4
Molecular Weight
–9 –37
<–50
–46
–48
–47
–25 –4
–45
5.5 –40 –35 –58 –35 –27.4
Melting Point (°C)
0.96 0.953
0.97
0.961
0.97
0.986
0.978 0.97
1.00
1.192 1.118 1.042 1.050 1.111 1.011
Specific Gravity (20 °C)
4 M.K. Stanley et al.
Introduction
5
The principle phthalate plasticizer application in conjunction with polymers is, by far, the plasticization of vinyl resins. Plasticized vinyl is unique because a wide range of resin/plasticizer ratios permits production of plastics that range from very hard to very flexible. The phthalate plasticizer is usually chosen to produce the desired performance characteristics with the greatest possible economy of material costs and the greatest ease and speed of processing [1]. High molecular weight phthalates, C6 and above, are generally used as vinyl plasticizers. Low molecular weight phthalates have a diverse set of uses. The largest use of DMP is as a stabilizing diluent for the shipping and storage of organic peroxides. The main use of DEP is in the compounding of cellulosic films. Cosmetic-grade DEP acts as a fixative or carrier for perfumes and fragrances. There is also a pharmaceutical grade used in time-released preparations. DBP is consumed primarily in vinyl acetate emulsion adhesives and in cellulose lacquers. BBP is normally used with other general-purpose plasticizers in PVC applications to improve the performance of the final product and to aid in its processing.
6 Disposal and Releases into the Environment The releases of phthalates to the environment are discussed in the chapter on Multimedia Modeling. Very little phthalate ester is released to the environment during the manufacturing process. Very little is released into the air. Essentially all of the phthalate released during production and processing is disposed of in wastewater that is treated in wastewater treatment plants where it is either biodegraded or adsorbed to sludge with very little going to air. The fate of phthalates during wastewater treatment is covered in more detail in the chapter on Environmental Degradation. Phthalate adsorbed to sewage sludge is usually either incinerated or landfilled in the United States. In some countries, application of sewage sludge to agricultural land is also common. The latter two disposal routes will result in some release of phthalates to soil. Degradation and fate of phthalates in soil is also treated in subsequent chapters. Poorly operating incinerators may also lead to some phthalate emission to air. The major portion of phthalate esters that are found in the environment are the result of the slow releases of phthalates from plastics and other phthalate containing articles due to weathering. As stated earlier, phthalates within such articles are not covalently bound and under conditions of high surface exposure and warm temperatures, as in the case of exterior building materials, phthalate esters can diffuse from the solid surface into the air, despite their rather low vapor pressures. Thus, phthalate-containing consumer items during their useful lifetime may continue to be a source of phthalate esters to the atmosphere. Burial of phthalate ester containing articles in landfills will preclude further emissions to the air.As discussed in the Degradation chapter, phthalate esters contained within plastics buried in soil are degraded at the surface of the plastic by molds and bacteria. Phthalate diesters themselves show poor mobility in soil but aqueous leachates from landfills may contain trace amounts of more soluble products of phthalate degradation. The overall concentration, distribution and disposition of
6
M.K. Stanley et al.
typical phthalate esters in the environment are topics covered in the Multimedia Modeling chapter.
7 Regulations and Phthalate Esters Because of their very large production volumes, phthalate esters are subject to considerable regulatory scrutiny world-wide. Regulations on phthalate esters cover all aspects of their production, transportation, use, and disposal. Phthalates are regulated under the Clean Water Act, so that at certain manufacturing facilities in the US, wastewater to be treated in municipal sewage treatment plants may be required to undergo pretreatment prior to leaving the facility (Pretreatment Standards).When they become waste products, certain phthalates are subject to Resource Conservation and Recovery Act (RCRA) requirements. Drinking water standards have been set under the Safe Drinking Water Act for several DEHP. Releases to the environment of several phthalate esters are required to be publicly reported in the US, Canada, and Japan. Like all commercial chemicals, phthalate esters are subject to regulation under the U.S. Toxic Substance Control Act. Pursuant to an enforceable consent agreement under that Act, the industry has generated considerable data on phthalate physical-chemical properties, environmental fate characteristics that govern distribution and concentration of phthalate esters, and toxicity studies. In addition, the industry has voluntarily conducted many additional such studies. Phthalate esters have undergone comprehensive risk assessments regarding virtually all aspects of environmental and human health under “existing substances” regulations in the US, Canada, the EU, and at the Organization for Economic Cooperation and Development (OECD) level. These risk assessments have used the extensive data that has been generated for phthalate esters. The results of the various risk assessments completed to date have led to varying conclusions ranging from no further information needed and/or no need for further restrictions on use, to proposed requirements for some use-specific risk reduction measures. Some of the recent reviews include cancer classifications for DEHP and BBP by the International Agency for Research on Cancer, reviews of seven phthalate esters by an Expert Panel of the U.S. National Toxicology Program Center for the Evaluation for Risk to Human Reproduction, and review of DINP in children’s toys by the U.S. Consumer Product Safety Commission Chronic Hazard Advisory Panel on Diisononyl Phthalate. The European Union is nearing completion of comprehensive risk assessments for DBP, BBP, DEHP, DINP and DIDP. The commercial and regulatory interest in the ortho phthalate esters has resulted in an extensive list of health and environmental information on these substances. The remaining chapters in this review organize and present this information. Physical-chemical data are provided for all of the commercially important phthalate esters. The fate and distribution of phthalate esters in the environment, particularly in aquatic systems, is described. The effects of phthalate esters on plants and aquatic animals are discussed as well as its accu-
Introduction
7
mulation in an aquatic food web. Finally, human exposure to phthalate esters and their effects on human health are reviewed. All of this information provides an excellent data set for the continued use of these chemicals in a safe and responsible manner.
8 References 1. Use Category Document – Plastics Additive; Building Research Establishment Ltd., June 1998, pp 44–45 2. K.N. Woodward (ed) (1988) Phthalate Esters: Toxicity and Metabolism, vol I; CRC Press 3. Use Category Document – Plastics Additive; Building Research Establishment Ltd., June 1998, pp 44–45 4. Chemical Economics Handbook – SRI International, January 2000
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 9– 56 DOI 10.1007/b11461
Analytical Methods Review Frank David 1 · Pat Sandra 2 · Bart Tienpont 2 · Freddy Vanwalleghem 3 Michael Ikonomou 4 1
Research Institute for Chromatography, Pres. Kennedypark 20, 8500 Kortrijk, Belgium E-mail:
[email protected] 2 Laboratory of Organic Chemistry, University of Gent, Krijgslaan 281 S4, 9000 Gent, Belgium 3 Proviron, NV, Stationstraat 123, 8400 Oostende, Belgium 4 Contaminants Science Section, Institute for Ocean Sciences, BC, Canada
In this chapter, an overview is given of the analytical methods developed for the determination of phthalates in different environmental samples, including water, soil, sediments, sludges, air and biota. The analytical methods based on gas chromatography coupled to mass spectroscopy (GC-MS) and liquid chromatography coupled to mass spectroscopy (LC-MS) are described and typical applications are presented. Dedicated sample preparation methods are described and compared. Special attention is paid to contamination problems. Keywords. Phthalates, GC-MS, LC-MS, Contamination, Sample preparation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1
Introduction
2
Analysis of Phthalates
2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.2 2.3 2.3.1 2.3.2
Phthalate Analysis by Capillary Gas Chromatography Injection . . . . . . . . . . . . . . . . . . . . . . . . . Detection . . . . . . . . . . . . . . . . . . . . . . . . Mixed Isomer Phthalates . . . . . . . . . . . . . . . . Chemical Ionisation Mass Spectrometry . . . . . . . Quantification . . . . . . . . . . . . . . . . . . . . . Phthalate Analysis by HPLC-MS . . . . . . . . . . . . Analysis of Phthalic Acid Mono-Esters . . . . . . . . Derivatisation and GC-MS . . . . . . . . . . . . . . . Analysis of Phthalic Acid Mono-Esters by HPLC-MS .
3
The Blank Problem . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.1 3.2 3.3
Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Chromatographic Analysis . . . . . . . . . . . . . . . . . . . . . . 27
4
Analysis of Phthalates in Water Samples
4.1 4.2 4.3
Liquid-Liquid Extraction (LLE) . . . . . . . . . . . . . . . . . . . 28 Solid-Phase Extraction (SPE) . . . . . . . . . . . . . . . . . . . . 29 Solventless Extraction Methods (SPME and SBSE) . . . . . . . . . 31
. . . . . . . . . . . . . . . . . . . . . . . . 12 . . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
12 14 15 17 17 19 19 22 22 23
. . . . . . . . . . . . . . 28
© Springer-Verlag Berlin Heidelberg 2003
10
F. David et al.
5
Analysis of Phthalates in Sediments, Soils and Sewage Sludges . . 33
5.1 5.2 5.3
Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Clean-Up Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 35 Determination of Phthalates in Sewage Sludge . . . . . . . . . . . 35
6
Analysis of Phthalates in Air . . . . . . . . . . . . . . . . . . . . . 35
6.1 6.2 6.3 6.4
Introduction . . . . . . . . . . . . . . . . . . . Analysis of Total Phthalate Concentrations in Air Passive Sorptive Sampling of Phthalates in Air . Dust Analysis . . . . . . . . . . . . . . . . . . .
7
Analysis of Phthalates in Biota (Vegetation, Milk, Fish) . . . . . . 46
7.1 7.2
7.4
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical Procedure for the Determination of Phthalates in Vegetation . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical Procedure for the Determination of Phthalates in Milk or Edible Oils and Fat . . . . . . . . . . . . . . . . . . . . . . . . Analytical Procedure for the Determination of Phthalates in Fish
8
Sample Preparation Methods for Phthalic Acid Mono-Esters . . . 50
9
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.3
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
35 37 41 42
. 46 . 47 . 48 . 49
1 Introduction Phthalic acid diesters (PAEs) are widely spread in the environment. During the past years, interest in monitoring this class of industrial chemicals has increased. Several methods have been developed for their determination in a very broad range of matrices, including water (drinking water, surface water, waste water), soil, sediment, sludge, dust, indoor and outdoor air, and biota (vegetation, milk, fish, etc.). In this chapter an overview is given of different techniques used for the analysis of phthalates in these matrices. Particular attention is paid to the determination of the most relevant phthalates in terms of their environmental distribution. These phthalates include diisobutyl phthalate (DiBP), dibutyl phthalate (DBP), butylbenzyl phthalate (BBzP), bis(2-ethylhexyl) phthalate (DEHP), diisononyl phthalate (DiNP) and diisodecyl phthalate (DiDP). Although the last two phthalates have industrial importance, they are often not included in generic methods for environmental analysis, such as EPA or ISO methods [1, 2]. The main reason for this is that these phthalates are not pure chemicals, but mixtures of several isomers. Consequently, the analysis is more difficult. Special attention will be paid to their analysis and alternative methods will be discussed. An overview of the most important phthalates, including molecular formula, abbreviation, molecular weight and CAS number, is given in Table 1.
Table 1. List of phthalates
Phthalates
Formula
Abbreviation
Molar mass
CAS No
C10H10O4 C12H14O4 C14H14O4 C14H18O4
DM b DE b DalP DiPP
194.2 222.4 246.3 250.3
131-11-3 84-66-2 131-17-9 605-45-8
C14H18O4 C16H22O4 C16H22O4 C16H22O4
DPP BMPP DBP b DiBP b
250.3 278.4 278.4 278.4
131-16-8 17851-53-5 84-74-2 84-69-5
C16H22O6 C18H22O4 C18H24O4 C18H26O4
DeoEP DCPeP BCHP DMBP
310.4 302.4 304.4 306.4
605-54-9 18699-38-2 84-64-0 605-50-5
C18H26O4 C19H20O4 C20H14O4 C20H26O4 C20H30O4 C20H30O4 C20H30O4 C20H30O4 C22H18O4 C22H30O4 C22H34O4 C22H34O4 C22H34O4 C23H28O4 C24H38O4 C24H38O4 C24H38O4 C26H26O4 C26H42O4 C26H42O4 C26H42O4 C28H46O4 C28H46O4 C30H50O4 C32H54O4
DPeP BBzP b DPhP DCHP BEHP BOP DEBP DHP DBzP DMCHP BDcP DHpP DMHP BzEHP DEHP b DOP HDcP ODcP DMOP DNP DTMHP DDcP DMNP DUP DDDP
306.4 312.4 318.3 330.4 334.5 334.5 334.5 334.5 346.3 358.5 362.6 362.5 362.5 368.6 390.6 390.6 390.6 418.6 418.6 418.6 418.6 446.7 446.7 474.7 502.8
131-18-0 85-68-7 84-62-8 84-61-7 85-69-8 84-78-6 7299-89-0 84-75-3 523-31-9 27987-25-3 89-19-0 3648-21-3 41451-28-9 18750-05-5 117-81-7 117-84-0 25724-58-7 119-07-3 28553-12-0 84-76-4 4628-60-8 84-77-5 89-16-17 3648-20-2 2438-90-8
Isomeric mixture phthalates Diisoheptyl phthalate Diisononyl phthalate b Diisodecyl phthalate b
C22H34O4 C26H42O4 C28H46O4
DiHpP DiNP b DiDP b
362.5 418.6 446.7
Phthalic acid monoesters Monomethyl phthalate Monoethyl phthalate Monobutyl phthalate Mono(2-ethylhexyl) phthalate
C9H8O4 C10H10O4 C12H14O4 C16H22O4
MMP MEP MBP MEHP
180.2 194.2 222.2 278.4
Single isomer phthalates Dimethyl phthalate b Diethyl phthalate b Diallyl phthalate Di(1-methylethyl) phthalate (diisopropyl phthalate) Dipropyl phthalate Butyl-2-methylpropyl phthalate Dibutyl phthalate b Di(2-methylpropyl) phthalate (diisobutyl phthalate) b Di(2-ethoxyethyl) phthalate Dicyclopentyl phthalate Butylcyclohexyl phthalate Di(3-methylbutyl) phthalate (diisopentyl phthalate) Dipentyl phthalate Butylbenzyl phthalate b Diphenyl phthalate Dicyclohexyl phthalate Butyl 2-ethylhexyl phthalate Butyloctyl phthalate Di(2-ethylbutyl) phthalate Dihexyl phtalate Dibenzyl phthalate Dimethyl cyclohexyl phthalate Butyldecyl phthalate Diheptyl phthalate Di(5-methylhexyl) phthalate a Benzyl 2-ethylhexyl phthalate Di(2-ethylhexyl) phthalate b Dioctyl phthalate Hexyldecyl phthalate Octyldecyl phthalate Di(7-methyloctyl) phthalate a Dinonyl phthalate Di(3,3,5-trimethylhexyl) phthalate Didecyl phthalate Di(8-methylnonyl) phthalate a Diundecyl phthalate Didodecyl phthalate
a b
4376-18-5 2306-33-4 131-70-4 4376-20-9
These compounds are present in the isomeric mixtures of DiHpP, DiNP and DiDP respectively. These are the most important phthalates.
12
F. David et al.
The analysis of phthalic acid esters is mostly performed by gas chromatography (GC). Phthalates ranging from the most volatile dimethyl phthalate (DMP) to didodecyl phthalate can be analysed by capillary gas chromatography (CGC) as they are sufficiently volatile and thermostable. High-pressure liquid chromatography (HPLC) can be used as an alternative technique and is especially useful for the analysis of isomeric mixtures. The possibilities of CGC and HPLC for the analysis of phthalates are discussed below. Recent environmental and toxicological concerns have also focused on the determination of phthalate metabolites. The primary metabolites of the diesters are the monoesters, formed by the hydrolysis of one ester function to give phthalic acid monoesters. These monoesters contain a free acid function and require derivatisation before GC analysis. These methods are also presented. The major problem in phthalate analysis is, however, not the analysis itself, but the risk of contamination. Contamination can occur in every stage of the whole analytical procedure including sampling, sample preparation (extraction, cleanup, concentration) and analysis. An overview is given below of possible sources of contamination (instrument, solvents, air etc.) and recommendations are given on how to avoid or control the contamination problem. Finally, an overview is given of different sample preparation techniques that are used to extract and isolate phthalates from different environmental matrices.
2 Analysis of Phthalates 2.1 Phthalate Analysis by Capillary Gas Chromatography
Capillary gas chromatography is the most widely used analytical technique for the determination of phthalates. Different capillary columns, inlet systems and detectors are used. For the gas chromatographic separation, capillary columns coated with apolar stationary phases (polydimethylsiloxanes or polymethylphenylsiloxanes) are preferred, since they provide sufficient resolution, a higher maximum operating temperature and a lower bleeding than columns coated with polar stationary phases such as poly(ethylene glycol)s (Wax- columns) or cyanopropyl stationary phases. An example of the separation of the most important phthalates is given in Fig. 1. Peak identification is given in Table 2. This separation is performed on a 30 m ¥ 0.25 mm i.d. capillary column coated with a 0.25 µm film of 5% phenyl-95% methyl polysiloxane (HP-5MS). The analytical conditions are summarised in Table 3. These conditions are typical for the analysis of phthalates by CGC-MS and result in sufficient resolution of the most important phthalates. These conditions are also a good compromise between resolution and speed of analysis. Longer columns or slower temperature programs can give higher resolution, but for most environmental applications this is not needed. As demonstrated in Fig. 1, a boiling point separation is obtained on the apolar column.Around 22–24 min and at the end of the chromatogram, two unresolved groups of peaks are observed. These groups of peaks result from the
Time
Fig. 1. GC-MS chromatogram of a phthalate standard mixture (identification: see Table 2, conditions: see Table 3)
13
Analytical Methods Review
mixed isomer phthalates diisoheptyl phthalate, diisononyl phthalate and diisodecyl phthalate. Long capillary columns and slow temperature programs can give better resolution, but no column is able to fully separate all isomers. Therefore, a rather fast oven program is preferred, since this results in compression of the groups of C7-, C9- and C10-isomers and a better signal-to-noise ratio. The sensitivity is therefore higher than with slow temperature programs, in which broader isomer distribution patterns are obtained.
14
F. David et al.
Table 2. Peak identification for Fig. 1 and ions monitored in GC-MS analysis a
Peak Nr (Fig. 1) Phthalate
Abbreviation
Target ion Qualifier ion
1 2 3 4 5a, 5b, 5c 6 7 8 9 10 11
DMP DEP DiBP DBP DPeP BBzP DiHpP DCHP DEHP DiNP DiDP
163 149 149 149 149 149 149 149 149 149 149
194 177 223 223 237 91, 206 265 167, 249 279 293 307
D4-DBP D4-DOP DalP
153 153 149
– – 189
Dimethyl phthalate Diethyl phthalate Diisobutyl phthalate Dibutyl phthalate Diisopentyl phthalate (isomers) Butylbenzyl phthalate Diisoheptyl phthalate Dicyclohexyl phthalate Di(2-ethylhexyl) phthalate Diisononyl phthalate Diisodecyl phthalate
Internal standards Not included D4-ring-dibutyl phthalate Not included D4-ring-dioctyl phthalate Not included Diallyl phthalate a
For high-resolution mass spectroscopy, the monitored masses are 149.0239 (phthalates) and 153.0490 for the d4-labelled phthalates [3, 4].
Table 3. Analytical conditions for GC-EIMS
Instrument Column Carrier gas Injection Oven temperature Detection
Agilent 6890 GC – Agilent 5973 MSD 30 m ¥ 0.25 mm i.d.¥ 0.25 µm HP-5MS (5% phenyl–95% methyl silicone) 1 mL min–1 helium, constant flow (50 kPa at 50 °C) 1 µL splitless, 280 °C, 0.75 min purge delay 50 °C–1 min–10 °C min–1 –320 °C–2 min MS in SIM mode, Monitored ions: see Table 2
2.1.1 Injection
For sample introduction, cool on-column injection, whereby the liquid sample is directly introduced into the capillary column, is the best choice in terms of solute degradation and discrimination. Moreover, in the case of phthalate analysis, cool on-column injection also results in the lowest possible instrumental phthalate contamination. The major limitation of cool on-column injection is, however, the contamination of the capillary column with non-volatile material. This nonvolatile material can be present in extracts from complex samples, such as wastewater, sludges, food or biota. In practice, cool on-column injection is therefore restricted to the analysis of clean samples, such as extracts from drinking water (or surface water). For routine analyses of other sample extracts, splitless injection
Analytical Methods Review
15
is the method of choice. Modern split/splitless inlets in combination with new autosamplers minimise sample degradation and discrimination. Splitless injection can be used for the injection of contaminated sample extracts, since most of the non-volatile contaminants remains in the liner. Recently, also programmed temperature vaporising (PTV) inlets have been used. PTV inlets can be compared to split/splitless inlets, but injection is performed at a relatively low temperature (around the boiling point of solvent), followed by a rapid heating (typically 10 °C s–1) to 250–300 °C for the vaporisation of the sample. The main advantage of PTV injection is the possibility of performing large volume injection. Injections up to 1000 µL can be performed by using the solvent vent mode. During sample introduction and an initial hold time at low temperature, the solvent is evaporated and vented via the split vent (at a high split ratio).After solvent evacuation, the split vent is closed (splitless mode) and the inlet temperature is increased to vaporise the solutes. 2.1.2 Detection
Detection of phthalates can be done by flame ionisation detection (FID), electron capture detection (ECD) or mass spectrometry (MS). GC-FID is not frequently used, since the detector is not specific for phthalates. Some official methods (US EPA methods 606 and 8060) describe the use of ECD for phthalate analysis. Although ECD detectors are relative sensitive for phthalates, the specificity is restricted, since ECD responds much more sensitively towards halogenated compounds. The most important detector for phthalate analysis is mass spectrometric detection. All types of MS analysers, including quadrupole analysers, triple quadrupole analysers, ion traps and magnetic sector instruments have been used. Benchtop quadrupole systems are generally preferred due to their robustness, stability, linear dynamic range and low cost. Ion trap detectors have similar sensitivity, but the dynamic range is often lower and at higher solute concentration, as is often encountered in phthalate analysis, spectral deviations have been observed. High-resolution magnetic sector instruments are also used [3, 4], but the high resolution is not required for most applications. For most applications, a benchtop single quadrupole system provides sufficient sensitivity in selected ion monitoring (SIM) mode in electron impact (EI) ionisation. Typical sensitivities obtained on state-of-the-art instruments are in the low picogram (1–10 pg) range for the single isomer phthalates. Due to the presence of several isomers, the sensitivity for isomeric mixture phthalates (DiHpP, DiNP, DiDP) is typically one order of magnitude lower (10–50 pg). The mass spectra for all phthalates (except dimethyl phthalate) are very similar. The mass spectrum of DEHP is shown in Fig. 2. The main ion is at m/z 149, resulting from fragmentation with loss of the alkyl ester groups and a furan ring formation (Fig. 3). Besides the most abundant ion at m/z 149, the spectrum is rather poor. The molecular ion (m/z 390) is not detected. The second most important ion is at m/z 279. This ion results from fragmentation by loss of one alkyl group. This ion is therefore different for each phthalate and can be used to dif-
Fig. 2. Electron impact mass spectrum of DEHP
16 F. David et al.
Analytical Methods Review
17
Fig. 3. Mass fragmentation of DEHP
ferentiate phthalates. Further fragmentation of the ion at m/z 279 results in ion at m/z 167. A similar fragmentation pattern is found for the other phthalates, except for dimethyl phthalate. In the mass spectrum of dimethyl phthalate, the molecular ion is detected at m/z 194. The most abundant ion is at m/z 163, corresponding to the loss of a methoxy group (M–31). A list of the most important ions that can be used for mass spectroscopic detection of phthalates is included in Table 2. 2.1.3 Mixed Isomer Phthalates
A specific problem is the quantitative determination of the mixed isomer phthalates. No capillary column is able to separate all isomers, or to separate all C9-isomers from the C10-isomers. Therefore, extracted ion chromatograms on the specific ions m/z 293 (DiNP) and m/z 307 (DiDP) can be used. This is illustrated in Fig. 4. Although the groups of peaks overlap, the extracted ion chromatograms can differentiate between both compounds, and the sum of the peak areas measured on each extracted ion trace can be used for quantification. 2.1.4 Chemical Ionisation Mass Spectrometry
As an alternative to electron impact ionisation, phthalates can also be analysed by CGC-MS by using chemical ionisation [5]. Positive ion chemical ionisation results in a softer ionisation and therefore the fragmentation is reduced and the relative abundance of high-mass ions is increased. This is illustrated for dibutyl phthalate in Fig. 5. With methane as the reagent gas, the fragmentation is still strong and the most abundant ion is still at m/z 149. However, now a pseudo-molecular ion at M+1 (m/z 279) is clearly detected.With ammonia as the reagent gas, the possibilities of positive ion chemical ionisation are clearly demonstrated. Now
18
F. David et al.
Abundance
Time Æ Fig. 4. Extracted ion chromatograms for DEHP (m/z 279), DiNP (m/z 293) and DiDP (m/z 307)
Fig. 5 A, B. PCI mass spectra of dibutyl phthalate (A methane reagent gas, B ammonia reagent
gas). Published with permission from Ref. [5]
Analytical Methods Review
19
the most abundant ion is the pseudo-molecular ion at m/z 279 (M+1). Positive chemical ionisation is therefore an interesting alternative, especially for the determination of the mixed isomer phthalates. 2.1.5 Quantification
The quantification of phthalates can be done by external or internal standard calibration. Internal standard calibration is preferred, since the internal standard can be used to correct for variations in injection, analysis (column performance) and condition of the mass spectrometer. The best internal standards for phthalate analysis by mass spectrometric detection are isotopically labelled phthalates. Both C13-labelled or ring D4-labelled (deuterated phthalates) are commercially available. If added before sample preparation, these internal standards can also be used to correct for extraction efficiency. The most important internal standards used in the analysis of phthalates are included in Table 2. 2.2 Phthalate Analysis by HPLC-MS
Although CGC-MS is widely used, LC-MS can provide interesting alternatives, especially in the analysis of the isomeric mixture phthalates and monoesters. An analytical method based on LC-electrospray ionisation mass spectrometry for the determination of phthalates in marine water, sediment and biota samples was presented by Lin et al. [6]. The LC-MS instrument is operated in the positive ion monitoring mode. The separation was achieved on a 25 cm ¥ 2 mm i.d. C8 column with a mobile phase of 90% methanol and 10% 0.5 mM sodium acetate. The mobile phase flow rate was 0.22 mL min–1 and a 1/10 flow split ratio (0.022 mL min–1 to MS) was used before the mass spectrometer (VG Quattro triple quadrupole). The C8 (octyl-silicagel) column is preferred because the isomeric mixtures elute as single, well-defined and narrower peaks than on a classical C18 (octadecyl silicagel, ODS) column. Phthalate esters are monitored as sodiated molecular adducts (M+Na, at m/z MW+23) under SIM conditions. The stable isotope dilution method was used for quantification. Compared to GC-MS analysis, a lower sensitivity is obtained, but the LC-MS approach was found to give the following advantages: superior selectivity with molecular weight information for the isomeric mixtures, more reliable quantification of the phthalate ester isomeric mixtures (such as DiHpP, DiNP and DiDP), simpler clean-up procedures and shorter analysis times. An example of a separation of mixed isomer phthalates by LC-MS is given in Fig. 6. This separation was obtained under slightly different conditions [7]. The analytical conditions are summarised in Table 4. The separation of the isomeric mixtures obtained on the cyanopropyl silicagel (CN) column is very good. On this column, the isomeric mixtures are also separated as well-defined peaks, allowing good quantification. In comparison to GC-MS, less resolution is obtained, but for quantification this is an advantage. The mass spectra obtained by electrospray ionisation MS for DEHP and DiDP are given in Fig. 7. The most abundant ions
Fig. 6. LC-MS analysis of isomeric mixture phthalates by electrospray ionisation. (Ions monitored: m/z 363 for DiHpP, m/z 391 for DEHP, m/z 419 for DiNP and m/z 447 for DiDP) (conditions: see Table 4)
20 F. David et al.
21
Analytical Methods Review Table 4. Analytical conditions for LC-ESMS analysis of isomeric mixture phthalates
Instrument Column Mobile Phase
Agilent 1100 LC–Agilent 1100 SL MSD 25 cm L¥4.6 mm i.d. ¥5 µm Alltima CN (cyanopropyl silica) A. 0.5% ammonium acetate in water B. methanol 60% B for 5 min, to 100% B at 35 min 1 mL min–1 10 µL Ionisation mode: API-ES (electrospray) Polarity: positive Monitored Ions: M+1 (363 (DiHP), 391 (DEHP), 394 (IS: d4-DEHP), 419 (DiNP), 447 (DiDP))
Gradient Flow Injection Detection
DEHP
DiDP
Fig. 7. Positive ion electrospray spectra obtained for DEHP and DiDP (conditions Table 4)
22
F. David et al.
are respectively ions at m/z 391 and m/z 447, corresponding to the M+H adducts (at m/z M+1). In addition, the acetate-adducts are detected (as M+59, respectively at m/z 449 and m/z 505). In this method, ammonium acetate is used as buffer in the mobile phase. This analysis was performed on an Agilent 1100 LCMSD using an orthogonal electrospray interface. No effluent splitting was used and the entire column flow (1 mL min–1) is introduced into the mass spectrometer. Since contamination of the interface and ion source is reduced by using a volatile buffer, an ammonium acetate buffer is preferred in this method over a sodium salt buffer. In addition to LC-MS, LC-MS/MS can also be used. By monitoring daughter ions from a secondary fragmentation of an isolated parent ion, an additional level of specificity is added. LC-MS/MS is especially suitable in the determination of phthalates in complex matrices such as biota extracts [6]. 2.3 Analysis of Phthalic Acid Mono-Esters
Phthalic acid mono-esters are the primary metabolites of phthalates. In addition to the analysis of the phthalic ester diesters, recent analytical work also focuses on the analysis of the monoesters. The determination of monoethylhexyl phthalate (MEHP) was first described in biological samples (mainly plasma), as it was observed that during blood transfusion DEHP was leached out of PVC blood storage bags and tubing for haemodialysis and extracorporeal blood circulation [8–10]. Only later have methods for the determination of monoesters in other biomatrices been described. More recently, methods have been developed for the analysis of the monoesters in environmental samples such as water, soil, and biota. 2.3.1 Derivatisation and GC-MS
In comparison to the analysis of the phthalic acid diesters, the analysis of the monoesters is more difficult. The free fatty acid group makes the compound more polar and this leads to adsorption during gas chromatographic analysis. For the analysis of trace levels of monoesters, it is therefore not possible to obtain quantitative data without blocking the acidic function prior to gas chromatographic analysis. Derivatisation of the acid (–COOH) function can be done by esterification (–COOR) or by silylation (–COOSiR3). In selecting an appropriate derivatisation technique, it is important that possible hydrolysis and/or transesterification are avoided. Monoethylhexyl phthalate can for instance be derivatised by classical esterification methods used for fatty acids (BF3/methanol, KOH/methanol, H2SO4/methanol, etc.) but these methods will lead to further hydrolysis, with partial or complete formation of phthalic acid dimethyl ester (DMP). It should also be kept in mind that in real samples the monoesters are normally present in combination with the diesters. Some derivatisation reagents also lead to (partial) hydrolysis of the diesters and therefore lead to false positive results.
23
Analytical Methods Review
Sjöberg et al. [8–10] described the use of a mixture of pentafluoropropanol and pentafluoropropionic anhydride as derivatisation reagent. The analysis was performed by GC-MS under conditions similar to those described for phthalate analysis. Marschall and Egestad [11, 12] used silylation for the derivatisation of MEHP and conjugates prior to GC-MS analysis. Esterification by triethyloxonium tetrafluoroborate was used by Haam et al. [13] for the derivatisation of four metabolites, including MEHP, 5-carboxy-2-ethylpentyl phthalate, 2-ethyl-5-oxohexyl phthalate and 2-ethyl-5-hydroxyhexyl phthalate in urine extracts. Suzuki et al. [14] used diazomethane to prepare the methyl esters of the monoesters prior to GC-MS analysis. All these derivatisation techniques work well, but they must be performed in anhydrous conditions. First the monoesters have to be extracted from the matrix (blood, urine, water, suspended solids) and subsequently the extracts are dried under nitrogen. Then, the reagent is added and the derivatisation is performed under heating (pentafluoropropanol/pentafluoropropionic anhydride or silylation) or at room temperature (diazomethane). These derivatisation reactions cannot be performed in aqueous matrices. Recently, ethylchloroformate was used for the derivatisation (ethylation) of monoesters in aqueous media [15]. This derivatisation works in situ and allows subsequent extraction of the derivatised phthalates as described below. 2.3.2 Analysis of Phthalic Acid Mono-Esters by HPLC-MS
Alternatively, phthalic acid monoesters can be analysed without derivatisation by HPLC [12, 16–19]. HPLC in combination with mass spectroscopic detection is preferred especially for the analysis of isomeric mixtures of phthalic acid monoesters. For spectrometric detection, atmospheric pressure chemical ionisation (APCI) gives the highest sensitivity [18, 19].An example of a separation obtained by LC-APCI-MS for a standard mixture is given in Fig. 8. The analytical conditions are summarized in Table 5. The mass spectra obtained by APCI-MS for MEHP and monoisodecyl phthalate (MiDP) are given in Fig. 9. The most abunTable 5. Analytical conditions for LC-APCIMS analysis of phthalic acid monoesters
Instrument Column Mobile Phase Gradient Flow Injection Detection
Agilent 1100 LC–Agilent 1100 SL MSD 25 cm L ¥ 4.6 mm i.d.¥ 5 µm Alltima CN (cyanopropyl silica) A. 0.5% ammonium acetate in water B. methanol Isocratic, 50% A–50% B 0.5 mL min-1 50 µL Ionisation mode: APCI (atmospheric pressure chemical ionisation) Polarity: negative Fragmentor: 70 V Vaporizer: 325 °C; Drying gas: 5 L min–1 nitrogen Monitored Ions: M–1 (277 (DEHP), 281 (IS: d4-MEHP), 291 (MiNP), 305 (MiDP))
and m/z 305 for MiDP) (analytical conditions: see Table 5)
Fig. 8. LC-MS analysis of phthalic acid monoesters by atmospheric pressure chemical ionisation. (Ions monitored: m/z 277 for MEHP, m/z 291 for MiNP
24 F. David et al.
Analytical Methods Review
Fig. 9. Mass spectra obtained for MEHP and MiDP by APCI ionisation
25
26
F. David et al.
dant ions are detected at M–1 (m/z 277 for MEHP and m/z 305 for MiDP). In the spectrum of MiDP, a trace of monoisoundecyl phthalate is also detected (m/z 319). Blount et al. [18] also used LC-APCI-MS for the analysis of phthalate metabolites in urine extracts. The analyses were performed on a ThermoQuest TSQ 7000 triple quadrupole analyzer. The mass spectrometer was operated in MS/MS mode. The parent ions at M–1 were isolated and specific daughter ions obtained after collision-induced dissociation with argon were monitored. This method allows very selective detection of phthalic acid monoesters in complex matrices. Prior to the analysis, the monoesters need to be extracted from the sample matrix. Examples of sample preparation methods for monoesters are given below.
3 The Blank Problem The major problem in phthalate analysis is the contamination problem, resulting in false positive results or over-estimated concentrations. The risk of contamination is present in the whole analytical scheme, including sampling, sample preparation and chromatographic analysis. Due to the fact that phthalates are widely used, they are present in air, water, organic solvents, plastics and adsorbed on glass or other materials. Some typical sources of possible contamination are listed below and recommendations are given for minimising contamination. 3.1 Sampling
Sampling is the critical first step. The sampling of liquids and solids is preferably done in glass containers. All plastic materials should be avoided. Although some materials do not contain phthalates as additives, phthalates might be adsorbed on the surface and cleaning can be difficult. Specially cleaned glass containers can be purchased (e.g. I-Chem sample bottles). Otherwise glass containers should be rinsed with solvents and dried at 400 °C. Containers should not be left open, since they can adsorb phthalates from laboratory air onto the wall surface. Also, stoppers for bottles or container lids can contain phthalates. The stoppers or lids should also be cleaned or blank checked. During sampling, contact between the sample and hands or plastic gloves should be avoided. Metal spatulas are preferred over plastic materials.After sampling, the containers should be closed. Storage of liquid samples can be done at 4 °C. Biota samples or soils and sediments are stored at –20 °C. Since phthalates undergo biodegradation, the storage of aqueous samples at 4 °C should be not longer than 4 days. Chemical preservation may be performed by addition of 500 mg sodium azide per litre sample [20]. A detailed description of sampling methods for water and sediments is given by Parkman and Remberger [21] and by Braaten et al. [22].
Analytical Methods Review
27
3.2 Sample Preparation
The most important rule is that the risk of contamination is reduced if the sample preparation is kept to a minimum, with minimal extraction steps, minimal extract concentration and minimal glassware use. For the different sample preparation techniques, the risks of contamination associated with each method are discussed. In general, glassware and solvents are the most important sources of contamination. Glassware can be cleaned by solvent rinsing and thermal treatment at 400 °C for 1–2 h [20]. After cooling, the glassware should be stored in a closed container or wrapped in aluminium foil to avoid adsorption of phthalates from the air. Prior to use, the glassware should be rinsed with a small portion of blank-tested organic solvent (cyclohexane or isooctane) to deactivate the surface. Due to the thermal treatment at 400 °C, the glass surface is more active and phthalates can be (irreversibly) adsorbed. Organic solvents and laboratory grade water also contain traces of phthalates. Even commercially available solvents for trace analysis (e.g. pesticide analysis grade) can contain ppb (µg L–1) levels of phthalates. Some laboratories use inhouse-distilled solvents, but this is not always possible in routine analysis and contamination of the solvents during and after distillation is still possible. Due to the ppb level of phthalates in solvents, concentration of the extracts should be minimised, since the concentration of phthalates will increase proportionally. If for instance liquid-liquid extraction is used and a 1 L sample is extracted three times with 40 mL solvent, and the extract is concentrated to 1 mL, a 1 µg L–1 contamination of phthalates in the extraction solvent will result in a 0.12 ppb background value. The limit of quantification can therefore not be lower than 0.12 ppb.Also, reagents need to be checked.Anhydrous sodium sulfate can, for instance, contain up to 0.5 mg kg–1 of DEHP. Contamination of glassware and solvents is also likely to occur due to laboratory air. Phthalate concentrations in laboratory air are often in the 100– 500 ng m–3 range and the phthalates can be concentrated in open solvent bottles or adsorbed on glassware. Glassware can be protected by aluminium foil, while solvent bottles should remain closed until use. As far as possible, plastic materials should be removed from the laboratory. Cleaning agents used in the laboratory (floor or furniture cleaning, glassware cleaning, etc.) may contain volatile phthalates (especially DEP, DiBP and DBP) and may severely contaminate the laboratory air. Finally, it should be noted that personal hygiene materials (soaps, handcreams, cosmetics) often contain phthalates (mainly DMP and DEP). 3.3 Chromatographic Analysis
Phthalates can be present in the chromatographic system. The most important contamination is located in the inlet and gas supply system. Split/splitless inlets may contain septa, liners and O-rings that are contaminated with phthalates. During splitless injection, a large solvent vapour cloud is created and these solvent
28
F. David et al.
vapours can escape from the liner and perform phthalate extraction from septa and O-rings. This problem is less important with cool on-column or PTV inlets, since the temperature at which injection is made is lower and no large vapour volumes are created.When using splitless injection, care should be taken that the volume of the sample vapour does not exceed the liner volume. If splitless injection is used, the quality of the septum should also be evaluated. Some septa contain phthalates and these should be avoided. Septumless inlets can be an alternative (e.g. Merlin Microseal). Another critical factor is the quality of caps for autosampler vials. These caps can also contain phthalates. The best solution is never to inject twice from the same vial. Once a vial cap has been punctured, phthalates leach out into the sample (in the organic solvent). It can be observed that the phthalate concentration in an extract will increase with the number of injections made from the same vial. In case special sample introduction systems such as thermal desorption are used, these systems also need to be checked for blanks. Cryogenic traps will concentrate phthalates and even very low concentrations in blank desorption tubes, in gas lines, etc. will lead to false positive results.
4 Analysis of Phthalates in Water Samples For the determination of phthalates in water samples, published methods can be divided in two classes: methods based on liquid-liquid extraction (LLE) and methods based on solid-phase extraction (SPE). Both methods can be used successfully, but depending on the expected phthalate concentration and on the complexity of the matrix (drinking water versus wastewater) one technique is preferred over the other. Recent developments in analytical methods for aqueous matrices include micro liquid-liquid extraction (in combination with large volume injection), solid-phase micro-extraction (SPME) and stir bar sorptive extraction (SBSE). These techniques are briefly discussed below. 4.1 Liquid-Liquid Extraction (LLE)
The extraction of relatively large volumes of water samples (1–2 L) with an apolar non-miscible solvent is the most straightforward method for the extraction of phthalates from aqueous samples. Good extraction recoveries are obtained by using dichloromethane, cyclohexane or hexane [22]. The ratio of the volume of the water sample to the volume of the organic solvent should be smaller than 20. Higher ratios will still result in high recovery for the high-molecular weight phthalates (log Kow = 7.73 for DEHP [23]), while low-molecular weight phthalates are more water-soluble and consequently the recovery drops (log Kow = 1.61 for DMP). After extraction, the organic phase is dried (over anhydrous sodium sulfate) and concentrated. If complex-contaminated samples are extracted, several compounds such as hydrocarbons, detergents and plant material (sterols) are co-extracted. The extract can then be purified on an activated aluminium oxide column. A glass column or cartridge is packed with 1 g alu-
Analytical Methods Review
29
minium oxide (50–200 µm, neutral, activated for 4 h at 400 °C) and rinsed with 2 mL extraction solvent. The column is dried under nitrogen and the extract is eluted through the cartridge and collected in a tube for further concentration or analysis. In general, liquid-liquid extraction is limited due to the presence of trace levels of phthalates in commercially available solvents, even in solvents for trace analysis (pesticide residue analysis). Therefore the concentration factor is limited. A contamination of 1 ng L–1 in the solvent and a concentration factor of 50 (extraction with 50 mL solvent and concentration to 1 mL) leads to a 50 ng L–1 background value. Accurate determinations below 100 ng L–1 (0.1 ppb) are therefore questionable with this method. Vikelsoe et al. [3] reported a 90 ng L–1 detection limit and 770 ng L–1 background values for DEHP by a liquid-liquid extraction method applied to wastewater samples. Recently, micro liquid-liquid extraction techniques have been presented [24]. These techniques have the advantage that the solvent consumption is reduced and the whole extraction can be performed in a small (10–50 mL vial). Micro liquid-liquid extraction can be combined with large volume injection, resulting in maintained sensitivity. Extraction of 25 mL with 5 mL solvent and injection of 50 µL gives the same sensitivity as extraction of 250 mL with 50 mL solvent, concentration to 1 mL and a 1-µL injection. However, with micro-LLE and large volume injection, the purity of solvents is again a limiting factor. Liquid-liquid extraction offers the advantage that water samples with particles or heavily contaminated water samples (waste water) can be extracted. For the analysis of relatively pure water samples (drinking water, river water, surface water, seawater) solid-phase extraction is preferred. 4.2 Solid-Phase Extraction (SPE)
In comparison with liquid-liquid extraction, the amount of solvent can be reduced drastically by using solid-phase extraction. The water sample (100 mL to 2 L) is passed through an apolar sorbent and the phthalates are concentrated on the sorbent. The most widely used sorbent for enrichment of phthalates is octadecyl silicagel (ODS, C18). For elution, different solvents can be used (dichloromethane, ethyl acetate, etc.). Solid-phase extraction can be performed with classical cartridges containing 100–500 mg ODS material with an average particle size of 40 µm. These cartridges are commercially available from different suppliers. Most commercial cartridges are made from polyethylene or polypropylene barrels and this can cause relatively high background levels. For trace analysis, glass barrels are available that can be packed with ODS material. As an alternative to classical cartridges, solid-phase extraction disks can be used. These disks are composed from small particle (7 µm) ODS material that is incorporated in a Teflon fibril. The advantage of the disks is that the sampling flow is higher and that (very) large samples can be processed faster. A typical SPE sample preparation method is proposed as a draft ISO method [2]. The method is based on the work of Furtmann et al. [20]. Glass SPE cartridges are packed with 250 mg of RP C18 material. The cartridges are installed on a SPE
30
F. David et al.
vacuum manifold and conditioned with ethyl acetate (one bed volume) and then dried under nitrogen (for 10 s). Nitrogen is preferred over air to reduce contamination. Then, the cartridges are conditioned with methanol (two bed volumes). After this conditioning, the ODS bed is not allowed to run dry. A (300 mL) reservoir is connected on top of the cartridge and the sample (250 mL) is loaded in the reservoir. Another pre-treated ODS cartridge is installed above the reservoir to reduce contamination by air. With the aid of vacuum, the sample is eluted through the cartridge at a flow rate of about 2–10 mL min–1. After elution, the cartridge is dried with nitrogen for about 5 min. Finally, the phthalate fraction is eluted from the cartridge into a collection tube with 2 mL of internal standard solution (in ethyl acetate) by using vacuum. The collected fraction can be analysed directly or an additional clean-up can be used (see above). The advantage of solid-phase extraction is that a large concentration factor can be obtained without (or with limited) solvent concentration, and consequently without the concentration of contaminants in the organic solvent. If a 250 mL water sample is extracted and the final volume is 2 mL, a concentration factor of 125 is obtained without solvent concentration. Background levels below 10 ng L–1 and detection limits below 50 ng L–1 can be obtained. The major limitation of the SPE method is the extraction of water samples containing solids or heavily contaminated samples. If the amount of particulates is low, the SPE method can still be used if the particles do not block the cartridges. If the eluate is allowed to soak the particulates for a few minutes (after the eluting solvent is added), quantitative extraction is obtained. If a high amount of particles is present, it is advised that the sample is filtered first and the phthalates measured separately on the aqueous and the solid phases. If the sample is highly contaminated with other organic material (solvents, oil) or contains high amounts of detergents, recovery may be low do to incomplete enrichment on the C18 material. For these samples liquid-liquid extraction is advised. Several groups have used similar SPE techniques successfully. Van der Velde et al. [24] used 500 mg C18 Polar Plus cartridges for the extraction of 250 mL samples. The cartridges were eluted with 5 mL of a 70% pentane-30% methyl tert-butyl ether mixture. The reported recoveries were between 63% (DEHP) and 99% (BBzP). The detection limit for the single isomer phthalates was around 0.1 µg L–1. Letinski [25] used C18 extraction disks for the analysis of 3 L water samples. Elution was done by 10 mL of an ethyl acetate/dichloromethane mixture. The extract was concentrated to 0.5 mL. The detection limit for DiNP and DiDP was 0.1 µg L–1 and the recovery for DiNP and DiDP was 94% and 99%, respectively. Lin et al. [6] also used C18 extraction disks.Very high sample volumes (30 L) were pumped through the cartridge (47 mm C18 3 M Empore disk) at 30 mL min–1. After extraction, the disks were extracted by sonication, as described for sediment samples. In this way, very low concentrations (down to 5 ng L–1) could be measured in seawater samples. Normally extracts from water samples (drinking water, river water, seawater, rainwater) obtained by solid-phase extraction do not require additional clean-up. Only wastewater samples can require additional clean-up (on aluminium oxide, see above).
Analytical Methods Review
31
4.3 Solventless Extraction Methods (SPME and SBSE)
Recently solventless extraction methods have been described for the extraction of semi-volatile compounds, such as pesticides, PAHs and phthalates, from aqueous matrices. Pawliszyn [26] developed solid-phase micro-extraction (SPME). In this extraction technique, a fused silica fibre coated with a thin (7–100 µm) layer of polymer phase (usually polydimehylsiloxane) is immersed in the aqueous sample, while the sample is stirred. After a certain extraction time (varying from minutes to one hour), the coated fibre is retracted in a holder, transferred to the GC and desorbed in a hot inlet (for instance a splitless inlet at 250 °C). The sorbed compounds are desorbed again and injected into the column for separation and analysis. This simple technique has been applied to different environmental applications. However, the major limitation is the small amount of coating on the fibre. The very small volume of the coating (around 0.5 µL) results in a very high ratio between sample volume (typically 10–20 mL) and the extraction phase. Consequently, only solutes with very high (Kow > 4) octanol-water partition coefficients can be extracted with high recovery. Baltussen et al. [27] introduced magnetic stir bars coated with silicones as alternative solventless extraction method (stir bar sorptive extraction, SBSE). The stir bars are placed in 10–50 mL samples and are stirred on a magnetic stirrer for 30–120 min.After extraction, the stir bar is removed from the sample and placed in a thermal desorption unit for thermal desorption and on-line GC-MS analysis. The same methodology is used for air monitoring (see below). For semivolatile compounds, such as phthalates, a 10 mm-long stir bar coated with a 0.5 mm film of dimethylpolysiloxane (PDMS volume = 24 µL) is used. The amount of coating is 50 times higher and consequently higher recoveries are obtained for compounds with lower Kow values. Stir bar sorptive extraction has been used for various environmental samples and was also used for the extraction of semi-volatile compounds, including phthalates, in beverages. The possibilities of stir bar sorptive extraction, followed by thermal desorption-GC-MS are demonstrated in Fig. 10.A 10 mL rainwater sample was collected directly in a 20 mL vial. The PDMS-coated stir bar was introduced and extraction was performed over 2 h. The resulting chromatogram (extracted ion chromatogram at m/z 149) shows the presence of DiBP (16.01 min), DBP (18.39 min) and DEHP (29.59 min). The concentrations were 4.60 ppb (DiBP), 0.53 ppb (DBP) and 0.25 ppb (DEHP). These data clearly illustrate that transport through air is an important pathway of distribution of phthalates in the environment. This analysis also shows that very high sensitivities can be obtained by SBSE. In this application, the mass spectrometer was operated in scan mode and the signal-to-noise ratio for DEHP is higher than 40. The detection limit is thus lower than 25 ppt (ng L–1) in scan mode and lower than 5 ppt in SIM mode.
29.59 min = DEHP)
Fig. 10. Extracted ion chromatogram of stir bar sorptive extraction of phthalates from rainwater (16.01 min = DiBP, 18.39 min = DBP,
32 F. David et al.
Analytical Methods Review
33
5 Analysis of Phthalates in Sediments, Soils and Sewage Sludges In comparison with water samples, the determination of phthalates in solid (or semi-solid) samples usually requires two steps: an extraction step and a clean-up step. Solid samples include soils, sediments, sludges and solid waste. Soil samples are normally collected as grab samples by using a stainless steel drill, which allows sampling at different depths. Typical concentration levels for phthalates in soil are in the order of 10–1000 µg kg–1 dry mass. Sampling of sediments is used for the study of historical phthalate contamination and for the determination of local contamination and biodegradation. Sampling techniques that can be used for sediment sampling are described by Parkman and Remberger [21] and by Braaten et al. [22]. Concentration levels of phthalates largely depend on the sampling site. The average concentrations are of the same order of magnitude as the soil samples. Sludge samples from wastewater treatment are important samples to monitor input sources and to study biodegradation. Moreover, determination of phthalates in treated (digested and dried) sludge may be important. In at least one European country (Denmark) there is a defined maximum allowable level for DEHP in sludge to be used as an agricultural fertiliser. Phthalates entering the wastewater treatment plant via the wastewater influent are in general well removed from the aqueous stream in part by biodegradation and in part by removal in the sludge. Typical phthalate concentrations in treated sewage sludge are between 10–100 mg kg-1 dry mass. Dried sludges usually have a 20–30% dry mass content and can be analysed in a similar way to soils and sediments. Wet sludges (dry mass < 5%) can be analysed as such or after concentration of the solids and removal of the aqueous phase (centrifugation, filtering). Solid samples can be stored at –20 °C during several weeks. 5.1 Extraction
For the extraction of phthalates from solid (or semi-solid) matrices, various techniques have been used. Soxhlet extraction, whereby an organic solvent is heated and the condensed vapours are percolated through the solid samples held in a filter cartridge (thimble), is still considered as the reference method for the extraction of semi-volatile pollutants from solid environmental samples. Soxhlet extraction is a slow extraction procedure (4–24 h) and uses relatively large amounts of solvent (100–250 mL per sample). Due to the large solvent consumption, contamination is therefore a major issue with Soxhlet extraction. Soxhlet extraction has been used by Steffen and Lach [28]. Sediment samples were first freeze-dried. Then 30–50 g samples were extracted with 200 mL toluene over 8 h. For some samples, sulfur compounds were removed prior to GC-MS analysis by sonication of the extract in combination with copper powder. During the past years, automated and miniaturised extraction techniques are slowly replacing classical Soxhlet extraction. These techniques include automated Soxhlet extraction (Soxtec), shaking, ultrasonic extraction, microwave assisted
34
F. David et al.
extraction (MASE), accelerated solvent extraction (ASE) and supercritical fluid extraction (SFE).Automated miniaturized Soxhlet extraction (Soxtec, Soxtherm) is faster than Soxhlet extraction and solvent consumption is reduced. Further reduction of solvent consumption is possible by using supercritical fluid extraction. In SFE, pressurized carbon dioxide is brought to a pressure and temperature above its critical pressure (75 bar) and critical temperature (35 °C), resulting in a supercritical fluid. A supercritical fluid can be considered as a dense gas, with gas-like viscosity and flow characteristics and liquid-like solvating characteristics. Supercritical fluid extraction has been used successfully for a broad range of applications. SFE was used by Kolb et al. [29] for the extraction of phthalates from sewage sludge. The extraction was performed at 60 °C with carbon dioxide modified with 5% hexane as the extraction solvent and a three-step pressure program (5 min at 200 bar, 5 min at 300 bar and 20 min at 400 bar). The extraction efficiencies were higher than 90% for DBP, BBzP and DEHP, and were 84% for DiNP and 79% for DiDP. The standard deviation was 2–7%. Currently, supercritical fluid extraction is replaced by accelerated solvent extraction (ASE). In ASE, similar equipment is used, but the supercritical fluid is replaced by a classical organic solvent (dichloromethane, hexane, etc.). The solvent is pumped into an extraction vessel containing the sample. The solvent in the thimble is pressurized (up to 5000 psi) and heated (typically at 100°C). After extraction, the solvent is eluted in a vial. The extraction is completely automated and in comparison to Soxhlet extraction, solvent consumption is drastically reduced. Accelerated solvent extraction was used by Lettinski et al. [25]. Typically 15 g soil sample is mixed with anhydrous sodium sulfate (or Hydromatrix). The sample is placed in an extraction thimble and extracted with dichloromethane at 2000 psi and 100 °C for15 min. Recoveries of phthalates were 80–100%. The main problem was the relatively high background values that were typically around 36 µg kg–1 for DiBP, 24 µg kg–1 for DBP and 35 µg kg–1 for DEHP. Probably one of the most simple and cheap extraction methods is shaking or sonication. Vikelsoe et al. [4] used a simple shaking method for the determination of phthalates in soil samples. The samples were not dried before extraction. After addition of internal standard, a 50 g sample was extracted with 100 mL dichloromethane by shaking for 4 h. The blank values were around 6 µg kg–1 for DEHP. Low solvent consumption in combination with relatively short extraction times are also obtained by ultrasonic extraction. Braaten et al. [22] used sonication for the determination of phthalates in sediments by using 5 g wet sample. Firstly, 2 mL acetonitrile was added and a 10 min ultrasonic extraction was performed. Then, 2 mL acetonitrile and 3 mL hexane were added and the ultrasonic extraction was repeated for 30 min. The organic phase was isolated, washed with water and the hexane phase was cleaned on aluminium oxide. A similar method was applied by Parkman and Remberger [21] to the determination of phthalates in Swedish and Dutch sediments. Liu [6] used a triple sonication extraction of 2 g sample (mixed with 20 g sodium sulfate) with each time 20 mL 1 : 1 dichloromethane : hexane for the determination of phthalates in marine sediments. Paxeus [30] also used ultrasonic extraction of dried sludge with MTBE or a 1 : 1 mixture of hexane and acetone. Recoveries were higher than 90% by
Analytical Methods Review
35
using three 10 min extractions. The organic phase (upper layer) is removed after each extraction and replaced by fresh solvent. Zurmuhl [31] demonstrated that similar extraction recoveries are obtained by ultrasonic extraction in comparison to Soxhlet extraction. 5.2 Clean-Up Procedures
Since extracts of soils, sediments and sludges often contain other contaminants or co-extracted compounds such as plant sterols, clean-up is often needed. A simple clean-up method based on solid-phase extraction is described by Braaten et al. [22]. The extract from sediments are applied on a 500 mg neutral alumina column (activated at 400 °C, deactivated with 9% water). The column is rinsed with 3 mL hexane. The phthalate fraction is eluted with 3 mL of a 75% hexane/25% MTBE mixture. Liu [6] used 15 g alumina column (300 ¥10 mm i.d., alumina deactivated with 15% water) with 1–2 cm anhydrous sodium sulfate on top. After applying the sample, the column is eluted with 30 mL hexane (waste), then with 30 mL 10% dichloromethane in hexane (PCB fraction) and finally with 30 mL 50% dichloromethane in hexane. This last fraction contains the phthalates. In some cases, an additional clean-up on a 7.5 g Florisil column was used. The column was rinsed with 30 mL dichloromethane (waste) and with 30 mL 5% acetone in dichloromethane. The phthalates are present in this last fraction. The method blanks for a 2 g sediment sample were: 1.3 µg kg–1 DiBP, 28.4 µg kg–1 DBP and 24.9 µg kg–1 DEHP. The detection limits were thus determined by the method blanks for these compounds. The reported recoveries were higher than 70% for DMP, higher than 80% for DEP, DiNP and DiDP, and higher than 90% for DiBP, DBP, BBzP and DEHP. 5.3 Determination of Phthalates in Sewage Sludge
The determination of phthalates in a sludge sample is demonstrated in Fig. 11. The sludge was obtained from a wastewater treatment plant. The extraction was performed by ultrasonic extraction with acetone : hexane according to the procedure described by Paxeus [30]. In this sample, DiBP, DBP, BBzP and DEHP are clearly detected. At the end of the chromatogram, the “hump” corresponding to DiNP and DiDP is also detected. In this sample, the concentration of DEHP, the most abundant phthalate, was around 30 mg g–1 dry mass.
6 Analysis of Phthalates in Air 6.1 Introduction
Transport in the atmosphere is an important pathway for the distribution of phthalates in the environment [32, 33]. Most of the phthalates are adsorbed onto
Fig. 11. GC-MS chromatogram obtained for a wastewater treatment plant sludge extract
36 F. David et al.
Analytical Methods Review
37
particulate matter [34–36] and earlier methods used for monitoring phthalates in air often focused on the collection and analysis of dust samples [37].Although the vapour pressures of phthalates are relatively low [23], phthalates can still be regarded as semi-volatiles and are also present in the aerosol or vapour phase. Since phthalates are present at trace levels (ng–µg m–3 level), direct analysis of air samples is not possible and enrichment techniques are needed. Enrichment of compounds in air can be done by active or passive concentration of the solutes on adsorbents or sorbents. For the determination of phthalates adsorbed on particulates, active or passive sampling on filters is used. 6.2 Analysis of Total Phthalate Concentrations in Air
For the determination of total phthalate concentrations (gaseous + aerosols + particulates), sampling on a sorbent tube is used. Most widely used adsorbents are activated carbon, porous polymers like Tenax (2,6-diphenylphenylene oxide polymer) and resins (XAD-2) [38–47]. Recommended methods for the sampling and analysis of phthalates (OSHA CIM, OSHA 104, NIOSH 5020) also include trapping of the phthalates present as aerosols or in the vapour phase on a combination of a polyurethane foam (PUF) plug and a resin cartridge. Phthalates adsorbed to particulate material can also be collected first on a glass fibre filter. In these generic methods, typically 60–240 L of air are sampled at a rate of 1000 mL min–1. It should be noted that with these methods contamination problems can occur due to PVC filter holders or plasticized rings that are used to hold the glass fibre filters. These materials can contain phthalates and consequently lead to high blank values. After sampling, the phthalates are desorbed from the adsorbents by liquid extraction, or by thermal desorption. In the case of particulates concentrated on filters, liquid extraction is used. The extraction procedure for particulates on filter media is very similar to the extraction of phthalates from solid samples (soil, sediment). Ultrasonic extraction, Soxhlet extraction, accelerated solvent extraction and others can be used [41, 47]. The advantage of liquid desorption is that quantitative extraction of the solutes can easily be obtained, the extract can be further fractionated or purified and the final extract can be analysed several times. Liquid desorption, however, lacks sensitivity. Assuming a limit of detection of 10 pg per compound with mass spectroscopic detection, a concentration of minimum 10 ppb is required in the final extract. By using liquid extraction of the sorbent or filter, concentration to 1 mL and a 1 µL injection, at least 1000 L of air should be sampled to quantify 10 ng m–3. Otake et al. [46] used for instance 100 mg charcoal tubes and sampled for 72 h at 1 L min-1 (total sample 4.3 m3). After sampling, desorption was done with 1 mL toluene and the extract was analysed by GC-MS. A detection limit of 4 ng m–3 could be obtained. Thermal desorption, on the other hand, has the advantage of higher sensitivity because all sorbed compounds can be quantitatively transferred to the GC and the detector. The same sensitivity (10 pg in detector) can thus be obtained by sampling only 1 L of air. The limiting factor in thermal desorption is the desorption efficiency. The adsorption of high-molecular weight compounds is strong and a high energy
38
F. David et al.
(high temperature) is needed for quantitative desorption especially when classical adsorbents are used (e.g. carbon-based materials, Tenax and Porapak). Tenax adsorption traps can be used successfully for the low-molecular weight phthalates (typically DEP, DiBP, DBP), but for the high-molecular weight isomeric mixtures (DiNP, DiDP) poor recoveries have been observed. Pre-concentration by sorptive enrichment with silicones offers a useful alternative and was introduced by Baltussen et al. [48, 49]. Air was sampled through a tube packed with a bed of 100% polydimethylsiloxane (PDMS). The polymeric material is above its glass transition temperature (Tg) at sampling temperatures (0–30 °C) and the solutes are, in contrast to classical sampling systems, sorbed into (dissolved in) the liquid stationary PDMS phase rather than adsorbed onto an active surface. The sampling traps allow sufficient enrichment at a relative high sampling speed and quantitative thermal desorption can be performed at moderate temperatures. Moreover, the material is highly inert and has a high thermal stability. However, for the enrichment of phthalates, it was observed that 100% PDMS traps resulted in too many bleeding compounds (low-molecular weight silicones) that interfered in the extracted ion chromatograms at m/z 149 [50]. This is the most abundant and the quantification ion for phthalate determination. The dimethylsiloxane oligomers give m/z 73 and m/z 147 as most abundant ions but because of the silicium isotopes the ion m/z 149 is also observed in the spectra.An alternative method was found by coating a silicone layer on an inert support in a 5% (w/w) concentration of 5%. Thermal desorption tubes (4 mm i.d. ¥ 180 mm) were filled with 100 mg of the material and were conditioned at 300 °C for 2 h. The ‘blank’ profile showed only some traces of volatile silicon fragments, but these did not interfere with the phthalate peaks. By using these traps, ng m–3 concentrations of phthalates can be measured. Since pre-concentration on PDMS tubes is based on sorptive enrichment, the breakthrough volumes (Vb) of phthalates can be calculated directly from the theory developed by Lövkist and Jönsson [51]. The values for some important phthalates are summarised in Table 6 and vary between 17 L for dimethyl phthalate up to 75 ¥ 106 L for diisodecyl phthalate when sampling is performed at a rate of 500 mL min–1. If the sample volume is limited to 15 L, all phthalates are quantitatively trapped. Assuming a mass spectrometric detection limit of 10 pg, the sensitivity of the thermal desorption-GC-MS method is 10 pg/15 L or 0.7 ng m–3. In practice, a Table 6. Breakthrough volumes (L) of several phthalates when
sampling is performed on PDMS tubes (100 mg 5% PDMS) at a flow rate of 500 mL min–1 at room temperature (20 °C) Compound
Breakthrough volume (L)
DMP DEP DIBP DBP DEHP DIDP
17 23 13 ¥ 102 19 ¥ 102 17 ¥ 106 75 ¥ 106
Analytical Methods Review
39
limit of quantitation of 3 ng m–3 was obtained due to typical background levels of DiBP, DBP and DEHP around 20–30 pg. A typical sampling and analysis procedure based on sorptive enrichment and thermal desorption-GC-MS is summarized as follows [50]. Samples are aspirated through a thermal desorption tube containing 100 mg 5% PDMS sorbent by using a universal air sampling pump (SKC, Dorset, UK) at a nominal flow of 500 mL min–1 for 30 min (sample volume = 15 L). After sampling, the sampling tubes are closed and stored in an airtight container or wrapped in aluminium foil. After sampling, the tubes are placed in a thermal desorption unit (TDS-A, Gerstel GmbH, Muelheim, Germany), mounted on an Agilent 6890 GC coupled to an Agilent 5973 mass selective detector. Desorption is started by programming the tube to 300 °C. The released solutes are transferred from the sampling tube through a heated fused silica capillary into a cryotrap (PTV inlet). During thermal desorption, the PTV is cooled to –100 °C and the phthalates are trapped in an empty liner. After 10 min, the thermal desorption is completed and the solutes are injected into the capillary column by rapidly heating the PTV to 300 °C at 10 °C s–1. The chromatographic and mass spectroscopic conditions are the same as described above (Table 3). As for other sample matrices, a potential problem in the trace analysis of phthalates in air is contamination. It is very important that the contact between stored sample tubes and laboratory air is minimised, since DiBP and DBP were found to be two important contaminants originating from the surrounding lab atmosphere. A typical chromatogram of the analysis of laboratory indoor air is presented in Fig. 12. The measured concentrations of DiBP and DBP are 200–700 ng m–3. If tubes are left unprotected in the laboratory, they will adsorb phthalates and this leads to overestimated data. The limit of detection (LOD) of the method with a sampling speed of 500 mL min–1 and a sampling time of 30 min (15 L sample) is of the order of 1 ng m–3 for the single isomer phthalates and around 10 ng m–3 for the mixed isomer phthalates. For DiBP, DBP and DEHP, background values under optimised (clean) conditions can range between 2–3 ng m–3 and values below these limits are questionable for these compounds. The sorptive enrichment-thermal desorption-GCMS method was validated in the concentration range 3 to 3000 ng m–3. The correlation coefficients of the linearity curves were higher than 0.9970 for the single isomer phthalates DMP, DEP, DiBP, DBP, BBzP and DEHP, and 0.9786 for DiNP and 0.9782 for DiDP. The relative standard deviation of five replicate analyses at a level of 100 ng m–3 was between 2.3% (for DEP) and 9.2% (for DEHP). As a typical example of phthalate monitoring in air, the concentration of phthalates was measured in a greenhouse and in the surrounding atmosphere. Outside samples were taken at 1, 10 and 100 m distances. The resulting data are summarized in Fig. 13. Inside the greenhouse, the concentrations of the detected phthalates were respectively: 226 ng m–3 DiBP, 156 ng m–3 DBP, 48 ng m–3 BBzP and 309 ng m–3 DEHP. At 100 m distance, only background levels of DiBP, DBP and DEHP were measured and the concentrations decrease with increasing distance from the greenhouse. It should also be noted that dynamic (active) sampling by using either adsorbents or sorbents measures total phthalate concentrations, since particulates will
Fig. 12. Phthalate of laboratory air at m/z 149 on a 5% PDMS tube. Peaks 1 DEP, 2 DiBP, 3 DBP, 4 DEHP
40 F. David et al.
Analytical Methods Review
41
Fig. 13. Phthalate concentrations (ng m–3) in air samples in and outside a greenhouse
also be trapped on the material that acts as a filter. If a glass filter is placed before the (ad)sorbent tube, the particulates can be monitored separately. However, the differentiation between particle-bound (and adsorbed) phthalates and the phthalates in gaseous and aerosol phase by such a sampling train (filter+sorbent) is questionable. It can be expected that phthalates initially present in the gaseous phase can also be concentrated on a filter or on dust collected on the filter during sampling due to adsorption (so the concentration in the sorption tube is too low). In contrast, phthalates initially adsorbed on particles and trapped on a glass fibre filter can be purged out towards the sorption tube (giving too high values for the “gaseous+aerosol fraction”). This effect has already been clearly demonstrated by Schulz and Püttmann [37]. With a 1 h sampling time, measured phthalate concentrations were higher than the values obtained by using a 24 h sampling. This effect was called the “blowing off ” problem. Therefore, we recommend the use of adsorbents or sorbents for the determination of total phthalate concentrations. 6.3 Passive Sorptive Sampling of Phthalates in Air
Passive sampling is a very suitable technique for the measurement of air quality in indoor and working environment [52]. For passive sampling, a sorbent is put in a holder. This may be placed in a specific place in the room or used as a personal sampler. Sampling is performed over several hours or days and concentrations are expressed as time weight averages (TWA). The (ad)sorbed solutes are desorbed with a suitable solvent and the extract is analysed. Recently, solid-phase micro-extraction (SPME) with a PDMS coating immobilized on a fused silica
42
F. David et al.
fibre, was introduced as an alternative sampling device for passive diffusion sampling [53–55]. After sampling, the fibre is introduced in a hot GC inlet and the concentrated compounds are thermally desorbed in the GC. The major constraint of the SPME method is the relatively small amount of sorbent that is present on the fibre (0.5 µL). For this reason, the recently developed PDMS-coated stir bars can be an interesting alternative. It has been demonstrated that these stir bars can be used to sample volatiles in the headspace of liquid or solid samples [56]. In passive sampling with sorptive extraction, the PDMS polymer is exposed to an air sample during a relative long period (hours, days) and the pollutants present at an initial concentration CS0 are enriched until equilibrium is reached between the gas and the PDMS polymer phase. This distribution can be defined for a given component at a given temperature as the distribution coefficient KPDMS/air : • CPDMS KPDMS/S = 9 (1) CS• • where CPDMS and CS• are the solute concentrations in the PDMS phase and air sample at equilibrium state, respectively. Since passive sampling is usually performed in open areas, where the sample volume is much larger than the volume of the PDMS phase, the concentration CS• at equilibrium does not significantly differ from the initial concentration CS0. This last value can thus easily be calculated: n (2) or n = VPDMS · CS0 · KPDMS/S (3) CS0 = 968 VPDMS · KPDMS/S
where n the absolute amount of solute that is enriched in the PDMS polymer and VPDMS is the volume of the PDMS phase. Equation (3) shows that n is directly proportional to the amount of sorbent VPDMS . This implies that if a coated bar containing 50 µL PDMS is used instead of an SPME fibre (0.5 µL PDMS) the sensitivity is increased by a factor 100. Passive sampling with a PDMS-coated bar was applied to the determination of phthalates in a room with PVC flooring. A clean PDMS-coated bar containing 50 µL PDMS was exposed to the indoor environment for 24 h.After sampling, the phthalates were analysed by thermal desorption-GC-MS. The resulting extracted ion chromatogram at m/z 149 is shown in Fig. 14. The high abundance of DiBP, DBP and DEHP illustrates the extremely high enrichment of the phthalates in the PDMS polymer phase. Since phthalates adsorbed on particulates are not enriched by this sampling method, passive sampling with sorptive extraction allows differentiation between phthalates that are present in the vapour or aerosol phase and phthalates adsorbed onto particulates. Only the phthalates in vapour or aerosol phase are measured. Therefore, passive sorptive sampling is complementary to dynamic sorptive enrichment. 6.4 Dust Analysis
An important part of the human phthalate intake from the atmosphere may be attributed to the inhalation of dust particles on which plasticizers are adsorbed.
Fig. 14. Passive diffusion extraction (PSSE, 50 µL PDMS) of phthalates in a carpeted house
Analytical Methods Review
43
44
F. David et al.
Hence, attention has been paid to the determination of phthalates in dust. The hazardous impact of phthalates adsorbed on dust particles depends on the particle size of the dust.According to the American Conference of Governmental Industrial Hygienists (ACGIH) [57] and European regulation [58], the inhalable dust fraction is defined as the fraction of particulates with a median particle size smaller than 100 µm. The thoraric fraction is defined as the fraction of particulates with a median particle size smaller than 10 µm. The respirable fraction is defined as the fraction of particulates with a median particle size smaller than 4 µm. Inhalable dust is deposited in the respiratory trajectory (mouth and gullet), while respirable particles are deposited anywhere in the gas-exchange region, including the lung bladders. Thoraric dust is retained before the lung area. Dust sampling can be done by using vacuum cleaners, glass fibre filters in combination with high-volume air sampling pumps or with special dust samplers. A detailed description of dust sampling is given in VDI method 4300 [59]. Special dust samplers allow differentiation between the total inhalable and respirable dust filtration [60]. By using these samplers, the total inhalable dust fraction is aspirated at a flow of about 2 L min–1 onto a glass fibre filter (pore size 1.0 µm) through a broad cassette inlet. This sampler discriminates larger particulate matter. Smaller respirable particles can be collected with a cyclone-type sampler. The recent models are constructed of plasticizer-free plastics that eliminate electrostatic problems. This avoids repellence of the particles and contributes to high sampling efficiencies. In both cases, sampling is typically performed over several hours. Total dust concentrations are measured gravimetrically. Phthalates adsorbed on the dust fraction can be determined after liquid extraction of the filter. However, in our experience, it is extremely difficult to obtain reliable data on phthalates on the dust fraction collected in these special samplers. More reliable data are obtained by sampling larger amounts of dust by highvolume sampling, followed by dust fractionation by sieving and phthalate determination by liquid extraction and GC-MS. Fractionation of dust samples can be done by using stainless steel sieves or by centrifugation. Extraction of phthalates from dust can be performed by the methods described for solid samples [47, 61]. Alternatively, thermal extraction can be used, especially if limited amounts of sample are available. A few mg of sample can be placed in a thermal desorption tube and the tube is analysed in the same way as for air samples. An example is given in Fig. 15, showing the thermal extraction-GC-MS analysis of a house dust sample. The measured phthalate concentrations were 90 mg kg–1 DBP, 200 mg kg–1 DiHP, 700 mg kg–1 DEHP and 300 mg kg–1 DiNP. This sample clearly shows the presence of the C7-isomeric mixture of diisoheptyl phthalate eluting before DEHP.
Fig. 15. Thermal extraction-GC-MS analysis of house dust
Analytical Methods Review
45
46
F. David et al.
7 Analysis of Phthalates in Biota (Vegetation, Milk, Fish) 7.1 Introduction
The determination of phthalates in biological matrices such as vegetation, fish or milk is more difficult due to the complexity of the matrix. In general, the methodology for the determination of phthalates in biota can be divided into two classes, depending on the fat content of the matrix. For samples with a relatively low fat content (< 1%), phthalates can be extracted by using the same methods as described for soils, sediment and other solid samples. During the extraction, other constituents such as sterols, pigments, flavanoids, waxes, fatty acids, etc will be co-extracted. After extraction, a cleanup method is needed. As clean-up methods, column chromatography or solidphase extraction can be used. In these clean-up methods, the separation of phthalates and co-extracted compounds is based on differences in polarity. For fatty matrices, the main problem is the co-extraction of fats. Since lipids have a polarity similar to the polarity of phthalates, it is difficult to remove them with methods based on column chromatography or solid-phase extraction. The fat matrix can be removed by a clean-up method based on size exclusion chromatography (gel permeation chromatography, GPC). In size exclusion chromatography, compounds are separated according to the molecular volume (ª molecular mass). Larger molecules cannot enter the pores of the GPC packing material and elute before smaller molecules that can enter the small pores. GPC was first described for the fractionation of pesticides from lipid matrices [62–63]. However, this liquid chromatography method is more and more considered as a very valuable sample preparation technique for all determinations of organic contaminants in lipid matrices. On GPC columns, phthalates with molecular weights in the range 200–400 Da are separated from lipids (MW around 800 Da). Classical GPC separations are performed on large (40 cm ¥ 25 mm i.d.) low-pressure columns, operated at 5 mL min–1 [62–63]. The method can be miniaturized and solvent consumption can be largely reduced by using HPLC equipment and high-pressure GPC columns. Automated GPC clean-up can be performed with a system consisting of an isocratic HPLC pump, an autosampler allowing the injection of 500 µL fat extract, a temperature-controlled column oven, a variable UV detector (optional) and a fraction collector. The separation is performed on small-bore columns (e.g. 300 mm ¥ 7.5 mm i.d. PL-Gel with 5 µm particles and 5 nm pore size), operated at 1 mL min–1 dichloromethane. The 5 nm pore size is important, since it allows the separation of organic compounds in the 100–1000 mass range. Best resolution is obtained on two columns in series (total length = 60 cm). Solvent consumption is approximately 5–10 times lower in comparison to the classical GPC method. The phthalate fraction (3–5 mL) is collected automatically.After collection, the fraction is concentrated and this clean-up step can eventually be followed by an additional clean-up method according to polarity by using column chromatography or SPE.
Analytical Methods Review
47
7.2 Analytical Procedure for the Determination of Phthalates in Vegetation
Vegetation samples normally do not contain high levels of lipids. For the determination of phthalates in plant material, first a liquid-solid extraction is used.All techniques described for solid samples (Soxhlet, shaking, ultrasonic extraction, ASE, SFE) can also be used for the extraction of phthalates from plant material. High recoveries are obtained by using a simple shaking or ultrasonic extraction method. These methods are also fast, cheap and relatively small amounts of solvents are used. For the extraction, two approaches can be used. In the first approach [6], vegetation samples are first homogenised with a blender. Approximately 5 g (wet) sample is spiked with internal standard(s) and mixed with 30 g anhydrous sodium sulfate (pre-baked at 450 °C) in a mortar until a dry powder is obtained. The dried sample is extracted with 30 mL of a 1 : 1 mixture of hexane and dichloromethane by sonication for 10 min. After the suspended particles are settled, the supernatant is removed. The extraction is repeated twice with fresh solvent and finally the three organic fractions are combined and concentrated to 1 mL under nitrogen. In the second approach [7], the sample is not dried with sodium sulfate. Approximately 10 g homogenised (wet) sample is weighed in a 40 mL I-Chem vial. After addition of the internal standard(s) and 10 mL acetone, the sample is extracted for 15 min by sonication. Then 10 mL cyclohexane is added and the sonication extraction is repeated for another 15 min. After this first extraction, the vials are placed on a shaking machine for 30 min. Finally, the vials are again placed in the ultrasonic bath for 15 min. After completion, the extraction vial is centrifuged. An aliquot (5 mL) of the (upper) cyclohexane phase is transferred to another tube and concentrated to 1 mL Clean-up of the extracts can be done by column chromatography or by solidphase extraction. The procedures are identical to the procedures described for solid samples [6,22]. By using column chromatography, the concentrated extract is transferred onto an alumina column packed with 15 g deactivated alumina (15% water w/w) and with a 1–2 cm bed of anhydrous sodium sulfate on top. The column is eluted with 30 mL hexane, followed by 30 mL 10% dichloromethane in hexane and finally 30 mL 50% dichloromethane in hexane. This last fraction contains the phthalates. This fraction is concentrated to 1 mL and analysed. The clean-up can be miniaturized by using a 500 mg alumina cartridge. After applying the extract, the cartridge is washed with 3 mL hexane and the phthalates are eluted with 3 mL 75% hexane-25% MTBE. The limit of detection of these methods is in the order of 2–10 µg kg–1 wet sample for the single isomer phthalates (and 20–100 µg kg–1 wet sample for the mixed isomer phthalates). However, method blanks are typically in the order of 10–20 ppb, especially for DiBP, DBP and DEHP. For these compounds, the limit of quantitation is around 20–40 µg kg–1 wet sample (set to two times the background level).
48
F. David et al.
7.3 Analytical Procedure for the Determination of Phthalates in Milk or Edible Oils and Fat
For the determination of phthalate in milk samples the following method was developed [7]. Milk samples are first homogenised by shaking. Milk powders are dissolved in water (1 g/10 mL). Approximately 5 g milk sample is then extracted with 20 mL of a 1 : 1 cyclohexane/acetone mixture in a 40 mL I-Chem vial. The vials are shaken on a shaking machine for 30 min.After completion, the vials are centrifuged and the supernatant is transferred to a pre-weighed I-Chem vial. The solvent is evaporated under nitrogen and the fat content is measured. The residue is dissolved in dichloromethane and internal standard (d4-DEHP) is added. The amounts of solvent and internal standard are adjusted to give approximately 50 mg fat and 100 ng internal standard per mL dichloromethane. The solution was then homogenised in a vortex agitator. The separation of the fat from the phthalate containing fractions is then performed by gel permeation chromatography (GPC). For the determination of phthalates in edible oils and fat, the oil or fat sample can by diluted directly in dichloromethane to a concentration of 50 mg fat per mL. If the sample solution is not clear, some water might be present and this can be removed by adding some anhydrous sodium sulfate to the sample. The clear solution is also fractionated by GPC. For the GPC separation of the dichloromethane solution, 500 µL is injected onto a two column combination. This combination consisted of a 5 cm ¥ 7.5 mm i.d. PL-Gel pre-column and two 30 cm ¥ 7.5 mm i.d. ¥ 5 µm PL-Gel 5 nm columns. The mobile phase is dichloromethane at 1 mL min–1 flow rate. UV detection at 220 nm is used to monitor the effluent. Phthalates typically elute in a window between 20 and 23 min (3 mL fraction). This fraction is automatically collected. The total run time is 30 min. To the collected fraction, 100 µL cyclohexane is added and the extract is concentrated to 100 µL. This extract is analysed by GC-MS. The limit of detection of this method is around 20 ng g–1 fat. DiBP, DBP and DEHP are detected in the method blanks, but the levels are constant around 20–50 ng g–1 fat. Hence, the limit of quantification for these phthalates is 40–100 ng g–1 fat (two times background level). For milk samples with 3% fat content, this corresponds to a quantification limit of 1–3 ng g–1 milk. The recovery of the GPC clean-up method was tested by spiking an olive oil sample at a 500 ppb level with DMP, DEP, DiBP, DBP, BBzP and DEHP and at a 5 ppm level with DiNP and DiDP. The olive oil was tested before and no phthalates were detected at concentrations above 50 ppb (500 ppb for DiNP and DiDP). The results from the non-spiked oil were therefore considered as the “procedure blanks”. D4-DEHP (internal standard) was added to the dichloromethane solution. The sample was analysed in triplicate. The linearity was tested by spiking an olive oil sample at seven levels (100, 250, 500, 1000, 2500, 5000 and 7500 ng g–1 fat) (¥ 10 for mixed isomers). The mean recovery, standard deviation (RSD %) and linearity (correlation coefficient r2) are listed in Table 7. In general, good recoveries (>80%) are obtained, except for butylbenzyl phthalate for which the recovery is lower. This is probably due to a slightly different behaviour of this
49
Analytical Methods Review Table 7. Validation data for the determination of phthalates in oil
Recovery (%) RSD (%) Linearity (r2)
DMP
DEP
DiBP
DBP
BBzP
DEHP
DiNP
DiDP
103 15.7 0.9956
134 15.7 0.9946
104 9.5 0.9992
138 11.8 0.9962
69 16.0 0.9947
127 12.1 0.9978
85 11.6 0.9841
81 10.6 0.9903
compound in sample clean-up (GPC). The correlation coefficients are better than 0.99 for all phthalates, except for DiNP (0.98). This correlation shows that the method can be used in the concentration range from 100–7500 ppb for the single isomer phthalates and from 1–75 ppm for the mixed isomer phthalates. An alternative and automated method for the determination of phthalates in oil and fat matrices was described by Pacciarelli et al. [64]. In this method, an online HPLC-GC method was used. The sample fractionation was performed by HPLC in straight (normal) phase mode. Fractionation of the triglycerides was done on a silica column (100 ¥ 4.6 mm i.d. Lichrosorb SI-100, 5 µm) using 1 : 1 dichloromethane/cyclohexane (with 0.5% acetonitrile) as mobile phase at 1 mL min–1. The fraction containing DEHP was automatically transferred in the GC and the triglyceride matrix is eluted afterwards. 7.4 Analytical Procedure for the Determination of Phthalates in Fish
Depending on the type of fish and available sample size, the fish samples can be analysed as whole fish, as fillet (muscle with skin removed) or the analysis can be performed on specific organ samples (liver, stomach). For diet studies, the fillet sample represents the edible part of the fish. For contamination studies, whole fish samples or organ analyses give relevant information. Due to their lipophilic character, phthalates accumulate in the fat. The analytical methods for the determination of phthalates in fish consist in general of a fat (+phthalate) extraction step, followed by a clean-up step. Different extraction techniques can be used. These include ultrasonic extraction, Soxhlet extraction,ASE, etc. For cleanup, gel permeation chromatography and column chromatography or solid phase extraction can be used. A fish matrix is quite complex and it is advisable to use both GPC plus an additional alumina column clean-up based on polarity differences to remove fats and other interfering compounds that are co-extracted. Liu et al. [6] used ultrasonic extraction. The samples are homogenised unfrozen. Analytical subsamples, 5 g wet weight, are spiked with d4-labelled standards and mixed with 30 g anhydrous sodium sulfate in a mortar until a dry powder is obtained. This homogenised and dried sample is extracted with 30 mL dichloromethane by sonication for 10 min and shaking for another 10 min. After the suspended particles are settled, the supernatant is removed and the extraction is repeated twice with fresh solvent. Finally, the three dichloromethane fractions are combined and concentrated to 1 mL. The extracts are first cleaned by GPC. The phthalate fraction is concentrated and then transferred onto an alu-
50
F. David et al.
mina column packed with 15 g deactivated alumina (15% water w/w) and with 1–2 cm bed of anhydrous sodium sulfate on top. The column is eluted with 30 mL hexane, 30 mL 10% dichloromethane in hexane and finally 30 mL 50% dichloromethane in hexane. This last fraction contains the phthalates. This fraction is concentrated to 1 mL and analysed. By using this method, Liu et al. [6] obtained detection limits around 2 ng g–1. Due to method blanks, the limit of quantification for DBP was around 20 ng g–1 and around 40 ng g–1 for DEHP. The recovery was 70–95% for the single isomer phthalates, measured by GC-MS and 75–110% for the mixed isomer phthalates measured by LC-MS. The same preparation procedure was used by David et al. [7]. In this procedure, 5 g homogenised fish sample was mixed with internal standard (d4-DEHP) and with 15 g pre-cleaned sodium sulfate. Extraction was performed by sonication with 20 mL cyclohexane for 15 min. The extract was centrifuged and the supernatant was removed. The extraction was performed another time with fresh solvent and the combined extracts were concentrated under nitrogen and fractionation by GPC, using the method described for milk samples. An example of the GPC separation for a fish extract is given in Fig. 16. The chromatogram obtained for a DEHP reference standard is overlayed. The DEHP peak elutes on the tail of the lipid peak. The fraction between 15 and 17 min is collected, concentrated and analysed by GC-MS. In this fraction, some sterols elute that cannot be separated from DEHP. These compounds do not interfere with the further analysis. The GC-MS chromatograms are shown in Fig. 17. The extracted ion chromatogram at m/z 149 shows the presence of DEHP. The extracted ion chromatogram at m/z 153 shows the presence of the internal standard (d4-DEHP). In this case, no additional clean-up by column chromatography or by SPE was used. The recovery for DEHP measured at three spike levels between 2 ng g–1 (wet sample) and 40 ng g–1 (wet sample) were 89–93%. The relative standard deviation at 40 ng g-1 (wet sample) was 9.7%. Letinski [25] used accelerated solvent extraction for the extraction of phthalates from fish tissue.Approximately 1 g fish sample was mixed with sodium sulfate and extracted with a 90% hexane/10% ethyl acetate mixture at 120 °C and 1600 psi. This extract was purified on alumina. The limit of quantification was 150 ng g–1 fish sample and recovery was 98% for the recovery matrix spike.
8 Sample Preparation Methods for Phthalic Acid Mono-Esters Phthalic acid monoesters are more polar than phthalates and therefore the sample preparation methods must be adapted. For the determination of monoesters in blood, plasma and urine samples, several methods are described in the literature [8–10, 12, 13, 17, 18]. For the determination of monoesters in environmental samples, only limited data are available. Suzuki et al. [14] used solid-phase extraction on polymeric phase disks (styrene divinylbenzene) for the determination of monoesters in riverwater samples. The waters were acidified to pH 2 before extraction to ensure the monoesters are in the protonated form (not ionised). Extraction efficiency was higher than 72% (except for monomethyl phthalate).
51
Fig. 16. GPC fractionation of phthalates in fish extract
Analytical Methods Review
After extraction, the monoesters must be derivatised prior to GC analysis. Alternatively, HPLC of non-derivatised monoesters can be performed (see above). Recently, a method has been developed for the analysis of phthalic acid monoesters in water samples by using an in situ derivatisation, followed by stir bar sorptive extraction and thermal desorption GC-MS [19]. One mL of water sample is placed in a 20 mL glass vial and spiked with internal standard solution. 500 µL of a 2 : 1 (v/v) mixture of ethanol and pyridine is added and the
52
F. David et al.
Fig. 17. GC-MS chromatograms of DEHP and internal standard (d4-DEHP) extracted from fish
mixture was homogenised by vortex agitation. After addition of 100 µL of ethyl chloroformate, the mixture was again vortex homogenised and placed in an ultrasonic bath for 15 min. During vortex and ultrasonic agitation gas evolved from the vial. The mixture was finally diluted with 10 mL water. The derivatised monoesters are extracted by stir bar sorptive extraction (SBSE), followed by thermal desorption-GC-MS. The limit of detection (LOD) is in the order of 0.05 µg L–1 for MBP and MEHP. The analysis of a sample spiked at 1 µg L–1 levels is shown in Fig. 18. The monoesters MBP, MEHP, MiNP and MiDP are easily detected.
Fig. 18. Determination of phthalic acid monoesters in water by in situ derivatisation and SBSE-thermal desorption-GC-MS
Analytical Methods Review
53
54
F. David et al.
9 Conclusions During the past years, various sample preparation and analytical methods have been developed for the determination of phthalic acid diesters in different environmental samples. The major problem in phthalate analysis is the risk of contamination. Precautions have to be taken into account to minimize this risk and to control the background values. Based on the authors’ personal experience, the following methods are recommended for the different matrices. Clean water samples can be extracted by methods based on solid-phase extraction. For contaminated water samples, miniaturized liquid-liquid extraction can be used. Sediments, sludges and soil samples can be extracted by using ultrasonic extraction. Biota samples can also be extracted by sonication. The extracts are purified by gel permeation chromatography, followed by column chromatography or solid-phase extraction. The analysis of the single isomer phthalates can be done by GC-MS. HPLC-MS offers an interesting alternative for the analysis of mixed isomer phthalates. For the analysis of monoesters in environmental samples, new methods are currently under development. The extraction and clean-up method used for phthalates need to be adapted. Specifically, the monoesters must be derivatised prior to GC-MS analysis. Alternatively, they can be analysed without derivatisation by using LC-MS.
10 References 1. US EPA methods 525 (drinking water), 606 and 625 (waste water) and 8060 and 8270 (solid waste). US Environmental Protection Agency, Cincinnati, Ohio 45268, USA 2. ISO draft international standard 18856, ISO. Geneva, Switzerland 3. Vikelsoe J, Thomsen M, Johansen E (1998) Sources of phthalates and nonylphenols in municipal waste water, technical report no 225 and 268. National Environmental Research Institute (NERI), Roskilde, Denmark 4. Vikelsoe J, Thomsen M, Johansen E, Carlsen L (1999) Phthalates and nonylphenols in soil, technical report no 268. National Environmental Research Institute (NERI), Roskilde, Denmark 5. George C, Prest H (2001) Agilent technologies application note Nr 5988–2244EN 6. Lin, Zhong-Ping, Ikonomou MG, Hongwu J, Macintosh C, Gobas FAPC (2002) Environ Sci Technol submitted for publication 7. David F, Tienpont B, Sandra T, Sandra P (2002) submitted for publication 8. Sjöberg P, Bondesson U (1985) J Chromatogr 344:167 9. Sjöberg P, Bondesson U, Sedin E, Gustafsson J (1985) Eur J Clin Invest (1985) 15:430 10. Sjöberg P, Bondesson U, Sedin E, Gustafsson J (1985) Transfusion 25:424 11. Marschall H-M, Green G, Egestad B, Sjövall J (1988) J Chromatogr 452:459 12. Egestad B, Green G, Sjöberg P, Klasson-Wehler E, Gustafsson J (1996) J Chromatogr B 677:99 13. Haam D, Vandenbroek P, Jongeneelen F (1993) Int Arch Occupat Environ Health 64:555 14. Suzuki T, Yaguchi K, Suzuki S, Suga T (2001) Environ Sci Technol 35:3757 15. Tienpont B, David F, Sandra P (2002) submitted for publication 16. Barry Y, Labow R, Keon W, Tocchi M, Rock G (1989) J Thorac Cardiovasc Surg 97, 900 17. Shintani H (2000) Chromatographia 52:721 18. Blount B, Milgram K, Silva M, Malek N, Reidy J, Needham L, Brock J (2000) Anal Chem 72:4127
Analytical Methods Review
55
19. Tienpont B, David F, Sandra P (2002) submitted for publication 20. Furtmann K (1994) Fresenius J Anal Chem 348:291 21. Parkman H, Remberger M (1995) Phthalates in Swedish sediments, IVL Report Publ no 1167, p 21+appendix 22. Braaten B, Berge JA, Berglind L, Baekken T (1996) Occurence of Phthalates and Organotins in sediments and water in Norway. Norwegian Institute for Water Research (NIVA) report SNO 3552–96, p 45 23. Cousins I, Mackay D (2000) Chemosphere 41:1389 24. Van der Velde E, Korte de G, Versteegh A (1998) Determination of phthalate esters in water by SPE or in-vial extraction with GC-MS analysis: how to avoid the contamination problem. Paper presented at 20th international symposium on capillary chromatography, Riva del Garda, Italy, 26–29 May 1998 25. Letinski DJ (1999) Lecture presented at Workshop Nov 4–5 26. Pawliszyn J (1999) Applications of solid phase micro-extraction, RSC chromatography monographs. Royal Society of Chemistry, Letchworth, UK (ISBN 0–85404–525–2) 27. Baltussen E, Sandra P, David F, Cramers C (1999) J Microcolumn Sep 11:737 28. Steffen D, Lach G (2000) Phthalate und Trichlosan in Sedimenten und Schwebstoffen Niedersächsischen Gewässer, Niedersächsisches Landesamt für Ökologie, report 10/2000) 29. Kolb M, Welte K, Mettenleiter S, Trinkmann A (1997) Wasser Boden 49:57 30. Paxéus N (1999) submitted for publication 31. Zurmühl T (1990) Analyst 115:1171 32. Thomas GH (1973) Environ Health Perspect 3:23 33. Giam CS, Chanm HS, Neff GS, Atlas EL (1978) Science 199:419 34. Cautreels W, van Cauwenberghe K (1978) Atm Environ 12:1133 35. Cautreels W, van Cauwenberghe K, Guzman LA (1993) Sci Total Environ 8:47 36. Thúren A, Larsson P (1990) Environ Sci Technol 24:554 37. Schulz H-M, Püttmann W (1993) Analysis of saturated hydrocarbons, fatty acids and phthalic acid esters in air particulate matter of a city area (Aachen). Wissenschaft und Umwelt 2/1993, p 131 38. Chang LW, Atlas E, Giam CS (1985) Int J Environ Anal Chem 19:145 39. Figge K, Rabel W, Wieck A (1987) Fresenius J Anal Chem 327:261 40. Vainiotalo S, Pfaffli P (1990) Ann Occup Hyg 34:585 41. California Environmental Protection Agency (1992) Monitoring of phthalates and PAHs in indoor and outdoor air samples in riverside, California. Contract no A933-144, Dec 1992 42. Fisher J, Ventura K, Prokes B, Jandera P (1993) Chromatographia 37:47 43. Fujimoto T, Takeda N, Taira T, Iikawa R (1995) Kurin Tekunoroji 5:45 44. Bartulewicz J, Bartulewicz E, Gawlowski J, Niedzielski J (1996) J Chem Anal (Warsaw) 41:753 45. Weiling G, Xiku W (1997) Huanjing Huaxue 16:382 46. Otake T, Yoshinaga J, Yanagisawa Y (2001) Environ Sci Technol 35:3099 47. Rudel RA, Brody JG, Sprengler JD, Vallarino J, Geno PW, Sun G, Yau A (2001) J Air Waste Manage Assoc 51:499 48. Baltussen E, Janssen HG, Sandra P, Cramers CA (1997) J High Resol Chromatogr 20:385 49. Baltussen E, David F, Sandra P, Janssen HG, Cramers CA (1998) J High Resol Chromatogr 21:332 50. Tienpont B, David F, Sandra P, Vanwalleghem F (2000) J Microcolumn Sep 12:194 51. Lövkist P, Jönsson JA (1987) Anal Chem 59:818 52. Namiesnik J, Gorecki T (2000) LC-GC Europe, September 2000, pp 678 53. Martos PA, Pawliszyn J (1997) Anal Chem 69:206 54. Martos PA, Pawliszyn J (1999) Anal Chem 71:1513 55. Khaled A, Pawliszyn J (2000) J Chromatogr 892:455 56. David F, Sandra P (2001) Passive sorptive sampling for indoor air monitoring using polydimethylsiloxane coated stir bas, paper 740, The Pittsburgh conference on analytical chemistry and applied spectroscopy, 4–9 March 2001
56
F. David et al.
57. ACGIH technical committee on air sampling procedures (1984) Particle size-selection sampling in the workplace. ACGIH, Cincinnati, Ohio 58. EN 481 (1993) European committee for standardisation CEN 1993-07–27 59. VDI 4300 (1999) Measurement of indoor air pollution, sampling of house dust. Verein Deutscher Ingenieure, Düsseldorf 60. Mark D, Vincent JH (1986) Ann Occup Hyg 30:89 61. Bruns-Weller E, Pfordt J (2000) Z Umweltchem Okotox 12:125 62. Specht W, Tilkes M (1980) Frezenius Z Anal Chem 301:300 63. Thier H-P, Zeumer H (1987) Manual of pesticide residue analysis, vol 1. VCH, Weinheim, p 75 64. Pacciarelli B, Müller E, Schneider R, Grob K, Steiner W, Fröhlich D (1988) J High Resol Chromatogr 11:135
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 57– 84 DOI 10.1007/b11463
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters Ian T. Cousins 1 · Donald Mackay 1 · Thomas F. Parkerton 2 1 2
Canadian Environmental Modelling Centre, Environmental and Resource Studies, Trent University, Peterborough, Ontario, K9J 7B8, Canada. E-mail:
[email protected] Exxon Mobil Biomedical Sci. Inc., Hermeslaan 2, 1831, Machelen, Belgium
A review is presented of the physical-chemical properties and reactivity of the phthalate esters including a discussion of how these properties control their partitioning and fate in the environment. The air and water solubilities decrease by orders of magnitude from the short alkyl chain phthalates such as dimethyl phthalate (DMP) to the long alkyl chain phthalates such as di-2-ethylhexyl phthalate (DEHP). The octanol-water partition coefficient, which is a measure of hydrophobicity, increases by orders of magnitude with increasing alkyl chain length and this increase is mainly controlled by the reduction in water solubility rather than an increase in octanol solubility. This increase in hydrophobicity results in strong sorption of the higher molecular weight phthalates to organic matter.Air-water partition coefficients (or Henry’s law constants) also increase with increasing alkyl chain length. However, the greater evaporative potential of higher molecular weight phthalate esters from water is offset by sorption to suspended matter in surface waters. Phthalates have high values of KOA suggesting that they will be appreciably sorbed to aerosol particles, soil and vegetation. From available data obtained under environmental conditions, half-lives of phthalates in environmental media are proposed. Systematic differences in reactivity or half-life are apparent, with the primary biodegradation half-life tending to increase with increasing alkyl chain length. In contrast, the opposite pattern is observed for the air oxidation half-life. A series of evaluative modelling calculations is described to illustrate how the physical-chemical properties result in differences in environmental partitioning behaviour, persistence and transport potential. In comparison to other organic chemical classes, model results indicate that phthalates are not environmentally persistent or subjected to significant long-range transport. Although the overall environmental persistence of the higher molecular weight phthalates tends to increase, KOA and thus the propensity to partition to aerosols, vegetation and soils also increases, thereby reducing the potential for long-range transport. Recommendations for future research on physical-chemical properties of phthalate esters for environmental fate assessment are discussed. Keywords. Phthalate ester, Structure, Physical-chemical property, Model, Fate
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
1
Introduction
2
Structure-Property Analysis of Physical-Chemical Properties
3
Observations on Physical-Chemical Properties
3.1 3.2 3.3
Physical State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Solubility in Water . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Vapour Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
. . 58
. . . . . . . . . . 64
© Springer-Verlag Berlin Heidelberg 2003
58
I.T. Cousins, D. Mackay and T.F. Parkerton
3.4 3.5 3.6 3.7
Air-Water Partition Coefficient . . Octanol-Water Partition Coefficient Octanol-Air Partition Coefficient . Data Gaps . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
67 68 71 72
4
Degrading Reactions . . . . . . . . . . . . . . . . . . . . . . . . . 72
5
Evaluative Fate Modelling with the EQC Model . . . . . . . . . . . 74
5.1 5.2 5.3 5.4
EQC Level I Modelling . . . . . . . . . . . . . . . . . . . . EQC Level II Modelling . . . . . . . . . . . . . . . . . . . . EQC Level III Modelling . . . . . . . . . . . . . . . . . . . Estimating Persistence and Long-Range Transport Potential with the TaPL3 Model . . . . . . . . . . . . . . . . . . . .
. . . . 75 . . . . 76 . . . . 76 . . . . 80
6
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
1 Introduction Phthalate esters are widely used as plasticizers, serving as important additives that impart flexibility to polymers including poly(vinyl chloride) (PVC), polyvinylacetates, cellulosics and polyurethanes [1]. The stability, fluidity and low volatility of high-molecular mass phthalate esters make them ideal for use as plasticizers. The variety of possible chemical structures of phthalate esters results in a wide range of physical-chemical properties and hence environmental partitioning behaviour for this class of compounds. This wide range of properties is principally a result of the variation in the length of the alkyl chains substituted on the diester groups. The names, molecular formulae, molar masses and melting points of 22 phthalate esters are listed in Table 1. The objectives of this chapter are to review the published physical-chemical and reactivity data for the phthalate esters, seek relationships between chemical structure and properties and determine how these properties will influence partitioning between abiotic media in the environment with the use of evaluative environmental fate models. The accumulation of phthalate esters in biotic media (i.e. food webs) is the focus of a separate chapter in this volume.
2 Structure-Property Analysis of Physical-Chemical Properties Physical-chemical properties which can be measured readily in the laboratory with a view to determining environmental partitioning include: solubility in water, vapour pressure, the Henry’s law constant (H), the octanol-water partition coefficient (KOW) and the octanol-air partition coefficient (KOA). There are few direct measurements of Henry’s law constants for the phthalate esters and no
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
59
Table 1. List of phthalate esters studied and their associated molar masses, molar volumes and melting points
Phthalate ester
Abbreviation
Molar mass (g mol–1)
Le Bas molar volume (cm3 mol–1)
Dimethyl phthalate Diethyl phthalate Diallyl phthalate Dipropyl phthalate Di-n-butyl phthalate Disiobutyl phthalate Di-n-propyl phthalate Butylbenzyl phthalate Diisohexyl phthalate Di-n-heptyl phthalate Di-n-octyl phthalate Butyl 2-ethylhexyl phthalate Di(n-hexyl, n-octyl, n-decyl) phthalate a Di(2-ethylhexyl) phthalate Diisooctyl phthalate Di-n-nonyl phthalate Diisononyl phthalate Di-n-decyl phthalate Diisodecyl phthalate Di(heptyl, nonyl, undecyl) phthalate a Diundecyl phthalate Ditridecyl phthalate
DMP DEP DAP DPP DnBP DIBP DnPP BBP DHP DIHpP DnOP BOP 610P DEHP DIOP DnNP DINP DnDP DIDP D711P DUP DTDP
194.2 222.2 246.2 250.3 278.4 278.4 250.3 312.4 334.4 362.5 390.6 334.4 404.6 390.6 390.6 418.6 418.6 446.7 446.7 418.7 447.7 530.8
206.4 254.0 283.6 298.4 342.8 342.8 387.2 364.8 431.6 476.0 520.4 416.6 542.6 520.4 520.4 564.8 564.8 609.2 609.2 564.8 653.6 742.4
a
Melting point (°C) 5.5 –40 – – –35 –58 – –35 –27.5 – – –37 –4 –46 –46 – –48 – –46 <–50 –9 –37
These are mixtures of three phthalate esters. Le Bas molar volumes were calculated using the method in ref. [3]. Melting points were taken from ref. [4].
measurements of KOA , although, as discussed later, both these properties are easily estimated as ratios of solubilities in air, water and octanol. Solubilities in water, vapour pressures and KOW for 22 phthalate esters, which were taken from the review by Cousins and Mackay [2], are summarized in Table 2. For a specific physical-chemical property the published values in Table 2 sometimes vary by several orders of magnitude. For example, the reported aqueous solubilities of di-n-octyl phthalate at 25 °C vary between 0.4 and 3000 µg L–1, that is by a factor of 7500. This variation in reported physical-chemical properties is initially surprising because the values are physical constants that should be known precisely. However, such variation in reported values is not unusual [36] and occurs because there are a wide range of different measurement techniques used for a given property, a wide range of laboratories undertaking the tests, a number of measurement difficulties that need to be overcome (e.g. analysis of concentrations close to detection limits, emulsion formation, sorption to glassware etc.) and a number of errors or oversights made by researchers. Selecting a “best value” for a given physical-chemical property is problematic; approaches include: (1) expert judgment, (2) using a list or checklist of criteria to assess the quality of published value (e.g. see ref. [37]) or (3) using structure-
60
I.T. Cousins, D. Mackay and T.F. Parkerton
Table 2. Physical-chemical properties of phthalate esters at 25 °C that were used in the “three-
solubility analysis” [2] Phthalate ester
C SWL a (mg L–1)
PL b (Pa)
Log KOW c
DMP
2810 [5] 4000 [6] 4248 [7] 4290 [8] 4320 [9]
0.22 [6]
DEP
680 [13] 896 [9] 928 [8] 1080 [6]
0.052 [14] 0.0813 [14] 0.22 [6]
DAP DPP DnBP
182 [8] 108 [8] 9.15 [16] 10.1 [8] 11.2 [6] 13 [9]
DIBP
6.2 [26] 20 [27] 20.3 [8]
1.46 [5] 1.47 [4] 1.53 [6] 1.56 [5] 1.61 [5] 1.61 [10] 1.62 [11] 1.66 [10] 1.74 [10] 1.90 [10] 1.60 [12] 2.21 [11] 2.21 [11] 2.24 [6] 2.29 [6] 2.35 [8] 2.67 [7] 3.00 [15] 2.42 [12] 3.23 [8] 3.27 [8] 3.74 [8] 4.08 [19] 4.11 [21] 4.13 [12] 4.30 [24] 4.39 [18] 4.56 [19] 4.57 [8] 4.72 [25] 4.72 [16] 4.79 [6] 5.15 [7] 4.50 [11] 4.11 [8]
DnPP BBP
DHP
2.27 ¥10–3 [17] 2.53¥10–3 [18] 2.77¥10–3 [20] 4.67¥10–3 [22] 4.80¥10–3 [23] 5.47¥10–3 [19] 9.73¥10–3 [6]
0.7 [26] 2.69 [6] 2.82 [8] 2.9 [9] 40.2 [7]
1.15 ¥10–3 [27] 1.20 ¥10–3 [6]
0.24 [6] 7.00 ¥10–2 [28] 4.60 ¥10–2 [29]
2.40 ¥10–4 [22] 1.87 ¥10–3 [6]
5.62 [12] 3.57 [6] 3.97 [11] 4.05 [7] 4.11 [21] 4.75 [7] 4.77 [21] 4.77 [27] 4.91 [8] 4.73 [12] 5.65 [6] 5.93 [6] 6.82 [12]
61
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters Table 2 (continued)
Phthalate ester
C SWL a (mg L–1)
DIHpP DnOP
1.70 ¥10–2 [28] 0.02 [16] 0.02 [26] 3 [9] 4.00 ¥10–4 [28] 5.10 ¥10–4 [29] 0.9 [6] 0.041 [8] 0.315 [16] 0.285 [26] 0.34 [24] 0.4 [13] 1.9 ¥10–3 [9]
610P DEHP
DIOP DnNP DINP DIDP DnDP D711P DUP DTDP a b c
0.09 [6] 6.00 ¥10–4 [35] 0.2 [6] 1.1 ¥10–4 [28] 0.28 [27] 1.7 ¥10–4 [28] 1.19 [6] 5.00 ¥10–5 [28] 2.20 ¥10–4 [29]
PL b (Pa)
Log KOW c
2.53 ¥10–2 [18]
5.22 [8] 7.06 [30] 8.06 [15] 8.10 [12] 8.18 [29] 7.25 [6] 5.22 [32] 5.11 [33] 7.06 [30] 7.14 [30] 7.45 [30] 7.45 [32] 8.35 [34] 8.05 [32] 8.05 [32] 7.94 [6] 8.06 [15] 7.27 [12]
6.53 ¥10–4 [6] 5.47 ¥10–4 [31] 1.31 ¥10–5 [21] 3.73 ¥10–5 [18] 4.40 ¥10–5 [21] 9.47 ¥10–5 [17] 8.53 ¥10–4 [6]
1.87 ¥10–4 [18] 7.47 ¥10–4 [6] 1.3 ¥10–4 [9] 7.20 ¥10–5 [6] 6.80 ¥10–6 [23] 8.91 [29] 6.00 [6]
1.1 [6] 0.34 [26]
C SWL is the solubility of liquid phthalate in water. PL is the liquid vapour pressure. KOW is the octanol-water partition coefficient.
property relationships (SPRs) to identify potential errors in reported data. The structure-property method is particularly attractive because it is less subjective than the other two approaches and is well suited to analysing structurally similar compounds or a homologous series of compounds such as the phthalate esters. Recently, Thomsen et al. [38] used SPR concepts for correlating water solubilities and partition coefficients of phthalate esters against a variety of molecular descriptors and Cousins and Mackay [2] used the “three-solubility” analysis, which correlates solubilities of phthalates in air, water and octanol against molar volume. In this chapter, we include the solubilities and partition coefficients that were estimated by Cousins and Mackay [2] by using the “three-solubility” approach
62
I.T. Cousins, D. Mackay and T.F. Parkerton
[39] (Table 3). Briefly, in the “three-solubility” approach correlations are sought between the solubilities or “apparent-solubilities” of liquid-state compounds in air, water and octanol and molecular descriptors. Cousins and Mackay [2] used Le Bas molar volume [3] as the molecular descriptor, which is readily calculated by summing atomic volumes with adjustment for the volume decrease arising from ring formation. Figures 1–3 show the relationships between molar volumes and the three solubilities with linear regression trend-lines plotted through the points. Data points that were identified as unreliable by Cousins and Mackay [2] have been removed from these plots. The air solubility regression line has been extrapolated beyond the data range to show the expected solubility in air of heavier phthalates. It is currently not possible to reliably measure such low vapour pressures; thus, estimates are used based on extrapolation of the linear regression equation. From the correlations between solubilities and Le Bas molar volume, the partition coefficients KAW (air-water), KOW (octanol-water) and KOA (octanolair) were deduced as the ratios of the “three-solubility” correlations. Further details of the analysis of the physical-chemical properties of phthalate esters by the “three-solubility” approach are described in ref. [2]. The “three-solubility” analysis was particularly useful in highlighting the measurement errors in vapour pressures and solubilities in water for the longer chain phthalate esters. Furthermore, it validated some of the recent meticulous measurements of solubility in water and log KOW carried out by Letinksi et al. [28] and Table 3. Calculated physical-chemical properties of phthalate esters at 25 °C [2]
Phthalate ester
C SWL (mg L–1)
PL (Pa)
Log KOW
Log KOA
Log KAW
H (Pa m3 mol–1)
DMP DEP DAP DPP DnBP DIBP DnPP BBP DHP DnHP DIHpP DnOP BOP 610P DEHP DIOP DnNP DINP DnDP DIDP D711P DUP DTDP
5220 591 156 77 9.9 9.9 1.3 3.8 0.159 0.159 2.00 ¥10–2 2.49 ¥10–3 0.385 8.76 ¥10–4 2.49 ¥10–3 2.49 ¥10–3 3.08 ¥10–4 3.08 ¥10–4 3.81 ¥10–5 3.81 ¥10–5 3.08 ¥10–4 4.41 ¥10–6 7.00 ¥10–8
0.263 6.48 ¥10–2 2.71 ¥10–2 1.75 ¥10–2 4.73 ¥10–3 4.73 ¥10–3 1.28 ¥10–3 2.49 ¥10–3 3.45 ¥10–4 3.45 ¥10–4 9.33 ¥10–5 2.52 ¥10–5 5.37 ¥10–4 1.31 ¥10–5 2.52 ¥10–5 2.52 ¥10–5 6.81 ¥10–6 6.81 ¥10–6 1.84 ¥10–6 1.84 ¥10–6 6.81 ¥10–6 4.97 ¥10–7 3.63 ¥10–8
1.61 2.54 3.11 3.40 4.27 4.27 5.12 4.70 6.00 6.00 6.87 7.73 5.64 8.17 7.73 7.73 8.60 8.60 9.46 9.46 8.60 10.33 12.06
7.01 7.55 7.87 8.04 8.54 8.54 9.03 8.78 9.53 9.53 10.04 10.53 9.37 10.78 10.53 10.53 11.03 11.03 11.52 11.52 11.03 12.02 13.01
–5.40 –5.01 –4.76 –4.64 –4.27 –4.27 –3.91 –4.08 –3.53 –3.53 –3.17 –2.80 –3.73 –2.61 –2.80 –2.80 –2.43 –2.43 –2.06 –2.06 –2.43 –1.69 –0.95
9.78 ¥10–3 2.44 ¥10–2 4.28 ¥10–2 5.69 ¥10–2 0.133 0.133 0.302 0.205 0.726 0.726 1.69 3.95 0.466 6.05 3.95 3.95 9.26 9.26 21.6 21.6 9.26 50.5 275
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
Fig. 1. Relationship between solubility of phthalate esters in water and molar volume
Fig. 2. Relationship between solubility of phthalate esters in air and molar volume
63
64
I.T. Cousins, D. Mackay and T.F. Parkerton
Fig. 3. Relationship between solubility of phthalate esters in octanol and molar volume
Ellington [29], which were found to be consistent with the values estimated by the “three-solubility” analysis (compare measured data listed in Table 2 with estimated values in Table 3).
3 Observations on Physical-Chemical Properties 3.1 Physical State
The phthalate esters contained in Table 1 are liquids at typical environmental temperatures. Melting points lie between 5.5 °C and –58 °C, and boiling points are between 230 and 486 °C [4]. Thus at low environmental temperatures some phthalates have the potential to be present in the solid state. 3.2 Solubility in Water
A declining trend in solubility in water is observed with increasing alkyl chain length or molar volume (Table 3 and Fig. 1). The phthalates have a remarkably large variation in their solubility in water with DMP being moderately soluble at ª 5 g L–1 and DTDP having an estimated solubility some 11 orders of magnitude
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
65
lower at 7 ¥10–11 g L–1. The variability in reported solubility data (Table 2) tends to increase as the measured solubilities decrease, mainly as a result of increased difficulty of measurement. Measuring solubilities in water can be undertaken fairly accurately by standard methods for compounds with solubilities greater than 1.0 mg L–1, but below this limit problems occur; measurement below 1.0 µg L–1 is often unreliable. As discussed by Staples et al. [4], measurement problems are caused by emulsion formation during flask shaking, contamination from apparatus and solvents, and insufficiently low analytical detection limits. Each of these problems leads to measurements that overestimate the true water solubility. Furthermore, Thomsen et al. [40] have argued that phthalate esters with long alkyl chains exhibit weak surface activity, which may cause them to have measured solubilities that are higher than the actual unimeric saturation due to formation of micelles or emulsions. While the water solubility estimate of 15 µg L–1 reported for DEHP by Thomsen et al. based on surface tension measurements is much lower than found in early studies, this technique provides only an indirect measure of solubility. This is due to the fact that surface tension measurements can be affected by impurities present in the test sample (i.e. traces of unreacted monoester). Solubilities in water were directly measured by a “slow-stir” technique to prevent the formation of emulsions and “state-of-the-art” analytical methodology to prevent contamination and to obtain low detection limits [28–30]. “Slow-stir” studies are believed to provide the most accurate measurements of unimeric saturation for phthalates with long alkyl chains, particularly because experimental data were validated by the “three-solubility” analysis of Cousins and Mackay [2]. Turner and Rawling [41] have recently provided additional experimental confirmation concluding that the true water solubility of DEHP is in the few mg L–1 range. Ellington [29] has suggested that phthalate esters with long alkyl side chains (e.g. DnDP) may rotate and fold in aqueous solution to assume conformations of lower energy that more closely resemble a branched-alkyl chain. Furthermore, it is argued that the “effective” molar volume of the “folded” configuration of these phthalates would be less than the molar volume of the unfolded compounds and thus their solubility in water would be higher. It is not possible at this point, with the available data, to properly test this hypothesis, but Cousins and Mackay [2] report a strong linear correlation with the Le Bas Molar volume, which does not take folding into account, for both short- and long-chain phthalates, suggesting that the “folding effect” does not significantly reduce solubilities. Branched-chain isomers of chemicals in general have greater water solubilities than the straight-chain isomers [42], but it was also not possible to observe systematic differences (Table 2) between the straight- and branched-chain phthalates, possibly because measurement error between the various studies is greater than the actual difference in solubility. However, in the study by Letinksi et al. [28], branched isomers were found to systematically exhibit higher solubilities. For example, di-2-ethylhexyl phthalate was found to have a fourfold greater measured solubility than di-n-octyl phthalate. Similar trends were observed for linear and branched analogues with alkyl chains of nine or ten carbons.
66
I.T. Cousins, D. Mackay and T.F. Parkerton
The influence of salinity on the aqueous solubility of non-ionic organic chemicals is given by the equation [43]: log [Csat, w/Csat, salt] = Ks Ms
(a)
where Csat,w denotes the solubility in deionised water, Csat, salt denotes the solubility in the presence of dissolved salts, KS is the Setschenow constant and MS is the total molar concentration of dissolved salts. Turner and Rawling [41] have recently conducted careful solubility experiments with DEHP at different salinities by using radiotracer techniques. Experimental data were fitted to the above equation yielding an estimate for KS of 1.25 L mol–1.Assuming that seawater has a dissolved salt concentration of 0.5 M application of the above equation indicates the DEHP is approximately four times less soluble in salt versus fresh water. This prediction is in reasonable agreement with the ratio of measured values reported for DEHP in lab water (1.9 µg L–1) and seawater (0.6 µg L–1) [20, 21]. 3.3 Vapour Pressure
Vapour pressures similarly show a declining trend with increasing alkyl chain length or molar volume, although the decreasing trend is not as pronounced as for aqueous solubilities. DMP has a vapour pressure of ª 0.3 Pa, that is some seven orders of magnitude higher than the estimated vapour pressure of 4 ¥ 10–8 Pa for DTDP. Measured vapour pressures of a given compound are also highly variable, with variability increasing with increasing alkyl chain length. Measurement problems are encountered below 10–4 Pa and measurement below 10–6 Pa is often unreliable. As discussed by Cousins and Mackay [2], it is probably more accurate to estimate vapour pressures below 10–4 by extrapolation of the relationship between vapour pressure and molar volume than to rely on measured values. It is also possible to extrapolate data from higher temperatures. Unpublished vapour pressure measurements at various temperatures above 60 °C were obtained for DEHP, DINP and DIDP from industry. These data were obtained by using a dynamic vapour pressure balance that is suitable for vapour pressure determinations for these substances at elevated temperatures. Experimental data were fitted to the Clausius-Clapeyron equation: ln PL = A + B/T
(b)
where PL is the vapour pressure in Pa, T is the absolute temperature (K) and A and B are empirical constants derived by linear regression. Results of regression analyses are summarized in Table 4 and indicated a good fit to the above equation as shown in Fig. 4. Extrapolation of the resulting equations to 25 °C yielded vapour pressure estimates which can be compared with recommended estimates derived from the “three-solubility” approach as shown in Table 4.While estimates obtained from high-temperature extrapolation differed by only a factor of three for DEHP, extrapolated values were more than an order of magnitude higher than values recommended by Cousins and Mackay [2]. Given the discrepancy in the vapour pressure estimates obtained from these independent approaches, it is recommended that vapour pressures of high-
67
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
Fig. 4. Vapour pressure data for selected phthalates at elevated temperature
Table 4. Results of regression analysis for vapour pressure measurements obtained at elevated
temperatures and comparison of estimates extrapolated to 25 °C to reported literature Phthalate
A
B
r2
n
PL @25 °C (Pa) Extrapolated
PL @25 °C (Pa) Recommended a
DEHP DINP DIDP
31.08 28.06 27.17
12,090 11,281 11,072
0.998 0.999 0.997
10 87 34
7.6 ¥10–5 5.6 ¥10–5 4.6 ¥10–5
2.5¥10–5 6.8¥10–6 1.8¥10–6
a
Based on “three-solubility” correlation approach [2].
molecular weight phthalates be investigated by using “state-of-the-art” methods at environmentally relevant temperatures. For example, a generator column method could be used [44] in which large volumes of air are passed through a temperature-controlled glass column containing glass wool coated with liquid phthalate. The flow rate through the column is such that the air becomes saturated with phthalate and a sorbent trap collects the phthalate in the air on exiting the column. 3.4 Air-Water Partition Coefficient
The equilibrium distribution of a substance between air and water is known as the air-water partition coefficient (KAW) and is used to indicate the tendency of a substance to escape from water to air. It is noteworthy that volatility of the pure substance is characterized by vapour pressure.Volatility from water is expressed as KAW and for volatilisation from organic media KOA is preferred. KAW is more usually measured in the form of a Henry’s law constant (H, Pa mol m–3), which
68
I.T. Cousins, D. Mackay and T.F. Parkerton
can also be calculated by dividing the vapour pressure (Pa) by the solubility in water (mol m–3). Cousins and Mackay [2] used the “three-solubility” analysis (see Table 3) to calculate values for H and derive values of KAW (from the relationship KAW = H/RT, where R is the gas constant J mol–1 K–1 and T is the absolute temperature). The lower molecular weight phthalate esters have fairly high vapour pressures, but because of their high solubility in water they have very low H or KAW values. Thus, they volatilise fairly rapidly from the pure state but only very slowly from aqueous solution. Because solubilities in water decrease more than vapour pressures (see Figs. 1 and 2) with increasing alkyl chain length (or molar volume), the air-water partition coefficients apparently increase with increasing molecular weight. Thus, the higher molecular weight phthalate esters will potentially evaporate more rapidly from water, but this behaviour is mitigated by sorption to suspended matter in the water column. To illustrate the combined effects of KAW and sorption on air-water distribution it is useful to calculate the concentration in air CA in equilibrium with unit total concentration in water CTW , that is the total of both dissolved and sorbed forms. Now, CA is KAW CW , where CW is the dissolved concentration. Furthermore, CW/CTW or Fdis is the fraction dissolved and can be evaluated by Fdis = 1/(1 + Msusp foc Koc)
(c)
where Msusp is the suspended solids concentration in kg L–1, foc is the organic carbon (OC) fraction of the suspended solids and Koc is the organic carbon normalized solids-water partition coefficient in L kg–1 OC. It follows that CA = KAW Fdis CTW
(d)
Using typical Msusp and foc values of 1.5 ¥ 10–5 kg L–1 and 0.1 and assuming as a first approximation KOC = 0.35 KOW [45], log CA is plotted in Fig. 5 as a function of molar volume for CTW = 1.0. The concentration in air therefore rises at lower molar volumes (because of increasing KAW), but above 400 cm3 mol–1 it drops because Fdis becomes smaller. The net result is that higher molecular weight phthalate esters volatilise only slowly from water. 3.5 Octanol-Water Partition Coefficient
The equilibrium distribution of a substance between water and octanol is known as the octanol-water partition coefficient and is often used to predict the expected partitioning in the environment between water and animal/plant lipids and water and sediment/soil organic matter. There is an abundance of correlations between KOW and measured environmental partition coefficients (e.g. bioconcentration factors, see review by Gobas et al. [46]; soil/sediment organic carbon-water partition coefficients, see reviews by Seth et al. [46] and Doucette [47]; and plant lipid-water partition coefficients, see review by McLachlan [48]). The most common method of measuring KOW is to shake the test substance in a two-phase mixture of octanol and water and to measure the resulting equilibrium concentration in both phases [21]. One of the problems with this technique is that for low solubility, hydrophobic compounds, shaking promotes
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
69
Fig. 5. Preliminary analysis illustrating influence of sorption processes on volatilisation potential of phthalates from surface waters. CA is the concentration in air in equilibrium with unit total (dissolved and sorbed) concentration in water
the formation of emulsions, which will cause concentrations in the water phase to be higher than the solubility in water. To minimize the formation of emulsions the “slow-stir” method has been developed and is believed to give more reliable results for the phthalate esters [29]. Several alternative methods for measuring KOW are available including the use of high-performance liquid chromatography [49]. Log KOW of phthalate esters increases with increasing alkyl chain length or molar volume indicating greater hydrophobicity. For example, the KOW of DMP of ª 66 is about ten orders of magnitude lower than the estimated KOW of DTDP of 1012. Cousins and Mackay [2] have shown that solubilities in octanol are much less sensitive to changes in molar volume than solubilities in water and air. Solubilities in octanol decrease by only about one order of magnitude from DMP to DTDP, whereas solubilities in water decrease by 11 orders of magnitude. Therefore, the reason for the increase in KOW with increasing alkyl chain length is that solubilities in water decrease more per unit volume in molar volume than the solubilities in octanol. It is noteworthy that lipophilicity (which may be thought of as lipid or octanol solubility) does not increase with increasing alkyl chain length as is often wrongly asserted, but actually decreases slightly. The high KOW values of the phthalates indicate these substances are very hydrophobic and will sorb strongly to organic matter and surfaces. However, the potential for bioaccumulation is mitigated by biotransformation as discussed in the chapter focussing on bioaccumulation of phthalates.
I.T. Cousins, D. Mackay and T.F. Parkerton Fig. 6. Relationship between KOW and experimental KOC data. The straight line shows the relationship KOC = 0.35 KOW [46]. Experimental KOC data were taken from ref. [4]
70
One of the major uses of KOW is to estimate the organic-carbon water partition coefficient (KOC) for soils and sediments and thus the soil/sediment-water distribution coefficient (KD). Staples et al. [2] reviewed KOC measurements for phthalates. When KOC and KOW are correlated there is a linear relationship between KOW and KOC for the lower molecular weight phthalates which follows the relationship proposed by Seth et al. [46]: KOC = 0.35 KOW . However, the linear relationship fails for phthalates with a KOW greater than 106 with the result that KOC is poorly predicted by KOW (Fig. 6).
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
71
This discrepancy may be the result of problems in measuring the truly dissolved concentration in water. The phthalates may also be in the form of emulsions or attached to colloidal material, thus the dissolved concentration is overestimated and KOC is underestimated. Recent support for the important role that colloids may serve in reducing apparent KOC values obtained in laboratory tests is provided by Zhou and Liu [50].An alternative explanation is given by Williams et al. [51] who have hypothesized that KOC is inversely dependent on solids concentrations due to particle-particle collisions that induce desorption. These authors calculated particle-corrected KOC values, which were one to three orders of magnitude higher than measured values. An inverse relationship between the DEHP solids-water partition coefficient and solids concentration has been reported in recent lab studies in both surface and marine water [41, 50]. This trend is also apparent from analysis of field monitoring data for DEHP presented by Long et al. [52]. However, it is difficult to differentiate the relative role of colloid binding versus particle interactions, since colloid and solids concentrations typically co-vary. A third explanation is that the time necessary to achieve equilibrium sorption in the organic carbon is long (i.e. months or years) because of the large molar volume of the phthalate and its correspondingly slow diffusion [53]. However, since both lab and field studies produce KOC values below theoretical estimates derived from KOW correlations, the previous two explanations appear more plausible. Turner and Rawling [41] also found that the higher KOC values obtained in seawater versus those in freshwater could not solely be explained by a “salting-out” effect. These authors have hypothesized that additional interactions between salt ions and organic matter facilitate partitioning of DEHP to solids. 3.6 Octanol-Air Partition Coefficient
Harner and Mackay [54] have suggested using the octanol-air partition coefficient (KOA) to describe the partitioning of organic chemicals between air and organic phases in soils, plants and atmospheric aerosols. Recently, highly statistically significant correlations have been found between KOA and measured gas-particle [55, 56], soil-air [57] and plant-air partition coefficients [58, 59] for a range of persistent organic pollutants.Although it is preferable to measure KOA directly, it can be calculated as the ratio KAW/KOW . Measured values are preferred because as a result of the partial miscibility of the octanol-water system KOW is not a true ratio of the solubilities in pure octanol and pure water, but rather the ratio of solubilities in octanol-saturated water and water-saturated octanol [60]. This complication can cause measured KOA values to be up to an order of magnitude higher than values estimated from KAW/KOW . However, since KOA measurements of phthalate esters have not yet been undertaken, the calculation method must be used here to give an indication of their magnitude. Table 3 shows that calculated values of KOA vary between 107 and 1013 and increase with increasing alkyl chain length or molar volume. These high KOA values result in a strong tendency of phthalate esters to partition to aerosols, plants and soils.
72
I.T. Cousins, D. Mackay and T.F. Parkerton
3.7 Data Gaps
Although there are an abundance of physical-chemical property data for the phthalates, the above analysis has revealed that there are many poor data in the literature and still some data gaps that it would be advisable to fill. For example, for phthalates with alkyl chains of more than six carbon atoms long, limited published vapour pressure data are available and new reliable measurements are needed using state-of-the-art methods. Furthermore, it would be advisable to measure KOA values in the laboratory, since calculated estimates may be in error by an order of magnitude. Finally, we recommend that such properties are measured at a range of temperatures representative of environmental conditions (i.e. –40 to +40 °C) because the environment is rarely at 25 °C. From this information it will be possible to calculate enthalpies of phase change, which allow calculation of a property at any temperature. Further research is also needed to better quantify the influence of particulate and colloidal organic carbon in freshwater and marine waters on KOC values selected for evaluative environmental fate modelling.
4 Degrading Reactions Biodegradation is the dominant loss process for phthalate esters in all media except the atmosphere where they are likely to be susceptible to rapid photo-oxidation by hydroxyl radicals [4]. Both aerobic and anaerobic microbial breakdown of phthalate esters is initiated by ester hydrolysis to form the monoester and the corresponding alcohol. The monoester will be further enzymatically degraded to phthalic acid, which will be mineralised following a series of steps. A full review of the reaction pathways has been reported by Ejlertsson and Svensson [61]. It is now common practice to simplify environmental reaction kinetics and assume that a first-order rate constant (or half-life) can be applied to estimate the loss from each medium. This is necessarily an approximation of the truth and involves a judgment that, in a particular type of soil or water, the compound is subject to biodegradation with a half-life of x hours [62]. Furthermore, only the transformation of the parent compound to the primary metabolite is considered; thus, primary rather than ultimate degradation half-lives are estimated. It is noteworthy that degradation half-lives of organic compounds cannot be viewed in the same way as radionuclide half-lives, which are a fundamental, reproducible property of the radionuclide. Degradation half-lives are functions of both the chemical and the environment. In a series of handbooks by Mackay and coworkers [62], half-lives are assigned on a semi-decade logarithmic scale into one of nine classes (Class 1: 5 h (3–10 h), 2: 17 h (10–30 h), 3: 55 h (30–100 h), 4: 170 h (100 – 300 h), 5: 550 h (300 – 1000 h), 6: 1700 h (1000 – 3000 h), 7: 5500 h (3000 – 10,000 h) 8: 17,000 h (10,000 – 30,000 h) and 9: 55,000 h (30,000–100,000 h)). A chemical can be assigned for example to a half-life class 3 with a geometric mean of 55 h and a range of 30–100 h, even though at different times and places the half-life may fall into class 2 or 4, with perhaps a 5% probability of reaching class 1 or 5.
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
73
Photo-oxidation half-lives, based on hydroxyl radical attack, for phthalate esters were predicted by Staples et al. [4] by using the Atmospheric Oxidation Program (AOP) [63] and these varied between 0.2-2 days for DEHP and 9.3–93 days for DMP. There is one experimental study to validate these predictions, which reported an atmospheric half-life for DEHP of about 1 day [64]. For the Level II and III simulations presented in this paper, half-lives in air were assigned based on the AOP predicted data. Primary aerobic biodegradation half-lives of phthalate esters in natural waters and soils have been estimated by Staples et al. [4] by an analysis of available measured data. Aerobic biodegradation half-lives in natural waters and soils tend to increase with increasing alkyl chain length. DMP, DEP, DnBP and DEHP were shown to have aerobic biodegradation half-lives in natural waters of 0.2–10, 0.3–12, 1.7–2.5 and 2–22 days, respectively, and half-lives in soils of 1–40, 1–75, 0.4–80 and 25–250 days, respectively. Rates of degradation in aerobic soils were shown to be 2-5 times slower than in aerobic aquatic environments. This is thought to be due to reduced bioavailability of phthalates resulting from sorption to soil organic matter. Staples et al. [4] reported that there are only limited data available on biodegradation of phthalate esters in sediments, but the data suggest that primary biodegradation of phthalate esters in sediments is slower than in soils and of the order of several months (>100 days). BBP is a special case in that it does not contain two straight alkyl chains in its structure and thus the mechanism for primary degradation is likely to be different. Measured data for BBP from Staples et al. [4] suggest that its aerobic biodegradation half-lives in natural waters, soils and sediments are 0.4 day to 8 ¥ 104 days, 9.6 days and 1.6–2.2 days, respectively. The value of 8 ¥ 104 days seems to be erroneous because there are five other data points in the range 0.4–1.4 days. There is only one data point for aerobic biodegradation in soil for BBP, which is particularly disappointing because soil is the primary medium of accumulation for BBP. Degradation rates of phthalate esters in anaerobic media are slower, but the models used in this chapter only treat aerobic environmental media. Only surface soils (top 5 cm) and sediments (top 3 cm) are treated. The above analysis of measured biodegradation half-life data from Staples et al. [4] has been used to allocate approximate half-lives by using a semi-decade logarithmic scale for water, soil and sediment compartments in the EQC Level II and III simulations (Table 5). This approximate allocation takes account of the large uncertainty in measured biodegradation half-lives. We have taken a conservative approach in our allocation of half-lives and assigned half-lives that are near to the top of the range reported by Staples et al. [4]. This conservative approach results in estimated half-lives that are higher than those suggested in the chapter of this handbook focussing on environmental degradation rates of phthalates. However, it is believed that a conservative approach is appropriate for allocation of half-lives because degradation studies are often conducted at a constant 25 °C, whereas the environment is often at a lower temperature, some studies use inoculums and some allow the microbial population to become acclimated. Furthermore, some microcosm studies may not separate losses from degradation from losses due to partitioning to sediments and volatilisation.
74
I.T. Cousins, D. Mackay and T.F. Parkerton
Table 5. Allocation of half-lives for use in Level II and III EQC simulations
Phthalate ester
DMP
DEP
DnBP
BBP
DEHP
Assumed reaction half-lives (h)
Class Mean Range Class Mean Range Class Mean Range Class Mean Range Class Mean Range
Air
Water
Soil
Sediment
5 550 300–1000 4 170 100–300 3 55 30–100 2 17 10–30 2 17 10–30
4 170 100–300 4 170 100–300 4 170 100–300 3 55 30–100 5 550 300–1000
5 550 300–1000 5 550 300–1000 6 1700 1000–3000 6 1700 1000–3000 7 5500 3000–10,000
6 1700 1000–3000 6 1700 1000–3000 7 5500 3000–10,000 6 1700 1000–3000 7 5500 3000–10,000
5 Evaluative Fate Modelling with the EQC Model Conducting evaluative assessments can provide invaluable insights into the characteristics of chemical behaviour in the environment. Because the environment considered is purely evaluative or hypothetical, there is no possibility of validation, but the equations used to describe partitioning, transport and transformation are identical to those used successfully in validated models of chemical fate in more defined environments. The aim is to establish the general features of chemical behaviour, namely, into which media the chemical will tend to partition, the primary loss mechanisms, the tendency for intermedia transport, the tendency to bioaccumulate, the tendency to undergo long-range transport and environmental persistence. Multimedia models of this type are widely used by the scientific community as useful tools for providing information on chemical fate and have also found acceptance in regulatory practice in a number of countries. The Equilibrium Criterion or EQC model, the model of choice here, has been described fully elsewhere [65]. Briefly, this model in the form of a computer program, deduces the fate of a chemical in Level I, II and III evaluative environments by using principles described by Mackay [36]. The EQC evaluative environment is an area of 100,000 km2 that is regarded as being representative of a jurisdictional region such as the US state of Ohio, or the country of Greece. EQC can simulate the chemical fate of a variety of different chemical class types, classified according to the data requirements to run a model simulation. Phthalate esters partition to all environmental media and are thus classified as type 1 chemicals
75
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
for which all partition coefficients and Z values (fugacity capacities) must be defined [36]. 5.1 EQC Level I Modelling
EQC Level I modelling has been performed for the 22 phthalate esters listed in Table 1 for which physical-chemical properties have previously been estimated (Table 3). Level I EQC model results indicate that under equilibrium, steady state conditions, with no reaction, the vast majority of phthalates released will reside in soil, sediment or water with over 99% being distributed to these three media (Table 6). The low vapour pressures ensure that only small percentages partition to air. Phthalate esters with alkyl chains containing greater than five carbons partition almost exclusively to the organic carbon component of soil and sediment, whereas those with short alkyl chains (<4 carbons), and hence lower hydrophobicities, partition readily to water. Table 6. Summary of results of Level I and II simulations using the EQC model
Phthalate ester
Level I distribution Air (%)
DMP DEP DAP DPP DnBP DIBP DnPP BBP DHP DnHP DIHpP DnOP BOP 610P DEHP DIOP DnNP DINP DnDP DIDP D711P DUP DTDP
0.2 0.4 0.4 0.4 0.2 0.2 0.1 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
Level II distribution
Water (%)
Soil (%)
Sediment (%)
Persistence (days)
Loss by reaction (%)
96.3 75.8 46.0 30.4 5.6 5.6 0.8 2.2 0.1 0.1 <0.1 <0.1 0.3 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
3.5 23.3 52.4 67.7 92.2 92.2 96.9 95.6 97.6 97.6 97.7 97.8 97.5 97.8 97.8 97.8 97.8 97.8 97.8 97.8 97.8 97.8 97.8
0.1 0.5 1.2 1.5 2.1 2.1 2.2 2.1 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2
8.4 9.9 – – 28 – – 27 – – – – – – 34 – – – – – – – –
80 81 – – 95 – – 98 – – – – – – 100 – – – – – – – –
76
I.T. Cousins, D. Mackay and T.F. Parkerton
5.2 EQC Level II Modelling
“Persistence in the environment” can be operationally defined as the overall residence time at steady state in a multimedia environment [66]. Level II and Level III calculations can both give estimates of persistence. Level II calculations give a preliminary estimate of persistence based on equilibrium partitioning, whereas Level III provides a more sophisticated albeit data-intensive estimate of persistence that accounts for medium of discharge and intermedia transport rates. A Level II approach is becoming an increasingly popular method because of its simplicity and low input data requirements [67]. Level II and III EQC simulations have been performed on five phthalate esters: DMP, DEP, DBP, DBP and DEHP. These representatives possess the best data sets for characterizing compartmental half-lives that are required for these calculations while spanning a wide range in physical chemical properties characteristic of this chemical class. Level II EQC modelling predicts an increasing environmental persistence with increasing alkyl chain length (Table 6). Increasing half-lives and tendency to partition to soils and sediments reduce advection and combine to increase the overall residence time as the alkyl chain length increases. 5.3 EQC Level III Modelling
Level III modelling is useful for determining how the medium of release affects environmental fate. Level III fugacity calculations allow non-equilibrium conditions to exist between connected media as steady state, and illustrate important transport and transformation processes. The tendency of chemicals to migrate between media can be assessed by modelling emissions to each individual medium and calculating the amount present at steady state. Table 7 shows the amount of chemical present in each medium of the EQC model environment for individual emissions of 1000 kg h–1 to the air, water and soil compartments, as well as a “total” for simultaneous emissions of 1000 kg h–1 to each compartment. Figures 7 and 8 show two example diagrams of the model output for two phthalates (DMP and DEHP) with contrasting alkyl chain lengths. The Level III analysis shows that DEHP tends to partition out of air and water and into soil and sediments respectively and this is likely to be the pattern for other high-molecular weight phthalates with long alkyl chains. The lower molecular weight phthalates tend to remain in the medium of release if emitted to air and water, but if emitted to air, they are deposited and accumulate in soil. Important transport processes for the phthalates that should be the focus of research are atmospheric deposition (both wet and dry) and sediment deposition. Important associated processes that should also be studied are gas-particle and sediment-water partitioning. The Level III model can also be used to derive residence times in the model world, which can be thought as “characteristic times” describing the dynamics of phthalates in the system (Table 7). For example, the model predicts that when to-
Air Water Soil “Total” Air Water Soil “Total” Air Water Soil “Total” Air Water Soil “Total” Air Water Soil “Total”
DMP
a
26,400 (5.8) 5 (<0.1) 755 (0.1) 27,100 (2.1) 39,300 (11.3) 19 (<0.1) 334 (<0.1) 39,600 (3.0) 38,700 (13.8) 99 (<0.1) 27.8 (<0.1) 38,900 (1.3) 18,600 (14.4) 26.7 (<0.1) 5 (<0.1) 18,600 (0.7) 13,200 (0.5) 20 (<0.1) <1 (<0.1) 13,200 (0.1)
Air 42,800 (9.4) 197,000 (99.7) 43,400 (6.6) 283,000 (21.6) 15,600 (4.5) 197,000 (99.3) 10,000 (1.3) 222,000 (17.0) 5650 (2.0) 195,000 (74.3) 699 (<0.1) 201,000 (6.7) 984 (0.7) 72,600 (74.6) 108 (<0.1) 73,700 (2.8) 5,460 (0.2) 161,000 (3.9) 128 (<0.1) 166,000 (1.1)
Water 385,000 (84.8) 75 (<0.1) 612,000 (93.3) 997,000 (76.2) 292,000 (84.2) 141 (0.1) 749,000 (98.6) 1,040,000 (79.8) 234,000 (83.5) 601 (0.2) 2,440,000 (>99.9) 2,680,000 (89.7) 109,000 (84.6) 51 (<0.1) 2,450,000 (>99.9) 2,560,000 (95.6) 2,340,000 (93.4) 3,520 (<0.1) 7,930,000 (>99.9) 10,300,000 (68.7)
Soil
Amount at steady state (kg) (percent in brackets)
The residence time multiplied by 3.
DEHP
BBP
DnBP
DEP
Emission medium
Phthalate ester 99 (<0.1) 455 (0.2) 100 (<0.1) 654 (0.1) 93.8 (<0.1) 1180 (0.6) 60 (<0.1) 1,340 (0.1) 1940 (0.7) 66,700 (25.5) 239 (<0.1) 68,900 (2.3) 335 (0.3) 24,700 (25.4) 36.8 (<0.1) 25,100 (0.9) 148,000 (5.9) 4,350,000 (96.4) 3,470 (<0.1) 4,500,000 (30.1)
Sediment
Table 7. EQC Level III results: chemical amounts in each medium based on single and multiple emissions
19 8.2 27 18 15 8.3 32 18 12 11 100 42 5.4 4.1 100 37 100 190 330 210
Residence time (d) 57 25 81 54 45 25 96 54 36 33 300 130 16 12 300 110 300 570 1000 630
Time to 97% steady state/clearance (d) a
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
77
78
0.399 kg/h
6.33 e · 4 kg/h
Sediment
0.456 kg/h
31.6 kg (6.57 e-3%) Fug. = 2.85 e · 3 mPa 0.049 ng/g
0.057 kg/h
1121 kg/h 88977 kg (18.5%) Fug. = 0.022 mPa 445 ng/L
Fig. 7. Diagram showing EQC Level III output for DMP
Persistence = 161 h = 6.688 days
INTERMEDIA EXCHANGE
ADVECTION
Total Mass = 4.82 e + 5 kg
1000 kg/h
EMISSION
Legend:
1000 kg/h
REACTION
1493 kg/h
Soil
3.66 e + 5 kg (76.0%) Fug. = 1.463 mPa 13.6 ng/g
641 kg/h
263 kg/h
79.9 kg/h
Water
0.088 kg/h
33.2 kg/h
1000 kg/h
EQC Model v. 1.0 Level III
17.4 kg/h
Air
26349 kg (5.472%) Fug. = 3.362 mPa 263 mg/m3
Chemical: DMP
263 kg/h
89.0 kg/h
I.T. Cousins, D. Mackay and T.F. Parkerton
tal emissions of DEHP to the environment are 1000 kg h–1 to air (or 8,760,000 kg year–1), the total inventory in the environment is 2,507,000 kg.As a result the characteristic time is 2,507,000/8,760,000 or 0.29 years or about 104 days. A dynamic model of DEHP under initial conditions of zero concentrations followed by sustained constant emissions would display an approach to a steady state that would be essentially 97% complete after three characteristic times or 0.85 years or 312 days. Similarly, if DEHP emissions were stopped entirely the system would be
79
673 kg/h
10.7 kg/h
Persistence = 1293 h = 53.9 days
INTERMEDIA EXCHANGE
ADVECTION
Total Mass = 3.88 e + 6 kg
1000 kg/h
REACTION
1295 kg/h
EMISSION
Legend:
1000 kg/h
Fig. 8. Diagram showing EQC Level III output for DEHP
Sediment
5.34 e + 5 kg (13.8%) Fug. = 0.026 mPa 835 ng/g
10.7 kg/h
195 kg/h 0.319 kg/h
694 kg/h
Water 33.8 kg/h
Soil
3.18 e + 6 kg (81.9%) Fug. = 3.38 mPa 118 ng/g
296 kg/h 1.25 e · 3 kg/h
1000 kg/h
EQC Model v. 1.0 Level III
1.54 + 5 kg (3.980%) Fug. = 0.213 mPa 772 ng/L
1.446 kg/h
540 kg/h
13238 kg (0.341%) Fug. = 0.146 mPa 132 mg/m3
Air
Chemical: DEHP
132 kg/h
154 kg/h
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
clear of DEHP almost entirely in 0.85 years. The implication is that, since DEHP has been in use for many years, it must have reached a steady-state condition in the environment, thus the use of a steady-state model is justified for regional mass balance calculations. This argument applies equally for all the phthalates (Table 7). The medium of release greatly affects the estimated environmental residence time with emissions to the relatively slowly reacting soil compartment resulting in the longest residence times.
80
I.T. Cousins, D. Mackay and T.F. Parkerton
5.4 Estimating Persistence and Long-Range Transport Potential with the TaPL3 Model
The TaPL3 model, which was developed specifically for estimating environmental persistence and long-range transport (LRT) potential [66, 68], is identical in properties to EQC, but it is set to mimic a closed system, from which no advective losses occur (via air or water flows). By considering only irreversible losses, a conservative estimate of persistence is obtained that may better reflect the removal of chemicals from the environment as a whole. Environmental persistence is evaluated by running various hypothetical releases to the TaPL3 model, usually 1000 kg h–1 to air, water and soil, one at a time, and then 1000 kg h–1 to all three media simultaneously. These hypothetical scenarios are used because they allow comparison with other chemicals on an equal basis and because release information is often not available. For phthalates, Parkerton and Konkel [69] have estimated the amounts of phthalates released to various environmental media and have concluded that the vast majority of phthalates are released to the environment from product end use to the atmosphere. Thus, the emission to air scenario may give the most realistic estimate of phthalate persistence in the environment. The calculated overall environmental residence times, which we define as “environmental persistence”, for emission to air and water are presented in Table 8 and compared to estimates for well-known persistent organics pollutant (POPs). It is concluded that phthalates are not persistent to the same extent as, for example, PCBs or dioxins. McLachlan and Horstmann [70] have suggested that uptake of organic compounds by vegetation from the atmosphere is important for compounds with octanol-air partition coefficients (KOA) greater than 108. Phthalate esters have KOA values that vary between 107 and 1013; thus, uptake by vegetation is likely to be important. The environmental persistence may be reduced due to the high reactivity of some organic compounds on vegetation surfaces. Unfortunately, there is currently a lack of data on metabolism rates of phthalate esters in vegetation making it difficult to assess the effect of vegetation uptake on persistence. Table 8. Overall environmental persistence and characteristic travel distances for total inputs
into either air or water calculated using TaPL3 [66, 68] Substance
Overall persistence for emission to air (d)
Characteristic travel distance in air (km)
Overall persistence for emission to water (d)
Characteristic travel distance in river water (km)
DMP DEP DBP BBP DEHP HCB a-HCH tetra-PCB 2,3,7,8-TCDD
26 24 19 6.6 120 1,300 150 2,500 860
520 1000 910 330 220 110,000 12,000 8,900 810
10 10 14 4.4 260 1,600 230 2,900 1,800
880 880 870 280 700 2,600 14,000 2,900 1,300
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
81
There are two main factors which control a chemical’s LRT potential: (1) persistence in the atmosphere or water column which can be characterized by a halflife and (2) “stickiness”, the propensity for a chemical to partition to the terrestrial surface if transported in air, or its propensity to partition to sediment if transported in water. TaPL3 [68] is used here to calculate the characteristic travel distance of the phthalate esters in air and water (Table 8). The comparison between phthalates and other compounds in Table 8 indicates that phthalates as a class do not exhibit unfavourable long-range transport characteristics. The lower molecular weight phthalates are reactive and the heavier phthalates are “sticky” ensuring that they do not travel far in air or water. The fact that phthalates are frequently detected at low concentrations in environmental media is likely to be a function of their high production volume and widespread usage in polymeric materials rather than a high potential for transport or persistence in the environment.
6 Conclusions The phthalate esters display a remarkable range of properties extending over 6–11 orders of magnitude. Especially important are the marked decreases in volatility and solubility in water with increased molar volume or alkyl chain length. Measuring these properties for the higher molecular weight phthalates is extremely difficult and new methodologies are required for reliable measurement. The very low solubilities in water of the octyl and high-molecular weight phthalates will cause bioconcentration and toxicity experiments to be extremely demanding, especially because the high KOW values will result in strong sorption to dissolved organic carbon and surfaces. The wide variations in physical-chemical properties are reflected in significant differences in environmental partitioning, which have been illustrated using an evaluative fate model. The lower molecular weight phthalate esters are quite volatile, but owing to their very low KAW values they will volatilise fairly rapidly from the pure state but only very slowly from aqueous solution. The log KOW values vary from 1.61 to 12.06; thus, the high-molecular weight esters are very hydrophobic and will sorb strongly to organic matter and surfaces. The high values of KOA suggest that any higher molecular weight esters present in the atmosphere will be appreciably sorbed to aerosol particles and to soil and vegetation. Air-water partition coefficients increase with increasing molecular weight; thus, the higher molecular weight phthalate esters will potentially evaporate more rapidly from water, but this will be mitigated by sorption to suspended matter in the water column. The phthalate esters also show significant and systematic differences in reactivity or half-life, with the primary biodegradation half-life increasing with increasing alkyl chain length and the photo-oxidation half-life showing the opposite trend. These changes in environmental persistence of phthalates have been illustrated by using the TaPL3 model. Based on these calculations it is concluded that phthalates as a class are not as persistent as other well-known chlorinated organic pollutants. Although the environmental persistence of higher molecular weight phthalates tends to be greater, KOA and thus propensity to partition to aerosols,
82
I.T. Cousins, D. Mackay and T.F. Parkerton
vegetation and soils also increases. This increased “stickiness” tends to limit the relative ability of the heavier phthalates to be transported in air or water as has been demonstrated with the TaPL3 model. The evaluative modelling exercise undertaken here is valuable for assisting the evaluation of experimental results on the fate of phthalates. Ultimately, these models require verification from well-designed monitoring programs, which seek to determine the actual fate of phthalates in our complex, variable and real environment. Evaluation of the performance of mass balance models in predicting environmental partitioning and fate is the focus of a later chapter. Acknowledgement. We are grateful to the Phthalates Esters Panel of the American Chemistry Council (ACC) for funding this research, and to NSERC and the consortium of chemical companies that provide finances to support the Canadian Environmental Modelling Centre.
7 References 1. ECETOC (1988) Technical Report No 29 2. Cousins IT, Mackay D (2000) Chemosphere 41:1389 3. Reid RC, Prausnitz JM, Polling BE (1987) The properties of gases and liquids, 4th edn. McGraw-Hill, New York 4. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:667 5. Nielsen NM, Bundgaard H (1989) J Med Chem 32:727 6. Howard PH, Banerjee S, Robillard KH (1985) Environ Toxicol Chem 4:653 7. Veith GD, Macek KJ, Petrocelli SR, Carroll J (1980) An evaluation of using partition coefficients and water solubility to estimate bioconcentration factors for organic chemicals. In: Eaton JG, Parrish PR, Hendricks AC (eds) Fish, aquatic toxicology ASTM STP 707.American Society for testing and Materials, pp 116–129 8. Leyder F, Boulanger P (1983) Bull Environ Contam Toxicol 30:152 9. Wolfe NL, Steen WC, Burns LA (1980) Chemosphere 9:403 10. Renberg LO, Sundström SG, Rosen-Olofsson A-C (1985) Toxicol Environ Chem 10:333 11. McDuffie B (1981) Chemosphere 10:73 12. Ellington JJ, Floyd TL (1996) Octanol/water partition coefficients for eight phthalate esters. United States Environmental Protection Agency: Environmental Research Brief. Report no EPA/600/S-96/006 13. Russell DJ, McDuffie B (1986) Chemosphere 15:1003 14. Grayson BT, Forsbraey LA (1982) Pest Sci 13:269 15. De Kock AC, Lord DA (1987) Chemosphere 16:133 16. DeFoe DL, Holcombe GW, Hammermeister DE (1990) Environ Toxicol Chem 9:623 17. Dobbs AJ (1984) Chemosphere 13:687 18. Werner AC (1952) Ind Eng Chem 44:2736 19. Harnisch M, Möckel HJ, Schulze G (1983) J Chromat 282:315 20. Gückel W, Kästel R, Lewerenz J, Synnatschke G (1982) Pest Sci 13:161 21. OECD guidelines for testing of chemicals (1981) “Partition coefficient (n-octanol-water) flask shaking method”, Section 107. Organisation for Economic Cooperation and Development, Paris, France 22. Frissel WJ (1955) Ind Eng Chem 18:1096 23. Quackenbos HM Jr (1954) Ind Eng Chem 46:1335 24. Haky JE, Leja B (1986) Anal Lett 19:123 25. Hansch C, Leo A (1987) LogP database. Pomona College, Claremont, CA 26. Hollifield HC (1979) Bull Environ Contam Toxicol 23:579 27. Gledhill WE, Kaley RG, Adams WJ, Hicks O, Michael PR, Saeger VW (1980) Environ Sci Technol 14:301
Physical-Chemical Properties and Evaluative Fate Modelling of Phthalate Esters
83
28. Letinksi DJ, Connelly MJ, Parkerton TF (1999) Slow-stir water solubility measurements for phthalate ester plasticisers. Paper presented at SETAC Europe, 9th Annual Meeting, Leipzig, Germany, May 25–29, p 159 29. Ellington JJ (2000) J Chem Eng Data 44:1414–1418 30. De Bruijn J, Busser F, Seinen W, Hermens J (1989) EnvironToxicol Chem 8:499 31. Dobbs AJ, Cull MR (1982) Environ Poll (Series B) 289 32. Brooke D, Nielsen I, De Bruijn J, Hermens H (1990) Chemosphere 21:119 33. OECD guidelines for testing of chemicals (1981) “Vapour pressure curve (dynamic method – static method – isoteniscope – vapour pressure balance – gas saturation method)”, Section 104. Organisation for Economic Cooperation and Development, Paris, France 34. Klein W, Kördel W, Weiss M, Poremski HJ (1988) Chemosphere 17:361 35. Exxon Biomedical Sciences Inc. (1996) Water solubility study no 199638. East Millstone, NJ, USA 36. Mackay D (2001) Multimedia environmental models: the fugacity approach, 2nd edn. Lewis, Boca Raton, London 37. Kollig HP (1988) Toxicol Environ Chem 17:287 38. Thomsen M, Rasmussen AJ, Carlsen L (1999) Chemosphere 38:2613 39. Cole JG, Mackay D (2000) Environ Toxicol Chem 19:265 40. Thomsen M, Carlsen L, Hvidt S (2001) Environ Toxicol Chem 20:127 41. Turner A, Rawling MC (2000) Marine Chem 68:203 42. Howard PH (2000) Biodegradation. In: Boethling S, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton 43. Schwarzenbach RP, Gschwend PM, Imboden DM (1993) Environmental organic chemistry. Wiley and Sons, New York 44. Wania F, Shiu WY, Mackay D (1994) J Chem Eng Data 39:572 45. Gobas FAPC, Morrison HA (2000) Bioconcentration and biomagnification in the environment. In: Boethling S, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton 46. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:2390 47. Doucette WJ (2000) Soil and sediment sorption coefficients. In: Boethling S, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton 48. McLachlan MS (2000) Vegetation-air partition coefficients. In: Boethling S, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton 49. Leo A (2000) Octanol-water partition coefficients. In: Boethling S, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. Lewis, Boca Raton 50. Zhou JL, Liu YP (2000) Marine Chem 71:165 51. Williams MD, Adams WJ, Parkerton TF, Biddinger GR, Robillard KA (1995) Environ Toxicol Chem 14:1477 52. Long JLA, House WA, Parker A, Rae JE (1998) Sci Tot Environ 210/211:229 53. Cornelissen G, van Noor PCM, Govers HAJ (1997) Environ Toxicol Chem 16:1351 54. Harner T Mackay D (1995) Environ Sci Technol 29:1599 55. Finizio A, Mackay D, Bidleman T, Harner T (1997) Atmos Environ 31:2289 56. Harner T, Bidleman T (1998) Environ Sci Technol 32:1494 57. Hippelein M, McLachlan MS (1998) Environ Sci Technol 32:310 58. Kömp P, McLachlan, MS (1997) Environ Sci Technol 31:2944 59. Hiatt MH (1999) Environ Sci Technol 33:4126 60. Chiou CT, Schmedding DW, Manes M (1982) Environ Sci Technol 16:4 61. Ejlertsson J, Svensson B (1995) A review of the possible degradation of polyvinyl chloride (PVC) plastics and its component phthalic acids and vinyl chloride under anaerobic conditions prevailing in landfills. Dept of Water and Environmental Studies, Linkoping University, Sweden
84
I.T. Cousins, D. Mackay and T.F. Parkerton
62. Mackay D, Shiu W-Y, Ma K-C (1999) Physical-chemical properties and environmental fate handbook, CRC netBASE CD-ROM. Chapman and Hall/CRC Press, Boca Raton, FL, USA 63. Meylan WM, Howard PH (1993) Chemosphere 26:2293 64. Behnke W, Nolting F, Zetzsch C (1987) An aerosol smog chamber for testing abiotic degradation. In: Greenhalgh R, Roberts TR (eds) Pesticide science biotechnology proceedings of sixth international congress on pesticide chemistry. Blackwell, Oxford, UK, pp 401–404 65. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:1627 66. Webster E, Mackay D, Wania F (1998) Environ Toxicol Chem 17:2148 67. Gouin T, Mackay D, Webster E, Wania F (2000) Environ Sci Technol 34:881 68. Beyer A, Mackay D, Matthies M, Wania F, Webster E (2000) Environ Sci Technol 34:699 69. Parkerton TF, Konkel WJ (2000) Evaluation of the production, consumption, end use and potential emissions of phthalate esters. Prepared for the American Chemistry Council (Draft of November 2000) by Exxon Mobil Biomedical Sciences Inc (EMBSI), Annandale, NJ, USA 70. McLachlan MS Horstmann M (1998) Environ Sci Technol 32:413
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 85– 124 DOI 10.1007/b11464
Degradation of Phthalate Esters in the Environment Dennis R. Peterson 1 · Charles A. Staples 2 1 2
ExxonMobil Biomedical Sciences, Inc., 1545 Route 22 East, P.O. Box 971, Annandale, New Jersey 08801–0971, USA. E-mail:
[email protected] Assessment Technologies, Inc., 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA
This chapter reviews the degradation data that are appropriate for modeling the environmental concentration of phthalate diesters. The necessary data are the first-order rates of primary degradation in air, water, and soil. The predominant fate of PDEs is biodegradation in wastewater, aerobic aquatic (water/sediment) environments, and soil. Anaerobic degradation may also play an important role for some phthalates under certain conditions. Of the abiotic processes, the rates of hydrolysis and direct photolysis are too low to have a significant influence on overall phthalate fate. On the other hand, indirect photolysis in air due to hydroxyl radical attack in both vapor phase and particle-sorbed phthalates may have a significant role in the overall environmental degradation of some phthalates. The literature data for these environmental degradation rates are evaluated with regard to their relevance to actual environmental conditions and recommendations are presented for the most likely degradation rates for specific phthalate diesters. The rates of further degradation of phthalate diester degradation products are also briefly discussed. Keywords. Phthalate ester, Hydrolysis, Photolysis, Biodegradation, First-order rate
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
1
Introduction
2
Selection of Data for the Evaluation of Environmental Persistence
2.1 2.2
Realism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Primary Versus Ultimate Degradation . . . . . . . . . . . . . . . . 87
3
Abiotic Degradation . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.1 3.2
Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4
Biodegradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.1 4.1.1 4.1.2 4.1.3 4.2 4.2.1
Surface Water . . . . . . . . . . . . . . . . Reported Rates . . . . . . . . . . . . . . . Bioavailability . . . . . . . . . . . . . . . . Biodegradation of Very Low Concentrations Soil . . . . . . . . . . . . . . . . . . . . . Reported Rates . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
87
. . . . . .
90 90 95 97 98 98
© Springer-Verlag Berlin Heidelberg 2003
86
D.R. Peterson and C.A. Staples
4.2.2 4.3 4.4 4.5
Conditions Affecting Soil Biodegradation Wastewater . . . . . . . . . . . . . . . . Anaerobic Rates . . . . . . . . . . . . . . Solid-Phase Biodegradation . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
100 105 110 113
5
Degradation of Phthalate Metabolites . . . . . . . . . . . . . . . . 114
6
Summary and Conclusions
7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
. . . . . . . . . . . . . . . . . . . . . 118
Abbreviations BBP DBP DAP DDP DEHP DEP DHP DIBP DIDP DIHP DINP DIUP DMP DnOP DNP DOP DPP DPrP
Butylbenzyl phthalate Dibutyl phthalate Diamyl phthalate Didecyl phthalate Bis(2-ethylhexyl) phthalate Diethyl phthalate Dihexyl phthalate Diisobutyl phthalate Diisodecyl phthalate Diisoheptyl phthalate Diisonyl phthalate Diisoundecyl phthalate Dimethyl phthalate Di-n-octyl phthalate Dinonyl phthalate Dioctyl phthalate Di-n-pentyl phthalate Dipropyl phthalate
1 Introduction Phthalate diesters (PDEs) are commonly released into the environment and are rapidly degraded. The main degradation route of PDEs is thought to be biodegradation. The biodegradability of PDEs has been reviewed extensively in the past. The main subject of this review will be data on the rate of degradation of PDEs in the environment, with particular emphasis on biodegradation. The rate of degradation of chemicals in the environment is important for modeling their overall environmental fate and, in particular, their overall degree of persistence and transport. The modeling of the environmental fate and concentration of PDEs is covered elsewhere in this volume. Recent reports on the results of standard biodegradation tests will only be mentioned briefly and more emphasis will be placed upon evaluation of the rate of primary biodegradation measured under environmentally realistic situations. Furthermore, the physical properties of PDEs generally result in their association
Degradation of Phthalate Esters in the Environment
87
with particulate matter in air and water and also to their strong partitioning to soil and sediment. Thus, particular attention will be paid to the studies of degradation rates in soil and sediment.
2 Selection of Data for the Evaluation of Environmental Persistence 2.1 Realism
The rate data need to be as relevant as possible to the natural environment, since these data are used to evaluate environmental persistence and provide data for environmental fate models. Data from simulation studies, microcosms, and field studies are the most applicable. Because the rate of biodegradation of a chemical is dependent both upon its concentration and the concentration of degrading microorganisms, the most useful results are those that are done at environmentally realistic concentrations of the chemical and with microorganism densities that are typical of a particular environmental compartment. Environmental conditions such as temperature, nutrient availability, oxygen availability, and substrate bioavailability all affect the rate of biodegradation. The influence of experimental conditions will be evaluated when reviewing rate data. 2.2 Primary Versus Ultimate Degradation
The degradation data that are of use for modeling and for evaluation of environmental persistence are the pseudo-first-order disappearance rates of the PDEs. Environmental fate models utilize physical-chemical properties such as water solubility and vapor pressure to calculate the partitioning behavior of a chemical in the environment. These physical chemical properties differ greatly between the “parent” PDE and its degradation products. The overall decrease in concentration of the parent chemical in the environment is evaluated from its degradation rates in the various environmental compartments. In the terminology of biodegradation, it is the rate of primary degradation that is needed. For abiotic degradation, such as hydrolysis and photolysis, the standard tests also measure the rate of decrease in parent chemical concentration, that is, primary degradation. In addition to the use of primary degradation rate data in environmental fate modeling, evaluation of the property of persistence within the context of “persistent organic pollutants” also implies primary degradation, since a chemical structure that is degraded cannot also persist. This does not mean that the degradation products are of no consequence. In assessing the overall risk of a chemical to the environment, the hazard and fate of the degradation products are important considerations. But these “daughter” chemicals will have different physical-chemical properties than the parent substance. In the case of PDEs, the biodegradation products are known and the further degradation of these will be discussed.
88
D.R. Peterson and C.A. Staples
3 Abiotic Degradation 3.1 Air
Direct photolysis probably does not play a major role in the atmospheric degradation of PDEs [1]. However, indirect photolysis by hydroxyl radical oxidation may contribute appreciably to overall fate. Staples et al. [1] calculated a range of air oxidation half-lives for PDEs based on the air oxidation program (AOP) of Atkinson [2] and a range of hydroxyl radical concentrations from 3 ¥ 105 to 3 ¥ 106 molecule cm–3. The atmospheric half-lives of a number of PDEs have been recalculated with an updated version of AOP (AOPWIN 1.89) as implemented in EPIWIN [3]. In this half-life calculation we also used a hydroxyl radical concentration of about 1 ¥ 106 molecule cm–3, a global average tropospheric concentration adjusted for diurnal and seasonal variation [4]. Although hydroxyl radical concentrations vary considerably depending upon the degree of atmospheric pollution and intensity of sunlight, Atkinson has recommended the use of this average value for global photooxidation estimates [4]. EPIWIN uses a 12-hour daylight period and a hydroxyl radical concentration of 1.5 ¥ 106. The rate constants (kOH) for hydroxyl radical attack and the calculated half-lives are shown in Table 1. Only DMP and DEP have appreciable (> 2 d) atmospheric half-lives. The half-lives shown in Table 1 for indirect photolysis in air are at the low end of those given by Staples et al. [1]. These values are still conservative, since emissions of PDEs are expected to be mainly in urbanized areas where hydroxyl radical concentrations are 3–10 times higher than the value used in the calculation. The less volatile PDEs, such as BBP and DOP, are reportedly about equally distributed in air between gas and particulate phases [5]. So the question arises Table 1. Atmospheric photodegradation rates and half-lives of selected PDEs
Phthalate
CAS No.
k (¥10–12 cm3 molecule–1 s–1)
T1/2 a (d)
T1/2 (h)
DMP DEP DBP DIBP BBP DHP DIHP DEHP DnOP DINP DIDP DIUP
131-11-3 84-66-2 84-74-2 84-69-5 85-68-7 84-75-3
0.574 3.466 9.277 9.260 11.049 14.929 18.719 21.955 20.581 23.408 26.217 31.847
14.41 2.39 0.89 0.89 0.75 0.55 0.44 0.38 0.40 0.35 0.32 0.26
346 57.3 21.4 21.4 18.0 13.3 10.6 9.0 9.6 8.5 7.6 6.2
a
117-81-7 117-84-0 68515-48-0 68515-49-1 3648-20-2
Based on a global, seasonal, and diurnal average hydroxyl radical concentration of 1 ¥ 106 molecule cm–3 [4].
Degradation of Phthalate Esters in the Environment
89
as to the photolysis half-lives of particulate-sorbed PDEs. Behnke et al. [6] have investigated the photolysis rate for DEHP adsorbed to various particulate aerosols. They report a first-order rate constant of 1.4 ¥ 10–11 cm3 molecule–1 s–1 for the reaction of DEHP with hydroxyl radicals when adsorbed as a monolayer on Fe2O3 or SiO2 aerosols. Photolysis on TiO2 was at a much higher rate due to catalytic activity of the carrier. This rate for inert particle sorbed photolysis corresponds to a half-life of 0.6 d, using the global average hydroxyl radical concentration of 9.7 ¥ 105 molecule cm–3. This half-life is not much longer than with the calculated gas-phase value of 0.38 d. Thus, sorption onto aerosol particles seems not to have a large effect on the overall rate of indirect photolysis of PDEs. 3.2 Water
Jin et al. [7] reported first-order aqueous photolysis rates for DBP and DEHP of 0.23 and 0.9 h–1, corresponding to half-lives of 3 and 0.75 h, respectively. The PDEs were present in a surface micro-layer in mg L–1 quantities. The reaction was dependent upon light intensity and oxygen. The rate was stimulated by the presence of TiO2 and H2O2 and the optimum pH was 6.0. The results under artificial light (72,000 lux) were essentially the same as with natural sunlight (83,000 lux). The photodegradation rates were higher in natural water than in simulated systems. The importance of these findings is difficult to assess but they certainly warrant further investigation. Due to their low water solubilities, surface films of PDEs are potentially quite important, as observed by Södergren [8] in laboratory ecosystems. Various estimates of the aquatic photolysis half-lives of PDEs are in the range of months to a year or more [1, 9]. Gledhill [10] reported a measured degradation of BBP exposed to sunlight of only 5% in 28 days. A possible reason for the apparent discrepancy between these values and the measurement by Jin et al. [7] is that the latter value, measured in natural water, is the consequence of indirect photolysis. Indirect photolysis in natural water may proceed through either hydroxyl radical or the photoactivation of organic matter or nitrate. Moreover, indirect aqueous photolysis half-lives of most classes of chemicals are expected to be of the order of days to weeks [11]. The hydrolysis rate of a number of PDEs has been reported by Wolfe et al. [12] and, with the exception of BBP, half-lives of many years are expected [1]. These finding are in general agreement with what is known regarding ester hydrolysis [13]. Thus, abiotic hydrolysis is not expected to play a significant role in determining the environmental fate of PDEs.
90
D.R. Peterson and C.A. Staples
4 Biodegradation 4.1 Surface Water 4.1.1 Reported Rates
There are a large number of reports that PDEs undergo rapid ultimate biodegradation (mineralization) in laboratory biodegradation tests and these have been reviewed by Staples et al. [1]. Since that review, Scholz et al. [14] published a report showing that DBP, DEHP, and DINP met all of the criteria for ready biodegradability, including the 10-day window, in tests strictly adhering to regulatory guidelines and criteria for ready biodegradability. Meeting the criteria for ready biodegradability implies that the chemical will rapidly degrade in the environment, but such tests are essentially pass/fail tests and do not provide quantitative data on the rate of environmental biodegradation [15, 16]. There are a number of reasons why these laboratory test systems are not predictive of environmental degradation rates: – The tests are designed to measure the extent of biodegradation and not the rate. – The tests measure the ultimate degradation (mineralization) of the test chemical and not its primary degradation rate. – The tests measure the population dynamics of batch culture of the degrading organisms, covering the adaptation, growth, and senescence phases, and not the steady-state kinetics attained in the environment. – The chemical being tested is the only organic chemical added and forms the sole carbon source for growth and metabolism of the microorganisms, while in the environment many carbon sources are available for growth. – The microorganisms in the test system are standardized as to type (aerobic sewage derived) and number and not intended to simulate typical surface water populations in type or number. – Adapted microbial cultures are not allowed in the ready test, but in the environment continuous emissions of a chemical would lead to adaptation. – The concentration of test chemical employed is generally much higher than would occur in the environment. The rate of biodegradation of a chemical is, in general, not a first-order reaction. However, under certain conditions it may approach first order (“pseudo-first-order”) and these conditions are often approximated in the environment. Mathematically, the degradation rate of substrate in a batch system may be described by the integrated Monod equation [15]: – ds/dt = mmax S(S0 + X0 – S)/(Ks + S)
(1)
Where mmax is the maximum specific growth rate of the degrading organisms, S is the substrate concentration at time t, S0 is the initial substrate concentration,
Degradation of Phthalate Esters in the Environment
91
X0 is the initial bacterial density divided by the growth yield, and Ks is the saturation constant, substrate concentration giving growth rate of mmax /2. Biodegradation rates depend upon both the concentration of the chemical and the concentration of the degrading organism and would be expected to be second order. However, when the number of organisms is rate limiting, their concentration is rapidly changing when they grow and multiply at the expense of substrate (test chemical). At low or intermediate chemical concentrations, when the test chemical concentration is limiting, as its concentration declines, uptake and enzyme binding of the chemical result in complex kinetics (Monod kinetics). At high test chemical concentrations, when its concentration is not rate limiting because the uptake and degradation mechanisms of the microorganisms are saturated, degradation rates will be zero order or logarithmic (depending only on microorganism number). Thus, if the number of degrading microorganisms is high in relation to test chemical concentration and the enzyme saturation constant is high in relation to the substrate concentration, then pseudo-first-order rates may be observed. That is, the degradation rate (k1) will be observed to be solely dependent on substrate concentration and a plot of the logarithm of substrate concentration versus time yields a straight line with a slope equal to k1 (and a half-life, t1/2 , equal to ln 2/k1). Under these conditions, the rate may be expressed as: – ds/dt = k1 S
(2)
Where k1 is equal to µmax X0/Ks . These conditions may be achieved in systems in which substrate concentrations are of the order of tens to hundreds of micrograms per liter and the population of degrading bacteria is about 108 cells L–1 or higher [15]. A secondary problem with regard to measured biodegradation rates of PDEs relates their hydrophobic nature. First-order reactions relate to the concentration of the chemical in solution, but PDEs may exist predominantly bound to particulates or dissolved organic carbon (DOC) and not truly in solution. Or they may be simply tested at concentrations above their water solubility. All of these situations are treated below under the topic of “bioavailability”. Considering the foregoing, it would seem that the best way to establish the rate of biodegradation in the environment is to rely upon measured rates of degradation of the chemical within the environment or within simulations of the environment. Aronson and Howard [17] have also made this recommendation. However, for chemicals in general there are few rate data measured under such circumstances and few standardized methods for collecting such data. There are much more available data from ready biodegradation tests or other screening tests. As a consequence of the lack of measured rate data, some regulatory authorities have established criteria for extrapolation from screening test results to environmental degradation rates. Within the EU risk assessment context, a chemical that passes the ready biodegradability test is assigned an aerobic aquatic biodegradation rate of 0.047 d–1 and one passing an inherent test is assigned a rate of 0.0047 d–1, corresponding to half-lives of 15 and 150 days. In the context of persistence for PBT chemicals, Boethling [18] has suggested rates of 0.14 d–1 and 0.0069 d–1 for chemicals that are readily and inherently biodegraded,
92
D.R. Peterson and C.A. Staples
respectively, corresponding to half-lives of 5 and 100 days. He also suggests intermediate rates of 0.069 d–1 and 0.023 d–1 for chemicals that give intermediate results in the ready test. As mentioned earlier, PDEs are readily biodegradable and thus, by EU or Boethling criteria would have assumed half-lives in the aquatic environment of 15 or 5 days. There are a number of published data for PDE pseudo-first-order biodegradation rates in aquatic microcosms. Most of these systems contain an added sediment layer or, as a result of containing field-collected water, contain suspended sediment and microorganisms. Since all of the PDEs have high adsorption coefficients for sediments and suspended solids [1], it is expected that a major fraction of PDEs in aquatic systems will be sediment associated and therefore, aquatic degradation studies for water and aerobic sediment are considered together. Saeger and Tucker [19] reported on the primary degradation of BBP, DEHP, DUP, and an isomeric mixture of heptyl, nonyl, and undecyl alcohol PDEs (“Sanitizer 711” or di-711) in Mississippi River water. The river water was settled and the incubations conducted unstirred in the dark at PDE concentrations of 1 mg L–1. The half-lives may be estimated from their graphic presentation of the chemical recovery data as about 1.5 d for BBP, 3.3 weeks for DUP, 4.3 weeks for DEHP, and 5.1 weeks for di-711. There was no apparent lag phase. These half lives correspond to rates of 0.46, 0.030, 0.023, and 0.019 d–1, respectively (see Table 2). A similar study was reported by Carson et al. [20] for BBP and di-711. They compared the results of a Mississippi River water biodegradation study design (river die-away) with two types of microcosms: (1) A lake/sediment system using a slow-stirred sediment-water system from a spring-fed freshwater lake (Lake 34, Bush wildlife area, St. Charles Co. MO), and (2) an unstirred river water/sediment (“ecocore” from Illinois River) design. PDE concentrations were at 10 and 100 µg L–1. They reported half-lives of 5 d and <3 d for BBP in the two systems compared with 2 d in the river die-away system. The di-711 had half-lives of 6–8 d in the river die-away design and 1–5 d in the lake water/ sediment microcosm. It is likely that the differences are due to differences in the number of degrading organisms in water/sediment from the two sample locations. Adams and Saeger [21] also reported a half-life of 1.4 d for BBP in a lake microcosm. Regional differences in the ability of water and sediment microorganisms to degrade DBP were investigated by Walker et al. [22]. They collected water and sediment samples from a number of rivers and estuaries in Florida, Mississippi, and Louisiana. Incubations were carried out in shake flasks (140 rpm, 25 °C) with DBP concentration of 0.5 mg L–1. Of the eight sites tested, seven of the water only systems showed overall (including lag phase) first-order rate constants of 0.07–0.98 d–1. The eighth system degraded too quickly to determine the rate. The mean of these seven rate constants was 0.472 d–1 and the median 0.512 d–1. For the eight sediment samples, rates ranged from 0.092 to 1.09 d–1 with a mean of 0.38 d–1 and a median of 0.29 d–1. There was no apparent relationship between rates in water and sediment from the same site or with the salinity at the sites. It is also evident that there is little difference in degradation rates between systems with and without suspended sediments.
93
Degradation of Phthalate Esters in the Environment
Table 2. Summary of first-order PDE biodegradation rates in aerobic aquatic environments
Phthalate DMP DEP
DPrP DBP
DIBP BBP
DPP DOP DEHP
Di-711
DDP DIUP a b c d
First-order rate (d–1) 0.5 0.23 1.8 0.16 0.98 b 1.3 0.8 0.51 c 0.29 c 0.22 0.14 d 0.12 d 0.8 0.46 b 0.50 b 0.35 b 0.14 b >0.23 b 2.2 1.3 0.7 0.023 b 0.2 1.73 0.019 0.13–0.09 b 0.69–0.14 b 0.5 0.030 b
Half-life (d)
Test conditions
Ref.
1.39 3.0 0.39 4.33 0.71 0.53 0.87 1.36 2.40 3.15 4.95 5.78 0.87 1.5 1.4 2 5 <3 0.32 0.53 1.0 30 3.5 0.4 36 6–8 1–5 1.4 23 R
RDAa, shaken, 25 °C ª 1 µg L–1 MITI inoculum RDA, shaken, 25 °C ª 1 µg L–1 MITI inoculum Microcosm periphiton RDA, shaken, 25 °C ª 1 µg L–1 RDA, shaken, 25 °C ª 1 µg L–1 RDE, estuarine & river Sediment microcosm MITI inoculum RDA, low sediment RDA, water only RDA, shaken, 25 °C ª 1 µg L–1 RDA, unstirred, 1 mg L–1 Microcosm, lake RDA Microcosm, un-impacted Microcosm, Illinois river RDA, shaken, 25 °C ª 1 µg L–1 RDA, shaken, 25 °C ª 1 µg L–1 RDA, shaken, 25 °C ª 1 µg L–1 RDA, unstirred, 1 mg L–1 RDA, shaken, 25 °C ª 1 µg L–1 Field data, estuarine sediment RDA, unstirred, 1 mg L–1 RDA Microcosm, lake RDA, shaken, 25 °C ª 1 µg L–1 DA, unstirred, 1 mg L–1
[23] [25] [23] [24] [27] [23] [23] [22] [22] [25] [29] [29] [23] [19] [21] [20] [20] [20] [23] [23] [23] [19] [23] [26] [19] [20] [20] [23] [19]
RDA “River Die-Away” study, flask containing un-inoculated river water. Value calculated from data presented in the referenced paper. Median result from 7 to 8 samples from different locations. Calculated from second-order rates.
Furtmann [23, 24] measured the rate of primary biodegradation of a series of 15 different PDEs in Rhine River water. Concentrations were in the low mg L–1 range for the lower PDEs and less than 1 µg L–1 for the less soluble PDEs. Incubations were performed in shake flasks at 25 °C. The first-order rate constants determined rates that ranged from a high of 2.2 d–1 for BBP to a low of 0.2 d–1 for DDP. There were no long lag phases as all of the PDEs tested degraded by 50% in <3 days. Ye and Tian [25] have reported the rates of primary biodegradation of DMP, DEP, DBP, and DAP. However, they used an inoculum derived from soil and sewage supernatant similar to that used in the MITI test. Half-lives of 0.23, 0.16, 0.22, and 0.22 d were reported for these four PDEs, corresponding to rates of 3.0, 4.3, 3.2, and 3.2 d–1.
94
D.R. Peterson and C.A. Staples
A different approach to the estimation of degradation rates has been made by Fauser et al. [26]. By using the concentration versus depth profile of DEHP in sediments in Roskilde Fjord, Denmark, they derived a model which incorporated rates of sedimentation and of aerobic and anaerobic biodegradation. Through fitting the model to observed concentrations, they arrived at a first-order aerobic degradation rate for DEHP in sediment of 2 ¥ 10–6 s–1 or 1.73 d–1, equivalent to a half-life of 0.4 d. The rates of chemical degradation in periphyton (aufwuchs) from microcosms were reported by Lewis et al. [27]. They correlated the rate of degradation with the presence of total bacteria and degrading bacteria. In the flow-through microcosms, DEP showed second-order rate constants of 0.03–2.2 nL cell–1 h–1. The mean of all of the experiments was 0.41 nL cell–1 h–1 (= 0.41 ¥ 10–9 L cell–1 h–1). These experiments contained 107 –109 total bacteria based on total plate count; thus, by using an average of 108 bacterial cells L–1, the average pseudo-first-order rate is 0.041 h–1 or 0.98 d–1 (t1/2 = 0.7 d). The typical number of bacteria in surface water is 108 –1010 cells L–1 [28], so that a microorganism number of 108 is also conveniently close to the average for surface water. Second-order rate constants for DBP primary biodegradation in Mississippi River water, with and without added sediments, were determined by Steen et al. [29]. Initial DBP concentrations were 1–2 mg L–1. Plate counts of total organisms were used to convert the observed pseudo-first-order rate constants to second-order rate constants. These ranged from 3.1 ¥ 10–11 L cell–1 h–1 in water alone to 6.1 ¥ 10–13 L cell–1 h–1 in the high-sediment content flasks. Unfortunately, the cell numbers are not given so the corresponding pseudo-first-order rates cannot be directly calculated. However, a logarithmic graph of the DBP concentrations versus time is presented and the slopes may be estimated to determine the pseudo-first-order rates. These first-order rates are calculated to be about 0.12, 0.14, and 0.07 d–1 for water only, low-sediment, and high-sediment systems, respectively. These rates correspond to half-lives of 5.8, 4.9, and 10.1 d. The larger differences between the reported second-order rates for water and for the high-sediment system than between these first-order rates we calculate for these same systems is evidently due to higher cell numbers in the sediment amended systems. A summary of these reported PDE biodegradation rates in surface water is presented in Table 2. The pseudo-first-order rates of primary biodegradation under environmentally realistic conditions, such as model ecosystems or incubation in river water, are in the range of 0.2–2.0 d–1 for most PDEs studied. The study of Saeger and Tucker [19] gave somewhat lower rates for the higher phthalates (DEHP, D711, and DIUP) but these studies were conducted at concentrations well above the solubilities of these phthalates and the systems were settled and not agitated at all. Studies conducted at more realistic environmental concentrations of PDEs, such as those of Furtman [23], show rates in the range of 0.2–2.0 d–1 for the higher molecular weight PDEs as well.
Degradation of Phthalate Esters in the Environment
95
4.1.2 Bioavailability
Biodegradation cannot occur if for some reason the chemical is unavailable to the microorganisms or if the concentration truly in solution is less than the total concentration, the biodegradation rate may be reduced. Factors that reduce a chemical’s effective concentration with regard to biodegradation and/or toxicity are said to reduce the “bioavailability” of the chemical. Factors affecting bioavailability are solubility, dissolution rate, and sorption to dissolved or particulate organic matter. Due to the relative hydrophobicity of the higher PDEs (DBP and larger), their biodegradation rates may be reduced by low bioavailability. One might expect that the phthalate esters would all undergo primary biodegradation at about the same rate. Kurane et al. [30] investigated the degradation rate of a series of PDEs by one (Pseudomonas acidovorans) of a number of isolated PDE degrading organisms. They found that for a series of nine n-alkyl PDEs, ranging in length from methyl to tridecyl, the rates were all between about 0.1 and 0.3 d–1. Most of the PDEs including DMP and DOP through DTDP had rates near 0.1 d–1, but higher rates were observed for DBP and DHpP. In a later study [31], the PDE hydrolyzing enzyme from another organism (Nocardia erythropolis) was isolated and its substrate specificity and other properties were determined. The enzyme had many of the properties of a typical lipase, including good degradation of triglycerides (olive oil and tributryn). This enzyme also degrades all of the phthalates at reasonably good rates, with the highest rates for DBP and DEHP. Based on enzyme specificity alone, one would expect to see only small differences in biodegradation rates among PDE, certainly not differences of orders of magnitude. The fact that in some test systems the higher molecular weight PDEs, such as DEHP, show much lower biodegradation rates than the lowmolecular weight PDEs, such as DMP, may be ascribed to a reduced bioavailability of the higher PDEs. This lower bioavailability may be due to lack of solubility or to sorption or sequestration by organic matter. Regarding solubility, the ready biodegradability screening tests, with the exception of the closed-bottle test, use chemical concentrations in the 10– 100 mg L–1 range. Within this concentration range, for the higher PDEs, a significant fraction is insoluble. Nyholm [32] reported on the results of biodegradation studies with a number of techniques to increase solubilization of some poorly water-soluble substances, including DEHP, in a manometric screening test. He found that by using either emulsifiers or increasing the surface area through the use of silica gel of glass fiber filters as carriers generally increased the biodegradability of these substances. For DEHP, it is evident from the degradation curves that the use of the solid carriers may have increased the early rate of degradation. However, the 28-day extent of degradation seems to be decreased by these materials, perhaps through adsorption. Scholz et al. [14] also noted that testing biodegradability at concentrations well above water solubility might have lead to some of the variable results reported for PDEs. On the other hand, Gibbons and Alexander [33] reported that some bacteria (Mycobacterium sp. and Nocardia sp.) excrete products that increase the solubility of DHP, DEHP, DIOP, and DIDP.
96
D.R. Peterson and C.A. Staples
Achinger et al. [34] reported on the use of respirometry to evaluate the effects of solubility on bacterial growth kinetics. They used enrichment cultures in a batch reactor, relating oxygen uptake to maximum growth rate and growth yield (Monod kinetics). They reported maximal growth rates of about 0.1 h–1 and growth yields of about 0.5 mg mg–1 for DMP, DEP, DBP, BBP, DnOP, and DEHP. The kinetic model fit the experimental data rather well for DMP, DEP, DBP, and BBP, even when BBP was tested above and below the its water solubility. On the other hand, DnOP and DEHP data fit the model more poorly as degradation proceeded. The authors concluded that this was most likely due to the rate of utilization exceeding the rate of solubilization.A modified model to account for solubility effects gave a better fit to the data. Rate constants for substrate utilization calculated from the data presented by Achinger et al. [34] in different experiments ranged from about 14 to 3.7 d–1 (t1/2 = 0.05–0.2 d) with little difference among the PDEs tested. Wang et al. [35, 36] reported on the maximal growth rate of DBP in a continuous culture system. They reported a value of 0.38 h–1 with substrate concentrations of 0.5–1 g L–1, well above the solubility of DBP of about 10–2 g L–1, which may be the reason the rate is so much lower than those of Achinger et al. [34]. Another factor that may reduce bioavailability, and hence apparent biodegradation rates, is adsorption to organic matter. Wang and Grady [37] extended the study of Achinger et al. [34] to include both dissolution kinetics and sorption to biomass in the kinetic analysis of DBP utilization by microorganisms. A bacterium (Pseudomonas fluorescens) which did not degrade DBP was added to batch cultures at a concentration of 2 g L–1 and degrading organisms from an enrichment culture were added at about 10–2 g L–1. Degradation parameters were evaluated from 14C mass balance in evolved CO2 , liquid phase, and biomass, by using various concentrations of carbonyl-labeled 14C-DBP. They found that when DBP was added at concentrations greater than its solubility, the presence of the living carrier organisms increased the rate of biodegradation compared to systems without carrier. They concluded that desorption from carrier was faster than the rate of dissolution of insoluble DBP. It is also apparent that sorptiondesorption is reversible and relatively fast compared to either the rate of dissolution or the rate of biodegradation. In the study by Steen at al. [29] mentioned earlier, the addition of sediment to test systems in amounts that reduced the dissolved concentrations of DBP by 15% and 90% had only a small effect on the pseudo-first-order biodegradation rates, increasing the rate slightly at 15% and reducing it only by half at 90%. However, this minimal effect of added sediment may possibly be due to a concomitant increase in the number of microorganisms present, since there was a much greater reduction in the second-order rate constant as a result of sediment addition. They also reported that the second-order rates were all approximately the same if they were corrected for the fraction of BBP actually in solution. In this study the concentration of non-degraders was much higher than the concentration of degraders. Of course, when the organism to which the PDE is sorbed is capable of PDE degradation as well, such sorption would be expected to increase the rate of degradation. Yan et al. [38] reported a study on the bioconcentration of DMP, DEP, and DBP by the algae Chlorella pyrenoidosa. They found that the algae not
Degradation of Phthalate Esters in the Environment
97
only sorbed but also degraded the PDEs at an appreciable rate and the increase in algal biomass resulted in logistic rates. It seems likely that biodegradation does occur to some extent in the sorbed state. The higher phthalates in aqueous systems are predominantly sorbed to suspended solids [1, 23]. If they were only biodegraded in the non-sorbed state, the effective rate of biodegradation would be expected to be reduced proportionately to the extent of adsorption as pointed out by Steen et al. [29] (assuming rapid, reversible adsorption). Since DEHP and DnOP have adsorption coefficients to organic carbon (Koc) about three orders of magnitude higher than those of DMP and DEP [1], one would expect considerably longer measured half-lives for the higher PDEs, which does not seem to be the case. A possible explanation is that PDEs sorbed to biomass still undergo biodegradation. This topic will be addressed again when discussing biodegradation rates in soil and sewage. 4.1.3 Biodegradation of Very Low Concentrations
Furtmann [23] has drawn attention to the fact that many PDEs are biodegraded in laboratory river die-away studies down to a level of several ng L–1 and then degradation stops (in particular DEP, DHxP, and DEHP). The levels reached often appear to be similar to background levels originally found in the water sample. The very low rate of biodegradation at very low concentrations has been observed by many for a range of chemicals including glucose. This phenomenon has been discussed by Rubin et al. [39] who also observed it for DEHP. As an explanation of this, they suggested the existence of two types of organisms involved in chemical degradation: eutrophs and oligotrophs. Eutrophs are capable of growing on relatively high concentrations of the chemical but have a low affinity for it and cannot sustain growth and metabolism at very low concentrations. Oligotrophs can degrade very low concentrations of a variety of chemicals but are less specific and may degrade other chemicals present in the water rather than the chemical of interest. Another explanation suggested by Pagga [40] is that bacteria are unable to produce the necessary degradative enzymes either because a minimum substrate concentration is necessary to induce them or because the substrate is at too low a level to be transported into the cell. Other possibilities for PDEs are that the chemically detected concentrations are not bioavailable to the bacteria due to binding to suspended or dissolved organic carbon or that measured background levels in the water are actually PDE backgrounds introduced during sample analysis do to the ubiquity of PDEs in the laboratory. The existence of such trace levels is often used as evidence that PDEs are persistent. However, Furtmann [24] concludes that the trace background levels of PDEs in water, which do not appear to be further biodegraded, are normally below 1 µg L–1 and that these levels are not relevant for toxicological concerns.
98
D.R. Peterson and C.A. Staples
4.2 Soil 4.2.1 Reported Rates
In soil mobility studies on DMP, DEP, and DBP, Russell et al. [41] noted that these chemicals biodegraded quickly in soil. They conducted biodegradability studies on these PDEs in soil from two locations in Broome County, New York that had been treated with landfill leachate. Soil suspensions of about 1–2 g soil in 50–60 mL of water were shaken at 25 °C with 1-3 mg L–1 PDE.At various time intervals, aliquots were extracted and analyzed by gas chromatography (GC). Similar experiments were done with uncontaminated soil from a nature preserve. The landfill leachate-treated soil gave biodegradation rates of about 0.06 h–1 for DMP, DEP, and 0.04 h–1 for DBP (half-lives of 0.5 and 0.7 d, respectively). The soil from the nature preserve gave biodegradation rates of 0.015, 0.016, and 0.067 h–1 for DMP, DEP, and DBP, respectively (t1/2 =1.9, 1.8, and 0.4 d). In a similar experiment, Kirchmann et al. [42] found DEHP to follow zero-order kinetics in soil with rates of 0.20 and 1.76 mg kg–1 d–1 at initial concentrations of 5 and 250 mg kg–1. The soil suspension method used by Russell [41] could be criticized for not being an exact simulation, since it is essentially an aqueous system with soil bacteria. Rüdel et al. [43] investigated the soil biodegradation rate of DEHP in simulation systems. They compared the first-order rates using 1 mg kg–1 DEHP in two different soils: a silty sand and a silty loam. They also compared rates between laboratory systems under standardized conditions of soil moisture (40% of holding capacity) and temperature (20 °C) with outdoor lysimeter experiments and laboratory flasks with varied moisture and temperature regimes to mimic outdoor conditions. The reported DEHP half-lives under laboratory conditions were 20 d for the loam and 68 d for the sand. The higher rate in loam (0.035 d–1) than sand (0.010 d–1) was thought to be due to the approximately threefold higher biomass in the loam. Bioavailability differences must not have played as great a role, since the loam had twice the organic carbon as the sand (2% vs. 1%), and would have been expected to reduce bioavailable concentrations proportionally and give a lower rate in the loam. Moisture and temperature had a significant effect on the rates, since the half-lives in the simulated outdoor systems were 31 d (first-order rate = 0.022) for the loam and 170 d (rate = 0.004) for the sand and the half-lives in the lysimeters were 21 d (rate = 0.033) and 54 d (rate = 0.013) in the two soils. The presence of plant growth (barley) had an uneven effect on DEHP degradation. These workers also reported on some noteworthy aspects of DEHA metabolism in soil. By using 14C-labeled DEHP, they found no metabolites besides CO2 , that is, the undegraded residue was probably parent DEHA. However, they also found that after long incubation times, a significant fraction of the radioactivity was non-extractable. An average extractability in the laboratory systems was 34 and 39% in 64 and 100 d, respectively, for the silty loam and an average of 11% and 13% in 64 and 100 d for the silty sand. Other chemicals (biocides) they tested showed similar behavior. The implications of this finding will be addressed later.
Degradation of Phthalate Esters in the Environment
99
Shanker et al. [44] published on the biodegradation of DMP, DBP, and DEHP in soil. They incubated samples of garden soil containing a reasonably high concentration of PDEs (500 mg kg–1) in Erlenmeyer flasks at 30 °C and 60% moisture holding capacity for varying time periods and then analyzed by HPLC for parent PDEs and metabolites. Their quantitative data may be analyzed to evaluate pseudo-first-order rates. The sample intervals were at time zero and then at 5 day intervals for up to 30 days. The rates we calculate from their data for DMP, DBP, and DEHP are 0.40, 0.39, and 0.12 d–1, respectively (t1/2 =1.7, 1.8, and 5.6 d). There was a lag phase of about 5 days for DEHP but not for the other two PDEs. It the lag phase is omitted, the rate for DEHP is 4.3 d–1. They found no appreciable amounts of metabolites, except for PA (phthalic acid), which was found in low concentrations during intermediate stages of degradation and never approached the concentrations of the parent PDE in the soil. The biodegradation rates of DBP and DEHP in three different soils were reported by Chen et al. [45]. They also used 500 mg kg–1 of PDE, incubating the soils at 30% soil moisture and 28 °C. They extracted and analyzed the PDEs by GC-FID at 5-day intervals up to 30 days. They found that the data fit first-order kinetics with no appreciable lag phase. The biodegradation rates were 0.103, 0.062, and 0.044 d–1 for DBP and 0.040, 0.019, and 0.015 d–1 for DEHP. The highest rates for both PDEs were in the soil with the intermediate organic carbon (2%), while the highest organic carbon (3.3%) soil gave the intermediate rates. There were no clear relationships between biodegradation rates and other soil properties. As Rüdel et al. [43] had earlier reported that they found biodegradation rates increased with increased soil temperature. They also reported that in sterile soil, extractable PDEs decreased with time, particularly for DEHP, which decreased by nearly 20% in 30 days in the low-organic carbon soil. Wang et al. [46] reported on the biodegradation of DBP in soil microcosms at 60% moisture holding capacity and 25 °C. DBP was initially present at 100 mg kg–1 soil. They report concentration versus time data that can be used to evaluate the first-order rate. The rate determined from their data is 0.036 d–1. Inoculation of the soil with a DBPdegrading bacterium greatly enhanced this rate. The microbial population of the un-inoculated soil was about 4 ¥105 colony forming units g–1 and only decreased slightly by the end of the 30-day incubation period. Cartwright et al. [47] recently reported on the effects of DEP and DEHP on soil microorganisms. As part of the study, they investigated the degradation of these two PDEs by indigenous soil organisms. The soil was sandy clay loam soil containing 3.78% organic carbon and the PDEs were added at concentrations of 0.1, 1, and 10 g kg–1. Sub-samples of the amended soils were incubated in 28 mL sealed bottles at 20 °C and 50% soil moisture. The bottles containing 1 and 10 g kg–1 PDE were opened for an hour every 5 days to allow for aeration. Separate containers were extracted and analyzed by HPLC for remaining PDE at various time intervals. They report that DEP degraded rapidly with “half-times” of 0.75, 5, and 17 d at 0.1, 1, and 10 g kg–1. DEHP only degraded by about 10% in 70 d. They proposed reduced bioavailability of DEHP, due to greater soil adsorption, as the reason for the difference in biodegradation rate between the two PDEs. Toxicity to soil organisms at the higher levels was not a reason for the difference between the biodegradability of the two PDEs, since the DEHP-containing soil
100
D.R. Peterson and C.A. Staples
showed no significant concentration related difference in total bacteria (colony forming units, about 3 ¥ 107 g–1 soil) at any of the sample intervals. The half-times for degradation are evidently not first-order half-lives but the time to reach 50% of nominal additions. The authors give a graph of the PDE versus time values for the 0.1 g kg–1 experiment and re-plotting values from this graph for DEP on a natural log scale gives a reasonably straight line (r2 = 0.98) with a rate constant of 0.28 d–1 (t1/2 = 2.5 d). The DEHP curve is quite interesting; after a decline to about 80% of its initial concentration in the first two weeks, it showed no further changes. This is further discussed in the next section. 4.2.2 Conditions Affecting Soil Biodegradation
The extent of biodegradation of DEHP over time observed by Cartwright et al. [47], as discussed above, shows an unexpected pattern. If adsorption to soil were responsible for decreased biodegradability, one would expect to see a continuing biodegradation at a rate consistent with the concentration of the equilibrium desorbed concentration over the entire 70 days, unless the adsorption was irreversible. In another recent paper (Jensen et al. [48]) on the toxicity of DBP and DEHP to a soil invertebrate (Folsomia fimetaria), the authors also measured DBP concentrations over time for soil with initial concentrations of 50 and 500 mg kg–1. The data over 25 days very closely fit a model the authors proposed wherein adsorption and degradation occur simultaneously but wherein desorption is negligible. The model predicts that after 30 days about 50% of the DBP will be sequestered and unavailable for biodegradation. The lack of linearity of logarithmic plots of PDE biodegradation in soil has been examined in a number of studies on PDE mineralization [49–52]. These studies are not strictly applicable to determining rates of primary biodegradation. When primary degradation and mineralization are measured in the same study, mineralization, as measured by carbon dioxide evolution, generally proceeds at a slower rate and to a lesser extent than primary degradation. This difference may be ascribed to the formation of metabolites that are further degraded at a slower rate than the parent substance. Another reason is the incorporation of carbon into growing biomass. For instance, Roslev et al. [51] found that a significant portion of 14C-DEHP was incorporated into phospholipid fatty acids of soil organisms. They also observed that a significant fraction of CO2 in long-term degradation assays originated from turnover of biomass. Nevertheless, such studies of mineralization may provide information on the conditions affecting the overall biodegradation process. They also constitute a lower limit for biodegradation rate, since the first biodegradation step (primary) will always be as fast or faster than complete conversion to CO2 . Fairbanks et al. [49] observed that mineralization of DEHP in three New Mexico soils gave an initial rapid rate, with little or no lag, followed by a slowing of the rate with time. They stated that they did not include first-order rate constants in their report due to the complexity of the curves, but only presented log DEHP concentration (calculated based on the percent of 14CO2 produced) versus time curves. The initial points of these curves are relatively linear and correspond to
Degradation of Phthalate Esters in the Environment
101
rates of about 0.035, 0.069, 0.058 d–1 at an initial soil concentration of 2 mg kg–1. At 20 mg kg–1, initial rates are 2- to 4-fold slower. These workers also reported data for rates in sewage sludge amended soil, which gave a similar pattern. Dörfler et al. [50] did analyze the kinetics and derived a complicated kinetic expression describing DEHP mineralization. Roslev et al. [51] studied DEHP mineralization in sludge-amended soils, including an analysis of the complex kinetics. They report that the mineralization of DEHP could be divided into two distinct kinetic phases: an initial phase over 28 days that followed first-order kinetics and a late phase that was slower and a function of an exponent of time (t–0.387). Half-time for the initial phase was 58 d (first-order rate = 0.012 d–1) and for the late phase was 147 d. This same group later published a more detailed report on the rates of the two phases [52]. The rates were similar for untreated and sludge-amended soil.At 20 °C the initial first-order rate of DEHP biodegradation in soil was 0.0134 d–1 (t1/2 = 52 d) and in sludge-amended soil it was 0.0127 d–1 (t1/2 = 55 d). The second phase of mineralization observed by these workers is difficult to interpret, since they are not measuring primary degradation of DEHP but production of CO2 . A number of other slow processes may limit ultimate degradation to CO2 of a fraction of the 14C originally in the DEHP, particularly since, as noted in the Roslev et al. [51] paper, a significant portion of the 14C was incorporated into lipids and eventually was released through turnover of the biomass. Another explanation of the second slow phase of release is the possible “sequestration” of the parent DEHP within soil solids. This phenomenon has been reported on and investigated by numerous researchers [53–56]. A wide range of hydrophobic chemicals have been shown to undergo such sequestration, wherein after aging of the chemicals in soil (or sediment), they penetrate into solid particles and are no longer surface adsorbed and available to some extraction solvents. The chemicals are also not available for biodegradation or toxicity. Unlike sorption-desorption equilibrium, which is relatively fast in relation to biodegradation rate, diffusion of the entrapped chemical out of the solid is a slow process with a half-life of the order of ten days or more [55]. The rates of first-order reactions are concentration independent. The above studies on mineralization of DEHP showed a considerable dependence of rates on initial test material concentration. Fairbanks et al. [49] used initial concentrations of 2 and 20 mg kg–1 and half-lives were 2- to 4-fold longer at the higher concentration in the three soils. Dörfler et al. [50] studied 0.5 and 10 mg kg–1 concentrations and report that higher initial concentration resulted in lower percent degradation in all three soils tested. Madsen et al. [52] reported very similar firstorder rate constants for DEHP at initial concentrations of 1.6, 3.2, and 9.9 mg kg–1 in sludge-amended soil. They reported rates of 0.0087, 0.0081, and 0.0078 d–1, respectively. However, an initial concentration of 35.1 mg kg–1 gave an order of magnitude higher rate of 0.090 d–1. The rate of biodegradation of a chemical in soil will differ considerably depending on how the concentrations are measured and how rapid desorption relative to the rate of biodegradation. Assuming for the moment that a chemical does not biodegrade when tightly adsorbed to soil solids but only when it is in pore water (bioavailable), then one would expect that biodegradation would be
102
D.R. Peterson and C.A. Staples
much slower for chemicals with a high Koc . For instance, comparing DEHP with DMP, the Koc value of DEHP is more than three orders of magnitude higher, so its bioavailable concentration in soil is expected to be three orders of magnitude lower. As a consequence, the apparent biodegradation rate, based on total soil concentration change would be three orders of magnitude lower if the two PDEs have the same aqueous biodegradation rate. If one measures only the concentration change in the aqueous phase, and if the desorption of DEHP from the soil is rapid relative to the biodegradation rate, then the apparent degradation rate is still slow because as DEHP is degraded from the water, it is quickly replenished from the sorbed phase. However, if the desorption is irreversible, then the apparent biodegradation rate is much faster if one is measuring concentration change in the water phase, as shown in Fig. 1a. Moreover, if one measures the total decrease in soil concentration when desorption is irreversible, a logarithmic plot of the data versus time gives a non-linear curve with a decreasing rate, Fig. 1b. If one assumes a slow rate of desorption relative to biodegradation, then that rate becomes limiting. In fact this probably explains the kinetics observed by Dörfler et al [50], Roslev et al. [51], and Madsen et al. [52]. However, it is apparent that there must also be some biodegradation of the PDEs in the sorbed phase. In soil, with water making up about 12% of its weight and organic matter 2%, one would expect that for the higher PDEs with Koc values greater than 10,000, the fraction dissolved in soil pore water would be less than 0.01%. One would expect that biodegradation rates of these PDEs would be many orders of magnitude lower in soil than in water if only this dissolved fraction were biodegradable, since the numbers of microorganisms in soil is not 1000-fold higher than in water. But, in fact the biodegradation rates of the PDEs in soil are very similar to those in water. There are also a number of reports of certain organisms degrading PDE in the solid phase, as will be described later. Temperature plays an important role in biodegradation rate. Rüdel et al. [43] ascribed the major reason for the differences between laboratory and simulated outdoor soil degradation systems to temperature. When they corrected for temperature differences by using a factor of 50% decrease in rate with a 6.5 °C decrease in temperature, the two systems agreed well for DEHP and the pesticides and biocides tested. Chen et al. [45] studied the biodegradation of DEHP at three different temperatures: 10, 28, and 35 °C. Unfortunately they did not report the rates at these temperatures; however, the extent of degradation at the three temperatures was 21.5, 28.5, and 33.2%, respectively. Madsen et al [52] also investigated the effect of temperature on the rate of mineralization of DEHP in sludgeamended and un-amended soil at 5, 10, and 20 °C. Each doubling of the temperature resulted in a doubling of the rate. Soil moisture has been reported in many of the studies as being an important condition for soil biodegradation, with higher soil moisture resulting in a greater rate of biodegradation or mineralization. This is expected since bacteria live in the water phase and require water for diffusion of nutrients into and out of the cell. It is also likely that the phthalates are degraded most rapidly in the aqueous phase and that increased volumes of interstitial water in the soil will result in a greater dissolved amount PDE. However, if the soil is flooded and its maximum moisture holding capacity exceeded, there is a possibility that anaerobic condi-
Degradation of Phthalate Esters in the Environment
103
a
b Fig. 1. a Modeled slurry biodegradation, concentration change with time in the water phase, comparing reversible (dashed line) and irreversible (solid line) adsorption. Conditions: Koc = 104, solids 600 g L–1, 5% organic carbon, k1 = 1 d–1. b Modeled slurry biodegradation, concentration change with time in entire system, comparing reversible (dashed line) and irreversible (solid line) adsorption. Conditions the same as Fig. 1a
104
D.R. Peterson and C.A. Staples
tions will occur. A lack of oxygen will be expected to result in a slower biodegradation rate in soil; for instance, Madsen et al. [52] report a rate of anaerobic degradation of DEHP in sludge-amended soil of 0.0023 d–1 as compared with 0.0127 d–1 in the same soil under aerobic conditions. Anaerobic biodegradation rate in sediment and soil is treated below as a separate topic. The literature values for the pseudo-first-order half-lives for PDEs in soil discussed in this section are summarized in Table 3.Although the data are sparse for DMP and DEP, it is expected that these two PDEs should not differ greatly in primary degradation rate due to their similar chemical properties. Their biodegradation rates range from 0.36 to 0.40 d–1. There are more data for DBP, ranging from 0.036 to 1.6 d–1. The value of 1.6 was determined on an aqueous suspension of soil rather than soil alone. A value of about 0.1 d–1 would seem to be representative for the remaining DBP rate data. Clearly, the higher extent of soil sorption of DEHP results in somewhat lower bioavailability and lower biodegradation rate. However, the primary rate data would indicate that degradation might occur to some extent in the sorbed phase. There are considerable data for a primary biodegradation rate of DEHP in the range of 0.01–0.1 d–1.A biodegradation rate of 0.03 d–1 would seem to be representative of these data. It must be remembered that the use of first-order rates is a simplification of the process of degradation. As we have seen, DEHP can become sequestered in soil and unavailable for biodegradation. In which case, biodegradation of this sequestered fraction would seem to be limited by the rate of reversal of this sequestration process. Table 3. Summary of pseudo-first-order PDE biodegradation rates in aerobic soil
Phthalate
First-order rate (d–1)
Half-life (d)
Test conditions
Ref.
DMP
0.36 a 0.40 0.38 a 1.61 a 0.39 0.103 0.044 0.062 0.036 0.035 0.010 0.033 0.013 0.12 0.040 0.019 0.015 0.012
1.93 1.7 1.83 0.43 1.8 6.7 11.2 15.8 19.3 2.0 69.3 21 53.3 5.6 17.3 36.5 46.2 58
Aqueous suspension, agitated Flask, 30 °C, 60% WHC b, garden soil Aqueous suspension, agitated Aqueous suspension, agitated Flask, 30 °C, 60% WHC, garden soil Flask, 28 °C, 30% WHC, 2% OC c Flask, 28 °C, 30% WHC, 3.3% OC Flask, 28 °C, 30% WHC, 1.6% OC Flask, 25 °C, 60% WHC Flask, 20 °C 49% WHC, loam Flask, 20 °C 49% WHC, sand Outdoor lysimeter, loam Outdoor lysimeter, sand Flask, 30 °C, 60% WHC, garden soil Flask, 28 °C, 30% WHC, 2% OC Flask, 28 °C, 30% WHC, 3.3% OC Flask, 28 °C, 30% WHC, 1.6% OC Sludge amended loam, 75% WHC
[41] [44] [41] [41] [44] [45] [45] [45] [46] [43] [43] [43] [43] [44] [45] [45] [45] [51]
DEP DBP
DEHP
a b c
Value calculated from data presented in the referenced paper. WHC water holding capacity. OC organic carbon content of the soil.
Degradation of Phthalate Esters in the Environment
105
4.3 Wastewater
A number of studies have shown that PDEs degrade rapidly in wastewater (for instance Furtmann [23]). The standard screening tests for biodegradability also generally use aerobic sludge from a wastewater treatment plant (WWTP) as an inoculum source and the ready biodegradability of PDEs would indicate that competent organisms are present in this sludge. The issue this section seeks to address is the rate and/or extent of biodegradation in a wastewater treatment system, which of course is dependent on the specific system employed. Wastewater treatment systems have a much higher concentration of microorganisms than standard screening test systems. One method that has been used to evaluate potential removal of chemicals from wastewater in a WWTP is a semi-continuous activated sludge (SCAS) test. Saeger and Tucker [19] evaluated removal of a number of PDEs (MBP, BBP, DEHP, di-711, and DUP) in a SCAS system. The system is a cylinder containing aerated activated sludge from a domestic treatment plant. Once daily, aeration is stopped and the sludge is allowed to settle. Supernatant liquid is drawn off and replaced with water containing nutrients (synthetic sewage) and test chemical.After an acclimation period, samples are withdrawn at various times subsequent to chemical addition, and the extent of degradation (or removal) evaluated.At PDE addition rates up to 200 mg d–1, Saeger and Tucker [19] found >99% primary biodegradation for MBP and BBP, as determined by sludge removal and analysis. The degradation of DEHP, di-711, and DUP was 78, 52, and 45%, respectively, at 5 mg d–1 additions. They also reported the time course of biodegradation of BBP, which declined to about 50% in slightly over an hour and by about 98% in 6 h. Soluble by-products reached a maximum by 5 h and were degraded by 14 h. Results of SCAS testing on DBP were also reported by Wang et al. [57]. They found essentially the same degradation rate at concentrations of 50–200 mg L–1 and concluded they were at zero-order kinetics and that the rate constant could be equated with the saturation biodegradation rate. The rate constant they determined was 0.015 h–1 (or 0.36 d–1). SCAS systems are not strictly simulations of WWTP because they are fed intermittently rather than continuously. Petrasek et al. [58] studied the removal efficiencies of a number of PDEs (DMP, DEP, DBP, BBP, DnOP, and DEHP) in a laboratory sewage simulation system at nominal concentrations of 50 µg L–1. The system was fed with raw sewage from a nearby domestic WWTP. Sludge retention time was seven days. The average concentrations of the PDEs in the activated sludge effluent was around 1 µg L–1 for all but DEHP, with a number of non-detects averaged in as the detection limit. DEHP was detected in all samples of activated sludge effluent at an average concentration of 11.3 µg L–1. For the more volatile DMP and DEP, a significant portion of the removal may be attributable to air stripping. Tokuz [59] has also reported on the removal of DMP and DEP in a laboratory-scale sewage treatment system. The system was fed with synthetic sewage. The mixed liquid suspended solids (MLSS) were at about 1100 mg L–1. Addition of DEP and DMP was slowly increased to 410 mg L–1 and 540 mg L–1, respectively, by day 19. The total influent COD (chemical oxygen demand) rose from about 600 to 2000 mg L–1, while the effluent COD remained low
106
D.R. Peterson and C.A. Staples
< 100 mg L–1. MLSS increased to 1600 mg L–1, indicating that a considerable portion of the PDEs was converted to biomass. Although indicating that DMP and DEP were appreciably removed by treatment, quantitative removals or rates are not possible for these data, since specific analyses were not conducted. Another approach to evaluating the removal rate of chemicals in a WWTP is to evaluate influent and effluent concentrations in an actual WWTP. Paxéus et al. [60] conducted such a study over the course of three years (1989–1991) at the Göteborg (Sweden) Regional Sewage Works. They found average influent wastewater concentrations in the low µg L–1 for DMP, DEP, BBP, and DnOP. Annual average influent concentrations were 36–86 µg L–1 for DBP and 30–40 µg L–1 for DEHP. Effluent concentrations were below detection levels for all PDEs except these last two. Effluent concentrations of these were 0.1–2.0 µg L–1 for DBP and 0.3–2.0 µg L–1 for DEHP. Removals from wastewater appear to be in the range of 90% or greater but the contribution of biodegradation to the total removal cannot be evaluated from these data. Furtmann [23] also analyzed influent and effluent concentrations of a number of PDEs from two WWTPs, one treating domestic sewage and the other industrial sewage. DEP, DIBP, DBP, BBP, DEHP, and DOP were all in the 1–10 µg L–1 concentration range in the influent to both plants. DEHP was at 25 µg L–1 in the domestic sewage and 71 µg L–1 in the industrial sewage. DOP concentration was low (1.3 µg L–1) in the domestic sewage and high (36 µg L–1; average of two sampling dates) in the industrial sewage. All of these PDEs were removed by 98% in the effluents of both plants, with the exception of DEP and DBP. These were removed by 88% (DEP) and 83% (DBP) in the domestic plant and 93% and 95% in the industrial WWTP. Of course, it is likely that the higher PDEs partition primarily with the waste sludge [23]. There are a number of papers on the occurrence of PDEs in sludges from WWTPs [24, 61, 62] but these data shed little light on the issue of biodegradation rate without having concurrent influent and effluent concentrations. At best, it can be concluded that a portion of the higher phthalates remains sorbed to the waste sludge. For instance, Zurmühl [61] reported sludge concentrations (in units of mg kg–1 dry weight (d.w.)) ranging from 2.6 to 263 for DBP, non-detectable to 0.7 for BBP, and 65.8–481 for DEHP, while no DMP or DEP was detected. Calculated mean values were 35, 0.3, and 179 for DBP, BBP, and DEHP. The DBP mean is considerably influenced by a single high value, as the median is 4.1 mg kg–1. Mickelson et al. [63] used a WWTP computer model (SimpleTreat) and compared the results with removals in three Danish WWTPs to evaluate the biodegradation rate constant. They assumed the rate constant was the major source of error in the model. Observed removals for DEHP in three WWTPs averaged 85% and the modeled results for the same three plants averaged 93%.A calculation of the biodegradation rate necessary to obtain this result, gave 0.55 L gss–1 d–1 at 20 °C. Assuming a standard sludge retention time of 8.5 d, the extent of biodegradation for a system with no primary settler was 11% and for a system with a primary settler was 6%. They calculated that the major portion of the influent DEHP, 78% and 56% for these two treatment systems, respectively, goes to waste sludge. Clark et al. [64] published results on a model that evaluates fate of PDEs in a WWTP (STP). They estimated half-lives of DBP and DEHP of 100 h in a 2 g L–1 MLSS system. This is equivalent to a rate of 0.083 L gss–1 d–1, somewhat lower than that
Degradation of Phthalate Esters in the Environment
107
determined by Mickelson et al. [63]. The resulting modeled removals were 81% for DBP and 91% for DEHP, with biodegradation accounting for 27% of the removal of each. We have also used an updated spreadsheet version (3.0) of SimpleTreat [65] to model the data given by Furtmann [23] for WWTPs, treating industrial wastewater and domestic wastewater. Furtmann analyzed the influent and effluent wastewater concentrations in addition to activated sludge concentrations for DEP, DIBP, DBP, BBP, DEHP, and DOP. A mass balance is not strictly possible, since the analyses were performed on grab samples that Furtman characterizes as “random,” although they were all taken on the same day. The problem is that the differences in sludge and water retention times may result in a lengthy delay before a change in influent concentration manifests itself in sludge concentration, as may be seen in the DEHP concentrations in the domestic WWTP in Table 4. The influent concentration of DEHP is quite low, while the sludge concentration is quite high. For the industrial WWTP, analyses from two sampling dates were averaged. Table 4 shows the concentrations in these two plants in influents, effluents, and sludges. Standardized European conditions [66], including the default WWTP biodegradation rate for ready biodegradable chemicals of 1 h–1, together with physical properties of the PDEs from Cousins and Mackay [67], were used in the model. The calculated concentrations shown in Table 4 were obtained for a WWTP without primary sedimentation. The default biodegradation rate of 1.0 h–1 greatly underestimates the total effluent concentrations, except for DEP. The amount calculated as un-biodegraded is greater for the higher molecular weight PDEs, although the rates are all the same. This is a result of the assumption that no degradation occurs in the adsorbed phase. The SimpleTreat model (version 3) also has the option of assuming the same first-order kinetic rate in both the aqueous and solid phase. By using the model in this mode, one must assume a lower rate constant to fit the measured data because 1.0 h–1 results in an overestimation of total removal. Adjustment of the rate constant to 0.03 h–1 gives approximately correct concentrations in sludge (77 mg kg–1) and effluent (2.3 µg L–1, mostly associated with suspended solids). However, the assumptions of degradation in the solid phase and a rate of 0.03 h–1 gives too low a total removal for DEP with 77% of the influent concentration still in the aqueous effluent. The rate has to be raised to nearly 1 h–1 to get agreement with the measured removals of DEP. Thus, the truth probably lies somewhere in the middle, with biodegradation occurring in solid and liquid phase at different rates. The main point is that degradation is probably occurring in the sorbed phase. This may be partly a result of sorption to organisms capable of degrading PDEs. It also may be a result of the fact that interfacial surfaces may serve to concentrate both the PDE and degrading organisms [8]. Wang et al. [68] have reported that microbial cells immobilized within gel beads biodegrade DBP faster than freely suspended cells. Another phase that may contain PDEs in sewage is small particles of abraded PVC. There is evidence that plastic particles, probably abraded from surfaces by mechanical action, are found in wastewater and sludge. Teinpoint et al. [69] used pyrolysis GC-MS to quantify PVC in sludges from nine different WWTPs. The concentration of PVC ranged from 18 to 508 mg kg–1 (d.w.). In these sludge sam-
c
b
a
Value Measured Modeled b, kb1 = 1.0 h–1 Modeled, kb2 = 0.03 h–1 Measured Modeled, kb1 = 1.0 h–1 Modeled, kb2 = 0.03 h–1 Measured Modeled, kb1 = 1.0 h–1 Modeled, kb2 = 0.03 h–1 Measured Modeled, kb1 = 1.0 h–1 Modeled, kb2 = 0.03 h–1 Measured Modeled, kb1 = 1.0 h–1 Modeled, kb2 = 0.03 h–1 Measured Modeled, kb1 = 1 h–1 Modeled, kb2 = 0.03 h–1 0.52 0.52 0.52 2.3 2.3 2.3 1.3 1.3 1.3 0.8 0.8 0.8 25.0 25.0 25.0 1.3 1.3 1.3
Influent (µ L–1)
0.06 0.07 0.40 0.07 0.33 0.36 0.22 0.19 0.21 n.d.c 0.12 0.07 n.d. 3.80 0.80 0.03 0.20 0.04
– 0.01 0.05 0.3 2.0 2.1 0.8 1.1 1.2 1.0 1.4 0.8 14 128 27 153 6.6 1.4
1.0 1.0 1.0 5.4 5.4 5.4 8.2 8.2 8.2 2.4 2.4 2.4 36.0 36.0 36.0 71.0 71.0 71.0
0.06 0.01 0.77 0.08 0.78 1.3 0.30 1.2 1.3 n.d. 0.35 0.20 n.d. 5.5 1.2 0.08 10.9 2.3
Effluent (µ L–1)
Influent (µ L–1)
Effluent (µ L–1)
Sludge (mg kg–1)
Industrial a
Domestic
Mean of measured values from two industrial WWTPs. kb1 first-order biodegradation rate applied only to aqueous phase, kb2 applied to both liquid and solid phases. n.d. not detected.
DEHP
DOP
BBP
DBP
DIBP
PDE DEP
Wastewater source: Concentrations:
Table 4. PDE removal by WWTPs, measured and modeled
< 0.01 0.02 0.10 0.02 4.6 7.4 1.4 6.9 7.4 <0.01 4.3 2.4 2.3 184 38 92 360 76
Sludge (mg kg–1)
108 D.R. Peterson and C.A. Staples
109
Degradation of Phthalate Esters in the Environment
ples, the sum of DEHP, DINP, and DIDP (phthalates commonly used as plasticizers in PVC) concentrations ranged from 14 to 297 mg kg–1. Furthermore, the PDE concentration (sum of these three) correlated very closely with the concentration of PVC (r2 = 0.89). If PDEs are contained within a plastic matrix, one would expect that they are not very bioavailable for biodegradation. The degradation of PDEs in the solid phase will be discussed further as a separate topic. For purposes of modeling the rate of biodegradation of PDEs in wastewater, the assumption of a biodegradation rate of 1 h–1 seems to provide a conservative estimate for the more water-soluble PDEs (e.g., DMP, DEP, DBP, and BBP). However, if the model algorithm does not provided for degradation in the sorbed phase, this rate provides too low an estimate of biodegradation for the higher PDEs. A more appropriate model is one that allows degradation in the sorbed phase, in which case, a rate of 0.03 h–1 for overall degradation seems to fit the higher PDE. Recently, Fauser et al. [70] published a report on a detailed investigation of the behavior a number of phthalates (and two surfactants) in a specific, municipal wastewater treatment plant (Bjermarken WWTP in Roskilde, Denmark.) They analyzed influent and effluent, together with sludge, for concentrations of DEHP, DPP, DBP, BBP, DOP, and DNP. Influent concentrations were daily composites, influent were daily grab samples, and sludge was a single day sample. Total hydraulic retention time of the plant is 46 h (19 h for biological treatment) and the sludge age (residence time) is 20 days. The analytical results are shown in Table 5. DEHP was present in influent wastewater at 35 µg L–1 and the other PDEs were all at or below about 1 µg L–1. Removals were >95% for all but BBP and DBP. A purpose of the Fauser et al. [70] study was to evaluate the applicability of a continuous flow model like SimpleTreat to a site specific model of a WWTP with alternating aerobic and anoxic cycles (4 h cycle) from separate treatment units. They found that the variation from the constant steady-state removal rate brought about by the cyclic nature of the process was far less for substances that are hydrophobic and adsorb to the sludge than for water-soluble substances. Their modeling also showed that for substances with half-lives longer than about Table 5. Measured concentrations and modeled biodegradation in Roskilde [70]
Phthalate
DBP BBP DPP DEHP DOP DNP a b c d
Measured Concentrations Inlet (µg L–1)
Outlet (µg L–1)
II° Sludge (mg kg–1)
1.03 0.39 0.07 35.4 0.57 0.44
0.91 0.13 0.008 0.96 0.013 0.013
0.16 0.01 n.d. d 3.51 0.05 0.05
k1 a (h–1)
–c 0.009 0.053 0.032 0.024 0.021
Disposition (%) Biodegraded
Aqueous effluent
Sludge b effluent
– 48 71 70 63 61
– 15 8 2 3 2
– 37 21 28 34 37
Pseudo-first order biodegradation rate for aerobic degradation obtained by modeling measured data in aggregate flow model. Disposition values for primary and secondary sludge were combined. Due to insufficient data, DBP was not modeled. n.d. not detected.
110
D.R. Peterson and C.A. Staples
2 h, a simpler model of the aggregate flows could be used. They assumed that biodegradation occurred only in the soluble state and represented the aerobic pseudo-first-order rate constant as k1/R where R is the sorbed fraction, equal to 1+ kd · [particulate mater]. Fitting the model to the measured data, they determined half-lives equivalent to the aggregate biodegradation rates shown in Table 5. These rates range from 0.009 h–1 for BBP to 0.053 h–1 for DPP. The corresponding fraction of the total influent biodegraded is also shown in Table 5. These ranged from 48% for BBP to 71% for DPP. 4.4 Anaerobic Rates
The fate of PDEs in anaerobic sewage sludge is important, since many WWTPs use an anaerobic digestion process prior to sludge disposal. Shelton et al. [71] reported on the biodegradation of a number PDEs in laboratory digesters containing undiluted sludge and 10% sludge. The systems were purged with a gas mix of 90% N2 and 10% CO2 . Initial PDE concentrations were 20 mg L–1.Aliquots were removed at various intervals, dried, and extracted with solvent, and the PDEs were determined by GC. Although they did not report rates of biodegradation, rates may be evaluated from their graphs of the data. For the undiluted sludge, DMP, DEP, and DBP all were essentially completely degraded by the first sampling interval (7 d), so the rates for these were >0.3 d–1. BBP had a degradation rate of 0.056 d–1. DEHP and DOP degraded considerably slower, with rates of 0.001 d–1 and 0.006 d–1, respectively. In the 10% sludge experiment the systems were sampled more frequently and the rates were slower; thus, rates of 0.42, 0.069, 0.073, and 0.096 d–1were calculated for DMP, DEP, DBP, and BBP, respectively. DEHP and DOP had quite inconsistent results, with all concentrations, including time zero, varying between 40 and 60% of nominal. There must have been some degradation in the 10% sludge, since both gave about 10% of theoretical methane production, while the other PDEs gave 80–100%. Subsequently, Ziogou et al. [72] determined the rates of biodegradation of six PDEs in anaerobic sewage sludge, amending the sludge at PDE concentrations in the range of 30–600 mg kg–1. They reported pseudo-first-order rates of 0.21, 0.14, 0.25, and 0.16 d–1 for DMP, DEP, DBP, and BBP, respectively. Except for DMP, these rates are about twice as fast as those derived from the data of Shelton et al. [71]. DOP and DEHP did not degrade measurably during the 32 day test. More recently, Painter and Jones [73] studied the anaerobic biodegradation of DBP, BBP, and DEHP in an anaerobic biodegradation test system by using 10% inocula from various sources. The PDEs were incubated at 0.2 mM and analyzed by hexane extraction-GC. The anaerobic sludge inoculated system gave degradation of DBP and BBP but not DEHP. The rates calculated from their data are 0.025 d–1 for DBP and 0.013 d–1 for BBP. Wang et al. [74] also studied the rate of PDE biodegradation in undiluted anaerobic digester sludge from a domestic WWTP. They report first-order rate constants of 0.696 d–1 for DMP, 0.518 d–1 for DBP, and 0.0336 d–1 for DOP. Johnson and Lulves [75] studied the biodegradation of DBP in anaerobic pond sediment. They used 14C-DBP and separated parent form metabolites
Degradation of Phthalate Esters in the Environment
111
by thin-layer chromatography. They present data on the percent recovery of DBP versus control after 1, 5, 7, and 14 days of incubation that may be used to calculate a first-order rate of degradation of 0.27 d–1. The anaerobic degradation of DMP, DBP, and DEHP in garden soil was studied by Shanker et al. [44]. They amended the soil with 500 mg kg–1 of PDE. The soil was then flooded with water and the flasks stoppered to achieve anoxic conditions. The PDEs and PA were extracted and analyzed at various intervals. Although they did not calculate rates, these may be calculated from their tabulated data. The calculated first-order rates are 0.033 d–1, 0.036 d–1, and 0.013 d–1 for DMP, DBP, and DEHP, respectively. Painter and Jones [73] reported on the use of freshwater and salt marsh inocula (10% v/v) in the anaerobic test system that they used for sludge. The rates of anaerobic degradation calculated from their data are 0.13 d–1 and 0.076 d–1 for DBP and BBP, respectively, with the freshwater sediment. The saltwater sediment showed similar activity, higher for DBP (0.31 d–1) and lower for BBP (0.051 d–1). Madsen et al [76] also studied the anaerobic biodegradation in sediment of a number of chemicals, including DMP and DIBP, but they only measured ultimate degradation by using methane production. This end point is unlikely to be equivalent to parent chemical degradation, since sulfate- and nitrate-reducing organisms may be present in sediment. Chauret et al. [77] reported on the biodegradation of DBP in subsurface soil microcosms under aerobic and nitrate-, iron-, and sulfate-reducing redox conditions. They report a rate of 2.05 nmol g-sediment–1 d–1 in aerobic systems and apparently zero-order rates of 0.86, 0.50, and 0.18 nmol g-sediment–1 d–1 for nitrate-, iron-, and sulfate-reducing conditions, respectively. It is apparent by comparison with the aerobic rate, that anaerobic biodegradation via nitrate- and sulfate-reducing pathways may be appreciable.Wang et al. [78] have reported on the isolation of a nitrate-reducing organism growing on and completely degrading DBP. For reasons of lack of environmental relevance, we omit these rates in laboratory systems using isolated pure strains of bacteria. Ejlertsson et al. [79–81] studied anaerobic degradation in laboratory systems inoculated with diluted (1%) samples from a digester treating municipal solid waste. For DEP, after an initial adaptation phase of up to 20 d, degradation proceeded rapidly at a rate of 0.7 mg g–1 d–1 [79]. In experiments terminated at 278 days they found complete removal of DBP, 77% removal of BBP, and 19% removal of DEHP. Ejlertsson et al. [82] studied the ultimate anaerobic degradation (methane production) of a range of PDEs and of their constituent alcohols. They found that PDE with water solubilities above about 50 µg L–1, such as DBP, DBP, and DHxP, were degraded within the 35–100 d incubation period, while DEHP, DOP, and DDP, with solubilities below about 3 µg L–1 showed no conversion to methane. The higher alcohols such as 2-ethylhexanol, octanol, and decanol were well degraded. The authors ascribed the difference in PDE biodegradability to lack of solubility of the higher PDEs, although it may as well be due to a lack of bioavailability due to their hydrophobicity and strong adsorption to the high solids content in anaerobic test systems. As discussed earlier, Fauser et al. [26] developed a model to reflect the concentration of DEHP in the depth profile of sediments in Roskilde Fjord, Den-
112
D.R. Peterson and C.A. Staples
mark. The model takes into account the rates of sedimentation, aerobic biodegradation, and anaerobic biodegradation. Fitting of the model to experimental sediment concentrations gave an aerobic rate constant for degradation of 2 ¥ 10–5 s–1 and an anaerobic rate constant of 8 ¥ 10–6 s–1 (below 5 cm.). This latter rate constant converts to a DEHP anaerobic biodegradation rate of 0.69 d–1 (t1/2 = 1.0 d). The reported rates of anaerobic biodegradation of PDEs are summarized in Table 6. For the less hydrophobic PDEs, the pseudo-first-order rates ranged from Table 6. Summary of biodegradation rates of PDEs in anaerobic environments
Phthalate
First-order rate (d–1)
Half-life (d)
Test conditions
Ref.
Anaerobic WWTP digester sludge, batch incubation DMP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DMP 2.8 Undiluted sludge, 0.5–10 mg L–1 DMP 0.25 a 0.696 1.0 Undiluted sludge, 10 mg L–1 DMP 0.42 3.3 10% diluted sludge, 20 mg L–1 DMP DEP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DEP 5.0 Undiluted sludge, 0.5–10 mg L–1 DEP 0.14 a 0.069 10.0 10% diluted sludge, 20 mg L–1 DEP DBP >0.3 <2.3 Undiluted sludge, 20 mg L–1 DBP 2.7 Undiluted sludge, 0.5–10 mg L–1 0.26 a 0.581 1.19 Undiluted sludge, 10 mg L–1 DBP 0.073 9.5 10% diluted sludge, 20 µg mL–1 DBP 0.025 27.7 10% diluted sludge, 0.2 mM DBP BBP 0.056 12.4 Undiluted sludge, 20 mg L–1 BBP 3.7 Undiluted sludge, 0.5–10 mg L–1 BBP 0.19 a 0.096 7.2 10% diluted sludge, 20 mg L–1 BBP DEHP 0.001 693 Undiluted sludge, 20 mg L–1 DEHP Nil – Undiluted sludge, 0.5–10 mg L–1 DEHP Nil – 10% sludge, 20 mg L–1 DEHP Nil – 10% sludge, 0.2 mM DEHP DOP 0.006 115 Undiluted sludge, 20 mg L–1 DOP Nil – Undiluted sludge, 0.5–10 mg L–1 DOP 0.0336 20.6 Undiluted sludge, 10 mg L–1 DO
[71] [72] [74] [71] [71] [72] [71] [71] [72] [74] [71] [73] [71] [72] [71] [71] [72] [71] [73] [71] [72] [74]
Anaerobic soil and sediment DMP 0.033 DBP 0.036 0.27 0.13 0.31 BBP 0.076 0.051 DEHP 0.013 Nil 0.69
[44] [44] [75] [73] [73] [73] [73] [44] [75] [26]
a
21.0 19.3 2.6 5.3 2.2 9.1 13.6 53.3 – 1.0
Flooded soil, 500 mg kg–1 Flooded soil, 500 mg kg–1 Pond sed.: water (1:2), 1 mg L–1 DBP 10% freshwater sed., 0.2 mM DBP 10% salt marsh sed., 0.2 mM DBP 10% freshwater sed., 0.2 mM DBP 10% salt marsh sed., 0.2 mM DBP Flooded soil, 500 mg kg–1 Pond sed.: water (1:2), 1 mg L–1 DEHP Field data, sed. DEHP concentrations
Mean of determinations at three PDE concentrations, 0.5, 1, and 10 mg L–1.
Degradation of Phthalate Esters in the Environment
113
about 0.2 to 0.6 d–1 for undiluted anaerobic sludge, corresponding to half-lives of about 1-5 days.A 10% sludge inoculum in an anaerobic, aqueous system resulted in somewhat lower rates. Anaerobic sludge biodegradation of BBP was a bit slower, with rates of about 0.06–0.2 d–1, while the more hydrophobic DEHP and DOP were degraded very slowly in anaerobic sewage sludge with reported halflives of nearly two years (693 d) for DEHP and 21–115 d for DOP.Anaerobic sediment studies of DBP and BBP gave biodegradation rates within the same ranges as seen for anaerobic sludge biodegradation even though most of these were carried out at 10% inoculum. The flooded soil studies of Shanker et al. [44] gave somewhat lower rates for DMP (0.033 d–1) and DBP (0.036 d–1) than was seen with anaerobic sludge, while DEHP showed a higher rate (0.013 d–1) than reported for sludge. 4.5 Solid-Phase Biodegradation
There have been numerous reports of the growth of microbes on the surface of fabricated PVC plastic materials. This susceptibility of PVC to biodeterioration is principally due to the presence of biodegradable plasticizers in the polymer [83]. Webb et al. [83] isolated and identified a large number of fungi that colonized the surface of dioctyl adipate (DOA)-plasticized PVC. The most active strains were able to utilize the plasticized PVC as a sole source of carbon, produced an extra-cellular esterase, and caused weight loss of the PVC. Other investigators have also isolated large numbers of fungi and yeasts capable of utilizing PDEs as a source of carbon [84, 85]. El-Sharouny [86] isolated plasticizer-degrading thermophilic fungi from Nile River mud. All but four of the 19 fungi isolated could degrade DBP and DOP, the two phthalate plasticizers used. He then measured weight loss in plastic strips containing 33% plasticizer, in liquid cultures inoculated with each of the six most active degraders. The percent weight loss was corrected for simple diffusional loss with a non-inoculated poisoned control. Losses in 14 days ranged from 2.3 to 11.9% for DBP and 2.8–12.5% for DOP.Yabannavar and Bartha [87] studied the biodegradability of a number of plasticized plastic films in soil microcosms. One of these films contained DOP (24.8%) as the plasticizer. The soil was a sandy loam containing 5% organic matter and maintained at 50% moisture capacity and 27 °C. The film was amended to the soil as 2¥3 mm pieces at 1% w/w and the soil incubated aerobically for three months.Weight determinations were made by Soxhlet extraction of the PVC with methyl ethyl ketone (MEK), and the plastic was precipitated and weighed. The PVC film containing DOP lost 20.8% of its weight. This is likely to all be DOP loss because the authors report that no plasticizer was detectable by GC in the MEK extracts. On a weight basis, this is 84% biodegradation of the DOP.Although the time course of biodegradation was not determined on the DOP-plasticized film, it was determined on dioctyl adipateplasticized films, which showed similar weight loss in three months. On average, these lost 64% in the first month and then very little in the two subsequent months. The authors speculate that this result may be due to a residual fraction trapped within the film.
114
D.R. Peterson and C.A. Staples
Gumargalieva et al. [88] evaluated the kinetics of PDE loss from PVC by using a culture of Aspergillus niger isolated from PVC insulation of wires. The overall rate of PDE loss depends on both diffusion through the PVC to the surface and desorption from the surface. They reported that under their conditions, desorption from the surface of the plastic was slower than diffusion. They measured overall loss rate of PDE from PVC over 12 months with and without surface growth of the fungus. They found rates of 8.6 ¥ 10–4 d–1 (t1/2 = 800 d) in the presence of the fungus and 1.38 ¥ 10–4 d–1 (t1/2 = 5000 d or 13.8 y) in its absence. By comparing these rates with calculated rates for diffusion and desorption in the absence of the fungus, it was shown that the volatilization of the PDE is rate limiting without the fungus. With the fungus, the rate of loss was limited by diffusion out of the PVC. The loss mechanism in the presence of the fungus is thus biodegradation rather than volatilization. Obviously, such solid-phase biodegradation of PDEs is very slow in comparison to their biodegradation rates in soil or water. The point is though, that PDEcontaining plastic articles in soil, landfills, and even in the open air are likely to be surface colonized by phthalate-degrading microorganisms and the PDE degraded at the surface as rapidly as it diffuses out. This diffusion rate is slow and dependent on the size and shape of the plastic items (Osmon et al. [89]). There is also the possibility that in some polymers a fraction of PDE that is trapped within the polymer matrix and cannot diffuse out [87]. The implication, however, is that plastics in soil and landfills are not likely to be a source of PDE emission to water or air because their surfaces may be colonized by slow growing fungi, yeast, and bacteria that can biodegrade the PDE as it diffuses out of the item.
5 Degradation of Phthalate Metabolites The metabolic pathways for PDE biodegradation have been reviewed [1]. There is abundant evidence that the pathway of PDE metabolism, at least by microorganisms, is via a stepwise hydrolysis of the two ester bonds, first giving the monoester plus the free alcohol, followed by hydrolysis of the second ester bond to give PA and alcohol. The PA is degraded aerobically via hydroxylation and decarboxylation to give protochatechuic acid, which is further degraded to carbon dioxide through either ortho or meta cleavage of the aromatic ring. PA is degraded anaerobically via decarboxylation to benzoic acid, which is further degraded through ring saturation. Most of the studies delineating these pathways were conducted using isolated PDE degrading bacteria [90–93]. In these studies, there were differences between the microorganisms in their further metabolism of the immediate degradation products, phthalate monoester (PME), and PA. For instance, Englehardt et al. [92] reported that four isolates from enrichment cultures (Penicillium lilacinum and three bacteria, two gram positive and one gram negative) together with three organisms from stock cultures (Corynebacterium petrophilum, Arthrobacter hydrocarboglutamicus, and Mycobacterium phlei) all formed MBP from DBP almost quantitatively. MBP was isolated by thin-layer chromatography (TLC) and no other metabolites were detected. On the other hand, they found that three other
Degradation of Phthalate Esters in the Environment
115
strains of coryneform bacteria degraded DBP quickly with only a transient appearance of MBP in the medium.All three strains could be grown on MBP or PA. Other workers have reported on individual microorganisms growing on PDEs with only a transient appearance of PME or PA in the medium [92, 93]. Kurane et al. [30] followed the time course of degradation of DEHP by Pseudomonas acidovarans and saw a transient appearance of soluble metabolites. They analyzed these by TLC and identified the soluble metabolites as PA and protocatechuic acid. Kurane et al. [31] also analyzed the intermediates produced from DEHP hydrolysis by a purified phthalate ester hydrolase from Nocardia erythropolis. They found only PA and no MEHP and concluded that “the purified enzyme rapidly converts phthalate diesters into phthalic acid without phthalate monoesters accumulating.” This implies that the hydrolysis is accomplished by the same enzyme and that hydrolysis of the monoester is faster than of the diester. However, studies on single strains of microorganisms in growth medium are not entirely relevant to the rate of biodegradation of PMEs under environmental conditions. There are some studies available on mixed inocula or microcosms, wherein the degradation of PME is reported, including its decrease in concentration over time. The mineralization of 14C-labeled DBP, MBP, and PA in soil was studied by Inman et al. [94]. Their graphs indicate that both MBP and DBP produced 14CO2 at about the same rate, while PA produced it much more quickly. These data seem to indicate that metabolism of the phthalic acid through decarboxylation is faster than hydrolysis of the ester linkages and that hydrolysis of DBP is as fast or faster than hydrolysis of MBP. On the other hand, Shanker et al. [44] only detected parent PDE and phthalic acid in garden soil degradation studies of DMP, BBP, and DEHP. At each time interval, for all three PDEs, the PA was always much lower in concentration than the PDE. In the aerobic studies, the maximum concentration of PA never exceeded 3% by weight of the original amount of PDE added. In the studies of the fate of DEHP in soil by Schmitzer et al. [95], radiolabeled DEHP was employed and soil residues were extracted and analyzed by TLC. In laboratory studies, after seven days, 89.8% of the radioactivity was unchanged DEHP and the other residues were too low in amount to quantify. In outdoor studies, after 111 days, 3% of the residual radioactivity was DEHP, 0.14% was MEHP, and 0.35% was PA. In laboratory soil microcosms and outdoor lysimeters, Rüdel et al. [43] were able to detect only DEHP and no MEHP or other metabolites by using radiolabeled DEHP and HPLC separation.All of these studies, with the exception that of Inman et al. [94], indicate that in aerobic soil, degradation of the PME is faster than degradation of the PDE. A simple first-order model may be constructed for the hydrolysis of PDE to PME and of PME to PA. If the rates of the two reactions are equal, and assuming irreversibility, the concentration of PME will reach a maximum of about 37% of the starting concentration of PDE. This is also the crossover point, after which the concentration of PME will exceed the concentration of PDE. To achieve a situation in which the concentration of PME is always less than the concentration of PDE, the rate of PME hydrolysis must be at least twice the rate of PDE hydrolysis. These situations are illustrated in Fig. 2a and 2b. Of course, if the enzyme systems for intermediate breakdown need to be induced, a lag in their further catabolism may occur. For the Micrococcus strain studied by Eaton and Ribbons
116
D.R. Peterson and C.A. Staples
a
b Fig. 2. a Modeled time course of degradation for a PDE (solid line) and intermediates. Rates: diester – 1.0 d–1, monoester – 1.0 d–1, phthalic acid – 1.0 d–1. b Modeled time course of degradation for a PDE (solid line) and intermediates. Rates: diester – 1.0 d–1, monoester – 2.0 d–1, phthalic acid – 1.0 d–1
Degradation of Phthalate Esters in the Environment
117
[91], the esterase enzyme activities were constitutive but the enzyme catalyzing the hydroxylation of PA was inducible. In the pure species (Nocardia sp.) work of Engelhardt and Wallnöfer [93], curves for DBP disappearance and transient concentrations of MBP appeared as in Fig. 2b, in which the rate of monoester hydrolysis is twice that of the diester. On the other hand, a similar graph presented by Karegoudar and Pujar [96] for DEP degradation by a Micrococcus sp. looked more like Fig. 2a, in which the two hydrolysis rates are equal. The initial, first-order rate of PA degradation in sludge-amended soil was determined by Roslev et al. [51] as 0.32 d–1, under aerobic conditions. Under anaerobic conditions, it appears that such a lag in PA degradation does occur. Shelton et al. [71] presented a graph of micromolar concentrations of BBP, MBuP, and PA versus time during anaerobic incubation in 10% sludge. Only a small amount of MBuP was found, reaching a maximum on day 3 of about 5% of the initial concentration of BBP (concentrations are not presented for MBnP). PA did not reach a maximum until day 10, when it was about 80% of initial BBP. Analysis of the rate of BBP disappearance (linear regression of the natural logarithm of concentration versus time) gives a rate of BBP degradation of 0.21 d–1. Curve “stripping” and regression analysis of the terminal rate of PA degradation gives a rate of 0.31 d–1. By using these rates in the kinetic model described in the previous paragraph, the maximum reached by the PA concentration would be 32% and occurs on day 4. However, if one assumes a 9-day lag before PA degradation begins, the maximum occurs on that day at about 80% of the initial BBP concentration. The data of Shanker et al. [44], on the other hand, show that for DMP, DBP, and DEHP in anaerobic soil systems, the transient appearance of PA never reached more than 10% of the amount of PDE present or more than 6% of the initial amount. Similarly, Johnson and Lulves [75] reported only low concentrations of PA during DBP biodegradation in anaerobic freshwater sediment. Ejlertsson et al. [79, 80] have observed very similar results to those of Shelton et al. [71] in diluted (1%) anaerobic municipal solid waste digester (biogas reactor). By using DEP in concentration ranges of 50–250 mg L–1, they found proportional amounts (70–75%) of PA produced which took considerable time to degrade beyond complete DEP degradation [79]. In addition, they observed a lag of about 20 days in the degradation of DEP. Degradation of PA commenced on about day 30 and was not complete until day 60–80. Only small amounts of MEP (<15%) were produced and quickly disappeared. However, this same group [80] also reported that in an incubation of DEP with anaerobic solid waste from a model landfill, only 43% was degraded in 110 days and more MEP was produced (40%) than PA (5.7%). By using this same inoculum, anaerobic degradation of BBP resulted in 78% degradation to MBuP (29%), MBnP (43%), and PA (14%). By contrast, PA itself was completely degraded in 110 days. Thus, the source of inoculum appears to influence the relative rates of biodegradation of parent PDE, PME, and PA. Ejlertsson and Svendsson [81] also studied the anaerobic degradation of metabolites of DEHP, namely MEHP, 2-ethylhexanol (2-EH), and 2-ethylhexanoic acid (2-EHA). Again, by using inocula from digestion of solid waste (biogas system), they found that MEHP degraded quickly with no lag phase and a half-life of about 7 days. The PA produced reached a stoichiometric amount and did not
118
D.R. Peterson and C.A. Staples
begin to degrade until about day 40. It then degraded rapidly in about 15 days. Rates are roughly 0.1 d–1 for each. 2-EH and 2-EHA both showed lags of 6 and 8 d, respectively, before rapid degradation began. After this lag phase, 2-EHA degraded in 4 days and in 5 days once degradation began. 2-EH also showed the appearance of a stoichiometric amount of 2-EHA as a transient intermediate, reaching a maximum at day 10 and completely degrading at day 15. The degradation of the intermediates, both the monoester and the alcohols are more rapid than the degradation of the parent DEHA. This may be partially due to the greater bioavailability of these more polar and more water-soluble metabolites. Given the difference in hydrophobicity between PMEs and PDEs, especially for the higher phthalates, models predict that the PMEs will partition much more strongly to water than to solids in WWTP effluents. Thus, one would expect much higher concentrations of PMEs than PDEs in aqueous effluent if PMEs were present at steady state. In fact, Paxéus [97] has reported on the concentrations of PMEs and PDEs from three large WWTPs in Sweden. Two of the plants had detectable amounts of DEP, DBP, BBP, and DEHP. All concentrations were in the range 3–22 µg L–1. Only two of the corresponding PMEs were detected, MEP and MBuP. At one plant these were present at lower concentrations than the corresponding PDEs (0.5 and 4 versus 3 and 11) and at the other at about similar concentrations (11 and 20 versus 8 and 22). At the third WWTP, the only PDE detected in aqueous effluent was DBP (6 µg L–1), while MBuP and MEP were detected at low levels (0.5 µg L–1, each).Again, these data suggest that much lower concentrations of PMEs, as compared to PDEs, appear to be present in activated sludge during wastewater treatment and that this must be a consequence of a more rapid degradation rate of PME than PDE. In conclusion, in almost every circumstance in which the kinetics of metabolite formation and disappearance has been quantified, the rate of PME degradation is twice that of the corresponding PDE or more. Routes of degradation of the PMEs in the environment, other than biodegradation, are not expected to be significant. As discussed previously, abiotic hydrolysis of PDEs is quite slow and not expected to be significant. Thus, the formation of PMEs will be mainly due to metabolism and will occur within biota and in aqueous systems. Since the PDEs are carboxylic acids, they will be mainly dissociated at neutral pH and hence are not expected to be volatile. Abiotic hydrolysis rates of the PMEs are at least an order of magnitude slower than hydrolysis of the parent PDE [12].
6 Summary and Conclusions Owing to their major use as components in fabricated plastic articles, it is expected that PDEs will be released to the environment through slow volatilization or leaching. It is apparent that the diffusion of PDEs out of such articles is slow and also that organisms may colonize surfaces of plastics and metabolize PDEs as fast as they diffuse to the surface. Such articles when exposed to the atmosphere may also emit PDEs through volatilization or leaching by rainwater. Indirect photooxidation in surface films may also play a role in PDE degradation.
Degradation of Phthalate Esters in the Environment
119
Phthalates emitted in household wastewater may still be contained within small particles of abraded plastic. PDEs contained within plastics in landfills are unlikely to migrate due to their hydrophobicity. In surface water, hydrolysis is not expected to contribute very greatly to the overall degradation of PDEs (or PMEs), since their hydrolysis half-lives are of the order of a year or longer. Photolysis on the other hand may be a significant pathway for abiotic degradation. In aqueous systems, PDEs in surface layers may be degraded at a significant rate (1–3 h–1) due to indirect photolysis. However, the bulk of PDEs in the aquatic environment are expected to be particulate bound. No data are available on the rate of photodegradation in this state but this is also unlikely to be a major fate of PDEs. For the more volatile, low-molecular weight PEs (DMP, DEP, and DBP) and for all of the PEs when released to air, indirect photodegradation via hydroxyl radicals may also be a significant route of abiotic degradation. Photodegradation half-lives, with the exception of DMP, are in the range of a day or less in typical urban air where they are likely to be emitted except for DMP. Although a sizeable percentage of the PEs in air are likely to be associated with particulates, data for DEHP show that indirect photolysis of particulate sorbed PDE is about the same as in the vapor phase [7]. Thus, it is reasonable to use the modeled values shown in Table 1 as the likely half-lives for PDE degradation in air. The relatively low vapor pressures of PDEs (and PMEs) are expected to result in a major portion of releases going to wastewater, surface water, and soil. The PDEs in surface water are also likely to be mainly associated with suspended particulates and sediments. Thus, the rates of biodegradation of PDEs in WWTPs, surface water, sediment, and soil are expected to be major determinants of their overall environmental fate. The rate of aqueous biodegradation of all of the PDEs is in the approximate range of 0.2–2.0 d–1 in test systems that are intended to simulate environmental conditions, including the presence of suspended sediments. There are data on the biodegradation rates in soil for fewer of the PDEs than in water or sediment. However, soil biodegradation rates appear to be lower than aquatic ecosystem rates and to be lower for the higher PDEs. The biodegradation rates in soil of DMP and DEP are about 0.4 d–1, while that of DBP is in the range of 0.04–0.4 d–1 and DEHP rates are lower, at about 0.01–0.1 d–1. This lower rate for DEHP biodegradation in soil is likely to be due to its reduced bioavailability as a result of adsorption or sequestration. The rates of anaerobic biodegradation do not seem to vary greatly between sludge, sediment, and soil for the same PDEs and were about 0.6–0.06 d–1 in these systems for DMP, DBP, and BBP. On the other hand, DEHP and DOP show a relatively low rate of anaerobic biodegradability.Again, lower bioavailability is likely to be responsible for these lower rates. A range of 0.01–0.001 d–1 was taken as representative for these higher PDEs. Regarding the rate of biodegradation in wastewater treatment plants, the default rate given in the EU Technical Guidance Documents [66] of 1 h–1 for ready biodegradable chemicals, seems approximately correct when applied to DEP. That is, the use of this rate in the SimpleTreat model results in calculated effluent concentrations in water and sludge that are in approximate agreement with measured values. However, for the other PDEs, the model overestimates the concentrations in aqueous effluent and sludge. This is the result of an underestimation
120
D.R. Peterson and C.A. Staples
of total biodegradation, since biodegradation of adsorbed chemical is not included. Running the model in a mode that utilizes the same rate irrespective of the degree of sludge adsorption, results in too high an extent of removal at a rate of 1 d–1. A rate of about 0.03 d–1 in this mode gives a better result for DOP and DEHP but overestimates effluent concentrations for DEP. The recommended ranges for half-lives of phthalates derived in this review are presented in Table 7. The data are grouped by “low”- and “high”-molecular weight PDEs, since the values are quite similar within these two groups and since the amount of data does not support separate values for each individual PDE. Boethling et al. [98] analyzed biodegradability data for a number of chemicals and concluded that degradation rates are generally about as fast in surface soil as in freshwater. They also observed that half-lives of chemicals were 2–4 times longer under anaerobic conditions than aerobic. The values in Table 7 are in apTable 7. Recommended environmental degradation rates of PDEs
Environmental compartment Surface water & sediment
Route
biodegradation (aerobic) biodegradation (aerobic) biodegradation (aerobic) hydrolysis photolysis (indirect) Soil biodegradation (aerobic) biodegradation (aerobic) Soil & sediment biodegradation (anaerobic) biodegradation (anaerobic) biodegradation (anaerobic) Wastewater water phase slurry phase Air photolysis (indirect) photolysis (indirect) photolysis (indirect) photolysis (indirect) a b
Phthalate
Rate (d–1)
Half-life (d)
Half-life (h)
low MW a
0.2–2.0
3.5–0.35
84–8.4
high MW b
0.2–2.0
3.5–0.35
84–8.4
PMEs
0.4–4.0
1.7–0.17
42–4.2
all all
<7 ¥10–4 ?
>103 ?
>2 ¥104 ?
low MW
0.1–0.4
6.9–1.7
166–41
high MW
0.01–0.1
69–7
1663–166
low MW
0.06–0.6
11.6–1.2
278–28
high MW
0.006-.01
116–69
2784–1656
PMEs
0.1–1.0
6.9–0.7
166–17
low MW high MW DMP
24 0.75 0.048
0.029 0.92 14.4
0.7 22 346
DBP
0.29
2.4
58
low MW
0.8–1
0.9–0.7
22–17
high MW
3.5–1.2
0.6–0.3
14.4–7.2
The low-MW PDEs are considered those with C1 to C4 alcohol side-chains (DMP, DEP, DBP) plus BBP. The high-MW PDEs are those with C6 alcohol side-chains and above.
Degradation of Phthalate Esters in the Environment
121
proximate agreement with these conclusions; however, soil biodegradation halflives are about 2–4 times longer than aquatic for the lower PDEs and 20-fold longer for the higher molecular weight PDEs. This lower soil biodegradability for the higher PDEs is most likely due to their sorption to soil organic matter, rendering them unavailable for biodegradation. Soil anaerobic biodegradation is about the same as in aerobic soil for the lower PDEs but has 2–5 times longer half-lives for the higher PDEs. The differences in measured soil and sediment half-lives between the lower and higher PDEs do not agree with the EU Technical Guidance Documents (TGD) values, which are based, in part, on soil sorption coefficient (Kp) [66]. For “readily” biodegradable chemicals with a Kp of <100 L kg–1, the TGD specifies a half-life in soil of 30 days. The TGDs also specify an increase of one order of magnitude in half-life for each order of magnitude increase in Kp . Thus, for a chemical with a Kp of between 100,000–1,000,000 L kg–1, the specified half-life is 300,000 d or 822 y. DMP has a Kp of <100 L kg–1 while that of DEHP is between about 100,000 and 1,000,000. This assumed low rate is clearly not correct, based on the measured rates presented in this chapter. The TGD specified increase in half-life with increase in Kp is apparently based on the assumption that adsorbed chemicals are not degraded. The data and discussions on bioavailability and biodegradation of sorbed PDEs clearly contradict this premise. The ranges for rates and half-lives given in Table 7 account approximately for the variation in results from laboratory tests. In the environment the actual expected range of variation in these rates is likely to be much larger due to the large range of types and numbers of bacteria in different locations and variation in conditions such as temperature and pH. However, to use such fate data in modeling one must assume a standard condition of the model that is typical or representative. The studies that were considered in this review were environmentally relevant, generally using natural water or inocula from areas likely to be impacted by PDE emissions. The degradation half-lives used in the Modeling Chapter are longer than those recommended here. These half-lives were based on the upper end of the range of degradation rates reviewed by Staples et al. [1]. The use of these half-lives in modeling is expected to result in conservative estimates of the persistence and concentrations of PDEs in the environment.
7 References 1. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:667 2. Atkinson R (1988) Environ Toxicol Chem 7:435 3. Meylan WM, Howard PH (1993) Chemosphere 26:2293; AOPWIN version 1.89 is part of EPIWIN, a QSPR computer program available from Syracuse Research Corporation, North Syracuse, NY 4. Atkinson R (2000) Atmospheric oxidation. In: Boethling RS, Mackay D (eds) Handbook of property estimation methods for chemicals, environmental and health sciences. Lewis, Boca Raton, FL, p 335, chap 14 5. Ligocki MP, Pankow JF (1988) Environ Sci Technol 23:75
122
D.R. Peterson and C.A. Staples
6. Behnke W, Nolting F, Zetzsch C (1987) An aerosol smog chamber for testing abiotic degradation of compounds with low volatility. In: Greenlegh R, Roberts TR (eds) Pestic Sci Biotechnol, Proc Int Congr Pestic Chem Blackwell, Oxford, UK, p 401 7. Jin Z, Huang G, Chai Y, Zhong Y, Wang D, Li H (1999) Huanjing Huaxue 18:109 8. Södergren A (1982) Environ Poll A27:263 9. Howard P (1991) Handbook of environmental degradation rates. Lewis, Chelsea, MI 10. Gledhill WE, Kaley RG, Adams WJ, Hicks O, Michael PR, Saeger VW, LeBlanc GA (1980) Environ Sci Technol 14:301 11. Mill T (2000) Photoreactions in surface waters. In: Boethling RS, Mackay D (eds) Handbook of property estimation methods for chemicals, environmental and health sciences. Lewis, Boca Raton, FL, p 355, chap 15 12. Wolfe NL, Steen WC, Burns LA (1980) Chemosphere 9:403 13. Wolfe NL, Jeffers PM (2000) Hydrolysis. In: Boethling RS, Mackay D (eds) Handbook of property estimation methods for chemicals, environmental and health sciences. Lewis, Boca Raton, FL, p 311, chap 13 14. Scholz N, Diefenbach R, Rademacher I, Linnemann D (1997) Bull Environ Contam Toxicol 58:527 15. Painter HA (1995) Detailed review paper on biodegradability testing. OECD guidelines for the testing of chemicals. Environment monograph No 98. Organisation for Economic Cooperation and Development, Paris 16. Blok J, Balk F (1994) Guidance document for the interpretation of biodegradability test data. Report to the European Communities, Contract No B-3040/93/001114. BKH Consulting Engineers, Delft, The Netherlands 17. Aronson D, Howard PH (1999) Evaluating potential POB/PBT compounds for environmental persistence. Report No SRC-TR-99–020. Syracuse Research Corp. North Syracuse NY. 18. Boethling B (2000) HPVC-screening tool: using ready and inherent biodegradability data to derive input data for the EQC model. In: environment Canada, environmental categorization for persistence bioaccumulation and inherent toxicity of substances on the Domestic Substance List using QSARs, results of an international workshop hosted by Chemicals Evaluation Division of Environment Canada in Philadelphia, PA 11–12 Nov 1999. Final Report, Appendix 10. Environment Canada 19. Saeger VW, Tucker ES (1976) Appl Environ Microbiol 31:29 20. Carson DB, Saeger VW, Gledhill WE (1990) Use of microcosms versus conventional biodegradation testing for estimating chemical persistence. In: Landis WG, van der Schalie WH (eds) Aquatic toxicology and risk assessment: thirteenth volume. ASTM STP 1096. American Society for testing and Materials, Philadelphia PA, p 48 21. Adams WJ, Saeger VW (1993) Utility of microcosms for predicting the environmental fate of chemical. In: Gorsuch JW, Dwyer FJ, Ingersol DG, La Pointe TW (eds) Environmental toxicology and risk assessment. American Society for testing and Materials, Philadelphia PA, p 48 22. Walker WW, Cripe CR, Pritchard PH, Bourquin AW (1984) Chemosphere 13:1283 23. Furtmann K (1993) Phthalate in der aquatischen Umwelt. PhD Thesis, Universität Gesamthochschule Duisenberg. English Translation prepared for European Council for Plasticizers and Intermediates, Brussels, 1996. 24. Furtmann K (1995) Anal Methods Instrum 2:254 25. Ye C, Tian K (1990) Water Treat 5:474 26. Fauser P, Sørensen PB, Vikelsøe J, Carlsen L (2000) Fate of di-2-ethylhexyl) phthalate (DEHP) in Roskilde Fjord. Poster presented at the 20th international symposium on halogenated environmental pollutants & POPs, Dioxin 2000, Monterey CA, August 2000 27. Lewis DL, Kellogg RB, Holm HW (1985) Comparison of microbial transformation rate coefficients of xenobiotic chemicals between field-collected and laboratory microbiota. In: Boyle TP (ed) Validation and predictability of laboratory methods for assessing the fate and effects of contaminants in aquatic ecosystems. American Society for Testing and Materials, Philadelphia PA, p 3
Degradation of Phthalate Esters in the Environment
123
28. Skow KM (1982) Rate of biodegradation. In: Lyman WJ, Reehl, WF, Rosenblatt DH (eds) Handbook of chemical property estimation methods. McGraw-Hill, New York, chap 9 29. Steen WC, Paris DF, Baughman GL (1980) Effects of sediment sorption on microbial degradation of toxic substances. In: Baker RA (ed) Contaminants in sediments, vol 1: Fate and transport, case studies, modeling toxicology. Ann Arbor Science, Ann Arbor MI, p 477 30. Kurane R, Suzuki T, Takahara Y (1977) Agric Biol Chem 41:2119 31. Kurane R, Suzuki T, Fukuolka S (1984) Appl Microbiol Biotechnol 20:378 32. Nyholm N (1990) Chemosphere 21:1477 33. Gibbons JA, Alexander M (1989) Environ Toxicol Chem 8:283 34. Aichinger G, Grady CPL Jr, Tabak HH (1992) Water Environ Res 64:890 35. Wang J, Liu P, Quan Y (1995) Water Treat 10:331 36. Wang J, Liu P, Shi H, Qian Y (1998) Chemosphere 37:257 37. Wang X, Grady CPL Jr (1995) Water Environ Res 67:863 38. Yan H, Ye C, Yin C (1995) Environ Toxicol Chem 14:931 39. Rubin HE, Subba-rao RV, Alexander M (1982) Appl Environ Microbiol 43:1133 40. Pagga U (1987) Wasser Abwasser Forsch 20:101 41. Russell DJ, McDuffie B, Fineberg S (1985) J Environ Health A20:927 42. Kirchmann H, Aastroem H, Joensaell G (1991) Swed J Agric Res 21:107 43. Rüdel H, Schmidt S, Kördel W, Klein W (1993) Sci Total Environ 132:181 44. Shanker R, Ramakrishna C, Seth PK (1985) Environ Poll A39:1 45. Chen Y, Shen D, Hu Z, Liu X,Wu D, Zhao D, Zhang J (1997) Huanjing Kexue Xuebao 17:340 46. Wang J, Liu P, Shi H, Qian Y (1997) Chemosphere 35:1747 47. Cartwright CD, Thompson IP, Burns RG (2000) Environ Toxicol Chem 19:1253 48. Jensen J, van Langevelde J, Pritzl G, Krogh PH (2001) Environ Toxicol Chem 20:1085 49. Fairbanks BC, O’Connor GA, Smith SE (1985) J Environ Qual 14:479 50. Dörfler U, Haala R, Matthies M, Scheunert I (1996) Ecotoxicol Environ Saf 34:216 51. Roslev P, Madsen PL, Thyme JB, Henriksen K (1998) Appl Environ Microbiol 64:4711 52. Madsen PL, Thyme JB, Moldrup P, Roslev P (1999) Environ Sci Technol 33:2601 53. Hatzinger PB, Alexander M (1997) Environ Toxicol Chem 16:2215 54. Cornelissen G, van Noort PCM, Govers HAJ (1997) Environ Toxicol Chem 16:1351 55. Pignatello JJ, Xing B (1996) Environ Sci Technol 30:1 56. Northcott GL, Jones KC (2001) Environ Sci Technol 35:1103 57. Wang J, Liu P, Shi H, Qian Y (1997) Process Biochem (Oxford) 32:567 58. Petrasek AC, Kugelman IJ,Austern BM, Pressley TA,Winslow LA,Wise RH (1983) J Wat Poll Cont Fed 55:1286 59. Tokuz RY (1991) Hazard Waste Hazard Matter 8:245 60. Paxéus N, Robinson P, Balmer P (1992) Water Sci Technol 25:249 61. Zurmühl T (1990) Analyst 115:1171 62. Merkel D, Appuhn H (1996) Korrespondenz Abwasser 43:578 63. Mikkelsen J, Nyholm N, Jacobsen BN, Fredenslund FC (1996) Water Sci Technol 33: 279 64. Clark B, Henry GLH, Mackay D (1995) Environ Sci Technol 29:1488 65. Struijs J, Stoltenkamp J, van de Meent D (1991) Water Res 25:891 (SimpleTreat Ver 3.0 from Struijs J, 1997) 66. European Commission (1996) Technical guidance documents in support of the Commission Directive 93/67/EEC on the risk assessment of new notified substances and the Commission Regulation (EC) 1488/94 on risk assessment of existing substances. European Chemicals Bureau, Ispra, Italy 67. Cousins I, Mackay D (2000) Chemosphere 41:1389 68. Wang J, Liu P, Quan Y (1997) Environ Int 23:775 69. Tienpont B, David F, Vanwalleghem F, Sandra P (1995) J Chromatogr A911:235 70. Fauser P, Sørensen PB, Carlsen L, Vikelsøe J (2001) Phthalates, nonophenol and LAS in Roskilde wastewater treatment plant. NERI technical report no 354. National Environmental Research Institute, Denmark 71. Shelton DR, Boyd SA, Tiedje JM (1984) Environ Sci Technol 18:93
124 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98.
D.R. Peterson and C.A. Staples
Zigou K, Kirk PWW, Lester JN (1989) Water Res 23:743 Painter SE, Jones WJ (1990) Environ Technol 11:1015 Wang J, Chen L, Shi H, Qian Y (2000) Chemosphere 41:1245 Johnson BT, Lulves W (1975) J Fish Res Board Canada 32:333 Madsen T, Rasmussen HB, Nilsson L (1995) Chemosphere 31:4243 Chauret C, Inniss WE, Mayfield CI (1996) Ground Water 35:791 Wang J, Liu P, Qian Y (1999) Process Biochem (Oxford) 34:745 Ejlertsson J, Houwen FP, Svensson BH (1996) Swed J Agric Res 26:53 Ejlertsson J, Meyerson U, Svensson BH (1996) Biodegradation 7:345 Ejlertsson J, Svensson BH (1996) Biodegradation 7:501 Ejlertsson J, Alnervik M, Jonsson S, Svensson BH (1997) Environ Sci Technol 31:2761 Webb JS, Nixon M, Eastwood IM, Greenhalgh M, Robson GD, Handley PS (2000) Appl Environ Microbiol 66:3194 Jackson MA, Labeda DP, Becker LA (1996) J Ind Microbiol 16:301 Peciulyte D (1997) Biologija 2:29 El-Sharouny HMM (1993) Sohag Pure Appl Sci Bull 9:117 Yabannavar A, Bartha R (1993) Soil Biol Biochem 25:1469 Gumargalieva KZ, Zaikov GG, Semenov SA, Zhdanova OA (1999) Polym Degrad Stab 63:111 Osmon JL, Klausmeier RE, Jamison EI (1972) Rate-limiting factors in biodeterioration of plastics. In: Biodegradation of materials, vol 2. Applied Science, London, p 66 Keser P, Pujar BG, Eaton RW, Ribbons DW (1976) Environ Health Persp 18:156 Eaton RW, Ribbons DW (1982) J Bacteriol 151:48 Engelhardt G, Wallnofer PR, Hutzinger O (1975) Bull Environ Contam Toxicol 13:342 Engelhardt G, Wallnofer PR (1978) Appl Environ Microbiol 35:243 Inman JC, Strachan SD, Sommers LE, Nelson DW (1984) J Environ Sci Health B19:245 Schmitzer JL, Scheunert I, Korte F (1988) J Agric Food Chem 36:210 Karegoudar TB, Pujar BG (1984) Current Microbiol 11:321 Paxéus N (1996) Water Res 30:1115 Boethling RS, Howard PH, Beauman JA, Larosche ME (1995) Chemosphere 30:741
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 125– 177 DOI 10.1007/b11465
Observed Concentrations in the Environment Kathryn Clark 1 · Ian T. Cousins 2 · Donald Mackay 2 · Kentaro Yamada 3 1 2 3
BEC Technologies Inc., 61 Catherine Avenue, Aurora, Ontario, L4G 1K6, Canada E-mail:
[email protected] Canadian Environmental Modelling Centre, Environmental and Resource Studies, Trent University, Peterborough, Ontario, K9J 7B8, Canada CG Ester Corporation, Landic Bldg. 8F, 2-16-13, Nihonbashi, Chuo-ku, Tokyo 103-0027, Japan
Measured concentrations of six phthalate esters in seven environmental media are compiled and analyzed. The data are predominantly from Europe, the United States, and Japan. The six phthalate esters are dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), butylbenzyl phthalate (BBP), bis(2-ethylhexyl) phthalate (DEHP), and di-n-octyl phthalate (DnOP). The media addressed are water, sediment, soil, air, dust, food, wastewater, sewage sludges, and rainwater. The reported concentrations vary widely as a result of several factors including analytical error, sample contamination, and proximity to a variety of past and present sources. To gain an impression of the absolute levels and distributions, histograms are prepared with binning on a semi-decade logarithmic scale. Cumulative histograms are also prepared to convey an impression of cumulative distribution. To gain an appreciation of the relative concentrations in various media, fugacities are estimated and plotted, thus revealing the relative equilibrium status between media and any biomagnification. These plots suggest that phthalate esters are not persistent in the environment and do not biomagnify, as they are rapidly metabolized in organisms. Keywords. Phthalate ester, Concentrations, Fugacity, Persistence
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
1
Introduction
2
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
2.1 2.2 2.3
Database Generation . . . . . . . . . . . . . . . . . . . . . . . . . 128 Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Transforming Concentrations into Fugacities . . . . . . . . . . . . 130
3
Dimethyl Phthalate (DMP) . . . . . . . . . . . . . . . . . . . . . . 132
3.1 3.1.1 3.1.2 3.1.3 3.2 3.3 3.4 3.5
Water . . . . . Surface Water . Groundwater . Drinking Water Sediment . . . Soil . . . . . . Air . . . . . . . Dust . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
132 132 132 136 136 136 136 136
© Springer-Verlag Berlin Heidelberg 2003
126
K. Clark et al.
3.6 3.7 3.7.1 3.7.2
Food . . . . Other Media Wastewater Sludge . . .
. . . .
. . . .
. . . .
4
Diethyl Phthalate (DEP) . . . . . . . . . . . . . . . . . . . . . . . 137
4.1 4.1.1 4.1.2 4.1.3 4.2 4.3 4.4 4.5 4.6 4.7 4.7.1 4.7.2
Water . . . . . Surface Water . Groundwater . Drinking Water Sediment . . . Soil . . . . . . Air . . . . . . . Dust . . . . . . Food . . . . . . Other Media . . Wastewater . . Sludge . . . . .
5
Dibutyl Phthalate (DBP) . . . . . . . . . . . . . . . . . . . . . . . 142
5.1 5.1.1 5.1.2 5.1.3 5.2 5.3 5.4 5.5 5.6 5.7 5.7.1 5.7.2
Water . . . . . Surface Water . Groundwater . Drinking Water Sediment . . . Soil . . . . . . Air . . . . . . . Dust . . . . . . Food . . . . . . Other Media . . Wastewater . . Sludge . . . . .
6
Butylbenzyl Phthalate (BBP) . . . . . . . . . . . . . . . . . . . . . 149
6.1 6.1.1 6.1.2 6.1.3 6.2 6.3 6.4 6.5 6.6 6.7 6.7.1 6.7.2
Water . . . . . Surface Water . Groundwater . Drinking Water Sediment . . . Soil . . . . . . Air . . . . . . . Dust . . . . . . Food . . . . . . Other Media . . Wastewater . . Sludge . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
136 137 137 137
137 137 137 141 141 141 141 141 142 142 142 142
142 142 147 147 147 147 148 148 148 148 148 148
149 149 149 149 153 153 153 153 153 154 154 154
127
Observed Concentrations in the Environment
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
154 154 154 159 159 159 159 160 160 160 160 160 161 161
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
161 161 161 161 164 164 164 164 164 164 164 165
9 9.1 9.2 9.3
Discussion and Recommendations Histograms . . . . . . . . . . . . . Fugacities in Various Media . . . . Conclusions . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
165 165 172 175
10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
7 7.1 7.1.1 7.1.2 7.1.3 7.2 7.3 7.4 7.5 7.6 7.7 7.7.1 7.7.2 7.7.3
Bis(2-Ethylhexyl) Phthalate (DEHP) Water . . . . . . . . . . . . . . . . Surface Water . . . . . . . . . . . . Groundwater . . . . . . . . . . . . Drinking Water . . . . . . . . . . . Sediment . . . . . . . . . . . . . . Soil . . . . . . . . . . . . . . . . . Air . . . . . . . . . . . . . . . . . . Dust . . . . . . . . . . . . . . . . . Food . . . . . . . . . . . . . . . . . Other Media . . . . . . . . . . . . . Wastewater . . . . . . . . . . . . . Sludge . . . . . . . . . . . . . . . . Rainwater . . . . . . . . . . . . . .
8 8.1 8.1.1 8.1.2 8.2 8.3 8.4 8.5 8.6 8.7 8.7.1 8.7.2
Di-n-Octyl Phthalate (DnOP) Water . . . . . . . . . . . . Surface Water . . . . . . . . Drinking Water . . . . . . . Sediment . . . . . . . . . . Soil . . . . . . . . . . . . . Air . . . . . . . . . . . . . . Dust . . . . . . . . . . . . . Food . . . . . . . . . . . . . Other Media . . . . . . . . . Sludge . . . . . . . . . . . . Wastewater . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
Abbreviations BBP C DBP DEHP DEP DMP DnOP f fOM fD
Butylbenzyl phthalate Concentration (mol m–3) Dibutyl phthalate Bis(2-ethylhexyl) phthalate Diethyl phthalate Dimethyl phthalate Di-n-octyl phthalate Fugacity (Pa) Organic matter content of the aerosol Fraction of chemical in the dissolved phase of water
128 fOC H K KOW KOA KOC KP M R T TSP vSP Z r j
K. Clark et al.
Fraction of organic carbon in the suspended sediment Henry’s law constant (Pa m3 mol–1) Kelvin Octanol-water partition coefficient Octanol-air partition coefficient Organic carbon-water partition coefficient Gas-particle partition coefficient (m3 µg–1) Molar mass (g mol–1) Gas constant (Pa m3 mol–1 K–1) Absolute temperature (K) Total suspended particle concentration (µg m–3) Volume fraction of suspended particles in water Fugacity capacity of the phase (mol m–3 Pa–1) Density of the phase (kg m–3) Fraction of chemical associated with particles in air
1 Introduction Measured concentrations of six phthalate esters in seven environmental media are compiled and analyzed. The six phthalate esters are dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), butylbenzyl phthalate (BBP), bis(2-ethylhexyl) phthalate (DEHP), and di-n-octyl phthalate (DnOP). The media addressed are water, sediment, soil, air, dust, food, and “other media” including wastewater, sewage sludges, and rainwater. The data are predominantly from Europe, the United States, and Japan. The monitoring data have been collected for a variety of reasons and by different groups (e.g., by regulators to support development of regulations, by industry for compliance purposes, by researchers to support modeling efforts, etc.). Due to the varying interests of the organizations that have collected the data, there is variation in the proximity of the measurements to sources of the phthalate esters. Where possible, data locations are classified as rural or urban to assist in evaluation of the data.
2 Methodology 2.1 Database Generation
Numerous data sources have been reviewed to determine the ranges and distributions of phthalate esters in the environment. A compilation of data was produced by Exxon Mobil Biomedical Sciences Inc. (EMBSI) [1] for five phthalate esters: DEP, DMP, DBP, BBP, and DEHP. The data were categorized by EMBSI into regions including Canada, United States, Europe, and Japan/Asia. The data were also categorized by EMBSI in terms of data quality. The following ranking scheme was employed:
129
Observed Concentrations in the Environment
1. 2. 3. 4.
Reliable without restrictions Reliable with restrictions Not reliable Unassignable
The EMBSI database is used as the starting point for the database presented herein. Additional data, available as of August 31, 2000, have been added to this database. These data include recent measurements of phthalates in air, dust, fish, milk, and vegetation in the Netherlands [2, 3], phthalates in surface water and wastewater in Germany [4], DEHP and DBP in milk, breast milk, baby food, and dust [5], DBP in surface water and wastewater in Europe [6], DEHP in surface water, sediments, and sludge in Europe [7], phthalates in Canadian drinking water and surface water [8, 9] and sludge [10, 11], and data from Environment Canada and Health Canada (EC &HC) [12–15]. In addition to the five phthalate esters in the original EMBSI database, some information on DnOP concentrations in the environment is available in the Canadian Environmental Protection Act (CEPA) Priority Substances List assessment report [16, 17] as well as from Alberta Environment [8, 9]. Measured concentrations, particularly for phthalates other than DEHP, are often reported as “not detected”. As a result, treatment of the detection limit significantly influences the characteristics (i.e., mean and standard deviation) of the data. For the results presented herein, the non-detectable data have been set to one-half the detection limit. Tables and histograms that summarize the data are presented herein. The complete database, with references, is presented in a report to the American Chemistry Council [18]. 2.2 Histograms
It is difficult to visualize the distribution of the monitoring data by examining the raw data contained in the monitoring database or by examining the summary tables and statistics. Therefore, it was decided to display the monitoring data graphically using both frequency histograms and cumulative distribution plots derived from these histograms. These are termed “cumulative histograms”. Environmental concentrations in each medium are divided logarithmically into classes or bins, with a factor of approximately three between adjacent bins. For example, for surface water concentrations, the following bins are allocated: Bin
Range (µg L–1)
Midpoint of range (µg L–1)
1 2 3 4 5 6 7 8
0.01–0.03 0.03–0.1 0.1–0.3 0.3–1.0 1.0–3.0 3.0–10 10– 30 30– 100
0.017 0.055 0.17 0.55 1.7 5.5 17 55
130
K. Clark et al.
The bins allocated can be adjusted to fit the range of environmental concentrations in the database. The frequency histograms are devised by recording the number of study averages in the monitoring database that fall within each bin.A weighted frequency histogram is devised by adding up the number of samples that contribute to each study average within a particular bin. For example, if there are three study averages, within a bin ranging from 0.1 to 0.3 µg L–1, of 0.12 µg L–1 (n=10), 0.20 µg L–1 (n=1), and 0.25 µg L–1 (n=5), then this bin will have a frequency of 3 in the unweighted histogram and a frequency of 16 in the weighted histogram. The cumulative histograms are devised by calculating the cumulative percentage contribution of each successive bin to the total number of study averages in the case of the unweighted cumulative histogram, or to the total number of samples in the case of the weighted cumulative histogram. Weighted and unweighted frequency and cumulative histograms are plotted for DEHP, surface water, sediment, and air concentrations in Europe and are discussed in Section 9.0. Insufficient data are available to produce a histogram for concentrations in soil. 2.3 Transforming Concentrations into Fugacities
The study of both aquatic and terrestrial ecosystems has shown that one useful approach for studying food chain bioaccumulation is through transforming environmental concentrations into fugacities [19]. Fugacities offer the advantage of being able to use a single currency to compare levels of contamination in different environmental media and organisms. The fugacity f (Pa) of a compound in a particular phase can be calculated from the concentration C (mol m–3) by using the following equation: f = C/Z
(1)
where Z is the fugacity capacity of the phase for the compound (mol m–3 Pa–1). Hence, if the fugacity capacities are known, then the fugacities can be calculated from measured concentrations on a volume basis. It is possible to derive concentrations in units of mol m–3 by using the molecular mass M (g mol–1) and density of the phase r (kg m–3). The fugacity capacities or Z values for the phases can be estimated by using methods outlined by Mackay [20]. In this report, concentrations of phthalates in Europe are transformed into their corresponding fugacities for the media of air, surface water, sewage sludge, vegetation, soil, sediments, fish, cow’s milk, and human milk.We use only concentration data from the recent RIVM/ECPI monitoring campaign [2, 3] for calculating fugacities, except for surface water concentrations [4] and human milk concentrations [5, 21]. This high quality subset of multimedia concentrations is selected to minimize data variability. Unfortunately, it is only possible to calculate the fugacities of DEHP and DBP with this limited data set. The equations used to estimate the Z values for these phases are briefly described below. Physical-chemical data used in the calculations such as KOW , H, and KOA are taken from Cousins and Mackay [22]. Environmental parameters used in the calculations are listed in Table 1.
131
Observed Concentrations in the Environment Table 1. Environmental parameters assumed for calculating media-specific fugacities
Parameter
Value
Total suspended particle concentration in air (µg m–3) Fraction of organic matter in air particles Volume fraction of suspended particles in water Fraction of soil organic carbon Fraction of sediment organic carbon Gas constant (Pa m3 mol–1 K–1) Environmental temperature (K) Temperature of cow’s milk (K) Temperature of human milk (K) Fraction of plant lipid Fraction of lipid in fish Fraction of lipid in cow’s milk Fraction of lipid in human milk
80 0.20 0.000015 0.02 0.05 8.314 273 310 310 0.01 0.05 0.033 0.04
The fugacity capacity of pure air ZA is given by ZA = 1/(RT)
(2)
where R is the gas constant (Pa m3 mol–1 K–1) and T is the absolute temperature (K).Air concentrations given in the database are total air concentrations (the sum of the amount in gaseous and particle phases). To calculate the fraction on the particles (j) and from this the gaseous phase concentration ((1– j) multiplied by the total air concentration), the following equation is used:
j = KP (TSP)/[1+KP (TSP)]
(3) (m3
µg–1)
where KP is the gas-particle partition coefficient and TSP is the total suspended particle concentration (µg m–3). KP is estimated from KP = 1.23 ¥ 10–12 fOM KOA
(4)
where fOM is the organic matter content of the aerosol and KOA is the octanol-air partition coefficient. The fugacity capacity of pure water ZW is given by ZW = 1/H
(5)
where H is the Henry’s law constant (Pa m3 mol–1). Surface water concentrations given in the database are total water concentrations (the sum of the amount in dissolved and suspended particle phases). To calculate the dissolved water concentration the following equation is used: fD = 1/(1 + fOC nSP KOC)
(6)
where fD is the fraction in the dissolved phase, fOC is the fraction of organic carbon in the suspended sediment, vSP is the volume fraction of suspended particles, and KOC is the organic carbon/water partition coefficient; it is assumed that KOC is equal to KOW .
132
K. Clark et al.
Concentrations in soil, sediments, cow’s milk, human milk, fish, and sewage sludge are all reported on a mass per lipid or mass per organic carbon basis. These concentrations are converted to units of mol m–3 (assuming a density of 1000 kg m–3 for all phases). Concentrations in plants are not given on a lipid basis, thus they are first converted to units on a mass per lipid basis by assuming the volume fraction of lipid in the plant is 0.01. The fugacity capacity of the lipid/organic carbon phase (ZO) in each medium is calculated as ZO = KOW ZW
(7)
This assumes that KOW is equivalent to the lipid-water and organic carbon-water partition coefficients. The fugacities of the phthalates in each medium are then calculated by using Eq. (1).
3 Dimethyl Phthalate (DMP) A summary of the reported concentrations of DMP is presented in Table 2. 3.1 Water 3.1.1 Surface Water
The overall mean concentration of DMP in surface water in Canada (1.40 µg L–1) is significantly higher than that calculated for the United States (0.0017 µg L–1) and Europe (0.034 µg L–1). As indicated in Table 2, the concentrations measured in rural and urban regions in Canada do not differ significantly. The maximum measured concentration in the United States surface water (0.003 µg L–1) is much less than the maximum measured concentration in Canada (33 µg L–1). The maximum concentration for water with substantial industrial sources in Canada (9 µg L–1) is slightly lower than the overall maximum concentration; however, the mean concentration for water with substantial industrial sources is 2.76 µg L–1, which significantly affects the overall mean concentration. No data are reported for Japan/Asia. 3.1.2 Groundwater
A maximum concentration of 355 µg L–1 is reported for the United States as a mean value; however, the median (27 µg L–1) is much lower than the mean. Only one reference is available for Europe, which indicates concentrations less than 0.1 µg L–1. Canadian data are available in a drinking water database reported below. Although some of these data represent Canadian groundwater, the data does not differentiate between surface water supplies and groundwater supplies.
Medium
Mean
Canada USA Europe Japan/Asia
Canada rural urban USA Europe Japan/Asia
USA Europe Japan/Asia
Canada
Sediments (mg kg NA 3.6 0.011 NA
–1)
Drinking water (µg L 0.5 0.5 0.5 NA NA NA
27 NA NA
NA 0.004 0.0001 NA
0.5 0.5 0.5 NA <0.1 NA
–1)
NA <0.1 NA
Groundwater (µg L–1) NA NA
0.05 0.05 0.05 0.1 0.001 <0.001 NA
Minimum
Concentration
Surface water (µg L–1) Canada 1.4 rural 0.8 urban 0.86 industrial sources 2.76 USA 0.0017 Europe 0.034 Japan/Asia NA
Region
NA 590 0.2 NA
1 1 1 NA <0.1 NA
NA <0.1 NA
NA
33 33 13 9 0.003 2 NA
Maximum
NA 0.17 0.0035 NA
NA NA NA NA NA NA
NA NA NA
NA
0.15 NAc NA NA 0.0013 0.009 NA
10th Percentile a
Table 2. Summary of concentrations for dimethyl phthalate
NA 1.6 0.0035 NA
0.5 0.5 0.5 NA NA NA
NA NA NA
NA
0.35 0.5 0.5 2.6 0.0017 0.009 NA
Median a
NA 11.8 0.045 NA
NA NA NA NA NA NA
NA NA NA
NA
1.2 NA NA NA 0.002 0.055 NA
90th Percentile a
1651 51
1273 637 636
2
1962 1026 184 625 509 138
no data references
no data references
no data references
data primarily from Alberta Environment [8] median=0.5; represents 1/2 the detection limit Median=0.5; represents 1/2 the detection limit no data references
no data references
Canadian data represented in drinking water summary
no data references
data primarily from Alberta Environment [9] median=0.5; represents 1/2 the detection limit median=0.5; represents 1/2 the detection limit
No. Data Comments Points b
Observed Concentrations in the Environment
133
Dust (indoors) (µg kg NA NA 1730 NA
Other media Wastewater (µg L–1) 2.76 13.7 1.5 NA
Canada USA Europe Japan/Asia
0.1 13.7 <0.02 NA
) NA NA <200 NA
<1 <1 NA
2.55 20.2 NA
–1
NA NA
NA NA NA NA
NA NA
Canada USA Europe Japan/Asia
Air (ng m–3)
Canada USA Europe Japan/Asia
Canada USA Europe outdoor indoor Japan/Asia
Soil (µg kg–1) NA NA NA NA
Mean
Minimum
Concentration
Medium
Region
Table 2 (continued)
9 207 13 NA
NA NA 5000 NA
23 129 NA
NA NA
NA NA NA NA
Maximum
NA NA 0.03 NA
NA NA 340 NA
0.5 2.1 NA
NA NA
NA NA NA NA
10th Percentile a
2.6 NA 0.14 NA
NA NA 500 NA
1 10 NA
NA NA
NA NA NA NA
Median a
NA NA 2.1 NA
NA NA 5000 NA
5 39 NA
NA NA
NA NA NA NA
90th Percentile a
625 1 76
7
30 26
no data references
includes sewage, slurry, effluent, and influent
no data references no data references detected in 2 of 7 samples no data references
no data references
no data references no data references
no data references no data references no data references no data references
No. Data Comments Points b
134 K. Clark et al.
c
b
a
NA NA NA NA NA NA NA 0.003 NA NA NA NA NA NA NA NA NA
NA NA NA NA 0.002 NA NA
NA
NA NA NA
NA NA
NA
130 NA 19,500 NA
NA NA NA
<10 NA 20 NA
Maximum
NA
NA NA
NA NA NA
NA
NA NA NA NA 0.002 NA NA
NA NA NA
NA NA NA NA
10th Percentile a
NA
0.005 NA
NA NA NA
NA
NA NA 0.005 NA 0.003 NA 0.00125
NA NA NA
NA NA NA NA
Median a
NA
NA
NA NA NA
NA
NA NA NA NA 0.003 NA NA
NA NA NA
NA NA NA NA
90th Percentile a
48
29
3
25
72 2 38
no data references
mean represents 1/2 the detection limit no data references
no data references no data references no data references
no data references
no data references no data references mean represents 1/2 the detection limit no data references data categorized as not reliable no data references mean represents one half detection
no data references no data references no data references
no data references
primary sewage sludge
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual datapoints. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available.
Food (µg g ) Beverages NA Cereal NA Dairy NA (excl. milk) Eggs NA Fat & oils NA Fish 0.005 Fruits NA Grains 0.003 Meat NA Milk 0.00125 limit Nuts & NA beans Other foods NA Poultry NA Processed NA meat Vegetables 0.005 Infant NA formulapowder Breast milk NA
–1
Sludge (µg kg–1) 30 39,800 863 NA
Canada USA Europe Japan/Asia
Mean
Minimum
Concentration
Medium
Region
Table 2 (continued) Observed Concentrations in the Environment
135
136
K. Clark et al.
3.1.3 Drinking Water
The calculated mean concentration of DMP in Canadian drinking water is equal to one-half the detection limit (0.5 µg L–1). One reference for Europe reports concentrations less than the detection limit (0.1 µg L–1). 3.2 Sediment
The mean concentration of DMP in sediment is highest in the United States (3.6 mg kg–1) and is two orders of magnitude lower for Europe (0.011 mg kg–1). The maximum concentration reported in the United States is 590 mg kg–1. No data references are reported for Canada or Japan/Asia. 3.3 Soil
No data are available for DMP measured in soil. 3.4 Air
Concentrations measured in air are only identified for Europe. The calculated mean concentration in indoor air is 20.2 ng m–3, with the data ranging from <1 ng m–3 to a maximum of 129 ng m–3. The calculated mean concentration in outdoor air is 2.55 ng m–3, with the data ranging from <1 ng m–3 to 23 ng m–3. 3.5 Dust
Measurements of DMP in dust are only available for Europe, where concentrations range between <200 µg kg–1 and 5000 µg kg–1. The calculated mean concentration is 1730 µg kg–1. 3.6 Food
Measurements of DMP concentrations in food are very limited. There are not enough data to prepare a summary by region, nor to complete each of the food categories reported for the other phthalates. DMP has not been detected in fish, milk, or vegetables.
Observed Concentrations in the Environment
137
3.7 Other Media 3.7.1 Wastewater
The calculated mean concentrations for Canada and Europe are comparable (2.76 µg L–1 and 1.5 µg L–1, respectively).A maximum concentration of 207 µg L–1 was reported for the United States, which is an order of magnitude higher than that of Canada or Europe. The calculated mean concentration for the United States is 13.7 µg L–1. 3.7.2 Sludge
Only one reference is available for DMP measured in sludge in the United States. This reference presents a mean concentration of 39,800 µg kg–1 with no reported range. In a study of 72 samples of sewage sludge from 12 locations across Canada, seven of the locations report non-detectable results [11]. The measured concentrations of DMP in sludge range between the detection limit (assumed equal to 10 µg kg–1) and 130 µg kg–1 with an overall calculated mean of 30 µg kg–1. Sludge data for Europe range between 20 µg kg–1 and 19 500 µg kg–1 with a calculated mean concentration of 863 µg kg–1.
4 Diethyl Phthalate (DEP) A summary of the reported concentrations of DEP is presented in Table 3. 4.1 Water 4.1.1 Surface Water
The overall mean concentration of DEP in surface water in Canada (1.42 µg L–1) is comparable with that calculated for the United States (1.33 µg L–1).As indicated in Table 3, the calculated mean concentration for Canada is influenced by the data for surface water with substantial industrial sources. The mean concentrations for rural and urban Canada are less than one-half of the calculated mean for the United States. The European and Japan/Asian mean concentrations (0.087 µg L–1 and 0.1 µg L–1, respectively) are lower than the North American data. 4.1.2 Groundwater
Only two references are available for measurements of DEP in groundwater in the United States. A maximum concentration of 147 µg L–1 was reported for the
Medium
Mean
0.084 0.084 0.42 0.01 0.05 NA
NA <0.002 <0.007 NA
Sediments (mg kg–1) NA 3.35 0.0075 NA
Canada USA Europe Japan/Asia
–1)
0.1 NA
Drinking water (µg L 0.52 0.53 0.52 6.57 0.05 NA
NA NA
Europe Japan/Asia
0.87
Canada rural urban USA Europe Japan/Asia
143
Groundwater (µg L ) NA NA
USA
Canada
–1
0.05 0.05 0.05 0.05 0.01 <0.008 NA
Minimum
Concentration
Surface water (µg L–1) Canada 1.42 rural 0.56 urban 0.5 industrial sources 3.04 USA 1.33 Europe 0.087 Japan/Asia 0.1
Region
NA 590 2 NA
3 3 3 10 0.1 NA
0.2 NA
147
NA
55 5 2.5 55 55 4 NA
Maximum
NA 0.03 0.0035 NA
NA NA NA 0.028 NA NA
NA NA
NA
NA
NAc NA NA NA 0.013 0.009 NA
10th Percentile a
Table 3. Summary of concentrations for diethyl phthalate
NA 0.07 0.004 NA
NA 0.5 0.5 0.1 NA NA
NA NA
NA
NA
NA 0.5 0.5 3.5 0.019 0.03 NA
Median a
NA 8.3 0.01 NA
NA NA NA 5.4 NA NA
NA NA
NA
NA
NA NA NA NA 5.7 0.14 NA
90th Percentile a
1674 46
1150 587 563 67 1
35
1742 963 171 608 657 321 5
no data references
no data references
average based on 1/2 the detection limit no data references
data primarily from Alberta Environment [8] median=0.5; represents 1/2 detection limit median =0.5; represents 1/2 the detection limit
no data references
Canadian data represented by drinking water summary 1 reference with 34 data points (median=11; mean=147)
average based on 1/2 the detection limit
data primarily from Alberta Environment [9] median =0.5; represents 1/2 the detection limit median =0.5; represents 1/2 the detection limit median =3.5
No. Data Comments Points b
138 K. Clark et al.
Other media Wastewater (µg L–1) 3.04 6.24 4.73 NA
Canada USA Europe Japan/Asia
0.05 0.01 <0.02 <0.2
) NA NA 900 NA
<1 25 NA
38.6 621 NA
Dust (indoors) (µg kg NA NA 2700 NA
0.2 NA
NA NA
–1
NA
NA NA NA NA
NA
Canada USA Europe Japan/Asia
Air (ng m–3)
Canada USA Europe Japan/Asia
Canada USA outdoor indoor Europe outdoor indoor Japan/Asia
Soil (µg kg–1) NA 7.50¥104 NA NA
Mean
Minimum
Concentration
Medium
Region
Table 3 (continued)
55 1220 76.9 3.7
NA NA 7000 NA
242 3234 NA
55 NA
NA
NA NA NA NA
Maximum
NA NA 0.03 NA
NA NA 500 NA
0.55 119 NA
NA NA
NA
NA NA NA NA
10th Percentile a
NA NA 0.21 NA
NA NA 3000 NA
11.5 171 NA
NA 340
NA
NA 620 NA NA
Median a
NA NA 9.1 NA
NA NA 5200 NA
76 333 NA
NA NA
NA
NA NA NA NA
90th Percentile a
608 14 51
7
32 26
37
includes sewage, slurry, effluent, and influent
no data references no data references detected in 5 of 7 samples no data references
no data references
no data references
no data references no data references
no data references
No. Data Comments Points b
Observed Concentrations in the Environment
139
c
b
a
<0.05 0.04 NA
NA NA <0.001 <0.04 NA NA <0.0025 NA
<0.01 NA NA
<0.01 NA
NA
0.027 0.11 0.05
0.05 0.25 0.059 0.076 0.05 0.05 0.002 0.045
0.58 0.05 0.05
0.005 NA
NA
<20 NA <0.02 NA
NA
0.026 NA
5.3 NA NA
NA NA 0.5 0.73 NA NA <0.01 NA
0.09 0.19 NA
5000 NA 2900 NA
Maximum
NA
NA NA
0.005 NA NA
NA NA 0.007 NA NA NA 0.00275 NA
0.025 0.043 NA
NA NA 0.01 NA
10th Percentile a
NA
NA NA
0.05 NA NA
NA NA 0.05 NA NA NA 0.005 NA
0.025 0.11 NA
NA NA 119 NA
Median a
NA
NA NA
1.5 NA NA
NA NA 0.5 NA NA NA 0.005 NA
0.038 0.18 NA
NA NA 2050 NA
90th Percentile a
48
19 2 15
1 3 63 13 9 9 33 3
40 4 6
77 1 18
no data references
no data references
detected in 5 of 19 samples mean represents 1/2 the detection limit mean represents one half detection limit
detected in 2 of 13 samples detected in 1 of 9 samples mean represents 1/2 the detection limit mean represents 1/2 the detection limit mean represents 1/2 the detection limi
mean represents one half detection limit mean represents 1/2 the detection limit
not separated into geographical locations detected in 1 of 40 samples detected in 3 of 4 samples mean represents 1/2 the detection limit
no data references
primary sewage sludge
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual datapoints. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available.
Food (µg ) Beverages Cereal Dairy (excl. milk) Eggs Fat & oils Fish Fruits Grains Meat Milk Nuts & beanst Other Foods Poultry Processed meat Vegetables Infant formulapowder Breast milk
g–1
Sludge (µg kg–1) 337 290 507 NA
Canada USA Europe Japan/Asia
Mean
Minimum
Concentration
Medium
Region
Table 3 (continued)
140 K. Clark et al.
Observed Concentrations in the Environment
141
United States. Only one reference is available for Europe and the reported concentrations range between 0.1 µg L–1 and 0.2 µg L–1. Canadian data are available in a drinking water database reported below. Although some of these data represent Canadian groundwater, the data does not differentiate between surface water supplies and groundwater supplies. 4.1.3 Drinking Water
The highest calculated mean concentration is for the United States (6.57 µg L–1). The calculated mean concentrations did not vary significantly between urban and rural areas in Canada (0.52 µg L–1 and 0.53 µg L–1, respectively). These mean values are comparable to the European maximum of 0.1 µg L–1. No data are available for Japan/Asia. 4.2 Sediment
References for DEP reported in sediments are only available for the United States and Europe. The mean concentration for DEP in sediment is highest in the United States (3.35 mg kg–1) and is over two orders of magnitude lower in Europe (0.0075 mg kg–1). The maximum reported concentration was in the United States (590 mg kg–1). 4.3 Soil
Only one reference is available reporting measurements of DEP in soil. This reference reports 37 samples with a mean concentration of 75,000 µg kg–1 and a median of 620 µg kg–1. The range of the data is not reported. 4.4 Air
Measured concentrations indoors are higher than concentrations outdoors. The maximum mean indoor air concentration is for Europe (621 ng m–3), compared with a median concentration of 340 ng m–3 in the United States. The European data indicated outdoor concentrations ranging between <1 ng m–3 and 242 ng m–3 with a calculated mean of 38.6 ng m–3. 4.5 Dust
Measurements of DEP in dust are only available for Europe. Reported concentrations range between 900 µg kg–1 and 7000 µg kg–1. The calculated mean concentration is 2700 µg kg–1.
142
K. Clark et al.
4.6 Food
Several references present DEP concentrations measured in food; however, there are insufficient food data to prepare a summary by region. Concentrations measured in the available food groups reveal the highest maximum concentration measured in other foods (5.3 µg g–1 in miscellaneous packaged food). DEP was not detected in dairy products, eggs, fats and oils, meats, milk, nuts and beans, poultry, or processed meat. No measurements of DEP in infant formula or breast milk are available. 4.7 Other Media 4.7.1 Wastewater
Canada, Europe, and the United States have comparable calculated mean concentrations (3.04 µg L–1, 4.73 µg L–1 and 6.24 µg L–1, respectively). A maximum concentration of 1220 µg L–1 was reported for the United States, which is more than an order of magnitude higher than that of Canada and Europe. The lowest maximum concentration reported for Japan/Asia was 3.7 µg L–1. 4.7.2 Sludge
The measured concentrations of DEP in sludge range between a trace amount and 5000 µg kg–1, with an overall calculated mean of 337 µg kg–1. One reference for the United States reports a mean concentration of 290 µg kg–1. The European data result in a calculated mean concentration of 507 µg kg–1.
5 Dibutyl Phthalate (DBP) A summary of the reported concentrations of DBP is presented in Table 4. 5.1 Water 5.1.1 Surface Water
The calculated mean concentration of DBP in surface water for Canada (1.42 µg L–1) is slightly higher than that calculated for the other regions. The maximum measured concentration was 350 µg L–1 (for Japan/Asia). The maximum for Canadian waters with substantial industrial sources (64 µg L–1) is comparable with the United States maximum (63 µg L–1).
rural urban USA Europe Japan/Asia
Canada
USA Europe Japan/Asia
Canada
2.9 0.41 0.24 0.79
0.68 0.45
0.01 0.0007 0.002 0.02
0.01 0.0003
–1)
0.5 0.12 0.5
0.73 0.85 4.6 NA 0.9
0.14 0.18 0.01 NA <1.0
Drinking water (µg L 0.8 0.14
4.5 0.25 2.14
Groundwater (µg L–1) NA NA
industrial sources USA Europe Japan/Asia
rural urban
Canada
Minimum
Surface water (µg L–1) 1.42 0.0003
Mean
Concentration
Medium
Region
8 7 50 NA 3.1
8
50 0.46 12
NA
64 63 170 350
5.2 6.98
100
Maximum
Table 4. Summary of concentrations of dibutyl phthalate
NA NA 0.03 NA 0.5
NA
NA NA 0.5
NA
0.5 0.5 0.1 NA 0.5
NA
4.74 NA 0.5
NA
1.0 0.014 0.11 0.5
0.5 0.5
NAc NA NA 0.0019 0.009 0.5
0.09
Median a
0.01
10th Percentile a
NA NA 5.8 NA 0.5
NA
NA NA 0.5
NA
NA 1.0 0.7 0.5
NA NA
1.4
90th Percentile a
36
354 351 1111
705
1044 1 21
580 1180 234 119
874 218
1672
no data references
data primarily from Alberta Environment [8] measurements in drinking water median=0.5; represents 1/2 detection limit median=0.5; represents 1/2 the detection limit
median=0.5; represents 1/2 the detection limit
Canadian data represented by drinking water summary
data primarily from Alberta Environment [9], EC & HC [14] median=0.5; represents 1/2 the detection limit median=0.5 for Alberta dataset (165 values in total); represents 1/2 the detection limit median=1.0
No. Data Comments Points b
Observed Concentrations in the Environment
143
Canada
0.1 NA
0.08 0.2
0.2 <3
0.0021 NA
2.5 0.2
15.8 1032
96 0.0009
0.01 NA
1.5
4.9 2.9
184 NA
Europe Japan/Asia
Canada outdoor indoor USA outdoor indoor Europe outdoor indoor Japan/Asia outdoor indoor
24
USA
Air (ng m
0.027
Soil (µg kg–1) NA
–3)
0.01 0.00007 0.0001 <0.005
Minimum
Sediments (mg kg–1) NA 14.6 0.22 0.12
Mean
Canada USA Europe Japan/Asia
Concentration
Medium
Region
Table 4 (continued)
370 NA
380 9445
92 420
700 NA
560 NA
280
1.52
10.4 3333 28.3 2.3
Maximum
NA NA
3.2 156
0.33 NA
2.5 NA
NA NA
NA
NA
NA 0.0076 0.0035 0.03
10th Percentile a
NA NA
8 551
2.6 NA
4.3 NA
NA NA
NA
NA
NA 0.12 0.016 0.08
Median a
NA NA
86 5900
17 NA
5.7 NA
NA NA
NA
NA
NA 18 0.25 0.28
90th Percentile a
2 1
83 26
38 6
44 9
13
1
3073 94 7
no data references
3 references with no data points; unable to calculate average number of data points unknown; average reported from reference
2 references; unable to calculate average
No. Data Comments Points b
144 K. Clark et al.
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
Food (µg g–1) Beverages 0.1 Cereal 0.3 Dairy 0.04 (excl. milk) Eggs 0.09 Fat & oils 2.5
Other media Wastewater (µg L–1) 2.8 4.7 22 0.1 Sludge (µg kg–1) 1.29¥104 5.60¥104 2.10¥104 NA Rainwater (µg L–1) NA NA 0.16 NA
10th Percentile a
NA 0.03 0.07 NA
0.56 0.5 <0.1 0.1 11
0.05 <0.5
NA NA 4.5 NA
0.059 0.48
0.025 0.1 0.04
NA NA 0.20 NA
1.61¥105 1.68¥104 3.20¥106 493 4.30¥105 0.36 NA NA
100 2265 1000 250
NA NA NA NA 6.78¥105 2500 NA NA
Maximum
0.025 0.1 <0.01
NA NA 0.03 NA
200 20 0.14 NA
0.01 0.004 <0.02 0.1
Minimum
Canada USA Europe Japan/Asia
Mean
Dust (indoors) (µg kg–1) NA NA NA NA 5.67¥104 5 NA NA
Concentration
Medium
Region
Table 4 (continued)
0.089 1.8
0.025 0.3 0.05
NA NA 0.29 NA
5.67¥104 2400 2480 NA
NA 0.08 0.27 NA
NA NA 5.6¥104 NA
Median a
0.099 6.5
0.325 0.5 0.05
NA NA 1.0 NA
5.69¥104 1.74¥105 2.82¥104 NA
NA 7.1 36 NA
NA NA 1.6¥105 NA
90th Percentile a
11 28
12 5 12
76
82 7 70
599 39 70 10
26
not separated into geographical locations detected in 4 of 12 samples
no data references
no data references no data references
no data references
primary sewage sludge
includes sewage, slurry, effluent, and influent
no data references
no data references no data references
No. Data Comments Points b
Observed Concentrations in the Environment
145
c
b
a
Food (µg g–1) Fish Fruits Grains Meat Milk Nuts & beans Other foods Poultry Processed meat Vegetables Infant formulapowder Infant formulaliquid Breast Milk Baby Food
Medium
<0.001 0.01 0.03 0.03 0.003 <0.09
<0.01 <0.1 <0.1
<0.01 <0.05
<0.003
0.01 <0.01
0.16 0.13 0.54
0.17 0.07
0.0029
0.03 0.028
Minimum
0.23 0.033 0.26 0.92 0.012 0.18
Mean
Concentration
0.08 0.03
0.007
10 0.4
62 0.2 6.2
35 0.16 1.9 7.3 0.2 0.57
Maximum
NA NA
NA
0.02 0.025
0.0085 0.05 NA
0.0005 NA 0.046 0.05 0.0036 0.045
10th Percentile a
NA NA
NA
0.08 0.065
0.09 0.05 NA
0.0045 NA 0.05 0.05 0.015 0.045
Median a
NA NA
NA
8.0 0.23
0.61 0.17 NA
0.34 NA 0.46 1.3 0.045 0.23
90th Percentile a
10 12
4
67 53
38 4 13
43 14 9 9 50 6
detected in 1 of 4 samples
detected in 3 of 6 samples
detected in 3 of 14 samples detected in 4 of 9 samples
not separated into geographical locations
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual data points. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available.
Region
Table 4 (continued)
146 K. Clark et al.
Observed Concentrations in the Environment
147
5.1.2 Groundwater
Groundwater data are primarily available for the United States and Japan/Asia. A maximum concentration of 50 µg L–1 is reported for the United States. The highest mean concentration is also calculated for the United States (4.5 µg L–1), which is approximately twice the mean for Japan/Asia (2.14 µg L–1). Only one reference for Europe reported a mean concentration (0.25 µg L–1). In addition, the maximum concentration for Europe (0.46 µg L–1) is considerably lower than the maximum concentrations for the United States and Japan/Asia. Canadian data are available in a drinking water database reported below.Although some of these data represent Canadian groundwater, the data do not differentiate between surface water supplies and groundwater supplies. 5.1.3 Drinking Water
The drinking water data reported for the United States results in a calculated mean concentration (4.6 µg L–1) higher than the mean concentrations calculated for Canada (0.8 µg L–1) and Japan/Asia (0.9 µg L–1). The maximum concentration was reported for the United States (50 µg L–1). No data are available for Europe. 5.2 Sediment
The mean concentration of DBP in sediment is highest in the United States (14.6 mg kg–1). The United States data represents the largest data set (3073 reported samples). The mean concentration is two orders of magnitude lower in Europe (0.22 mg kg–1) and Japan/Asia (0.12 mg kg–1), although there are significantly less data references. Two data references are reported for Canada [15], however, a mean concentration is not reported. The Canadian data range between 0.01 mg kg–1 and 10.4 mg kg–1, which are comparable with the European and Japan/Asian data. The maximum concentration was reported for the United States (3333 mg kg–1). 5.3 Soil
Few references report measured concentrations of DBP in soil. An average concentration in soil of 24 µg kg–1 was reported for the United States, which is lower than the European mean of 184 µg kg–1. The maximum measured concentration was recorded in Europe (560 µg kg–1). Maximum concentrations are comparable for the United States and Europe, while the Canadian maximum (1.52 µg kg–1) is considerably lower.
148
K. Clark et al.
5.4 Air
Concentrations measured in air are reported to be the highest in Europe. The calculated mean indoor air concentration in Europe is 1032 ng m–3, while the maximum is 9445 ng m–3. The calculated means in the other regions are lower, although they are based upon less data. The maximum mean concentration of DBP in outdoor air is 96 ng m–3 (in Japan/Asia). The calculated mean concentration in Europe is 15.8 ng m–3, while the mean concentrations in the United States and Canada are 2.5 ng m–3 and 4.9 ng m–3, respectively. 5.5 Dust
References reporting measured concentrations in dust are only available for Europe. The calculated mean concentration is 56,700 µg kg–1, with data ranging between 5 µg kg–1 and 6.78 ¥105 µg kg–1. 5.6 Food
Several references present DBP concentrations measured in food; however, there are not sufficient food data to prepare a summary by region. Concentrations measured in the available food groups reveal the highest maximum concentration measured in “other foods” (62 µg g–1, measured in gravy and Parmesan cheese) and fish (maximum of 35 µg g–1). The highest mean concentration is for fats and oils (2.5 µg g–1). 5.7 Other Media 5.7.1 Wastewater
The maximum mean concentration is calculated for Europe (22 µg L–1), although this mean is comparable with the mean concentrations calculated for the other regions. The lowest calculated mean is for Japan/Asia (0.1 µg L–1). The Canadian maximum concentration of 100 µg L–1 is the same order of magnitude as the Japan/Asian maximum of 250 µg L–1. The highest reported maximum concentration was measured in the United States (2265 µg L–1). 5.7.2 Sludge
The calculated mean concentrations of DBP in sludge are in the same order of magnitude (104 µg kg–1) for the three regions referencing data (Canada, United
Observed Concentrations in the Environment
149
States, and Europe). The highest concentration is reported for the United States (3.2 ¥106 µg kg–1). No data are referenced for Japan/Asia.
6 Butylbenzyl Phthalate (BBP) A summary of the reported concentrations of BBP is presented in Table 5. 6.1 Water 6.1.1 Surface Water
The overall calculated mean concentration of BBP in surface water in Canada (1.51 µg L–1) is higher than that calculated for the United States (0.38 µg L–1). The majority of the data points for the United States are non-detectable. Similarly, the median concentrations for Canadian rural and urban waters are both 0.5 µg L–1, which represent one half of the detection limit. The maximum concentration (66 µg L–1) for the United States is comparable with the Canadian maximum for rural and surface waters with substantial industrial input (50 µg L–1 and 84 µg L–1, respectively). The maximum measured concentration in urban Canadian surface water (6 µg L–1) is comparable with the maximum measured concentration in Europe (13.9 µg L–1). Only one data reference is available for Japan/Asia. 6.1.2 Groundwater
Very little data have been identified for BBP in groundwater. A maximum concentration of 38 µg L–1 is reported in one of two references for the United States. These references did not report an average concentration. Canadian data are available in a drinking water database reported below. Although some of these data represent Canadian groundwater, the data do not differentiate between surface water supplies and groundwater supplies. 6.1.3 Drinking Water
The maximum measured concentration in the United States is 38 µg L–1. Analysis of the Canadian data is primarily dependent upon treatment of the detection limit, as most measurements were less than the detection limit. The maximum measured value is 2.8 µg L–1. One reference for the Netherlands reports non-detectable concentrations (<0.1 µg L–1).
Medium
Mean
NA 0.012 0.0025 NA
Sediments (mg kg–1) NA 0.19 0.059 NA
Canada rural urban USA Europe Japan/Asia
Canada USA Europe Japan/Asia
370 5.5 18.2 0.02
NA NA NA NA NA NA
Drinking water (µg L–1) 0.5 <0.00002 2.8 0.5 0.5 1 0.5 0.5 1 NA NA 38 <0.1 <0.1 <0.1 NA NA NA
<100 0.005 <0.007 ND
NA NA NA
38 0.24 NA
NA NA NA
NDd 0.04 NA
USA Europe Japan/Asia
Canada
NA
NAc NA NA NA 0.002 0.009 NA
10th Percentile a
NA
84 50 6 84 66 13.9 1
Maximum
Groundwater (µg L–1) NA NA
0.02 0.05 0.05 0.02 0.001 <0.01 <0.2
Minimum
Concentration
Surface water (µg L–1) Canada 1.51 rural 0.61 urban 0.55 industrial sources 3.05 USA 0.38 Europe 0.06 Japan/Asia NA
Region
Table 5. Summary of concentrations of butylbenzyl phthalate
NA 0.099 0.007 NA
NA 0.5 0.5 NA NA NA
NA NA NA
NA
2.47 0.5 0.5 1.0 0.003 0.009 NA
Median a
NA 0.27 0.08 NA
NA NA NA NA NA NA
NA NA NA
NA
NA NA NA NA 0.59 0.21 NA
90th Percentile a
1485 86
1239 619 620
1884 1010 183 629 2745 85
detection limit not provided
data primarily from Alberta Environment [8] median =0.5; represents 1/2 the detection limit median =0.5; represents 1/2 the detection limit
no data references
Canadian data represented by drinking water summary below detection limit not provided
data primarily from Alberta Environment [9] median =0.5; represents 1/2 the detection limit median=0.5; represents 1/2 detection limit median =1.0
No. Data Comments Points b
150 K. Clark et al.
Canada USA Europe Japan/Asia
84 449 30 1.5
<1 <3 NA
1.7 35 NA
Dust (indoors) (µg kg–1) NA NA NA NA 3.33 ¥105 <1000 NA NA
Other media Wastewater (µg L–1) 3.05 299 0.76 0.1
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
0.02 0.1 <0.06 <0.2
NA NA 1.7¥106 NA
1 NA
9.5 NA <10 465 NA
20 140
0.38 NA
1.78 NA
NA NA NA NA
NA ND
NA NA NA NA
Maximum
Canada outdoor indoor USA outdoor indoor Europe outdoor indoor Japan/Asia
Air (ng m–3)
Soil (µg kg–1) NA NA NA NA
Mean
Minimum
Concentration
Medium
Region
Table 5 (continued)
NA NA 0.071 NA
NA NA 2250 NA
0.5 1.5 NA
1.9 NA
NA NA
NA NA NA NA
10th Percentile a
1.0 NA 0.61 NA
NA NA 1.9¥104 NA
1.0 13 NA
4.55 35
NA NA
NA NA NA NA
Median a
NA NA 1.72 NA
NA NA 1.1¥106 NA
4.7 163 NA
15.5 120
NA NA
NA NA NA NA
90th Percentile a
629 3 47 10
8
32 26
89
includes sewage, slurry, effluent, and influent
no data references
no data references no data references
no data references
min. and max. are referenced averages 250 indoor air samples
detection limit not provided
no data references no data references no data references no data references
No. Data Comments Points b
Observed Concentrations in the Environment
151
d
c
b
a
<0.05 NA <0.1
<0.1 <0.5 <0.001 NA NA <0.1 <0.0025 NA
<0.01 0.03 NA
<0.01 <0.001
NA
0.039 0.05 0.4
0.08 7.4 0.01 0.02 0.05 0.13 0.0012 0.045
0.09 0.04 0.05
0.005 0.044
NA
50 NA 0.14 NA
10th Percentile a
NA
0.018 0.25
0.48 <0.1 NA
0.09 47.8 0.018 NA NA 0.8 0.008 NA
0.11 NA 1.6
NA
NA 0.0015
0.005 0.034 NA
NA 0.25 0.02 NA NA 0.05 0.0016 NA
0.025 NA 0.05
1.40¥104 NA NA NA 2.10¥105 1.3 NA NA
Maximum
NA
NA 0.0065
0.05 0.05 NA
NA 0.64 0.05 NA 0.05 0.05 0.005 0.045
0.025 0.05 0.05
NA NA 670 NA
Median a
NA
NA 0.12
0.13 0.05 NA
NA 8.5 0.05 NA NA 0.14 0.005 NA
0.068 NA 1.1
NA NA 7600 NA
90th Percentile a
48 57
19 4 3
3 23 31 13 9 9 35 3
6 4 6
79 3 62
no data references
mean represents 1/2 the detection limit
detected in 1 of 19 samples
mean represents 1/2 the detection limit
mean represents 1/2 the detection limit mean represents 1/2 the detection limit
mean represents 1/2 the detection limit
not separated into geographical locations
no data references
primary sewage sludge
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual data points. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available. ND not detected.
Food (µg Beverages Cereal Dairy (excl. milk) Eggs Fat & oils Fish Fruits Grains Meat Milk Nuts & beans Other foods Poultry Processed meat Vegetables Infant formulapowder Breast milk
g–1)
Sludge (µg kg–1) 3200 1.12¥105 1800 NA
Canada USA Europe Japan/Asia
Mean
Minimum
Concentration
Medium
Region
Table 5 (continued)
152 K. Clark et al.
Observed Concentrations in the Environment
153
6.2 Sediment
The mean concentration for sediment is higher in the United States (0.19 mg kg–1) as compared with that of Europe (0.059 mg kg–1), although there are less data references for Europe. Three references for Canada are available that cite a maximum concentration of 370 mg kg–1. This is higher than the maximum concentrations reported for the United States (5.5 mg kg–1) and Europe (18.2 mg kg–1). Only two references are reported for Japan/Asia, with significantly lower values (maximum concentration of 0.02 mg kg–1). 6.3 Soil
No data are available for BBP measured in soil. 6.4 Air
Concentrations in indoor air are higher than concentrations in outdoor air. The calculated mean indoor air concentration is 35 ng m–3 in Europe, which corresponds with the median of a United States study. The maximum indoor air concentration was in Europe and is 465 ng m–3. A concentration of 120 ng m–3 is reported as the 90th percentile of a study of indoor air in California. The mean outdoor concentration in Europe is 1.7 ng m–3, while the calculated mean concentration in the United States is 9.5 ng m–3. Canadian data range between 0.38 ng m–3 and 1.78 ng m–3. 6.5 Dust
Measurements of BBP in dust are only available for Europe. The concentrations range between <1000 µg kg–1 and 1.75 ¥106 µg kg–1, with a calculated mean concentration of 3.33 ¥105 µg kg–1. 6.6 Food
Several references report BBP concentrations measured in food; however, there are insufficient data to prepare a summary by region. Concentrations measured in the available food groups reveal the highest mean concentration for fats and oils (7.4 µg g–1). For the food categories of cereal, fruits, grains, nuts and beans, and processed meats, BBP was not detected and the mean value in Table 5 represents one half the detection limit. BBP was detected in only 1 of 19 samples of “other foods”. No measurements of BBP in breast milk are available.
154
K. Clark et al.
6.7 Other Media 6.7.1 Wastewater
The comparative results of BBP measured in various wastewaters, including industrial effluents and influents, leachates, and storm water, etc., are summarized in Table 5. The maximum mean concentration is calculated for the United States (299 µg L–1). This mean value is two orders of magnitude higher than the mean concentrations calculated for the other regions. The lowest calculated mean is for Japan/Asia (0.1 µg L–1). The Canadian maximum concentration of 84 µg L–1 is the same order of magnitude as the European maximum of 30 µg L–1. The highest reported maximum was measured in the United States (449 µg L–1). 6.7.2 Sludge
The majority of the reported data for sludges and composts are available for Canada and Europe. The maximum mean concentration is calculated for the United States (112,000 µg kg–1). The calculated mean for Canada is less (3200 µg kg–1), as is the calculated mean concentration for Europe (1800 µg kg–1). The maximum measured concentration is reported for Europe (210,000 µg kg–1); however, a maximum concentration is not reported in the study referenced for the United States.
7 Bis(2-Ethylhexyl) Phthalate (DEHP) A summary of the reported concentrations of DEHP is presented in Table 6. 7.1 Water 7.1.1 Surface Water
The lowest mean concentration (0.21 µg L–1) is for the United States. The calculated mean concentrations for Europe and Japan/Asia (0.93 µg L–1 and 0.9 µg L–1, respectively) are within the range of the Canadian urban and rural mean concentrations of 1.15 µg L–1 and 0.78 µg L–1, respectively. The overall Canadian mean is 3.78 µg L–1. Note that the Canadian means are strongly influenced by the assumption that non-detect values are equal to one half the detection limit. The median concentration for both urban and rural waters is 0.5 µg L–1, which represents one half the detection limit. The detection limits for the United States and European data are more than an order of magnitude lower than the lowest Canadian detection limit.
Drinking water (µg L 1.72 1.56 1.97 0.55 NA 0.91
Sediments (mg kg–1) NA 1.4 4.9 0.48
Canada USA Europe Japan/Asia
15.7 0.26 0.79
Canada rural urban USA Europe Japan/Asia
USA Europe Japan/Asia
Canada
0.78 1.15 5.69 0.21 0.93 0.9
0.05 0.05 0.05 <0.002 <0.008 <0.02
–1
NA 0.00027 0.0001 0.009
) 1.0 0.4 0.5 0.16 0.18 1.0
ND d <0.07 <1.0
Groundwater (µg L–1) NA NA
rural urban industrial sources USA Europe Japan/Asia
Canada
Minimum
Surface water (µg L–1) 3.78 0.05
Mean
Concentration
Medium
Region
NA 218 487 17
54 54 42 170 3.5 4.7
470 1.4 18.4
NA
14 11 336 137 50 15
336
Maximum
NA 0.041 0.019 NA
0.003 NA NA 0.22 NA 0.5
4.3 0.19 0.5
NA 0.16 0.29 NA
0.25 0.5 1.0 0.55 NA 0.5
15.7 0.67 0.5
NA
0.5 0.5 3.0 0.05 0.21 0.5
NAc NA NA 0.012 0.018 0.5 NA
2.87
Median a
0.29
10th Percentile a
Table 6. Summary of concentrations of bis(2-ethylhexyl) phthalate
NA 3.4 8.2 NA
1.4 NA NA 0.88 NA 3.0
27 1.1 11
NA
NA NA NA 1.6 1.9 3.2
6.6
90th Percentile a
83 405 1
959 422 499 4 NA 61
2 9 61
367 78 607 1309 397 198
1081
no data available
range only available
data primarily from Alberta Environment [8] median =0.5; represents 1/2 the detection limit median=1.0
maximum is presented as an average
Canadian data represented in drinking water summary below detection limit not specified
data primarily from Alberta Environment [9]; high/low data values removed median =0.5; represents 1/2 the detection limit median =0.5; represents 1/2 the detection limit median = 3.0
No. Data Comments Points b
Observed Concentrations in the Environment
155
Europe
Canada USA Europe Japan/Asia
Canada outdoor indoor USA outdoor indoor Europe outdoor indoor Japan/Asia outdoor indoor
0.5 <500
<0.4 20
0.28 18
<17 NA
2.0 NA
5.0 109
21.9 245
1450 1000
<0.1 0.03 4 10
2800 NA
1090 1046
65 240
5.0 3100
11 1280 5100 500
Maximum
NA NA
1.3 20
0.76 27
NA NA
NA NA 29 NA
10th Percentile a
Vegetation (µg NA
kg–1
dry wt.) 1200 11,300
NA
Dust (indoors) (µg kg–1) NA NA NA NA 3.24¥106 2.38¥106 4.10¥106 NA 4.58¥106 3.67¥104 6.62¥105 2000 NA NA NA NA
Air (ng m–3)
Soil (µg kg–1) NA 0.03 48 NA
Canada USA Europe Japan/Asia
Mean
Minimum
Concentration
Medium
Region
Table 6 (continued)
NA
NA NA 4.71¥105 NA
1450 NA
17.5 111
2.3 55
NA NA
NA NA 50 NA
Median a
NA
NA NA 2.09¥106 NA
NA NA
126 398
16 99
NA NA
NA NA 66 NA
90th Percentile a
2 55
2 1
85 26
34 107
3 NA
NA 1 3 NA
no data references
no data references
30 data points; only range available
No. Data Comments Points b
156 K. Clark et al.
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
Region
Mean
Food (µg g–1) Beverages 0.077 Cereal 0.53 Dairy 1.5 (excl. milk) Eggs 0.21 Fat & oils 4.1 Fish 0.46 Fruits 0.03 Grains 0.5 Meat 0.35 Milk 0.08 Nuts & beans 0.21
2.6 3.0 0.23 NA
10th Percentile a
0.045
<0.08
0.8
0.019 1.0 32 0.02 0.05 0.05 0.01
<0.01 0.6 0.7 11.9 5.00¥10–5 <0.02 0.11 <0.1 1.5 <0.01 0.8 <0.005 1.4
NA 0.039 0.43 0.4 0.015 0.032 0.076
0.01 0.68 54 16.5
4.40¥105 1.90¥104 5.83¥107 NA 2.60¥106 1.40¥104 170 NA
336 4400 1800 40
Maximum
1.7 1.7 16.8
0.006 0.02 0.059
0.004 0.004 0.0083 0.0053
3000 420 900 8
<0.1 0.01 0.068 <0.2
Minimum
Concentration
Other Media Wastewater (µg L–1) 5.72 27 34.4 5.6 Sludge (µg kg–1) 1.39¥105 3.01¥105 5.31¥104 48 Rainwater (µg L–1) 0.006 0.17 14 6.1
Medium
Table 6 (continued)
0.045
0.12 2.4 0.001 0.02 0.14 0.05 0.035
0.043 0.05 0.96
NA 0.17 0.85 1.8
6.80¥104 NA 4.80¥104 NA
5.2 8.3 1.74 NA
Median a
0.27
0.47 4.6 0.02 0.062 1.1 0.75 0.14
0.15 1.3 2.5
NA 0.26 38 14
1.60¥105 NA 1.60¥105 NA
8.8 61 53 NA
90th Percentile a
6
4 36 1.3 15 11 27 108
72 5 107
1 22 4 3
84 16 62 1
662 8 76 2
64
not separated into geographical locations
1 set of measurements in 1974
primary sewage sludge
includes sewage, slurry, effluent, and influent
No. Data Comments Points b
Observed Concentrations in the Environment
157
d
c
b
a
Mean
25 2.6 4.32 2.2 0.98 0.15 0.6 0.6
0.0098 <0.012
<0.005
0.01 0.01
Maximum
<0.01 0.05 <0.1
Minimum
Concentration
Food (µg g–1) Other foods 0.28 Poultry 1.1 Processed 0.94 meat Vegetables 0.17 Infant 0.2 formulapowder Infant 0.007 formulaliquid Breast milk 0.062 Baby food 0.12
Medium
NA 0.039
0.006
0.045 0.006
0.005 0.25 0.11
10th Percentile a
0.062 0.12
0.006
0.048 0.12
0.05 0.9 0.45
Median a
NA 0.23
0.007
1.1 0.64
1.3 2.2 1.9
90th Percentile a
10 16
9
84 66
63 4 25 minimum represents median from one study
not separated into geographical locations
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual datapoints. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available. ND not detected.
Region
Table 6 (continued)
158 K. Clark et al.
Observed Concentrations in the Environment
159
The maximum concentration measured in Canada is 336 µg L–1 (for surface water with substantial industrial sources). The maximum concentrations, measured in the United States, Europe, and Japan/Asia, are 137 µg L–1, 50 µg L–1, and 15 µg L–1, respectively. 7.1.2 Groundwater
The database for groundwater is considerably more limited compared to that for surface water. The highest maximum and mean concentrations for DEHP in groundwater have been measured in the United States as 470 µg L–1 and 15.7 µg L–1, respectively. The mean concentration for Europe is 0.26 µg L–1, which is comparable to the Japan/Asian calculated mean concentration of 0.79 µg L–1. Canadian data are available in a drinking water database reported below; although some of these data represent Canadian groundwater, the data do not differentiate between surface water supplies and groundwater supplies. 7.1.3 Drinking Water
The calculated mean concentration of DEHP in drinking water is highest in Canada (1.72 µg L–1). Although, considerably less data are available for the other regions, the calculated mean concentrations in the United States and Japan/Asia are slightly less (0.55 µg L–1 and 0.91 µg L–1, respectively). However, the maximum concentration reported for the United States (170 µg L–1) is three times higher than the Canadian maximum (54 µg L–1). The European and Japan/Asian maximums are 3.5 µg L–1 and 4.7 µg L–1, respectively. 7.2 Sediment
The mean concentration in sediment is highest in Europe (4.9 mg kg–1). The mean concentrations are somewhat lower in Japan/Asia (0.48 mg kg–1) and the United States (1.4 mg kg–1) although there are less data references. Maximum referenced concentrations range between 17 mg kg–1 (Japan/Asia) and 487 mg kg–1 (Europe). 7.3 Soil
Very little data are available for DEHP measured in soil. Of the data referenced, DEHP concentrations in soil are higher in Europe with a mean concentration of 48 µg kg–1. The maximum concentrations range between 11.0 µg kg–1, reported for Canada, and 5100 µg kg–1, reported for Europe. Data for the United States are reported to range between 0.03 µg kg–1 and 1280 µg kg–1. Only one reference is reported for Japan/Asia, with concentrations in soil ranging between 10 µg kg–1 and 500 µg kg–1.
160
K. Clark et al.
7.4 Air
In general, measured indoor concentrations are higher than concentrations outdoors. The Canadian and Japan/Asian datasets are very limited compared to the United States and European datasets. The mean outdoor concentration ranges from 2.0 ng m–3 in Canada to 1450 ng m–3 in Japan/Asia. The mean indoor concentration ranges from 109 ng m–3 in the United States to 1000 ng m–3 in Japan/Asia. 7.5 Dust
DEHP concentrations measured in dust are only available for Europe and the United States. Europe reports concentrations ranging between 2000 µg kg–1 and 4.58 ¥106 µg kg–1, with a mean of 6.62 ¥105 µg kg–1. The United States reports concentrations ranging between 2.38 ¥106 µg kg–1 and 4.10 ¥106 µg kg–1, with a mean of 3.24 ¥106 µg kg–1. 7.6 Food
DEHP concentrations measured in food are the most extensive of all of the phthalates reported; however, there are insufficient food data to prepare a summary by region. DEHP concentrations measured in the various food groups in Table 6 reveal the highest mean concentration for fats and oils (4.1 µg g–1). The highest maximum concentrations are in fish (32 µg g–1) and “other foods” (25 µg g–1). Note that the data have not been separated by year of sampling, although it should be noted that food processing practices may have changed since the time of sampling. 7.7 Other Media 7.7.1 Wastewater
The comparative results of DEHP measured in various wastewaters, including industrial effluents and influents, leachates, and storm water, etc., are summarized in Table 6. The maximum mean concentration is calculated for Europe (34.4 µg L–1). This mean is only slightly higher than the mean concentrations calculated for the other regions. Japan/Asia and Canada have the lowest calculated means (5.6 µg L–1 and 5.72 µg L–1, respectively). The highest reported maximum was measured in the United States (4400 µg L–1).
Observed Concentrations in the Environment
161
7.7.2 Sludge
The majority of the reported data for sludges and composts are available for Canada and Europe. The maximum mean concentration is calculated for the United States (301,000 µg kg–1). The calculated mean for Canada is in the same order of magnitude (139,000 µg kg–1), while the calculated mean concentration for Europe is slightly lower (53,000 µg kg–1). A single reference is available for Japan/Asia, which reports a much lower mean concentration (48 µg kg–1). 7.7.3 Rainwater
The highest calculated mean concentration is for Europe (14 µg L–1), while the lowest mean concentration is for Canada (0.006 µg L–1). The highest maximum concentration was also measured in Europe (54 µg L–1).
8 Di-n-Octyl Phthalate (DnOP) A summary of the reported concentrations of DnOP is presented in Table 7. 8.1 Water 8.1.1 Surface Water
The overall mean concentration for Canada is calculated to be 1.35 µg L–1. The data include rural and urban locations, as well as surface waters with substantial industrial sources. The maximum concentration (45 µg L–1) was recorded in an urban area; however, the highest mean concentration (5.5 µg L–1) is calculated for surface water with substantial industrial input. The calculated mean concentration for urban areas (1.10 µg L–1) is comparable with that of the rural areas (0.91 µg L–1). 8.1.2 Drinking Water
Measured concentrations in Canada range between the detection limit (<1.0 µg L–1) and 11 µg L–1. The mean concentrations are 0.54 µg L–1 and 0.53 µg L–1, for urban and rural areas, respectively. Analysis of 25 drinking water supplies in the United States is reported by Health Canada [17]. The results did not exceed the detection limit. DnOP was detected, but not quantified, in three aquifer-derived drinking water supplies in New York State [17]. DnOP was detected in drinking water in Europe; however, levels were not quantified [17].
Canada USA Europe Japan/Asia
rural urban USA Europe Japan/Asia
Canada
Sediment (mg kg–1) 15 NA NA NA
<0.015 NA NA NA
0.5 0.5 NA NA NA
15 NA NA NA
9 11 NA NA NA
11
Drinking water (µg L–1) 0.53 0.5
0.53 0.54 NA NA NA
18 45 10 NA NA NA
0.91 1.10 5.50 NA NA NA
45
Maximum
0.05 0.05 0.1 NA NA NA
rural urban industrial sources USA Europe Japan/Asia
Canada
Minimum
Surface water (µg L–1) 1.35 0.05
Mean
Concentration
Medium
Region
NA NA NA NA
NA NA NA NA NA
NA NA NA NA
0.5 0.5 NA NA NA
NA
0.5 0.5 5.0 NA NA NA
NAc NA NA NA NA NA
NA
0.0071
Median a
0.0036
10th Percentile a
Table 7. Summary of concentrations of di-n-octyl phthalate
NA NA NA NA
NA NA NA NA NA
NA
NA NA NA NA NA NA
1.0
90th Percentile a
1
625 657
1276
1115 165 621
1901
no data reported no data reported no data reported
data primarily from Alberta Environment [8], EC & HC [16] and HC [17] median=0.5; represents 1/2 the detection limit median=0.5; represents 1/2 the detection limit no data reported no data reported no data reported
data primarily from Alberta Environment [9], EC & HC [16] and HC [17] median=0.5; represents 1/2 the detection limit median=0.5; represents 1/2 the detection limit median=5.0 no data reported no data reported no data reported
No. Data Comments Points b
162 K. Clark et al.
c
b
a
NA NA NA NA
Mean
<10 NA NA NA
NA NA NA NA
Minimum
Concentration
Other media Sludge (µg kg–1) 5420 NA NA NA
Air (ng m–3)
Medium
63,000 NA NA NA
NA NA NA NA
Maximum
NA NA NA NA
NA NA NA NA
10th Percentile a
NA NA NA NA
NA NA NA NA
Median a
NA NA NA NA
NA NA NA NA
90th Percentile a
79
1
no data reported no data reported no data reported
primary sewage sludge
no data reported reference for flyash only no data reported no data reported
No. Data Comments Points b
The 10th percentile, median, and 90th percentile values were calculated by using the means of the studies, not the individual data points. The number of data points represents the number of data points available with sufficient information to calculate an arithmetic mean. NA not available.
Canada USA Europe Japan/Asia
Canada USA Europe Japan/Asia
Region
Table 7 (continued)
Observed Concentrations in the Environment
163
164
K. Clark et al.
8.2 Sediment
Environment Canada and Health Canada [16] includes three references to the analyses of DnOP in sediments. One of the references reports data collected 0.5 km and 1 km downstream of a sewage outfall. The data range between <0.015 mg kg–1 and 15 mg kg–1. 8.3 Soil
No data have been identified. 8.4 Air
The unquantified presence of DnOP in air in Europe is reported [17]. In addition, a study of air quality from a coal-fired power station in the United States reports levels of DnOP of 517,000 ng m–3 and 751,000 ng m–3 [17]. 8.5 Dust
No data have been identified. 8.6 Food
Concentrations of 0.081 µg g–1 and 0.11 µg g–1 are reported in two out of ten fish samples; however, the results are considered suspect due to possible sources of contamination in sampling and analysis [16]. 8.7 Other Media 8.7.1 Sludge
Webber and Nichols [11] evaluated 72 samples of sewage sludge analyzed from 12 locations across Canada. Webber and Lesage [10] reported seven sludge samples analyzed for DnOP. The measured concentrations of DnOP in sludge range between the detection limit (assumed to equal 10 µg kg–1) and 63,000 µg kg–1 with an overall calculated mean of 5420 µg kg–1.
Observed Concentrations in the Environment
165
8.7.2 Wastewater
As noted for surface water, the industrial surface water data reported by Alberta Environment [9] range between 0.1 µg L–1 and 10 µg L–1 with a calculated mean concentration of 5.5 µg L–1.
9 Discussion and Recommendations An in-depth analysis and discussion of the DEHP concentration ranges in Europe is presented here. DEHP is selected for this analysis as there are more monitoring data available for DEHP than for any other phthalate ester. The analysis can be applied to other phthalate esters and different geographical regions, where sufficient monitoring data are available. 9.1 Histograms
Four types of histograms are plotted to show the distributions of concentrations of DEHP in air, surface water, and sediment in Europe, that is, unweighted frequency, weighted frequency, unweighted cumulative, and weighted cumulative. Thus, in total, 12 histograms are plotted for DEHP (see Figs. 1–6). The location of the calculated weighted average is shown on each histogram. Where possible, data corresponding to a laboratory detection limit are identified. The weighting of study averages to the number of samples in the study has a large effect on the distributions in the histograms and can dictate the magnitude of the calculated weighted average. Furthermore, as the weighted average is an arithmetic average, a few high concentrations can bias the concentrations upwards. This is evidenced by the cumulative histograms, which show that the 50th percentile is usually at a lower concentration than the weighted average concentration. If all of the raw data from every study were available (and not just the study arithmetic mean and number of samples), the preferred approach would be to calculate a geometric mean (i.e., a mean of the logarithms of each reported concentration). This approach would help to reduce the effect of outliers on the calculated average and would make weighting unnecessary. The histograms show that DEHP concentrations are characterized as extending over a very wide range, which makes interpretation and averaging difficult. For example, water concentrations vary from 0.003 to 10 µg L–1, that is, by a factor of 3000. We believe that this wide distribution is due to variable proximity to sources, that is, the concentrations on the right-hand side of the histograms are probably representative of impacted sites and those on the left-hand side are likely to be representative of regional background environmental concentrations. Separation of the monitoring data into groups based on proximity to sources would require going back to the source references and allocating concentrations to different classifications such as “urban/industrial” or “rural/background”. Even after careful scrutiny of source references, separation of the data may not be pos-
Fig. 1. Frequency histogram of the distribution of DEHP water concentrations in Europe from published studies
166 K. Clark et al.
Fig. 2. Cumulative histogram of the distribution of DEHP water concentrations in Europe from published studies
Observed Concentrations in the Environment
167
Fig. 3. Frequency histogram of the distribution of DEHP sediment concentrations in Europe from published studies
168 K. Clark et al.
Fig. 4. Cumulative histogram of the distribution of DEHP sediment concentrations in Europe from published studies
Observed Concentrations in the Environment
169
Fig. 5. Frequency histogram of the distribution of DEHP air concentrations in Europe from published studies
170 K. Clark et al.
Fig. 6. Cumulative histogram of the distribution of DEHP air concentrations in Europe from published studies
Observed Concentrations in the Environment
171
172
K. Clark et al.
sible. Therefore, we have sought to interpret the monitoring data based on the cumulative histograms as follows: – One third represent “remote from source” samples. – One third represent “close to source” samples. – One third represent intermediate samples. Based on the cumulative histograms for DEHP (Figs. 2, 4, and 6), concentration ranges and average values are assigned to the first two categories for surface waters, sediments, and air in Europe. The corresponding fugacities are also calculated. It is not possible to undertake a similar analysis for soil because of the lack of monitoring data available. For water, 90% of the reported data lie between 0.01 and 3.00 µg L–1 with a weighted average of 0.93 µg L–1. We interpret the distribution to indicate that, in water which is not subject to direct discharge of DEHP, the concentration probably lies in the range 0.01–0.10 µg L–1, that is, typical values of 0.03 µg L–1 plus or minus a factor of three. In waters that are in the vicinity of discharges, concentrations probably lie in the range 0.3–3.0 µg L–1 with a typical value of 1.0 µg L–1, plus or minus a factor of three. The corresponding fugacities of these averages are 12 and 390 nPa for the “remote from source” and “close to source” samples, respectively. For sediments, 90% of the reported data lie between 0.001 and 10 mg kg–1 with a weighted average of 4.9 mg kg–1. We interpret the distribution to indicate that, in water that is not subject to direct discharge of DEHP, the concentration probably lies in the range 0.001–0.1 mg kg–1, that is, typical values of 0.01 mg kg–1 plus or minus a factor of ten. In sediments that are in the vicinity of discharges, concentrations probably lie in the range 1.0–10 mg kg–1 with a typical value of 3.0 mg kg–1, plus or minus a factor of three. The corresponding fugacities of these averages are 0.12 and 37 nPa for the “remote from source” and “close to source” samples, respectively. For air, 90% of the reported data lie between 0.3 and 100 ng m–3 with a weighted average of 21.9 ng m–3. We interpret the distribution to indicate that, in air which is not subject to direct discharge of DEHP, the concentration probably lies in the range 0.3–3.0 ng m–3, that is, typical values of 1.0 ng m–3 plus or minus a factor of three. In air that is in the vicinity of discharges, concentrations probably lie in the range 10–100 ng m–3 with a typical value of 30 ng m–3, plus or minus a factor of three. The corresponding fugacities of these averages are 3.7 and 110 nPa for the “remote from source” and “close to source” samples, respectively. 9.2 Fugacities in Various Media
A plot of DEHP and DBP fugacities for a range of environmental media is shown in Fig. 7. The ranges of the reported environmental concentrations in the European dataset are represented by the bars on Fig. 7. Fugacities of DEHP in different media decrease by a factor of about 1000 from left to right on the plot, from air to fish. The fugacities of DEHP in
Observed Concentrations in the Environment
173
Fig. 7. Estimated fugacities of DEHP and DBP in environmental media
Fig. 8. Comparison of food chain biomagnification for two contrasting organic compounds
174
K. Clark et al.
fish, cow’s milk, and human breast milk are considerably lower than in other media. We hypothesize that this decrease in fugacity represents a biodilution effect, as DEHP is transferred through aquatic and terrestrial food chains. As discussed by McLachlan [19], three factors combine to cause these biodilution effects: 1. Chemical degradation. 2. Reduced absorption of highly hydrophobic compounds in the digestive tract. 3. Kinetically limited uptake of involatile, hydrophobic compounds by plants. It is believed that for DEHP, all three factors are likely to have some contribution to the observed biodilution effects. McLachlan [19] focused on the air/soil-plant-cow-milk-human food chain, which is important for human exposure to polychlorinated biphenyls (PCBs) and polychlorinated dibenzo-p-dioxins and furans (PCDD/Fs), as well as the phthalate esters. In Fig. 8, the fugacities of DEHP in European air, soil, plants, cow’s milk, and human milk are compared to fugacities reported by McLachlan [19] for hexachlorobenzene (HCB), in samples collected in Germany. This provides an interesting comparison of different chemicals in the same food chain. The fugacities of HCB clearly increase by a factor of 300 (or 2.5 log units) between air and human breast milk indicating strong biomagnification. In contrast, DEHP biodilutes by a factor of 1000 (or 3.0 log units). There is strong biodilution from air to plant (a factor of 13) and from plant to cow (a factor of 32). The drop in fugacity from air to plant is probably the result of the third factor listed above: kinetically limited uptake by plants. The drop in fugacity from plant to cow may be a combination of the first and second factors: degradation and reduced absorption in the digestive tract. McLachlan [19] found that the 2,3,7,8-substituted polychlorinated dibenzo-p-dioxins and furans, which have similar hydrophobicities (KOW of 106 –108) to DEHP (KOW of 107.65), but are highly persistent, may biodilute by a factor of 10–100 between air and human breast milk. This is lower than the biodilution effect observed for DEHP of about 1000 and suggests that for DEHP degradation (i. e., biotransformation) contributes to the biodilution effect (Fig. 9). In Fig. 7, the fugacities of DEHP and DBP in the same media are compared. This is the only meaningful comparison that we can make with another phthalate ester because of a lack of comparable quality data for other phthalates. DBP fugacities only drop by a factor of 10 from air to fish and are thus close to equifugacity in all media. Because DBP has a lower hydrophobicity than DEHP, its moderate biodilution is likely to be mediated by metabolism only. The observed biodilution of DEHP must be regarded as fortuitous because it reduces human exposure, which would be much greater if the food supply were in equilibrium with the air. Furthermore, the lack of DEHP biomagnification will limit the exposure of upper trophic level mammals to the monoethylhexyl ester (MEHP), a metabolite of DEHP, and the putative toxic species in mammals. MEHP is formed by enzymatic hydrolysis of DEHP in the intestine and liver of organisms. MEHP is water-soluble and will not bioaccumulate or biomagnify in the food chain, but will instead be excreted (or conjugated and excreted) in the urine of the organism in which it is formed.
Observed Concentrations in the Environment
175
Fig. 9. Comparison of the biodilution of two compounds with similar hydrophobicities but different metabolism rates
This approach of plotting fugacity as a function of medium can, we believe, provide interesting insights into chemical fate. As discussed in the chapter by Cousins and Mackay entitled “multimedia mass balance modeling of two phthalate esters by the regional population-based model (RPM)”, for DEHP and DBP, emissions to air represent more than 90% of the total emissions to the environment. A hydrophobic substance that is less persistent (i.e., more reactive) will display a decreasing fugacity from source to sink because the rate of loss in “sink” media is sufficient to reduce fugacities, since there is insufficient time for the solute to diffuse in adequate quantities to equalize the fugacity. A fugacity gradient approach of this type thus indicates two features of the chemical behavior: 1. It is likely that the highest fugacity medium is the primary source and the lowest fugacity media tend to be sinks subject to transport from the source. 2. The downward source to sink slope is an indication of reactivity, that is, a high negative slope corresponds to high reactivity and short half-lives or persistence. 9.3 Conclusions
For all phthalates, there is a considerable amount of data reporting measured concentrations in surface waters and sediments in Europe and the United States. However, there is a lack of outdoor air data available and a severe paucity of data reporting measured concentrations in soil. Given the estimated importance of
176
K. Clark et al.
soil in terms of environmental fate (i.e., it is the primary medium of accumulation for phthalate esters), additional measurements in soil would be helpful for model verification. Concentrations of phthalate esters are characterized by a very wide range, which makes interpretation and averaging difficult.We believe that this wide distribution is due to variable proximity to sources. Therefore, we have sought to interpret the monitoring data as follows: one third represent “remote from source” samples, one third represent “close to source” samples, and one third represent intermediate samples. The EUSES (European Union System for the Evaluation of Substance) software, which is the preferred tool in Europe for the risk assessment of chemicals, is designed to undertake a three-tier risk assessment, that is, local-scale, regionalscale, and continental-scale [23, 24]. This three-tier approach helps to determine risks to humans and organisms in differently impacted regions. The EUSES modeling approach would fit quite well with the proposed separation of monitoring data into different classifications; however, the approach is most successful when large amounts of high quality data are available. By transforming environmental concentrations into fugacities, it becomes clear that DEHP appreciably biodilutes as it is transferred through aquatic and terrestrial food chains. Three factors combine to cause these biodilution effects: chemical degradation, reduced absorption of highly hydrophobic compounds in the digestive tract, and kinetically limited uptake of involatile, hydrophobic compounds by plants. However, degradation is thought to have the largest influence on the observed biodilution of phthalate esters. DBP does not biodilute as appreciably. Acknowledgement. We are grateful to the Environmental Research Task Group (ERTG) of the Phthalate Ester Panel of the American Chemistry Council (ACC) for funding this research.
10 References 1. Exxon Mobil Biomedical Sciences Inc (1999) Compilation of data for five phthalate esters. East Millstone, NJ, Prepared for American Chemistry Council, Arlington, VA 2. RIC (1999) DBP risk assessment – Environmental sampling 3. RIC (2000) RIVM monitoring program for milk, vegetation, cattle feed, fish and air. Subproject E2. ECPI\2000–07 and 2000–14 4. Alberti J, Brull U, Furtmann K, Braun G (2000) Occurrence of Phthalates in German surface and wastewater. Presented at SETAC World Congress, Brighton 5. Bruns-Weller E, Pfordt J (2000) Z Umweltchem Okotox 12:125 6. VROM (1998) Risk assessment – Dibutylphthalate. Prepared for the Ministry of Housing, Spatial Planning and the Environment (VROM). Prepared by the Netherlands Organization for Applied Scientific Research (TNO) and the National Institute of Public Health and the Environment (RIVM). Draft, 17 November 7. National Chemicals Inspectorate (NCI) (2000) Risk assessment – Bis(2-ethylhexyl) phthalate. Sweden, Draft, May 8. Alberta Environment (1999) Data analysed for phthalates in drinking water in Alberta. Enforcement and Monitoring Division. Edmonton, AB, Canada 9. Alberta Environment (1999) Data analysed for phthalates in surface water in Alberta. Water Sciences Branch, Water Data Management Section. Edmonton, AB, Canada
Observed Concentrations in the Environment
177
10. Webber MD, Lesage S (1989) Waste Manage Res 7:63 11. Webber MD, Nichols JA (1995) Organic and Metal contaminants in Canadian municipal sludges and a sludge compost. Wastewater Technology Centre, Burlington, ON, February 12. Environment Canada and Health Canada (2000) Canadian Environmental Protection Act, Priority Substances List assessment report – Butylbenzyl phthalate. Ottawa, ON 13. Environment Canada and Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List assessment report – Bis(2-ethylhexyl) phthalate. Ottawa, ON 14. Environment Canada and Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List assessment report – Dibutyl phthalate. Ottawa, ON 15. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List – Supporting documentation health-related sections – Di-n-Butyl phthalate. Ottawa, ON 16. Environment Canada and Health Canada (1993) Canadian Environmental Protection Act, Priority Substances List assessment report – Di-n-octyl phthalate. Ottawa, ON 17. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List – Supporting documentation health-related sections – Di-n-octyl Phthalate. Ottawa, ON 18. Clark K, Cousins I, Mackay D (2001) Multimedia modelling and exposure assessment for phthalate esters – Observed concentrations in the environment. Prepared for American Chemistry Council 19. McLachlan MS (1996) Environ Sci Technol 30:252 20. Mackay D (2001) Multimedia environmental models: The fugacity approach, 2nd edn. Lewis Publishers, Boca Raton, FA 21. Gruber L, Wolz G, Piringer O (1998) Deutsche Lebensmittel-Rundschau 94:177 22. Cousins I, Mackay D (2000) Chemosphere 41:1389 23. Berding V, Schwartz S, Matthies M (1999) Environ Sci Pollut Res 6:37 24. European Chemical Bureau (EC) (1997) EUSES documentation – The European Union system for the evaluation of substances. National Institute of Public Health and Environment (RIVM), the Netherlands, available from European Chemical Bureau (EC/DGXI), Ispra
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 179– 200 DOI 10.1007/b11466
Multimedia Mass Balance Modelling of Two Phthalate Esters by the Regional Population-Based Model (RPM) Ian T. Cousins · Donald Mackay Canadian Environmental Modelling Centre, Environmental and Resource Studies, Trent University, Peterborough, Ontario, K9J 7B8, Canada. E-mail:
[email protected]
Achieving an adequate understanding of the fate of commercial chemicals in the environment is best demonstrated by assembling a comprehensive multimedia mass balance model of the chemical’s behaviour in a specified region. This approach is demonstrated for two phthalate esters, di-n-butyl phthalate (DBP) and di-2-ethylhexyl phthalate (DEHP), in an industrialized region with similar characteristics (including population density) to the Netherlands and the Eastern United States. To accomplish this task the relevant physical-chemical properties of these substances are compiled and emission rates are estimated on a per capita basis. The “regional population-based model” (RPM) is described and concentrations of the phthalate esters are estimated and compared with available monitoring data from Europe.A sensitivity and uncertainty analysis is also included. Comparison of model predictions and monitoring data suggests that the major uncertainties in the multimedia fate assessment are the degradation halflives and emission rates, but despite this, most reported concentration ranges are within a factor of four of the median predicted concentrations. Improved agreement between predicted and observed water, sediment and fish concentrations of DEHP is obtained by using a measured value of KOC for the suspended particles in the water column. The overall persistence or residence time of DBP and DEHP attributable to reaction only (not including advection) are estimated to be 25 and 47 days, respectively; thus historical accumulation of these chemicals over periods of years or decades is unlikely. It is concluded that, in general, the model captures the key processes that control the behaviour of these substances and predicts environmental concentrations that are of the correct order of magnitude. Consequently, this analysis appears to provide a reasonable quantitative description of the environmental fate of these chemicals. Keywords. Phthalate ester, DBP, DEHP, Multimedia, Model, Mass balance
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
1
Introduction
2
Regional Population-Based Model (RPM)
3
Chemical Input Data . . . . . . . . . . . . . . . . . . . . . . . . . 185
4
Environmental Emissions . . . . . . . . . . . . . . . . . . . . . . 186
5
Monitoring Data Used for Model Evaluation . . . . . . . . . . . . 190
6
Model Results and Discussion . . . . . . . . . . . . . . . . . . . . 191
7
Conclusions and Recommendations
8
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
. . . . . . . . . . . . . 181
. . . . . . . . . . . . . . . . 199
© Springer-Verlag Berlin Heidelberg 2003
180
I.T. Cousins and D. Mackay
1 Introduction The objective of this chapter is to demonstrate how a regional multimedia mass balance model can be successfully used to reconcile observed environmental concentrations of two commercial phthalate esters with reported emission rates. A satisfactory reconciliation has several benefits indicating that emissions are fully accounted for, and that the dominant fate processes (including degradation rates) are in accord with model predictions. Furthermore, once a reasonable model evaluation has been demonstrated, future changes in emission scenarios can be translated into projected changes in environmental concentrations. It can also, with appropriate modification, be applied to other phthalate esters, and indeed to other chemicals. In general, the availability of a validated mass balance model provides confidence that the fate of chemicals in the environment can be predicted, given sufficiently accurate data on their properties and emissions. To accomplish this objective, a model is developed that is regarded as intermediate between the use of a region-specific model such as CalTOX [1] or ChemCAN 4.0 [2] applied to a specific region, and a purely evaluative model such as EQC [3]. This approach is similar to that taken by the SimpleBox 1.0 regional model [4], which is designed to represent a typical region of the European Union (EU). SimpleBox 1.0 is used in the EUSES (European Union System for the Evaluation of Substances) software, which is the preferred modelling tool for risk assessments of new and existing substances, performed by EU Member States [5]. The model presented here, the “regional population-based model” (RPM), is parameterised to represent an industrially developed region, typical of Western Europe or the Eastern United States. The aim is not to validate the model against actual concentration data from a specific geographical region, rather it is to reconcile model predicted concentrations with observed values from regions of similar population density. This approach has the advantage that it permits a larger database of concentrations to be used. This model is regarded as being particularly suitable for substances which are consumed on a fairly constant per capita basis such as components of plastics and “down the drain” chemicals. The model was evaluated by using European emission and monitoring data, mainly because of data availability. This chapter is structured into five sections. First, the RPM model is described. Second, the required physical-chemical property and reactivity model inputs are summarized, drawing on information presented in the physical-chemical property chapter. Model calculations are presented for two phthalate esters, di-n-butyl phthalate (DBP) and di(2-ethylhexyl) phthalate (DEHP), for which we have the most extensive and highest quality chemical input data available. Moreover, for these two substances, observed environmental concentrations are consequently above detection limits thereby allowing a thorough model evaluation. Third, per capita emission estimates taken from Parkerton et al. [6] are used to calculate emissions to the model region. Fourth, the monitoring data used for comparison with the model predictions are reviewed. Finally, RPM-predicted concentrations are compared with environmental measurements, and sensitivity and uncertainty analysis is used to identify the most sensitive and important input data.
181
Multimedia Mass Balance Modelling of Two Phthalate Esters
2 Regional Population-Based Model (RPM) RPM is used to evaluate the environmental fate of DBP and DEHP by relating their emissions on a per capita basis to observed environmental concentrations. The region modelled is designed to be representative of densely populated, industrialized areas. Therefore, the first task of model development was to identify a variety of industrialized regions and collect data for surface area, population, population density and surface water coverage (Table 1). Experience with designing model environments has shown that a region should not be too large, since this results in excessive “dilution”, or too small, since losses by advection in air leaving the region become relatively too important. The EQC model [3], developed as a tool for evaluating chemical fate, has a surface area of 100,000 km2, which is about the size of the US state of Ohio or England, Greece or Portugal.We decided that this was a reasonable geographic area for the RPM model. Population densities in the selected industrialized regions are 101–419 people km–2 (Table 1). Japan, New Jersey and the heartlands of industrial Europe have the highest population densities of 300–450 people km–2. It is noteworthy that within these larger regions there will be population “hotspots” in cities in which population densities can be as high as 1000–5000 people km–2, and low populations in rural areas of less than 10 people km–2. The population density selected is largely dependent on the region from which the monitoring data used for model evaluation are collected. We decided to use a population density of 400 people km–2 (comparable to the Netherlands or New Jersey) for the RPM, which equates to a population for the model environment of 40 million people. This population density is higher than average, but the most reliable and comprehensive field monitoring data that are presently available for comparison to model predictions are available for this characteristic region. RPM is a Level III, steady-state model that treats five bulk compartments: air, surface water, bottom sediments, surface soil and terrestrial vegetation. Figure 1 summarizes the model compartments, transport processes and reaction losses. Table 1. Surface area, population and percent water cover of industrialized regions
Country/Province/ State
Area (km2)
New Jersey New York Ohio Pennsylvania UK France Germany Italy Netherlands Belgium Japan
20,000 128,000 107,000 117,000 245,000 547,000 357,000 301,000 37,000 31,000 378,000
Population 8,314,000 17,783,000 10,744,000 11,853,000 57,592,000 58,609,000 82,072,000 56,831,000 15,650,000 10,165,000 125,733,000
Pop. density (per km2)
% Water
412 138 101 101 235 107 230 189 419 333 333
2.8 0.7 1.4 2.5 1.3 0.3 2.1 2.4 9.1 0.9 0.8
182
I.T. Cousins and D. Mackay
Fig. 1. Schematic diagram of the RPM model showing transport and reaction processes
A mass balance equation is written for each compartment and the set of equations is solved algebraically. The equations are written in fugacity format and solved for the fugacities in each of the bulk media (Table 2). In fugacity format, the chemical concentration in a medium (C, mol m–3) equals the fugacity of chemical in the medium (f, Pa) multiplied by the fugacity capacity or Z value (Z, mol m–3 Pa–1) of the medium for that chemical; thus, C = Zf. The Z value quantifies the relative affinity of the chemical for a particular medium. The processes that govern chemical movement and reaction are quantified by D values (D, mol h–1 Pa–1). The D values, in conjunction with the chemical fugacity in a medium, are used to calculate the chemical flux (N, mol h–1) that is either transported or reacted. A full description of the formulation of the Z and D values is available in a series of articles by Mackay et al. [3, 7–9]. The method of including vegetation as an additional compartment in the mass balance is discussed in detail by Cousins and Mackay [10]. It was deemed necessary to include a vegetation compartment because it is expected that phthalate esters will partition appreciably to foliage because of their high octanol-air partition coefficients (see the chapter discussing physical-chemical properties). Environmental characteristics of the RPM model region are given in Table 3. Mass transfer coefficients are listed in Table 4. A notable difference between RPM and the EQC model [3] is the
Multimedia Mass Balance Modelling of Two Phthalate Esters
183
Table 2. Steady-state mass balance equations used in RPM and their algebraic solution
Air Water Sediment Vegetation Soil compartment
f1 DT1 = I1 + f2 D2,1 + f4 D4,1 + f5 D5,1 f2 DT2 = I2 + f1 D1,2 + f3 D3,2 + f5 D5,2 f3 DT3 = I3 + f2 D2,3 f4 DT4 = I4 + f1 D1,4 + f5 D5,4 f5 DT5 = I5 + f1 D1,5 + f4 D4,5
where f1 to f5 are the fugacities for each compartment (Pa), DT1 to DT5 are the sum of all the D-processes that transport chemical out of each compartment (reaction and advection losses) (mol Pa–1 h–1), I1 to I5 are the inputs (emissions and chemical flowing into the model in background air or water) to each compartment (mol h–1) and the D values on the right hand side of the equations are intercompartment D values (mol Pa–1 h–1) Solution f1 = (IAEV + f2 D2,1)/DTC1 f2 = IT/(DT2 –D3,2 D2,3/DT3 – D2,1/DTC1 (D1,2 + DTC2 D5,2/DTC3)) f3 = I3/DT3 + f2D2,3/DT3 f4 = (I4 + I5 D5,4/DT5 + f1(D1,4 + D5,4D1,5/DT5))/(DT4 – D4,5D5,4/DT5) f5 = (I5 + I4 D4,5/DT4 + f1 (D1,5 + D4,5 D1,4/DT4))/(DT5 – D5,4D4,5/DT4) where IT = I2 + I3 D3,2/DT3 + I4 D4,5 D5,2/(DT4DTC5) + I5 D5,2/DTC5 + IAEV DTC2/DTC1 IAEV = I1 + I4 D4,1/DTC4 + I4 D4,5 D5,1/(DTC5DT4) + I5 D5,1/DTC5 + I5 D5,4 D4,1/(DTC4DT5) DTC1 = DT1 –D4,1 (D1,4 + D5,4 D1,5/DT5)/DTC4 – D5,1 (D1,5 + D4,5 D1,4/DT4)/DTC5 DTC2 = D1,2 + (D5,2/DTC5) (D1,5 + D4,5 D1,4 DT4) DTC4 = DT4 – D4,5 D5,4/DT5 DTC5 = DT5 – D5,4 D4,5/DT4
assumed percentage of surface area covered by freshwater. For RPM it is 3%, chosen to be representative of an area drained by rivers, and not containing large lakes (see Table 1). The EQC model has a large water surface area of 10%, which is chosen to account for the effect of large lakes and coastal waters on chemical fate. The 3% surface area selected for RPM is the same as in SimpleBox [4]. Surface water predictions obtained from the RPM model were used to estimate concentrations in fish using the equation: CFISH (mg/kg lipid) = BCF ¥ CWATER (mg/L) ¥ FDISSOLVED ¥ frLIPID
(a)
where BCF is the measured bioconcentration factor for fish, CWATER is the total water concentration, FDISSOLVED is the fraction of chemical in the dissolved phase, which is calculated (as explained previously in the chapter discussing physicalchemical properties), and frLIPID is the mass fraction of lipid in the fish (kg lipid kg–1 fish), which is assumed to be 0.05 (or 5%). In addition to choosing single ‘preferred’ values for the model input parameters, a range of values has been selected to give a more realistic expression of uncertainty in the output. Uncertainty and sensitivity analysis of the model output to all model input parameters (most of which are listed in Tables 3–6) were carried out by simultaneously varying them by using Crystal Ball 4.0* (Deci-
184
I.T. Cousins and D. Mackay
Table 3. Environmental input parameters used in RPM
Parameter
Median value
Unit
Confidence factor a
Total area of region Water surface area Environmental temperature Water depth Air mixing height Soil mixing depth Sediment mixing depth Percent terrestrial surface covered by vegetation Leaf area index Plant biomass Fraction of rain intercepted by foliage Rain scavenging ratio Volume fraction of particles in air Volume fraction of particles in water Volume fraction of fish in water Volume fraction of air in soil Volume fraction of water in soil Volume fraction of soil solids Volume fraction of sediment pore water Volume fraction of sediment solids Volume fraction of water in vegetation Organic carbon fraction of particles in water Organic carbon fraction of soil solids Organic carbon fraction of sediment solids Lipid fraction of vegetation Residence time of air Residence time of water Residence time of annual vegetation cycle Vegetation density Water and aquatic biota density Soil solids, sediment solids, atmospheric aerosol and water particle density
1011 3 ¥ 109 12 3 1000 0.05 0.03 80 4.0 1.0 0.1 200,000 2 ¥ 10–11 10–5 10–6 0.2 0.3 0.5 0.8 0.2 0.8 0.1 0.02 0.05 0.01 4.17 41.67 180 950 1000 1500
m2 m2 °C m m m m % m2 m–2 kg m–3 – – – – – – – – – – – – – – – days days days kg m–3 kg m–3 kg m–3
– 5 1.5 3 3 3 3 3 3 3 3 3 3 3 3 1.5 1.5 1.5 1.5 1.5 1.5 3 3 3 3 3 3 3 1.5 1.5 1.5
a
95% confidence factors (Cf) were used as convenient expressions of variance where the standard deviation (s) is equal to 0.5 ln Cf (or Cf = e2 s) for a lognormal distribution. A Cf of n implies that 95% of the data will be between n times and 1/n of the median value.
sioneering, Colorado, USA). Lognormal distributions were assumed for all input variables to reduce skewness and were parameterised by the median (m) and standard deviation (s) of the corresponding normal distribution on a log scale. 95% confidence factors (Cf) were used as convenient expressions of variance where s is equal to 0.5 ln Cf for a lognormal distribution. A confidence factor of n implies that 95% of the data will be between n times and 1/n of the median. Input variables were sampled from their 95% confidence factors shown in these tables for 1000 simulations. With this many simulations, the standard error of estimate for the median was always less than 5% of the standard deviation, which was considered adequate.
185
Multimedia Mass Balance Modelling of Two Phthalate Esters Table 4. Mass transfer coefficients used in RPM
Mass Transfer Coefficient (MTC)
Median value (m h–1)
Confidence factor a
Air side air-water MTC Water side air-water MTC Rain rate Aerosol deposition Soil-air phase diffusion MTC Soil water phase diffusion MTC Soil air boundary layer MTC Sediment-water MTC Sediment deposition Sediment resuspension Soil water runoff Soil solids runoff Sediment burial Diffusion to stratosphere Leaching from soil Air side air-vegetation MTC Vegetation water uptake velocity
5.0 0.05 1 ¥ 10–4 10.8 0.02 1 ¥ 10–5 5.0 1.0 ¥ 10–4 5.0 ¥ 10–7 2.0 ¥ 10–7 5.0 ¥ 10–5 1.0 ¥ 10–8 3.0 ¥ 10–7 0.01 1.0 ¥ 10–5 5 8 ¥ 10–4
3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
a
95% confidence factors (Cf) were used as convenient expressions of variance where the standard deviation (s) is equal to 0.5lnCf (or Cf = e2s) for a lognormal distribution. A Cf of n implies that 95% of the data will be between n times and 1/n of the median value.
Crystal Ball™ was used to provide both a sensitivity analysis and a contribution to variance analysis. In a classical sensitivity analysis the objective is to determine the change in an output parameter that results from a fixed change in each input parameter, which can be illustrated with the following equation: S = (DO/O)/(DI/I)
(b)
where S is the sensitivity, O and I are the best estimates for input and output parameters and DO and DI are changes in the output and input parameters. The contribution to variance analysis also assesses the change in an output parameter that results from changes in the input parameters. However, the change in each input parameter is not fixed, since it varies according to the confidence factors that have been assigned to each input parameter. The application of sensitivity and uncertainty analysis to multimedia models is discussed in detail by MacLeod et al. [11].
3 Chemical Input Data The key chemical input data are vapour pressures, solubilities in water, octanolwater partition coefficients (KOW) and reaction half-lives in air, water, soil, sediment and vegetation. Phthalates are in the liquid state at environmental temperature ranges so correction of solid vapour pressures to sub-cooled
186
I.T. Cousins and D. Mackay
liquid vapour pressures is not required. Physical-chemical properties and degradation half-lives have been recently reviewed by Staples et al. [12], Cousins and Mackay [13] and in the chapter of this handbook discussing environmental degradation rates of phthalates. The recommended values presented in the last of these sources are used as inputs to RPM (see Table 5). An adjustment of vapour pressure and aqueous solubility to a mean annual temperature of a temperate geographical region of 12 °C has been included by using Clausius-Clapeyron equations [9]. The enthalpies of phase change for solution (DHSOL) and vaporization (DHVAP) were assumed to be 10,000 and 50,000 J mol–1, respectively. No temperature correction has been applied to KOW or the reaction half-lives. There is a paucity of plant metabolism data on chemical reaction rates on plant surfaces and in plant tissues in the literature and certainly no data available for phthalate esters. It would appear, however, that metabolism is viable in plant tissues because plant cells contain a complement of active enzymes [14]. Furthermore, photodegradation of compounds sorbed to the surface of plants may be rapid as a result of the tendency of leaves to align to achieve maximum exposure to sunlight [14]. Here, it is assumed that the half-life of a chemical in vegetation is intermediate in value – the same as the water half-life. Vegetation reaction rates are seldom measured, but for certain chemicals for which partitioning to vegetation is known to be important (e.g. the phthalate esters), measurement of degradation rates is a research need. For the higher molecular weight phthalates, measured values of KOC are often much higher than predicted from KOW (see the physical-chemical property chapter). For example, the relationship proposed by Seth [15] (KOC = 0.35 KOW) calculates a KOC value for DEHP of 1.9¥107 L kg–1 for KOC, whereas measured values of KOC for DEHP are much lower and of the order of 104 –106 L kg–1 [12]. This overestimation of KOC is usually observed for suspended particles in the water column. Suspended particles may collide inducing desorption, and equilibrium between the particles and the dissolved-phase may not have been achieved [16]. The DEHP simulation has been run with both a measured (105 L kg–1) and a predicted KOC (1.9 ¥ 107 L kg–1) for the suspended sediments in the water column. The KOC for the bottom sediments is fixed at the default predicted value of 1.9 ¥ 107 L kg–1. Mean fish bioconcentration factors for DBP and DEHP of 167 and 280 L kg–1 (wet fish) were taken from Staples et al. [12] and used to estimate regional concentrations in fish, as described earlier.
4 Environmental Emissions Worldwide “production” and “consumption” data have recently been compiled by Parkerton et al. [6] in a comprehensive study on the environmental emissions of phthalates during their life cycle (see Table 6). This study was used as the basis for estimating the environmental emissions used in the modelling calculations.“Production” indicates the amount of a substance that is annually manufactured within a region, whereas “consumption” reflects the amount of the
b
a
median value confidence factor b median value confidence factor b
8.24 2 2.07 ¥ 10–3 2
Log KOW 4.27 1.1 7.73 1.1
VP a (Pa) 1.88 ¥10–3 2 1.00 ¥10–5 2 10,000 1.5 10,000 1.5
DHSOL (J mol–1) 50,000 1.5 50,000 1.5
DHVAP (J mol–1) 55 5 17 5
Air 170 5 550 5
Water
1700 5 5500 5
Soil
5500 5 5500 5
Sediment
Assumed reaction half-lives (h)
170 5 550 5
Vegetation
SW is the solubility of the liquid in water, VP is the liquid vapour pressure. 95% confidence factors (Cf) were used as convenient expressions of variance where the standard deviation (s) is equal to 0.5 ln Cf (or Cf = e2s) for a lognormal distribution. A Cf of n implies that 95% of the data will be between n times and 1/n of the median value.
DEHP
DBP
Phthalate Physical-chemical property ester SW a (mg L–1)
Table 5. Estimated chemical properties of DBP and DEHP at 12 °C
Multimedia Mass Balance Modelling of Two Phthalate Esters
187
188
I.T. Cousins and D. Mackay
Table 6. Summary of emission factors to different environmental compartments at each step
in the life cycle for DBP and DEHP (these unitless fractions are multiplied by production and consumption tonnages to estimate emissions) Phthalate ester
Life stage
Air
Water
Soil
Confidence factor a
DBP
production and industrial use transportation product end use product disposal total production and industrial use transportation product end use product disposal total
0.5 0 23.5 0 24.0 0.25 0 1.0 0 1.25
0.25 0.01 0.095 0.008 0.363 0.015 0.01 0.029 0.002 0.056
0.05 0 0.132 0 0.182 0.005 0 0.065 0 0.07
3 3 3 3 – 3 3 3 3 –
DEHP
a
95% confidence factors (Cf) were used as convenient expressions of variance where the standard deviation (s) is equal to 0.5 ln Cf (or Cf = e2s) for a lognormal distribution. A Cf of n implies that 95% of the data will be between n times and 1/n of the median value.
substance that is converted into end-use products. Differences between total production tonnage and consumption tonnage estimates reflect import/export of these commodity substances into or out of a region to meet worldwide demands. Phthalates can be released to the environment industrially from facilities that produce or process phthalates, and during consumption, as a result of evaporation and possible leaching of phthalates from products during use. Despite the low volatility of the heavier phthalates, their vapour pressures are sufficient to cause evaporation over the long term. This is commonly observed as plastic materials become less flexible with age. In the report by Parkerton et al. [6], the estimated percentage losses of phthalate esters from each stage of their life cycle, from manufacture to waste disposal, have been estimated (reproduced in Table 6). Table 7 gives estimated environmental emissions for DBP and DEHP to the RPM model region from different stages of the phthalate life cycle. First, the emissions to the whole EU were calculated by multiplying the estimated percentage losses from each life stage reported [6] by the EU production tonnage (for losses from production and industrial use and losses from transportation) or EU consumption tonnage (for losses from product end use and product disposal). The total annual production in metric tonnages of DBP and DEHP in the EU are estimated to be 37,000 and 595,000, respectively, whereas the consumption tonnages are estimated to be 21,000 and 476,000, respectively [6]. Confidence factors of 1.1 were applied to these production and consumption data. The estimated total emissions to the EU on a per capita basis were calculated by dividing the total estimated emissions for the region by the total population of the region (370 million for the EU, assumed to be accurate). The emissions to RPM were estimated by multiplying the per capita emissions for the EU by the assumed population of RPM (40 million).
Life stage
production and industrial use transportation product end use product disposal total emission advecting in from outside region emission and advection
production and industrial use transportation product end use product disposal total emission advecting in from outside region emission and advection
Phthalate ester
DBP
DEHP
168,449 0 539,038 0 707,487 11,380 718,867
20,950 0 558,855 0 579,805 28,009 607,814
Emission to air (kg year–1)
10,107 6738 15,632 1078 33,555 1420 34,975
10,475 419 2259 190 12,924 710 13,634
Emission to water (kg year–1)
Table 7. Estimated environmental emissions to the RPM model region (based on EU emission data)
3369 0 35,037 0 38,406 – 38,406
2095 0 3139 0 5234 – 5234
Emission to soil (kg year–1)
181,925 6738 589,707 1078 779,448 12,800 792,248
33,520 419 564,253 190 597,963 28,700 626,682
Total (kg year–1)
Multimedia Mass Balance Modelling of Two Phthalate Esters
189
190
I.T. Cousins and D. Mackay
5 Monitoring Data Used for Model Evaluation Over the last few years, a regional field monitoring program has been conducted in the Netherlands to obtain concentration measurements of selected phthalates in different environmental media for comparison to multimedia fate model predictions. Samples of outdoor air, soil, sediment, vegetation and fish were collected by the RIVM and analysed with state-of-the-art methodologies. These results have been reported by the two contract laboratories that performed these analyses (Alcontrol [17] and the Research Institute of Chromatography [18]). Regional monitoring data for surface water concentrations of DBP and DEHP were taken from the analyses of German rivers reported by Furtmann [19] and Alberti [20]. In each of these studies, considerable precautions were taken to minimize the confounding problems associated with laboratory contamination. Statistical summaries of these data including median concentrations, and 10th and 90th percentile concentrations are given in Table 8. The comprehensive monitoring database developed for the American Chemistry Council (described in detail in the chapter of this handbook discussing environmental concentrations) was used to estimate background concentrations in air and water that are advected into a region (see Table 5 for advection estimates). It was decided to take the 10th percentile concentrations for air and water from this monitoring database as boundary conditions for this modelling exercise (see Table 3 for values). A confidence factor of three was applied to the air and water background concentrations. Table 8. Summary of monitoring data used for model evaluation
Phthalate ester
DBP
DEHP
Medium
air (ng m–3) water (µg L–1) fish (µg kg–1 fat) soil (µg kg–1 dry) sediment (µg kg–1 dry) vegetation (µg kg–1 dry) air (ng m–3) water (µg L–1) fish (µg kg–1 fat) soil (µg kg–1 dry) sediment (µg kg–1 dry) vegetation (µg kg–1 dry)
Regional European monitoring data 10th percentile
Median
90th percentile
5 0.052 <50 5 34 5 1.0 0.087 <50 13 66 82
10.3 0.12 449 12 95 5 16.7 0.46 449 28 250 224
33.9 0.25 1778 30 160 36 85.2 1.3 5566 61 730 484
Multimedia Mass Balance Modelling of Two Phthalate Esters
191
6 Model Results and Discussion The model was run for steady-state conditions for DBP and DEHP; mass balance diagrams are given in Figs. 2 and 3. The majority of environmental releases for all phthalates, including the relatively higher molecular weight DEHP, are to the air compartment. The diagram shows that DEHP partitions from air to vegetation, air to soil and water to sediment and this is likely to be the pattern for other phthalates with long alkyl chains. Furthermore, the lighter DBP also tends to partition out of the air and into the soil, but does not partition as appreciably to bottom sediments. For DBP the uptake by vegetation from the atmosphere is balanced by volatilisation from the plant surface and thus accumulation in vegetation is not as significant as for DEHP. The predicted environmental concentrations are given in Table 9 with error or variation limits (10th and 90th percentiles) corresponding to model uncertainty. Comparisons between fugacities calculated from monitoring data and fugacities calculated from model predicted concentrations are shown in Figs. 4 and 5. For consistency it was decided to use the same methods for converting concentrations into fugacities that were used in the chapter of this handbook discussing environmental concentrations of phthalates. Direct comparison between the con-
Fig. 2. RPM steady-state mass balance for DBP (all fluxes in mol h–1)
192
I.T. Cousins and D. Mackay
Fig. 3. RPM steady-state mass balance for DEHP (all fluxes in mol h–1). (In this simulation a
measured KOC of 105 was used for the suspended particles in the water column)
centration values can be made by comparing Tables 8 and 9. The 10th and 90th percentiles of the monitoring data are approximately a factor of three greater or less than the median predicted concentration. In general, predicted environmental concentrations (or fugacities) compare favourably with the monitoring data, with most of the predictions being within an order of magnitude of observed data. The reduction in fugacity of DEHP from abiotic media, such as air and water, to biotic media, such as vegetation and fish, is thought to be due primarily to metabolism in the biotic media and possibly, in the case of fish, trophic dilution. These findings are consistent with observations made in the chapter of this handbook discussing environmental concentrations of phthalates. For DBP, predicted concentrations in all media are within a factor of four of observed concentrations. Agreement within a factor of four is considered satisfactory in a regional mass balance modelling exercise in which there are often large uncertainties in chemical and environmental model input parameters. For DEHP, air, soil, vegetation and sediment predicted concentrations are within a factor of three of the median concentrations from the two monitoring datasets. Water concentrations of DEHP, are underestimated by an order of magnitude if an estimated KOC is used for the suspended particles in the water column. This suggests that the model does not accurately describe the partitioning and/or transport between bottom sediments and the overlying water. However, the
10th percentile median 90th percentile
10th percentile median 90th percentile
DEHP a
DEHP b
b
4.5 15 45
4.6 15 44
8.4 25 64
Air (ng m–3)
0.04 0.12 0.41
0.011 0.037 0.13
0.017 0.059 0.21
Water (µg L–1)
180 650 2100
5.0 18 63
55 190 670
Fish (µg kg–1 fat)
Using an estimated KOC from the equation KOC = 0.41KOW (0.41 ¥107.73 = 2.2 ¥107 L kg–1). Using a measured KOC of 105 L kg–1.
10th percentile median 90th percentile
DBP
a
Statistic
Phthalate ester
Table 9. RPM predicted environmental concentrations
2.9 12 48
3.1 12 44
1.1 4.2 17
Soil (µg kg–1 dry)
28 150 950
95 440 2000
4.0 31 170
Sediment (µg kg–1 dry)
20 120 810
23 110 490
3.4 14 73
Vegetation (µg kg–1 dry)
Multimedia Mass Balance Modelling of Two Phthalate Esters
193
194
I.T. Cousins and D. Mackay
Fig. 4. Comparison of RPM predicted and observed median concentrations for DBP (error bars represent 10th and 90th percentiles)
reconciliation between observed and measured water concentrations improves considerable, to within a factor of four, if a measured KOC value is used for suspended particles. The agreement between the predicted and observed fish and sediment concentrations is also much improved if the measured KOC value is used. A validation issue that arises with these, and indeed with all mass balance models is that there may be some cancellation of errors. For example, a concentration may be correctly predicted by using an over-estimate of emission rate and an over-estimate of reactivity (or correspondingly an under-estimate of half-life). The model provides a constraint on the relationship between these variables, for example, in this case the product of emission rate and half-life is known. There are also constraints of “reasonableness” for each variable that tend to narrow the range of possible values. In the present situation the various parameter values selected are judged to be both reasonable and in accord with monitoring data. We do, however, preclude the possibility of some error cancellation. As discussed earlier, sensitivity and contribution of variance analyses have been undertaken by using Crystal Ball™ (Figs. 6 and 7). For DBP, KOW is the most sensitive model input parameter controlling the total environmental inventory, followed by the estimated consumption of DBP in Europe, the emission to air from product end use and the soil reaction half-life. It is noteworthy that the most sensitive model input parameters are chemical parameters rather than environmental parameters. The contribution of variance analysis, which relies on
Multimedia Mass Balance Modelling of Two Phthalate Esters
195
Fig. 5. Comparison of RPM predicted and observed median concentrations for DEHP (error bars represent 10th and 90th percentiles)
our estimated confidence factors ascribed to input parameters, finds that the four most important input parameters controlling output variance of the total environmental inventory of DBP are: soil reaction rate, the emission factor to air from product end use, the average air height and the air reaction rate. In both analyses for DBP, reaction rates and parameters associated with environmental emissions are shown to be the key input parameters controlling the predicted environmental inventory. Similar results were obtained for DEHP. The high sensitivity of KOW, solubility in water, vapour pressure and the enthalpy of phase change for vaporization illustrate the importance of obtaining accurate physicalchemical properties. It is worth noting that given the same model input data other environmental models such as SimpleBox [1], ChemCAN [2] or CalTOX [4] will give similar predictions as RPM. Recently, an intercomparison exercise between SimpleBox 1.0 (used in EUSES) and a Level III fugacity model [21] yielded near-identical predicted environmental concentrations of DEHP from the same set of input parameters. The Level III fugacity model was calibrated by changing the bulk properties of the environmental compartments and mass-transfer coefficients in the model to those used in the EUSES regional environment. Equations used to describe environmental partitioning, however, were not altered in the fugacity model. This agreement between different models is not surprising, since a wide range of multimedia models has previously been shown to give very similar pre-
Fig. 6. Sensitivity analysis for DBP showing the percentage contribution to output variance resulting from a fixed variance in each input parameter
196 I.T. Cousins and D. Mackay
variance in each input parameter
Fig. 7. Contribution to variance analysis for DBP showing the percentage contribution to output variance resulting from estimated
Multimedia Mass Balance Modelling of Two Phthalate Esters
197
198
I.T. Cousins and D. Mackay
dictions when using the same chemical input data and a standardized environment [22]. It is noteworthy that these models are based on a similar set of equations describing partitioning and transport. RPM predicts that when total emissions of DEHP to the environment are 792,000 kg year–1 the total inventory in the environment is 83,000 kg. As a result, the residence time is 83,000/792,000 or 0.10 years (38 days). Likewise, the overall residence time of DBP in the RPM model is 26,000/627,000 or 0.16 years (15 days). Losses due to reaction dominate over losses due to advection out of the model environment. The estimated residence times of DBP and DEHP attributable to reaction only are 25 and 47 days, respectively. DEHP is estimated to reside longer in the environment than DBP because it has longer degradation half-lives in most environmental media and a greater proportion of the more hydrophobic DEHP partitions to soils and sediments in which degradation tends to be slower. These results are consistent with the results of the evaluative fate modelling presented in the physical-chemical property chapter. Mathematically, the calculated residence time in the model is a “characteristic time” describing chemical dynamics in the system. A dynamic model under initial conditions of zero concentrations, followed by sustained constant emissions would display an approach to a steady state that would be essentially 97% complete after three characteristic times. Thus, steady state would be achieved for DBP and DEHP after 45 days (0.12 years) and 114 days (0.31 years), respectively. Similarly, if emissions were stopped the system would be clear of the chemicals almost entirely in the same periods of time. The implication is that, since phthalates have been in use for many decades, they must have reached a steady-state condition in the environment; thus, the use of a steady-state model is justified. The fact that emission rates and observed concentrations were reconciled with a steady-state model is further support for this assertion. Fears that conditions may be getting progressively worse as a result of accumulation of phthalates from past discharges are unfounded. This hypothesis is supported by published monitoring data that have shown that concentrations of phthalate esters in surface water and sediments over the last decade (1990–2000) have remained constant and have reached a steady-state condition (see the chapter discussing environmental concentrations of phthalates for more details). The validity of scaling emissions on a per capita basis should be better tested by applying the RPM model to a number of regions with different population densities. The predicted concentrations are linearly related to population; thus, if the population density is doubled, then so will the predicted concentrations. There are no other datasets that compare in quality to the multimedia monitoring dataset used here for the model evaluation, but it is possible to make a few general observations from the data contained in the database described in the chapter of this handbook discussing the environmental concentrations of phthalates. For example, concentrations of phthalates in Western Europe are 2–3 times higher than concentrations in Northern Europe (i.e. the Scandinavian countries). National average population densities, however, are greater than ten times higher in Western European countries than in the Scandinavian countries. The problem with applying the RPM model to a region such as Scandinavia is that the population is not evenly distributed. People are concentrated in urban
Multimedia Mass Balance Modelling of Two Phthalate Esters
199
areas, and rural populations densities are low. The low transport potential of phthalates is likely to create pronounced urban-rural gradients in environmental concentrations and exposures in such regions. Environmental concentrations in cities should be 2-3 times higher than in the RPM region because of the high population densities found in urban areas (>1000 km–2). Alberta Environment measured surface water concentrations of phthalates in urban and rural areas of Alberta and they found that environmental concentrations in urban areas were on average higher than in rural areas, but usually less than two times higher [23]. Urban areas may act as sources to surrounding rural areas as the air and water in cities will have fairly short residence times. This discussion suggests that although there is certainly a qualitative relationship between environmental concentrations of phthalate esters and population density, this relationship may not be perfectly linear and requires further detailed investigation as more high quality multimedia datasets become available. It would appear that the RPM model is best applied to homogenously and densely populated regions in Western Europe and the Eastern United States.
7 Conclusions and Recommendations The RPM model has successfully reconciled the known properties of two phthalate esters with the latest emission rate estimates and monitoring data. Reported concentration ranges are within a factor of four of the median predicted concentrations. There is a caveat that there has been some cancellation of errors within the model to give a good agreement between predictions and monitoring data, but such errors are believed to be relatively small in magnitude given the constraints between the various parameters. The sensitivity and uncertainty analysis suggests that the major uncertainties are the degradation half-lives and emission rates. Improved agreement between predicted and observed water, sediment and fish concentrations of DEHP is obtained by using a measured value of KOC for the suspended particles in the water column. The residence time of DBP and DEHP including advective losses are 15 and 38 days, respectively; thus, accumulation of these chemicals over periods of years and decades is unlikely. The residence times attributable to reaction only are 25 and 47 days, respectively. It is concluded that, in general, the model captures the key processes that control their behaviour and predicts concentrations of the correct order of magnitude. We are thus close to having an adequate quantitative description of the environmental fate of these chemicals. The following are recommendations for further research: – It would be advisable to experimentally determine the degradation rate in plants because a significant fraction of DEHP partitions to vegetation and potentially degrades there. – RPM should be applied to other regions with different population densities. The further application of the RPM model requires that further high quality multimedia monitoring surveys are undertaken. – More phthalate esters and other chemicals used on a per capita basis should be modelled by the RPM model.
200
I.T. Cousins and D. Mackay
Acknowledgement. We are grateful to the Phthalate Ester Panel of the American Chemistry Council (ACC) for funding this research, and to NSERC and the consortium of chemical companies that support the Canadian Environmental Modelling Centre.
8 References 1. McKone TE (1993) CalTOX,A multi-media total-exposure model for hazardous waste sites, Part II: the dynamic multi-media transport and transformation model. Lawrence Livermore National Laboratory, Livermore, CA, No UCRL-CR-111456 PtII 2. Mackay D, Paterson S, Tam DD (1991) Assessments of chemical fate in Canada: continued development of a fugacity model. A report prepared for Health and Welfare Canada 3. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:1618 4. Meent D Van De (1993) SimpleBox: a generic multimedia fate evaluation model. Bilthoven, National Institute of Public Health and Environmental Protection. Report No 672720001 5. European Chemical Bureau (ECB) (1997) EUSES documentation – the European Union System for the Evaluation of Substances. National Institute of Public Health and Environment (RIVM), the Netherlands, available from European Chemicals Bureau (EC/DGXI), Ispra 6. Parkerton T, Konkel W (2000) Evaluation of the production, consumption, end use and potential emissions of phthalate esters. Prepared for the American Chemistry Council (Draft of November 2000) by Exxon Mobil Biomedical Sciences Inc (EMBSI), East Millstone, NJ, USA 7. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:1627 8. Mackay D, Di Guardo A, Paterson S, Cowan C (1996) Environ Toxicol Chem 15:1638 9. Mackay D (2001) Multimedia models: the fugacity approach, 2nd edn. Lewis, Boca Raton 10. Cousins IT, Mackay D (2001) Chemosphere 44:643 11. MacLeod M, Fraser A, Mackay D (2002) Environ Toxicol Chem 21:700–709 12. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:667 13. Cousins IT, Mackay D (2000) Chemosphere 41:1389 14. McFarlane JC (1995) In: Trapp S, McFarlane JC (eds) Plant contamination – modeling and simulation of organic chemical processes. Lewis, Boca Raton, FL, pp 13–34 15. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:2390 16. Williams MD, Adams WJ, Parkerton TF, Biddinger GR, Robillard KA (1995) Environ Toxicol Chem 14:1477 17. Alberti J, Brull U, Furtmann K, Braun G (2000) Occurrence of phthalates in German surface and wastewater. Poster presentation at the 10th annual SETAC Europe meeting, Brighton, UK, 21–25 May 18. Alcontrol Biochem Laboratoria (1999) The analysis of phthalates in soil and sediment. Hoogvliet, The Netherlands 19. Furtmann K (1993) Phthalates in the aquatic environment. PhD dissertation, Regional Water and Waste Water Authority, Nordrhein-Westfalen, Düsseldorf, Germany 20. Research Institute of Chromatography (2001) Final report on analysis of selected phthalates in the Dutch environment, Report ECPI-2001-XX. Kortrik, Belgium 21. Cousins IT, Mackay D (2000) Review of EUSES modelling for di-2-ethylhexyl phthalate (DEHP). Final report prepared for the European Chemical Industry Council (CEFIC),April 2001, CEMC Report No 200101 22. Cowan CE, Mackay D, Feijtel TCJ, van de Meent D, Di Guardo A, Davies J, Mackay N (1995) The multi-media model: a vital tool for predicting the fate of chemicals. Proceedings of a workshop organized by the Society of Environmental Toxicology and Chemistry (SETAC). Based on an international task force which addressed the application of multi-media fate models to regulatory decision-making, held at Leuven, Belgium, April 14–16, 1994 and Denver, Colorado, November 4–5, 1994 23. Alberta Environment (1999) Data analysed for phthalates in surface water in Alberta. Water Sciences Branch, Water Data Management Section. Edmonton, AB, Canada
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 201– 225 DOI 10.1007/b11467
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs Frank A.P. C. Gobas 1 · Cheryl E. Mackintosh 1 · Glenys Webster 1 Michael Ikonomou 2 · Thomas F. Parkerton 3 · Kenneth Robillard 4 1 2 3 4
School of Resource and Environmental Management, Simon Fraser University, Burnaby, British Columbia, V5A 1S6, Canada. E-mail:
[email protected] Department of Fisheries and Oceans, Contaminants Science Section, Institute of Ocean Sciences, Sidney, British Columbia, V8L 4B2, Canada Exxon Mobil Biomedical Sciences, Inc., Machelen, Belgium Eastman Kodak Company, 1100 Ridgeway Ave., Rochester, NY 14652-6276, USA
This chapter explores the bioaccumulation behavior of several phthalate esters in aquatic foodwebs. It includes: (i) a compilation of bioconcentration data from reported laboratory studies in the literature, (ii) an overview and discussion of the results from a recently completed foodweb bioaccumulation field study, and (iii) an analysis of the results of a bioaccumulation modeling study. The study concludes that laboratory and field studies indicate that phthalate esters do not biomagnify in aquatic food-webs. Higher molecular weight phthalate esters (DEHP, DnOP, and DnNP) show evidence of trophic dilution in aquatic food-webs, which is consistent with findings from laboratory and modeling studies which indicate that metabolic transformation is a key mitigating factor. Bioaccumulation patterns of DBP, DiBP, and BBP indicate no significant relationship with trophic position consistent with a lipid-water partitioning model. The lowest molecular weight phthalate esters (DMP and DEP) show bioaccumulation factors in laboratory and field studies that are greater than predicted from a lipid-water partitioning model. The considerable variability in the field-derived bioaccumulation factors (BAFs) for lower molecular weight phthalate esters across aquatic species suggests that species-specific differences in metabolic transformation can have a significant effect on observed bioaccumulation. With some exceptions discussed below, the bioconcentration and bioaccumulation factors of the phthalate esters discussed in this paper are below the UNEP bioaccumulation criterion of 5000. The low bioavailability of the high-molecular weight phthalate esters in natural waters is the main reason why the BAFs of the higher molecular weight phthalate esters are below the UNEP bioaccumulation criterion. Since the intention of the bioaccumulation criteria is to identify substances as being “bioaccumulative”, if they (like PCBs) biomagnify in the foodweb then current evidence supports the conclusion that phthalate esters do not appear to be “bioaccumulative”. Keywords. Bioaccumulation, Biomagnification, Phthalate Esters, Aquatic Food-Webs, Fish
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
1
Introduction
2
Bioaccumulation Nomenclature . . . . . . . . . . . . . . . . . . . 203
3
Phthalate Ester Nomenclature . . . . . . . . . . . . . . . . . . . . 204
4
Bioconcentration Studies
5
Dietary Bioaccumulation Studies . . . . . . . . . . . . . . . . . . 207
6
Bioaccumulation Studies from Sediments
. . . . . . . . . . . . . . . . . . . . . . 204
. . . . . . . . . . . . . 208
© Springer-Verlag Berlin Heidelberg 2003
202
F.A.P.C. Gobas et al.
7
Food-Web Bioaccumulation Studies . . . . . . . . . . . . . . . . . 208
7.1 7.2 7.3
Biota-Sediment-Accumulation . . . . . . . . . . . . . . . . . . . . 208 Food-Web Bioaccumulation . . . . . . . . . . . . . . . . . . . . . 212 BAFs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
8
Bioaccumulation Models . . . . . . . . . . . . . . . . . . . . . . . 217
9
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Abbreviations BAFs BBP BCFs BSAF DBP DEHP DEP DiBP DMP DnNP DnOP LRTAP MS-MS detection PCBs POPs QA/QC UNEP
Bioaccumulation Factors Butylbenzyl Phthalate Bioconcentration Factors Biota-Sediment Accumulation Factor Di-n-Butyl Phthalate Di-2(Ethylhexyl) Phthalate Diethyl Phthalate Diisobutyl Phthalate Dimethyl Phthalate Di-n-Nonyl Phthalate Di-n-Octyl Phthalate Long-Range Transboundary Air Pollution Protocol Mass Spectrometry-Mass Spectrometry detection Polychlorinated Biphenyls Persistent Organic Pollutants Quality Assurance/Quality Control United Nations Environmental Program
1 Introduction Dialkyl phthalate esters are hydrophobic substances with octanol-water partition coefficients ranging between 101.61 for dimethyl phthalate esters to values exceeding 108 for congeners like diundecyl phthalate ester and ditridecyl phthalate ester [1]. Due to their hydrophobicity, phthalate esters are often believed to have a high potential to bioconcentrate and bioaccumulate in aquatic organisms. The degree of bioaccumulation and the mechanism by which phthalate esters are absorbed and retained by aquatic organisms is of considerable importance as the United Nations Environmental Program (UNEP) long-range transboundary air pollution protocol (LRTAP) on POPs as well as domestic legislation in Canada (Canadian Environmental Protection Act, 1999) and several countries [2] aim to eliminate substances from commerce that are bioaccumulative, persistent, and toxic. The bioaccumulation criterion identifies chemicals as “bioaccumula-
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
203
tive” if they exhibit bioaccumulation or bioconcentration factors (BAFs or BCFs) greater than 5000 in aquatic organisms. In absence of BAF or BCF data “bioaccumulative” substances are defined as compounds with octanol-water partition coefficients (KOWs) greater than 105. The intent of these legislative efforts is to identify substances that biomagnify in aquatic food-webs. Biomagnification is the process in which the lipid-normalized concentration of the chemical increases with each step in the food-web. The significance of biomagnification is that organisms at the top of the food-web are exposed to chemical concentrations that are greater than those at lower trophic levels. The scientific rationale for the bioaccumulation criterion is based on findings for persistent organochlorines, such as PCBs and chlorobenzenes, which indicate that persistent substances biomagnify in aquatic food-webs if laboratory-derived bioconcentration factors exceed approximately 5000 or the octanol-water partition coefficient of the substance exceeds approximately 105. Several authors have suggested that the bioaccumulation behavior of phthalate esters is not comparable to that of persistent organochlorines such as PCBs. Laboratory studies, which in most cases involved one particular phthalate ester (i.e., diethylhexyl phthalate ester), have pointed out that the bioaccumulation factors of phthalate esters are typically less than expected from their lipid-water partitioning properties [3]. Metabolism and a reduced bioavailability of phthalate esters have been proposed to be the main factors causing the lower than expected bioaccumulation factors of phthalate esters [3–8]. However, field studies to confirm this hypothesis have not previously been reported. In this chapter, we will explore the bioaccumulation behavior of several phthalate esters in aquatic food-webs. We will present a compilation of bioaccumulation data from reported laboratory studies in the literature and from a recently completed bioaccumulation field study that we conducted in a marine food-web. The objective of our analysis is to gain insights into the mechanisms of phthalate ester uptake, elimination, and bioaccumulation in aquatic foodwebs. This information can be useful in assessing the bioaccumulative potential of this group of ubiquitous and widely used substances relative to other chemical classes.
2 Bioaccumulation Nomenclature Bioconcentration is defined as the process in which an aquatic organism achieves a concentration level that exceeds that in the surrounding water as a result of exposure of the organism via the respiratory surface and the skin [9]. Bioconcentration refers to a condition, usually achieved under laboratory conditions, in which the organism is exposed to a chemical substance in the water, but not in its diet. The underlying mechanism of this process is the lipid-water partitioning property of the substance [10]. A number of depuration processes including egestion in fecal matter, deposition in eggs, growth, and metabolic transformation can interfere with the lipid-water partitioning behavior of the chemical substance, typically resulting in bioconcentration factors that are less than the corresponding lipid-water partition coefficient [9].
204
F.A.P.C. Gobas et al.
Bioaccumulation refers to the process by which the chemical concentration in an aquatic organism achieves a level that exceeds that in the water as a result of chemical uptake through all routes of chemical exposure (e.g., dietary absorption, transport across the respiratory surface, dermal absorption). Bioaccumulation takes place under field conditions. It is a combination of chemical bioconcentration and biomagnification [9]. Biomagnification refers to the process by which the chemical concentration in the predator exceeds that in the prey organisms it consumes. In this chapter biomagnification refers to field conditions where the predator and prey organisms are simultaneously exposed to chemical via both water and diet. Biomagnification has been observed in aquatic and terrestrial food-webs in the field [11, 12] and under laboratory conditions and mechanisms for this process have been postulated [13, 14]. Food-web bioaccumulation is the process by which the chemical concentration in organisms increases with increasing trophic level. It is the result of a sequential series of biomagnification events. Trophic dilution is the process by which the chemical concentration in organisms drops with increasing trophic level. It occurs when the chemical concentration in the predator remains below the concentration in the prey, typically as a result of metabolic transformation of the chemical in the predator.
3 Phthalate Ester Nomenclature Phthalate esters include a large number of substances that share a common chemical core structure. The phthalate esters discussed in this document and their chemical structures and acronyms are listed in Table 1.
4 Bioconcentration Studies Several laboratory studies have investigated the bioconcentration of phthalate esters in various fish species, algae, macrophytes, polychaetes, molluscs, crustacean, aquatic insects, and other organisms. The data reported in these studies have been compiled and reviewed in Staples et al. [3]. When evaluating the results from laboratory bioconcentration studies, it is important to recognize some of the characteristics and experimental artifacts of bioconcentration studies for phthalate esters. First, the majority of reported bioconcentration studies involve only one particular phthalate ester, that is, DEHP (Table 1). Bioconcentration data for other phthalate esters are scarce, causing much of the experimental evidence on the bioaccumulation of phthalate esters to rely on observations for a single congener. Secondly, the majority of the reported studies use radiolabeled phthalate ester congeners. Because phthalate esters can be subject to metabolic transformation in organisms, BCFs based on total radioactivity (i.e., radioactivity from the parent substance and its metabolites) can be greater than BCFs based on radioactivity of the parent (i.e., unmetabolized) compound alone. BCFs determined with the use of radioactive test sub-
Table 1. Chemical structure and acronyms of phthalate esters
Phthalate Ester
Abbreviation
Dimethyl Phthalate
DMP
Diethyl Phthalate
DEP
Di-iso-butyl Phthalate
DiBP
Di-n-Butyl Phthalate
DBP
Butyl Benzyl Phthalate
BBP
Di(2-Ethylhexyl) Phthalate
DEHP
Di-n-Octyl Phthalate
DnOP
Di-n-Nonyl Phthalate
DnNP
Chemical Structure
206
F.A.P.C. Gobas et al.
stances may therefore overestimate the BCF of the parent substance. Third, the aqueous solubility of especially the higher molecular weight phthalate esters is low and the water concentrations used in some of the bioconcentration tests significantly exceed the aqueous solubility. Results from these bioconcentration tests are difficult to interpret. On one hand, water concentrations above the aqueous solubility indicate that a considerable fraction of the total chemical concentration in the water is not available for uptake via the respiratory surface. On the other hand, phthalate esters can form emulsions when concentrations are in excess of the water solubility due to their surface-active properties. These emulsions can create micelles that may adhere to the outer surface of the organism and which may be also ingested by organisms. As a result, it is unclear to what degree BCFs determined at concentrations above the water solubility are representative of field conditions. Fourth, water concentrations of phthalate ester that are constant over the exposure duration are typically not achieved in the bioconcentration tests due to the low phthalate ester concentrations in the water, rapid absorption by fish, and degradation in the water phase. Ignoring this experimental artifact in deriving the BCF can lead to an underestimate of the BCF, especially when nominal or initial water concentrations are used to derive the BCF or uptake rate constants [15]. Fifth, the exposure duration in most bioconcentration tests is relatively short and typically much shorter than exposure conditions in the field. For example, a number of bioconcentration tests have been conducted over a period of one day or less while a 28 day period is generally recommended for bioconcentration studies [16]. Experiments that use short exposure times have a tendency to underestimate the actual BCF, since steady-state conditions may not be achieved. To eliminate some of the experimental artifacts of laboratory bioconcentration tests in the analysis of reported bioconcentration data, we plotted BCFs in aquatic macrophytes, algae, benthic invertebrates, and fish (Fig. 1), and then eliminated BCF data determined under conditions in which (i) the water concentration exceeded the water solubility and (ii) the exposure time was less than three days. The remaining BCF data are presented as a function the chemical’s octanolwater partition coefficient in Fig. 1. Figure 1 also shows the BCF expected if phthalate esters simply partition between the water and lipids of the organisms. A 5% lipid content is assumed. Figure 1 illustrates a number of characteristics of the bioaccumulation behavior of phthalate esters. First, despite the availability of a large number of experimental BCF data, there are few data that meet basic data quality criteria. This illustrates the experimental difficulties of measuring BCFs for phthalate esters. Secondly, reported BCFs for individual phthalate esters exhibit a large variability. This variability has also been noticed by other authors. For example, Karara and Hayton [17–18] report BCFs for DEHP in 1-5 g Sheepshead Minnow (Cyprinodon variegatus) of 6–637 L kg-1 wet weight within a temperature range of 10–23 °C. Thirdly, the reported BCFs of the higher molecular weight phthalate esters are below those expected from lipid-water partitioning. This has been explained by metabolic transformation of phthalate esters [4] and by a reduced bioavailability of phthalate esters in the bioconcentration tests [5–8, 20, 21]. Fourth, the BCFs for DMP and DEP are approximately an order of magnitude greater than expected from simple lipid-to-water partitioning.
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
207
Fig. 1. Laboratory-derived bioconcentration factors (BCFs) of parent () and total (i.e.,
parent phthalate ester and metabolites), (+) phthalate esters in various aquatic organisms. The solid line represents the lipid-water partitioning assuming a 5% lipid content (i.e., BCF= 0.05 KOW)
Fifth, when bioconcentration factors of individual congeners are compared between different taxa, it appears that bioconcentration factors for the higher molecular weight phthalate esters in benthic organisms are greater than those in fish. The latter has also been observed by Staples et al. [3] and Wofford et al. [21] who explained these observations by inter-species differences in metabolic transformation rates. Finally, the bioconcentration factors are generally less than 5000.
5 Dietary Bioaccumulation Studies While the bioconcentration of phthalate esters has been investigated in many studies, the dietary bioaccumulation of phthalate esters has received little attention. Macek et al. [23] examined the dietary transfer of DEHP from daphnids to bluegills (Lepomis macrochirus) and concluded that the contribution of the dietary route to the equilibrium body burden in the bluegill may be small. Gloss and Biddinger [24] investigated dietary transfer in daphnids that were feeding on dihexyl phthalate ester-contaminated algae. Perez et al. [25] suggested that dietary exposure was responsible for seasonal differences in the accumulation of DEHP in marine biota in microcosm studies. Staples et al. [3] conducted theoretical calculations to show that as much as 60% of the DEHP exposure in predators could be derived from the diet. They further argued that the general increase
208
F.A.P.C. Gobas et al.
in the rate of metabolic transformation with increasing trophic level may result in trophic dilution in which organisms at the top of the food-web contain lower concentrations than those in organisms at lower levels.
6 Bioaccumulation Studies from Sediments Uptake and bioaccumulation of sediment-associated phthalate esters has been investigated by Woin and Larson [26] and Brown et al. [27] in dragon flies and chironomid larvae. Based on these data, Staples et al. [3] estimated that the lipid and organic carbon-normalized biota-sediment-accumulation factor (BSAF) for DEHP is 0.1 kg organic carbon (OC) kg–1 lipid in dragonflies and 0.5 kg organic carbon kg–1 lipid for DEHP and diisodecyl phthalate ester in chironomids. They concluded that these values are lower than the theoretical BSAFs based on an equilibrium partitioning value of 1.0 kg organic carbon kg–1 lipid [28], and suggested that metabolic transformation is a plausible explanation for this discrepancy.
7 Food-Web Bioaccumulation Studies A field study to assess the food-web bioaccumulation of a range of phthalate esters was recently carried out by our research group. The details of the study can be found in Mackintosh [29]. The study involved the collection and subsequent chemical analysis of phthalate esters in water, sediments, algae, plankton, filter feeders (mussels), deposit feeders, forage fish, benthic feeding fish and predatory fish, and carnivorous water fowl in a marine embayment referred to as False Creek. Table 2 lists the species included in the study. With the exception of the dogfish, all species selected can be considered resident species. Environmental media were collected from three different stations in the embayment with a sampling frequency of three or four samples per site. Since inter-site variability in concentration was not a significant factor, concentration data were reported for all stations combined, representing a sample size of 12 for sediment and water and nine for the biota investigated. The trophic status of the organisms (Table 2) was identified by applying the trophic positioning model by Vander Zanden and Rasmussen [30] to dietary composition data from various studies [31–38]. A conceptual diagram of the food-web is presented in Fig. 2. The study focused on eight individual phthalate esters, that is, DMP, DEP, DiBP, DBP, BBP, DEHP, DnOP, and DnNP. Water concentration measurements identified dissolved and particulate-bound phthalate ester fractions in the water [29]. 7.1 Biota-Sediment-Accumulation
Figure 3 shows the biota-sediment-accumulation factors (BSAFs) of phthalate esters in a range of benthic invertebrate species as a function of the seawatercorrected octanol-water partition coefficient. Octanol-water partition coeffi-
209
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
Table 2. Names of species and their trophic position included in the bioaccumulation field
study Species common name
Species Latin name
Trophic position
Green algae Brown algae Phytoplankton Minnows Shiner perch Pacific staghorn sculpin Cutthroat trout Three spine stickleback Whitespotted greenling Starry flounder Manila clams Blue mussels Pacific oysters Geoduck clams Pile perch Striped seaperch Forage fish Pacific herring Surf smelt Northern anchovy Purple starfish Surf scoters Pacific staghorn sculpin Dungeness crabs Flatfish English sole Starry flounder Whitespotted greenling Spiny dogfish
Enteromorpha intestinalis Nereocystis luetkeana & Fucus gardneri
1.00 1.00 1.00 2.33
Cymatogaster aggregata Leptocottus armatus Salmo clarki clarki Gasterosteus aculeatus Hexogrammos stelleri Platichthys stellatus Tapes philippinarum Mytilus edulis Crassostrea gigas Panope abrupta Rhacochilus vacca Embiotoca lateralis
2.40 2.48 2.48 2.53 3.05 3.25
Clupea harengus pallasi Hypomesus pretiosos pretiosus Engraulis mordax mordax Pisaster ochraccus Melanitta perspicilata Leptocottus armatus Cancer magister Pleuronectes ventulus Platichthys stellatus Hexogrammos stelleri Squalus Acanthias
3.47 3.49 3.51 3.55 3.64 3.81 4.07
cients of phthalate esters in seawater were derived from those measured in freshwater [1] following Xie [39]. The observed BSAFs are shown in relation to the BSAF based on simple organic carbon-lipid partitioning (i.e., BSAF = 1/0.35 = 2.86 kg OC kg–1 lipid) [28, 40–43]. It shows that among the various benthic species, the BSAFs in geoduck clams are the highest. The BSAFs in geoduck clams are the lowest for DMP and then appear to increase with increasing log KOW to values that, with the exception of DEHP, are not statistically different from 2.86. BSAFs for the other benthic species are significantly lower than those in geoduck clams and show a parabolic relationship with KOW with maximum BSAFs for DBP, DiBP, and BBP. Figure 3 illustrates that there is a substantial variability in the BSAFs among benthic organisms. Sediment burying invertebrates like the geoduck clams and Manila clams appear to exhibit higher BSAFs than the invertebrates (e.g., Dungeness crabs) inhabiting surficial sediment. Filter feeding benthic invertebrates such as the mussels and oysters exhibit intermediary BSAFs.
F.A.P.C. Gobas et al.
Fig. 2. A simplified conceptual diagram of the food-web interactions in the False Creek food-web
210
One of the contributing factors to the differences in the observed BSAFs between the species is the chemical disequilibrium that appears to exist between the sediments and the overlying water. This disequilibrium is illustrated in Fig. 4, which shows the observed sediment-water distribution coefficients (expressed in terms of L kg–1 organic carbon) in relation to the sediment-seawater partition coefficients (L kg–1 organic carbon), derived from the seawater corrected octanolwater partition coefficients according to Seth et al. [44]. A sediment-water disequilibrium occurs if the sediment-water distribution coefficient exceeds the chemical’s sediment-water partition coefficient. It can be expressed by the degree
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
211
Fig. 3. Biota-sediment-accumulation factors (BSAFs) in units of kg organic carbon kg–1 lipid of phthalate esters in a range of benthic invertebrate species as a function of the octanol-seawater partition coefficient. The solid line represents the sediment-organism equilibrium partition coefficient (BSAF=2.86)
Fig. 4. Sediment-water distribution coefficients in units of L kg–1 organic carbon in relation to
sediment-seawater partition coefficients (L kg–1 organic carbon), derived as 0.35 KOW according to Seth et al. [44]
212
F.A.P.C. Gobas et al.
to which the observed sediment-water distribution coefficient (KSW) exceeds the sediment-water equilibrium partition coefficient (KSW, EQ), that is, KSW/KSW, EQ . It represents a situation in which the sediments are at a higher concentration than sediment-water partitioning thermodynamics dictates. Figure 4 illustrates that sediment-water disequilibria fall with increasing KOW from values as high as 229,000 for DMP to 41 for DEP and reach a constant value between approximately two and ten for the higher molecular weight phthalate esters. A disequilibrium of ten indicates that the sediment pore water concentration is an order of magnitude greater than the concentration in the overlying water.A value above unity also suggests that the sediments are serving as an exposure source to the water column. From a bioaccumulation perspective, the significance of the apparent disequilibrium between sediments and overlying water is that the degree of direct exposure of the organism to sediments and associated interstitial water versus exposure to the overlying water will have a significant effect on the body burden of the exposed organism. A sediment burying invertebrate (such as geoduck clams) with greater contact to sediments is therefore expected to be exposed to a higher effective concentration than epibenthic organisms that inhabit the epilimnion (e.g., the Dungeness crab), where they are exposed to the overlying water. Differences in an organism’s habitat utilization are therefore expected to be partly responsible for the differences in the BSAFs that are observed. Other factors, such as metabolic transformation, growth dilution, and low dietary uptake efficiencies of phthalate esters may also play a role. 7.2 Food-Web Bioaccumulation
Figure 5 illustrates the relationship between the lipid equivalent concentration (CL in ng g–1 lipid) of phthalate esters and trophic position for the organisms in the False Creek food-web. For fish and shellfish, lipid equivalent concentrations were derived by dividing the wet weight-based concentration CB (ng g–1 wet weight) by the lipid content L (kg lipid kg–1 organism or tissue on a wet weight basis): CL = CB/L
(1)
For algae and plankton, the calculation of the lipid equivalent concentration was conducted as: CL = CB*/(LB* + 0.35 fOC)
(2)
where CB* is the chemical concentration on a dry weight basis, LB* is the lipid content on a dry weight basis (kg lipid kg–1 sample, dry weight), fOC is the organic carbon content (kg OC kg–1 sample, dry weight), and 0.35 is a proportionality constant reflecting the differences in the sorptive capacities between organic carbon and octanol [44]. The reason for the difference in the methodology for lipid normalization between algae, plankton, fish, and shellfish is that due to the low lipid content of algae and plankton (i.e., 0.1–0.4%) but high organic carbon content (i.e., fOC = 33–40%) lipids are not the main site for chemical accumulation [45]. The purpose of the lipid normalization is to remove the effect of differences
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
213
in lipid content among organisms of different trophic levels on phthalate ester concentrations. Figure 5 illustrates that there are no statistically significant relationships between the lipid equivalent concentrations and trophic position for DMP, DEP, DiBP, DBP, and BBP. Analysis of covariance shows that lipid equivalent concentrations do not appear to increase or drop significantly (i.e., P >0.05) between trophic levels. This indicates that these phthalate esters do not biomagnify in the food-web. Biomagnification in the food-web is defined as the increase in the lipid equivalent concentration with increasing trophic level. The apparent constancy of the lipid equivalent concentrations with increasing trophic position suggests that the accumulation of these phthalate esters is due to simple water-to-lipid partitioning, which produces approximately equal lipid equivalent concentrations in the various organisms of the food-web. For the higher molecular weight phthalate esters (i.e., DEHP, DnOP, and DnNP) there appears to be a statistically significant drop (i.e., P <0.05) in the lipid equivalent concentration with increasing trophic position. The latter indicates trophic dilution, in which lipid equivalent concentrations in organisms decline with increasing trophic level. The observations indicate that higher trophic level organisms are exposed to lower concentrations of these higher molecular weight phthalate esters than organisms of lower trophic levels. 7.3 BAFs
Figure 6 illustrates the observed relationships between the BAF (expressed in units of L kg–1 equivalent lipid) and KOW for all the species included in the field bioaccumulation study. To simplify Fig. 6, the species are grouped into trophic guilds. For the purpose of this analysis, we distinguished between algae, plankton, benthic invertebrates, small forage fish, predatory fish, and aquatic birds. To provide a basis for comparison, Fig. 6 also presents the expected BAFs assuming that only simple lipid-water partitioning of the chemicals between the organisms and the water controls bioaccumulation [10], that is, BCF (L kg–1 equivalent lipid) = KOW. This simple model ignores the potential role of dietary uptake, biomagnification, metabolism, growth dilution, and the reduction of the chemical bioavailability due to sorption in the water phase. Figure 6 illustrates a number of characteristics of the bioaccumulation behavior of phthalate esters in the field. Firstly, BAFs of individual phthalate esters exhibit a considerable variability. The variability in BAFs ranges from a factor of approximately 30 for the lower molecular weight phthalate esters to a factor of 1000 for DEHP, DnOP, and DnNP. There appears to be no apparent relationship between the BAF and the trophic position of the organism for the lower molecular weight phthalate esters. This indicates that the observed variability in the BAF is not due to differences in trophic position among the organisms sampled. However, for the higher molecular weight phthalate esters, there appears to be a trend for the BAFs to drop with increasing trophic position. This trend is the main reason that the BAFs of the higher molecular weight phthalate esters show a greater variability than the BAFs of the lower molecular weight phthalate esters.
214
F.A.P.C. Gobas et al.
Fig. 5. Relationship between the lipid equivalent concentrations of phthalate esters and trophic
position for a range of organisms in a marine food-web
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
215
Fig. 6. Relationships between the observed BAF, expressed relative to the total concentration
in the water, in units of L kg–1 equivalent lipid in algae, plankton, benthos, small fish, predatory fish, and birds, and the octanol-seawater partition coefficient for phthalate esters. The dashed line represents the Canadian Environmental Protection Act’s bioaccumulation criterion expressed on a lipid-normalized basis assuming a 5% lipid content
Secondly, the comparison of the observed BAFs with the BCFs derived by simple lipid-water partitioning shows that the BAFs for DMP and DEP are greater than expected from simple lipid-to-water partitioning. In particular, the BAF of DMP is much greater (i.e., on average a factor of 100) than that expected based on lipid-water partitioning. Considering that (i) all biota and water concentration exceed the method detection limits by many fold, (ii) DMP showed high extraction recoveries, negligible degradation in and evaporative losses from the water samples (due to immediate analysis at low temperature), and (iii) positive MS-MS confirmation in water and biota samples, it is unlikely that the much higher than expected BAFs are due to analytical error. Also, due to the mass-specific analysis, metabolites can be ruled out as a factor. A possible explanation for the high BAFs for DMP is the large disequilibrium between sediments and overlying water. Contact of organisms with the sediments (e.g., sculpins burying in sediments) may elevate the body burden of DMP in organisms over that absorbed from the overlying water. Thirdly, the BAFs of DBP, DiBP, and BBP are generally in reasonable agreement with the BCFs based on simple lipid-water partitioning. This suggests that the bioaccumulation of these substances is mainly the result of chemical exchange between the organism and the water via the respiratory surface of the organisms. Dietary uptake, metabolic transformation, and growth dilution appear to play a
216
F.A.P.C. Gobas et al.
secondary role, but may contribute to the variability in the observed BAFs among the different organisms. The BAFs of the higher molecular weight phthalate esters (i.e., DEHP, DOP, DnNP) are lower than anticipated based on lipid-water partitioning. The low bioavailability of the total water concentration is a key factor causing the BAFs of the higher molecular weight phthalate esters to fall below the lipid-water partition coefficients. For example, our study suggests that approximately 0.1% of the total water concentration of DEHP is freely dissolved and hence assumed to be available for uptake via the respiratory surface. The freely dissolved water concentration is believed to represent the phthalate ester concentration that can be absorbed by organisms via the respiratory surface area as a result of lipid-water partitioning. Figure 7 illustrates the BAF based on freely dissolved concentrations BAFfd (L kg–1 lipid) normalized to the lipid-water partition coefficient, that is, BAFfd/KOW. It shows that the BAFs in plankton and green algae approach the lipid-water partition coefficients, that is, BAFfd/KOW is approximately 1.0. This result appears to be reasonable as algae can be expected to lack a metabolic transformation capability for lipid-like molecules such as
Fig. 7. Relationship between the observed BAFs, expressed in terms of the freely dissolved wa-
ter concentration, in units of L kg–1 equivalent lipid, divided by the octanol-seawater partition coefficient (KOW) in algae (shaded triangles), plankton (), benthos (+), small fish (shaded circles), predatory fish (), and birds (–) for phthalate esters. The solid line represents the organism-water equilibrium partition coefficient (BAFL = KOW)
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
217
phthalate esters. The BAFs of the higher molecular weight phthalate esters in the majority of benthic invertebrates, forage fish, predatory fish, and birds are less than the lipid-water partition coefficients (i.e., BAFfd /KOW < 1), indicating that bioaccumulation factors for the freely dissolved phthalate ester are significantly less than their octanol-water partition coefficients. This suggests that processes other than lipid-water partitioning (e.g., metabolic transformation, growth, fecal excretion) have a significant effect on the BAF. Fourthly, Fig. 6 further shows that when the BAFs are compared to the Canadian Environmental Protection Act’s cut-off value of 5000 (if the BAF is expressed on a wet weight) or a 100,000 (if the BAF is expressed on a lipid weight basis), the BAFs of all the phthalate esters generally fall below the cut-off value. The only exceptions are the BAFs of BBP for green algae, plankton, geoduck clams, striped seaperch, pile perch, staghorn sculpins, and surfscoters.
8 Bioaccumulation Models Bioaccumulation models can be useful tools in the investigation of the mechanism of phthalate ester bioaccumulation. The merit of such models is in investigating the role that uptake and elimination processes contribute to the observed BAFs. For example, these models can be used to estimate whether a chemical substance is predominantly absorbed by an organism from the water via the respiratory surface or through the dietary route. These models can also be applied to assess to what degree chemical substances can be expected to be eliminated via water or feces and to what degree metabolic transformation and growth affect tissue concentrations. Most importantly, the model can be used to test hypotheses regarding the mechanisms contributing to the bioaccumulation process. The hypothesis can be tested by comparing model predictions to observed data. In this section, we will discuss a general bioaccumulation model [46] for fish and test the model against the data from the field bioaccumulation study. The model esti-
Table 3. Model equations, parameters, and their units of the fish bioaccumulation model of
Gobas [13]. BAF = CF/CW = [k1 fDW + kD CD/CW)]/(k2 + kE + kM + kG) Parameter Units
fDW CD CF CW k1 k2 kD kE kG kM
fraction g kg–1 wet weight g kg–1 wet weight g L–1 L water kg–1 organism wet weight d–1 d–1 kg food kg–1 organism wet weight d–1 d–1 d–1 d–1
Definition Fraction of the water concentration that is freely dissolved Chemical concentration in diet Chemical concentration in fish Total chemical concentration in overlying water Uptake clearance rate from water Elimination rate constant from fish Dietary uptake clearance rate Fecal egestion elimination rate constant from fish Growth dilution rate constant Metabolic transformation rate constant from fish
Log KOW CW
1.78 2.74 4.52 4.52 4.98 8.08 8.08 8.98
Dogfish Phthalate ester
DMP DEP DBP DiBP BBP DEHP DnOP DnNP
a
1.78 2.74 4.52 4.52 4.98 8.08 8.08 8.98
DMP DEP DBP DiBP BBP DEHP DnOP DnNP
2430 16,900 3480 82,700 12,800 114,000 9800 15,100
CD
2000 11,500 32,800 2750 11,400 131,000 8700 26,900
CD
100 99 75.4 75.4 53.3 0.11 0.11 0.01
fDW (%)
100 99 75.4 75.4 53.3 0.11 0.11 0.01
fDW (%)
21.4 48.1 56.7 56.7 56.8 56.9 56.9 56.9
k1
83 188 221 221 222 222 222 222
k1
2.37 0.584 0.0114 0.0114 0.00397 0.00000316 0.00000316 0.000000397
k2
27.7 6.83 0.237 0.134 0.0464 0.0000369 0.0000369 0.00000465
k2
0.0148 0.0148 0.0148 0.0148 0.0147 0.00392 0.00392 0.000642
kD
0.0246 0.0246 0.0246 0.0246 0.0246 0.00653 0.00653 0.00107
kD
0.00370 0.00370 0.00367 0.00369 0.00369 0.000980 0.000980 0.000161
kE
0.00616 0.00616 0.00616 0.00615 0.00614 0.00163 0.00163 0.000268
kE
0 0 0.17 0.09 0.04 0.017 0.052 0.03
kM
0 0 0 0 0 0.0011 0.0052 0.0001
kM
0.0011 0.0011 0.0011 0.0011 0.0011 0.0011 0.0011 0.0011
kG
0.0021 0.0021 0.0021 0.0021 0.0021 0.0021 0.0021 0.0021
kG
89 560 189 3240 11,800 580 374 28
BAFL, P
70.4 552 14,200 25,300 72,700 16,000 10,000 3100
BAFL, P
9040 891 189 3240 11,800 580 374 28
BAFL, O
19,000 3900 22,000 28,000 204,000 16,000 10,000 3100
BAFL, O
The model input parameters are the octanol-seawater partition coefficient KOW , the observed total water concentration CW (ng L–1), the dietary concentration CD (ng kg–1 wet weight), the freely dissolved fraction of the water concentration fDW (%). The model output parameters are the gill uptake rate constant k1 (L kg–1 wet weight), the gill elimination rate constant k2 (day–1), the dietary uptake rate constant kD (kg food kg–1 wet weight day–1), the fecal egestion rate constant (day–1), the metabolic transformation rate constant kM (day–1), the growth rate constant kG (day–1), the model-predicted lipid equivalent bioaccumulation factor expressed based on the total water concentration BAFL, P (L kg–1 lipid) and the observed lipid equivalent bioaccumulation factor expressed based on the total water concentration BAFL, O (L kg–1 lipid).
3.51 127 110 5.15 3.48 228 12.8 77.7
3.51 127 110 5.15 3.48 228 12.8 77.7
Log KOW CW
Sculpins Phthalate ester
Table 4. Bioaccumulation model input parameters and the results of the model calculations of the lipid-normalized BAFL for several phthalate esters in a field exposed 0.1 kg staghorn sculpin (lipid content is 5.0%) and a 3 kg dogfish (lipid content is 15%) in relation to the observed BAFs a
218 F.A.P.C. Gobas et al.
219
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
mates a whole organism BAF in units of L kg–1 wet weight based on the total water concentration as: BAF = CF /CW = [k1 fDW +kD CD /CW]/(k2 + kE + kM +kG) –1
(3)
The corresponding lipid equivalent BAFL (L kg equivalent lipid) is BAF/L, where L is the lipid content of the fish (kg lipid kg–1 wet weight organism). The model parameters are explained in Table 3. The methods for the calculation of the uptake and elimination rate constants can be found in Gobas [46]. The model calculations are illustrated for several phthalate esters in a 0.1 kg staghorn sculpin (lipid content of 5.0%) and a 3 kg dogfish (lipid content is 15%) at a water temperature of 10 °C. The weight, lipid content, and temperature of the species correspond to the animals that were sampled as part of the field bioaccumulation study. The model input parameters and results are listed in Table 4 and a comparison of model-predicted and observed BAFLs are given in Fig. 8. The model calculations were conducted by using (i) a metabolic transformation rate constant kM of 0 and (ii) a metabolic transformation rate constants that was fitted to produce a perfect agreement between model-predicted and field-observed BAFLs. The latter method is essentially an approximation of the potential magnitude of the metabolic transformation rate constant kM . The model calculations illustrate that the lower molecular weight phthalate esters (i.e., DMP, DEP, DiBP, DBP, and BBP) in sculpin and DMP, DEP, DiBP, and DBP in dogfish are almost exclusively absorbed from the water via the gills. Dietary uptake of these phthalate esters is small compared to the uptake from the water, and model-predicted biomagnification factors are less than 1.0, indicating that biomagnification is not expected to occur. The latter is supported by the results of the bioaccumulation field study, which shows that lipid equivalent concentrations as well as lipid equivalent BAFs do not increase with increasing trophic position. The model calculations further show that in sculpins the lower molecular weight phthalate esters are virtually completely depurated through gill elimination. Growth and fecal egestion do not have a significant effect on the BAF in sculpins. Metabolic transformation rate is not required in the model to explain the observed BAFs in sculpins in the field. This does not mean that metabolic transformation does not occur; only that its rate may be too low to have a significant effect on the BAF. The model calculations show that with the exception of DMP and DEP, the lower molecular weight phthalate esters in dogfish require a substantial metabolic transformation rate to explain the observed BAFs. The model results suggest that metabolic transformation is the main route of depuration of the DBP, DiBP, and BBP in the upper-trophic level dogfish. As a result, the BAFLs of these substances are lower than their KOW . The model calculations show that the exposure of sculpins and dogfish to higher molecular weight phthalate esters (DEHP, DnOP, DnNP) is a combination of both direct exposure to the water and dietary uptake. Dietary uptake appears to be more important than direct uptake from the water. However, the uncertainty in the determination of the freely dissolved water concentrations prevents a more conclusive assessment of whether the diet or the water is the main source of uptake in these fish species. Staples et al. [3] predicted that as much as 60% of the
220
F.A.P.C. Gobas et al.
Fig. 8. Model-predicted lipid-normalized BAFs, expressed relative to the total concentration in the water, in units of L kg–1 lipid in relation to the observed values for staghorn sculpins and spiny dogfish. Grey bars represent model predictions assuming metabolism (see Table 4). Black bars represent model predictions assuming no (kM = 0) metabolism. White bars represent the observed values
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
221
DEHP exposure in predators could be derived from the diet. The model shows that the depuration rates for the higher molecular weight phthalate esters are substantially smaller than those for the lower molecular weight phthalate esters. This means that even low rates of metabolic transformation can have a significant effect on the BAF. The model calculations show that for sculpins only small metabolic transformation rates need to be invoked to explain the observed BAFs in sculpins. Considering the error in the model parameterization and in the observed BAFs, it is unclear whether metabolic transformation has a significant effect on the BAF. Assuming no metabolic transformation, model predicted BAFs in sculpins are in good agreement with the observed values. In dogfish, however, high rates of metabolic transformation need to be used to explain the observed BAFs. The latter suggests that metabolic transformation is an important depuration process in dogfish causing the BAFs to be much lower than the lipid-water partition coefficients of these phthalate esters. Metabolic transformation rates of the higher molecular weight phthalate esters in dogfish may be between 0.02 and 0.05 d–1. These rates are on average about an order of magnitude lower than depuration rates derived from laboratory bioconcentration studies with fish [3]. The lower rate of metabolic transformation may be due to (i) the larger size and higher lipid content of the dogfish compared to the fish used in the laboratory experiments and (ii) the much longer exposure duration in the field compared to that in laboratory tests. The high lipid content and size of the organism may generate a large lipid storage compartment for phthalate esters and reduce the fraction of phthalate in the fish that is available for metabolic transformation compared to that in smaller, less lipid-rich fish. The longer exposure duration in the field-exposed fish is likely to increase the fraction of the total amount of phthalate ester in less accessible “slow” storage compartments. The greater fraction of phthalate ester stored in these less accessible compartments (such as the lipids) may cause a smaller fraction of the total amount of phthalate ester in the fish to be available for metabolic transformation. For DMP and DEP the model underestimates the observed BAFs, which may be explained by the apparent sediment-water disequilibria. The sediment-water disequilibria may cause the exposure concentration of this benthic fish species to exceed that measured in the overlying water. Use of the overlying water concentration can therefore be expected to underestimate the actual BAF in this fish species.
9 Conclusions Currently, there exists considerable information on the bioaccumulation behavior of phthalate esters in aquatic systems. Laboratory experiments, field studies, and mathematical modeling studies have all been carried out. A number of conclusions can be drawn from the information available. Firstly, there is no evidence from laboratory and field bioaccumulation studies to support the hypothesis that phthalate esters biomagnify in aquatic foodwebs. Dietary bioaccumulation studies, sediment bioaccumulation studies in the lab and the field as well as the food-web bioaccumulation study discussed in this
222
F.A.P.C. Gobas et al.
chapter all indicate that food-web bioaccumulation (i.e., the increase of the lipid equivalent concentration with increasing trophic level) does not occur. This indicates that despite their high octanol-water partition coefficients, phthalate esters do not appear to biomagnify in aquatic food-webs. Secondly, it is interesting that the lowest molecular weight phthalate esters DMP and DEP exhibit BAFs that are greater than their lipid-to-water partition coefficients. The results from laboratory bioconcentration experiments show similar results (Fig. 1). These observations suggest that DMP and DEP have a greater bioaccumulation and bioconcentration potential than indicated by their octanol-water partition coefficient. The laboratory observations may be explained by the possible formation of metabolic products of DMP and DEP that, due to the method of detection used, were indistinguishable from the parent compounds. However, due to the more specific detection methodology in the bioaccumulation field study (i.e., MS-MS detection), the formation of metabolic products of DMP and DEP is unlikely to explain the higher-than-expected BAFs. While analytical error of the water concentration of DMP and DEP is a likely cause for the higher-than-expected BAFs, the QA/QC procedures applied indicate that analytical error cannot explain the observations either. The model calculations indicate that DMP and DEP are almost exclusively absorbed from the water via the respiratory surface. The large apparent sediment-to-water disequilibria for DMP and DEP is the most likely explanation of the higher-than-expected BAFs in the field. Thirdly, the lipid equivalent BAFs of DBP, DiBP, and BBP, appear to be fairly uniform among the organisms of the marine food-web investigated in this study (Fig. 5). This suggests that the organisms are exposed to a common source and that dietary uptake of phthalate esters has little effect on the BAF of these phthalate esters. Model calculations support this, by demonstrating that direct uptake of these phthalate esters from the water via the respiratory surface of the organisms can be expected to be the main exposure route and that dietary uptake is less important. The general agreement between lipid-normalized BAFs and lipid-water partition coefficients (Fig. 6) also support this conclusion and further indicates that the bioaccumulation of the lower molecular weight phthalate esters generally follows the lipid-water partitioning model. Laboratory bioconcentration studies suggest that the BCFs can reach values up to the lipidwater partition coefficients of these phthalate esters. However, several observed BCFs appear to be lower than the lipid-water partition coefficients. An interesting observation from the laboratory bioconcentration tests and the bioaccumulation field study is that there is a substantial variability in the observed bioconcentration and bioaccumulation factors. While experimental artifacts can be expected to be an important cause of the variability in the observed BCFs, they are an unlikely source of the variability in the BAFs. One potential cause of the observed variability is metabolic transformation. Several authors have implicated metabolic transformation as an important factor controlling the bioaccumulation factors of phthalate esters. The model calculations illustrate that for the metabolic transformation to have a significant effect on the BAFs of the lower molecular weight phthalate esters, the rates of metabolic transformation have to be relatively high. It is possible that metabolic transformation differ among or-
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
223
ganisms and that in certain organisms, metabolic transformation rates are sufficiently large to affect the BAF and cause some of the variability in the observed BAFs among organisms of the food-web. A second cause of the variability in BAFs may relate to the concentration gradients between sediment and water and differences among organisms in their interaction with the sediments and water. Water and sediment concentrations indicate that the sediments may provide higher phthalate esters exposure concentrations than the overlying water. Hence, the interaction of the organisms with the sediments and its pore water is likely to be responsible for some of the variability in the observed BAFs. Fourthly, the BAFs of the higher molecular weight phthalate esters (i.e., DEHP, DnOP, and DnNP) show a tendency to decrease with increasing trophic position. This suggests that organisms at higher trophic levels are exposed to lower phthalate ester concentrations via prey. A similar apparent relationship between BCF and trophic status has been found in laboratory experiments in which BCFs were highest for algae and lowest for fish with invertebrates exhibiting intermediate values [3]. Assessment of the freely dissolved concentrations indicates that the higher molecular weight phthalate esters exhibit a very low bioavailability, that is, only a very small fraction of these phthalate esters in natural waters can be absorbed via the respiratory surface of aquatic organisms. When expressed relative to the freely dissolved concentration in the water, the BAFs in algae and plankton appear to be within an order of magnitude of the lipid-water partition coefficients. This suggests that partitioning is likely an important mechanism for bioaccumulation in algae and plankton. In higher trophic level organisms such as fish, model calculations indicate that dietary uptake is likely to be an important route of exposure as bioavailable concentrations in the water are expected to be very low. The inability of dietary uptake to cause biomagnification is therefore an interesting characteristic of the high-molecular weight phthalate esters in particular and phthalate esters in general. The model calculations provide two possible explanations for this phenomenon. First, it is possible that after these phthalate esters have been absorbed, the phthalate esters are metabolized in the fish. This explanation has been proposed by several authors and supported by the detection of some phthalate ester metabolites [3–8]. A greater rate of metabolic transformation has been suggested to explain the drop in BCFs with increasing trophic level. The other possible explanation is that phthalate esters ingested with the diet are very effectively metabolized in the gastro-intestinal tract even before they are absorbed (i.e., effectively decreasing kD in Eq. (3)). This first-pass effect essentially prevents a significant rate of dietary uptake of the parent phthalate esters. The structural similarity between lipids and phthalate esters may favor such a process as pH and enzymatic conditions in the gastro-intestinal tract are tailor-made for the hydrolysis of lipids and perhaps phthalate esters. The uptake that would still occur is directly from the water. Model calculations illustrate that in the absence of dietary uptake the BAF can be expected to drop with increasing organism size (which correlates well with trophic level) as has been observed in the field study. This is due to the fact that with increasing organism size (and reducing area-to-volume ratio), the gill elimination and fecal egestion rates drop and become negligible compared to growth rates or even small metabolic transformation rates. This results in smaller BCFs for larger organisms. This second
224
F.A.P.C. Gobas et al.
hypothesis does not require the occurrence of a high rate of metabolism in the fish. The bioaccumulation behavior of the lower molecular weight phthalate esters is not consistent with a high rate of metabolism. It is therefore possible that phthalate esters are fairly slowly metabolized after they have been absorbed, but they are effectively metabolized in the gastro-intestinal tract before they are absorbed. We are currently carrying out laboratory experiments to distinguish between these two possible explanations. The toxicological significance of these different mechanisms is that metabolic transformation in organisms has the potential to create metabolic products, while an effective first-pass effect may prevent dietary uptake and the formation of potentially reactive metabolic products within the organism. The majority of observed BAFs for phthalate esters did not exceed the bioaccumulation criterion of 5000 L kg–1 wet weight or 100,000 L kg–1 lipid if expressed on a lipid equivalent basis. Only BAFs of BBP in green algae, plankton, geoduck clams, striped seaperch, pile perch, staghorn sculpins, and surfscoters exceeded the bioaccumulation criterion The results of the field study also confirmed the hypothesis that these substances do not appear to biomagnify in the food-web. Since the intention of the bioaccumulation criteria is to identify substances that, like PCBs, exhibit biomagnification, current evidence in the literature and from our study support the conclusion that phthalate esters do not appear to be bioaccumulative. Acknowledgement. The authors wish to acknowledge the Natural Sciences and Engineering Research of Canada and the American Chemistry Council for sponsoring this research. We further thank the contributions of Audrey Chong, Judy Carlow, Zhongping Lin, Jing Hongwu, and Natatsha Hoover for their contributions to the research.
10 References 1. Cousins I, Mackay D (2000) Chemosphere 41:1389 2. OECD (1998) Harmonized integrated hazard classification system for human health and environmental effects of chemical substances. Organization for Economic Cooperation and Development, Paris 3. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997) Chemosphere 35:667 4. Barron MG, Schultz IR, Hayton WL (1988) Toxol Appl Pharmacol 98:49 5. Hogan JW (1977) In: Johnson BT, Stalling DL, Hogan JW, Schoettger RA (eds) Pollutants in the air and water environments. Wiley, New York, p 292 6. Metcalf RL, Booth GM, Schuth CK, Hansen DJ, Lu PY (1973) Environ Health Perspect June:27 7. Carr KH, Coyle GT, Kimerle RA (1992) 13th annual Society of Environmental Toxicology & Chemistry meeting. Seattle, Washington 8. Barron MG, Albro PW, Hayton WL (1995) Environ Toxicol Chem 14:873 9. Gobas FAPC, Morrison HA (1999) Bioconcentration & bioaccumulation in the aquatic environment. In: Boethling R, Mackay D (eds) Handbook of property estimation methods for chemicals: environmental and health sciences. CRC Press, Boca Raton, p 139 10. Mackay D (1982) Environ Sci Technol 16:274 11. Connolly JP, Pedersen CJ (1988) Environ Sci Technol 22:99 12. Kelly BC, Gobas FAPC (2001) Environ Sci Technol 35:325 13. Gobas FAPC, Zhang X, Wells R (1993) Environ Sci Technol 27:2855
Bioaccumulation of Phthalate Esters in Aquatic Food-Webs
225
14. Gobas FAPC, Wilcockson JWB, Russell RW, Haffner GD (1999) Environ Sci Technol 33: 133 15. Gobas FAPC, Zhang X (1995) Chemosphere 25:1961 16. Organization for Economic Co-operation and Development (1996) Bioaccumulation: flowthrough fish test, 305 E. OECD guideline for testing chemicals 17. Karara AH, Hayton WL (1984) Aquat Toxicol 5:181 18. Karara AH, Hayton WL (1989) Aquat Toxicol 15:27 19. Gobas FAPC, Clark KE, Shiu WY (1989) Environ Toxicol Chem 8:231 20. Mayer FL (1976) Fish Res Board Can 33:2610 21. Boese BL (1984) Can J Fish Aquat Sci 41:1713 22. Wofford H, Wilsey CD, Neff GS, Giam CS, Neff JM (1981) Ecotoxicol Environ Safety 5:202 23. Macek KJ, Petrocelli SR, Sleight BH (1979) Considerations in assessing the potential for and significance of biomagnification of chemical residues. In: Marking LL, Kimerle RA (eds) Aquatic food chains, aquatic toxicology.American Society for Testing and Materials, p 251 24. Gloss SP, Biddinger GR (1985) Comparison and system design and reproducibility to estimate bioconcentration of di-n-hexylphthalate by Daphnia magna. In: Cardwell RD, Purdy R, Bahner RC (eds) Aquatic toxicology and hazard assessment: seventh symposium. American Society for Testing and Materials, Philadelphia, p 202 25. Perez KT, Davey EW, Lackie NF, Morrison GE, Murphy PG, Soper AE, Winslow DL (1985) Environmental assessment of phthalate ester di-(2-ethylhexyl) phthalate derived from a marine microcosm. US Environmental Protection Agency Report, Naragannsett RI, EPA 600/D-85/070 26. Woin P, Larsson P (1987) Bull Environ Contam Toxicol 38:220 27. Brown D, Thompson RS, Stewart KM, Croudace CP, Gillings E (1996) Chemosphere 32:2177 28. DiToro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE, Thomas NA, Paquin PR (1991) Environ Toxicol Chem 10:1541 29. Mackintosh CE (2001) MSci thesis, Simon Fraser University 30. Vander Zanden MJ, Rasmussen JB (1996) Ecological Monographs 66:451 31. Butler TH (1980) Shrimps of the Pacific coast of Canada. Canadian Bulletin of Fisheries and Aquatic Sciences no 202. Department of Fisheries and Oceans 32. Forrester CR (1969) Life history information on some groundfish species. Fisheries Research Board of Canada. Technical Report no 105 33. Hart JL (1973) Pacific fishes of Canada. Fisheries Research Board of Canada. Bulletin 180 34. Jamieson GS, Francis K (eds) (1986) Invertebrate and marine plant fishery resources of British Columbia. Canadian Special Publication of Fisheries and Aquatic Sciences 91. Department of Fisheries and Oceans, Ottawa, Ontario 35. Jones BC (1976) PhD thesis, Simon Fraser University 36. Miller BS (1967) J Fish Res Board Can 24:2515 37. Ricketts EF, Calvin J, Hedgpeth JW, Phillips DW (1985) Between Pacific tides, 5th edn. Stanford University Press, Stanford 38. Vermeer K, Ydenburg RC (1989) Feeding ecology of marine birds in the Strait of Georgia. In: Vermeer K, Butler RW (eds) The ecology and status of marine and shoreline birds in the Strait of Georgia, British Columbia. Canadian Wildlife Service, Ottawa, p 62 39. Mallhot H (1987) Environ Sci Technol 21:1009 40. Shea D (1988) Environ Sci Technol 22:1256 41. Gobas FAPC, Bedard DC, Ciborowski C, Jan JH (1989) J Great Lakes Res 15:581 42. Bierman VJ Jr (1990) Environ Sci Technol 24:1407 43. Parkerton TF (1993) PhD thesis, Rutgers University 44. Seth R, Mackay D, Muncke (1999) Environ Sci Technol 33:2390 45. Swackhamer DL, Skoglund RS (1993) Environ Toxicol Chem 12:831 46. Gobas FAPC (1993) Ecol Modell 69:1
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 227– 262 DOI 10.1007/b11468
Assessment of Critical Exposure Pathways Kathryn Clark 1 · Ian T. Cousins 2 · Donald Mackay 2 1 2
BEC Technologies Inc., 61 Catherine Avenue, Aurora, Ontario, L4G 1K6, Canada E-mail:
[email protected] Canadian Environmental Modelling Centre, Environmental and Resource Studies, Trent University, Peterborough, Ontario, K9J 7B8, Canada
Human exposure to phthalate esters for five different age classes is evaluated for the following routes of exposure: inhalation of air (indoors and outdoors), ingestion of drinking water, incidental ingestion of soil, ingestion of dust (indoors), and ingestion of food. Exposure is estimated for: dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), butylbenzyl phthalate (BBP), and bis(2-ethylhexyl) phthalate (DEHP). For the five phthalate esters evaluated, the median estimated daily intake is highest for toddlers and lowest for infants. For all five phthalates evaluated (except BBP exposure for formula-fed infants), food represents the most important source of exposure. The food categories contributing most to exposure depend upon the phthalate ester and the age group evaluated. Ingestion of dust and inhalation of indoor air represent the most important non-food sources of exposure to phthalate esters. Detection limits have a large influence on the estimated intakes. A comparison of the results of the present study with studies that back-calculate phthalate ester intake from urinary metabolite data suggests that exposure in the present study may be overestimated for DEHP, BBP, and DBP due to changes in food processing over time (many of the measured concentrations of phthalates in food are not recent), loss of phthalates due to cooking has not been accounted for in the present study, and some measured concentrations in food may be elevated due to background contamination. Conversely, exposure to DEP is underestimated in the present study because direct exposure to personal care products is not included. The overestimate of exposure to BBP and DBP from food, referred to above, may be partially cancelled by the lack of inclusion of personal care products. Keywords. Phthalate ester, Human exposure, Probabilistic analysis
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
1
Introduction
1.1 1.2 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5 1.2.6 1.2.7 1.3
Methodology . . . . . . . . . Receptor Characteristics . . . Body Weight . . . . . . . . . . Inhalation Rate . . . . . . . . Food Consumption . . . . . . Drinking Water Consumption Soil Ingestion . . . . . . . . . Dust Ingestion . . . . . . . . Time Spent Indoors . . . . . . Chemical Concentrations . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
229 229 230 230 230 232 233 233 233 233
© Springer-Verlag Berlin Heidelberg 2003
228
K. Clark et al.
2
Dimethyl Phthalate (DMP) . . . . . . . . . . . . . . . . . . . . . . 234
2.1 2.2
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 235 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 236
3
Diethyl Phthalate (DEP) . . . . . . . . . . . . . . . . . . . . . . . 236
3.1 3.2
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 238 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 239
4
Dibutyl Phthalate (DBP) . . . . . . . . . . . . . . . . . . . . . . . 239
4.1 4.2
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 242 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 243
5
Butylbenzyl Phthalate (BBP) . . . . . . . . . . . . . . . . . . . . . 246
5.1 5.2
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 246 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 248
6
Bis(2-Ethylhexyl) Phthalate (DEHP) . . . . . . . . . . . . . . . . . 252
6.1 6.2
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 252 Comparison to Other Studies . . . . . . . . . . . . . . . . . . . . 254
7
Discussion
8
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Abbreviations BBP DBP DEHP DEP DMP
Butylbenzyl phthalate Dibutyl phthalate Bis(2-ethylhexyl) phthalate Diethyl phthalate Dimethyl phthalate
1 Introduction Human exposure to five phthalate esters is evaluated for the following routes of exposure: inhalation of air (indoors and outdoors), ingestion of drinking water, incidental ingestion of soil, ingestion of dust (indoors), and ingestion of food. Food is separated into the following categories: beverages excluding water, milk, cereals, dairy products excluding milk, eggs, fats and oils, fish, fruit products, grains, meats, nuts and beans, poultry, processed meats, vegetable products, “other” foods (includes soup, desserts, snacks), and, for infants, infant formula and breast milk. Exposure to phthalate esters contained in children’s products or other consumer products is not evaluated in this chapter. The following phthalate esters are evaluated: DMP, DEP, DBP, BBP, and DEHP.
Assessment of Critical Exposure Pathways
229
1.1 Methodology
Probabilistic analysis is used to estimate human exposure to each phthalate ester. The physical and behavioral characteristics of receptors in various age classes are first specified. The concentrations of each phthalate ester in each medium are also specified. A spreadsheet containing the appropriate receptor characteristics and chemical concentrations is then prepared and used to calculate the intake of each phthalate in micrograms per kilogram of body weight per day (µg kg–1 d–1). Excel (Microsoft Corporation) and Crystal Ball (Decisioneering Inc.) are then used to generate probability distributions for each exposure estimate, based on a simulation comprising 3000 iterations, and to produce a summary of the results. A sensitivity analysis is performed by Crystal Ball, to identify the input variables contributing most to the variance in the exposure estimates. The following three probability distributions are used in the present study: – the lognormal distribution – the triangular distribution – the uniform distribution Where the lognormal distribution is used, it is positively-skewed (i.e., most values lie closer to the minimum value than the maximum). This distribution is used to define most of the parameters in the present study, including: body weight and inhalation rate of the receptors; rates of food, drinking water, soil, and dust ingestion; and the concentration of each phthalate ester in drinking water and in the various food groups. In most cases, the standard deviation is set equal to 65% of the mean value. This results in a shape with a skewness coefficient of approximately 2.2. A triangular probability distribution is typically used when only the minimum, most likely, and maximum values of a variable are known. The distribution is described by a triangle, with the minimum, most likely, and maximum values forming the vertices. Use of a triangular distribution tends to maximize the variability in the parameter and, in comparison to the lognormal distribution, results in more frequent selection of values in the extremes of the distribution [1].Variables for which a triangular distribution is used include the concentrations of some phthalate esters in indoor and outdoor air and soil. The uniform distribution describes a situation in which any value between a specified minimum and a maximum is equally likely. This distribution is used to describe the amount of time spent indoors (any value between 20 and 24 h d–1 is considered equally probable). 1.2 Receptor Characteristics
The population is divided into five age groups, which is consistent with previous evaluations undertaken by Environment Canada and Health Canada (EC &HC) [2, 3] and a previous evaluation of exposure to DEHP [4]:
230 Adults Teens Children Toddlers Infants
K. Clark et al.
– – – – –
20–70 years of age 12–19 years of age 5–11 years of age 7 months to 4 years of age 0–6 months of age
Information describing the physical characteristics and habits of the Canadian population is obtained primarily from Health and Welfare Canada (HWC) [5]. In most cases this document summarizes only mean or median values, so it is necessary to make assumptions regarding the shape of the probability distributions for a given characteristic. The assumed probability distributions assigned to each characteristic are summarized in Table 1 and are described in the following sections.Although some of the receptor characteristics may be correlated, correlations are not used in the present study. Finley et al. [1] found that correlating interdependent variables had little effect on lifetime risk estimates provided that age-specific distributions were used. 1.2.1 Body Weight
For the adult, teen, and child age groups, mean body weights (males and females combined) are derived from Stephens and Craig [6] who reported 1988 data. Stephens and Craig [6] did not report data for toddlers or infants. For these latter age groups, mean body weights are modified from 1970 Nutrition Canada Survey data (Health Canada, unpublished), adjusting for a 2.9% increase in mean body weight in children between 1970 (Health Canada, unpublished) and 1988 [6]. Standard deviations for toddler and infant body weight are assumed to be equivalent to those measured in 1970. 1.2.2 Inhalation Rate
Lognormal probability distributions are used to describe ranges of inhalation rates for the Canadian population. Recommended air intake rates reported to Health Canada [7] are used for each age group. 1.2.3 Food Consumption
HWC [5] used the results of a Nutrition Canada survey, conducted during 1970 to 1972, as the basis for estimating the rate of ingestion of various individual foods, for the different age groups. The survey involved detailed dietary surveys of over 13,000 individuals, based on a 24-hour recall method. HWC [5] also used information from Statistics Canada to comment on changes in food consumption patterns since the Nutrition Canada survey, in order that appropriate adjustments may be made to the assumed ingestion rates. Important changes in food consumption patterns between 1970 and 1989 include decreases in consumption of whole milk and eggs, and increases in consumption of 2% milk, poultry, and fish.
c
b
a
LN (71; 14) LN (16; 3.9) LN (0.80; 0.52) LN (0.96; 0.62) LN (27; 16) LN (53; 34) LN (32; 21) LN (25; 16) LN (14; 9.0) LN (190; 120) LN (160; 100) LN (95; 61) LN (0.230; 0.15) LN (28; 18) LN (220; 144) LN (21; 14) LN (22; 14) LN (230; 150) NA NA 2300 LN (40; 100) LN (40; 100) U (20; 24)
L d–1 L d–1 g d–1 g d–1 g d–1 g d–1 g d–1 g d–1 g d–1 g d–1 L d–1 g d–1 g d–1 g d–1 g d–1 g d–1 g d–1 L d–1 g d–1 mg d–1 mg d–1
h d–1
Adult 20–70 y
kg m3 d–1
Units
U (20; 24)
LN (1.0; 0.67) LN (0.43; 0.28) LN (24; 15) LN (50; 33) LN (22; 14) LN (29; 19) LN (11; 7.3) LN (160; 100) LN (210; 130) LN (93; 60) LN (0.523; 0.34) LN (31; 20) LN (250; 160) LN (20; 13) LN (23; 15) LN (240; 150) NA NA 2100 LN (40; 100) LN (40; 100)
LN (60; 14) LN (16; 4.0)
Teen 12–19 y
U (20; 24)
LN (1.1; 0.70) LN (0.23; 0.15) LN (34; 22) LN (45; 29) LN (21; 14) LN (21; 14) LN (8.4; 5.5) LN (200; 130) LN (190; 120) LN (55; 36) LN (0.564; 0.37) LN (24; 15) LN (210; 140) LN (17; 11) LN (19; 12) LN (190; 120) NA NA 1800 LN (40; 100) LN (40; 100)
LN (27; 7.3) LN (15; 3.2)
Child 5–11 y
U (20; 24)
LN (0.70; 0.46) LN (0.12; 0.08) LN (42; 27) LN (38; 25) LN (24; 16) LN (11; 7.1) LN (3.4; 2.2) LN (190; 120) LN (90; 58) LN (38; 25) LN (0.632;0.41) LN (15; 9.7) LN (270; 180) LN (13; 8.6) LN (11; 7.0) LN (120; 76) NA NA 1500 LN (40; 100) LN (40; 100)
LN (15; 3.8) LN (9.3;2.6)
Toddler 0.5-4 y
U (20; 24)
LN (0.8; 0.52) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA LN (130; 85) LN (0.750; 0.49) 820 LN (40; 100) LN (40; 100)
LN (7.5; 3.2) LN (2.1; 0.57)
Infant 0–0.5 y
See text for references; NA not applicable; U (minimum; maximum) uniform distribution; LN (mean; standard deviation) lognormal distribution. Standard deviation assumed to be 65% of the mean value. Lognormal distributions for food group consumption and beverage consumption have been truncated with an upper limit of the total maximum consumption of food.
General receptor characteristics Body weight Inhalation rate Receptor ingestion rates b Tap water Beverages Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruits Grains Meats Milk Nuts and beans Other foods Poultry Processed meats Vegetables Infant formula (powder) Breast milk Total max. food consumption c Incidental soil Incidental dust Exposure frequency Time spent indoors
Input parameter
Table1. Receptor characteristics a Assessment of Critical Exposure Pathways
231
232
K. Clark et al.
Although the type of milk consumed has changed, the overall rate of milk consumption has not changed substantially [8]. Similarly, although there have been changes in the type of meat consumed, overall meat consumption has not changed markedly. Infant feeding practices have also changed since the Nutrition Canada survey. In the period from 1965 to 1971, HWC [5] reports that only 25% of mothers breast-fed their babies and more than three quarters of this group had discontinued by one to two months. Data as of 1990 indicate that breast-feeding initiation rates are close to 80%, with 30% still continuing to breast feed at six months [5]. HWC reports that the Nutrition Canada survey likely overestimates the consumption of solid foods by infants due to changes in the age when solid foods are introduced to infants. Solid foods are introduced to approximately 50% of babies by four months, and 89.5% of babies by six months, in contrast to the results of the Nutrition Canada survey in which solid foods are generally introduced to infants’ diets at a much earlier age. HWC recommends that a typical infant should be considered to be exclusively breast-fed up to six months of age and to consume approximately 750 mL of breast milk per day. Alternatively, the typical infant could be considered to be exclusively formula-fed for the first six months. Such is the approach taken in this exposure assessment; for infants up to age six months, two types of exposure are calculated – one for infants consuming breast milk exclusively and the other for infants consuming formula exclusively. Measured phthalate concentrations are available for both ready-to-feed liquid formula and powdered formula. Consistent with a study of phthalate exposure to children, performed by Zaleski et al. [9], the mass of powdered formula is assumed to represent one-seventh of the total mass of formula ingested. Only very limited data are available for “baby food” for DEHP and, as these data fall within the range of available measurements for DEHP in fruit and vegetable products used for all age classes, the data for “baby food” are not used in the exposure assessment. For the present study, composite food groups are developed for all age groups, other than infants. The composite food groups are: beverages excluding water, milk, cereals, dairy products excluding milk, eggs, fats and oils, fish, fruit products, grains, meats, nuts and beans, other foods (includes soups, desserts, snacks), poultry, processed meats, and vegetable products. Mean consumption rates for each age group (summarized in Table 1) are calculated by adding the consumption rates for all foods obtained from [5], in each composite food group. Lognormal distributions are used to define food consumption rates for each age group. The standard deviations are defined as being 65% of the mean values to obtain positively skewed distributions. Maximum values are stipulated for each distribution; maxima are set equal to the age-appropriate total daily food consumption (2300 g d–1 for adults, 2100 g d–1 for teens, 1800 g d–1 for children, 1500 g d–1 for toddlers, and 820 g d–1 for infants) [5,10]. 1.2.4 Drinking Water Consumption
Tap water consumption rates for each age group are determined by calculating age-weighted averages from the data in HWC [5] and subtracting the tea and cof-
Assessment of Critical Exposure Pathways
233
fee consumption rates from the same document (because tea and coffee consumption are included in the “other beverages” food group). The resulting mean values (summarized in Table 1) are used to define positively skewed lognormal distributions, with standard deviations assumed equal to 65% of the mean values. For infants, the mean tap water ingestion rate is assumed to be 800 mL d–1 for infants consuming powdered infant formula [5]. 1.2.5 Soil Ingestion
The distribution for assumed soil ingestion is derived from Stanek and Calabrese [11]. A lognormal distribution is created with an arithmetic average intake of 40 mg d–1 and a standard deviation of 100, which gives a 95th percentile intake rate of approximately 200 mg d–1. There are insufficient age-related data to define different rates of soil ingestion for different age groups [11]. Most quantitative soil ingestion data have been collected for children [12]. Therefore, in the absence of any quantitative data suggesting otherwise, the same assumption is applied to all age groups. 1.2.6 Dust Ingestion
Due to a lack of available data, the average intake and standard deviation for dust ingestion are assumed to be the same as that for soil ingestion. Both the dust and the soil ingestion rates are highly uncertain due to difficulties in measuring these rates. Also, there is a large degree of variability in these ingestion rates, from one individual to another. 1.2.7 Time Spent Indoors
The number of hours a person spends indoors each day is represented by a uniform distribution. It is assumed that the total time spent indoors (at all locations) is 20–24 hours per day, with any value in the range considered equally likely. This distribution is used for all age groups. The amount of time spent indoors is used to calculate exposure from breathing indoor air and the remainder is used to calculate exposure from breathing outdoor air. 1.3 Chemical Concentrations
Distributions of concentrations of each phthalate ester, in the various exposure media, are assigned by using the compilation of data reported elsewhere [13] and summarized in Chapter 5. These concentrations, and the assigned distribution characteristics, are further described below. Note that, due to limited data, the datasets from different regions (Canada, United States, Europe, and Japan/Asia) are combined. Exposure estimates are not made for individual regions.
234
K. Clark et al.
2 Dimethyl Phthalate (DMP) The assigned distributions for the concentration of DMP in each medium used for the exposure assessment are summarized in Table 2. These concentration distributions are obtained from the report prepared for the American Chemistry Council [13]. For DMP, few data are available for food. Measurements of DMP in fish, milk, and vegetables are all reported as not detected. One of the larger studies of phthalate ester concentrations in food [14] indicated that the presence of DMP was evaluated, but was not found. Therefore, to obtain an approximate estimate of DMP exposure, the mean concentration of DMP is assumed to be equal to one half of the reported detection limit for DEP, for the food categories for which no data are available. As shown in Table 2, the median estimated daily intake of DMP for each group in mg kg–1 d–1 is: adults (0.7), teens (0.7), children (1.4), and toddlers (1.6). The estimated intake for infants, from non-food sources, is 0.05 µg kg–1 d–1 and 0.01 µg kg–1 d–1, for formula-fed and breast-fed infants, respectively. Table 2 also presents the estimated percentage intake of DMP for each age group via each medium. As shown, for all age groups, food represents the domTable 2. Dimethyl phthalate – exposure estimates a
Concentration Medium
Units
Outdoor air Indoor air Drinking water Ingested soil Ingested dust
µg m–3 µg m–3 µg L–1 µg g–1 µg g–1
LN LN LN C LN
0.0026 0.02 0.5 0 1.73
0.0017 0.013 0.33
Food Beverages excl. water Milk Cereals Dairy products Eggs Fats and oils Fish Fruit products Grains Meats Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula – powder Breast milk
µg L–1 µg L–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN C C
25 b 1.25 0.02 b 0.05 b 0.05 b 0.25 b 0.005 0.02 b 0.05 b 0.05 b 0.045 b 0.005 b 0.05 b 0.05 b 0.005 0 0
16 0.8 0.013 0.03 0.03 0.16 0.003 0.013 0.03 0.03 0.029 0.003 0.03 0.03 0.003
a b
Dist.
Min.
Mean
Max.
Std. Dev.
1.12
Dist. distribution type; LN log normal; C constant; NA not available or not applicable. Concentration based on one half of assumed detection limit.
235
Assessment of Critical Exposure Pathways Table 2 (continued)
Intake
Adult
Teen
Child
Toddler
Infant Formula- Breastfed fed
Total daily in- mg kg–1 d–1 take (median)
0.7
0.7
1.4
1.6
0.0 0.5 0.7 0.0 0.1
0.0 0.6 1.0 0.0 0.1
0.0 0.7 1.2 0.0 0.2
0.0 0.7 1.3 0.0 0.3
0.1 8.2 77.3 0.0 14.4
0.4 36.0 0.0 0.0 63.5
Food Beverages excl. water Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats Milk Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula/breast milk
39.0 1.0 4.7 2.9 11.0 0.2 6.6 14.2 8.4 0.5 2.2 2.0 1.9 1.9 2.1 NA
22.4 1.0 5.1 2.3 15.3 0.1 6.7 21.6 9.8 1.4 2.9 2.6 2.1 2.4 2.5 NA
15.3 1.8 5.7 2.8 14.1 0.1 10.7 25.3 7.3 1.9 2.8 2.8 2.2 2.5 2.5 NA
11.7 3.4 7.0 4.9 11.0 0.1 15.3 18.0 7.8 3.2 2.7 5.5 2.7 2.2 2.3 NA
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 NA
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 NA
Total food
98.7
98.3
97.9
97.8
NA
NA
Total daily intake (% of total) Outdoor Air Indoor Air Drinking water Ingested soil Ingested dust
Total
100
100
100
100
0.05
100
0.01
100
inant source of exposure, accounting for more than 97% of exposure. It is not appropriate to identify any particular food group as more important than another, as the relative importance of the various foods is controlled by the detection limits assumed. 2.1 Sensitivity Analysis
The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball and are shown below. The values in parentheses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed:
236
K. Clark et al.
– Adult – concentration of DMP in beverages (26.9%), body weight (24.6%), ingestion rate of beverages (24.2%), ingestion rate of grains (5.2%). – Teen – body weight (37.3%), concentration of DMP in beverages (11.8%), ingestion rate of beverages (9.3%), ingestion rate of grains (8.7%), concentration in grains (8.6%). – Child – body weight (47.4%), ingestion rate of grains (9.9%), concentration in grains (9.7%), ingestion rate of fats and oils (5.4%). – Toddler – body weight (47.4%), ingestion rate of grains (7.3%), concentration in fruit (7.1%), concentration in grains (6.6%), ingestion rate of fruit (5.7%). – Formula-fed infant – concentration in drinking water (31.3%), ingestion rate of drinking water (31.0%), body weight (27.4%), ingestion rate of dust (5.2%). – Breast-fed infant – ingestion rate of dust (47.3%), body weight (22.2%), concentration in indoor air (15.4%), and concentration in dust (7.3%). The parameters listed above are used to quantify exposure for the pathways found to be the dominant sources of DMP exposure but, as described above, the relative importance of the various pathways is largely controlled by the detection limits assumed for each food group/pathway. Reduction in the variability and/or the uncertainty in the above parameters would reduce the uncertainty in the exposure estimates. 2.2 Comparison to Other Studies
No other studies that evaluated human exposure to DMP have been identified.
3 Diethyl Phthalate (DEP) The assigned distributions for the concentration of DEP in each medium used for the exposure assessment are summarized in Table 3. These concentration distributions are obtained from the report prepared for the American Chemistry Council [13]. For DEP, data are not available for infant formula or breast milk. As shown in Table 3, the median estimated daily intake of DEP for each group in mg kg–1 d–1 is: adults (2.5), teens (3.0), children (5.7), and toddlers (10.6). The estimated intake for infants, from non-food sources, is 0.2 µg kg–1d–1. Table 3 also presents the estimated percentage intake of DEP for each age group via each medium. As shown, for all age groups, food represents the dominant source of exposure, accounting for more than 95% of exposure. The other foods category is the largest source of exposure, representing 61–78% of exposure. For the non-infant, inhalation of indoor air represents the most important source of non-food exposure, accounting for 4.4% of the exposure to DEP for adults. Inhalation of outdoor air, ingestion of drinking water, soil, and dust, combined, represent less than 1% of exposure. For the infant, where food data are not available, inhalation of indoor air is the most important pathway of exposure, followed by ingestion of dust.
237
Assessment of Critical Exposure Pathways Table 3. Diethyl phthalate – exposure estimates a
Concentration Medium
Units
Dist.
Outdoor air Indoor air Drinking water Ingested soil Ingested dust
µg m–3 µg m–3 µg L–1 µg g–1 µg g–1
LN LN LN LN LN
0.039 0.621 0.5 0.62 2.7
0.025 0.40 0.33 0.40 1.8
Food Beverages excl. water Milk Cereals Dairy products Eggs Fats and oils Fish Fruit products Grains Meats Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula – powder Breast milk
µg L–1 µg L–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µgg–1 µg g–1 µg g–1 µg g–1
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN C C
27 2 0.11 0.05 0.05 0.25 0.059 0.076 0.05 0.05 0.045 0.58 0.05 0.05 0.005 0 0
18 1.3 0.07 0.03 0.03 0.16 0.038 0.049 0.03 0.03 0.029 0.38 0.03 0.03 0.003
Intake
Min.
Mean
Adult Teen Child Toddler
Max.
Std. Dev.
Infant Formula-fed Breast-fed
Total daily inµg kg–1 d–1 take (median) Total daily intake (% of total) Outdoor air Indoor air Drinking water Ingested soil Ingested dust Food Beverages excl. water Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats a
2.5
3.0
5.7
10.6
0.2
0.2
0.0 4.4 0.2 0.0 0.1
0.0 4.2 0.2 0.0 0.1
0.0 4.5 0.2 0.0 0.1
0.0 2.6 0.2 0.0 0.1
0.4 70.6 21.4 1.4 6.2
0.5 89.9 0.0 1.8 7.8
11.4 1.4 1.3 0.8 3.0 0.6 6.8 3.8 2.3
5.4 1.2 1.2 0.5 3.4 0.3 5.7 4.8 2.2
3.4 2.1 1.2 0.6 2.9 0.3 8.3 5.2 1.5
1.6 2.3 0.9 0.6 1.3 0.1 7.1 2.2 0.9
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Dist. distribution type; LN log normal; C constant; NA not available or not applicable.
238
K. Clark et al.
Table 3 (continued)
Intake
Adult Teen Child Toddler
Infant Formula-fed Breast-fed
Milk Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula/breast milk
0.2 0.6 61.5 0.5 0.5 0.6 NA
0.5 0.7 68.1 0.5 0.5 0.6 NA
0.6 0.6 67.2 0.5 0.5 0.5 NA
0.6 0.3 78.3 0.3 0.3 0.3 NA
0.0 0.0 0.0 0.0 0.0 0.0 NA
0.0 0.0 0.0 0.0 0.0 0.0 NA
Total food
95.3
95.5
95.2
97.1
NA
NA
Total
100
100
100
100
100
100
3.1 Sensitivity Analysis
The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball and are shown below. The values in parentheses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed: – Adult – ingestion rate of other foods (36.9%), concentration of DEP in other foods (36.9%), body weight (15.2%). – Teen – concentration in other foods (39.3%), ingestion rate of other foods (37.9%), body weight (15.1%). – Child – concentration in other foods (36.4%), ingestion rate of other foods (34.9%), body weight (20.8%). – Toddler – concentration in other foods (40.2%), ingestion rate of other foods (38.6%), body weight (15.6%). – Formula-fed infant – concentration in indoor air (38.9%), body weight (38.0%), inhalation rate (9.6%). – Breast-fed infant – concentration in indoor air (52.7%), body weight (29.2%), inhalation rate (11.9%). The parameters listed above are used to quantify exposure for the pathways found to be the dominant sources of DEP exposure (i.e., ingestion of other foods, inhalation of indoor air). Body weight is also an important parameter, as the intakes are expressed on a “per body weight” basis. Reduction in the variability and/or the uncertainty in the above parameters would reduce the uncertainty in the exposure estimates.
Assessment of Critical Exposure Pathways
239
3.2 Comparison to Other Studies
Several other studies have evaluated human exposure to DEP. Table 4 presents a comparison of the results of these studies with those of the present study. The Centers for Disease Control and Prevention (CDC) measured monoester metabolites of phthalate esters, including DEP, in urine from a population of 289 adult humans [15]. These measurements integrate exposure from all routes (oral, inhalation, dermal, and intravenous) and all sources (environmental media, food, and consumer products). David [16] and Kohn et al. [17] used the CDC urinary metabolite data to back-calculate the intake of various phthalate esters. As shown in Table 4, the calculated geometric mean intake from David [16] and the median intake from Kohn et al. [17] are approximately a factor of five greater than the estimated median intake in the present study. As described below, these results are in contrast to those for DEHP, BBP, and DBP, for which the present study overestimates the intake relative to the intake estimated from urinary metabolite data. The most likely reason for the discrepancy, for DEP, is that DEP is used in consumer products and direct exposure to consumer products is not included in the present study. In two more recent studies, the CDC reports measurements of monoester metabolites of phthalate esters in urine from 1000 individuals, age 6 through adult [18] and in the urine of toddlers aged 12 to 18 months [19]. David has back-calculated the intake of DEP for these studies [20, 21] by using the same method as in David [16] and the results are presented in Table 4. As shown in Table 4, the intakes back-calculated for the CDC studies [18, 19] are comparable to the intakes estimated in the present study.
4 Dibutyl Phthalate (DBP) The assigned distributions for the concentration of DBP in each medium used for the exposure assessment are summarized in Table 5. These concentration distributions are obtained from the report prepared for the American Chemistry Council [13]. For DBP, data are available for all media and food categories. As shown in Table 5, the median estimated daily intake of DBP for each group in mg kg–1 d–1 is: adults (5.6), teens (6.4), children (11), toddlers (14), formula-fed infants (1.5), and breast-fed infants (2.9). These results are compared to the results of other studies in Sect. 4.2. Table 5 also presents the estimated percentage intake of DBP for each age group via each medium. As shown, for all age groups, food represents the dominant source of exposure, accounting for 63–96% of exposure. The food category contributing most to total exposure depends upon the age group considered but, in general, the most important sources of exposure are: meats, other foods, beverages excluding water, fats and oils, grains, vegetables, and infant formula and breast milk. Inhalation of indoor air and ingestion of dust represent the most important sources of non-food exposure, accounting for between 4% and 34% of exposure, depending upon the age group. Inhalation of outdoor air, ingestion of drinking
12.34 (geometric mean) 93.33 (95th percentile) 242.81 (highest value) 12 (median) 110 (95th percentile) 320 (highest value)
David [16] Intake calculated from urinary metabolite data (Blount et al. [15])
Kohn et al. [17] Intake calculated from urinary metabolite data (Blount et al. [15])
a
NA not available.
David [21] Intake calculated from urinary metabolite data [19]
ACC [20] Intake calculated from urinary metabolite data [18]
2.9
2.4 a
NA
NA
5.4
5.7
Child
Age 6 through adult: 5.42 (geometric mean) 31.90 (90th percentile) 50.06 (95th percentile)
NA
NA
3.0
Teen
2.5
Adult
DEP intake (µg kg–1 d–1)
Present study Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only
Study
Table 4. Comparison of DEP exposure estimates with estimates from other studies
6.29 (geometric mean) 36.72 (95th percentile) 42.95 (highest value)
NA
NA
10.3
10.6
Toddler
NA
NA
NA
Non-food only: 0.2 (formula-fed) 0.2 (breast-fed)
Infant
240 K. Clark et al.
241
Assessment of Critical Exposure Pathways Table 5. Dibutyl phthalate – exposure estimates
Concentration Medium
Units
Dist.
Min.
Mean
Max.
Outdoor air Indoor air Drinking water Ingested soil Ingested dust
µg m–3 µg m–3 µg L–1 µg g–1 µg g–1
T LN LN LN LN
0.00008
0.016 1.0 0.5 0.18 56.7
0.38
Food Beverages excl. water Milk Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula – powder Breast milk
µg L–1 µg L–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN
Intake
Adult
Teen
0.33
100 12 0.3 0.04 0.09 2.5 0.23 0.033 0.26 0.92 0.18 0.16 0.13 0.54 0.17 0.07 0.03 Child
Toddler
Std. Dev.
65 7.8 0.20 0.03 0.06 1.6 0.15 0.02 0.17 0.60 0.12 0.10 0.08 0.35 0.111 0.05 0.02 Infant Formulafed
Total daily inµg kg–1 d–1 take (median) Total daily intake (% of total) Outdoor air Indoor air Drinking water Ingested soil Ingested dust Food Beverages excl. water Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats
5.6
6.4
11
0.0 3.6 0.1 0.0 0.5
0.0 3.6 0.1 0.0 0.6
0.1 4.4 0.1 0.0 0.7
21.0 1.9 0.5 0.7 14.8 1.2 1.5 10.0 20.8
10.7 1.8 0.5 0.5 18.3 0.6 1.3 13.5 21.5
7.6 3.4 0.6 0.6 17.4 0.6 2.2 16.3 16.6
14
Breastfed
1.5
2.9
0.0 4.0 0.1 0.0 1.1
0.2 15.2 2.9 0.1 19.1
0.1 7.7 0.0 0.0 9.7
5.5 5.9 0.7 1.0 12.8 0.4 2.9 10.9 16.6
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
242
K. Clark et al.
Table 5 (continued)
Intake
Adult
Teen
Child
Toddler
Infant Formulafed
Breastfed
Food Milk Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula/breast milk
0.7 1.2 8.5 0.7 2.8 9.5 NA
1.6 1.4 10.0 0.7 3.2 10.1 NA
2.2 1.4 11.2 0.7 3.4 10.5 NA
3.6 1.3 20.5 0.8 2.7 9.3 NA
0.0 0.0 0.0 0.0 0.0 0.0 62.7
0.0 0.0 0.0 0.0 0.0 0.0 82.6
Total food
95.8
95.6
94.7
94.7
62.7
82.6
Total a
100
100
100
100
100
100
Dist. distribution type; T triangular; LN log normal; NA not available or not applicable.
water, and ingestion of soil, combined, represent less than 1% of exposure (except for the formula-fed infant where ingestion of drinking water accounts for 2.9% of exposure). 4.1 Sensitivity Analysis
The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball and are shown below. The values in parentheses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed: – Adult – body weight (31.0%), ingestion rate of meats (10.6%), concentration of DBP in beverages (10.5%), ingestion rate of beverages (9.0%), concentration in meats (8.9%), concentration in fats and oils (6.8%). – Teen – body weight (36.1%), concentration of DBP in meats (10.1%), ingestion rate of meats (10.0%), ingestion rate of fats and oils (7.3%), concentration in fats and oils (6.9%), and concentration in grains (5.0%). – Child – body weight (46.3%), ingestion rate of fats and oils (6.5%), ingestion rate of meats (6.0%), concentration of DBP in grains (5.8%), ingestion rate of grains (5.0%), concentration in fats and oils (5.0%). – Toddler – body weight (45.9%), concentration of DBP in other foods (8.7%), ingestion rate of other foods (8.4%), concentration in meats (6.3%), ingestion rate of meats (5.5%). – Formula-fed infant – body weight (33.7%), concentration in formula (25.8%), ingestion rate of infant formula (22.6%), ingestion rate of dust (7.8%). – Breast-fed infant – concentration in breast milk (37.5%), ingestion rate of breast milk (29.0%), body weight (25.7%).
Assessment of Critical Exposure Pathways
243
The parameters listed above are used to quantify exposure for the pathways found to be the dominant sources of DBP exposure (i.e., ingestion of meats, other foods, beverages excluding water, fats and oils, and infant formula and breast milk). Body weight is also an important parameter, as the intakes are expressed on a “per body weight” basis. Reduction in the variability and/or the uncertainty in the above parameters would reduce the uncertainty in the exposure estimates. 4.2 Comparison to Other Studies
Several other studies have evaluated human exposure to DBP [3, 16, 17, 20–24]. Table 6 presents a comparison of the results of these studies with those of the present study. As described in Sect. 3.2, David [16] and Kohn et al. [17] used the CDC urinary metabolite data to back-calculate the intake of various phthalate esters.As shown in Table 6, the calculated geometric mean intake from David [16] is a factor of 3.6 less than the estimated median intake in the present study and the 95th percentile intake is slightly larger than the median intake in the present study. The median and 95th percentile intakes from Kohn et al. [17] are comparable to the geometric mean and 95th percentile intakes in David [16]. Similarly, the median intakes in the present study are larger than the intakes back-calculated [20, 21] from the more recent CDC studies [18, 19]. Comparison of the results of the present study with the studies using urinary metabolite data suggests that human exposure to DBP may be overestimated in the present study in terms of exposure via food, but may be underestimated in terms of exposure to consumer products. Possible sources of the overestimation include: – Changes in food processing technology with time have not been evaluated.All available data on measurements of DBP in food have been pooled and a large portion of the data was obtained in the late 1980s. The data may not reflect current concentrations in food. – The reported concentrations in food are generally prior to any food preparation. Cooking of the food, and removal of fats during cooking, may reduce the concentration of DBP in the food that is ultimately consumed. – Background contamination in analysis of food samples. Background concentrations may be determined in the analysis, but often the measured concentrations are not corrected for the blank concentrations. MAFF [22] estimated the intake of DBP to infants at birth and at six months of age, due to ingestion of infant formula. MAFF found that the estimated intake decreased by a factor of about six from 1996 to 1998. The estimated intake for a six month old, using the 1998 data, is comparable to the estimated intake for the infant in the present study. The International Program on Chemical Safety (IPCS) [24] estimated adult exposure to DBP using the 1986 Canadian market basket survey data (the same data used by EC &HC [3]) and found food to be the dominant source of exposure. As shown in Table 6, the results of the present study are comparable to those of IPCS [24].
IPCS [24] Used data for concentrations in food as in EC&HC [3] Food only
ACC [20] Intake calculated from urinary metabolite data [18] David [21] Intake calculated from urinary metabolite data [19] MAFF [22] Estimated intake due to ingestion of infant formula)
David [16] Intake calculated from urinary metabolite data (Blount et al. [15]) Kohn et al. [17] Intake calculated from urinary metabolite data (Blount et al. [15])
Present study Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only
Study
NA NA
7
NA
7.1
NA
NA
NA
NA
NA
NA
NA
1998 data: 2.4 (at birth) 1.4 (at age 6 months 1996 data: 14 (at birth) 9.3 (at age 6 months) NA
NA
2.45 (geometric mean) 16.57 (95th percentile) 180.7 (highest value) NA
NA
NA
NA 1.56 (geometric mean) NA a 6.87 (95th percentile) 116.96 (highest value) 0.084 (minimum) NA NA 1.5 (median) 7.2 (95th percentile) 110 (highest value) Age 6 through adult: 0.90 (geometric mean) 2.70 (90th percentile); 3.64 (95th percentile)
1.5 (formula-fed) 2.9 (breast-fed)
Infant
0.9 (formula-fed) 2.4 (breast-fed) NA
14
Toddler
13
6.1
5.4
11
Child
10
6.4
Teen
5.6
Adult
DBP intake (µg kg–1 d–1)
Table 6. Comparison of DBP exposure estimates with estimates from other studies
244 K. Clark et al.
b
3.2
4.3
NA
Child
4.1
5.0
NA
Toddler
1.6
2.4
NA
Infant
NA not available. EC&HC [3] estimate obtained using concentrations in individual food items, rather than food categories, as was done in the present study (see Sect. 4.2).
1.4
1.1
a
2.3
1.9
EC&HC [3] b Deterministic analysis (all exposure pathways excluding children’s and other consumer products) Food only
NA
Teen
0.19 (mean) 0.44 (97.5th percentile)
Adult
DBP intake (µg kg–1 d–1)
MAFF [23] Estimated intake due to ingestion of carcass meat, poultry, eggs and milk. Converted intake in mg/person/day by assuming a body weight of 70 kg.
Study
Table 6 (continued)
Assessment of Critical Exposure Pathways
245
246
K. Clark et al.
MAFF [23] estimated the intake of DBP to adults from ingestion of carcass meat, poultry, eggs, and milk. The mean estimated intake is approximately a factor of 30 lower than the estimated median total intake in the present study, because not all sources of exposure are included. As shown in Table 6, the estimated intakes in the present study are approximately three times greater than the estimated intakes in the EC&HC [3] study for all ages except the infant. In both studies, food is the most important source of exposure; however, because the estimated intake from food is lower in the EC &HC study, the relative contribution from indoor air to total intake is greater in the EC&HC study. Examination of the supporting documentation for the EC &HC study [25] shows that intake of DBP via food is calculated for individual foods (rather than for food categories as is done in the present study). This methodology is consistent with that used by EC &HC [26] for BBP (explained in Sect. 5.2). In cases where measurements are limited to only a few food items, this method may result in an underestimate of the total intake, as concentrations in many food items are assigned a zero or laboratory detection limit rather than the concentration measured in a related food item. For the infant, the estimated intake is comparable between the present study and EC&HC [3].
5 Butylbenzyl Phthalate (BBP) The assigned distributions for the concentration of BBP in each medium used for the exposure assessment are summarized in Table 7. These concentration distributions are obtained from the report prepared for the American Chemistry Council [13]. For BBP, data are not available for soil or breast milk. As shown in Table 7, the median estimated daily intake of BBP for each group in mg kg–1 d–1 is: adults (3.7), teens (5.7), children (7.9), toddlers (9.3), and formula-fed infants (1.5). These results are compared to the results of other studies in Sect. 5.2. Table 7 also presents the estimated percentage intake of BBP for each age group via each medium. As shown, for the non-infants, food represents between 91% and 96% of exposure. The food category contributing most to total exposure is fats and oils, representing 51–67% of exposure depending upon the age group. For the formula-fed infant, ingestion of formula represents 27% of exposure, while ingestion of dust accounts for 70% of exposure, and ingestion of drinking water accounts for 2% of exposure. Ingested dust represents the most important source of non-food exposure for the other age groups, accounting for 4–9% of the exposure to BBP depending upon the age group. Inhalation of air and ingestion of drinking water, combined, represent less than 1% of exposure. As mentioned above, no data are available with which to evaluate exposure to soil. 5.1 Sensitivity Analysis
The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball and are shown below. The values in paren-
247
Assessment of Critical Exposure Pathways Table 7. Butyl benzyl phthalate – exposure estimates a
Concentration Medium
Units
Dist.
Min.
Mean
Max.
Outdoor air Indoor air Drinking water Ingested soil Ingested dust
µg m–3 µg m–3 µg L–1 µg g–1 µg g–1
T LN LN C LN
0.0005
0.0017 0.035 0.5 0 333
0.02
Food Beverages excl. water Milk Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula – powder Breast milk
µg L–1 µg L–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN C
Intake
Adult
0.023 0.33 216
39 1.2 0.05 0.4 0.08 7.4 0.01 0.02 0.05 0.13 0.045 0.09 0.04 0.05 0.005 0.044 0 Teen
Child
Std. Dev.
25 0.8 0.03 0.26 0.05 4.8 0.007 0.013 0.03 0.08 0.029 0.06 0.03 0.03 0.003 0.029
Toddler
Infant Formulafed
Total daily intake (median) µg kg–1 d–1
3.7
5.7
7.9
9.3
1.5
Total daily intake (% of total) Outdoor air Indoor air Drinking water Ingested soil Ingested dust
0.0 0.2 0.1 NA 4.2
0.0 0.2 0.1 NA 4.2
0.0 0.2 0.2 NA 5.3
0.0 0.2 0.2 NA 8.5
0.0 0.4 2.0 NA 70.2
Food Beverages excl. water Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats
11.2 0.4 6.9 0.8 60.1 0.1 1.2 2.6 4.0
5.2 0.4 6.1 0.5 66.7 0.0 1.0 3.2 3.8
3.7 0.7 7.0 0.7 64.2 0.0 1.6 3.9 2.9
2.9 1.3 8.8 1.2 51.0 0.0 2.4 2.8 3.2
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
a
Dist. distribution type; T triangular; LN log normal; C constant; NA not available or not applicable.
248
K. Clark et al.
Table 7 (continued)
Intake
Adult
Teen
Child
Toddler
Infant Formulafed
Food Milk Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula
0.1 0.4 6.5 0.3 0.4 0.4 NA
0.2 0.4 7.0 0.3 0.4 0.4 NA
0.3 0.4 7.8 0.3 0.4 0.4 NA
0.5 0.4 15.5 0.3 0.3 0.4 NA
0.0 0.0 0.0 0.0 0.0 0.0 27.4
Total food
95.5
95.5
94.3
91.1
27.4
100.0
100.0
100.0
100.0
100.0
Total
theses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed: – Adult – concentration of BBP in fats and oils (35.3%), ingestion rate of fats and oils (34.6%), body weight (16.7%). – Teen – concentration of BBP in fats and oils (37.2%), ingestion rate of fats and oils (36.1%), body weight (17.6%). – Child – ingestion rate of fats and oils (35.3%), concentration in fats and oils (34.2%), body weight (20.2%). – Toddler – concentration in fats and oils (28.7%), body weight (25.7%), ingestion rate of fats and oils (24.8%). – Formula-fed infant – ingestion rate of dust (51.9%), body weight (17.5%), concentration in dust (9.8%), concentration in formula (8.9%), ingestion rate of formula (7.7%). The parameters listed above are used to quantify exposure for the pathways found to be the dominant sources of BBP exposure (i.e., ingestion of fats and oils, infant formula, and dust). Body weight is also an important parameter, as the intakes are expressed on a “per body weight” basis. Reduction in the variability and/or the uncertainty in the above parameters would reduce the uncertainty in the exposure estimates. 5.2 Comparison to Other Studies
Several other studies have evaluated human exposure to BBP [16, 17, 20–23, 26, 27]. Table 8 presents a comparison of the results of these studies with those of the present study. As shown in Table 8, for all age groups, the estimated median intake of BBP is a factor of three or more greater in the present study compared to the “average
Adult
BBP Intake (µg kg–1 d–1)
0.4 (formula-fed)
b
a
NA
NAc
0.73 (geometric mean) 3.34 (95th percentile) 19.79 (highest value)
40–47
21–25
11–14 (20–59 year olds); 9–12 (60+years)
40–47
21–25
11–14 (20–59 year olds); 9–12 (60+years)
NA
71–82
71–82
NA
0 (formula-fed) 129–145 (not formula-fed)
0.27–0.38 (formula-fed) 129–145 (not formula-fed)
EC&HC [26], estimate of “average intake” used concentrations in individual food items, rather than food categories (see Sect. 5.2). Reasonable worst case” estimate used food categories. cNA not available.
David [16] Intake calculated from urinary metabolite data (Blount et al. [15])
Reasonable worst caseb (all exposure pathways excluding children’s and other consumer products) Food only
0 (formula-fed) 0.28–3.15 (not formula-fed)
8.5
1.5 (formula-fed)
Infant
Deterministic analysis; average intake a (all exposure pathways excluding children’s and other consumer products) Food only 0.63–2.01 (20–59 year olds); 0.72–2.29 0.97–3.70 1.27–4.88 0.40–1.55 (60+years)
7.4
5.4
9.3
Toddler
0.01–0.14 (formula-fed) 0.29–3.21 (not formula-fed)
7.9
Child
5.7
Teen
0.64–2.05 (20–59 year olds); 0.74–2.33 0.98–3.77 1.29–4.96 0.42–1.58 (60+ years)
EC&HC [26]
Present study 3.7 Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only 3.5
Study
Table 8. Comparison of BBP exposure estimates with estimates from other studies
Assessment of Critical Exposure Pathways
249
MAFF [23] 0.1 (mean) Estimated intake due to ingestion of carcass 0.3 (97.5th percentile) meat, poultry, eggs and milk. Converted intake in mg/person/day by assuming a body weight of 70 kg
NA
NA
MAFF [22] Estimated intake due to ingestion of infant formula
NA
NA
IPCS [27] 2 Intake based on ingestion of BBP measured in four foods (butter, cheddar cheese, crackers, yoghurt)
David [21] Intake calculated from urinary metabolite data [19]
NA
NA
NA
Infant
NA
NA
NA
1998 data: 0.2 (at birth) 0.1 (at age 6 months) 1996 data: 8.7 (at birth) 5.6 (at age 6 months)
NA
1.51 (geometric mean) 6.42 (95th percentile) 7.75 (highest value)
NA
Toddler
Approx. 6 NA
Age 6 through adult: 0.66 (geometric mean) 2.52 (90th percentile); 3.18 (95th percentile)
NA
Child
ACC [20] Intake calculated from urinary metabolite data [18]
NA
Teen
0.094 (minimum) 0.88 (median) 4.0 (95th percentile) 29 (highest value)
Adult
BBP Intake (µg kg–1 d–1)
Kohn et al. [17] Intake calculated from urinary metabolite data (Blount et al. [15])
Study
Table 8 (continued)
250 K. Clark et al.
Assessment of Critical Exposure Pathways
251
intake” in the EC&HC [26] study. The “reasonable worst-case” intake in the EC&HC [26] study is larger than the median intake in the present study. In both studies, food is the dominant source of exposure. The EC&HC study is a deterministic analysis, using receptor characteristics similar to the mean receptor characteristics employed in the present study. There are some important differences between the two studies in terms of the concentrations of BBP in food and their treatment. To determine the average intake of BBP, EC&HC [26] calculated the intake from ingestion of only 8 of 181 food items. The calculations are based on the results of measurements of BBP in six foods (yogurt, cheddar cheese, butter, pork, mixed vegetable juice, and crackers). For the remaining 173 food items, a concentration of zero or an assigned laboratory detection limit is used to calculate a range of intakes. In the present study, the 181 food items are grouped into 15 food categories. If a phthalate is measured in one or more of the items in that group, then that concentration is used for the entire group. This is done to overcome the fact that phthalates have not been tested in every food item. EC&HC employed a similar strategy in their estimate of “reasonable worst-case” intakes. The measured concentrations of BBP from four food items are assumed to represent the average concentration of BBP in each of four food groups. A concentration of zero or a laboratory detection limit is assumed for the remaining eight food groups.Another difference between the two studies is that EC&HC [26] assumed a concentration of 0.64 µg g–1 BBP in butter based on Page and Lacroix [14]. In an earlier paper [28], the concentration of BBP in butter and margarine is reported as 3.1–47.8 µg g–1, based on measurements in 20 samples. By using the data from Page and Lacroix [14,28], a mean concentration of 7.4 µg g–1 is calculated and is used in the present study. This results in a large portion of the intake of BBP (51–67%, depending upon the age category) coming from ingestion of fats and oils. As described in Sect. 3.2, David [16] and Kohn et al. [17] used the CDC urinary metabolite data to back-calculate the intake of various phthalate esters. As shown in Table 8, the geometric mean intake of BBP, calculated by David [16], is a factor of five less than the estimated median intake in the present study and the 95th percentile intake is comparable to the median intake in the present study. The median and 95th percentile intakes from Kohn et al. [17] are comparable to the geometric mean and 95th percentile intakes in David [16]. Similarly, the median intakes in the present study are larger than the intakes back-calculated [20, 21] from the more recent CDC studies [18, 19]. Comparison of the results of the present study with the studies using urinary metabolite data suggests that human exposure to BBP may be overestimated in the present study. Possible sources of the overestimation include changes in food processing technology with time, a reduction in the concentration of BBP in food following cooking (especially since fats and oils are estimated to be the largest source of BBP exposure in the present study), and measured concentrations in food may not be corrected for background contamination. Conversely, total human exposure to BBP may be underestimated in the present study because BBP is used in personal care products [16] and such exposures have not been included in the present study, but would be accounted for in studies using urinary metabolite data.
252
K. Clark et al.
The International Program on Chemical Safety [24] estimated adult exposure to BBP using the 1986 Canadian market basket survey data (the data used in [26]). The estimated intake for an adult is less than that in the present study and comparable to the estimated “average intake” by EC&HC [26]. MAFF [22] estimated the intake of BBP to infants at birth and at six months of age, due to ingestion of infant formula. MAFF found that the estimated intake decreased by a factor of approximately 50 from 1996 to 1998. The estimated intake for the infant in the present study is greater than the MAFF estimate using 1998 data, but less than the intake using the 1996 data. MAFF [23] estimated the intake of BBP to adults from ingestion of carcass meat, poultry, eggs, and milk. The mean estimated intake is a factor of 37 lower than the estimated median total intake in the present study, because not all sources of exposure are included.
6 Bis(2-Ethylhexyl) Phthalate (DEHP) The assigned distributions for the concentration of DEHP in each medium used for the exposure assessment are summarized in Table 9. These concentration distributions are obtained from the report prepared for the American Chemistry Council [13]. For DEHP, data are available for all required media. As shown in Table 9, the median estimated daily intake of DEHP for each group in mg kg–1 d–1 is: adults (8.2), teens (10), children (19), toddlers (26), formula-fed infants (5.0), and breast-fed infants (7.3). These results are compared to the results of other studies in Sect. 6.2. Table 9 also presents the estimated percentage intake of DEHP for each age group via each medium. As shown, for the non-infants, food represents between 92% and 95% of the total exposure. The food category contributing most to total exposure depends upon the age group considered but, in general, the most important sources of exposure are: beverages excluding water, dairy products, fats and oils, grains, milk, and other foods. For both the formula-fed and breast-fed infants, food (infant formula or breast milk) represents approximately one half of the exposure, with ingestion of dust accounting for the remainder. Ingested dust represents the most important source of non-food exposure for the other age groups, accounting for between 4.2% and 6.6% of exposure to DEHP. Inhalation of indoor air represents about 1% of exposure for all age groups, while inhalation of outdoor air, ingestion of drinking water, and ingestion of soil, combined, represent less than 1% of exposure. 6.1 Sensitivity Analysis
The input parameters contributing most to the variation in the exposure estimates are identified by Crystal Ball™ and are shown below. The values in parentheses indicate the percentage contribution to variance. Note that only parameters contributing 5% or more to the variance are listed:
253
Assessment of Critical Exposure Pathways Table 9. Bis(2-ethylhexyl) phthalate – exposure estimates a
Concentration Medium
Units
Dist.
Min.
Mean
Max.
Outdoor air Indoor air Drinking water Ingested soil Ingested dust
µg m-3 µg m–3 µg L–1 µg g–1 µg g–1
T T LN T LN
0.0003 0.02
0.02 0.2 0.5 0.048 662
0.5 1.0
Food Beverages excl. water Milk Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains Meats Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula – powder – ready-to-feed liquid Breast milk
µg L–1 µg L–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1 µg g–1
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN
Intake
Adult
Teen
0.00003
0.325 1.0 430
77 80 0.53 1.5 0.21 4.1 0.46 0.03 0.5 0.35 0.21 0.28 1.1 0.94 0.17 0.2 0.007 0.062 Child
Toddler
Std. Dev.
50 52 0.34 0.98 0.14 2.7 0.30 0.02 0.33 0.23 0.14 0.18 0.7 0.61 0.11 0.13 0.005 0.040 Infant Formulafed
Total daily intake (median)
µg kg–1 d–1
Total daily intake (% of Total) Outdoor air Indoor air Drinking water Ingested soil Ingested dust Food Beverages excl. water Cereals Dairy products (excl. milk) Eggs Fats and oils Fish Fruit products Grains a
Breastfed
8.2
10.0
18.9
25.8
5.0
7.3
0.0 1.0 0.1 0.0 4.3
0.0 0.9 0.1 0.0 4.2
0.0 1.0 0.1 0.0 5.0
0.0 0.9 0.1 0.0 6.6
0.1 1.5 0.7 0.0 54.1
0.0 1.1 0.0 0.0 39.3
11.2 2.4 13.2 1.1 16.9 1.6 0.9 13.4
5.2 2.0 11.7 0.7 19.1 0.8 0.8 16.6
3.3 3.5 12.2 0.8 16.5 0.7 1.1 18.1
2.2 5.5 12.9 1.3 11.1 0.4 1.4 11.1
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Dist. distribution type; T triangular; LN log normal; NA not available or not applicable.
254
K. Clark et al.
Table 9 (continued)
Intake
Adult
Teen
Child
Toddler
Infant Formulafed
Breastfed
Food Meats Milk Nuts and beans Other foods Poultry Processed meats Vegetable products Infant formula/breast milk
5.5 3.1 1.0 10.3 3.9 3.4 6.6 NA
5.2 6.7 1.0 11.2 3.6 3.5 6.5 NA
3.7 8.6 0.9 11.3 3.5 3.4 6.1 NA
3.3 12.6 0.8 18.9 3.6 2.5 4.9 NA
0.0 0.0 0.0 0.0 0.0 0.0 0.0 43.7
0.0 0.0 0.0 0.0 0.0 0.0 0.0 59.6
Total food
94.6
94.7
93.8
92.4
43.7
59.6
Total
100
100
100
100
100
100
– Adult – body weight (35.9%), concentration of DEHP in fats and oils (9.1%), ingestion rate of fats and oils (6.2%), ingestion rate of grains (6.0%), and concentration of DEHP in beverages (5.5%). – Teen – body weight (40.9%), ingestion rate of fats and oils (9.1%), concentration of DEHP in fats and oils (8.3%), concentration in grains (7.6%), ingestion rate of grains (6.4%). – Child – body weight (48.0%), concentration in grains (7.8%), ingestion rate of grains (6.6%), ingestion rate of fats and oils (6.5%). – Toddler – body weight (43.9%), ingestion rate of other foods (8.2%), concentration in other foods (7.7%), ingestion rate of milk (5.1%). – Formula-fed infant – ingestion rate of dust (34.1%), body weight (24.7%), concentration in infant formula (18.6%), ingestion rate of formula (12.9%), concentration in dust (5.5%). – Breast-fed infant – concentration in breast milk (24.9%), body weight (23.6%), ingestion rate of breast milk (22.5%), ingestion rate of dust (20.7%). As expected, the parameters listed above are used to quantify exposure for the pathways found to be the dominant sources of DEHP exposure (i.e., ingestion of beverages excluding water, infant formula and breast milk, fats and oils, and grains). Body weight is also an important parameter, as the intakes are expressed on a “per body weight” basis. Reduction in the variability and/or the uncertainty in the above parameters would reduce the uncertainty in the exposure estimates. 6.2 Comparison to Other Studies
Several other studies have evaluated human exposure to DEHP [2, 4, 9, 16, 17, 20–23]. Table 10 presents a comparison of the results of these studies with those of the present study.
Assessment of Critical Exposure Pathways
255
David [16] used the CDC data to back-calculate the intake of each of the phthalate esters and the results are shown in Table 10 for DEHP. As shown in Table 10, the calculated geometric mean intake is approximately an order of magnitude less than the estimated median intake in the present study and the 95th percentile intake is 2.7 times less than the median intake in the present study. Kohn et al. [17] also back-calculated the intake of DEHP using the CDC data, using an alternate model. As shown in Table 10, the median and 95th percentile intakes from Kohn et al. [17] are less than those in the present study and are comparable to those in David [16]. Similarly, the median intakes in the present study are larger than the intakes back-calculated [20, 21] from the more recent CDC studies [18, 19]. These results suggest that human exposure to DEHP may be overestimated in the present study. As for DBP and BBP, possible sources of the overestimation include changes in food processing technology with time, a reduction in the concentration of DEHP in food, following cooking, and measured concentrations in food may not be corrected for background contamination. Zaleski et al. [9] estimated the intake of DEHP for toddlers and infants. Both the Zaleski et al. study and the present study are probabilistic and use a similar concentration database; however, the age groupings employed are slightly different. For the present study, the toddler is defined as being 0.5-4 years old, while in the Zaleski et al. study the toddler is 1.5–4.5 years old. Despite the age difference, the total estimated intake of DEHP, and the estimated intake from food, for the toddler is comparable in both studies. The infant is defined in the present study as being 0-6 months of age and is either exclusively formula-fed or exclusively breast-fed. In contrast, in the Zaleski et al. study, the infant is defined as 0–1 year in age and consumes other foods in addition to either infant formula or breast milk. This is the reason why the estimated intake for the infant in the Zaleski et al. study is about a factor of four greater than the estimated intake for the infant in the present study. MAFF [22] estimated the intake of DEHP to infants at birth and at six months of age, due to ingestion of infant formula. MAFF found that the estimated intake decreased by a factor of about three from 1996 to 1998. The estimated intake for a six month old, using the 1998 data, is comparable to the estimated intake for the infant in the present study. MAFF [23] estimated the intake of DEHP to adults from ingestion of carcass meat, poultry, eggs, and milk. The mean estimated intake is about a factor of four lower than the estimated median total intake in the present study, because not all sources of exposure are included. Huber et al. [29] reviewed several references that evaluated exposure of adults to DEHP in food. The estimated daily intake ranged between 0.3 µg kg–1 d–1 and 30 µg kg–1 d–1, which is in agreement with the estimate in the present study. Huber et al. [29] also presented an estimated “worst-case” intake of 485 µg kg–1 d–1. For all age groups, the estimated intake of DEHP is lower in the present study compared with the estimated intake in Health Canada [4]. The present study used the same distributions for receptor characteristics as in Health Canada [4], but used updated distributions for concentrations based on considerably more (recent) data.
0.71 (median) 3.6 (95th percentile) 46 (highest value) Age 6 through adult: 0.65 (50th percentile) 2.13 (90th percentile); 2.62 (95th percentile)
Kohn et al. [17] Intake calculated from urinary metabolite data (Blount et al. [15])
ACC [20] Intake calculated from urinary metabolite data [18]
NA
NA
NA
NA
18–20
22–32
NA
Zaleski et al. [9] a Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only NA
2.76 (geometric mean) 12.91 (95th percentile) 21.38 (highest value)
NA
NA
24
26
Toddler
David [21] Intake calculated from urinary metabolite data [19]
NA
NA
NAd
0.60 (geometric mean) 3.05 (95th percentile) 38.48 (highest value)
David [16] Intake calculated from urinary metabolite data (Blount et al. [15]) NA
18
9.5
7.8
19
Child
10
Teen
8.2
Adult
DEHP Intake (µg kg–1 d–1)
Present study Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only
Study
Table 10. Comparison of DEHP exposure estimates with estimates from other studies
13 (formula-fed) 18–19 (breast-fed)
17–26 (formula-fed) 23–31 (breast-fed)
NA
NA
2.2 (formula-fed) 4.4 (breast-fed)
5.0 (formula-fed) 7.3 (breast-fed)
Infant
256 K. Clark et al.
38
26
d
c
b
a
8.15–8.20
7.18
5.78–5.82
4.91
133
134
NA
NA
Toddler
12.85
17.81
14.03–14.10 18.86–18.98
78
79
NA
NA
Child
7.88
8.87–9.12
152
153
NA
1998 data: 13.8 (at birth) 7.7 (at age 6 months) 1996 data: 35 (at birth) 23 (at age 6 months)
Infant
Infant in Zaleski et al. [9] is 0–1 years old and consumes a variety of foods in addition to either infant formula or breast milk. In the present study, the infant is 0–6 months old and consumes only infant formula or breast milk. Health Canada [4] used same receptor characteristics as present study, but concentration database was expanded for present study. EC&HC [2] used lower concentrations for DEHP in many foods compared to the present study (see Sect. 6.2). NA not available.
EC&HC [2] Deterministic analysis (all exposure pathways excluding children’s and other consumer products) Food only
c
38
27
Health Canada [4] b Probabilistic analysis; median intake (all exposure pathways excluding children’s and other consumer products) Food only
NA
2.1 (mean) 4.3 (97.5th percentile)
MAFF [23] Estimated intake due to ingestion of carcass meat, poultry, eggs and milk. Converted intake in mg/person/day by assuming a body weight of 70 kg.
NA
Teen
NA
Adult
DEHP Intake (µg kg–1 d–1)
MAFF [22] Estimated intake due to ingestion of infant formula.
Study
Table 10 (continued)
Assessment of Critical Exposure Pathways
257
258
K. Clark et al.
The estimated intake of DEHP for all age groups, except the infant, is slightly larger in the present study compared to EC &HC [2]. In addition to being a deterministic analysis rather than a probabilistic analysis, the EC &HC [2] study used lower concentrations than are used in the present study, for many foods, based upon the data that were available at that time. Differences between the EC&HC study and the present study include: the concentrations of DEHP in dairy products, fats and oils, vegetables, and meat, are a factor of 3–12 lower in EC &HC [2] than the mean concentrations used in the present study, and exposure to DEHP in beverages other than drinking water is not evaluated in EC&HC [2]. These differences in concentrations account for the differences in the estimated intake of DEHP.
7 Discussion For the five phthalate esters evaluated in this chapter, the median estimated daily intake is highest for toddlers (ranging from 1.6 µg kg–1 d–1 for DMP to 26 µg kg–1 d–1 for DEHP) and lowest for infants (ranging from 1.5 µg kg–1 d–1 for DBP and BBP to 7.3 µg kg–1 d–1 for DEHP); exposure to infants from food could not be evaluated for DEP or DMP. For adults, the median estimated intake ranges from 0.7 µg kg–1 d–1 for DMP to 8.2 µg kg–1 d–1 for DEHP. Table 11 presents a summary of estimated total daily intake for each of the five phthalate esters for each age group. For all five phthalate esters evaluated (except BBP exposure for formula-fed infants), food represents the most important source of exposure. The food categories contributing most to exposure depend upon the phthalate ester and the age group evaluated. For all phthalate esters, except DMP, eggs, fish, and nuts and beans each represent less than 2% of exposure for all age groups. The exposure estimates presented are considered to be representative of typical population exposures, as the estimates employed “average” exposure factors [5]. Individuals who consume large quantities of fish are sometimes at greater risk of exposure to chemicals than are other members of the population.
Table 11. Estimated total daily intake by age group for a series of phthalate esters
Phthalate ester
DMP DEP DBP BBP DEHP a b
Estimated total daily intake (median value) in mg kg–1 d–1 Adult
Teen
Child
Toddler
Infant (formula-fed)
Infant (breast-fed)
0.7 2.5 5.6 3.7 8.2
0.7 3.0 6.4 5.7 10.0
1.4 5.7 11 7.9 18.9
1.6 10.6 14 9.3 25.8
0.05 a 0.2 a 1.5 1.5 5.0
0.01 a 0.2 a 2.9 NA b 7.3
Non-food exposure only. NA not available.
Assessment of Critical Exposure Pathways
259
Richardson [8] reported a lognormal distribution for consumption of fish by members of the Amerindian and Inuit communities. The mean ingestion rate (excluding individuals who reported no fish consumption) is 220 g d–1 for adults. By using this mean fish ingestion rate, the estimated intake of each phthalate ester in mg kg–1 d–1 due only to fish consumption is: 0.015 (DMP), 0.18 (DEP), 0.7 (DBP), 0.03 (BBP), and 1.4 (DEHP). This estimated intake from high fish consumption represents between 0.8% and 17% of the total estimated intake for each phthalate (reported in Tables 2, 4, 6, 8, and 10).As high fish consumers would have lower ingestion rates for some foods compared to the general population (e.g., meat, processed meat), the estimated intake of phthalates for the high fish consumers is not likely to be higher than the estimated intake for the general population. By using data obtained in 1996 and 1998, MAFF [22] estimated the intake of DBP, BBP, and DEHP in infants due to ingestion of infant formula. MAFF found that the estimated intake decreased by factors of 2.5–56 over the two-year period (see Tables 6, 8, and 10). The influences of changes in food processing methodologies and storage have not been evaluated in the present study, as recent measurements are combined with older measurements to obtain a more complete dataset. Many of the measurements of phthalate esters in food were obtained in the late 1980s. Estimates of phthalate ester exposure would be improved by obtaining recent measurements of phthalate esters in a variety of foods. Ingestion of dust and inhalation of indoor air represent the most important non-food sources of exposure to phthalate esters. These estimates rely heavily on European data, due to a lack of North American data. Detection limits have a large influence on the estimated intake of phthalate esters, particularly for DMP. The estimated intake for DMP should be regarded as approximate and, quite possibly, an overestimate, as it is based primarily on concentrations equal to one half of a detection limit. The results of the present study are compared to results from other studies for DEP, DBP, BBP, and DEHP. The estimated intakes in the present study are higher than those in EC&HC [2, 3, 26] due primarily to methodological differences in the treatment of food items versus food categories. EC&HC calculated intakes for individual food items (up to 181 items). Where data are lacking for a particular food item, the concentration is assigned a zero or a laboratory detection limit. In contrast, in the present study, intake is evaluated for food categories. Measured concentrations of one or more food items in a category are assigned to the entire category. For chemicals with a large database of measured concentrations (i.e., concentrations have been measured in most food items), the difference between the two methods may not be large but, if little data are available, an analysis by individual items may underestimate exposure and analysis by food categories may overestimate exposure. This observation leads to the recommendation for additional testing of a variety of food items. A comparison of the results of the present study with studies that backcalculate phthalate ester intake from urinary metabolite data suggests that exposure in the present study may be overestimated for DEHP, BBP, and DBP due to changes in food processing over time (many of the measured concentrations of phthalates in food are not recent), loss of phthalates due to cooking has not
260
K. Clark et al.
Table 12. Comparison of estimated total daily intake in present study with studies that back-
calculate intake from metabolite data Phthalate ester
DMP DEP DBP BBP DEHP a
Estimated total daily intake (µg kg–1 d–1) Present study (median)
Intake calculated from urinary metabolite data (mean or median)
Adult
Toddler
Adult [15–17]
Age 6 to Adult [18, 20]
Toddler [19, 21]
0.7 2.5 5.6 3.7 8.2
1.6 10.6 14 9.3 25.8
NAa 12–12.34 1.5–1.56 0.73–0.88 0.6–0.71
NA 5.42 0.90 0.66 0.65
NA 6.29 2.45 1.51 2.76
NA not available.
been accounted for in the present study, and some measured concentrations in food may be elevated due to background contamination. The comparison also suggests that the significant sources of exposure to DEHP have been accounted for in the present study. Conversely, exposure to DEP (and possibly BBP and DBP) is underestimated in the present study because direct exposure to personal care products has not been included. The overestimate of exposure to BBP and DBP from food, referred to above, may be partially cancelled by the lack of inclusion of personal care products as a source of exposure. Table 12 presents a summary of the comparison of the intakes estimated in the present study with the backcalculated intakes. A link between human exposure and multimedia modeling of phthalate esters has not yet been established. This link has been difficult to estimate, for several reasons, including: – Phthalate ester exposure may arise from industrial or consumer use in addition to environmental sources; – As discussed in Chapter 5, phthalate esters appear to be metabolized as they progress up the food chain (i.e., they biodilute), making it difficult to estimate concentrations in food from concentrations in abiotic media; – Some foods that contribute significantly to human exposure are imported from regions outside the region being modeled. As a result, human exposure estimates will not be substantially improved by measuring phthalate ester concentrations in foods from different regions. Regional differences in exposure are more likely to be due to dietary differences (i.e., differences in ingestion rates of different foods) than in the concentrations of phthalate esters. Acknowledgement. We are grateful to the Environmental Research Task Group (ERTG) of the Phthalate Ester Panel of the American Chemistry Council (ACC) for funding this research.
Assessment of Critical Exposure Pathways
261
8 References 1. Finley B, Proctor D, Scott P, Harrington N, Paustenbach D, Price P (1994) Risk Anal 14:533 2. Environment Canada and Health Canada (EC&HC) (1994) Canadian Environmental Protection Act, Priority Substances List assessment report – bis(2-ethylhexyl) phthalate. Ottawa, ON 3. Environment Canada and Health Canada (EC&HC) (1994) Canadian Environmental Protection Act, Priority Substances List assessment report – dibutyl phthalate. Ottawa, ON 4. Health Canada (1996) Environmental pathways analysis for diethylhexyl phthalate (DEHP) and other phthalate esters. Prepared by O’Connor Associates Environmental Inc. for the Health Protection Branch, Ottawa, ON 5. Health and Welfare Canada (HWC) (1993) Reference values for Canadian populations. Prepared by the Environmental Health Directorate Working Group on Reference Values. July 1988; updated May 1993 6. Stephens T, Craig CL (1990) The well-being of Canadians: highlights of the 1988 Campbell’s survey. Canadian Fitness and Lifestyle Research Institute, Ottawa, p 95, appendices, data 7. Health Canada (1995) Probabilistic assessment of 24-hour breathing rates. Prepared by Cornerstone Engineering and Consulting Inc. for the Health Protection Branch, Ottawa, ON. October 8. Richardson GM (1997) Compendium of Canadian human exposure factors for risk assessment. O’Connor Associates Environmental, Ottawa, ON 9. Zaleski RT, Parkerton TF, Konkel WJ (2000) An assessment of di-2-ethylhexyl phthalate exposure for children. Report prepared for the European Council of Plasticisers & Intermediates (ECPI), a Sector Group of CEFIC, Brussels, Belgium, by Exxon Biomedical Sciences, Inc, Annandale, NJ, p 35, appendices 10. Health Canada (1994) Probabilistic exposure assessments for twenty contaminants in the Canadian environment. Prepared by O’Connor Associates Environmental Inc. for the Health Protection Branch, Ottawa, ON. October 11. Stanek EJ III, Calabrese EJ (1995) Hum Ecol Risk Assess J 1:133 12. Calabrese EJ, Stanek EJ, Gilbert CE, Barnes RM (1990) Regul Toxicol Pharmacol 12: 88 13. Clark K, Cousins I, Mackay D (2001) Multimedia modelling and exposure assessment for phthalate esters – observed concentrations in the environment. Prepared for American Chemistry Council. December 14. Page BD, Lacroix GM (1995) Food Addit Contam 12:129 15. Blount BC, Silva MJ, Caudill SP, Needham LL, Pirkle JL, Sampson EJ, Lucier GW, Jackson RJ, Brock JW (2000) Environ Health Perspect 108:979 16. David RM (2000) Environ Health Perspect 108:A440 17. Kohn MC, Parham F, Masten SA, Portier CJ, Shelby MD, Brock JW, Needham LL (2000) Environ Health Perspect 108:A440 18. Centers for Disease Control and Prevention (CDC) (2001) National report on human exposure to environmental chemicals. National Center for Environmental Health.Available online at http://www.cdc.gov/nceh/dls/report/contact.htm 19. Brock JW, Caudill SP, Silva MJ, Needham LL, Hilborn ED (2002) Bull Environ Contam Toxicol 68:309 20. American Chemistry Council (ACC) (2001) What the CDC national report says about phthalate exposures. Phthalate Esters Panel.Available online at http://www.phthalates.org/ mediacenter/pep_2001-6-7.html 21. David RM (2002) Personal communication 22. Ministry of Agriculture, Fisheries and Food (MAFF) (1998) MAFF UK – phthalates in infant formulae – follow-up survey. Joint Food Safety and Standards Group, Food Surveillance Information Sheet #168, December 23. Ministry of Agriculture, Fisheries and Food (MAFF) (1996) MAFF UK – phthalates in food. Joint Food Safety and Standards Group, Food Surveillance Information Sheet #82, March
262
K. Clark et al.: Assessment of Critical Exposure Pathways
24. International Program on Chemical Safety (IPCS) (1997) Environmental health criteria – 189 – di-n-butyl phthalate.World Health Organization (WHO), Geneva, Switzerland. ISBN 92 4 157189 6, as cited in National Toxicology Program – Center for the Evaluation of Risks to Human Reproduction (NTP-CERHR) (2000) NTP-CERHR expert panel report on di-nbutyl phthalate. US Department of Health and Human Services. NTP-CERHR-DBP-00. October 25. Health Canada (1994) Canadian Environmental Protection Act, Priority Substances List – supporting documentation health-related sections – di-n-butyl phthalate. Ottawa, ON 26. Environment Canada and Health Canada (EC&HC) (2000) Canadian Environmental Protection Act, Priority Substances List assessment report – butylbenzyl phthalate. Ottawa, ON 27. International Program on Chemical Safety (IPCS) (1999) Concise international chemical assessment document 17 – butyl benzyl phthalate. World Health Organization (WHO), Geneva, Switzerland, as cited in National toxicology program – Center for the Evaluation of Risks to Human Reproduction (NTP-CERHR) (2000) NTP-CERHR expert panel report on butyl benzyl phthalate. US Department of Health and Human Services. NTP-CERHRBBP-00. October. 28. Page BD, Lacroix GM (1992) Food Addit Contam 9:197 29. Huber WW, Grasl-Kraupp B, Schulte-Hermann R (1996) Crit Rev Toxicol 26:365
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 263– 298 DOI 10.1007/b11469
Aquatic Toxicity of Phthalate Esters Christopher A. Bradlee 1 · Paul Thomas 2 1 2
BASF Corporation, Corporate Ecology & Safety, Wyandotte, MI, USA Akzo Nobel, Chemicals Research Arnhem, Arnhem, The Netherlands
The aquatic toxicity of Phthalate esters is discussed in this chapter. This chapter begins with an examination of the physico-chemical properties of phthalate esters that have a significant influence on toxicity.Acute and chronic toxicity are then reviewed with a focus on chronic effects. This chapter concludes with a discussion of studies that have examined the potential for endocrine modulating effects of phthalate esters in aquatic organisms. Perhaps what is most notable about phthalate esters with regard to their physico-chemical properties is that they demonstrate a solubility cut-point (threshold) so that phthalate esters with alkyl chain length of C6 or greater have water solubilities less than 1 mg/L. The acute and chronic toxicity data show that while the lower phthalates (
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
1
Introduction
1.1
Solubility Influence on Toxicity
2
Acute Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
3
Chronic Toxicity
3.1 3.2 3.3
Algae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 Invertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
. . . . . . . . . . . . . . . . . . . 264
. . . . . . . . . . . . . . . . . . . . . . . . . . . 266
© Springer-Verlag Berlin Heidelberg 2003
264
C.A. Bradlee and P. Thomas
3.4 3.4.1 3.4.2
Endocrine Modulating Effects . . . . . . . . . . . . . . . . . . . . 284 Lower Phthalates . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Higher Phthalates . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
6
Discussion
7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
1 Introduction The aquatic toxicity of phthalate esters is discussed in this chapter. A comprehensive review on the aquatic toxicity of phthalate esters was previously presented in Staples et al. [1] and this chapter provides a summary of that information and reviews the literature either not included in, or published after, the Staples et al. [1] review. This chapter begins with an examination of the physicochemical properties of phthalate esters that have a significant influence on toxicity.Acute and chronic toxicity are then reviewed with a focus on chronic effects. This chapter concludes with a discussion of studies that have examined the potential for endocrine modulating effects of phthalate esters in aquatic organisms. 1.1 Solubility Influence on Toxicity
For an in-depth discussion of the physico-chemical properties of phthalate esters the reader is referred to Chapter 3 of this handbook. Perhaps most notable about phthalate esters in regards to their physico-chemical properties is that they demonstrate a solubility cut-point, and like many organic chemicals the aquatic toxicity of phthalate esters is strongly influenced by water solubility.As shown in Table 1, phthalate esters have solubilities ranging from 5200 mg/L to less than 1¥10–7 mg/L, with solubility decreasing with increasing molecular weight. There is a solubility cut-point (threshold) for phthalate esters with an alkyl chain length of C6 or greater where the water solubility decreases to less than 1 mg/L. The solubility of BBP is 3.8 mg/L while for DHP it is only 0.159 mg/L. It is appropriate to treat phthalate esters as two groups when discussed in terms of their aquatic toxicological properties; the “lower phthalates” with side chains of less than C6 and “higher phthalates” of C6 or greater side chains. The data show that higher phthalate esters with alkyl chain lengths ≥C6 do not pose intrinsic toxicity to aquatic organisms. For the higher phthalates, in many aquatic toxicity studies neither acute nor chronic effects are found even when the test concentrations are 2–3 times higher than the true water solubility. Parkerton and Konkel [2] in their paper on the application of quantitative structure-activity relationships for assessing the toxicity of phthalate esters have provided a theoretical basis for understanding the lack of acute or chronic aquatic toxicity effects from the higher phthalates. It is well established that phthalate esters are rapidly metabolized in biota and thus the actual bioconcentration factor for these products is considerably lower than would be predicted from their octanol/wa-
265
Aquatic Toxicity of Phthalate Esters Table 1. Solubility of phthalate esters
Name
CAS no.
Abbreviation
Aqueous solubility (mg/L)
Dimethyl phthalate Diethyl phthalate Dibutyl phthalate Butyl benzyl phthalate Dihexyl phthalate Butyl(2-ethylhexyl) phthalate
131-11-3 84-66-2 84-74-2 85-68-7 84-75-3 85-69-8 68151-50-4 25724-58-7 68515-51-5 117-81-7 27554-26-3 26761-40-0 28553-12-0 26761-40-0 68515-49-1 111381-89-6 111381-90-9 111381-91-0 3648-20-2 68515-44-6 68515-45-7 3648-20-2 119-06-2 68515-47-9
DMP DEP DBP BBP DHP BOP
5220 591 9.9 3.8 0.159 0.385
610P
0.0009
DEHP DIOP DINP
0.0025 0.0025 0.0003
DIDP
0.00004
711P
0.0003
DUP DTDP
0.000004 0.00000007
Di-(n-hexyl,n-octyl,n-decyl) phthalate Di(2-ethylhexyl) phthalate Di-iso-octyl phthalate Di-iso-nonyl phthalate Di-iso-decyl phthalate Di(heptyl,nonyl,undecyl) phthalate
Di-undecyl phthalate Di-tridecyl phthalate
Data regarding physical properties of phthalate esters from Cousins and Mackay, this volume.
ter partition coefficients. This fact, together with the low water solubility, suggests that the critical body burden for toxicity is not reached. Another important physico-chemical property of phthalate esters, in relation to aquatic toxicity, is that the higher phthalates appear to form stable emulsions in water due to a self-dispersing property. The higher phthalates dissolve in water up to the limit of solubility and anything beyond this point forms a stable emulsion. This property may account for the considerable range of published data for both water solubility and octanol/water partition coefficient values. More importantly, the stable emulsions may lead to artifactual toxicity in laboratory experiments when, for example, a Daphnia dies after it becomes entrapped in the water-phthalate-air interface.Almost all aquatic toxicity testing conducted on the higher phthalates has been carried out a concentrations far above the water solubility where artifactual toxicity from emulsions is relevant.
2 Acute Toxicity Acute aquatic test results on eighteen different phthalate esters are reviewed in Staples et al. [1], which includes data on both freshwater and saltwater species of
266
C.A. Bradlee and P. Thomas
algae, invertebrates and fish.A discussion on studies with microorganisms is also included. Of the eighteen phthalate esters for which acute data are available, only six esters (DMP, DEP, DAP, DBP, DIBP, BBP) have acute effects consistently across test type and species. Therefore, acute toxicity is demonstrated only in the lower phthalate esters (
3 Chronic Toxicity The results of chronic toxicity studies with fourteen phthalate esters are shown in Tables 3–7. The data in Table 3 are sorted by individual phthalate ester, presented in order of increasing molecular weight and are divided by test species. Tables 4–7 are sorted by tests species and are divided by individual phthalate ester. Whether the exposure concentrations were measured or nominal, test duration, endpoint(s) or criteria and NOEC/LOEC are also presented in Tables 3–7. 3.1 Algae
Chronic toxicity test results on algae are shown in Table 4. Algal assays that have at least a 72-h duration are considered chronic assays [13] because the duration
Phthalate esters ≥C6 Water flea Daphnia magna
Invertebrates Phthalate esters
F F
F F F
F F F F F
DHP, BOP, 610P, F DEHP, DIOP, DINP, DIDP, 711P DUP
DMP DEP DAP DBP BBP
M
M M N M M
M
M M
M M N
Fresh/salt Measured/ (F/S) nominal (M/N)
DHP, BOP, 610P, F DEHP, DIOP, DINP, DIDP, 711P DUP
DBP BBP
S. subspicatus S. capricornutum
Phthalate esters ≥C6 S. capricornutum,
DMP DEP DAP
Phthalate ester
Algae Phthalate esters
Test species
48 h
48 h 48 h 24 h 48 h 48 h
7 d static
72 h 96 h
6 d static 96 h 96 h
Test duration
Survival
Survival Survival Survival Survival Survival
Cell number
Cell number Cell number Cell multiplication inhibition Cell growth Cell number
Test endpoint or criteria
Table 2. Select EC/LC50 values for the acute aquatic toxicity of phthalate esters, as presented in staples et al. [1]
no effect at or above solubility limit
45.9 37.5 26 3.0 1.8
no effect at or above solubility limit
1.2 0.4
142* 85.6 4.5
EC or LC50 (mg/L)
[5, 6]
[5, 6] [4, 5] [48] [5, 6] [3]
[5, 43, 46, 47]
[15] [16]
[5, 43] [44] [45]
Ref.
Aquatic Toxicity of Phthalate Esters
267
F F F F F
Fresh/salt (F/S)
DHP, BOP, 610P, F DEHP, DIOP, DINP, DIDP, 711P DUP
DMP DEP DBP DIBP BBP
Phthalate ester
M
M M M M M
Measured/ nominal (M/N)
96 h static
96 h static 96 h static 96 h static 96 h flow-through 96 h static
Test duration
Survival
Survival Survival Survival Survival Survival
Test endpoint or criteria
no effect at or above solubility limit
39 16.8 1.54 0.9 >0.78
EC or LC50 (mg/L)
[5, 21, 48, 50]
[5, 49] [5, 50] [5, 50] [48] [5, 50]
Ref.
Note: EC50=Effect concentration 50%, LC50=Lethal concentration 50%, M=Measured exposure concentrations used to calculate results, N =Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater. * Values from reference [5] were recalculated from original contract laboratory reports per current USEPA practice. Original values were based on measured initial test concentration; current values are based on averaging initial and final measured test concentrations.
Phthalate esters ≥C6 Fathead minnow P. promelas
Fathead minnow P. promelas Fathead minnow P. promelas Fathead minnow P. promelas Fathead minnow P. promelas Fathead minnow P. promelas
Fish Phthalate esters
Test species
Table 2 (continued)
268 C.A. Bradlee and P. Thomas
Dibutyl phthalate (DBP) Algae S. capricornutum Scenedesmus subspicatus S. subspicatus S. subspicatus Invertebrates Water flea D. magna Water flea D. magna
Invertebrates Water flea D. magna Fish
Algae S. capricornutum
M M M M M M
F F
M
M
F F F F
F
F
F
Fish Rainbow trout Oncorhynchus mykiss
Diethyl phthalate (DEP)
M M
F S M
M
Measured/ nominal (M/N)
F
Fresh/salt (F/S)
Algae Selenastrum capricornutum Invertebrates Water flea Daphnia magna Grass shrimp Palaemonetes pugio
Dimethyl phthalate (DMP)
Test species
Table 3. Chronic toxicity of phthalate esters to aquatic organisms
21 d 21 d
10 d static 7d 72 h 72 h
21 d –
8 d static
60 d post-hatch
21 d 30 d
6 d static
Test duration
Survival Survival/reproduction
Cell number Growth rate Cell growth Growth rate
Survival/reproduction –
Cell number
Growth/survival
Survival Larval mortality
Cell number
Test endpoint or criteria
0.96 (2.5) 0.11 (0.20)
0.21 6.1 0.5 0.5
25 (59) No information
3.65
11 (24)
9.6 (23.0) 10(100)
(64.7)*
NOEC (LOEC) (mg/L)
[7, 8] [21]
[5, 43] [14] [15]
[7, 8]
[5, 43]
[8]
[7, 8] [18]
[5, 43]
Ref.
Aquatic Toxicity of Phthalate Esters
269
Mysid shrimp Mysidopsis bahia Fish Rainbow trout O. mykiss Fathead minnow P. promelas
M M M
F F
M M M M
F F F F
S
M M M M M
M N M
M M M M
Measured/ nominal (M/N)
F F F S S
F F F
Fish Rainbow trout O. mykiss Fathead minnow P. promelas Fathead minnow P. promelas
Butylbenzyl phthalate (BBP) Algae Navicula pelliculosa S. capricornutum S. capricornutum Dunaliella tertiolecta Skeletonema costatum Invertebrates Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna
F F S S
Fresh/salt (F/S)
Invertebrates Water flea D. magna Water flea D. magna Grass shrimp P. pugio Estuarine microcosm
Test species
Table 3 (continued)
109 d 30 d
28 d flow-thru
42 d 21 d 21 d static 21 d flow-thru
96 h 96 h 6 d static 96 h 96 h
60 d post-hatch 20 d flow-thru 144 h flow-thru
21 d 21 d 30 d 2 wk
Test duration
Survival/growth Survival/growth
Survival/reproduction Survival/reproduction Growth/reproduction Reproduction Growth Reproduction/growth
Cell number Cell number Cell number Cell number Cell number
Growth/survival Hatchability Mortality
Survival/reproduction Survival/reproduction Larval mortality Abundance and diversity
Test endpoint or criteria
0.20 0.14 (0.36)
0.26 (0.76) 0.28 (1.4) 0.35 (0.70) 0.26 (0.76) 0.76 0.075 (0.17)
0.3 0.1 (0.10) 0.3 0.1
0.1 (0.19) 0.56 (1.0) 0.32
1.05 (1.91) 0.16 (0.64) 10.0 (50.0) 0.04 (0.34)
NOEC (LOEC) (mg/L)
[8] [16]
[52]
[16] [7, 8] [3] [3]
[16] [16] [5, 43] [16] [16]
[8] [9] [48]
[21] [21] [18] [19, 51]
Ref.
270 C.A. Bradlee and P. Thomas
Di-(2-ethylhexyl) Phthalate (DEHP) Algae S. capricornutum Invertebrates Water flea D. magna Water flea D. magna M M M
F F
21 d static 21 d
6 d static
21 d –
M
F
6 d static
21 d No information
6 d static
111 d
21 d
7 d static
Test duration
M
M –
F –
M
F
M
M
F
F
M
Measured/ nominal (M/N)
F
Fresh/ salt (F/S)
Di-(n-hexyl,n-octyl,n-decyl) phthalate (610P) Algae S. capricornutum F Invertebrates Water flea D. magna F Fish
Butyl(2-ethylhexyl) phthalate (BOP) Algae S. capricornutum Invertebrates Water flea D. magna Fish
Dihexyl phthalate (DHP) Algae S. capricornutum Invertebrates Water flea D. magna Fish Rainbow trout O. mykiss
Test species
Table 3 (continued)
Survival/ growth/reproduction Survival/reproduction
Cell number
Survival/reproduction –
Cell number
Survival/reproduction
Cell number
Survival /growth
Survival/reproduction
Cell number
Test endpoint or criteria
no effect at solubility no effect at solubility
no effect at solubility
no effect at solubility No information
no effect at solubility
no effect at solubility
no effect at solubility
no effect at solubility
no effect at solubility
no effect at solubility
NOEC (LOEC) (mg/L)
[3] [53]
[5, 43]
[7, 8]
[5, 43]
[7, 8]
[5, 43]
[8]
[7, 8]
[5, 43]
Ref.
Aquatic Toxicity of Phthalate Esters
271
Diisononyl phthalate (DINP) Algae Selenastrum capricornutum
Diisooctyl phthalate (DIOP) Algae S. capricornutum Invertebrates Water flea D. magna Fish
Invertebrates Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Grass Shrimp P. pugio Mussel Mytilus edulis Fish Rainbow trout O. mykiss Rainbow trout O. mykiss Brook trout Salvelinus fontinalis Fathead minnow P. promelas Medaka Orizias latipes Stickleback Gasterosteus aculeatus Zebrafish Brachydanio rerio Medaka O. latipes Flagfish Jordanella floridae Guppy Poecilia reticulata
Test species
Table 3 (continued)
M
M
F
F
M
N M N N M N N N N N
F F F F F F F F F F
F
M M M M M M
Measured/ nominal (M/N)
F F F F S S
Fresh/ salt (F/S)
5 d static
21 d –
6 d static
90 d flow-thru 90 d flow-thru 150 d flow-thru 127 d flow-thru 168 d flow-thru 28 d 28 d 28 d 28 d 28 d
21 d 21 d 21 d semi-static 21 d 28 d 28 d
Test duration
Cell number
Survival/reproduction –
Cell number
Growth Hatchability, survival, growth Growth Growth inhibition Growth mortality, growth, sublethal effects mortality, sublethal effects, growth mortality, sublethal effects, growth mortality, sublethal effects, growth mortality, sublethal effects, growth
Survival Survival/reproduction Reproduction Survival/growth/reprod. Larval mortality Mortality, byssal thread attachment
Test endpoint or criteria
no effect at solubility
no effect at solubility No information
no effect at solubility
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
NOEC (LOEC) (mg/L)
[5, 43]
[7, 8]
[5, 43]
[23] [21] [23] [23] [21] [22] [22] [22] [22] [22]
[7, 8] [11] [17] [12] [18] [54]
Ref.
272 C.A. Bradlee and P. Thomas
M M M M M
F F F S
M M
Measured/ nominal (M/N)
F
F F
Fresh/ salt (F/S)
M M
F
F
F
Di-undecyl phthalate (DUP) Algae S. capricornutum M
M
F
Algae S. capricornutum Invertebrates Water flea D. magna Fish Rainbow trout O. mykiss
Di(heptyl,nonyl,undecyl) phthalate (711P)
Fish
Algae S. capricornutum Invertebrates Water flea D. magna Water flea D. magna Water flea D. magna Mussel Mytilus edulis
Diisodecyl phthalate (DIDP)
Fish
Invertebrates Water flea Daphnia magna Water flea Daphia magna
Test species
Table 3 (continued)
8 d static
120 d
21 d
7 d static
–
21 d 21 d 21 d 28 d
8 d static
–
21 d 21 d
Test duration
Cell number
Survival/growth
Survival
Cell number
–
Survival Survival/reproduction Survival/reproduction Mortality, byssal thread attachment
Cell number
–
Survival Survival/growth/reprod.
Test endpoint or criteria
no effect at solubility
no effect at solubility
no effect at solubility
no effect at solubility
No information
no effect at solubility no effect at solubility no effect at solubility no effect at solubility
no effect at solubility
No information
no effect at solubility no effect at solubility
NOEC (LOEC) (mg/L)
[5, 43]
[8]
[7, 8]
[5, 43]
[7, 8] [11] [12] [54]
[5, 43]
[7, 8] [12]
Ref.
Aquatic Toxicity of Phthalate Esters
273
Fresh/ salt (F/S)
M M
F
21 d –
8 d static
120 d
M
F
21 d
Test duration
M
Measured/ nominal (M/N)
Survival –
Cell number
Survival/growth
Survival/reproduction
Test endpoint or criteria
no effect at solubility No information
no effect at solubility
no effect at solubility
no effect at solubility
NOEC (LOEC) (mg/L)
[7, 8]
[5, 43]
[8]
[7, 8]
Ref.
Note: NOEC=No observed effect concentration, LOEC=Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater. * Values from reference [5] were recalculated from original contract laboratory reports per current USEPA practice. Original values were based on measured initial test concentration; current values are based on averaging initial and final measured test concentrations.
Ditridecyl phthalate (DTDP) Algae Selenastrum capricornutum Invertebrates Water flea Daphnia magna Fish
Invertebrates Water flea D. magna F Fish Rainbow trout Oncorhynchus mykiss F
Test species
Table 3 (continued)
274 C.A. Bradlee and P. Thomas
DHP BOP 610P DEHP DIOP DINP DIDP 711P DUP DTDP
Phthalate esters ≥C6 Algae Selenastrum capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum S. capricornutum F F F F F F F F F F
F F F F F F F F F S S
Fresh/salt (F/S)
M M M M M M M M M M
M M M M M M M M M M M
Measured/ nominal (M/N)
7 d static 6 d static 6 d static 6 d static 6 d static 5 d static 8 d static 7 d static 8 d static 8 d static
6 d static 8 d static 10 d static 7d 72 h 72 h 96 h 96 h 6 d static 96 h 96 h
Test duration
Cell number Cell number Cell number Cell number Cell number Cell number Cell number Cell number Cell number Cell number
Cell number Cell number Cell number Growth rate Cell growth Growth rate Cell number Cell number Cell number Cell number Cell number
Test endpoint or criteria
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
(64.7)* 3.65 0.21 6.1 0.5 0.5 0.3 0.1 (0.10) 0.3 0.1
NOEC (LOEC) (mg/L)
[5, 43] [5, 43] [5, 43] [5, 43] [5, 43] [5, 43] [5, 43] [5, 43] [5, 43] [5, 43]
[5, 43] [5, 43] [5, 43] [14] [15] [15] [16] [16] [5, 43] [16] [16]
Ref.
Note: NOEC=No observed effect concentration, LOEC = Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater * Values from reference [5] were recalculated from original contract laboratory reports per current USEPA practice. Original values were based on measured initial test concentration; current values are based on averaging initial and final measured test concentrations.
DMP DEP DBP DBP DBP DBP BBP BBP BBP BBP BBP
Phthalate ester
Phthalate esters
Test species
Table 4. Chronic toxicity of phthalate esters to algae Aquatic Toxicity of Phthalate Esters
275
276
C.A. Bradlee and P. Thomas
of the assay covers several generations of cell growth and, therefore, a no observable effect concentrations (NOEC) or lowest observable effect concentration (LOEC) is reported. Considering ≥ 72-h duration assays as chronic is consistent with the monitored endpoints of population growth rates, cell number and chlorophyll A content, whereas the typical acute endpoint of mortality is not measurable in algal assays. There is chronic toxicity data on fourteen phthalate esters with five different species of algae, including both freshwater and saltwater species. While the lower phthalates have chronic toxicity, the data clearly show that phthalate esters with side chains ≥C6 are not chronically toxic to algae; even at concentrations up to the limit of solubility [5, 14–16]. In the Adams et al. [5] study, fourteen phthalate esters were tested with the freshwater alga Selenastrum capricornutum and only four phthalate esters were chronically toxic (DMP, DEP, DBP and BBP). Studies by Huels [14], Scholz [15] and Gledhill et al. [16] also showed chronic algae toxicity for DBP and BBP. Overall, for the phthalate esters that were chronically toxic to algae the reported NOECs ranged from 0.1 mg/L for BBP to 3.65 mg/L for DEP, which suggests that BBP has the highest chronic toxicity to algae. BBP was also found to have the highest acute toxicity of the
Toxicity test results on invertebrates are shown in Table 5. There is chronic toxicity data on twenty-two phthalate esters with five different species of invertebrates, including both freshwater and saltwater species. While the lower phthalates have chronic toxicity, the data clearly show that phthalate esters with side chains ≥C6 are not chronically toxic to invertebrates; even at concentrations up to the limit of solubility [3, 7, 8, 11, 12, 17, 18]. For phthalate esters
S F F F F F S S F F F F
DMP DEP DBP DBP DBP DBP DBP DBP BBP BBP BBP BBP
BBP
DHP BOP 610P DEHP DEHP
Mysid shrimp Mysidopsis bahia
Phthalate esters ≥C6 Invertebrates Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna
F F F F F
S
F
DMP
Phthalate Fresh/ Ester salt (F/S)
Phthalate esters
Test species
M M M M M
M
M M M M M M M M M M M M
M
21 d 21 d 21 d 21 d static 21 d
28 d flow-thru
30 d 21 d 21 d 21 d 21 d 21 d 30 d 2 wk 42 d 21 d 21 d static 21 d flow-thru
21 d
Measured/ Test nominal duration (M/N)
Table 5. Chronic toxicity of phthalate esters to invertebrates
Survival/reproduction Survival/reproduction Survival/reproduction Survival/ growth/reproduction Survival/reproduction
Reproduction/growth
Larval mortality Survival/reproduction Survival Survival/reproduction Survival/reproduction Survival/reproduction Larval mortality Abundance and diversity Survival/reproduction Survival/reproduction Growth/reproduction Reproduction Growth
Survival
Test endpoint or criteria
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
0.075 (0.17)
10(100) 25 (59) 0.96 (2.5) 0.11 (0.20) 1.05 (1.91) 0.16 (0.64) 10.0 (50.0) 0.04 (0.34) 0.26 (0.76) 0.28 (1.4) 0.35 (0.70) 0.26 (0.76) 0.76
9.6 (23.0)
NOEC (LOEC) (mg/L)
[7, 8] [7, 8] [7, 8] [3] [53]
[52]
[18] [7, 8] [7, 8] [21] [21] [21] [18] [19, 51] [16] [7, 8] [3] [3] [3]
[7, 8]
Ref.
Aquatic Toxicity of Phthalate Esters
277
DEHP DEHP DEHP DEHP DEHP DEHP DIHP DIOP DINP DINP DIDP DIDP DIDP DIDP 711P DUP DTDP L911P L810P DIHP DPHP L79P DIUP DTDP
Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Grass Shrimp P. pugio Mussel Mytilus edulis Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Mussel Mytilus edulis Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna Water flea D. magna
F F F F S S F F F F F F F S F F F F F F F F F F
Fresh/ salt (F/S) M M M M M M M M M M M M M M M M M M M M M M M M
21 d 21 d 21 d semi-static 21 d 28 d 28 d 21 d static 21 d 21 d 21 d 21 d 21 d 21 d 28 d 21 d 21 d 21 d 21 d static 21 d static 21 d static 21 d static 21 d static 21 d static 21 d static
Measured/ Test nominal duration (M/N) Survival Survival/reproduction Reproduction Survival/growth/reprod. Larval mortality Mortality, byssal thread attachment Survival/growth/reproduction Survival/reproduction Survival Survival/growth/reproduction Survival Survival/reproduction Survival/reproduction Mortality, byssal thread attachment Survival Survival/reproduction Survival Survival/growth/reproduction Survival/growth/reproduction Survival/growth/reproduction Survival/growth/reproduction Survival/growth/reproduction Survival/growth/reproduction Survival/growth/reproduction
Test endpoint or criteria
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
NOEC (LOEC) (mg/L)
[7, 8] [11] [17] [12] [18] [54] [12] [7, 8] [7, 8] [12] [7, 8] [11] [12] [54] [7, 8] [7, 8] [7, 8] [12] [12] [12] [12] [12] [12] [12]
Ref.
Note: NOEC=No observed effect concentration, LOEC=Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater.
Phthalate Ester
Test species
Table 5 (continued)
278 C.A. Bradlee and P. Thomas
Aquatic Toxicity of Phthalate Esters
279
The laboratory and field aquaria were allowed to be colonized for 8 weeks (October 5 – November 30, 1981). After colonization, the field tanks were collected and both the field tanks and laboratory tanks were exposed to DBP for 2 weeks (December 1–15, 1981). DBP dissolved into a carrier solution of 60% acetone and 40% distilled water was dosed at nominal concentrations of 0.05, 0.5 and 5 mg/L into a continuous supply of seawater that was pumped to the aquaria at 1.5 L/min. An equal amount of carrier solution (0.5 ml/L) was metered in to the water that entered the control aquaria. Water samples were taken three times weekly and analyzed for DBP content. The measured average and standard deviation of DBP concentrations in the laboratory-colonized aquaria were 0.044 mg/L ±0.014, 0.34 mg/L ±0.12, and 3.7 mg/L ±0.70 and field-colonized aquaria were 0.036 mg/L ±0.005, 0.45 mg/L ±0.14 and 3.8 ±0.62. After a 2-week exposure to DBP, organisms were collected using a 1-mm mesh sieve and were preserved and identified. In the laboratory-colonized communities a total of 2331 animals representing 29 species in 6 phyla were collected, including Chordata, Mollusca, Anthropoda, Annelida, Echinodermata, and Coelenterates. Community structure was clearly reduced at the highest concentration (3.7 mg/L), which showed statistically significant (p=0.05) decreases in the average density and average number of species for all phyla. At the mid-dose (0.34 mg/L) the only phylum affected was Anthropoda which had a reduced average density. The average number of species was not significantly reduced and, moreover, none of the other phyla demonstrated any toxicity. In several instances the density and number of species in the DBP treatments were higher than the control.Arthropods were identified as the most sensitive phyla with a NOEC and LOEC of 0.04 mg/L and 0.34 mg/L, respectively. For the other phylum the NOEC and LOEC are 0.34 mg/L and 3.7 mg/L, respectively. In the field study, significantly fewer organisms were present. Only 181 animals representing 46 species and 7 phyla were collected, which included; Mollusca, Annelida, Rhynchoceia, Anthropoda, Cordata, Echinodermata, and Coelenterates. However, mollusks and annelids dominated in number and relatively few animals from other phyla were collected. The results of this part of the study showed that DBP exposure resulted in a statistically significant reduction in the average density and average number of species only in the phylum Mollusca, and only in the highest concentration. Therefore, the field experiment did not reflect the observations from the laboratory experiment. It should be noted that there was a study design limitation that had the potential to significantly confound the findings. The authors used acetone as a solvent but did not use a solvent-free control, so that any effects of acetone in the dilution water (0.3 mL/L) on the macrobenthic communities were not determined. Moreover, the authors acknowledge that acetone may have effected the bioavailability of DBP thereby artificially altering the toxicity of DBP to the benthic communities. A more relevant study with Daphnia magna further confirms that ≥C6 phthalate esters are not chronically toxic to invertebrates. Brown et al. [12] published results from a 21-day reproduction test performed on Daphnia magna. Twelve phthalate esters were tested, which included: DEHP, L911P, DINP, L810P, DIDP, DIHP, DPHP, 610P, DUP, L79P, DIUP and DTDP. The test was conducted
280
C.A. Bradlee and P. Thomas
using OECD 202 part ii methods that were modified to provide one daphnid per test vessel, as recommended in the more recently modified OECD 211 guideline. Each phthalate ester was tested at a single nominal concentration of 1 mg/L, except DEHP which was tested at 0.25 mg/L. The endpoints for this test were survival of parental organisms (Po), the number of offspring (F1) and the mean body length of surviving Po. Previous attempts to perform chronic studies on daphnids resulted in “floaters” when the concentrations of ≥ C6 phthalates exceeded the stable colloidal suspension concentration of the phthalate (approximately between 200 µg/L for phthalates greater than C6). In order to avoid this problem, the authors used a dispersant, Marlowet R40 (a castor oil 40 mole ethoxylate), with a ratio of phthalate to dispersant of 1:10. The test was conducted in M4 Elendt Schneider medium, using a concentration of 1 mg/L of DEHP in 10 mg/L dispersant and a solvent control at 10 mg/L Marlowet R40 was run beside the negative control. Daphnids were fed daily with a specified quantity of suspension of Chlorella vulgaris cells and Frippak booster. Water quality parameters (O2 concentration, pH, etc) were maintained within the required limits throughout the study. With only one exception (L911P), the concentrations of the phthalates were maintained to plus or minus 20% of the nominal concentration as confirmed by analysis. The results of this study showed there was no significant difference between both controls and any the test concentrations for survival of parental organisms (Po), the number of offspring (F1) and the mean body length of surviving Po. In fact the number of offspring were almost always higher in the phthalate exposures than in the control. Therefore a NOEC >1 mg/L is established for the tested phthalate esters except DEHP that had a NOEC >0.25 mg/L, which was the highest concentration tested. Furthermore, these results demonstrate that mortality in previous studies carried out on D. magna, at concentrations exceeding colloidal stability in water, was likely related to physical effects of phthalate in suspension or as a surface film and not to any toxicological properties of the phthalates. Typically, the route of administration for the phthalate esters in invertebrate studies is via the water; however, there has been a study in which Penaeid shrimp were exposed via food. Hobson et al. [20] conducted a study in which Penaeid shrimp (Penaeus vannamei) were fed diets containing 40 ppm to 50,000 ppm DEHP for 14 days at 4% body weight per day. The results of this study showed no increase in mortality or histopathological alterations in any of the doses. Moreover, whole-body DEHP residues in the shrimp were 18 ppm at the highest dose, and bioconcentration factors were inversely proportional to dose. These findings, as well as findings of other dietary exposure studies, are summarized in Table 6. In conclusion, phthalate esters with alkyl chain lengths equal to or greater than C6 were not toxic to invertebrates at concentrations at or above their solubility limit. For the lower phthalates (
DBP >DMP >DEP and that the saltwater species Mysid shrimp (Mysidopsis bahia) may be a more sensitive species than the freshwater Daphnia magana.
F
F
F
F
DEHP
DEHP
DEHP
DINP
DIDP
Medaka O. latipes
M
M
M
N
N
N
M
Measured/ nominal (M/N)
3 generation
3 generation
4 wks postyolk-sac resorption 7 wks
4 wks
14 d
2 generation
Mortality, histopathology, growth, gonadal: somatic index, sexual development, fecundity, embryonal development, microsomal testosterone metabolism Mortality, histopathology, growth, gonadal: somatic index, sexual development, fecundity, embryonal development, microsomal testosterone metabolism
Peroxisome proliferation
Growth, survival, sex ratio, liver :somatic index
Sex ratio, liver :somatic index
Mortality, molting
Mortality, histopathology, growth, gonadal: somatic index, sexual development, fecundity, embryonal development, vitellogenin induction, hepatic microsomal testosterone metabolism
Test duration Test endpoint or criteria
>20 *
>20 *
>20,000*
>1500 * food mg/kg
300 (1500)
>50,000 *
5 (50)
NOEC (LOEC) (mg/kg food)
[33]
[33]
[24]
[29]
[34]
20]
[
[28]
Ref.
Note: NOEC=No observed effect concentration, LOEC=Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater. * No treatment related effects at highest feed concentration tested.
F
S
F
DEHP
DBP
Phthalate Fresh/ Ester salt (F/S)
Phthalate esters ≥C6 Invertebrates Penaeid shrimp Penaeus vannamei Fish Atlantic salmon Salmo salar Atlantic salmon Salmo salar (Repeat of study Ref. 120) Rainbow trout Oncorhynchus mykiss Medaka O. latipes
Phthalate esters
Test species
Table 6. Feeding studies using phthalate esters with aquatic organisms Aquatic Toxicity of Phthalate Esters
281
282
C.A. Bradlee and P. Thomas
3.3 Fish
Toxicity test results on fish are shown in Table 7. Data are presented on seven phthalate esters with eight different species of fish. While the lower phthalates show chronic toxicity, the data clearly demonstrate that phthalate esters with side chains of ≥ C6 are not chronically toxic to fish, even at concentrations up to the limit of solubility. For phthalates with C6. In chronic studies with Brook trout, Fathead minnow, Medaka, Stickleback, Zebrafish, and Guppy no effects on mortality or growth were seen with exposure to DEHP at or above the solubility limit [21–23]. The results summarized in Table 7 also show that survival can be equally as sensitive and sometimes a more sensitive endpoint than growth; however, other endpoints such has lipid metabolism in fish have been studied also. As shown in Table 6 a dietary study by Henderson and Sargent [24] was performed on DEHP to determine its effects on lipid metabolism in Rainbow trout. The authors carried out the study as DEHP is a known peroxisome proliferator in rat liver and they wished to examine potential effects of this kind in the aquatic environment from exposure via contaminated food. The trout were fed on a diet of freeze dried zooplankton mixed with DEHP at a concentration of 20,000 ppm for a period of seven weeks before sacrifice. Liver homogenates were prepared and assayed for mitochondria and peroxisomes. Lipids were extracted from specific fish tissues and the lipid profile analyzed. Overall percent lipid content was significantly lower than the control in the adipose tissue and the liver; however, the authors did not observe any significant difference in body weight between controls and the exposed group or in liver weight (peroxisome proliferation in rats classically increases liver weight). Prepared liver peroxisomes from DEHP exposed fish did not influence oxidising activity in-vitro compared to the control. The authors concluded that this finding makes it is unlikely that DEHP pose a significant threat to lipid catabolism in fish in the natural environment. In summary, the results of the studies on fish show those phthalate esters with alkyl chain lengths greater than C6 were not toxic to fish at concentrations at or above their solubility limit. This includes mortality, growth and other sub-lethal effects. For the lower phthalates (DBP >DMP >DEP.
DHP DEHP DEHP DEHP DEHP DEHP DEHP
DEHP
DEHP
DEHP
DEHP
711P DUP
Phthalate esters ≥C6 Fish Rainbow trout O. mykiss Rainbow trout O. mykiss Rainbow trout O. mykiss Brook trout Salvelinus fontinalis Fathead minnow P. promelas Medaka Orizias latipes Stickleback Gasterosteus aculeatus
Zebrafish Brachydanio rerio
Medaka O. latipes
Flagfish Jordanella floridae
Guppy Poecilia reticulata
Rainbow trout O. mykiss Rainbow trout O. mykiss
F F
F
F
F
F
F F F F F F F
F F F F F F
M M
N
N
N
N
M N M N N M N
M M N M M M
120 d 120 d
28 d
28 d
28 d
28 d
111 d 90 d flow-thru 90 d flow-thru 150 d flow-thru 127 d flow-thru 168 d flow-thru 28 d
60 d post-hatch 60 d post-hatch 20 d flow-thru 144 h 109 d 30 d
Survival/growth Growth Hatchability, survival, growth Growth Growth Growth mortality, growth, sublethal effects mortality, growth, sublethal effects mortality, growth, sublethal effects mortality, growth, sublethal effects mortality, growth, sublethal effects Survival/growth Survival/growth
Survival Growth Hatchability flow-thru Mortality Survival/growth Survival/growth
Test endpoint or criteria
[8] [23] [21] [23] [23] [21] [22]
[8] [8] [9] [46] [8] [16]
Ref.
no effect at solubility [8] no effect at solubility [8]
no effect at solubility [22]
no effect at solubility [22]
no effect at solubility [22]
no effect at solubility [22]
no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility no effect at solubility
11 (24) 0.1 (0.19) 0.56 (1.0) 0.32 >0.20 0.14 (0.36)
NOEC (LOEC) (mg/L)
Note: NOEC=No observed effect concentration, LOEC=Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater.
DMP DBP DBP DBP BBP BBP
Phthalate Fresh/ Measured/ Test duration ester salt nominal (F/S) (M/N)
Phthalate esters
Test species
Table 7. Chronic toxicity of phthalate esters to fish Aquatic Toxicity of Phthalate Esters
283
284
C.A. Bradlee and P. Thomas
3.4 Endocrine Modulating Effects
Recently, much attention has been given to the potential for endocrine modulating effects of phthalate esters on aquatic organisms. A number of studies have been published on phthalates since 1996 that have shifted focus to estrogenic properties and away from standard chronic endpoints such as growth. The following section summarizes the findings of studies with endocrine endpoints for both the higher and lower phthalates, with the results summarized in Table 8. 3.4.1 Lower Phthalates
For DBP a number of in-vitro and in-vivo studies have been performed suggesting equivocal estrogenic effects of this substance. An in-vitro receptor binding trout hepatocytes assay resulted in an IC50 of 1 mM (0.28 mg/L), a reporter gene assay on yeast was positive above 1 µM but on trout hepatocytes DBP was inactive up to 100 µM (highest concentration tested) [25]. An in-vivo study on trout, DBP intra-peritoneally injected did not increase yolk protein precursor synthesis at a very high dose level (50 mg/kg), while an in-vivo frog sex determination study [26] it did result in sex reversal and inter-sex when tadpoles were exposed to 2.8 mg/L (10 µM) for a 5 day period at the critical life stage. Ohtani et al. [26] examined the estrogenic potential of DBP by examining phenotypically male populations of the Japanese wrinkled frog (Rana rugosa). The sex-determining chromosomes of this species vary from one population to another allowing the authors to produce laboratory bred F1 generation, genetic males (XZ) making these animals particularly suitable for determination of intersex and sex-reversal effects due to external sources. Fifty genetically male eggs were hatched and maintained in water until 19 days after fertilization. Test concentrations of DBP and 17b-Estradiol (E2), as a positive control, were prepared at 0.028, 0.28 and 2.8/L respectively in acetone (100 µl/L) and tadpoles were exposed to these solutions from day 19 to 23 at a temperature of approximately 25 °C. The concentration of DBP was chosen as one tenth of the five minutes LC100 of the tadpoles. Controls were exposed to 100 µl/L acetone only at this time. The tadpoles were then replaced and maintained in uncontaminated laboratory water until day 40 when the gonads of selected tadpoles (approximately 30 per group) were removed for histological inspection. Results from this study are presented in Table 9. At the highest concentration of DBP (2.8 mg/L) the structure of one tadpole gonad was found to be entirely ovarian while four were mixed ovarian and testicular. At 0.28 mg/L two tadpole gonads were observed to contain ovarian and testicular tissue although none were entirely ovarian. For E2, the highest concentration (0.38 mg/L) resulted in gonadal transformation to ovaries in all tadpoles exposed. Despite the relative increase in number of cases of observations of meiotic cells, the authors do not consider this to be an objective measurement of feminization as they noted these cells in many of the tadpoles from the natural population of this species. The authors note that the level of feminization brought
Atlantic salmon Salmo salar (Repeat of study Ref. 120) Rainbow trout Oncorhynchus mykiss Medaka Oryzias latipes
Phthalate esters ≥C6 Fish Atlantic salmon Salmo salar
Medaka Oryzias latipes
Phthalate esters
Test species
F
F
F
DEHP
DEHP
F
DEHP
DEHP
F
F
BBP
DBP
F
DBP
N
M
N
N
M
N
M
90 d
4 wks postyolk-sac resorption 4 wks postyolk-sac resorption 7 wks
2 generation
6 wks
4 d postfertilization
Phthalate Fresh/ Measured/ Test duration ester salt nominal (F/S) (M/N)
Gonadal sex differentiation, total length, wet weight
Peroxisome proliferation
Growth, survival, sex ratio, liver:somatic index
Sex ratio, liver:somatic index
Number of eggs spawned, number of spawnings, egg batch size, gonado-somatic index, vitellogen n synthesis and the secondary sexual characteristics, male fatpad and tubercle formation Mortality, histopathology, growth, gonadal:somatic index, sexual development, fecundity, embryonal development, vitellogenin induction, hepatic microsomal testosterone metabolism
Gonadal sex differentiation (feminization of males)
Test endpoint or criteria
Table 8. Supplemental endpoints addressing potential mechanisms of chronic toxicity using aquatic organisms
Ref.
>20,000* mg/kg food >5,000* mg/L
>1500* mg/kg food
300 (1500) mg/kg food
5 (50) mg/kg food
>100 * mg/L
[31]
[24]
[29]
[34]
[28]
[29]
0.28 (2.8) mg/L [26]
NOEC (LOEC) (mg/L)
Aquatic Toxicity of Phthalate Esters
285
DIDP
Medaka O. latipes
F
F
M
M
3 generation
3 generation
Mortality, histopathology, growth, gonadal:somatic index, sexual development, fecundity, embryonal development, microsomal testosterone metabolism Mortality, histopathology, growth, gonadal:somatic index, sexual development, fecundity,embryonal development, microsomal testosterone metabolism
Test endpoint or criteria
>20 * mg/kg food
>20 * mg/kg food
NOEC (LOEC) (mg/L)
[33]
[33]
Ref.
Note: NOEC=No observed effect concentration, LOEC=Lowest observed effect concentration, M=Measured exposure concentrations used to calculate results, N=Nominal concentrations used to calculate results (whether or not concentrations were analytically confirmed), F=Freshwater, S=Saltwater. * No treatment related effects at highest feed concentration tested.
DINP
Phthalate Fresh/ Measured/ Test duration ester salt nominal (F/S) (M/N)
Phthalate esters ≥C6 Fish Medaka O. latipes
Test species
Table 8 (continued)
286 C.A. Bradlee and P. Thomas
287
Aquatic Toxicity of Phthalate Esters
Table 9. Histological evaluation of genetically male Rana rugosa gonads following E2 and DBP
exposure Concentration (mg/L)
Histological evaluation of gonads Ovarian throughout
DBP 0.028 0.28 2.8 17b-Estradiol 0.0038 0.038 0.38 Solvent control
Ovarian and testicular
Testicular (many meiotic germ cells)
Testicular (many meiotic germ cells)
Total
0 0 1
0 2 4
1 8 14
29 20 11
30 30 30
1 5 28 0
4 14 0 0
10 3 0 2
13 8 0 28
28 30 28 30
Data from Ohtani et al. [26]
about by the highest concentration of DBP is similar to that of the lowest concentration of E2 used, and they conclude that in this study DBP is about 1000 times less potent than E2. Overall the results suggest that DBP does have an estrogenic effect under specific conditions. However, conclusions on environmental and population effects of such endocrine modulation cannot be drawn from the frog study as the sex reversal observed on frog gonads occurred at concentrations of DBP that are known to be detrimental to other fauna based on other direct toxicological effects. For example, the high concentration in the frog study was 2.8 mg/L which is comparable to many LC50s from acute studies on DBP. Moreover, a LOEC of 0.28 mg/L is established based on effects in the frog testes (2 intersex out of thirty frogs), which is nearly thirty times higher than the recommended maximum permissible concentration (MPC) of 10 µg/L DBP [27] (based on the 60 d NOEC for growth of rainbow trout, divided by 10). The relevance of the frog study can also be questioned at another level. Monobutyl phthalate is the primary metabolite of DBP and it is not active in invitro studies; therefore, the metabolism of DBP may be of primary importance in deactivating the diester to the more estrogenically inactive monoester. The exposure concentrations in the frog study were high enough to saturate the enzymes systems such that the gonads may have been exposed to the unmetabolized DBP. At lower, environmentally relevant, concentrations the DBP is sufficiently metabolized that endocrine effects may not occur.A more pertinent study on possible endocrine effects of DBP was conducted by Patyna et al. [28]. Medaka was used to examine the multigenerational effects of DBP in a dietary exposure study by Patyna et al. [28]. Also, a positive control consisting of 17bEstradiol (E2) was conducted as part of this study. Seven treatment groups were studied, which included an ethanol control, 0.5, 5, 50 mg DBP/kg food, and 0.05, 0.5, 5 mg E2/kg food. An ethanol control was tested because each phthalate was
288
C.A. Bradlee and P. Thomas
spiked into the food via an ethanol solution. Each treatment group was divided into two replicate tanks of 20 fish each. The F0 generation was first exposed as 14day old larvae by feeding either DBP or E2 in dry flake food at a daily ration of 5% body weight. The F0 and F1 generations were fed each treatment through sexual maturation. The endpoints for the study were mortality, histopathology, growth, gonadal:somatic index (GSI), sexual development, fecundity, embryonal development, vitellogenin induction, and hepatic microsomal testosterone metabolism. The authors found that all the fish exposed to 0.05 mg E2/kg food were phenotypic females and that no eggs were produced by the E2 treated fish. In contrast, DBP had no significant treatment related effects on the F0 generation; however, the authors found that in females 50 mg DBP/kg food decreased F1 generation ovary weight, GSI, and egg production and in males it reduced testes weight and GSI in the F1 generation. DBP at 50 mg DBP/kg food increased microsomal protein levels, liver weight and hepatic somatic index in both F1 generation males and females. No adverse effects were reported in the F1 and F2 eggs from the DBP treated groups and, furthermore, they showed normal embryonic development. The sex ratios in DBP treated groups were similar to the control. As with DBP, BBP equivocal estrogenicity has been observed between in-vitro and in-vivo studies [25]. A trout hepatocyte receptor binding assay had an IC50 for E2 of 10 µM and a yeast reporter gene assay was positive above 1 µM. For the monoesters monobutyl and monobenzyl phthalate, metabolites of BBP were inactive at concentrations between 500 µM and 1 mM. But while in one in-vivo study on trout injected intra-peritoneally with 50 mg DEHP/kg an induction of vitellogenin in one fish out of six was observed, and in a water-borne in-vivo BBP study on Pimephales promelas by Harries et al. [29] no reproductive effects were observed up to concentrations of 82 µg/L. These data suggest that metabolism of BBP may occur in-vivo and under environmental conditions the substance does not arrive intact at the estrogen receptors. In a review of estrogenic assays by Moore [25] the relevance of in-vitro tests on yeast strains, trout hepatocytes and in-vivo studies on trout by intra-peritoneal injection to phthalates was discussed. For the lower phthalates, receptor binding assay data using trout hepatocytes exist for DBP with an IC50 at 1 mM BBP had an IC50 for E2 binding of approximately 10 µM and maximum inhibition of 60%. Relative binding of the phthalates compared to E2 in this study was 0.1% supporting the in-vivo results found by Ohtani et al. [26] for DBP. Results from yeast (Saccharomyces cerevisiae) reporter gene assays transfected with human estrogen receptor performed on DEP, DBP, DiBP and BBP were found to demonstrate estrogenic activity. DBP, DiBP and BBP elicited responses at concentrations above 1 µM reaching a plateau at about 50% implying they are partial agonists for E2. Mono-n-butyl phthalate and monobenzyl phthalate were found to be inactive at concentrations between 500 µM and 1 mM. Trout hepatocytes are also used in reporter gene assays and will synthesize vitellogenin when exposed to estrogens but DBP did not elicit a response at concentrations up to 100 µM. In-vivo studies on the synthesis of fish oocyte proteins, vitellogenin and zona radiata have been performed on juvenile Rainbow trout using DBP and BBP [25]. DBP did not elicit oocyte protein synthesis at a dose of DBP injected intra-peri-
Aquatic Toxicity of Phthalate Esters
289
toneally (i.p.), and BBP induced vitellogenin production above control level in one out of six trout injected. Some doubt exists as to the reason for this effect reported on one fish as neither the age nor the sex of the animals is reported. At i.p. doses of 5 and 50 mg/kg BBP increased the level of hepatic estrogen receptors (but this was not statistically significant) but significantly decreased the level of zona radiata in plasma while E2 was found to increase both oocyte proteins. Moore [25] concludes that it is essential that the substance tested should be in the form in which it will be found in the organism considered. Phthalate diesters are hydrolyzed to the monoester form in fish gills, for example, and so the use of intra-peritoneal injection of the diester in these in-vivo studies is questionable. The same is true for the in-vitro studies which may not be relevant to the situation in the environment. Harries et al. [29] measured the reproductive performance of pair breeding fathead minnow (Pimephales promelas) over six weeks using BBP and nonylphenol (NP). Number of eggs spawned, number of spawnings, egg batch size, gonado-somatic index, vitellogenin synthesis, secondary sexual characteristics, male fatpad and tubercle formation were assessed prior to and after three week exposure. The concentrations selected, up to 100 µg/L nominal (69 to 82 µg/L measured), were lower than the aquatic solubility of both BBP and NP. Stock solutions of both substances were made up using methanol as a carrier solvent to allow use of a flow-through testing apparatus. The authors stated that BBP had no significant effect on fecundity although there was a significant decrease in number of spawnings compared to the control there was also a significant increase of egg batch size per spawning. No discernible effects of BBP were observed on any of the other reproductive parameters, while NP was found to elicit a response on fecundity, GSI and secondary sexual characteristics even at the lowest concentration used (between 1 and 10 µgNP/L measured). It should be noted that, based on the published bar charts, the results from the solvent control more closely resembled results from BBP exposure than those of the negative control. In conclusion, the studies have demonstrated equivocal findings from in-vitro and in-vivo studies on the endocrine modulating effects of DBP and BBP. Insufficient information exists on other lower phthalates to draw conclusions about the environmental relevance of assays suggesting estrogenic activity for the diesters, DEP and DIBP. 3.4.2 Higher Phthalates
For the higher phthalates, an in-vitro assay using DEHP did not result in receptor binding even at one order of magnitude above the water solubility of the diester and was negative in a reporter gene assay. DHP, DIDP and the monoester MEHP were negative in yeast reporter gene assays while results for DINP were at best equivocal. In his review on in-vitro studies on the higher phthalate esters, Moore [25] stated that assays on receptor binding using trout hepatocytes did not result in a response for DEHP at 2 µM but reduced E2 binding at higher concentrations with a maximum of 25% at 1 mM. However, the solubility limit of this substance
290
C.A. Bradlee and P. Thomas
is 0.008 µM (3 µg/L) and positive results above this level should be treated with caution, as non-selective binding is likely to occur. Using yeast reported gene assays, Harris et al. [30] showed that DINP demonstrated an equivocal response; however, the results are confounded by the fact that the DINP used in the study may have contained a supplement of nonyl phenol. DHP, DEHP, DIDP and several monoesters, mono(2-ethylhexyl) phthalate (MEHP), and its oxidative metabolites, were all inactive between 500 nM and 1 mM. A trout hepatocyte reporter gene assay using DEHP failed to elicit an estrogenic response at concentrations up to 100 µM. Metcalfe et al. [31] performed an in-vitro yeast estrogenicity reporter gene screening assay and also an in-vivo “90 day” Japanese medaka (Oryzias latipes) assay. In both cases, the organisms were exposed to 500, 1000 and 5000 µg/L concentrations of DEHP. Other substances tested were estriol, 17b-estradiol, 17aethinylestradiol, NPEO, 4-NP and bisphenol A (all at doses considerably lower than for DEHP). Exposure was initiated 1 day after hatch in a static renewal system. Stock solutions were prepared by dissolving the tests substance in acetone (maximum 5 µl/L). Sixty fish were exposed to the tests substances in aqueous solution. For DEHP concentrations of 0, 500, 1000 and 5000 µg/L were employed despite the fact that DEHP is only water soluble at concentrations of up to about 3 µg/L. While it may be possible to generate emulsions of DEHP in laboratory tests that greatly exceed aqueous solubility, such an experimental design confounds test interpretation because toxicity due to physical effects may be misinterpreted as intrinsic toxicity of DEHP. Moreover, concentrations that far exceed aqueous solubility have little environmental relevance. Solutions were replaced every 48-h over the duration of the assay. Study termination time was based on fish size (1.5 cm) and not on the duration so the length of the assay was between 85 and 110 days post hatch. The authors report that a background contamination level of DEHP was found in all solutions in which acetone was used and they consider that this came from the tubing and joints of the test system which were made of PVC. All measured concentrations reported were equivalent to the nominal plus the background. Fish were embedded whole in paraffin and microtomed for light microscopical examination of the testes or ova. The authors examined two end points: change in expected sex ratio of the fish, and presence of testis-ova (at least one oogonia in the testes), hypothesized to signal the presence of estrogenic substances (although this could not be confirmed). For the yeast assay no responses were observed for DEHP or for certain NPEOs while NP and bisphenol A were found to be positive at 5 orders of magnitude greater than E2. In the medaka in-vivo study the results were generally consistent with those of the in-vitro assay. DEHP did not induce a positive response and the authors conclude that this is in line with other in-vitro assays that show that DEHP is not an estrogen agonist. Other studies examining endocrine effects following aquatic exposure to higher phthalates have been performed on fish. Shioda and Wakabayashi [32] exposed adult male medaka (Oryzias latipes) to concentrations of DEHP (0.1, 0.3 and 1 µM equivalent to 10 to 100 times the water solubility), nonylphenol (up to 0.3 µM) and bisphenol-A (up to 10 µM) for a period of two weeks. After this time each male fish was placed together with two females and resulting
Aquatic Toxicity of Phthalate Esters
291
spawning parameters were observed. The number of eggs per spawn and the number of eggs hatched was measured every day for a week. DEHP had no adverse effects on the number of eggs spawned or on hatch compared to the control. Bisphenol-A significantly decreased the number of eggs and hatch from 10 µM but for NP no significant differences could be observed because group variation was so high. Patyna et al. [33] performed a study on Japanese medaka that shows neither DIDP nor DINP elicit an effect on reproduction or development after three generations of dietary exposure. Four treatment groups were studied that included an untreated control, acetone control, 20 mg DIDP/kg food, and 20 mg DINP/kg food. An acetone control was tested because each phthalate was spiked into the food via an acetone solution. Each treatment group was divided into five replicate tanks of 50 fish each. The F0 generation was first exposed as 14-day old larvae by feeding either DIDP or DINP in dry flake food at a daily ration of 5% body weight. The F0 and F1 generations were fed each treatment through sexual maturation and oviposition (140 days-post-hatch). The test ended during the F2 generation prior to sexual maturation and the F0 adults were terminated at day 123. The endpoints for the study were mortality, histopathology, growth, gonadal:somatic index (GSI), sexual development, fecundity, embryonal development, and microsomal testosterone metabolism. The authors found that there were no statistically significant changes in either fecundity or mortality in both the DIDP and DINP treatment groups. Normal embryonic development of the F1 generation was observed except for a transient decrease in red blood cell pigmentation in the DIDP, DINP and acetone treatment groups. The presence of this effect in the acetone groups suggests the transient pigmentation effects are not related to phthalate exposure. The only histopathologic change observed in the F0 adults was a slight alteration in hepatocellular staining around the central vein. The sex ratios and microsomal testosterone metabolism in DIDP and DINP treated groups were similar to the untreated control. Overall, the authors concluded that neither of the phthalate esters elicited an effect on reproduction or development of Japanese medaka after three generations of dietary exposure. Two in-vivo dietary studies on Atlantic salmon (Salmo salar) using DEHP have been performed. In 1999, Norrgren et al. [34] published a study that indicated some increase in the proportion of female fish at the highest dose tested. The second study, Norman et al. [35], was conducted by the same laboratory in 2001 as a direct repeat of the 1999 study but used an improved methodology. Improvement included analyzing the amount of DEHP in the feed and using a broader range of doses to better define the NOEC. The primary purpose of these studies was to evaluate if DEHP exposure would alter the normal sex ratio of males to females, and although a 1:1 (50%:50%) male:female may be considered the normal sex ratio for Atlantic salmon the authors provide no information on the range of normal variations to the 1:1 sex ratio within populations of Atlantic salmon. Overall, the findings of the first study in 1999 that showed some higher proportion of females; however, these findings were not confirmed in the repeat study of 2001. In the 1999, study exposure was initiated by providing one replicate per group of 200 fish with food spiked with DEHP at nominal concentrations of 300 and
292
C.A. Bradlee and P. Thomas
1500 mg DEHP/kg food four weeks after hatch (immediately after yolk-sac resorption), and continued for four weeks. The dietary concentration of DEHP was not measured in this study and as such is considered to be a study design limitation because the actual exposure doses of DEHP are not confirmed. Spiked concentrations of 17b-Estradiol (E2) at 15 and 30 mgE2/kg and 4-n-nonylphenol (NP) at 300 and 1500 mg NP/kg were also run in this first test. The fish were maintained for a further four months by providing them with uncontaminated food until maturity such that histological results on gonadal development could be obtained. Between 82 and 184 individuals per group were dissected to remove the liver and gonads and the sex was determined by light microscopic evaluation. In an extension of this study, salmonids were injected intra-peritoneally with various concentrations of DEHP, NP and PCB two to four times over a 17-day exposure period. The authors determined the liver :somatic index (LSI) and found that for the two concentrations of E2 and the highest concentration of DEHP, (but not for NP) there was a significant increase in this parameter compared to the control. This finding was contrary to the findings of Henderson and Sargent [24] who exposed trout to a far greater concentration and for a longer time period, although based on the weight (30–50 g) of the fish used in the experiment exposure was started at a later life stage. The environmental significance of increased LSI for fish is unclear but this is not an indicator of estrogenic activity. E2 was found to increase female:male sex ratio at 15 and 30 mg E2/kg to 88 and 100% females respectively. At the highest concentration of DEHP a statistically significant ratio change was also noted with 64% of the population determined histologically as female. In the extension to the 1999 study, intra-peritoneally injected fish were analyzed for vitellogenin at the end of the exposure period. The results showed no statistically significant difference in vitellogenin levels between the control and any of the tested substances. Overall in the 1999 study [34] the authors concluded that for the endpoints of sex ratio, LSI, and vitellogenin the NOEC is considered as 300 mg DEHP/kg. Owing to the large difference between the NOEC and the LOEC and also to the fact that there was no true dose response relationship, Norrgren initiated a repeat this study. The 2001 repeat study focused on re-examining the results of the feeding study with DEHP and its apparent effect on sex ratio; therefore the repeat study did not include E2 or NP exposures and interperitoneal injections to examine for vitellogenin were not conducted. The repeat study utilized a similar dosing regime to the first study but used dietary concentrations of 400, 800 and 1500 mg DEHP/kg food. The concentration of DEHP in the food was measured at the beginning and end of the study and was found to have excellent agreement with the targeted concentrations. The endpoints for the repeat study included survival, length and weight in addition to sex ratio and LSI index. The findings of the repeat study did not concur with the first study with regards to effects on the LSI or sex ratio. The first study showed a statistically significant (p<0.05) difference in the sex ratio between the control and the 1500 mg DEHP/kg food dose level where there were 49% and 64% females, respectively. A 50%:50% sex ratio of male:female is
Aquatic Toxicity of Phthalate Esters
293
thought to be normal; however, the authors provide no information on the normal range of variation within populations of Atlantic salmon. A second finding from the 1999 study showed a statistically significant (p<0.01) difference between the control and the 1500 mg DEHP/kg food dose level for the LSI, where the LSI for the control and DEHP were 1.74 and 2.22, respectively. In the repeat study neither the sex ratio nor LSI were statistically different as compared to the control. The number of females in the control and 1500 mg DEHP/kg food were 50% and 51%, respectively; while the LSI for the control and the high DEHP dose were 1.20 and 1.11, respectively. Moreover, no effects on survival or mean length and weight were seen at the high DEHP dose level (1500 mg DEHP/kg food). Another finding of the 2001 study was a slight, reversible, increase in testisovo.An examination of the gonads from the 1500 mg/kg exposure concentration showed a small (3%), but statistically significant increase in the incidence of testis-ovo. This occurred only in the highest concentration. This effect was only temporary as a second sampling of the exposed fish after a five month recovery period where the fish were maintained in DEHP-free water showed no significant incidence of testis-ovo. There is no clear understanding of the significance of the observed, reversible, testis-ovo; however, significant long-term adverse impacts on populations Atlantic salmon is not expected. In study by Freeman et al. [36] performed on Atlantic cod testes and head kidneys in which the organs were incubated in-vitro with equimolar amounts of radiolabeled pregnenolone and progesterone with various concentrations of cold or radiolabeled DEHP. They examined the steroid metabolites and analyzed for DEHP metabolic breakdown products. The authors found no differences from controls at any of the concentrations used, up to 1000 mg DEHP/kg tissue. Moreover, perhaps unexpectedly, DEHP was not metabolized to MEHP or any other degradation product, echoing Moore’s [25] cautionary conclusion on in-vitro studies when using hydrolyzable test substances. The authors subsequently tested DEHP on steroid metabolism of Atlantic cod in vivo in order to determine effects of potential metabolites. Fish were fed gelatine capsules containing a dietary concentration of 0, 10, 100 or 1000 mg DEHP/kg wet food, twice weekly, over 121 days. The testes, ovaries and corresponding head kidneys were prepared and assayed with radiolabeled pregnenolone and progesterone prior to X-ray auto-radiographical analysis while the remaining fractions were isolated and identified by sequential thin layer and paper chromatography followed by scintillation spectrometry. Histological examinations were also carried out on certain tissues. No abnormalities were found in male fish compared to the control in any of the DEHP dosed groups, both in terms of the steroid metabolism and histological parameters examined. In female cod the authors reported a significantly, and dose related, increased synthesis of one of the steroid intermediates (designated S) at 100 and 1000 mg/kg wet weight, suggesting a change in biosynthetic pathway, but they were unclear as to the significance of this change and no histological changes in female fish were reported. No changes were found in male cod. Van Wezel et al. [27] proposed Environmental Risk Limits (ERL) or Maximum Permissible Concentrations (MPC) for the higher phthalate, DEHP, coming to the conclusion that 0.19 µg/L would be a suitable ERL for this compound based on
294
C.A. Bradlee and P. Thomas
ecotoxicology and environmental chemistry. The ERL (or MPC) value is based on the original Larsson and Thurén [37] frog egg study using a NOEC of 10 mg/kg fresh weight and an uncertainty factor of 10 to provide a PNEC of 1 mg/kg f.w. The authors consider that, in standard sediment the organic carbon content at 1 mg/kg f.w. would be 17.5 mg/kg o.c. (based on a Koc of 87¥103 l/kg o.c.) and that this would be in equilibrium with 0.2 µg/L in the aquatic phase. The authors were aware of a second study performed by Wennberg et al. [38], but discounted it due to the difference in study length (60 days for the former as opposed to 29 days for the latter). They also mention problems of counting and bacterial infection with the Wennberg et al. [38] study. They do not mention that the reason for the difference in study time was due to the low temperature employed (5 °C in the first study as opposed to 10 °C in the second) which virtually doubled egg incubation time. Moreover, they do not mention the numerous problems and unknown quantities (ethanol concentration, spiking method, control mortality compared to the later study) encountered in the Larrson and Thurén [37] study. Finally, they did not take into account the second repeat frog egg study performed at both 5 and 10 °C, at two different organic carbon concentrations and accepted by the EU commission as the definitive frog egg study on DEHP. Neither the Wennberg et al. [38] nor the IVL study found any effects at any concentration used up to 450 mg/kg in the former and 1000 mg/kg in the latter study. The EU risk assessment states that no study on DEHP conclusively shows any effect in the aquatic compartment and so the ERL recommended by Van Wezel et al. [27] is not appropriate for DEHP.
6 Discussion The aquatic toxicity of phthalate esters varies widely across the class of compounds, with only the lower molecular weight esters (C1 to C4) consistently posing toxicity. Phthalate esters higher than about C6 pose no toxicity up to their aqueous solubility limits. The lack of toxicity for the higher phthalates is related to their relative insolubility in water and their ready metabolism by aquatic organisms, so that the critical body burden for toxicity is not reached [2]. The toxicity of the lower molecular weight (C1 to C4) phthalate esters to aquatic organisms has been widely documented [1, 2]. These publications have focused on studies that measured ecologically relevant endpoints related to survival, growth and development, and reproductive fitness of populations of aquatic organisms. Parkerton and Konkel [2] calculated predicted no effect concentrations for DMP, DEP, DBP and BBP using the abundant acute and chronic toxicity data, acute to chronic ratios and statistical calculation procedures. The PNECs calculated by the authors using three slightly different procedures were 3109 to 4780 µg/L for DMP, 865 to 1173 µg/L for DEP, 43 to 62 µg/L for DBP and 38 to 60 µg/L for BBP. The range of chronic NOEC values were 9600 to 10,000 µg/L for DMP, 3650 to 25,000 µg/L for DEP, 40 to 10,000 µg/L for DBP and 75 to 350 µg/L for BBP. Thus the PNEC values calculated by Parkerton and Konkel [2] are demonstrated to be amply protective of the aquatic environment as they are below the range of available valid chronic data for these compounds.
Aquatic Toxicity of Phthalate Esters
295
Recently, the issue of endocrine disruption in wildlife species has generated considerable scientific, political, and public interest [39–41]. Endocrine-disrupting chemicals can, as a consequence of their molecular structure, bind to hormone receptors and may mimic or antagonize the action of the natural hormonal ligand. Adverse impact should be judged on the basis of environmentally relevant endpoints that affect populations and communities of organisms, such as survival, growth, and reproductive success. Moreover, studies that utilize inter-peritoneal injection (i.p.) as the route of administration do not accurately evaluate effects from phthalates esters because they do not account for metabolism. Inter-peritoneal injection may be a relevant route of exposure for substances that do not undergo metabolism before they enter the systemic circulatory system or those substances otherwise found systemically in the administered form. However, i.p. administration does not accurately represent the plasma metabolite profile for phthalates, and consequently the systemic response, resulting from environmental exposure to phthalates, because it bypasses the normal “firstpass” metabolic pathways for the phthalates (esterase activity in the gill), nor does it offer a measure of the inherent activity of the diesters themselves, because they are metabolized in the plasma and liver. The evidence for endocrine modulation effects caused by phthalate esters is equivocal at best. Studies from a few in-vitro and in-vivo assays that were designed to identify endocrine modulating compounds have suggested that DBP and BBP, but no other phthalate ester, were capable of interacting with estrogenic receptors. In addition, another study suggested that BBP was also anti-androgenic [42]. Testis-ova in male gonads and induction of hepatic vitellogenin have occasionally been reported for some phthalate esters at very high doses. These findings occurred at concentrations already considered to be toxic based on conventional studies, were not supported upon repetition of the original study or were reported in studies that employed injection of test material into individual fish. The abundant numbers of high quality studies testing a range of phthalate esters that have examined ecologically relevant endpoints have consistently reported some toxicity with the lower esters and no toxicity in the higher esters. It is using these studies measuring ecologically relevant endpoints that should be used for risk assessment purposes and for developing appropriate water quality criteria.
7 References 1. Staples CA,Adams WJ, Parkerton TF, Gorsuch JW, Biddinger GR, Reinert KH (1997) Aquatic toxicity of eighteen phthalates esters. Environ Toxicol Chem 15:875–981 2. Parkerton TF, Konkel WJ (2000) Application of quantitative structure – activity relationships for assessing the aquatic toxicity of phthalate esters. Ecotoxicol Environ Saf 45:61–78 3. Adams WJ, Heidolph BB (1985) Short-cut chronic toxicity estimates using Daphnia magna. Aquatic Toxicology and Hazard Assessment: Seventh Symposium,ASTM STP 854, Cardwell RD, Purdy R, Bahner RC (eds) American Society for Testing and Materials, Philadelphia, PA, pp 87–103 4. Passino DRM, Smith SB (1987) Acute bioassays and hazard evaluation of representative contaminants detected in Great Lakes fish. Environ Toxicol Chem 6(11):901–907
296
C.A. Bradlee and P. Thomas
5. Adams WJ, Biddinger GR, Robillard KA, Gorsuch JW (1995) A summary of the acute toxicity of 14 phthalate esters to representative aquatic organisms. Environ Toxicol Chem 14:1569–1574 6. Springborn Bionomics, Inc (1984) Acute toxicity of fourteen phthalate esters to (Daphnia magna). Chemical Manufacturers Association, Washington, DC 7. Springborn Bionomics, Inc (1984) Chronic toxicity of fourteen phthalate esters to Daphnia magna. Chemical Manufacturers Association, Washington, DC 8. Rhodes J, Adams WJ, Biddinger GR, Robillard KA, Gorsuch JW (1995) Chronic toxicity of 14 phthalate esters to Daphnia magna and rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem 14:1967–1976 9. McCarthy JF, Whitmore DK (1985) Chronic toxicity of di-n-butyl and di-n-octyl phthalate to Daphnia magna and the fathead minnow. Environ Toxicol Chem 4:167–179 10. Scholz N (1995) Determination of the effect of Vestinol AH (DEHP) on the swimming behavior of Daphnia magna. Complies with Directive 92/69/EEC. Final Report DK-631, Marl, Germany 11. Brown D, Thompson RS (1982) Phthalates and the aquatic environment: Part I. The effect of di-2-ethylhexyl phthalate (DEHP) and di-isodecyl phthalate (DIDP) on the reproduction of Daphnia magna and observations on their bioconcentration. Chemosphere 11: 417–426 12. Brown D, Croudace CP, Williams JJ, Shearing JM, Johnson PA (1998) The effect of phthalate ester plasticisers tested as surfactant stabilised dispersions on the reproduction of the Daphnia magna. Chemosphere 36:1367–1379 13. Versteeg D.J (1990) Comparison of short- and long-term toxicity test results for the green alga, Selenastrum capricornutum. Plants for Toxicity Assessment. ASTM STP 1091, Wang W, Gorsuch JW, Lower WR (eds) American Society for Testing and Materials, Philadelphia, PA, pp 40–48 14. Huels AG (1991) Untersuchung über den Einfluss von Di-n-butylphthalat auf Scenedesmus subspicatus. Huels AG, unveroffentliche (12.03.91), Marl, Germany 15. Scholz N (1995) Determination of the effect of Vestinol C (DBP) on the growth of Scenedesmus subspicatus 86.81. SAG. Complies with Directive 92/69/EEC. Final Report AW392, Marl, Germany 16. Gledhill WE, Kaley RG,Adams WJ, Hicks O, Michael PR, Saeger VW, LeBlanc GA (1980) An environmental safety assessment of butyl benzyl phthalate. Environ Sci Tech 14:301–305 17. Scholz, N (1995) Determination of the effects of Vestinol AH (DEHP) on the reproduction rate of Daphnia magna. OECD Guideline 202 Part II. Final Report DL-160, Marl, Germany 18. Laughlin RB Jr, Neff JM, Hrung YC, Goodwin TC, Giam CS (1978) Effects of three phthalate esters on the larval development of the grass shrimp Palaemonetes pugio (Holthuis). Water, Air, Soil Pollut 9:323–336 19. Tagatz ME, Plaia GR, Deans CH (1986) Toxicity of dibutyl phthalate-contaminated sediment to laboratory and field-colonized estuarine benthic communities. Bull Environ Contam Toxicol 37:141–150 20. Hobson JF, Carter DE, Lighter DV (1984) Toxicity of a phthalate ester in the diet of a penaied shrimp. J Toxicol Environ Health 13:959–968 21. DeFoe DL, Holcombe GW, Hammermeister DE, Biesinger KE (1990) Solubility and toxicity of eight phthalate esters to four aquatic organisms. Environ Toxicol Chem 9:623–636 22. van den Dikkenberg RP, Canton HH, Mathijssen-Spiekman LAM, Roghair CJ (1989) Usefulness of Gasterosteus aculeatus – the three-spined stickleback – as a test organism in routine toxicity tests. Rijksinstitut voor de Volksgezondheid en Milieuhygiene, Bilthoven, Netherlands, pp 28 23. Mayer FL, Mehrle PM, Schoettger RA (1977) Collagen metabolism in fish exposed to organic chemicals. In: Taub RA (ed) Recent Advances in Fish Toxicology (ed) EPA 600/3–77–085, U.S. Environmental Protection Agency, Corvallis, OR, pp 31–54 24. Henderson RJ, Sargent JR (1983) Studies of the effects of Di-2-ehthylexyl Phthalate on Lipid Metabolism in Rainbow Trout Fed Zooplankton Rich Wax Esters. Comp Biochem Physiol 74C:325–330 (1983)
Aquatic Toxicity of Phthalate Esters
297
25. Moore NP (2000) The oestrogenic potential of the phthalate esters. Repro Toxicol 14:183–192 26. Ohtani H, Miura I, Ichikawa Y (2000) Effects of Dibutyl Phthalate as an environmental endocrine disruptor on gonadal sex differentiation of genetic males of the frog Rana rugosa. Environ Health Perspec 108:1189–1193 27. van Wezel AP, van Vlaardingen P, Posthumus R, Crommentuijn GH, Sijm DT (2000) Environmental risk limits for two phthalates, with special emphasis on endocrine disruptive properties. Ecotoxicol Environ Saf 46:305–321 28. Patyna PJ, Thomas PE, Cooper KR (1999) Multigeneration reproductive effects of di-nbutyl phthalate in Japanese medaka (Oryzias latipes). Toxicologist. 48(1-S):262 29. Harries JE, Runnalls T, Hill E, Harris C, Maddix S, Sumpter JP, Tyler CR (2000) Development of a reproductive performance test for endocrine disrupting chemicals using pair-breeding fathead minnow (Pimephales promelas). Environ Sci Technol 34:3003– 3011 30. Harris CA, Henttu P, Parker MG, Sumpter JP (1997) The estrogenic activity of phthalate esters in vitro. Environ Health Perspect 105:802–811 31. Metcalfe CD, Metcalfe TL, Kiparissis Y, Koenig BG, Khan C, Hughes RJ, Croley TR, March RE, Potter T (2001) Estrogenic potency of chemicals detected in sewage treatment plant effluents as determined by in-vivo assays with Japanese medaka (Oryzias latipes). Environ Toxicol Chem 20:297–308 32. Shioda T,Wakabayashi M (2000) Effect of certain chemicals on the reproduction of medaka (Oryzias latipes). Chemosphere. 40:239–243 33. Patyna PJ, Parkerton TF, Davi RA, Thomas PE, Cooper KR (1998) Evaluation of two phthalate ester mixtures in a three generation reproduction study using Japanese medaka (Oryzias latipes). Toxicologist (1998 42(1-S):338–339 34. Norrgren L, Blom A, Andersson PL, Borjeson H, Larsson DGJ, Olsson PE (1999) Effects of potential xenoestrogens (DEHP, nonylphenol and PCB) on Sexual Differentiation in Juvenile Atlantic Salmon (Salmo salar). Aquatic Ecosys Health Manag 2:311–317 35. Norman L, Norrgren L, Borjeson H (2001) Exposure of Atlantic salmon (Salmo salar) to DEHP-contaminated food. Manuscript in Preparation 36. Freeman HC, Sangalang GB, Burns BG, McMenemy M (1980) The Effects of Di-(2-ethylhexyl) phthalate (DEHP) on Steroid Metabolism in the Atlantic Cod Gadus morhua. Proceedings of the 7th Annual Aquatic Toxicity Workshop, Montreal, Canada 11/5–7/1980, pp 198–226 37. Larsson P, Thuren A (1987) Di-2-ethylhexylphthalate inhibits the hatching of frog eggs and is bioaccumulated by tadpoles. Environ Toxicol Chem 6:417–422 38. Wennberg L, Parkman H, Remberger M,Viktor T,Williams C (1997) Influence of sedimentassociated phthalate esters (DEHP and DIDP) on hatching and survival of the moorfrog, Rana arvalis. Govt Reports Announcements & Index (GRA&I), Issue 19 39. Coburn T, vom Saal FS, Soto AM (1993) Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environ Health Perspect 101:378–84 40. Kendall RJ, Dickerson RL, Giesy JP et al. (eds) (1998) Principles and Processes for Evaluating Endocrine Disruption in Wildlife. SETAC Press, Pensacola, FL, USA 41. Tattersfield L, Matthiessen P, Campbell P et al. (eds) (1997) SETAC – Europe/OECD/EC Expert Workshop on Endocrine Modulators and Wildlife: Assessment and Testing, WMWAT; Veldhoven, The Netherlands, 10–13 April 1997. Brussels: Society of Environmental Toxicology and Chemistry-Europe 42. Sohoni P, Sumpter JP (1998) Several environmental oestrogens are also anti-androgens. J Endrocrinol 158:327–339 43. Springborn Bionomics, Inc (1984) Toxicity of fourteen phthalate esters to the freshwater green algae (Selenastrum capricornutum). Chemical Manufacturers Association, Washington, DC 44. Suggatt RH, Foote K (1981) Comprehensive review of acute aquatic toxicity data on phthalate esters. Final Report, Contract SRC TR 81–537. Syracuse Research Corp., April 1981
298
C.A. Bradlee and P. Thomas: Aquatic Toxicity of Phthalate Esters
45. Kuhn R, Pattard M (1990) Results of the harmful effects of water pollutants to green algae (Scenedesmus subspicatus) in the cell multiplication inihibition test. Water Res 24:31–38 46. Scholz N (1995) Determination of the effect of Vestinol AH (DEHP) on the growth of Scenedesmus subspicatus 86.81. SAG. Complies with Directive 92/69/EEC. Final Report AW391, Marl, Germany 47. Wilson WB, Giam CS, Goodwin TE, Aldrich A, Carpenter V, Hrung YC (1978) The toxicity of phthalates to the marine dinoflagellate Gymnodinium breve. Bull Environ Contam Toxicol 20:149–154 48. Springborn Bionomics, Inc (1983) Acute toxicity of thirteen phthalate esters to fathead minnow (Pimephales promelas) under flow-through conditions. Chemical Manufacturers Association, Washington, DC 49. Springborn Bionomics, Inc (1983) Acute toxicity of fourteen phthalate esters to fathead minnows (Pimephales promelas). Chemical Manufacturers Association, Washington, DC 50. Geiger DL, Northcott CE, Call DJ, Brooke LT (1985. Acute toxicities of organic chemicals to fathead minnows (Pimephales promelas) vol 2. Center for Lake Superior Environmental Studies, University of Wisconsin, Superior, WI, pp 326 51. Tagatz ME, Deans CH, Moore JC, Plaia GR (1983) Alterations in composition of field- and laboratory-developed estuarine benthic communities exposed to di-n-butyl phthalate. Aquat Toxicol 3:239–248 52. Springborn Bionomics, Inc (1986) Chronic toxicity of butylbenzyl phthalate to mysid shrimp (Mysidophsis bahia). Final report, Springborn Bionomics, Inc. Monsanto Company, St. Louis, MO 53. Knowles CO, McKee MJ, Palawski DU (1987) Chronic effects of di-2-ethylhexyl phthalate on biochemical composition, survival and reproduction of Daphnia magna. Environ Toxicol Chem 63:201–208 54. Brown D, Thompson RS (1982) Phthalates and the aquatic environment: Part II. The bioconcentration and depuration of di-2-ethylhexyl phthalate (DEHP) and di-isodecyl phthalate (DIDP) in mussels (Mytilus edulis). Chemosphere. 11(4):427–435
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 299– 316 DOI 10.1007/b11470
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters Raymond M. David 1 · Gerhard Gans 2 1 2
Health and Environment Laboratories, Eastman Kodak Company, 1100 Ridgeway Avenue, Rochester, NY 14652-6272, USA. E-mail: [email protected] BASF Corporation, Washington, DC, 20005, USA
Phthalate esters can be divided into three general categories based on size. These categories are low-MW esters used as solvents in cosmetics or as plasticizers of cellulose acetate polymers, mid-MW esters used as solvents in some PVC products but only in combination with other plasticizers in floor coverings, or as a solvent or plasticizer in the cosmetic and pharmaceutical industries, and high-MW esters used as plasticizers of PVC for wire and cable coverings, medical products, and other consumer products. Phthalate esters are data-rich for effects in laboratory animals, having been the focus of much research because of their effects on the biochemistry of liver cells, the effects on the testes, and the effects on the development of laboratory animals. All phthalate esters have little or no toxicity following single (acute) exposures. These substances are not dermal sensitizers, but may produce minor skin irritation with prolonged exposure to the neat chemical. Long-term hazards from short-term exposures are either minimal or reversible because many long-term effects are observed only following continuous exposure. Long-term effects such as liver cancer only occur in laboratory animals following life-time or near life-time exposure to doses of >100 mg kg–1 d–1 of high-MW esters. Cancer is thought to occur through a mechanism that involves biochemical changes in the liver cells of rats and mice. These biochemical changes are not seen in primates. As a result, scientists do not regard humans to be at risk of cancer from exposure to phthalate esters. Reproductive toxicity or developmental effects in the offspring of laboratory animals exposed to midMW phthalates during gestation have been reported. Reproductive toxicity is the result of damage to the testes causing lack of sperm production. It is not known if such effects occur in humans, but adult primates have been resistant to the effects seen in adult rodents. Developmental effects can also occur with high exposure to mid-MW esters. The effects include incomplete skeletal formation in the head, spinal cord, tail, and ribs. In addition, male rats demonstrate incomplete formation of the urogenital tract. It is not known if primates or humans are also susceptible to these effects, and the mechanism in laboratory animals is unknown. Exposure to phthalate esters is not thought to cause respiratory diseases such as asthma because these substances are hypoallergenic, but some have tried to associate exposure with an increased sensitivity to respiratory allergens. Sufficient data are lacking to make such a correlation. Phthalate esters are not neurotoxic. In evaluating the concern for humans, many principles of toxicology are discussed to provide the reader with sufficient understanding of species extrapolation. Primates are not as sensitive to phthalate esters as are rodents. There may be a variety of reasons for this lack of sensitivity, for example, lower absorption and different metabolic pathways. There are also intrinsic differences in the responses of primate and human cells to the biochemical effects of phthalate esters. While this difference relates directly to the likelihood of cancer, it may also impact the sensitivity to other effects seen in animals. Thus, predicting the effects in humans must convey some level of uncertainty. Keywords. Phthalate esters, Toxicity
© Springer-Verlag Berlin Heidelberg 2003
300
R.M. David and G. Gans
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
1
Introduction
1.1 1.2 1.3 1.4
Assumptions Used in Toxicology . . . Exposure . . . . . . . . . . . . . . . . Uncertainties in the Assessment of Risk Uses . . . . . . . . . . . . . . . . . . .
2
Short-Term Exposure
2.1 2.2 2.3 2.3.1 2.3.2 2.3.3
Oral/Dermal/Inhalation . . . . . . . . . . . Irritation/Sensitization . . . . . . . . . . . . Long-Term Risks from Short-Term Exposure Metabolic Effects . . . . . . . . . . . . . . . Regenerative Effects . . . . . . . . . . . . . Developmental Effects and Cancer . . . . . .
3
Prolonged Exposure
3.1 3.2 3.3 3.4
Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reproductive/Developmental Toxicity and Endocrine-Related Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Asthma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neurotoxicity . . . . . . . . . . . . . . . . . . . . . . . . . .
4
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
301 302 303 305
. . . . . . . . . . . . . . . . . . . . . . . . 306 . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
306 306 307 307 308 309
. . . . . . . . . . . . . . . . . . . . . . . . . 309 . . . 309 . . . 311 . . . 313 . . . 313
Abbreviations NTP DEHP DOP DnOP DINP DMP DEP BBP DIDP DiBP DUP MW PVC CDC TDI LOEL NOEL BUA DNA
National Toxicology Program di(2-ethylhexyl) phthalate dioctyl phthalate di-n-octyl phthalate diisononyl phthalate dimethyl phthalate diethyl phthalate butylbenzyl phthalate diisodecyl phthalate diisobutyl phthalate diundecyl phthalate molecular weight poly(vinyl chloride) Centers for Disease Control and Prevention tolerable daily intake lowest observed effect level no-observable-effect level Beratergremium für Umweltrelevante Altstoffe (advisor committee for environmental relevant old materials) deoxyribonucleic acid
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
IARC PPAR RfD ATSDR
301
International Agency for Research on Cancer peroxisome proliferator activated receptor oral reference dose Agency for Toxic Substances and Disease Registry
1 Introduction 1.1 Assumptions Used in Toxicology
Toxicology is not an exact science. It deals with probability or the likelihood that some event or biological response will occur.A toxicologist predicts biological responses in one species, or in one individual, based on the responses of another species or based on the response of a population. To predict the effects on human health, a toxicologist relies on laboratory animals as surrogates for humans. Laboratory animals are smaller, have a shorter lifespan, and can be bred to a homogenous population. They are inexpensive and relatively easy to maintain. The rat has become a popular test species for toxicological study. Over the past 50 years, scientists have come to realize that not all rats are the same, and breeding laboratories have purposely bred laboratory animals with slightly different characteristics to be used in different types of studies. In addition, testing laboratories have become more aware of animal diseases, diet, and husbandry procedures, which result in better survival with fewer background effects. Because of improved survival, fewer background lesions, and better husbandry of animals, data from recent toxicity studies are likely to be more reliable predictors of human health effects than data from 50 years ago. Although the rat is a cornerstone of toxicology, scientists have recognized in recent years that responses in rats do not always reflect responses in humans, that is, rats are not small people [1–3]. This fact was first recognized for phthalate esters in the 1980s. After the National Toxicology Program (NTP) reported liver cancer in rats and mice exposed to high levels of one phthalate ester, di(2-ethylhexyl)phthalate (DEHP), also known in the industry as DOP, which is different from DnOP, researchers began to understand the mechanism of this cancer. It is thought that liver cancer is the result of biochemical changes in the liver cells, a process called peroxisome proliferation [4]. Because several therapeutic drugs used to lower cholesterol also produce these biochemical effects in rats, scientists began looking at responses in human cells to determine if patients receiving these drugs were at risk for cancer. Some of the first studies using cultured human liver cells demonstrated that the biochemical changes that occurred in rats did not occur in human cells [5–8]. These studies were followed by others using monkeys, which also failed to demonstrate the same biochemical changes observed in rats [9–12]. Thus, it appears that humans may not be at risk for cancer via this mechanism because human liver cells do not undergo the kind of biochemical changes that rat cells do. Some of the most recent information suggests that other effects observed in rats, such as effects on the testes, also do not occur in monkeys [11, 12]. Thus, a rat cannot always predict effects in humans, at least
302
R.M. David and G. Gans
not for phthalate esters. Why that is true is not clear. There are many differences between rats and humans (or monkeys) in how phthalate esters are absorbed, broken down (metabolized) in the body, and excreted [13, 10]. There may also be intrinsic differences between rat cells and human cells in their capacity to respond, that is, rat cells may be far more susceptible to the effects of phthalate esters than are human cells. Whatever the reason, it is not appropriate to automatically assume that effects in laboratory animals accurately predict effects in humans, at least not for phthalate esters. 1.2 Exposure
Exposure is a key element in assessing the potential risk of adverse health effects. Unfortunately, determining the exposure of humans to phthalate esters has always been difficult because, in part, phthalate esters are found in many products, especially products in the laboratory, that lead to false positive or suspicious analytical results. In addition, the medium (air, water, food, article) that contains the phthalate ester greatly influences the biological response because phthalate esters are very water-insoluble (especially high-molecular weight (MW) esters) and will be absorbed only slowly through any membrane (intestinal wall from ingestion of water or food, lung from inhalation of air particles, skin from contact with articles). These issues are discussed in Chapter 2. Hydrolysis of the phthalate diester to a monoester enhances absorption. Studies have shown that for high-MW esters, breakdown (metabolism; hydrolysis) of the ester bond to liberate one alcohol and a remaining monoester greatly increases the absorption, since the monoester and alcohol are absorbed more rapidly than the diester [14–16]. The more hydrolysis occurs, the more monoester is available for absorption. Once absorbed, the monoester continues to be metabolized into substances that are excreted in the urine [17]. The fact that high-MW phthalate esters need to be hydrolyzed to be absorbed is an important factor that relates to the medium (route) of exposure. For example, if exposure is through the skin from contact with poly(vinyl chloride) (PVC) articles containing phthalate esters, the absorption is very slow because the ester is not hydrolyzed and must be absorbed intact [18]. Experiments with laboratory animals and human skin have demonstrated that the absorption rate of phthalate esters is slow for high-molecular weight esters [19, 20]. Therefore, high levels of exposure of the skin to neat chemical might not result in adverse health effects because the absorption is so slow and the metabolism of the ester is minimal. Likewise for inhalation exposure, absorption is likely to be slow but faster than absorption through the skin and slower than absorption from ingestion. This is because the lungs have some capacity to hydrolyze the ester [21], but it is not as predominant as it is in the intestine [22]. Not only is the medium (or route) of exposure (inhalation, ingestion, dermal) important in assessing the potential human health effects, the form of the ester is also important. For example, air might contain vapors of neat chemical or particles of PVC containing phthalate esters. Inhaling vapors of neat chemical would have a greater chance for absorption than inhaling PVC dust because the ester is
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
303
released from the PVC matrix very slowly [18]. In addition, particles such as PVC dust may get lodged in the nose or other airways rather than reaching the deep lung, further reducing the possibility of absorption. Laboratory experiments have determined that dermal contact with articles that contain phthalate esters results in minute quantities transferred to the skin, much less than if a worker is exposed to the neat chemical. Thus, the form of the exposure has a significant impact on any potential health effect. The route of exposure that results in the most efficient absorption of phthalate esters is ingestion. Laboratory studies have demonstrated, however, that rats are far more efficient at hydrolyzing the esters and, subsequently, absorbing the monoester than primates (and presumably humans) [9, 10]. This means that when studies of phthalate esters are conducted in laboratory animals where health effects are observed following very high doses of an ester, it is very difficult to reproduce such effects in primates (and presumably humans) because primates do not absorb phthalate esters as efficiently as other laboratory animals. Primates and humans absorb about seven times less phthalate than do rats (especially for high-MW esters) [9].At low doses, the absorption may be more comparable. Exposures to phthalate esters have been estimated by numerous government agencies. Based on concentrations of phthalates in the air, water, and food supply, the primary source of exposure is thought to be food. Minor levels of low- and high-MW phthalates can migrate into food from packaging or inks. These levels and the total exposures from food are generally low (estimated to be less than 30 µg kg–1 d–1). Recent studies in the United States conducted by the Centers for Disease Control and Prevention (CDC) have measured the levels of excreted monoester in approximately 300 individuals as a means of evaluating total exposure [23]. The study was extended to about an additional 1000 individuals with similar findings. The CDC results indicate that exposure to highMW phthalate esters is lower than estimated by consumption of food. Conversely, exposure to low-MW phthalates is higher than estimated from levels in food. This finding should not be startling because low-MW phthalate esters are used in consumer products such as pharmaceuticals and cosmetics. In any event, the levels are at or below calculated tolerable daily intake (TDI) values in the United States (Table 1). Exposure levels are presented and discussed in detail in Chapters 5 and 8. The concept of TDI and how it is derived is discussed below. 1.3 Uncertainties in the Assessment of Risk
Risk is the predicted frequency of occurrence of an adverse effect of a chemical substance from a given exposure to humans. Therefore, an indispensable prerequisite is a dose-response relationship, which correlates an adverse effect (hazard) with an internal or external dose (exposure) allowing the identification of a dose without any effect. The stages of a risk assessment include hazard identification, exposure assessment, and risk estimation. Each of these steps provides a source of uncertainty to the whole process.
304
R.M. David and G. Gans
Table 1. Exposure levels for common phthalate esters based on urinary metabolites
Phthalate
Geo. mean
95th percentile
Estimated intake
RfD
DEP DBP BBP DEHP DINP
12.34 a 1.45 0.73 0.60 0.21
93.33 6.37 3.34 3.05 1.08
57 b 7 6 30 c 10.8
800 100 200 20 ND d
a b c d
All values in mg kg–1 d–1 based on a maximum creatinine clearance of 20 mg kg–1 d–1. Estimated intake taken from ATSDR, IPCS, or EU draft risk assessments. From Doull J, Cattley R, Elcombe C, Lake BG, Swenberg J, Wilkinson C, Williams G, Van Gemert M (1999) Reg Toxicol Pharmacol 29:327–357 using ATSDR estimates. ND not determined.
Hazard identification can be obtained from human case reports, studies of human volunteers, mechanistic information, and animal studies. Each of these methods has a drawback and introduces some level of uncertainty into the evaluation process. For example, the major problem in using case reports is the uncertainty with respect to exposure levels. Case studies rarely provide adequate data on actual exposure concentrations. Only studies of human volunteers, performed under well-defined exposure conditions, are best able to correlate actual exposure with human responses. On the other hand, for studies of human volunteers often only subjective parameters are reported. Therefore, these studies rely heavily on the personal judgment of the test subjects, which can be biased by perception. For example, in a study of a malodorous chemical, subjects exposed to the malodorous substance will likely report more effects than subjects exposed to the same chemical but with an additive that masks the odor. Uncertainty may also be introduced when relying on the results of mechanistic studies. Special attention should be paid to mechanistic studies in order to identify a relevant end point for the human situation. Mechanistic studies provide a good basis for extrapolation from lower species to higher species, or they may provide a basis for discounting effects in lower species when considering the potential to occur in higher species. An example is the effect of some phthalate esters to modify the biochemistry of the liver of rats and mice that results in liver cancer. Mechanistic studies demonstrate convincingly that humans are not affected in the same way as rats and mice. The use of mechanistic studies to identify hazards in lower species, which pertain to higher species, assumes that all relevant interspecies physiologic processes are the same. Uncertainty is introduced when we assume that the processes are the same, but we have not demonstrated such equivalency. For animal studies, the most important information to be obtained is the doseresponse relationship, identifying effect (lowest observed effect level or LOEL) and no-effect level (no observed effect level or NOEL). Determining these regulatory values is dependent upon the selection of dose levels used in the experiment. Typically, three dose levels are used in animal studies: one at a maximum level that can just be tolerated by the animal, one at an intermediate level, and one that should be at a no-effect level. Selecting the dose levels for any study is sci-
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
305
entific judgment based on preliminary information, but there is a practice of using multiples of three between dose levels. Not every investigator follows this practice. Occasionally, investigators will use multiples of ten to establish dose levels. When evaluating the data from multiple studies, one should not assume that the lowest NOEL is the best value to establish risk.A reasonable NOEL should be based on all available studies, not on only the lowest dose without any effect from a single study. Once a reasonable NOEL is determined, the next step in risk assessment is to compare the NOEL from animal studies to the exposure levels for humans. Assessment of exposure levels also introduces some uncertainty because exposures of the general public are only estimated. One way in which exposures can be estimated is to determine the levels in products/foods/environment and extrapolate to exposure level based on contact/consumption/physiology. The measurement of phthalate esters in a matrix carries with it inherent difficulties. The application of an analytical method, which may liberate matrix-bound phthalate esters by using organic solvents, will overestimate the biologically available proportion. Furthermore, there could be significant differences between the external and internal exposure levels. For example, an air concentration measurement that includes dust particles will overestimate the exposure because some particles will be large enough that they will be impacted or retained in the nasal cavity.Another example is the absorption of a substance from the gut into the bloodstream may not reach 100% and a certain amount will be excreted in the feces without being absorbed. These examples need to be taken into account when developing a reasonable and scientifically based risk assessment. A goal of the comparisons between NOEL values from animal studies and human exposure is to determine the margin of safety (also called the margin of exposure) to determine if humans are at risk of adverse health effects. In the absence of variable data for human exposure, regulators can establish a TDI or in the case of EPA, a reference dose (RfD) which sets an upper limit for human exposure. These values take into account the NOEL value and the severity of the response, and assign a margin of safety for human exposure by using certain assumptions about the variability within a species and variability between species. Thus, a TDI value is typically 100 times lower than the NOEL from animal studies. This assumes that variability among animals can be accounted for by a factor of ten, that is, the most sensitive individual animal is likely no more than ten times more sensitive than the rest of the population, and variability between species can be accounted for be a factor of ten, that is, humans are no more than ten times more sensitive than rats. 1.4 Uses
Phthalate esters comprise the diesters of phthalic anhydride with alcohol moieties ranging from C1 to C13. These esters were developed to serve various technical demands. Table 2 summarizes the main characteristics of the phthalate esters [24]. This list is not comprehensive. Some low-MW phthalates can appear in a variety of products, some of which are consumer products. The manu-
306
R.M. David and G. Gans
Table 2. Uses of common phthalate esters
Phthalate Ester
Use
DMP DEP DBP
Specialty plasticizer used in cellulose esters and in cosmetics
DBP DiBP BBP
Very good gelling agent for PVC, used in combination with other plasticizers; in floor coverings; as a plasticizer in the cosmetic and pharmaceutical industries
DEHP DINP DIDP
Standard plasticizer for PVC, moderate volatility, DEHP used in medical applications
DUP
Specialty plasticizer
facturers of these products incorporate the phthalates because of the properties they impart to the product. Generally, such applications use small quantities of phthalate esters.
2 Short-Term Exposure 2.1 Oral/Dermal/Inhalation
The primary, common aim of toxicological testing is to provide information that can be used to assess the potential for adverse effects in humans. Short-term tests identify effects after a single high-dose administration and an observation period of 14 days. These tests reflect the possible hazards resulting from a single high exposure, which can occur in the workplace after an accidental release of phthalates during production, maintenance, or transportation. In general, phthalates show little acute toxicity. A summary of LD50 values for the most important commercial phthalates is given in the following table (compiled from producer material safety data sheets and various reports from the German regulatory agency, BUA, Table 3). Because most regulatory bodies and scientists consider LD50 of greater than 5000 mg kg–1 to reflect a lack of toxicity, these data demonstrate that phthalates are relatively non-toxic following acute exposure. 2.2 Irritation/Sensitization
The ability of a chemical to irritate the skin or eyes after local administration is usually determined in rabbits. For the phthalates, there is only slight or moderate irritation reported after administration of the undiluted substance (compiled from producer material safety data sheets and various BUA reports). There is also
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
307
Table 3. Acute lethal oral doses for common phthalate esters
Phthalate ester
Oral LD50 (rat) (mg kg–1 bw)
Dermal LD50 (rabbit) (mg kg–1 bw)
DMP DEP DBP DiBP DEHP DINP DIDP DUP
> 5,000 > 9,000 > 8,000 >15,000 >30,000 >10,000 >20,000 >15,800
>12,000 >20,000 >20,000 >10,000 >24,500 > 3,100 > 3,600 > 7,900
no evidence for a sensitization potential of the phthalate esters from animal studies. There are a few case reports describing potential skin sensitization after dermal contact with phthalate-containing products. Based on limited information of possible interference with other substances, these reports are not conclusive. In a comprehensive volunteer study with seven dialkyl (C6-C13) phthalate esters in a human-repeated, insult patch test using the modified Draize procedure, no evidence of dermal irritation or sensitization for any of the seven phthalates was observed [25]. These data suggest that phthalate esters are not likely to play a role in the induction of asthma, or other allergic or irritation-induced response (vide infra). 2.3 Long-Term Risks from Short-Term Exposure
Toxicologists evaluate the short-term consequences of single exposures and the effects of prolonged exposure, up to a lifetime. Because some tissues repair damage easily, the question of the long-term consequences of brief, sub-lethal exposures may be difficult to answer. Typically, tissues that grow rapidly or have a high rate of turnover, repair damage well. Tissues which repair easily include the blood synthesizing system (hematopoietic), gastrointestinal tract, testes (not ovaries), liver, and kidneys [26]. The nervous system repairs itself very poorly. All of this assumes that the damage to the cell, tissue, or organ is not lethal and does not irreversibly alter the architecture of the organ. 2.3.1 Metabolic Effects
Phthalate esters are a class of chemicals that in laboratory rodents cause changes in metabolism. Some enzyme systems are induced, while others appear to be suppressed. Such effects occur primarily in the liver (the primary drug-metabolizing organ of the body) or in the kidneys [4, 27–29]. There is little evidence that adverse effects occur in other tissues or organs [30]. The effect can be observed macroscopically as larger liver or kidneys (or increased organ weight), micro-
308
R.M. David and G. Gans
scopically as liver cells that contain more endoplasmic reticulum and increased size and number of organelles called peroxisomes, and biochemically as increased enzyme activity. Short-term exposure in animals of as little as five days is adequate to produce such effects. These effects will persist as long as the exposure and will subside within one to two weeks following cessation of exposure [4]. Because the effects are transient and metabolic, they are considered to be “adaptive” in that the organs adapt to the presence of the phthalate. Not all phthalate esters produce such metabolic changes. Low-MW phthalates do not induce these metabolic changes [31, 32]; mid-MW phthalates are weak inducers, while high-MW phthalates are the best inducers among phthalates [33, 32]. Exceptions to this rule exist [34]. It is known that some of the metabolic changes are unique to laboratory animals, specifically rats and mice [5, 11, 10, 35–38]. The relevance for metabolic changes in humans has been debated for nearly 15 years [3, 6, 8, 39–41]. 2.3.2 Regenerative Effects
Occasionally, short-term exposure can result in immediate tissue damage, but the nature of the tissue is such that the damage can be repaired. Such regeneration is limited to organs that have a high capacity for cell turnover and growth. For example, exposure of very young laboratory animals to high-dose levels (>1000 mg kg–1) of mid-MW phthalate esters (including DEHP) causes a reduction in sperm production or aspermatogenesis.While this can be a severe adverse effect, over time the testes are able to recover and produce normal amounts of sperm [42]. For any organ such as the testis, the timing of exposure relative to the life cycle of the organism is an important factor in recovery. Studies comparing the effect on the testes of young (immature) laboratory animals to old (mature) animals indicate that lower dose levels of transitional phthalates and DEHP have effects in young animals, but that older animals are insensitive to such lower levels [43]. The reason for the age-related difference in sensitivity is not known, but is likely related to the formation of the blood-testis barrier, which forms prior to adolescence. Following formation of this barrier, less toxicant would likely reach the testes to cause effects. Another organ capable of repair is the liver. Studies in laboratory animals using long-term exposure to high-MW phthalates followed by a period of nonexposure indicate that some liver effects are reversible. For example, a lesion associated with the Ito cell of the liver (spongiosis hepatis) or pigmentation of Kupffer cells is reversed during recovery even after prolonged exposure of threequarters of a lifetime [44]. In addition, some effects considered to be end-stage, such as neoplastic changes in the liver, may be partially reversible [45]. This suggests that short-term exposure would not be expected to result in permanent damage such as these types of cellular changes in later life. Studies with DEHP have demonstrated that short-term exposure does not cause long-term effects in the liver or testes [46, 42]. However, not all possible health effects have been investigated.
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
309
Another organ that has been identified as a target for high-MW phthalate esters is the kidney. Effects on the kidneys vary depending on the length of exposure. Following short-term exposure, there is evidence of renal tubular degeneration and proliferation, and cyst formation [47]. Over an extended period of exposure, mineralization of the renal tubules may occur, as well as chronic nephropathy. These cellular changes also occur in untreated animals and are part of a normal aging process [48, 49]. Thus, exposure to high-MW phthalates may not be solely responsible for the effects observed. Instead, it is possible that these larger substances enhance the normal process of aging in laboratory animals. Therefore, the relevance to humans is unclear [50]. 2.3.3 Developmental Effects and Cancer
Although covered in a later section specifically on developmental toxicity, it bears noting that mid-MW phthalates including DEHP can have an effect on the developing embryo/fetus of laboratory animals following short-term exposure. Rats appear to be more sensitive than mice with dose levels of approximately 100 mg kg–1 demonstrating no overt effect. Some very subtle effects have been seen in rats. Low-MW phthalates and high-MW phthalates (excluding DEHP) produce effects on the developing embryo at dose levels at least ten times higher than the effect dose levels for mid-MW phthalates. This will be discussed further in a later section. The potential of phthalate esters to produce cancer following long-term exposure will be discussed below. Occasionally, questions arise about the potential to develop cancer following short-term exposure. Based on animal studies, liver cancer associated with exposure to some high-MW phthalates occurs only after prolonged exposure [45, 51, 52].
3 Prolonged Exposure 3.1 Cancer
Carcinogens can be classified into two different groups: DNA-reactive carcinogens (genotoxic) and epigenetic carcinogens. The genotoxic category comprises carcinogens that chemically interact with DNA, based on an inherent reactivity or on metabolism-mediated reactivity. The second category is comprised of carcinogens for which there is evidence of a lack of direct interaction with genetic material, and for which another biologic process has been identified that could be the basis for carcinogenicity. This differentiation has major implications for human risk assessment. Genotoxic carcinogens may induce a mutagenic or precancerous effect after a single exposure. Conversely, epigenetic carcinogens produce cancer only after a sustained level of exposure. Thus, for epigenetic carcinogens, it may be possible to establish a “safe” threshold or even demonstrate that the underlying biochemical mechanism is not relevant for humans.
310
R.M. David and G. Gans
In studies in rats and mice, DEHP and DINP induce liver tumors after high exposure [45, 53–55]. These phthalates belong to a diverse group of chemicals and therapeutics, which induce specific changes called peroxisome proliferation in the liver of rats and mice [4]. In the last decade, a better understanding of the role of peroxisome proliferation in the carcinogenic process has been achieved. This understanding has led to the consensus that cancer induced by peroxisome proliferation is not relevant for humans [1, 39]. Recently, the IARC re-assessed the cancer classification of DEHP taking into account the mechanism of carcinogenicity [27, 56, 57]. In summary, they concluded: – “The weight of evidence for DEHP and its metabolic products demonstrates that they do not act as direct DNA-damaging agents. – DEHP produces liver tumors in rats and mice. – Under conditions of the bioassay, DEHP induces peroxisome proliferation and cell replication in liver that are characteristic of a peroxisome proliferator in mice and rats. – Rodent peroxisome proliferators exercise their pleiotropic effects in liver due to activation of PPARa. This process is essential for liver hypertrophy and hyperplassia and eventual hepatocarcinogenesis in response to peroxisome proliferation. – Hepatic peroxisome proliferation has not been adequately evaluated in studies of human livers following exposure of DEHP in vivo; however, the effect of treatment of human and mouse hepatocytes with DEHP metabolites which are active in rat hepatocytes, as well as other peroxisome proliferators, indicate that humans can reasonably be predicted to be refractory to induction of peroxisome proliferation and hepatocellular proliferation by DEHP. The evidence indicates that the mechanism of peroxisome proliferation induced by DEHP in rat hepatocytes does not operate in humans. – The absence of a significant response of human liver to induction of peroxisome proliferation and hepatocellular proliferation is explained by several aspects of PPARa-mediated regulation in gene expression. – Overall these findings indicate that the increased incidence of liver tumors in mice and rats treated with DEHP results from a mechanism that does not operate in humans.” This assessment of DEHP applies to all the other phthalate esters that are nongenotoxic but can induce peroxisome proliferation. Yet, there are marked differences in the potency of peroxisome proliferation. Most of the commercial phthalate esters were compared for their potential to induce peroxisome proliferation [33]. In general, phthalate esters are weak peroxisome proliferators compared with other peroxisome proliferators such as hypolidemic drugs. The lowMW phthalate esters do not show any peroxisome proliferator induction capacity or, at best, minimal capacity. The mid- and high-MW phthalates are peroxisome proliferators with the highest potency seen with DEHP, DINP, and DIDP [33, 58]. In accordance with this ranking, there is no evidence or equivocal evidence of carcinogenic potential for DEP, DMP, and BBP [59, 60], but evidence of carcinogenic potential in animals for DEHP and DINP.Again, cancer in animals through this mechanism is not considered relevant for humans.
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
311
3.2 Reproductive/Developmental Toxicity and Endocrine-Related Events
The male reproductive tract was identified as a target organ for some phthalate esters in the late 1970s. Investigators found that mid-MW phthalates could decrease sperm production in the testes of rats [61–64]. Studies of low- to highMW phthalates have not demonstrated these effects, so testicular damage in adults occurs primarily with mid-MW phthalates, including DEHP, rather than with other phthalates. Dose levels that cause effects in adult male rats are generally > 250 mg kg–1 d–1. The most prominent effect of DEHP exposure is loss of sperm production. This effect translates into reduced fertility for rodents (rats and mice) at dose levels of >250 mg kg–1 d–1. The exact mechanism for the testicular damage is unknown, but there are data to suggest that the effect is on the Sertoli cell rather than on the Leydig cells or directly on spermatogonia [65]. The relevance to humans of this effect in rodents has been debated. Two factors modify the assessment. One factor indicates that absorption of at least highMW phthalates is far less in monkeys (and probably humans) than in rodents [9]. This means that much less of the active metabolite can reach the target organ to cause the adverse effect. The other factor is the lack of testicular effects in monkeys treated with dose levels that cause testicular effects in rodents [11, 12]. Based on these two factors, the adult human male may not be at risk for the same antifertility effects seen in rodents. This same conclusion has been voiced by some expert groups [57, 66]. There are additional data for pre-adolescent (pre-pubertal) male rodents, which indicate that pre-pubescent males may have a greater susceptibility than adolescent or adult males [43]. Testicular damage in pre-pubertal males can occur at lower dose levels than testicular damage in adult males. The reason for this difference in age susceptibility is not known, but may be related to the formation of the blood-testis barrier during the puberty. This barrier provides some extra protection to the adult testes. Conversely, very young males that are affected (having exhibited aspermia, for example) also exhibit the ability to regenerate and recover [42] to the point that the testes of treated animals are indistinguishable from untreated when the animals are adults. Thus, pre-pubertal male rodents, while being more susceptible to the reproductive effects of some phthalate esters, do not exhibit permanent damage. Effects on the reproduction of female rodents have been reported based on the reduced fertility following breeding of treated females with untreated males. However, unlike studies in males, there are no cellular changes in the ovaries or reproductive organs of the female rodents that can be associated with these effects [26]. There are data to suggest that DEHP, a high-MW phthalate, given to female rats results in a disturbance of the estrous cycle (ovulation and fertility cycle), an effect that might cause reduced fertility in rodents [67]. Only very high dose levels are known to cause such effects, and it is not known if these disturbances can be related to humans who have a menstrual cycle rather than an estrous cycle. The mechanism suggested for these disturbances is alterations in the enzymes that form the sex hormones in the ovaries. Whether or not this
312
R.M. David and G. Gans
mechanism is applicable for humans is not known, nor is the dose level at which these effects do not occur in rodents (a NOEL). Toxicity to the developing embryo or neonate has also been investigated for most phthalate esters. In general, disturbances in the normal development of the rodent fetus have been observed for mid-MW phthalates and DEHP at dose levels of >100 mg kg–1, as stated previously. The nature of the effects observed vary with the timing of the exposure and the dose level. Exposure throughout gestation at dose levels above 500 mg kg–1 results in craniofacial abnormalities (cleft palate, cleft lip, exposed brain, or incomplete formation of the skull or spine) and skeletal abnormalities (incomplete bone formation) [68–73]. The mechanism(s) for these effects is(are) not known. Recent studies have also indicated that exposure of pregnant rodents during late gestation to dose levels of >300 mg kg–1 mid-MW phthalates and DEHP affect the developing male reproductive tract resulting in incomplete formation of secondary sex organs or testicular malformation [74–76]. No effects have been identified in the reproductive organs of female rodent offspring. The mechanism for these effects on male offspring is thought to be related to the hormone signals necessary for proper development of the urogenital tract in males. There is evidence to suggest that some phthalate esters interfere with the formation of testosterone, although the exact mechanism is not clear. How much this interference is attributable to changes in biochemistry, which is unique to rodents (as has been observed for the effects in the liver), has been debated. One investigator has demonstrated that the biochemical changes in the liver are not totally independent of effects in the testes [27]. It is of interest that the same phthalates that have effects on the male reproductive tract also affect development of the embryo. One possible explanation is that both the male reproductive tract and the developing embryo are rapidly dividing tissues that can be affected by exposure to some phthalate esters.Another explanation is that there is a common mechanism for both reproductive effects and developmental effects. At present, there is no information to suggest a common mechanism. Some have attributed the similarity in response to endocrinerelated events. Interference with endocrine-associated events in the developing reproductive tract and the pre-pubertal reproductive tract has been suggested. Associations with other developmental effects have not been established. However, the evidence, along with data from tissue culture studies, has suggested to some that these phthalate esters are endocrine disruptors. Regrettably, the term endocrine disruptor is not well defined. Phthalate esters are not estrogens (or estrogen mimics). Some diesters have been reported to bind to, or weakly activate, the estrogen receptor in tissue culture [77, 78]. This has been misinterpreted as indicating that phthalate esters are estrogens. Further study using intact animals has demonstrated that no estrogenic activity can be detected following oral exposure to doses of 20–2000 mg kg–1 d–1 phthalate esters [79, 80]. Neither are phthalate esters androgens [81, 82]. These substances do not either bind to the androgen receptor, or interfere with, the activation of the androgen receptor by testosterone.Yet, some of the phthalate esters cause effects that are similar to the effects caused by substances that interfere with testosterone-triggered events. For this reason, the effects have been termed “anti-androgen-like,” albeit through an unknown mechanism.
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
313
The question: “Are phthalate esters endocrine disruptors?” is continually raised. Because these substances do not act on the hormone receptors (either to activate or deactivate them), it is a difficult question to answer. In the sense that the normal function of the endocrine system is somehow altered to produce the effect observed because there is less testosterone produced, then these substances may be endocrine disruptors. How that function is distinguished from other metabolic changes that occur in rats and mice, and whether such changes can occur in humans, remains to be answered. 3.3 Asthma
Recently, a question regarding a possible role of phthalate esters in the pathogenesis of asthma has been raised [83, 84]. The studies in question found phthalate esters from vinyl products in the air of households with asthmatic children. They speculated on a possible mechanism for DEHP to cause asthma. The hypothesis is that DEHP and its metabolites have some structural similarities to, and can mimic, some prostaglandins and thromboxanes, and the widespread use of DEHP in wall coverings, flooring, and other construction material leads to the induction or exacerbation of asthma. The merit of this hypothesis is questionable. There is also no direct evidence that DEHP or MEHP acts like prostaglandin, the hormone associated with inflammatory responses. Furthermore, there is no indication of any sensitization potential from exposure to DEHP, based on animal experiments or human patch tests [25]. Phthalates are not the large, complex molecules which are more commonly associated with allergy, and are not among the substances such as insect parts, animal dander, and pollen, which authorities have associated with the induction of asthma.Vinyl flooring is easy to clean; it is often recommended for use in homes where asthmatics live. If an association between phthalate esters used in vinyl exists, it is not clear whether the use of vinyl flooring is a cause of the asthma or a remedy to reduce dust and animal dander that cause asthma. Further work is necessary to evaluate this phenomenon. 3.3 Neurotoxicity
Only one phthalate ester has been tested in a stand-alone neurotoxicity studies. One study under EPA guidelines was conducted and the results indicated no evidence of neurotoxicity [85]. Other studies using only functional tests of neurobehavior support this finding even in laboratory animals exposed during development [86]. These stand in contrast to a single report of altered neurobehavioral activity in a single, invalidated beam-walking test following in utero exposure [87]. The significance of this finding is questionable. As for the other phthalate esters, none of the recent 13-week and/or chronic toxicity studies of some low-, mid-, and high-MW phthalates have indicated abnormalities in clinical observations or behavioral assessments. While the focus of these studies was not neurotoxicity, they incorporated detailed daily observa-
314
R.M. David and G. Gans
tion of the animals that should have detected neurotoxicity. Thus, phthalate esters do not appear to be neurotoxic even following short-term or prolonged exposure.
4 References 1. IARC (1995) Peroxisome proliferation and its role in carcinogenesis. IARC technical report no 24. IARC Press, Lyon 2. IPCS (1992) Environmental health criteria document no 131, WHO 3. Cattley RC, DeLuca J, Elcombe C, Fenner-Crisp P, Lake BG, Marsman DS, Pastoor TA, Popp JA, Robinson DE, Schwetz B, Tugwood J, Wahli W (1998) Regul Toxicol Pharmacol 27:47 4. Moody DE, Reddy JK (1978) Toxicol Appl Pharmacol 45:497 5. Benford DJ, Patel S, Reavy HJ, Mitchell A, Sarginson NJ (1986) Food Chem Toxicol 24: 799 6. Elcombe CR, Mitchell AM (1986) Environ Health Perspect 70:211 7. Bichet N, Cahard D, Fabre G, Remandet B, Gouy D, Cano JP (1990) Toxicol Appl Pharmacol 106:509 8. Elcombe CR, Bell DR, Elias E, Hasmall SC, Plant NJ (1996) Ann NY Acad Sci 804:628 9. Rhodes C, Orton TC, Pratt IS, Batten PL, Bratt H, Jackson SJ, Elcombe CR (1986) Environ Health Perspect 65:299 10. Short RD, Robinson EC, Lington AW, Chin AE (1987) Toxicol Ind Health 3:185 11. Kurata Y, Kidachi F,Yokoyama M, Toyota N, Tsuchitani M, Katoh M (1998) Toxicol Sci 42:49 12. Pugh G Jr, Isenberg JS, Kamendulis LM, Ackley DC, Clare LJ, Brown R, Lington AW, Smith JH, Klaunig JE (2000) Toxicol Sci 56:181 13. Astill BD (1989) Drug Metab Rev 21:35 14. Rowland LR, Cottrell RC, Phillips JC (1974) Food Cosmet Toxicol 12:293 15. Rowland LR, Cottrell RC, Phillips JC (1977) Food Cosmet Toxicol 15:17 16. White RD, Carter DE, Earnest D, Mueller J (1980) Food Cosmet Toxicol 18:383 17. Albro PW, Moore B (1974) J Chromatogr 94:209 18. Deisinger PJ, Perry LG, Guest D (1998) Fd Chem Toxic 36:521 19. Elsisi AE, Carter DE, Sipes IG (1989) Fundam Appl Toxicol 12:70 20. Scott RC, Dugard PH, Ramsey JD, Rhodes C (1987) Environ Health Perspect 74:223 21. Bogdanffy MS, Randall HW, Morgan KT (1986) Toxicol Appl Pharmacol 82:560 22. Carter JE, Roll DB, Petersen RV (1974) Drug Metab Dispos 2:341 23. Blount BD, Silva MJ, Caudill SP, Neeham LL, Pirkle JL, Sampson EJ, Lucier GW, Jackson RJ, Brock JW (2000) Environ Health Perspect 108:979 24. Meier L (1990) Plasticizers. In: Gächter R, Müller H (eds) Plastics additives, 3rd edn. Hanser Publishers 25. Medeiros AM, Devlin DJ, Keller LH (1999) Cont Dermat 41:287 26. Lamb JC IV, Chapin RE, Teague J, Lawton AD, Reel JR (1987) Toxicol Appl Pharmacol 88:255 27. Ward JM, Peters JM, Perella CM, Gonzales FJ (1998) Toxicol Pathol 26:240 28. Cimini AM, Sulli A, Stefanini S, Serafini B, Moreno S, Rossi L, Giorgi M, Ceru MP (1994) Cell Mol Biol 40:1063 29. Ohno S, Fujii Y, Usuda N, Murata F, Nagata T (1982) Acta Histochem Cytochem 15:40 30. Bell FP, Gillies PJ (1977) Lipids 12:581 31. Bell FP (1982) Environ Health Perspect 45:41 32. Gray TJB, Lake BG, Beamand JA, Foster JR, Gangolli SD (1983) Toxicology 28:167 33. Barber ED, Astill BD, Moran EJ, Schneider BF, Gray TJB, Lake BG, Evans JG (1987) Toxicol Ind Health 3:7 34. DeAngelo AB, Garrett CT, Manolukas LA, Yario T (1986) Toxicology 41:279 35. Butterworth BE, Smith-Oliver T, Earle L, Loury DJ, White RD, Doolittle DJ, Working PK, Cattley RC, Jirtle R, Michalopoulos G, Strom S (1989) Cancer Res 49:1075
Summary of Mammalian Toxicology and Health Effects of Phthalate Esters
315
36. Deangelo AB, Daniel FB, Mcmillan L, Wernsing P, Savage RE Jr (1989) Toxicol Appl Pharmacol 101:285 37. Dirven HAAM, van den Broek PHH, Peeters MCE, Peters JGP, Mennes WC, Blaauboer BJ, Noordhoek J, Jongeneelen FJ (1993) Biochem Pharmacol 45:2425 38. Hall M, Matthews A, Webley L, Harling R (1999) Toxicol Sci 24:237 39. Ashby J, Brady A, Elcombe CR, Elliott BM, Ishmael, J, Odum J, Tugwood, JD, Kettle S, Purchase LF (1994) Hum Exp Toxicol 13, Suppl 2:S1 40. Bentley P, Calder I, Elcombe C, Grasso P, Stringer D, Wiegand HJ (1993) Food Chem Toxicol 31:857 41. Lake BG (1995) Ann Rev Pharmacol Toxicol 35:483 42. Dostal LA, Chapin RE, Stefanski SA, Harris MW, Schwetz BA (1988) Toxicol Appl Pharmacol 95:104 43. Sjöberg P, Lindquist NG, Plöen L (1986) Environ Health Perspect 65:237 44. David RM, Moore MR, Finney DC, Guest D (2001) Toxicol Pathol 29: 45. David RM, Moore MR, Cifone MA, Finney DC, Guest D (1999) Toxicol Sci 50:195 46. Ward JM, Diwan BA, Ohshima M, Hu H, Schuller HM, Rice JM (1986) Environ Health Perspect 65:279 47. Crocker J F, Safe SH, Acott P (1988) J Toxicol Environ Health 23:433 48. David RM, Moore MR, Finney DC, Guest D (2000) Toxicol Sci 55:433 49. David RM, Moore MR, Finney DC, Guest D (2000) Toxicol Sci 58:377 50. Woodward KN (1990) Hum Exp Toxicol 9:397 51. Ward JM, Hagiwara A,Anderson LM, Lindsey K, Diwan BA Toxicol Appl Pharmacol 96:494 52. Marsman DS, Cattley RC, Conway JG, Popp JA (1988) Cancer Res 48:6739 53. Kluwe WM, McConnell EE, Huff JE, Haseman JK, Douglasw JF, Hartwell WV (1982) Environ Health Perspect 45:129 54. Butala JH, Moore MR, Cifone MA, Bankston JR, Astill B (1996) Toxicologist 30:202 55. Butala JH, Moore MR, Cifone MA, Bankston JR, Astill B (1996) Toxicologist 36:173 56. Peters JM, Cattley RC, Glonzalez FJ (1997) Carcinogenesis 18:2029 57. IARC Monographs on the evaluation of carcinogeic risks to humans (2000) Some industrial chemicals, vol 77. IARC Press, Lyon 58. Huber WW, Grasl-Kraupp B, Schulte-Hermann R (1996) Crit Rev Toxicol 26:365 59. National Toxicology Program (NTP) NTP TR429, 1995 60. National Toxicology Program (NTP) NTP TR458, 1997 61. Agarwal DK, Maronpot RR, Lamb JC IV, Kluwe WM (1985) Toxicology 35:189 62. Agarwal DK, Eustis S, Lamb JC IV, Reel JR, Kluwe WM (1986) Environ Health Perspect 65:343 63. Gray TJB, Butterworth KR (1980) Arch Toxicol Suppl 4:452 64. Gray TJB, Beamand JA (1984) Food Chem Toxicol 22:123 65. Lloyd SC, Foster PMD (1988) Toxicol Appl Pharmacol 95:484 66. Center for the Evaluation of Risks to Human Reproduction (CERHR) NTP-CERHR expert panel report on di(2-ethylhexyl) phthalate. NTP-CERHR-DEHP-00, 2000 67. Davis BJ, Maronpot RR, Heindel JJ (1994) Toxicol Appl Pharmacol 128:216 68. Ema M, Harazono A, Miyawaki E, Ogawa Y (1997) Bull Environ Contam Toxicol 58:636 69. Ema M, Amano H, Ogawa Y (1994) Toxicology 86:1630 70. Ema M, Miyawaki E, Kawashima K (1998) Toxicology Lett 98:87 71. Ema M, Miyawaki E, Kawashima K (2000) Toxicology Lett 111:271 72. Ema M, Kurosaka R, Harazono A, Amano H, Ogawa Y (1996) Arch Environ Contam Toxicol 31:170 73. Ema M, Kurosaka R, Amano H, Ogawa Y (1994) Reprod Toxicol 8:231 74. Wine RN, Li LH, Barnes LH, Gulati DK, Chapin RE (1997) Environ Health Perspect 105:102 75. Mylchreest E, Cattley RC, Foster PMD (1998) Toxicol Sci 43:47 76. Gray LE, Ostby J, Furr J, Price M, Rao Veeramachaneni DN, Parks L (2000) Toxicol Sci 58:350 77. Jobling S, Reynolds T, White R, Parker MG, Sumpter JP (1995) Environ Health Perspect 103:582
316
R.M. David and G. Gans: Summary of Mammalian Toxicology and Health Effects
78. Coldham NG, Dave M, Sivapathasundaram S, McDonnell DP, Connor C, Sauer MJ (1997) Environ Health Perspect 105:734 79. Milligan SR, Balasubramanian AV, Kalita JC (1998) Environ Health Perspect 106:1 80. Zacharewski TR, Meek MD, Clemons JH, Wu ZF, Fielden MR, Matthews JB (1998) Toxicol Sci 46:282 81. Gray LE Jr, Wolf C, Lambright C, Mann P, Price M, Cooper RL, Ostby J (1999) Toxicol Ind Health 15:94 82. Sohoni P, Sumpter JP (1998) J Endocrinol 158:327 83. Oie L, Hersoug L-G, Madsen JO (1997) Environ Health Perspect 105:972 84. Jaakola JJ, Oie L, Nafstad P, Botten G, Samuelsen SO, Magnus P (1999) Am J Pub Health 89:188 85. Moser VC, Cheek BM, MacPhail RC (1995) J Toxicol Environ Health 45:173 86. Merkle J, Klimisch HJ, Jaeckh R (1988) Toxicol Lett 42:215 87. Arcadi FA, Costa C, Imperatore C, Marchese A, Rapidisarda A, Salemi M, Trimarchi GR, Costa G (1998) Food Chem Toxicol 36:963
The Handbook of Environmental Chemistry Vol. 3, Part Q (2003): 317– 349 DOI 10.1007/b11471
An Assessment of the Potential Environmental Risks Posed by Phthalates in Soil and Sediment Thomas F. Parkerton 1 · Charles A. Staples 2 1 2
ExxonMobil Biomedical Sciences Inc., Hermeslaan 2, 1831 Machelen, Belgium. E-mail: [email protected] Assessment Technologies, Inc., 10201 Lee Highway, Suite 580, Fairfax, VA 22030, USA. E-mail: [email protected]
To assess the potential environmental concerns associated with phthalate esters (PEs) in sediments and native- as well as sludge-amended soils a screening risk assessment was performed using the risk quotient paradigm. Five single isomers, dimethyl, diethyl, di-n-butyl, butylbenzyl and di-2-ethylhexyl, and two commercial mixed isomers, di-isononyl and di-isodecyl, were specifically investigated. Application of statistical extrapolation techniques to aquatic effects data coupled with Equilibrium Partitioning (EqP) theory were used to derive Predicted No Effect Concentrations (PNECs) intended to protect terrestrial and benthic organisms from direct toxicity posed by PEs in soil or sediment. The resultant PNECs were found to be protective when compared to the wealth of available soil and sediment toxicity data for these compounds. PNECs intended to protect wildlife consumers from indirect effects associated with exposure via the terrestrial/benthic food chain were also calculated for each PE. Comparison of riskbased criteria revealed that direct toxicity to soil or sediment-dwelling organisms dictates PNEC derivation for low molecular weight PEs while potential indirect effects on wildlife consumers via food chain exposure determine PNECs for higher molecular weight PEs. A comprehensive literature review indicated extensive field monitoring data are available characterizing PE concentrations in sediments from Europe, North America and Japan. While less exposure data were available for characterizing the soil compartment, predicted and observed concentrations were lower than in sediments. Results of the screening risk assessment found that for all PEs investigated, none of the observed soil concentrations exceeded risk-based limits even in the case of soils that were heavily amended with sewage sludge. Similarly, no study reported concentrations in field sediments that exceeded the PNEC for either BBP or DINP. For the remaining PEs, at least one study indicated a maximum sediment concentration above the PNEC. However, the number of sediment samples exceeding the PNEC was typically less than 1% of the available monitoring database. It is concluded that the environmental concerns posed by soil and sediment-associated PEs are at worst, restricted to infrequent, localized hot spots of contaminated sediment. The conservative assumptions invoked in this screening risk analysis and implications of this work in future regulatory decision-making are also discussed.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
1
Introduction
2
Hazard Characterization . . . . . . . . . . . . . . . . . . . . . . . 319
2.1 2.1.1 2.1.2 2.1.3
Direct Effects on Soil and Sediment-Dwelling Organisms . . Soil and Sediment Toxicity Tests . . . . . . . . . . . . . . . . Extrapolation from Aquatic Toxicity Data Using EqP Theory Association-Based Methods Based on Field Data . . . . . . .
. . . .
. . . .
. . . .
319 319 327 328
© Springer-Verlag Berlin Heidelberg 2003
318
T.F. Parkerton and C.A. Staples
2.1.4 2.2
PNEC Selection for Direct Effects . . . . . . . . . . . . . . . . . . 329 Indirect Effects on Wildlife via the Food Chain . . . . . . . . . . . 329
3
Exposure Characterization . . . . . . . . . . . . . . . . . . . . . . 334
3.1 3.1.1 3.1.2 3.1.3 3.2 3.2.1
Soil . . . . . . . . . . . . . . . . . . Native Soil . . . . . . . . . . . . . . . Sludge-Amended Soil . . . . . . . . . Field Monitoring Data for Soil . . . . Sediment . . . . . . . . . . . . . . . Field Monitoring Data for Sediments
4
Risk Assessment
5
Summary and Discussion
6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
334 334 334 336 336 337
. . . . . . . . . . . . . . . . . . . . . . . . . . . 339 . . . . . . . . . . . . . . . . . . . . . . 342
1 Introduction During the 1950s the potential commercial benefit of phthalate esters (PEs) became increasingly recognized. Due to excellent performance as cost-effective plasticizers in a broad range of applications, demand for PEs burgeoned. As PE use continued increasing concern was raised regarding the possible risks that made-made chemicals could pose to the environment. As a result, this class of important industrial chemicals has repeatedly been the focus of environmental research for several decades. The aquatic toxicity database for phthalate esters is extensive [1]. These data have recently been used to develop a number of species and endpoint-specific quantitative structure activity relationships (QSARs) that describe PE aquatic toxicity. Application of statistical extrapolation procedures to these data has enabled risk-based surface water concentrations (i.e. Predicted No Effect Concentrations or PNECs) to be developed for four commercially important PEs: dimethyl (DMP), diethyl (DEP), di-n-butyl (DBP) and butylbenzyl (BBP) phthalate [2]. To assess the potential risks that these substances pose to the aquatic environment, Staples et al. [3] prepared a comprehensive compilation of historical exposure monitoring data. Comparison of observed or predicted surface water concentrations to PNECs indicated environmental concentrations that were typically several orders of magnitude below risk-based environmental quality objectives. For higher molecular weight PEs such as di-2-ethylhexyl phthalate (DEHP), no acute or chronic toxicity is evident at the water solubility limit. This lack of aquatic toxicity hazard may be explained by the combined role of low water solubility and limited bioconcentration potential due to biotransformation. These two factors prevent the accumulation of tissue residues above a critical threshold. Thus, aqueous exposure is not expected to result in an internal critical body residue that elicits adverse effects. Consequently, surface water concentrations of these substances are not expected to pose a direct concern to aquatic life [2].
An Assessment of the Potential Environmental Risks Posed by Phthalates
319
The above studies indicate PEs are unlikely to pose harm to aquatic biota. However, due to the range of physico-chemical properties exhibited by PEs, soil and sediments may also serve as a significant, if not predominant compartment dictating environmental fate behavior [4]. Moreover, ingestion of contaminated soil or sediment by terrestrial or benthic organisms, respectively, may serve as an additional route of exposure relative to that provided by pore water, especially for poorly water soluble substances such as high molecular weight phthalates. Since degradation rates in soils and sediments typically are slower than in surface water [5], these compartments also have the potential to serve as long-term sources of indirect exposure via trophic transfer (e.g. via the food chain). Therefore, the potential environmental risks posed by phthalates in soil and sediment logically warrant further investigation. The objective of the present study is to provide an assessment of the direct risks posed by soil and sediment-associated phthalates on benthic and terrestrial organisms as well as the indirect risks (via the food chain) to wildlife. In addition to the single isomer PEs reported in the surface water risk assessment by Staples et al. [3], two additional mixed isomers, diisononyl (DINP) and diisodecyl (DIDP) phthalate are considered in the present study. These substances are included due to commercial significance and the expected importance that soil and sediment compartments play in the environmental fate of these poorly water soluble PEs. The remainder of this paper is organized into the following sections. First, the direct and indirect hazard of soil and sediment-associated phthalates is critically reviewed. Based on this analysis, risk-based soil and sediment quality objectives are derived. Environmental monitoring data obtained from field studies are compiled to characterize soil and sediment concentrations of the selected PEs in different regions of the world. This information is then used as the basis for risk characterization.A discussion of the assumptions and uncertainties in this analysis are also presented in the concluding section.
2 Hazard Characterization 2.1 Direct Effects on Soil and Sediment-Dwelling Organisms 2.1.1 Soil and Sediment Toxicity Tests
A compilation of available soil and sediment toxicity test data is provided in Table 1. Results are tabulated for broad taxonomic groups representing microbe, plant, invertebrate, vertebrate and multi-species (mesocosm) tests by endpoint type (i.e. L/EC50 , LOEC, NOEC). Test duration, and when available, soil organic carbon content (or soil type) is also provided. Available acute or short-term chronic data for DMP and DEP, while limited, are in the range of 100 to >1000 mg/kg dry. In the case of DBP, considerable toxicity data are available across trophic levels. Acute toxicity is observed at concentrations similiar to DMP and DEP. Several chronic NOECs for DBP based on growth
Diethyl phthalate (DEP) Microorganisms Soil microbes Plants Lactuca sativa (lettuce) Lactuca sativa (lettuce) Spinacea oleracea (spinach) Pisium sativum (peas) Sediment invertebrates Chironomus tentans (midge) Dibutyl phthalate (DBP) Plants Zea mays (corn) Zea mays (corn) Lactuca sativa (lettuce) Lactuca sativa (lettuce)
Dimethyl phthalate (DMP) Microorganisms Soil microbes Plants Spinacea oleracea (spinach) Pisium sativum (peas) Soil invertebrates Allolobophora tuberculata Eisenia foetida Eudrilus eugeniae Perionyx excavatus
Test species (common name)
shoot weight shoot weight seedling height seedling height survival, growth
seed germination height, shoot length shoot weight shoot weight
7 14 16 14
10
21 21 7 14
survival survival survival survival
14 14 14 14
bacteria number
seedling height seedling height
16 14
1
bacteria number
Test endpoint
1
Test duration in days except where stated
Table 1. Summary of soil/sediment toxicity test data for phthalate esters
Sand Sand 1.4 1.4
2.45
1.4 1.4 NR NR
3.8
OECD soil? OECD soil? OECD soil? OECD soil?
NR NR
3.8
Soil/sediment organic carbon (% dry)
387 >1000
>3100
106 134 >1000 >1000
1000*
3335 3160 2000 1064
<1000* ca. 1000*
1000*
EC or LC50 (mg/kg dry)
>20,000* 2000*/200* NR NR
3100/843
NR NR NR NR
1000*/100*
NR NR NR NR
NR NR
NR
LOEC/NOEC (mg/kg dry)
[11] [11] [9] [9]
[10]
[9] [9] [7] [7]
[6]
[8] [8] [8] [8]
[7] [7]
[6]
Ref.
320 T.F. Parkerton and C.A. Staples
adult survival adult survival adult reproduction adult reproduction juvenile survival juvenile growth juvenile development survival, growth survival, growth survival, growth survival, growth survival, growth survival, growth community structure
21 21 21 21 42 42 42
10 10 10 10 10 10
8 wks
14
8h
Benzyl butyl phthalate (BBP) Soil invertebrates Eisenia foetida
Di-2 ethylhexyl phthalate (DEHP) Microorganisms Soil microbes
respiration inhibition
survival and growth
seedling height seedling height
Test endpoint
16 14
Test duration in days except where stated
Dibutyl phthalate (DBP) Spinacea oleracea (spinach) Pisium sativum (peas) Soil invertebrates Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Folsomia fimetaria (springtails) Sediment invertebrates Chironomus tentans (midge) Chironomus tentans (midge) Chironomus tentans (midge) Hyalella azteca (amphipod) Hyalella azteca (amphipod) Hyalella azteca (amphipod) Multi-species Sediment Mesocosm
Test species (common name)
Table 1 (continued)
NR
Artificial?
NR
2.45 4.8 14.1 2.45 4.8 14.1
<1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5
NR NR
Soil/sediment organic carbon (% dry)
826 1664 4730
>10
305 277 68 84 19,4
EC or LC50 (mg/kg dry)
49,000/NR
>1000*
1000/100?
315/50 3090/423 3550/508 >17,400 >29,500 >71,900
33** 34** 14** 50** 11.3** >10 1.0/0.5**
>1000 >1000
LOEC/NOEC (mg/kg dry)
[15]
[14]
[13]
[10] [10] [10] [10] [10] [10]
[12] [12] [12] [12] [12] [12] [12]
[7] [7]
Ref.
An Assessment of the Potential Environmental Risks Posed by Phthalates
321
Plants Festuca arundinacea (tall fescue) Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lactuca sativa (lettuce) Danucus carota L. (carrot) Capsicum annum L. (chili pepper) Triticum aestivum (wheat) Lepidium sativum (cress) Brassica napas (mustard) Brassica rapa (turnip) Avena sativa (oats) Spinacea oleracea (spinach) Pisium sativum (peas) Soil invertebrates
Di-2 ethylhexyl phthalate (DEHP) Soil microbes Soil microbes Soil microbes Soil microbes Soil microbes Soil microbes Soil microbes Soil microbes Soil microbes Sediment microbes Sediment microbes
Test species (common name)
Table 1 (continued)
life cycle life cycle 7 14 life cycle life cycle 14 14 14 14 14 16 14
1–16 94 28 28 60 14, 28 14, 28 7, 28 7, 28 NR 2.5
Test duration in days except where stated
growth growth shoot weight shoot weight growth growth germination, shoot weight germination, shoot weight germination, shoot weight shoot weight shoot weight seedling height seedling height
structural and functional diversity respiration inhibition respiration inhibition respiration inhibition nitrogen mineralization inhibition nitrogen mineralization inhibition nitrogen mineralization inhibition dehydrogenase inhibition dehydrogenase inhibition respiration inhibition? respiration inhibition
Test endpoint
1 1 1.4 1.4 1 1 OECD OECD OECD NR NR NR NR
3.8 1.8 2.3 5.9 1.8 1.8 5.9 1.8 5.9 NR 9.2#
Soil/sediment organic carbon (% dry)
>1000* >1000*
>1000 >1000
EC or LC50 (mg/kg dry)
>14 >14 NR NR >14 >14 >100* >100* >100* >1000* 10/100–1000*x >1000 >1000
>100,000* >250 >573 >829 >250 >731 >686 >573 >829 >100 84+
LOEC/NOEC (mg/kg dry)
[22] [22] [9] [9] [22] [22] [23, 24] [23, 24] [23, 24] [25] [25] [7] [7]
[6] [16] [17] [17] [16] [18] [18] [19] [19] [20] [21]
Ref.
322 T.F. Parkerton and C.A. Staples
predation efficieny emergence, sex ratio survival, growth survival, growth egg hatching (5 C) tadople survival (5 C) egg hatching (10 C) egg hatching (10 C) egg hatching (10 C) egg hatching (10 C) tadople survival and growth (10 C) tadople survival and growth (10 C) tadople survival and growth (10 C) tadople survival and growth (10 C) egg hatching, hatching time (5 C) tadpole survival, growth and development (5 C) egg hatching, hatching time (5 C) tadpole survival, growth and development (5 C)
30 60 14 14 14 14 29 29 29 29 22–25 35
22–25 35
Rana arvalis (moor frog) Rana arvalis (moor frog)
survival adult survival and reproduction juvenile survival, growth and development
Test endpoint
40 28 10 10
14 21 42
Test duration in days except where stated
Sediment invertebrates Aeshna (dragonfly larvae) Chironomus tentans (midge) Chironomus tentans (midge) Hyalella azteca (amphipod) Sediment vertebrates Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog) Rana arvalis (moor frog)
Di-2 ethylhexyl phthalate (DEHP) Eisenia foetida Folsomia fimetaria (springtails) Folsomia fimetaria (springtails)
Test species (common name)
Table 1 (continued)
17.3 17.3
8.4–13.2# 8.4–13.2# 1.2# 9.0# 16.8# 30.6# 1.2# 9.0# 16.8# 30.6# 16 16
16 3.6 4.8 4.8
OECD <1.5 <1.5
Soil/sediment organic carbon (% dry)
ca. 450+
NR
EC or LC50 (mg/kg dry)
>1028 >1028
NR >2600 >205*** >433*** >699*** >255*** >205*** >433*** >699*** >255*** >999 >999
1468 >10,000 >3070 >3170
>1000* >5000 >1000
LOEC/NOEC (mg/kg dry)
[29] [29]
[21] [21] [28] [28] [28] [28] [28] [28] [28] [28] [29] [29]
[26] [27] [10] [10]
[23] [12] [12]
Ref.
An Assessment of the Potential Environmental Risks Posed by Phthalates
323
seed germination seed germination seed germination seed germination seed germination, growth seed germination seed germination survival survival survival, growth survival, growth
5 5 5 5 28 5 5
14 14
10 10
community structure
inhibition of glucose utilization
30
Multi-species Sediment mesocosm
egg hatching, hatching time (10 C) tadpole survival, growth and development (10 C) egg hatching, hatching time (10 C) tadpole survival, growth and development (10 C)
Test endpoint
33
9–21 26
Rana arvalis (moor frog) Rana arvalis (moor frog)
Di-isononyl phthalate (DINP) Microorganisms Soil microbes Plants Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lolium sp. (rye grass) Lolium sp. (rye grass) Soil invertebrates Eisenia foetida Eisenia foetida Sediment invertebrates Chironomus tentans (midge) Hyalella azteca (amphipod) Sediment vertebrates
9–21 26
Test duration in days except where stated
Di-2 ethylhexyl phthalate (DEHP) Rana arvalis (moor frog) Rana arvalis (moor frog)
Test species (common name)
Table 1 (continued)
4.8 4.8
4.0# 1.7
4.0# 1.7 4.0# 1.7 1.7 4.0# 1.7
1.7
NR
17.3 17.3
16 16
Soil/sediment organic carbon (% dry)
EC or LC50 (mg/kg dry)
>2680 >2900
>9000 >7900
<10,000 <10,000 3000/1000 3000/1000 >1387 >10,000 >10,000
>9616
>6.2
>1164 >1164
>844 >844
LOEC/NOEC (mg/kg dry)
[10] [10]
[35] [35]
[32] [32] [33] [33] [34] [32] [32]
[31]
[30]
[29] [29]
[29] [29]
Ref.
324 T.F. Parkerton and C.A. Staples
survival survival emergence, sex ratio survival, growth survival, growth egg hatching (10 C) tadople survival and growth (10 C)
14 14
28 10 10
14 29
9.0# 9.0#
3.6 4.8 4.8
4.0# 1.7
4.0# 1.7 4.0# 1.7
17.3 17.3
16 16
Soil/sediment organic carbon (% dry)
EC or LC50 (mg/kg dry)
>657*** >657***
>10,000 >2630 >2090
>9000 >7900
>10,000 >10,000 >10,000 >10,000
>1009 >1009
>707 >707
LOEC/NOEC (mg/kg dry)
[28] [28]
[27] [10] [10]
[35] [35]
[32] [32] [32] [32]
[29] [29]
[29] [ 29]
Ref.
Note: Exposure concentrations are reported as mean values based on analytical measurements reported over the exposure period unless otherwise indicated. NR=Not reported. * Based on nominal exposure concentrations reported. ** NOEC indicated corresponds to reported EC10 . *** Measured exposure concentration in sediment at end of test. # Organic carbon content estimated by multiplying % loss on ignition by 0.4. + Reported fresh weight concentrations corrected to dry weight by assuming a 0.4 dry to wet weight ratio. x – A non-dose dependent reduction in growth was observed in both 100 and 1000 mg/kg treatments.
seed germination seed germination seed germination seed germination
egg hatching, hatching time (10 C) tadpole survival, growth and development (10 C) egg hatching, hatching time (10 C) tadpole survival, growth and development (10 C)
Test endpoint
5 5 5 5
9–21 26
Rana arvalis (moor frog) Rana arvalis (moor frog)
Di-isodecyl phthalate (DIDP) Plants Lactuca sativa (lettuce) Lactuca sativa (lettuce) Lolium sp. (rye grass) Lolium sp. (rye grass) Soil invertebrates Eisenia foetida Eisenia foetida Sediment invertebrates Chironomus riparius (midge) Chironomus tentans (midge) Hyalella azteca (amphipod) Sediment vertebrates Rana arvalis (moor frog) Rana arvalis (moor frog)
9–21 26
Test duration in days except where stated
Di-isononyl phthalate (DINP) Rana arvalis (moor frog) Rana arvalis (moor frog)
Test species (common name)
Table 1 (continued) An Assessment of the Potential Environmental Risks Posed by Phthalates
325
326
T.F. Parkerton and C.A. Staples
or reproduction endpoints are in the range of 10–100 mg/kg dry. A statistically significant NOEC value below 1 mg/kg dry is reported for development of juvenile springtails (i.e. number of cuticles). However, the authors question the ecological significance of this reported effect. Moreover, interpretation of this endpoint was further complicated by the high variation in molting frequency observed in control animals. For BBP, although limited toxicity data are available, no acute toxicity was reported in earthworms exposed to a soil concentration of 1000 mg/kg dry. Considerable soil and sediment toxicity data are available for high molecular weight PEs.With a few exceptions discussed below, no acute or chronic effects are reported at the highest concentrations investigated, typically >100 mg/kg dry. Early studies by Swedish investigators reported that DEHP caused adverse effects on microbial respiration and hatching of moor frog eggs at sediment concentrations below 100 mg/kg dry [21, 36]. However, a critical review of these studies reveals a number of technical problems. In these tests DEHP was spiked to wet sediment by first dissolving the test substance into ethanol. The introduction of ethanol to wet sediment is known to significantly alter the nature of sediment organic carbon as evidenced by a marked increase in the concentration of dissolved organic carbon in the pore water (David Mount, USEPA, personal communication). Thus, ethanol functions as a solvent to extract particulate organic carbon from sediment particles. This perturbation of the test sediment can significantly alter normal partitioning behavior and confound toxicity test interpretation. Furthermore, since no analytical measurements were provided at the start of toxicity tests it is possible that considerable heterogeneity in sediment concentrations resulted as a result of this spiking procedure, again complicating test interpretation. Given the non-standardized nature of these tests with this species and lack of experience with normal control variation in the toxicity test endpoints examined the reliability of these tests is uncertain. To address these concerns, subsequent toxicity studies with microbes [6, 17] and moor frogs [28, 29] have been reported. In these follow-up studies in which the use of ethanol as a carrier solvent was typically avoided, no effects were observed. Moreover, further experiments using ethanol as a carrier solvent did not replicate any of the findings reported in the original studies [28]. Consequently, the early studies by Thuren and coworkers cannot be regarded as reliable for risk assessment purposes nor serve as an appropriate technical basis for derivation of environmental risk limits as recently proposed [37]. An analogous situation is represented by the soil toxicity study conducted by Stanley and Tapp [25] since anomalous test results were reported relative to numerous other test data available (Table 1). These authors spiked 1, 10, 100 and 1000 mg/kg of DEHP to quartz sand and then examined shoot growth of pregerminated seeds of turnips (Brassica rapa) and oats (Avena sativa) after 14 days relative to an untreated control group. No test substance related effects were reported for turnips up to 1000 mg/kg dry but statistical analysis of the raw shoot weight data indicated that the growth of oats was significantly reduced at both the 100 and 1000 mg/kg dry DEHP treatments. However, no concentrationdependent response was evident since both concentrations elicited the same degree of growth reduction (ca. 30%) questioning the interpretation and relia-
An Assessment of the Potential Environmental Risks Posed by Phthalates
327
bility of these findings. If one excludes as unreliable the studies mentioned above, none of the numerous soil and sediment toxicity tests available for DEHP demonstrated an adverse effect at the highest concentration tested (Table 1). As in the case of DEHP, numerous soil and sediment toxicity studies show no adverse effects for DINP and DIDP at the highest concentrations tested. However, one exception has been reported for DINP in studies with lettuce. Lettuce seed germination after 5 days was significantly reduced in a concentration-dependent manner in two soils resulting in a NOEC and LOEC of 1000 and 3000 mg/kg, respectively. A follow-up 28-day chronic toxicity study with lettuce seeds failed to reveal any growth effects at the highest DINP concentration tested (i.e. 1387 mg/kg dry). The above review of the available ecotoxicological data suggests that high molecular weight phthalates may cause adverse effects on plants at extreme exposure concentrations (e.g. >1000 mg/kg dry). Curiously, such effects are however not reported for DIDP (Table 1). If the effects observed are genuinely test substance related, it is hypothesized that such effects are likely due to a physical explanation (e.g. hydrophobic effect on soil influencing water uptake by seeds) rather than a systemic toxicity mechanism. Such physical effects have been reported previously for soils contaminated with petroleum hydrocarbons [38]. 2.1.2 Extrapolation from Aquatic Toxicity Data Using EqP Theory
The extensive aquatic toxicity database that is available for PEs can be extrapolated to predict the hazard to soil and sediment-dwelling organisms using the Equilibrium Partitioning (EqP) model: PNEC(direct) =Koc PNECaquatic
(1)
Where: PNEC(direct) predicted no effect concentration in soil/sediment (mg/kg oc) organic carbon-normalized partition coefficient (l/kg oc) Koc PNECaquatic predicted no effect concentration in surface water (mg/l) The PNECsoil/sediment can be expressed on a dry weight basis by simply multiplying by the organic carbon fraction of the soil or sediment. The technical basis supporting this approach for deriving sediment or soil quality criteria has previously been described [39–42]. This approach is currently used in a variety of regulatory programs in both North America and Europe [43, 44]. The Koc in Eq. (1) can be estimated from the octanol-water partition coefficient (Kow) using the correlation reported by Seth et al. [45]: Koc =0.35 Kow
(2)
For the low molecular weight phthalates DMP, DEP, DBP and BBP, a statistical extrapolation procedure has been recently applied to available aquatic toxicity data to derive PNECaquatic [2]. However, in the case of higher molecular weight phthalates, i.e. alkyl chain length of six or more carbons, no aquatic toxicity is observed at aqueous solubility. As noted earlier, the lack of hazard is attributed to the
328
T.F. Parkerton and C.A. Staples
Table 2. Derivation of PNECsoil/sediment (direct) based on equilibrium partitioning theory
PE
DMP DEP DBP BBP DEHP DINP DIDP a b
Aquatic PNEC (mg/L)
3.109 0.865 0.043 0.038 2.49 E-03 b 3.08 E-04 b 3.80 E-05 b
Log Kow
1.61 2.54 4.27 4.70 7.73 8.60 9.46
Soil/sediment PNEC (mg/kg OC)
PNEC a (mg/kg dry)
44 105 280 667 46802 42916 38358
0.44 1.05 2.80 6.67 >468 >429 >384
Assuming an organic carbon content of 1%. Water solubility limit.
combined role of low aqueous solubility and limited bioconcentration potential that prevent achieving tissue concentrations in biota needed to elicit adverse effects [2]. Consequently, for these substances the water solubility limit can be substituted into Eq. (1) to estimate a lower-bound concentration below which an ecotoxicity effect in soil or sediment is precluded. Water solubility and Log Kow values required in these calculations were taken from Cousins and Mackay [46]. In the case of a low organic carbon fraction (0.01) soil or sediment in which high bioavailability is expected, the PNEC(direct) is calculated to range from 0.44 mg/kg dry for DMP to 6.67 mg/kg dry for BBP (Table 2). In contrast for higher molecular weight phthalates, EqP predictions indicate that chronic effects are not expected at concentrations in the hundred parts per million range even in soils or sediments with low organic carbon content. 2.1.3 Association-Based Methods Based on Field Data
Association-based methods have also been used to derive sediment quality criteria for chemicals including selected phthalates (Table 3) as summarized in the U.S. EPA’s national sediment quality survey [44]. These methods are based on the empirical association between a specific biological endpoint (sediment toxicity, benthic diversity) and the concentration of the sediment contaminant determined in concurrent field samples. Barrick et al. [47] developed apparent effect thresholds (AETs) for several phthalates using concurrent chemical and biological effect data from the Puget Sound Estuary. AETs were defined for each biological indicator as the highest detected concentration among sediment samples that did not exhibit statistically significant effects. In other words, AETs characterize the highest observed sediment concentration for a given chemical that is tolerated without empirical evidence of adverse effect. A somewhat different method was used by the Florida Department of Environmental Protection [48] to calculate a probable effect level (PEL) for DEHP. The PEL was defined as the geometric mean of the 50th percentile concentration of the effects data (sediment
329
An Assessment of the Potential Environmental Risks Posed by Phthalates Table 3. Comparison of predicted no effect concentrations for sediment
PE
Causality-based PNEC (mg/kg dry) a
Association-based PNEC (mg/kg dry) b
Reference
DMP DEP [47] BBP DEHP
0.44 1.05 2.80 6.67 >468
0.16 0.2 1.4 0.9 1.3–1.9 2.65
[47] [47] [47] [47] [47] [48]
a b
Derived using Eq. 1 and data provided in Table 2. Derived using field data.
samples exhibiting biological effects) and the 85th percentile concentration of the non-effects data (sediment samples not exhibiting a statistically significant biological response). 2.1.4 PNEC Selection for Direct Effects
Comparison of association-based PNECs (Table 3) with causality-based PNECs derived using EqP indicate that the former values are 2 to 360 times lower (i.e. more conservative) with DEHP showing the greatest discrepancy. Associationbased PNECs for DEHP are clearly inconsistent with the results of laboratory toxicity tests summarized in Table 1 and thus do not provide a sound basis for risk assessment. The principle limitation of association-based PNECs is that causal relationships between concentration and biological responses cannot be established due to the confounding influence of other contaminant and non-contaminant factors that influence biological endpoints in field samples.Additional flaws in this methodology based on statistical considerations have recently been described by von Stackelberg and Menzie [49]. In contrast, lower-bound PNECs derived using EqP for DEHP, DINP and DIDP are fully consistent with the lack of toxicity observed for high molecular weight PEs. Moreover, soil and sediment toxicity test results, summarized in Table 1 provide empirical evidence that the PNECs presented in Table 2 for lower molecular weight PEs are protective for terrestrial and benthic species. For example, the PNEC for DBP is estimated to be 280 mg/kg oc whereas the chronic NOAEL for the most sensitive test species (Springtails) is >750 mg/kg oc. Further support for the use of EqP in the derivation of sediment PNECs is provided by Call et al. [10]. Based on the above discussion, PNECs obtained by extrapolation of aquatic toxicity data using EqP theory were used to quantify risks posed by direct effects. 2.2 Indirect Effects on Wildlife via the Food Chain
To assess the hazard posed to wildlife that consume terrestrial or benthic organisms that have been exposed to PEs in soil or sediment a no observed adverse
330
T.F. Parkerton and C.A. Staples
Table 4. Long-term laboratory toxicity studies with rats
PE
Study type
Effect endpoint a
NOAELrat (mg/kg/day)
Ref.
DMP DEP DBP BBP DEHP DINP DIDP
Cancer Teratology 2-Gen. Repro. b 2-Gen. Repro. b 2-Gen. Repro. b 2-Gen. Repro. b 2-Gen. Repro. b
Growth Material survival Litter size No effects Pup survival during lactation No effects Pup survival during post-partum
1440 1800 60 >100 113 >600 108
[51] [52] [53] [54] [55] [56] [57]
a b
Most sensitive population-based endpoint reported to show a significant effect. Two-generation reproductive toxicity test.
effect level (NOAELwildlife) must be defined. This value should be derived from long-term dietary toxicity studies with mammals or birds and be based on effect endpoints relevant to wildlife populations, i.e. survival, growth and reproduction. David et al. [50], has recently provided a detailed review of the available laboratory toxicity studies for PEs with mammals. Long-term dietary toxicity studies with rats are available for all the PEs considered in this analysis. Based on the available toxicological database, a NOAELrat was selected from the most relevant study that demonstrated population-based effects (Table 4). In contrast to the extensive toxicological database available for PEs in mammals, limited toxicological data are available in avian species. As in the case of mammals, PEs are reported to exhibit low acute toxicity to birds [58]. Chronic data in avian species are available for DEHP. O’Shea and Stafford [59] reported no adverse effects on survival or growth of European starlings fed DEHP at a dietary concentration of 250 mg/kg for 30 days corresponding to a NOAEL >30 mg/kg body wt/day. In a 4-week feeding study with chickens, egg production and growth were decreased at ca. 300 mg/kg body wt/day [60]. In another 230 day feeding study with chickens, cessation of egg production was reported at ca. 600 mg/kg body wt/day [61]. These studies suggest long-term effects for DEHP in avian species occur in the same range as reported for rats (i.e. NOAELrat =113 mg/kg body wt/day, Table 4). The derivation of NOAELwildlife from NOAELrat requires extrapolation factors for allometric scaling of dose as well as uncertainty regarding species sensitivity. These considerations can be expressed in equation form [62] as:
Where:
Wrat NOAELrat NOAELwildlife = 394 94 UF Wwildlife
0.33
(3)
Wrat body weight of rat (kg) Wwildlife body weight of wildlife (kg) UF uncertainty factor for interspecies sensitivity A recent review of ecological risk assessments conducted in the U.S. revealed an UF of 10 is typically assumed for interspecies extrapolation [62]. This study
An Assessment of the Potential Environmental Risks Posed by Phthalates
331
also reported that the weight of mammalian wildlife receptors ranged from 0.025 (deer mouse) to 100 kg (harbor seal). For derivation of a risk-based PNECsoil/sediment (indirect) intended to protect wildlife from indirect exposure via the food chain, the NOAELwildlife is equated to the dietary dose derived via this pathway: NOAELwildlife =Iprey Rwd Flip BSAF PNEC(indirect)
(4)
Where: Iprey Rwd Flip BSAF
wildlife ingestion rate to prey (kg prey dry/kg wildlife/day) wet to dry weight ratio of prey (kg wet/kg dry) lipid fraction of prey (kg lipid/kg wet) Biota to soil/sediment accumulation factor normalized to lipid and organic carbon (kg oc/kg lipid)
The ingestion rate of prey can be estimated based on the allometric equation provided by Nagy [63]: –0.18 Iprey =0.07 Wwildlife
(5)
Where: Wwildlife is the body weight of the wildlife receptor in kg wet Substituting Eqs. (3) and (5) into (4) and solving for the PNEC yields: 0.33 14 NOAELrat Wrat PNEC(indirect) = 389993 0.15 UF Rwd Flip BSAF W wildlife
(6)
Due to the susceptibility of PEs to biotransformation, these substances are not expected to undergo biomagnification [5]. In fact in a recent field study decreasing concentrations of phthalates in biota (i.e. biodilution) have been demonstrated with increasing trophic position for high molecular weight PEs [64]. Consequently, organisms at the base of the food web possessing limited metabolic capability (e.g. mollusks) are expected to exhibit the highest concentration of PEs. For these organisms, the Equilibrium Partitioning model provides a conservative characterization of the BSAF in Eq. (6). In order to apply Eq. (6), typical values are assumed for all input parameters except Wwildlife in which the maximum value reported by Duke and Taggart [62] is selected. An extreme value for this input was chosen to ensure calculated PNECs are conservative. Based on the following assumptions: Wrat Wwildlife Rwd Flip UF BSAF
0.48 kg 100 kg 5 kg wet/kg dry 0.01 kg lipid/kg wet 10 1
Substitution into Eq. (6) results in the following approximation: PNEC(indirect) = 10 NOAELrat
(7)
332
T.F. Parkerton and C.A. Staples
An alternative approach to wildlife effect assessment is provided by the European Technical Guidance Document of new and existing substances [65]. The approach used to assess the potential for “secondary poisoning” via the food chain first involves calculation of a predicted no effect concentration in the diet of a wildlife consumer: NOAELrat CF PNECoral = 399 AF
(8)
Where: PNECoral CF NOAELrat AF
Predicted no effect concentration in the diet (mg/kg diet) Conversion factor (kg body wt – day/kg diet) No adverse effect level from chronic rat study (mg/kg body wt/day) Application factor to account for interspecies variation and lab to field extrapolations
The default value for the conversion factor varies between 10–20 for rats depending on test animal size while the default application factor of 30 is applied to a rat chronic study for extrapolation purposes. If the Equilibrium Partitioning model is applied in conjunction with Eq. (8) the following equation is obtained for soil/sediment: NOAELrat CF PNEC(indirect) = 399 AF Flip BSAF
(9)
As in Eq. (7), the PNEC is expressed on an organic carbon basis and other variables are as previously defined. Given the following inputs: CONV Flip AF BSAF
10 0.01 kg lipid/kg wet 30 1
Substitution into Eq. (9) yields the following result: PNEC(indirect) =33 NOAELrat
(10)
Hence the EU TGD approach for wildlife effect assessment gives a similar, albeit slightly less conservative, PNEC than obtained using the methodology outlined for deriving Eq. (7). The above analysis has focused on potential adverse effects to wildlife predators that ingest soil or sediment-dwelling biota. However herbivores should also be considered. Past research suggests that phthalates are very inefficiently transferred from soil to plants hence this is not expected to be a significant wildlife exposure pathway of concern [1]. However, herbivorous wildlife or domestic livestock may ingest significant amounts of soil. In a recent study by Rhind et al. [66], the amount of DEHP ingested by sheep via soil from pastures amended with sewage sludge was investigated. This study found that sheep weighting 25–80 kg ingested 28 to 135 g dry soil per day depending on season.A maximum daily soil ingestion rate of 314 g dry soil was also reported. Applying Eq. (3) for derivation
An Assessment of the Potential Environmental Risks Posed by Phthalates
333
of a NOAEL for sheep and equating this to the maximum dose that a sheep would receive via soil ingestion enables a soil PNEC to be derived:
NOAELrat Wrat PNEC(indirect) = 394 91 UF Isoil foc Wsheep
0.33
(11)
Where: Isoil ƒoc
Soil ingestion rate for sheep (kg soil/kg body wt/day) Organic carbon fraction of soil (kg oc/kg dry)
Applying the same defaults for Wrat and UF as previously described and assuming a 0.01 organic carbon fraction and a maximum soil ingestion rate of 0.314 kg/day for a 25 kg sheep yields: PNEC(indirect) =216 NOAELrat
(12)
Comparison of Eqs. (12) with (7) suggests that the risks posed to wildlife by ingestion of contaminated prey will be greater than that posed via direct soil ingestion thus dictating PNEC derivation. PNECs intended to protect wildlife derived using Eq. (7) are compared to PNECs intended to protect soil and sediment-dwelling organisms in Table 5. Results indicate that direct effects drive environmental concerns for lower molecular weight PEs while indirect effects dictate the environmental hazard for the higher molecular weight PEs. Table 5. Comparison of PNECsoil/sediment for direct and indirect effects a
PE
PNECdirect b (mg/kg dry)
PNECindirect c (mg/kg dry)
DMP DEP DBP BBP DEHP DINP DIDP
0.44 1.05 2.80 6.67 >468 >429 >383
144 180 6.0 >10.0 11.3 >60 10.8
a b c
PNEC values are normalized to a 1% organic carbon content. Derived using Eq. 1, Table 1. Derived using Eq. 7.
334
T.F. Parkerton and C.A. Staples
3 Exposure Characterization 3.1 Soil 3.1.1 Native Soil
The principle source of PEs to soils that are not amended with sewage sludge is atmospheric deposition. Several studies have reported atmospheric deposition rates of PEs in different locations (Table 6). Based on these data an atmospheric deposition of 1 µgdry m–2 d–1 is typical for DMP, DEP, DBP and DEHP. Given this estimate and assuming an average mixing depth for non-agricultural soil of 0.05 m and a soil density of 1700 kgdry m–3 [65] and ignoring the mitigating role of biodegradation the resulting annual background soil concentration is esti–1 . Since atmospheric deposition of BBP appears to be mated to be 0.004 mg kgdry about an order of magnitude lower, even lower background soil concentrations are expected. 3.1.2 Sludge-Amended Soil
Sludge from municipal wastewater treatment plants is typically disposed of via incineration, placement in landfills, or via land application to agricultural fields, forested land or other sites e.g. parks, golf courses, and reclamation projects. This latter disposal method is often viewed as the most cost-effective and environmentally beneficial option [72]. The enhanced use of sludge for agricultural purposes is also a policy endorsed by the EU [73]. Certain PEs are commonly detected in sewage sludge from municipal wastewater treatment plants (Table 6). Thus, an examination of potential exposure and risks to soil-dwelling organisms and terrestrial wildlife that results from sludge application is warranted. Sludge application rates differ regionally and by type of application. For example in the US, typical sludge application rates to agricultural soils are 1 kgdry m–2 yr–1 (=10 t ha–1 yr–1) while a higher rate of 1.8 kgdry m–2 yr–1 is used for Table 6. Atmospheric deposition rates reported for phthalate esters
Location
US Great Lakes Sweden Denmark Germany Germany NR = Not reported.
Deposition flux (mg/m2/d)
Deposition Type
DMP
DEP
DBP
BBP
DEHP
Ref.
Wet+Dry Wet+Dry Wet+Dry Wet Wet
NR NR NR NR 1.15
NR NR NR NR 1.07
0.53 0.56 0.31 0.66 1.57
NR NR 0.05 NR 0.10
0.53 0.79 0.56 1.56 2.88
[67] [68] [69] [70] [71]
335
An Assessment of the Potential Environmental Risks Posed by Phthalates
forested or public lands [72]. In Canada, sludge application of 0.8 kgdry m–2 yr–1 is allowed over a 5-year period. In Europe, typical sludge application rates to agricultural soil and grassland are 0.5 and 0.1 kgdry m–2 yr–1, respectively [65] with application rates of as high as 1.7 kgdry m–2 yr–1 also reported [74]. Neglecting the role of degradation processes and background concentration in native soil, a conservative estimate of the soil concentration resulting from sludge amendment is given by: Csludge Xapp Csoil = 99 Zsoil rsoil
(13)
Table 7. Summary of phthalate concentrations in sludges and estimated upper-bound soil con-
centrations in sludge amended soils PE
Location
Year
Average sludge concentration (mg/kgdry)
No. of Samples
Ref.
Estimated soil concentration (mg/kgdry) a
DMP
Canada Denmark Europe b Canada Denmark Europe b Canada Germany Germany Germany Denmark Europe b Canada Germany Germany Germany Denmark Europe b Canada Germany Germany Germany Denmark Norway Europe b Germany Europe b Europe b
93/94 95/96 99 93/94 95/96 99 93/94 93/94 97 97 95/96 99 93/94 93/94 97 97 95/96 99 93/94 93/94 97 97 95/96 NR 99 97 99 99
0.030 0.034 <0.040 b 0.228 0.238 <0.160 c 6.84 22.5 0.5 1.83 3.88 6.50 c 2.97 11.7 <0.02 0.82 0.18 3.6 c 150 23.4 67.3 19.2 37.9 58 85.4 c 9.1 <0.5 c <1.3 c
72 11 35 72 11 35 72 50 15 7 20 35 72 50 15 28 20 35 72 50 15 7 20 36 35 5 35 35
[75] [76] [77] [75] [76] [77] [75] [78] [79] [80] [76] [77] [75] [78] [79] [80] [76] [77] [75] [78] [79] [80] [76] [81] [77] [80] [77] [77]
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.034 0.112 0.002 0.009 0.019 0.032 0.015 0.058 <0.001 0.004 <0.001 0.018 0.750 0.117 0.336 0.096 0.189 0.290 0.427 0.045 0.003 0.007
DEP DBP
BBP
DEHP
DINP DIDP a
Soil concentration estimated based on application of Eq. (13) using a upper bound sludge application rate of 1.8 kgdry m–2 yr–1. b Survey included sludge samples collected from UK, France, Germany, Sweden and The Netherlands. c Median value. NR=Not reported.
336
T.F. Parkerton and C.A. Staples
Where: Csoil Csludge Xapp Zsoil rsoil
Concentration in sludge amended soil (mg kg–1 dry) Concentration in sewage sludge (mg kg–1 dry) Sludge application rate (kgdry m–2 yr–1) Soil mixing depth (m) Soil bulk density (kgdry m–3)
Assuming a typical agricultural soil mixing depth of 0.2 m and a bulk density of 1700 kgdry m–3 [65], sludge concentrations will be diluted by two hundred-fold at an upper-bound sludge application rate of 1.7 kgdry m–2 yr–1. Repeated application in subsequent years is not expected to result in a progressive accumulation over time due to the biodegradable nature of PEs. Given average PE concentrations in sludge and the assumptions outlined above, upper-bound concentrations in sludge amended soil are calculated in Table 7. Based on a comparison to predicted exposure concentrations discussed in Sect. 3.1.1, these estimates suggest sludge amendments do not serve as a major contributor to DMP or DEP in soil. In contrast, sludge addition appears to be the dominant source of the more hydrophobic, higher molecular weight PEs. 3.1.3 Field Monitoring Data for Soil
Several recent field studies have reported PE concentrations in European soils. Monitoring data for individual PEs are summarized graphically in Figs. 1–6. In a Dutch regional monitoring survey [83], median soil concentrations of all PEs were about an order of magnitude lower than observed in an industrial area [82]. A similar range of DEHP soil concentrations was reported in sludge-amended soil in the United Kingdom [66]. The large variation in DBP, BBP, DEHP and DINP soil concentrations reported in Denmark [97] reflects the dramatic differences in sludge application rates to the soils investigated. For the higher molecular weight PEs (DBP and higher), the observed range in soil concentrations across these studies appears to correspond with the range in concentrations expected from inputs derived from background atmospheric deposition to sludge application. Moreover, DMP and DEP concentrations reported in the Dutch monitoring survey are consistently low and at a level reflective of atmospheric sources as anticipated. However, observed DMP and DEP soil concentrations reported in Germany are much higher than expected based on exposure calculations presented in the previous sections. This discrepancy suggests that the overall magnitude of emissions to soil in this local industrialized area is significantly higher than presumed for these PEs. 3.2 Sediment
Sources of phthalates to the aquatic environment include industrial and domestic wastewater effluents as well as non-point source inputs such as urban runoff
An Assessment of the Potential Environmental Risks Posed by Phthalates
337
Concentration (mg/kg dry wt) Fig. 1. Median and range of dimethyl phthalate (DMP) soil and sediment concentrations re-
ported in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line. Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. BDL indicates all samples examined were below the analytical detection limit(s). Letters correspond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83]; C=Furtmann 1993 [71]; D=Alcontrol 1999 [83]; E=Vethaak et al. 2002 [84]; F=Niva 1996 [85]; G=Parkmann and Remberger 1995, 1996 [86, 92]; H=Tan 1995 [87]; I=Lopes and Furlong 2001 [88]; J=USEPA 1997 [44]; K=Garrett 2002 [89]; L=Mackintosh et al. 2002 [90]
and atmospheric deposition [69]. While phthalates can undergo anaerobic biodegradation, as is the case for many organic chemicals, degradation generally occurs more slowly than in aerobic soils [5]. Differences in expected half-lives between soil and sediment largely explain the higher concentrations predicted in sediment by multimedia fugacity models [4]. 3.2.1 Field Monitoring Data for Sediments
Field data published from the 1990’s are shown in Figs. 1–6. Data compilation was restricted to recent studies to reduce the confounding problem of laboratory contamination that has plagued interpretation of early PE measurements while also facilitating direct comparison of contemporary sediment measurements with recent soil surveys discussed in Sect. 3.1.3.
338
T.F. Parkerton and C.A. Staples
Concentration (mg/kg dry wt) Fig. 2. Median and range of diethyl phthalate (DEP) soil and sediment concentrations reported
in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line. Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. BDL indicates all samples examined were below the analytical detection limit(s). Letters correspond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83]; C=Furtmann 1993 [71]; D=Vitali et al. 1997 [71]; E=Alcontrol 1999 [83]; F=Vethaak et al. 2002 [84]; G=Niva 1996 [85]; H=Parkmann and Remberger 1995, 1996 [92] [86]; I=Tan 1995 [87]; J=JAE 1993 [93]; K=MOC 1999a, b, c [94–96]; L=Lopes and Furlong 2001 [88]; M=USEPA 1997 [44]; N=Garrett 2002 [89]; O=Mackintosh et al. 2002 [90]
Several generalizations are apparent from this analysis. First, an extensive monitoring database for single isomer PE concentrations in sediments from North America, Europe and Asia is available. Limited data are also available for the mixed isomers, DINP and DIDP. Median sediment concentrations of the individual PEs follow the general trend: DEHP >DBP, DINP, DIDP >BBP, DMP, DEP. Second, PE sediment concentrations often exhibit a several order of magnitude range both within and between monitoring studies. This variance reflects sitespecific factors associated with sampling locations (e.g. vicinity of point source inputs, degree of industrial activity, receiving water dilution), the non-persistent nature of PEs (e.g. degradation as one proceeds from a source) as well as potential differences in analytical methodology (e.g. method detection limits, extraction efficiencies). Third, there are no obvious geographical differences in sediment concentrations which is likely due to the large scatter in results reported
An Assessment of the Potential Environmental Risks Posed by Phthalates
339
Concentration (mg/kg dry wt) Fig. 3. Median and range of dibutyl phthalate (DBP) soil and sediment concentrations reported
in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line. Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. Letters correspond to the following references: A=UMEG 1999 [82]; B=Alcontrol 1999 [83]; C=Vikelsoe et al. 1999 [97]; D=Vitali et al. 1997 [71]; E=Furtmann 1993 [71]; F=Fromme et al. 2002 [79]; G=Vondracek et al. 2001 [98]; H=Alcontrol 1999 [83]; I=Vethaak et all 2002 [84]; J=Vikelsoe et al. 2001 [99]; K=Parkmann and Remberger 1995, 1996 [86, 92]; L=NIVA 1996 [85]; M=Tan 1995 [87]; N=JAE 1993 [93]; O=MOC 1999a, b, c [94–96]; P=Lopes and Furlong 2001 [88]; Q=USEPA 1997 [44]; R=Garrett 2002 [89]; S=Mackintosh et al. 2002 [90]
across different world regions. Lastly, median sediment concentrations typically bracket median soil concentrations. However, comparison of the regional survey of sediment and soil PE concentrations in the Netherlands clearly shows that median concentrations in sediment are higher than soil [83]. Moreover, maximum concentrations exceed those reported in soils often by orders of magnitude (Figs. 1–6).
4 Risk Assessment To assess the potential risk that soil and sediment associated phthalates pose to terrestrial and benthic organisms by direct exposure or to wildlife by indirect ex-
340
T.F. Parkerton and C.A. Staples
Concentration (mg/kg dry wt) Fig. 4. Median and range of butylbenzyl phthalate (BBP) soil and sediment concentrations re-
ported in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line. Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. Letters correspond to the following references: A = UMEG 1999 [82]; B = Alcontrol 1999 [83]; C = Vikelsoe et al. 1999 [97]; D = Vitali et al. 1997 [71]; E = Furtmann 1993 [71]; F = Fromme et al. 2002 [79]; G = Alcontrol 1999 [83]; H = Vethaak et all 2002 [84]; I = Vikelsoe et al. 2001 [99]; J = NIVA 1996 [85]; K = Parkmann and Remberger 1995, 1996 [86, 92]; L = JAE 1993 [93]; M = MOC 1999a, b, c [94–96]; N = Lopes and Furlong 2001 [88]; O = USEPA 1997 [44]; P = Garrett 2002 [89]; Q = Mackintosh et al. 2002 [90]
posure via the food chain, observed concentrations were compared to the lowest PNEC given in Table 5. Results of this comparison are illustrated in Fig. 1-6 for each of the PEs investigated. In cases where the maximum concentration for a given study exceeds the PNEC (denoted by a solid vertical line), the number of exceedances (indicated in parenthesis) is specified next to the total number of samples analyzed (shown on the right hand axis of the plot). For all PEs investigated, none of the observed soil concentrations exceeded risk-based limits even in the case of soils that were heavily amended with sewage sludge. Similarly, for BBP and DINP, none of the maximum sediment concentrations reported in field surveys exceeded the PNEC. For the remaining phthalates, at least one monitoring study indicated maximum reported sediment concentrations above the PNEC. In the case of DMP, studies from the US and the Nether-
An Assessment of the Potential Environmental Risks Posed by Phthalates
341
Concentration (mg/kg dry wt) Fig. 5. Median and range of di-2-ethylhexy phthalate (DEHP) soil and sediment concentrations
reported in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line. Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. Letters correspond to the following references: A = UMEG 1999 [82]; B = Alcontrol 1999 [83]; C = Rhind et al. 2002 [66]; D = Vikelsoe et al. 1999 [97]; E = Vitali et al. 1997 [71]; F = Furtmann 1993 [71]; G = Fromme et al. 2002 [79]; H = Vondracek et al. 2001 [98]; I = Alcontrol 1999 [83]; J = Vethaak et all 2002 [84]; K = Long et al. 1998 [100]; L = Boutrup et al. 1998 [101]; M = Vikelsoe et al. 2001 [99]; N = NIVA 1996 [85]; O = Parkmann and Remberger 1995, 1996 [86, 92]; P = Tan 1995 [87]; Q = JAE 1993 [93]; R = MOC 1999a, b, c [94–96]; S = USEPA 1997 [44]; T = Lopes and Furlong 2001 [88]; U = Garrett 2002 [89]; V = Mackintosh et al. 2002 [90]
lands indicated 34 sediment samples were above the PNEC thus representing less than 2% of the reported sediment measurements included in Fig. 1. For DEP, two studies from North America yielded 5 samples exceeding the PNEC thus representing less than 0.3% of the available sediment monitoring database. Four monitoring studies from both North America and Europe indicated 14 sediment samples with DBP concentrations above the PNEC which translates to 0.6% of the reported sediment measurements. In the case of DEHP, six studies representing field data sets from North America, Europe and Asia were shown to include a limited number of measurements that exceeded the PNEC. The elevated concentrations reported in Sweden corresponded to samples taken near a production site. Collectively across all monitoring studies, 28 sediment samples were found to be above the PNEC representing 1.4% of the reported DEHP measurements. Three
342
T.F. Parkerton and C.A. Staples
Concentration (mg/kg dry wt) Fig. 6. Median and range of di-isononyl and di-isodecyl phthalate (DINP, DIDP) soil and sed-
iment concentrations reported in field surveys. The total number of samples analysed in each study is provided on the right hand axis. The predicted no effect concentration (PNEC) is indicated by a solid vertical line.Values in parenthesis indicate the number of samples for a given study that exceed the PNEC. BDL indicates all samples examined were below the analytical detection limit(s). Letters correspond to the following references: A, E = Mackintosh et al. 2002 [90]; B, D, F, G = Alcontrol 1999 [83]; C = Vikelsoe et al. 1999 [97]
sediment samples in the immediate vicinity of a European DIDP production plant with limited wastewater treatment facilities were found to exceed the PNEC for this substance. However median concentrations reflecting the regional exposure situation for DIDP are 2–3 orders of magnitude lower.
5 Summary and Discussion Risk assessment should form the logical basis for rationale management of contaminated soils and sediments. Derivation of transparent, scientifically defensible, causal, risk-based soil and sediment quality criteria is a critical aspect of the risk assessment process. In this study, the technical basis used in developing PNECs for each PE is explained. These values are then used for screening available monitoring data using the simple hazard quotient (i.e. PEC/PNEC) paradigm. Our analysis suggests that in the case of the low molecular weight PEs, direct toxicity to soil or sediment-dwelling organisms dictates PNEC derivation.
An Assessment of the Potential Environmental Risks Posed by Phthalates
343
In contrast, indirect effects on wildlife via food chain exposure determine PNECs for higher molecular weight PEs. However, despite the hydrophobicity of higher molecular weight PEs, trophic transfer is limited by metabolism at higher trophic levels [5, 64, 102, 103]. As a result, the PNECs presented are intended to protect wildlife that exclusively consume benthic and terrestrial prey which are at the base of the food chain and lack metabolic capabilities (e. g. mollusks, oligochaetes). Screening calculations suggest dietary exposure via ingestion of contaminated prey is a more important pathway then direct ingestion of contaminated soil or sediment. These calculations were based on the extreme soil ingestion rate reported for sheep. However, weight normalized soil ingestion rates for other domestic animals appear similar to sheep further supporting this general conclusion [102]. It is important to highlight some key assumptions invoked in the PNEC derivation since these are likely to contribute conservatism in the numerical values obtained. First, PNECs were calculated using an organic carbon content of only 1%. Often, contaminated soil or sediments have a much higher organic carbon content. Therefore, in accordance with the EqP theory, this assumption is likely to overstate the bioavailability of highly contaminated field samples. This conservatism can be illustrated by reexamining the number of sediment measurements reported in the USEPA national sediment quality inventory that exceed the organic carbon normalized PNEC. In the case of DMP, only 8 sediment samples out of 580 where organic carbon measurements were also reported, exceeded the PNEC. This represents a decline in the exceedance frequency from 3% (30/1001 on a dry weight basis) to 1.4% (8/580 on an organic carbon basis). In the case of DBP, only 2 out of 520 organic carbon normalized sediment concentrations were above the PNEC reflecting a reduction in the exceedance frequency from 0.9% (dry weight basis) to 0.2% (organic carbon basis). None of the organic carbon normalized concentrations of either DEP or DEHP exceeded the corresponding PNEC in contrast to dry weight based concentrations reported in this data set (Figs. 2, 5). A second assumption likely to introduce conservatism is that factors affecting the bioavailability of PEs in field samples have been ignored. While the EqP theory may be applicable to freshly spiked laboratory tests [104], such assumptions may introduce significant conservatism when applied to field samples that have been subjected to sequestration processes and/or contain PEs in an inert form. The importance of these considerations in risk assessment of soils and sediments is well recognized [105–107]. Recent laboratory and field studies clearly demonstrate the reduction in DEHP bioavailability to soil microbes over time [74, 108]. The relationship between PE bioavailability and environmental persistence is further reviewed by Peterson and Staples [109]. Recent work also suggests that certain PEs may be present in sewage sludge as abraded PVC particles thus occurring in an occluded state within the polymer matrix [110]. In this form, PEs may be less bioavailable than EqP predictions for either toxicity or bioaccumulation indicate. A third aspect of the PNEC derivation affording conservatism relates specifically to the protection of wildlife consumers. In the screening risk assessment presented, the spatial scale of PE contamination in soil/sediment is not consid-
344
T.F. Parkerton and C.A. Staples
ered. However, PNECs are based on the assumption that the wildlife consumers receive their entire diet from an area corresponding to the PNEC concentration. Since individual grab samples may represent only very localized “hot spots”, the actual exposure that resident wildlife consumers receive can in fact be much lower since prey from less contaminated areas also comprise a portion of the diet. Thus, neglecting spatial scale may overstate potential risks to wildlife consumers. With respect to exposure assessment, considerable field data are available characterizing PEs concentrations in sediments in different regions. Less information is generally available on PE concentrations in soil although available monitoring data suggest maximum concentrations found in sediments exceed concentrations observed in the soil compartment. For a number of the monitoring studies available, sampling has specifically targeted or at least included sites that are expected to exhibit high PE contamination (e.g. production sites). Thus, observed exposure measurements reflect both regional and local exposure scenarios which in part explains the wide range of concentrations reported as noted previously. Differences in exposure concentrations may also reflect differences in analytical protocol for PE determination that were taken to preventing or correcting for background laboratory contamination. In a number of the studies little information is provided on these critical analytical details. For example, in the study by Long [100], the issue of blank determination is not even mentioned. It is interesting to note that median sediment concentrations reported in this study are higher than in any of the other studies. The uncertainty in the analytical methods employed in this study at least questions the reliability of these data for use in the context of risk assessment. Despite the conservatism in PNEC derivation and the potential bias in exposure characterization, the key conclusion from this screening risk assessment is that environmental concerns posed by soil and sediment-associated PEs are indeed low. Given the available hazard and exposure data, environmental concerns are, at worst, restricted to infrequent, localized hot spots of contaminated sediments. At these sites, a more detailed risk assessment would be needed to refine the conservative assumptions used in the present analysis. Additional exposure data on PE concentration in soils and mixed isomer concentrations in sediments (particularly from the US and Japan) are suggested as future needs for complimenting the existing monitoring database and further confirming the conclusion reached in the present study. The general conclusion of this screening risk analysis for the soil/sediment compartment is consistent with the findings of a similar PE risk assessment focused on the surface water compartment [3]. This study concluded that PEs do not pose a threat to aquatic organisms in North American and Western European surface waters. The results of this study indicate that current pollution controls are generally quite effective in limiting PEs in soil and sediment to concentrations that are well below risk-based thresholds. Consequently, future regulatory initiatives that are intended to improve environmental quality by imposition of additional controls or restrictions, specifically on phthalates, are expected to provide little benefit. For example, based on the analysis presented in this study the proposed EU limit for DEHP of 100 mg/kg in sewage sludge [73] not only lacks a sound risk-based
An Assessment of the Potential Environmental Risks Posed by Phthalates
345
justification but also restricts the practical societal use of a cost-effective and environmentally beneficial waste disposal option. Similarly, the risk analysis presented in this study has shown that application of association-based effect thresholds for phthalates can lead to misdirected policy priorities for the management of contaminated sediments. These examples highlight the role that a environmental risk assessment can serve in guiding rational decision-making.
6 References 1. Staples CA, Parkerton TF, Adams WJ, Gorsuch JW, Biddinger GR, Reinhert K (1997a) Environ Toxicol Chem 16:875 2. Parkerton TF, Konkel WJ (2000) Ecotox Environ Saf 45:61 3. Staples CA, Parkerton TF, Peterson TR (2000) Chemosphere 40:885 4. Cousins I, Mackay D, Parkerton TF (2002) Physical Chemical Properties and Evaluative Fate Modeling of Phthalate Esters, Chapter submitted to Springer-Verlag Environmental Chemistry Handbook Volume for Phthalate Esters 5. Staples CA, Peterson DR, Parkerton TF, Adams WJ (1997b) Chemosphere 35:667 6. Cartwright CD, Thomas IP, Burns RC (2000) Environ Sci Technol 19:1253 7. Herring R, Bering L (1988) Contam Toxicol 40:626 8. Neuhauser EF, Durkin PR, Malecki MR, Anatra M (1986) Comparative Biochemistry and Physiology. 83C.1.197–200 9. Hulzebos EM, Adema DMM, Dirven-Van Breemen EM, Henzen L, van Dis WW, Herbold HA, Hoekstra JA, Baerselman R, van Gestel CAM (1993) Environ Toxicol Chem 12:1079 10. Call DJ, Cox CA, Geiger DL, Genisot KI, Markee TP, Brooke LT, Polkinghorne CN, Vande Venter FA, Robillard KA, Gorsuch JW, Parkerton TF, Reiley MC, Ankley GT, Mount DR (2001) Environ Toxicol Chem 20:1805 11. Shea PJ, Weber JB, Overcash MR (1982) Bull Environ Contam Toxicol 29:153 12. Jensen J, van Langevelde J, Putze G, Henning Krough P (2001) Environ Toxicol Chem 20:1085 13. Tagatz ME, Plaia GR, Deans CH (1986) Bull Environ Contam Toxicol 37:141. 14. Huntingdon Life Science (SLU 001/983882) Butylbenzylphthalate, acute toxicity (LC50) to the earthworm (Eisenia foetida) (1998) 15. Mathur SP (1974) J Environ Quality 3:207 16. Kirchmann H, Aström H, Jönsäll G (1991) Swedish J Agric Res 21:107 17. Kaiser T, Felgentreu D (2000) Determination of the influence of Di-2 ethylhexyl phthalate (DEHP) on soil quality – Part I: Determination of soil microflora activity. Berlin, Federal Biological Research Centre for Agriculture and Forestry; Institute of Ecological Chemistry 18. Kaiser T, Felgentreu D (2000a) Determination of the influence of Di-2 ethylhexyl phthalate (DEHP) on soil quality – Part II: Determination of dehydrogenase activity. Berlin, Federal Biological Research Centre for Agriculture and Forestry; Institute of Ecological Chemistry 19. Kaiser T, Felgentreu D (2000b) Determination of the influence of Di-2 ethylhexyl phthalate (DEHP) on soil quality – Part III: Determination of nitrogen mineralization in soils and the influence of DEHP on these processes. Berlin, Federal Biological Research Centre for Agriculture and Forestry; Institute of Ecological Chemistry 20. Mutz RC, Jones JR (1977) The effects of phthalate esters on geochemical cycles in freshwater hydrosoil. Transactions of the Missouri Academy of Science. Vol 10 and 11, p 296 21. Larsson P, Thurén A, Gahnström G (1986) Environ Poll (Series A) 42:223 22. Aranda JM, O’Connor GA, Eiceman GA (1989) J Environ Qual 18:45 23. Difenbach R (1998a) Bestimmung der Auswirkungen von Vestinol AH auf Regenwürmer (Eisenia foetida), RW-71, Huls Infracor GmbH, Marl, Germany, 15 pp
346
T.F. Parkerton and C.A. Staples
24. Difenbach R (1998b) Bestimmung der Auswirkungen von Vestinol AH auf das Wachstum Terrestrischer Planzen, Huls Infracor GmbH, PF-56, Marl, Germany, 20 pp 25. Stanley RD, Tapp JF, Williams BRH (1982) An assessment of ecotoxicological test methods: part VIII. The effect of nine chemicals on the growth of Avena sativa and Brassica rapa, Brixham Laboratory, Imperial Chemical Industries, Report BL/A/2164, 26 pp 26. Woin P, Larsson P (1987) Bull Environ Contam Toxicol 38:220 27. Brown D, Thompson RS, Stewart KM, Croudace CP, Gillings E (1996) Chemosphere 32:2177 28. Wennberg L, Parkman H, Remberger M,Viktor T,Williams C (1997) The influence of sediment-associated phthalate esters (DEHP and DIDP) on hatching and survival of the moorfrog, Rana arvalis. IVL, Box 21060, S-100 31 Stockholm, Sweden. IVL report B1260, 26 pp 29. Solyom P, Remberger M, Viktor T (2001) Further Investigations on the Influence of Sediment-Associated Phthalate Esters (DEHP and DINP) on Hatching and Survival of the Moorfrog Rana arvalis, IVL Swedish Environmental Research Institute Report No. A20173, 42 pp 30. Perez KT, Davey EW, Lackie NF, Morrison GE, Murphy PG, Soper AE,Windslow DL (1983) Environmental assessment of a phthalate ester, Di(2-ethylhexyl) phthalate (DEHP), derived from a marine microcosm. Aquatic Toxicology and Hazard Assessment: Sixth Symposium, ASTM STP 802, WE Bishop, RD Cardwell, and BB Heidolph (Eds), American Society for Testing and Materials, Philadelphia, pp 180– 191. 1983 (ASTM Special technical publ. 0066–0558:802) 31. Exxon Biomedical Sciences, Inc. (2002) Microbial Respiration Inhibition in Soil. Unpublished draft report 199694B. May 7, 1999 32. Exxon Biomedical Sciences, Inc. (1996a) “Seed germination limit test with rye grass and lettuce, East Millstone, NJ. Unpublished report No 199674 33. Exxon Biomedical Sciences, Inc. (1996b) “Seed germination test with lettuce, East Millstone, NJ. Unpublished report No 199674A 34. Exxon Biomedical Sciences, Inc. (2000) Lettuce, Seed Germination and Growth Test, Unpublished draft report 199674C, June 23, 1999 35. Exxon Biomedical Sciences, Inc. (1996) “Earthworm limit test,” East Millstone, NJ. Unpublished report No 1996 92 36. Larsson P, Thurén T (1987) Environ Toxicol Chem 6:417 37. Van Wezel AP, van Vaardingen P, Posthumus R, Crommentuijn GH, Sijm DT (2000) Ecotox Environ Saf 46:305 38. Yeung P (1990) A method for measuring water repellent soil, Proceedings of the 27th Annual Alberta Soil Science Workshop, Edmonton, Alberta, Canada, pp 59–64 39. Van Gestel CAM, Ma M (1990) Chemosphere 21:1023 40. Di Toro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE, Thomas NA, Paquin PR (1991) Environ Toxicol Chem 10:1541 41. van Leeuwen CJ, Van der Zandt PTJ, Aldenberg T, Verhaar HJM, Hermens JLM (1992) Environ Toxicol Chem 11:267 42. Belfroid AC, Seinen W, Gestel CAM, Hermens JLH, van Leeuwen CJ (1995) Environ Sci Pollut Res 2:5 43. EC. 1996. Commission of the European Communities. Technical Guidance Documents in Support of the Commission Directive 93/67/EEC on risk assessment for new substances and the Commission Regulation (EC) No 1488/94 on risk assessment for existing substances. Office of Publications of the European Communities, Luxembourg 44. USEPA (1997) The incidence and seventy of sediment contamination in surface waters of the United States, volume 1: National Sediment Quality Survey, EPA 823/R-97/006, Washington DC, USA 45. Seth R, Mackay D, Muncke J (1999) Environ Sci Technol 33:2390 46. Cousins IT, Mackay D (2000) Chemosphere 41:1389 47. Barrick R, Becker S, Brown L, Beller H, Pastouk R (1988) Sediment quality values refinement: 1988 Update and evaluation of Puget Sound AET, volume 1, Puget Sound estuary program, office of Puget Sound, Seattle, WA, USA
An Assessment of the Potential Environmental Risks Posed by Phthalates
347
48. Florida Department of Environmental Protection (1994) Approach to the assessment of sediment quality in Florida coastal waters, volume 1, Development and evaluation of sediment quality assessment guidelines, office of water policy, Talahasse, FL, USA 49. Von Stackelberg, Menzie CA (2002) Environ Toxicol Chem 21:466 50. David RM, McKee RH, Butala JH, Barter RA, Kayser M (2001) In: Bingham E, Cohrosen B, Powell CH (eds) Esters of aromatic mono-, di-, and tricarboxylic acids, aromatic diacids and di-, tri-, or polyalcohols, Pattys Toxicology, 5th edn, vol 6, pp 635–932 51. Lehman, AJ (1955) Insect repellents. Food and Drug Officials of the United States, Quarterly Bulletin 19, pp 87–99 52. Field AE et al. (1993) Teratology 48:33 53. Patel R, Wolfe G, Rouselle S, Cherian G, Ko K, Moore R, Bishop J, Chapin R (2001) The Toxicologist 60:385 54. Nagao T, Ohta R, Marumo H, Shindo T, Yoshimura S, Ono H (2000) Reproductive Toxicology 14:513 55. Schilling K, Gembart C, Hellwig J (2001) Di-2-ethylhexyl phthalate – Two-generation reproduction toxicity study in Wistar rats. Continuous Dietary Administration. Unpublished laboratory report 56. Waterman S, Keller L, Trimmer G, Freeman J, Nikiforov A, Harris S, Nicolich M, McKee R (2000) Reproductive Toxicology 14:21 57. Hushka L, Waterman S, Keller L, Trimmer G, Freeman J, Ambroso J, Nicolich M, McKee R (2001) Reproductive Toxicology 15:153 58. Hill EF, Heath RG, Spann JW,Williams JD (1975) Lethal dietary toxicities of environmental pollutants to birds, Washington DC, US Fish and Wildlife Service, Special Scientific Report Wildlife No. 191, pp 61 59. O’Shea TJ, Stafford CJ (1980) Bull Environ Contam Toxicol 25:345 60. Wood DL, Bitman J (1984) Poultry Sci 36:469 61. Ishida M, Suyama K, Adachi S (1981) J Agriculture Food Chem 29:72 62. Duke LD, Taggart M (2000) Environ Toxicol Chem 19:1668 63. Nagy KA (1987) Ecol Monogn 57:111 64. Gobas FAPC, Mackintosh CE, Webster G, Ikonomou M, Parkerton TF, Robillard K (2002) Bioaccumulation of Phthalate Esters in Aquatic Food-Webs, Chapter submitted to Springer-Verlag Environmental Chemistry Handbook Volume for Phthalate Esters 65. ECB (2002) Draft Revision of the Technical Guidance 7 Document on Risk Assessment. Environmental Risk Assessment, Chapter 3, European Chemicals Bureau, Ispra, Italy, 211 pp 66. Rhind SM, Smith A, Kyle CE, Telfer G, Martin G, Duff E, Mayes RW (2002) J Environ Monit 4:142 67. Eisenreich SJ, Looney BB, Thornton JD (1981) Environ Sci Technol 15:30 68. Thuren A, Larsson P (1990) Environ Sci Technol 24:554 69. Vikelsoe J, Thomsen M, Johansen E (1998) Sources of phthalates and nonylphenol in municipal waste water: a study in a local environment, NERU Technical Report No 225, Ministry of Environment and Energy, National Environmental Research Institute, Roskilde, Denmark, 52 pp 70. Schleyer R, Renner I, Mühlhausen D (1991) Immisionsbelastung (1991) – Konsequenzen für die Grundwasserqualität BGA 71. Furtmann K (1993) Phthalates in the Aquatic Environment, PhD dissertation, Regional Water and Waste Water Authority, Nordrhein-Westfalen 72. USEPA (1995) Process Design Manual: Land Application of Sewage Sludge and Domestic Seepage, Office of Research and Development, US Environmental Protection Agency, Washington, DC, EPA 625/K-95/001, 301 pp 73. EC (2001) Organic Contaminants in Sewage Sludge for Agricultural Use, European Commission Joint Research Center, Institute for Environment and Sustainability and Waste Unit, 73 pp 74. Vikelsoe J, Thomsen M, Carlsen L (2002) Sci Tot Environ (in press) 75. Weber MD, Nichols JA (1995) “Organic and Metal Contaminants in Canadian Municipal Sludges and a Sludge Compost.” Wastewater Technology Centre, Burlington, ON, Canada
348
T.F. Parkerton and C.A. Staples
76. Kristensen P, Torslov J, Samsoe-Peterson L, Rasmussen JO (1996) “Sludge and Waste Products to be used on Arable Land.” MilJoProject No. 328. MilJo-09, Energiministeriet, Miljostyvelsen. Copenhagen, Denmark,165 pp 77. Research Institute for Chromatography (RIC) (2000) Kortrijk, Belgium. RIC report ECPI\2000–05–0007 78. Schnaak W, John T (1994) “Investigations of organic pollutants in sewage sludge and the ecotoxicological evaluation of sewage sludge amended soils.” Prepared by Fraunhofer – Institut for Umweltchemie und Okotoxikologie for Landesumweitamt Brandenburg, Germany 79. Fromme H, Kuchler T, Otto T, Pilz K, Muller J, Wenzel A (2002) Wat Res 36:1429 80. Kolb M, Welte K, Mettenleiter S, Trinkmann A, Bestimmung von Phthalsäureestern in Klärschlämmen mittels GC/MS, Wasser and Boden 49:5/1997 81. Paulsrud BA, Wien and Nedland KT (2000) A Survey of Toxic Organics in Norwegian Sewage Sludge, Compost and Manure. – Aquateam, Norwegian Water Technology Centre ASOSLO 82. UMEG (1999) Bodenzustandsbericht Grossraum Stuttgart. – Herausgeber: Ministerium für Umwelt und Verkehr, Baden-Württemberg, 107S 83. AL Control Biochem Laboratoria (1999) The Analysis Of Phthalates In Soil And Sediment, Hoogvliet, The Netherlands, 11 pp + appendices 84. Vethaak AD, Rijs GBJ, Schrap SM, Ruiter H, Gerristen A, Lahr J (2002) Estrogens and Xenoestrogens in the Aquatic Environment of the Netherlands, Occurrence, Potency and Biological Effects, Directorate-General for Public Works and Waste Management, RIZA/RIKZ Report no 2002.001, 293 pp 85. Norwegian Institute for Water Research (NIVA), “Occurrence of Phthalates and Organotins in Sediments and Water in Norway,” Report SNO 3552–96 86. Parkman H, Remberger M (1996) “Phthalates in Water and Sediments in Major Cities and Remote Lakes in Sweden,” IVL, Swedish Environmental Research Institute, Stockholm, Sweden 87. Tan GH (1995) Bull Environ Contam Toxicol 54:171 88. Lopes TJ, Furlong ET (2001) Environ Toxicol Chem 20:727 89. Garrett CL (2002) Phthalate Esters in Harbour Areas of South Coastal British Columbia. Environment Canada, Environmental Protection Branch Pacific and Yukon Region, Regional Program Report, 100 pp 90. Mackintosh CE, Gobas FAPC, Ikonomou M (2002) Distribution of Phthalate Esters in a Marine Food Web, Report Prepared for the American Chemistry Council, Washington, DC, 209 pp 91. Vitali M, Guidotti M, Macilenti G, Cremisini C (1997) Env Int 23:337 92. Parkman H, Remberger M (1995) “Phthalates in Swedish Sediments,” IVL, Swedish Environmental Research Institute, Stockholm, Sweden 93. Japan Environment Agency (JAE) (1999) “Results of measurements of endocrine disrupting chemicals in the environment”, Tokyo, Japan 94. Ministry of Construction (MOC) (1999a) “Fiscal 1999 Fact-Finding Study of Endocrine Disrupting Chemical Substances in the Water Environment (Spring and Summer Studies)” River Bureau, Sewerage and Sewage Purification Department, City Bureau, Ministry of Construction, Japan. November 95. Ministry of Construction (MOC) (1999b) “Fiscal 1999 Fact-Finding Studies Regarding Endocrine Disrupting Chemical Substances in the Water Environment (Spring Study)” River Bureau, Sewerage and Sewage Purification Department, City Bureau, Ministry of Construction, Japan 96. Ministry of Construction (MOC) (1999c) “Fiscal 1999 Results of the Fact-Finding Studies Regarding Endocrine Disrupting Chemical Substances in the Water Environment (Summer Studies)” River Bureau, Sewerage and Sewage Purification Department, City Bureau, Ministry of Construction, Japan 97. Vikelsoe J, Thomsen M, Johansen E, Carlsen L (1999) “Phthalates and Nonylphenols in Soil. A Field Study of Different Soil Profiles.” National Environmental Research Institute, Denmark. 128 pp. NERI Technical Report No. 268
An Assessment of the Potential Environmental Risks Posed by Phthalates
349
98. Vondracek J, Machala M, Minksova K, Blaha L, Murk AJ, Kozubik A, Hofmanova J, Hilscherova K, Ulrich R, Ciganek M, Neca J, Svrckova D, Holoubek I (2001) Environ Toxicol Chem 20:1499 99. Vikelsoe J, Fauser P, Sorensen PB, Carlsen L (2001) “Phthalates and Nonylphenols in Roskilde Fjord. A Field Study and Mathematical Modelling of Transport and Fate in Water and Sediment. The Aquatic Environment.” National Environmental Research Institute, Roskilde. 106 pp. NERI Technical Report No. 339 100. Long J, House W, Parker A, Rae J (1998) Sci Tot Environ 210/211:229 101. Boutrup S, Erichsen PC, Wiggers L, Jensen A (1998) Miljöfremmende stoffer in Århus Amt – fase 2 og 3, 1997–1998. Århus Amt Natur og Miljökontoret, Lyseng Allé, DK-8270 Höjbjerg, Denmark 102. Fries G (1996) Sci Tot Environ 185:93 103. Cousins I, MacKay D (2003) Multimedia Mass Balance Modeling of two Phthalate Esters Using the Regional Population-Based Model (RPm), Chapter submitted to Springer-Verlag Environmental Chemistry Handbook Volume for Phthalate Esters 104. Sverdrup L, Nielsen ET, Krogh PH (2002) Environ Sci Technol 36:2429 105. Alexander M (1995) Environ Sci Technol 29:2713 106. Beck AJ, Johnson DL, Jones KC (1996) Sci Tot Environ 185:125 107. Kraaij RH, Tolls J, Sijm D, Cornelissen G, Heikens A, Belfroid A (2002) Environ Toxicol Chem 21:752 108. Madsen PL, Thyme JB, Moldrup P, Roslev P (1999) Environ Sci Technol 33:2601 109. Peterson DR, Staples CA (2003) Degradation of Phthalate Esters in the Environment, Chapter submitted to Springer-Verlag Environmental Chemistry Handbook Volume for Phthalate Esters 110. Tienpont B, David F, Vanwalleghem F, Sandra P (2001) J Chromatogr A 911:235