Critical Review and Interpretation of Environmental Data for Volatile

Sep 19, 2014 - *Phone: 989-496-5961; fax: 989-496-5956; e-mail: [email protected]. Cite this:Environ. Sci. Technol. 48, 20 ... Ingjerd Sunde Kro...
1 downloads 3 Views 962KB Size
Critical Review pubs.acs.org/est

Critical Review and Interpretation of Environmental Data for Volatile Methylsiloxanes: Partition Properties Shihe Xu,*,† Gary Kozerski,† and Donald Mackay‡ †

Dow Corning Corporation, Health and Environmental Sciences, Auburn, Michigan 48686, United States Centre for Environmental Modeling and Chemistry, Trent University, Peterborough, ON K9J 7B8, Canada



S Supporting Information *

ABSTRACT: Volatile methylsiloxanes (VMS) enter the environment through industrial activities and the use of various consumer products. Reliable measurements of environmental partition properties for these compounds are critical for accurate prediction of their environmental fate, distribution, transport, exposure and potential effects. In this study, the measured partition properties including air/water (KAW), octanol/water (KOW), and octanol/air partitioning coefficients (KOA), soil organic carbon/ water distribution coefficient (KOC), and biological medium/fluid partition coefficients, and their temperature dependence were critically reviewed. Based on these results, organosilicon compounds such as methylsiloxanes are expected to behave differently in the environment compared to conventional hydrophobic environmental contaminants, as a result of their inherent characteristics related to molecular size and capacity for different types of molecular interactions that control partitioning. The differences are critical and need to be taken into consideration in environmental exposure and risk analyses of these compounds.

I. INTRODUCTION Methylsiloxanes are widely used in industrial and commercial applications.1,2 For example, low molar mass cyclic volatile methylsiloxanes (cVMS) such as hexamethylcyclotrisiloxane (D3, where D = Si[(CH3)2]O] and octamethylcyclotetrasiloxane (D4) are used as intermediates for synthesis of highmolecular-weight polydimethylsiloxanes (PDMS). As a result, several cVMS including D3, D4, and their rearrangement products such as decamethylcyclopentasiloxane (D5), dodecamethylcyclohexasiloxane (D6), as well as linear VMS (lVMS) such as hexamethyldisiloxane (L2), octamethyltrisiloxane (L3) and decamethyltetrasiloxane (L4), may be present as impurities in silicone products. In addition, D5 is a major ingredient in personal care products. VMS have relatively high saturated vapor pressures (P°)3,4 and low water solubilities (Sw)5 at ambient temperature and thus readily become airborne once released. Approximately 90% of VMS materials in personal care products are volatilized during use, with the remaining being discharged to wastewater treatment systems.1,6 In the wastewater treatment processes, the majority of VMS will partition to air and sludge with a small fraction (1−2%) released to waters.7,8 Some of the VMS in © 2014 American Chemical Society

sludge is transferred to soil through land application of biosolids. Increasingly, VMS have become the subject of scientific studies and regulatory scrutiny with numerous reports of environmental exposure, effects and risk assessment, published in the last 5 years.9,10 However, for assessing the environmental fate and exposure of chemicals, accurate physicochemical properties are essential. Included in this review are P°, Sw, and partitioning properties such as air/water (KAW), octanol/air (KOA) and octanol/water partition coefficient (KOW), soil organic carbon/water distribution coefficient (KOC) and lipid/ water partition coefficient (KLW). As compiled elsewhere,9−12 a wide range of often inconsistent values have been reported for key properties of the VMS compounds. Close examination of the empirical methods, critical analysis of the partition property data, and selection of a single set of the best values for future use has not been presented. Received: Revised: Accepted: Published: 11748

July 22, 2014 September 12, 2014 September 19, 2014 September 19, 2014 dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

ΔSOW = ΔSAW + ΔSOA

Accordingly, the aims of this work are (1) to critically review the methods, determine the best available values of the physicochemical properties for VMS, and identify areas of future study; (2) to evaluate the effects of those properties on environmental fate; (3) to investigate how those properties may influence the environmental monitoring and toxicity tests of VMS.

II.3. Quantification Methods for Methylsiloxanes in Partition Media. For partition property measurements, analytical difficulties for organosilicons in partition media have long been recognized, largely because of the low water solubility, extreme partition coefficients, ubiquitous background contamination, and chemical reactivity with the partition media.14 For example, VMS are often quantified by gas chromatography (GC)-based techniques in many partition coefficient measurement methods. As major ingredients of personal care products, VMS are ubiquitous in the indoor air of analytical laboratories. In addition, VMS-containing silicone (polydimethylsiloxanes) fluids and elastomers are often used as lubricants, and can be released slowly from components (e.g., tubing and septa) of laboratory equipment. The stationary phases of most of nonpolar to medium polarity capillary GC columns often contain polydimethylsiloxane polymers and copolymers. cVMS can be generated by reaction of water or alcohol vapor with the stationary phase, especially under elevated column oven temperatures.14 II.4. Solvolysis of Organosiloxanes and Organosilanols in Various Partition Media. Methylsiloxanes have variable stability in the environment and may undergo solvolysis in partition media. These processes can result in erroneous partition coefficient determinations if not recognized and addressed in the experimental design. In water, hydrolysis of dissolved VMS has been reported.15 Under controlled laboratory conditions, the hydrolysis of VMS was found to follow pseudo-first order kinetics with rates dependent on pH and temperature.15 The hydrolysis half-life of D4 in water at pH 7 and 25 °C is 3−4 days; an increase or decrease in pH both accelerate the hydrolysis reactions. For other cVMS, the hydrolysis half-life increases as molecular weight increases. D5 has a measured hydrolysis half-life of 66 days, while the value for D6 is estimated to exceed 1 year under these conditions.16 In addition, the apparent Arrhenius activation energy for D4 hydrolysis in water15 is ∼90 kJ mol−1, corresponding to a 50% decrease in half-life for every six-degree Celsius increase in temperature. In the 1-octanol/water partition system, water is usually saturated with 1-octanol and also may contain water-soluble impurities from 1-octanol, such as 1-octanoic acid. Hydrolysis of organosiloxanes in this system appeared more rapid than anticipated17 based on hydrolysis rates in water disclosed by Kozerski et al.15 For example, in an attempt to measure KOW of D4, 14C−D4 with a radiochemical purity >98% was dissolved into a water-saturated 1-octanol. When this D4/1-octanol solution was brought into a container sharing a common headspace with 1-octanol-saturated water phase, the three major hydrolysis products (SI Figure S1) increased rapidly in the water phase.17 This was also the case for D5. While hydrolysis is a major concern for VMS, complex condensation reactions in 1-octanol phase are more prominent for organosilanols such as dimethylsilanediol (DMSD), a hydrolysis product of VMS. For example, two types of condensations may take place in 1-octanol: (1) silanol-alcohol condensation to form an alkoxylsilane, and (2) homocondensation of silanols to form larger oligomeric diols such as tetramethyldisiloxane-α, ω-diol (dimer diol) and hexamethyltrisiloxane-α, ω-diol (trimer diol). Both types of condensations have been observed in partition coefficient measurements such

II. BACKGROUND II.1. Partition Coefficients and Their Concentration and Temperature Dependence. The air−water partition coefficient KAW can be shown to be13 KAW = γWνW P°/RT

(1)

where νW is the molar volume of water; R and T are the ideal gas constant and absolute temperature; γW is the activity coefficient of the VMS in water. Equation 1 can be applied to the saturation condition: KAW,s = CA,s /C W,s = γW,sνW P°/RT

(2)

or at any other lesser concentration: KAW,i = CA,i /C W,i = γW,iνW P°/RT

(3)

where KAW, s and KAW,i are concentration-based air−water partition coefficients; CA, CW, and γW are concentrations in air and water phases and the activity coefficient of the VMS in water at the water solubility limit (s), or below (i). If different values are measured for KAW, s and KAW, i, the cause must be a change in γW. In other words, KAW,i /KAW,s = γW,i /γW,s

(4)

Similarly, the concentration dependence for KOA and KOW can be expressed as K OA,i /K OA,s = γO,i /γO,s

(5)

K OW,i /K OW,s = γO,iγW,s/(γO,sγW,i)

(6)

where γO refers to the activity coefficient in 1-octanol. In the range of 0−40 °C, the linear form of the Clausius− Clapeyron equation is generally adequate to describe the temperature dependence of the partition coefficients, log KXY = AXY + BXY /T

(7)

where KXY is the concentration-based partition coefficient (i.e., KAW, KOA or KOW) for a X-to-Y partition equilibrium at temperature T (K). AXY and BXY are constants related to standard entropy (ΔSXY) and enthalpy (ΔHXY) changes for the transfer of the compound in partitioning processes: ΔHXY = −2.303RBXY

(8)

ΔSXY = 2.303RAXY

(9)

II.2. Internal Consistency of Environmental Partition Coefficients. Based on the above relationships, the following equations are usually assumed to apply: log K OW = log KAW + log K OA

(10)

This is referred to as internal consistency of the partition coefficients. In addition, internal consistency also applies to the entropy and enthalpy changes: ΔHOW = ΔHAW + ΔHOA

(12)

(11) 11749

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

Table 1. Saturated Vapor Pressures of VMS at 25 °C (Log P° in Pa) and Their Temperature Dependence, Where A, B, r2 Are the Intercept, Slope, And Coefficient of Determination for the Relationship between Log P° and Temperature Based on eq 14a

a

VMS

mol. M (g mol−1)

A

B

r2

ΔvapS (kJ mol−1) K−1

ΔvapH (kJ mol−1)

log P°/ (Pa)

L2 L3 L4 D4 D5 D6

162 236 310 296 370 444

10.30 11.05 11.73 11.11 11.87 12.02

−1955 −2484 −2972 −2680 −3135 −3441

0.9999 0.9999 0.9999 0.9955 0.9920 0.9485

0.197 0.212 0.225 0.213 0.227 0.230

37.4 47.6 56.9 51.3 60.0 65.9

3.75 2.72 1.76 2.12 1.36 0.48

The data for lVMS are from Flaningam3 and those for cVMS are from Flaningam3 and Lei et al.4

as KOW, as observed in HPLC chromatograms of the 1-octanol and water phases.17

Δ vapS(kJ mol−1 K−1) = 1.18 × 10−4M + 0.181(r 2 = 0.882, n = 6, p = 0.0200)

III. METHODS FOR DETERMINING PARTITION PROPERTIES III.1. Separate Determination of Partition Coefficients. Traditionally, KOW, KOA, and KAW have been determined separately in separate pairs of partition media using different methodologies as described below. III.1.1. Air/Water. In the literature, there are three common methods for determining KAW:18,19 (1) estimation using P° and Sw of the test compound; (2) direct measurement of the equilibrium concentrations of the test compound in air and water phases; (3) measurement of related change in concentration of the test compound within one phase (e.g., air) as affected by the change of the concentration in another (e.g., water) based on a mass balance assumption. Experimental and predicted values of log KAW for VMS in the literature, based on the above approaches, are generally not consistent with each other as discussed below. III.1.1.1. Method 1: Estimation of KAW using P° and Sw. There are two sets of relevant P° data in the literature. Flaningam3 used a capacitance manometer to measure vapor pressure for a large number of lVMS and cVMS. In a wide temperature range from 5 to 255 °C, these vapor pressures are related to temperature by log P° = a + b/T + cln T + dT e

Measurement of water solubility for KAW estimation is very difficult for VMS because of their low solubilities and low cohesive energy in their liquid form. VMS are liquids at room temperature. Due to their low cohesive energies, the formation of oily microdroplets of VMS in water is common even by mild mechanical disturbance, or as a result of temperature fluctuation. These microdroplets are difficult to detect and remove, and can cause significant overestimation of water solubility.5 In addition, accurate measurement of trace or ultratrace levels of methylsiloxanes in water is extremely challenging due to analytical issues discussed in Section II. 3. Table 2 lists the measured solubility values for the six VMS by Varaprath et al.,5 who prepared droplet-free aqueous Table 2. Standard KAW for VMS at Water Solubility Limits (Log KAW,s) Based on the Saturated Vapor Pressures from Table 1 and Water Solubilities from Varaprath et al. 5 and Comparison with Those in Literature.21

(13)

where Po is in Pascal, and T is the temperature in Kelvin; a, b, c, d and e are all fitted parameters. At temperatures between 0 and 40 °C, eq 13 can be reduced to log P° = A + B /T

VMS

log P° at 23 °C (Pa)

SW at 23 °C (μmol m−3)5

log KAW,s (23 °C)

log KAW,s (25 °C)

log KAW,s by Mazzoni et al.21 (25 °C)

L2 L3 L4 D4 D5 D6

3.70 2.66 1.69 2.06 1.28 0.40

5744 146 21.7 190 46 11.6

2.55 3.11 2.97 2.39 2.23 1.95

2.62 3.15 3.04 2.48 2.38

1.98 2.52 3.12 2.69 2.43

solutions of VMS at 23 °C. Using these water solubility data and log P° values at 23 °C, calculated from the temperature dependence parameters in Table 1, log KAW values at 23 °C were calculated for various compounds (Table 2). These values are slightly different from those reported by Mazzoni et al.21 obtained using the same approach, but with estimated P° values. III.1.1.2. Method 2: Direct Measurement of Test Compound Concentrations in Air and Water Phases. Kochetkov et al.22 and Hamelink et al.23 have reported KAW values for several VMS by equilibrium methods with direct determinations of VMS concentrations in the water and air phases. Those direct measurement methods can be reliable, but only when attainment of equilibrium between the phases is verified, and when the analytical challenges of siloxanes14 are appropriately addressed. In the study by Kochetkov et al.,22 KAW was supposedly determined at VMS concentrations less than their water solubilities. The KAW values measured at 28 °C, however, were less than those calculated from saturated P° and

(14)

where A and B are constants (Table 1). The intercept (A) and slope (B) values of eq 14 reported for L2, L3, and L4 in Table 1 were based on Flaningam’s data. For D4, D5, and D6, vapor pressure data obtained using a GC retention technique20 were also available from Lei et al.4 The values A and B for D4, D5 and D6 were calculated using eq 14 after combining data from both studies.3,4 As expected, the temperature dependence of vapor pressure for VMS increases with the molar mass (M) as indicated by the slopes. For example, the enthalpy change (ΔvapH) calculated from the slopes is related linearly to M: Δ vapH(kJ mol−1) = 0.100M + 22.86 (r 2 = 0.968, n = 6, p = 0.0015)

(16)

(15)

Interestingly, entropy changes (ΔvapS) calculated from the intercepts is also related to M: 11750

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

Sw at 23 °C, by at least 1 order of magnitude for L2, L3, D4 and D5 (see Table 3). The large discrepancies could not be caused

water as discussed in Section II.3. In the study by Kocketkov et al.,22 silicone septa were used in the experiments, and cVMS in water was analyzed by direct injection of the water sample into the GC/FID with a capillary column having a stationary phase composition of 80% polydimethylsiloxane. No mention was made of the analysis of blanks, and the generation of cVMS in such a GC column due to interaction of hot water vapor with silicone material is expected to be very significant. In the study by Hamelink and co-workers, the KAW of D4 was measured under various conditions including in both pure water and water with humic acid.23 The analytical approach used by Hamelink et al. provided confusing results. The addition of Na-humic acid, for example, should decrease the apparent KAW because the dissolved organic matter in water is well-known to increase the total concentration of hydrophobic organic compound in water. However, the opposite was reported − the addition of 100 ppm humic acid to pure water actually increased the apparent log KAW value for D4 by >10-fold. III.1.1.3. Method 3: Indirect Measurement of Test Compound Concentrations in Air Phase only. The equilibrium partitioning in a closed system (EPICS) method used by David et al.24 to obtain the KAW value for D5 listed in Table 3 is an indirect method originally developed by Lincoff and Gossett.25 Briefly, two identical vessels containing different gas/water volume ratios (Vg/Vw) are fortified with the same mass of a test compound. After equilibration, the air from both vessels is sampled and analyzed to determine the concentration ratio r (r = CA1/CA2, where CA1 and CA2 are concentrations, or simply analytical responses, of cVMS in the headspace of Vial 1 and 2, respectively). The dimensionless KAW value is calculated as

Table 3. Measured Log KAW,i Values at Room Temperature for VMS and the Comparison to Their Measured Log KAW,s Values at the Same Temperaturesa VMS L2

L3

L4

D4

D5

D6

a

temp (°C)

method 22

log KAW, i

log KAW, s

γW,i/γW,s

headspace equilibrium vapor entry loop equilibrium22 3-phase equilibrium/HPLC12

27

0.11 0.38

2.68

0.003 0.005

25

2.49

2.61

0.759

headspace equilibrium22 vapor entry loop equilibrium22 3-phase equilibrium/HPLC12

27

2.17 2.08

3.20

0.093 0.076

25

3.07

3.15

0.832

headspace equilibrium22 vapor entry loop equilibrium22 3-phase equilibrium/HPLC12

27

ND 2.84

3.12

25

3.45

3.04

2.570

headspace equilibrium22 vapor entry loop equilibrium22 direct equilibrium23 direct equilibrium with humic acid addition23 3-phase equilibrium/HPLC12 3-phase equilibrium/HPLC17

28

1.36 1.38

2.61

0.056 0.059

20

0.53 1.53

2.26

0.019 0.186

25 22

2.74 2.69

2.48 2.35

1.820 2.19

headspace equilibrium22 vapor entry loop equilibrium22 equilibrium partitioning in closed system (EPICS)24 3-phase equilibrium/HPLC12 3-phase equilibrium/HPLC17

26

1.11 1.08

2.45

0.046 0.043

23

0.74

2.23

0.032

25

3.16 3.13

2.38

6.026 5.62

headspace equilibrium22 vapor entry loop equilibrium22 3-phase equilibrium/HPLC17

26

0.43 0.77

1.95b

0.030 0.066

24

3.01

0.525

KAW = (Vw2 − r Vw1)/(r Vg1 − Vg2)

(17)

The first problem with this method for VMS is the implicit mass balance assumption. That is, the method does not take into consideration any removal mechanisms for test compounds such as adsorption and transformation that occur during equilibration. Hydrolysis of VMS can be significant for long equilibrium times as discussed earlier. Furthermore, possible sorption of VMS by the stir bar, septum, container, air/water interface, etc., all contribute to the error in KAW calculation using eq 17. Next, EPICS is not suitable for compounds with KAW values that deviate considerable from unity because the calculation of KAW is too sensitive to analytical error in quantification of CA1 and CA2 when KAW is either too high (>100, solutes mostly in air) or too low ( 4.28 The

11.48

See eq 2 for the definition of γW,i/γW,s. At 23 °C. b

by the small temperature difference. The authors attributed this discrepancy to the effects of the “activity coefficient” of cVMS in water. However, this explanation is suspect. This is because the calculated γW,i/γW,s ratios have to be between 0.003 and 0.093 to account for their measured log KAW values (Table 3), implying that the effect felt by each solvated VMS molecule as a consequence of the presence of neighboring VMS molecules is greater at lower concentration. This alleged behavior of VMS violates the mass action law. Some discrepancy between log KAW values measured experimentally can be attributed to the deficiencies of the methods employed. As discussed in detail by Varaprath et al.,5 VMS tend to form oily microdroplets in water when they are mixed under mechanical disturbance, such as with stirring and bubbling in Kochetkov’s experiments. The microdroplets can be stable and may contribute to overdetermination of dissolved aqueous phase concentrations. In addition, sample contamination is a major challenge in analysis of trace level cVMS in 11751

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

Table 4. Directly Measured KOW Values at Room Temperature for VMS VMS L2

L3

L4

D4

D5

D6

temp (°C) 25

log KOW 4.20

25

5.20

slow stirring/mcroextraction/GC/MS9 HPLC retention time with alkylbenzenes as calibration standards26 3-phase equilibrium/HPLC12

25 25

6.60 4.80

25

6.79

slow stirring/mcroextraction/GC/MS9 HPLC retention time with alkylbenzenes as calibration standards26 3-phase equilibrium/HPLC12

25 25

8.21 5.40

25

8.14

report by Sible and Miller,27 was also disclosed. Although the equilibration method was slow-stirring, the resulting log KOW value is more than two log-units lower than that for D5 measured using the same equilibration method, but different analytical method from Kozerski et al.9 In Sible and Miller’s study, D5 concentrations were inferred from nonspecific analysis of total radioactivity in each phase by liquid scintillation counting; therefore, the accuracy of this KOW value relied on the assumption that all 14C activity in the aqueous phase represented only 14C−D5; obviously, this ignored the influence of any trace level polar impurities or degradation products (e.g., 14 C-DMSD) present in the radio-labeled D5 test material, or produced via solvolysis during equilibration as discussed in Section II.4. III.1.3. 1-Octanol/Air. In a recent study, log KOA values of VMS (Table 5) were measured using 14C-labeled VMS

slow stirring/Microextraction/GC/MS9 HPLC retention time with alkylbenzenes as calibration standards26 3-phase equilibrium/HPLC17 3-phase equilibrium/HPLC12

25 25

6.49 4.45

Table 5. Directly Measured KOA Values for VMS at Room Temperature Using Two Different Methods

22 25

6.98 6.98

slow stirring/mcroextraction/GC/MS9 HPLC retention time with alkylbenzenes as calibration standards26 shaking flask/liquid scintillation counting27 3-phase equilibrium/HPLC17 3-phase equilibrium/HPLC12

25 25

8.03 5.20

25 25 25

4.67 8.07 8.09

D4

HPLC retention time with alkylbenzenes as calibration standards26 3-phase equilibrium/HPLC17

25

5.86

D5

24

8.87

method HPLC retention time with alkylbenzenes as calibration standards26 3-phase equilibrium/HPLC12

VMS L2 L3 L4

D6

reported log KOW values as listed in Table 4 are much higher than those obtained by other approaches, and closely match those obtained by 3-phase equilibrium method discussed in Section III.2. Bruggeman et al.26 reported another experimental study of log KOW for VMS in the open literature. The measurements were made using an indirect method that relied upon empirical correlation between the measured HPLC retention times or indices on an octadecylsilyl (ODS)-bonded silica column and the known log KOW values for a set of reference compounds, in this case a homologous series of n-alkylbenzenes. Log KOW values of VMS were determined for several VMS,26 which gave much smaller values than those obtained recently.9,12,17 A comprehensive understanding of the methods employed by Bruggeman et al.26 is complicated by the absence of experimental details in the report. However, there are two potential issues with this HPLC retention time correlation method, which render the log KOW values for VMS suspect. First, the combination of mobile phase and octadecylfunctionalized stationary phase used for determination was unlikely a good surrogate for 1-octanol-to-water partitioning.29,30 Second, alkylbenzenes do not resemble VMS compounds structurally and thus should not have been used as calibration standards in determination of KOW for VMS compounds by a retention time based method without validation. For these reasons, the reported KOW values for VMS must be considered dubious. In the UK Environmental Agency’s risk assessment,9 a log KOW of 4.67 for 14C−D5, based on an unpublished technical

method 31

syringe headspace analysis/HPLC 3-phase equilibrium/HPLC12 syringe headspace analysis/HPLC31 3-phase equilibrium/HPLC12 syringe headspace analysis/HPLC31 3-phase equilibrium/HPLC12 syringe headspace analysis/HPLC31 3-phase equilibrium/HPLC17 3-phase equilibrium/HPLC12 syringe headspace analysis/HPLC31 3-phase equilibrium/HPLC17 3-phase equilibrium/HPLC12 syringe headspace analysis/HPLC31 3-phase equilibrium/HPLC17

temp (°C)

log KOA

25 25 25 25 25 25 25 22 25 25 25 25 25 24

2.98 2.89 3.77 3.75 4.64 4.66 4.31 4.29 4.28 4.95 4.94 4.95 5.77 5.86

compounds in an airtight syringe as the equilibration vessel.31 Radio-labeled VMS dissolved in dry 1-octanol were loaded into the system, and the VMS concentrations in 1-octanol and air were analyzed using HPLC equipped with an in-line radiochemical analyzer. Generally, the log KOA values measured at room temperature were almost identical to those determined with the water-saturated 1-octanol by the 3-phase equilibrium method12,17 (Table 5). This suggested that the effect of water content in the 1-octanol on the measured KOA values was very small. The aforementioned studies represent the only empirical values for log KOA currently available for VMS. All methods discussed above are based on separate determination of individual partition coefficients. In this approach, two types of 1-octanol phase and two types of water phase are encountered in the measurements of the three partition properties, due to mutual solubilization of water and 1-octanol: water saturated 1-octanol in KOW measurement vs dry 1-octanol in KOA measurement, and 1-octanol saturated water in KOW measurement vs pure water in KAW measurement. This mutual solubilization effect, plus the cumulative error from separate measurement of each partition coefficient, makes internal consistency difficult to achieve.12,17 As shown in Table 6, the three partition coefficients measured by those approaches do not achieve internal consistency defined in eq 10. The discrepancy of the partition coefficients, δ, defined as the difference between log KOW and the sum of log KAW and log KOA at the same temperature (i.e., δ = log KOW − (log KAW + 11752

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

Table 6. Internal Consistency of Directly Measured Partition Coefficients at Room Temperaturea VMS L2 L3 L4 D4 D5 D6

L2 L3 L4 D4 D5 D6 a

log KAW (T °C) 0.11−0.38 (27)

log KOA (T °C)

log KOW (T °C)

δ

Separate Measurements 2.98 (25) 4.20 (25)

+0.84 to +1.11 2.08−2.17 (27) 3.77 (25) 4.80 − −1.14 to 6.60 (25) +0.75 2.84 (27) 4.64 (25) 5.40 − −2.08 to 8.13 (25) +0.65 0.53−1.38 (20−28) 4.31 (25) 4.45 − −1.24 to 6.49 (25) +1.65 0.74−1.11 (23−26) 4.95 (25) 4.67 − −1.39 to 8.03 (25) +2.34 0.43−0.77 (26) 5.77 (25) 5.86 (25) −0.34 to +0.68 Simultaneous Measurements (3-Phase Equilibrium Method) 2.49 (25) 2.89 (25) 5.20 (25) −0.18 3.07 (25) 3.75 (25) 6.79 (25) −0.03 3.45 (25) 4.66 (25) 8.14 (25) 0.03 2.69 (22), 2.74 (25) 4.29 (22), 6.98 (22, 25) −0.04−0 4.28 (25) 3.13 −3.16 (25) 4.94 − 8.07 − −0.04 to 4.95 (25) 8.09 (25) +0.02 3.01 (24) 5.86 (24) 8.87 (24) 0

Figure 1. Three-phase equilibrium apparatus for simultaneous determination of KOW, KOA, and KAW for volatile and semivolatile compounds like VMS (upper) or water-soluble compounds like dimethylsilanediol (DMSD) (lower). Reproduced from Xu and Kropscott, 201217 with permission.

T = temperature; δ = log KOW − (log KAW + log KOA).

log KOA)), varied up to two log units for the separate methods due to the combination of inaccurate measurements and the mutual solubility effect. III.2. Simultaneous Determination of Partition Coefficients. The 3-phase equilibrium method is a novel method developed17 and evaluated12,17 recently for simultaneous determination of three partition coefficients involving two liquid phases and air. Briefly, a test compound is introduced to a double-syringe apparatus for equilibration in water, 1-octanol and air phases simultaneously (Figure 1), and KAW, KOA, and KOW are obtained from measured concentrations of the test compound in the three phases. The major differences between this method and the separate methods discussed earlier are two uniformities in the 3-phase method that are not found in separate measurement methods. First is the uniformity in partition media: there is only one type of water (1-octanolsaturated), and 1-octanol (water saturated) phase for measuring all three partition coefficients. The second uniformity is use of the same quantitation method for test compounds in all media. These two characteristics are the basis for the self-consistency of the three partition coefficients. For example, at room temperature δ is zero or close to zero by the 3-phase equilibrium method (Table 6). Using the same 3-phase approach, the self-consistency of temperature dependence, as measured by the entropy and enthalpy changes of the partition processes, was also verified for five VMS compounds and trimethylsilanol,12 which will be discussed later.

to those determined with the commercial 1-octanol phase (non-water saturated)31 (Table 5). All of the measured log KAW values by 3-phase equilibrium method12,17 are much higher than those reported in earlier studies22−24 with obvious design flaws that could overestimate the VMS concentration in water as discussed earlier (Table 3). Furthermore, the log KAW values by the 3-phase equilibrium method usually are obtained at concentrations much less than their corresponding Sw (0.1−1% of Sw12,17). For conventional contaminants, such divergence in concentrations from their Sw does not result in large changes in aqueous activity coefficients.32 For VMS, the effect of concentration on partition properties is more complex, as demonstrated by the activity coefficient ratios listed in Table 3. Activity coefficient ratios in Table 3 were calculated using partition coefficients measured at low concentrations by the 3phase equilibrium method and those at the water solubility limits (eq 4). The activity coefficient ratios are less than unity for L2 and L3 due to the smaller measured log KAW values compared to those calculated based on Sw (up to −0.12 log units). The small difference is within the range of measurement errors, and thus the activity coefficient ratio may not be considered different from 1, suggesting that concentration effect for small VMS is similar to that for the common contaminants. For larger VMS such as D6, the activity coefficients increase by up to 12-fold when concentration changes from 100% to 1% of Sw (Table 3). Nevertheless, log KAW values measured by the 3-phase equilibrium method are more consistent with those estimated from P° and Sw, compared with the measurements by other methods. Compared to prominent environmental contaminants such as PCBs and pesticides (Table S1 in Supporting Information (SI)), VMS have large KAW values. At room temperature, log KAW values for six low molecular weight VMS measured by the

IV. ROOM TEMPERATURE PARTITION COEFFICIENTS Compared with all other measured partition coefficients from the literature for VMS, room temperature log KOW values by the 3-phase equilibrium method are consistent only with those values measured using the slow-stirring method coupled with direct analysis9 (Table 4). Values of log KOA at room temperature using 3-phase equilibrium12,17 are almost identical 11753

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

compared to the increase in solubility as temperature increases. This behavior is commonly found for most known environmental contaminants. In fact, the range of ΔHAW values for VMS and are similar to those for alcohols, ketones, carboxylic acids, some polyfluorinated compounds such as fluoro- telomer alcohols (FTOHs), and some polychlorinated compounds such as PCBs.12 The ΔHAW values for lVMS are similar to their ΔvapH values (within ±10 kJ mol−1), although the cVMS show greater differences that imply higher negative excess enthalpies of dissolution in water. Excluding L2, these values are more negative for the higher molar mass oligomers within each group. This result is consistent with the additional polar interactions that are available in the aqueous phase due to the hydrogen bond acceptor capacity of VMS, similar to the other classes of monopolar (i.e., H-bond donor or acceptor) compounds listed above. For 1-octanol/air partitioning, VMS have negative ΔHOA values proportional to their molar masses.31,12 The negative ΔHOA can be regarded as a result of the greater temperature sensitivity of volatilization of pure liquid VMS, compared with dissolution in 1-octanol. This behavior is commonly found for other known environmental contaminants such as polyfluorinated compounds (PFCs) and polychlorinated compounds such as PCBs and PCNs.12 However, ΔHOA values of VMS are much smaller than those of most polychlorinated compounds, indicating that 1-octanol/air partitioning of VMS is less sensitive to temperature variation than other environmental pollutants. For any given organosilicon compounds, KOW is less sensitive to temperature change than KAW and KOA (Figure 2). Similar to other hydrogen bonding acceptors such as benzoates, phthalates, and some phenolic compounds, VMS have positive ΔHOW values for octanol/water partitioning, suggesting that increasing temperature could increase dissolution in 1-octanol more than in water. This is consistent with the preceding discussions of dissolution enthalpies. Two additional points are noteworthy regarding those enthalpy and entropy changes. The first point is related to internal consistency of the changes for each compound. As shown in Table 7, small discrepancies in thermodynamic consistency, δ(ΔH) [defined by δ(ΔH) = ΔHOW − (ΔHAW + ΔHOA)] and δ(ΔS) [defined by δ(ΔS) = ΔSOW − (ΔSAW + ΔSOA)], do exist for VMS. In the environmentally relevant temperature range, those discrepancies can result in differences of the partition coefficients in the range from −0.25 to +0.15 log units. Practically, those discrepancies are relatively small. The second point concerns the enthalpy−entropy compensation effect observed for VMS. This effect is manifested as a linear relationship between the enthalpy and entropy changes of chemical and biological processes, often observed for groups of similar compounds.38−40 For VMS, enthalpy and entropy changes of all partitioning processes, including vaporization of the pure liquid, are positively and linearly related.12 The exact mechanism for such enthalpy−entropy compensation is still a controversial subject. However, enthalpy−entropy relations for VMS in air/water partitioning are similar to those of other highly hydrophobic compounds such as n-alkanes and Freons.12 On the other hand, trimethylsilanol, a polar degradation product of lVMS, behaves more like alcohols or ketones.12 However, VMS behave more like polyfluorinated compounds in octanol/air partitioning, and more like benzoates and phenolic compounds (net hydrogen bonding acceptor) in octanol/water partitioning.12

3-phase equilibrium method (Table 6) are all greater than 2 at room temperature. These KAW values are similar to those for nalkanes,13,33−35 but several orders of magnitude higher than those for the common environmental contaminants with comparable molar masses such as PCB-153,13 hexachlorodibenzodioxin (HexaCDD), 1 3 di(ethylhexyl)phthalate (DEHP),13 and several pesticides (SI Table S1). In addition, most “volatile organic compounds” such as monoaromatics and trihalomethanes have log KAW values in the range −1 to 0.3.36 Oxygen has a log KAW value of 1.48.37 In comparison, the VMS have KAW values 1−2 orders of magnitude greater that must be reflected in profound differences in their environmental fate, transport and distribution.

V. TEMPERATURE DEPENDENCE Standard partition coefficients at 25 °C and 1 atm pressure can be used to reveal the differences in intrinsic solute properties of chemicals, both within the category of organosilicon compounds and over a broad range of non-Si-bearing environmental contaminants. However, in the real environment with varying temperatures, the temperature dependence of the partition coefficients is also an important consideration. In Figure 2, logarithmic values of the three partition coefficients by 3-phase methods12,17 were plotted together

Figure 2. Effect of equilibrium temperature (T in Kelvin) on three partition coefficients KOW, KAW, and KOA for selected volatile methylsiloxanes (L2, L3, L4, and D4) simultaneously measured by 3phase equilibrium method (blue)12, and KOA measured by syringe method (black)31.

with the log KOA values by the syringe method31 for selected VMS. As shown in Figure 2, temperature does have substantial effects on the partition coefficients of VMS, especially log KAW and log KOA values. In the narrow range (0−40 °C) tested, the logarithm of all three partition coefficients for any given compounds are linearly related to the reciprocal of the absolute temperature (eq 7). The enthalpy and entropy changes calculated based on eq 8 and 9 are listed in Table 7. For air/water partitioning, VMS have positive ΔHAW values (Table 7) that indicate a greater increase in volatility of VMS 11754

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

Table 7. Enthalpy and Entropy Changes, And Their Standard Deviations in the Parentheses, For Three Partitioning Processes and Their Overall Consistencya comp

octanol/water ΔHOW

L2 L3 L4 D4 D5 a

19.4 1.0 11.3 31.9 68.8

(6.4) (5.1) (1.3) (4.0)

octanol/air

ΔSOW 0.164 0.133 0.194 0.241 0.386

(0.022) (0.018) (0.004) (0.014)

air/water

ΔHOA −26.7 −37.5 −46.8 −43.8 −47.3

(4.2) (2.3) (3.0) (2.6) (2.7)

ΔS −0.034 −0.054 −0.068 −0.065 −0.064

ΔHAW

OA

(0.015) (0.008) (0.011) (0.009) (0.009)

53.0 39.5 65.5 73.9 123.9

overall consistency ΔS

(2.0) (5.5) (1.4) (6.7)

0.225 0.191 0.286 0.300 0.476

AW

(0.007) (0.019) (0.005) (0.023)

δ(ΔH)

δ(ΔS)

−6.9 −1.0 −7.4 1.8 −7.8

−0.027 −0.004 −0.024 0.005 −0.027

No temperature dependence for D6 has been determined.

Table 8. Comparison of Measured and Estimated Log KOC Values and Their Log KOC/KOW Ratios term contributions, eq 19 compound D4 D5 L3 L4 chlorobenzene 1,2,3,4tetrachlorobenzene tetracene PCB-153

measured log KOC (reference)

estimated log KOC (log KOW) (eq 18 with b = 0.4145)

vV

eE + sS

bB

4.2242 5.1742 4.3442 5.1642 2.2747,50 3.8447,50

6.59 7.69 6.40 7.75 2.45 4.25

(6.98) (8.08) (6.79) (8.14) (2.84) (4.64)

−2.64 −3.29 −2.48 −3.11 −0.53 −0.82

−3.02 −3.78 −2.70 −3.45 −1.08 −1.56

−0.16 −0.22 −0.16 −0.27 0.56 0.82

0.64 0.80 0.46 0.69 0.09 0.00

5.8147,50 6.4047

5.51 (5.90) 6.76 (7.15)

−0.37 −1.06

−2.35 −2.66

1.67 1.55

0.40 0.14

hydrophobic and polar organic compounds.47 Using the ppLFER model, log KOC values were calculated for various VMS,48−50 which matched the measured values better, although the original model was not trained with any Si-bearing compounds. Based on pp-LFERs of this kind for both KOC and KOW, the following pp-LFER equation was obtained:

VI. SOIL/WATER DISTRIBUTION COEFFICIENTS Values of KOC for VMS were measured using various approaches. In a laboratory study, a batch equilibrium method based on OECD test guideline 10641 was employed.42 The measured log KOC values as listed in Table 8 demonstrated a general trend of increasing log KOC as the molar mass of VMS increases, similar to that of log KOW. In addition to the wellcontrolled laboratory study cited above, KOC values were also measured in isolated soil organic matter/water systems,43 and river water/sediment systems.44 Generally speaking, those values are consistent with each other, with variations possibly arising from different sources of the organic matter or other aspects of the study designs. Overall, the measured log KOC values of any given VMS are much lower than their corresponding log KOW values.42 1-Octanol is used as a surrogate for organic-rich environmental phases such as soil organic matter and lipids from biota, and KOW is often used to represent KOC in soil/water partitioning and KLW for biota/water partitioning.45 For example, in environmental modeling, values of KOC for nonionic compounds are often estimated from their KOW values using a single parameter quantitative structure−property relationship (QSPR) of the form,45,46 K OC = b·K OW

estimated log KOC/KOW, eq 19

log(K OC/K OW ) = − 1.29V + 0.22E + 0.65S + 1.20B − 0.11

(19)

where V is McGowan’s characteristic volume, which represents molecular size; E is the excess molar refraction describing the polarizability of a molecule, and S primarily describes solute dipolarity, but it also contains a polarizability component. The B descriptor refers to solute hydrogen bond basicity. The coefficients represent the complementary properties of the system29,30 − in this case, the difference in solvation characteristics of soil organic matter relative to 1-octanol. Specifically, the negative coefficient (−1.29) in the V term indicates that wet organic matter, a 3-dimensional polymer network, is a more cohesive “solvent” relative to wet 1-octanol, a free expandable fluid under the test temperature. The positive coefficients for E (+0.22) and S terms (+0.65) indicate that wet organic matter engages solutes more effectively through interactions involving permanent or induced dipoles. Wet organic matter is also a stronger hydrogen bond donor than wet 1-octanol (positive B term). From eq 19, it is also clear that the value of log (KOC/KOW) can only be constant as that described in eq 18 for a series of compounds when variation of the negative V-term is offset by summation of all variation in the E, S, and B (interaction) terms. This is the case for most of the traditional hydrophobic organic compounds used to develop the simple KOC−KOW correlations (Table 8). For VMS, the reduction in free energy of interaction from the combined weaker van der Waals and hydrogen bonding interactions between VMS and organic

(18)

Representative values for b are 0.4145 and 0.3546 for traditional hydrophobic contaminants. Compared with the predicted values using eq 18, the measured KOC values for VMS are lower by two or more orders of magnitude, suggesting that the KOW-based QSPR model fails to describe the soil/water partitioning process for these compounds. Among non-Si-bearing organic compounds, KOW-based QSPRs have been found to be poor predictive models for polar organic compounds.47 In this case, a poly-parameter linear free energy relationship (pp-LFER) model can quantitatively describe all KOC values of non-Si-bearing 11755

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

consider how differences in lipid types might influence VMS partitioning into, and distribution within, biota. Endo and coworkers have quantified the differences in solvation characteristics of membrane lipids and storage lipids.64 Seston et al.65 demonstrated that these differences in lipid properties could lead to substantially larger errors in lipid−water partitioning of cVMS, estimated from KOW, when membrane lipids are present. Specifically, the calculated value of the hypothetical lipidoctanol partition coefficient (KLO) ranged from −1.2 (D4) to −1.9 (D 6 ) log units for membrane lipid, while the corresponding values for storage lipid were +0.1 to 0.2 log units.65 As with KOC, this type of analysis again demonstrated that 1-octanol may not be a good surrogate for all lipids in assessing biota/air or biota/water partitioning of VMS. Recently, Geisler and co-workers demonstrated high similarity between olive oil and various animal and plantderived oils and fats, including fish oil, as partitioning media for diverse organic chemicals.66 Given the good agreement between the measured olive oil/air and fat/air partition coefficients for L2 and D4, and general concurrence with the model estimates, these values seem appropriate for use in estimating partitioning of these VMS into storage lipids, and the calculations of Seston et al. suggest that octanol is a reasonable surrogate for storage lipids for VMS.65 However, further work on D5 seems warranted given the greater discrepancies between empirical values and the model estimates, and more studies are needed to verify the correct partition coefficients of all VMS for other relevant tissues or fluids where storage lipids may not be the dominant sorbing phase.

matter (summation of E, S, and B terms) fails to compensate for the energy penalty associated with their large molecular size.42 As a result, the KOW-based simple model does not describe accurately the soil/water partitioning of VMS.

VII. BIOLOGICAL MEDIUM/FLUID PARTITION COEFFICIENTS Partitioning into lipids or lipid-containing biological components is an important process influencing the bioaccumulation and toxicity of lipophilic organic chemicals. Partitioning between air and olive oil (a surrogate lipid) were reported in several studies describing the development of pharmacokinetic models for methylsiloxanes (L2, D4, and D5) in rats, and in vitro measurements of partitioning to fat, blood, and other tissues51−55 were also made; these data can be used to calculate other relevant parameters such as tissue/blood or tissue/water partition coefficient. The experiments for determining those partition coefficients involved static equilibration between siloxane vapor and various fluid or tissue samples contained in separate open containers within a sealed chamber. Concentration measurements were made by GC for both the air and fluid/tissue samples. However, evaluation of the quality of those data is difficult because only the final partition coefficients were reported, and few details were given regarding analytical quality control or evidence that true equilibrium was attained. The data, shown in Figure 3, were compared to

VIII. ENVIRONMENTAL IMPLICATIONS VIII.1. Risks of Using Single-Parameter QSPRs Outside Domain. Better to Use pp-LFERs. In environmental modeling, KOW is often used to estimate both KOC and KLW using single-parameter QSPRs. Those QSPRs are based on the assumption that 1-octanol is a valid surrogate for soil organic matter and biological lipids. The differences between soil organic matter or membrane lipids and 1-octanol have been clearly demonstrated using pp-LFER models. The effects of such differences on the accuracy of single-parameter QSPRs vary, mainly depending on the properties of the compounds used to derive the models compared to those to which they are applied. For this reason, the application of single-parameter QSPRs is only valid within the group of compounds defined by the training set. pp-LFERs make no such assumptions regarding the similarity of the known partition coefficient (e.g., KOW) and the coefficients being predicted (e.g., KOC and KLW), and can account for inherent differences in the partition media. Thus, when adequately trained, LFERs have wider domains of applicability, which can include VMS.42 The same conclusion has been reached by other researchers for different organic compounds.64,67 VIII.2. Exposure Limits in Soil and Sediments. In the evaluation of the ecological effects of VMS in soil and sediment, the maximum sorption capacity (MSC) at the solubility limit of the soil solution or pore water of the sediment is very useful in interpreting test results. For example, in a study on ecotoxicity of D5 in soil,68 nominal test concentrations for D5 were from 0 to 10 000 mg kg−1 dw. The test material D5 was delivered to the soil by spiking the biosolids with neat D5 and then mixing the spiked biosolids into the soil (“Gr10”) at the rate of 5 g kg−1. What is the form of D5 in the test soil: sorbed or neat D5?

Figure 3. Comparison of measured in vitro51−55 (bars) and LFER estimated 48,49,56−62 (+) partition coefficients for three VMS compounds. For D4, some bars represent the average of 2, and in one case 3, reported values.52−54 Bars represent in order from left to right: olive oil, perirenal fat, liver, kidney, muscle, lung, and blood. Note lack of measured values for lung-air partition coefficient of L2 and D5.

predicted values from pp-LFER models for these systems56−62 using no VMS data for model training. In general, the best agreement between measured and predicted values was observed for L2, with D4 and D5 showing increasingly greater discrepancies. That the measured D5 values are always less than the measured D4 values, even for the olive oil/air system, suggests that the systems might not have reached equilibrium. The poorer agreement for the more water-rich media (blood, liver and muscle) might indicate that chemical transformations of the VMS solutes also influenced the determinations. Biological fluids and tissues such as blood, liver and muscle, have a much greater ratio of phospholipids to neutral storage lipids in their compositions.63 Therefore, it is necessary to 11756

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

temperate zone.73,74 For the same period in the Arctic atmosphere, the average airborne D5 was 0.5 ng m−3,75 representing a several hundred-fold decrease compared with that in the source region, and a 40-fold decrease compared to that in the rural region in the N-temperate zone. Although the airborne D5 concentrations in Arctic air can increase in the winter time to an average near 4 ng m−3,75,76 the overall fraction of the released D5 transport into the remote atmosphere is still small (∼1%).71 As shown in Example S2 of the SI, the soil concentration of D5 in the Arctic from longrange transport under these conditions is estimated to be 1.1 × 10−5 and 1.4 × 10−3 μg kg−1 at 25 and 0 °C, respectively, lower than the lowest detection limit (7 μg kg−1) of D5 in soil8 by 3 to 5 orders of magnitude. For that reason, any D5 detected in terrestrial media in remote regions based on current analytical methods likely reflects the influence of local sources rather than the long-range transport.

To answer this question, the MSC of the soil can be estimated as

MSC = S W fOC K OC

(20)

where SW is the aqueous solubility of D5, and f OC is the mass fraction of OC in soil or sediment. For the specific “Gr10” soil used in the aforementioned ecotoxicity study,68 MSC is estimated to be 34 mg kg−1 based on eq 20 (see S1 in SI). In other words, the sorbed D5 accounted for only 19% of the total mass of D5 in the soil at the lowest non-zero dose level, 14% of the total mass of D5 in the soil at the reported IC50 for barley root, 4% at the IC50 for Springtail juvenile production, and 6.5 and log KAW < −1.72 The smaller molecule weight VMS such as L2, L3, L4, L5, D3, D4, D5, and D6 have room temperature log KOA < 6.5 and KAW ≫ −1 (by about 4 orders of magnitude). As predicted in a modeling assessment,71 a fraction of emitted cVMS can travel a long distance in the atmosphere, but their deposition potential from remote air to surface media is small because of their small log KOA and large log KAW values. This can be readily demonstrated with environmental monitoring data for different regions. In recent studies, the average airborne D5 concentration was reported in the range of 162−230, 52, and 18 ng m−3 during summer in urban, suburban, and rural atmospheres in the N-



ASSOCIATED CONTENT

S Supporting Information *

Additional graph on analytical complication, table of KAW values of conventional contaminants, calculations of maximum sorption capacity and expected concentrations of VMS in remote surface media. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: 989-496-5961; fax: 989-496-5956; e-mail: shihe.xu@ dowcorning.com. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Allen, R. B.; Kochs, P.; Chandra, G. Organosilicon materials. In The Handbook of Environmental Chemistry, Part H; Chandra, G Ed.; Springer-Verlag: Berlin, 1997; Vol 3. pp 1−15. (2) Horii, Y.; Kannan, K. Survey of organosilicone compounds,Including cyclic and linear siloxanes, in personal-care and household products. Arch. Environ. Contam. Toxicol. 2008, 55, 701−710. (3) Flaningam, O. L. Vapor pressure of poly(dimethylsiloxane) oligomers. J. Chem. Eng. Data 1996, 31, 266−272. (4) Lei, Y. D.; Wania, F.; Mathers, D. Temperature-dependent vapor pressure of selected cyclic and linear polydimethylsiloxane oligomers. J. Chem. Eng. Data 2010, 55 (12), 5868−5873. (5) Varaprath, S.; Frye, C. L.; Hamelink, J. Aqueous solubility of permethylsiloxanes (silicones). Environ. Toxicol. Chem. 1996, 15, 1263−1265. (6) Montemayor, B. P.; Price, B. B.; van Egmond, R. A. Evaporative fate of cyclopentasiloxane (D5) from personal care products during product use: Antiperspirants, skin care products and hair care products. Chemosphere 2013, 93, 735−740. (7) Parker, W. J.; Shi, J. C.; Fendinger, N. J.; Monteith, H. D.; Chandra, G. Pilot plant study to assess the fate of two volatile methyl siloxane compounds during municipal wastewater treatment. Environ. Chem. Toxicol. 1999, 18, 172−181. (8) Wang, D.; Steer, H.; Tait, T.; Williams, Z.; Pacepavicius, G.; Young, T.; Ng, T.; Smyth, S. A.; Kinsman, L.; Alaee, M. Concentrations of cyclic volatile methylsiloxanes in biosolid amended soil, influent, effluent, receiving water, and sediment of wastewater treatment plants in Canada. Chemosphere 2013, 93, 766−773. (9) Brook, D. N.; Crookes, M. J.; Gray. D.; Robertson, S. Environmental Risk Assessment Report: Decamethylcyclopentasiloxane; Environment Agency of England and Wales: Bristol, 2009; http://

11757

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

chemicals with the “slow-stirring” method. Environ. Toxicol. Chem. 1989, 8, 499−512. (29) Poole, S. K.; Poole, C. F. Separation methods for estimating octanol-water partition coefficients. J. Chromatogr., B 2003, 797, 3−19. (30) Poole, C. F.; Ariyasena, T. C.; Lenca, N. Estimation of the environmental properties of compounds from chromatographic measurements and the solvation parameter model. J. Chromatogr., A 2013, 1317, 85−104. (31) Xu, S.; Kropscott, B. Octanol-air partition coefficients of volatile methylsiloxanes and their temperature dependence. J. Chem. Eng. Data 2013, 58, 136−142. (32) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental Organic Chemistry, 2ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2003. (33) Jonsson, J. A.; Vejrosta, J.; Novak, J. Air/water partition coefficients for normal alkanes (n-pentane to n-nonane). Fluid Phase Equilib. 1982, 9, 279−286. (34) Bell, G. H. Solubilities of normal aliphatic acids, alcohols and alkanes in water. Chem. Phys. Lipids 1973, 10, 1−10. Poole, S. K.; Poole, C. F. Chromatographic models for the sorption of neutral organic compounds by soil from water and air. J. Chromatogr., A 1999, 845, 381−400. (35) Razzouk, A.; Naccoul, R. A.; Mokbel, I.; Saab, J.; Jose, J. Vapor and sublimation pressures of three normal alkanes: C20, C24, and C28. J. Chem. Eng. Data 2009, 54, 1214−1219. (36) Staudinger, J.; Roberts, P. V. A critical compilation of Henry’s law constants temperature dependence relations for organic compounds in dilute aqueous solutions. Chemosphere 2001, 44, 561−576. (37) Sander, R. Compilation of Henry’s Law Constants for Inorganic and Organic Species of Potential Importance in Environmental Chemistry, 1999. http://www.mpch-mainz.mpg.de/∼sander/res/ henry.html. (38) Leffler, J. E.; Grunwald, E. Rates and Equilibria of Organic Reactions; Wiley-Intersceince: New York, 1963. (39) Freed, K. F. Entropy-enthalpy compensation in chemical reactions and adsorption: An exactly solvable model. J. Phys. Chem. B 2011, 115, 1689−1692. (40) Perez-Benito, J. F. Some tentative explanations for the enthalpy−entropy compensation effect in chemical kinetics: From experimental errors to the Hinshelwood-like model. Monatsh. Chem. 2013, 144, 49−58. (41) OECD. OECD Guideline for Testing Chemicals No. 106: Adsorption-desorption using a batch equilibrium method. 2000. http://www.epa.gov/scipoly/sap/meetings/2008/october/106_ adsorption_desorption_using.pdf. (42) Kozerski, G.; Xu, S.; Miller, J.; Durham, J. Determination of soil organic carbon-water sorption coefficients for volatile methylsiloxanes. Environ. Toxicol. Chem. 2014, DOI: 10.1002/etc.2640. (43) Whelan, M. J.; Sanders, D.; van Egmond, R. Effect of Aldrich humic acid on water-atmosphere transfer of decamethylcyclopentasiloxane. Chemosphere 2009, 74, 111−1116. (44) van Egmond, R.; Sanders, D. Investigations into the effect of aging on the bioavailability of decamethylcyclopentasiloxane in artificial and natural sediments, Poster Presentation at SETAC Europe 20th Annual Meeting, Seville, Spain. 23−27 May 2010. (45) Karickhoff, S. W. Organic pollutant sorption in aquatic systems. J. Hydraul. Eng., ASCE 1984, 110, 707−735. (46) Seth, R.; Mackay, D.; Muncke, J. Estimating the organic carbon partition coefficient and its variability for hydrophobic chemicals. Environ. Sci. Technol. 1999, 33, 2390−2394. (47) Nguyen, T. H.; Goss, K.-U.; Ball, W. P. Polyparameter linear free energy relationships for estimating the equilibrium partition of organic compounds between water and the natural organic matter in soils and sediments. Environ. Sci. Technol. 2005, 39, 913−924. (48) Ahmed, H.; Poole, C. F.; Kozerski, G. E. Determination of descriptors for organosilicon compounds by gas chromatography and non-aqueous liquid-liquid partitioning. J. Chromatogr., A 2007, 1169, 179−192.

publications.environment-agency.gov.uk/PDF/SCHO0309BPQX-E-E. pdf. (10) Environment Canada, Health Canada. Screening assessment for the challenge decamethylcyclopentasiloxane (D5). 2008. http://www. ec.gc.ca/substances/ese/eng/challenge/batch2/batch2-541-02-6_en. pdf. (11) Wang, D. G.; Norwood, W.; Alaee, M.; Byer, J. D.; Brimble, S. Review of recent advances in research on the toxicity, detection, occurrence and fate of cyclic volatile methyl siloxanes in the environment. Chemospehere 2013, 93, 711−725. (12) Xu, S.; Kropscott, B. Evaluation of the three-phase equilibrium method for measuring temperature dependence of internally consistent partition coefficients (KOW, KOA and KAW) for volatile methylsiloxanes and trimethylsilanol. Environ. Toxicol. Chem. 2014, accepted. (13) Mackay, D. Multimedia Environmental Models: The Fugacity Approach, 2nd ed. ; CRC Press, Taylor & Francis Group: Boca Raton, FL, 2001. (14) Varaprath, S.; Stutts, D. H.; Kozerski, G. E. A primer on the analytical aspects of silicones at trace levels-challenges and artifactsA review. Silicon Chem. 2006, 3, 79−102. (15) Kozerski, G. E.; Durham, J. A. Kinetics and mechanism of the abiotic hydrolysis reactions of permethylated cyclosiloxanes in aqueous solution. In SETAC Northern America 27th Annual Meeting, Montreal, Quebec, Canada, November 5−9, 2006. (16) Europen Chemicals Agency. Dodecamethylcyclohexasiloxanes. REACh dossier. http://apps.echa.europa.eu/registered/data/dossiers/ DISS-9875b402-454f-57f6-e044-00144f67d031/DISS-9875b402-454f57f6-e044-00144f67d031_DISS-9875b402-454f-57f6-e04400144f67d031.html. (17) Xu, S.; Kropscott, B. Method for simultaneous determination of partition coefficients for cyclic volatile methylsiloxanes and dimethylsilanediol. Anal. Chem. 2012, 84, 1948−1955. (18) Mackay, D.; Shiu, W. Y. A critical review of Henry’s law constants for chemicals of environmental interests. J. Phys. Chem. Ref. Data 1981, 10, 1175−1199. (19) Gossett, J. M. Measurement of Henry’s law constants for C1 and C2 chlorinated hydrocarbons. Environ. Sci. Technol. 1987, 21 (2), 202−208. (20) Bidleman, T. F. Estimation of vapor pressure for nonpolar organic compounds by capillary gas chromatography. Anal. Chem. 1984, 56, 2490−2496. (21) Mazzoni, S. M.; Roy, S.; Grigoras, S. Eco-relevant properties of selected organosilicon materials. In Chandra, G, Ed.; The Handbook of Environmental Chemistry, Part H; Springer-Verlag: Berlin, 1997; Vol. 3, pp 53−81. (22) Kochetkov, A.; Smith, J. S.; Ravikrishna, R.; Valsaraj, K. T.; Thibodeaux, L. J. Air-water partition constants for volatile methyl siloxanes. Environ. Toxicol. Chem. 2001, 20 (10), 2184−2188. (23) Hamelink, J. L.; Simon, P. B.; Silberhorn, E. M. Henry’s law constant, volatilization rate and aquatic half-life of octamethylcylotetrasiloxanes. Environ. Sci. Technol. 1996, 30, 1946−1952. (24) David, M.; Fendinger, N.; Hand, V. C. Determination of Henry’s law constants for organosilicones in actual and simulated wastewater. Environ. Sci. Technol. 2000, 34, 4554−4559. (25) Lincoff, A. H.; Gossett, J. M. Gas Transfer at Water Surfaces; Brutsaert, W.; Jirka, G. H., ed.; Redel: Dordrecht, Holland, 1984; pp17−25. (26) Bruggeman, W. A.; Opperhuizen, A.; Wijbenga, A.; Hutzinger, O. Bioaccumulation of super-lipophilic compounds in fish. Toxicol. Environ. Chem. 1984, 7, 173−189. (27) Sible, V. S. & Miller, J. R. Determination of n-Octanol/Water Partition Coefficient of 14C-Decamethylcyclopentasiloxane (14C-D5) by Liquid Scintillation Counting, Final Study Report to the Silicones Environmental, Health and Safety Council (SEHSC); Dow Corning Corporation: Midland, MI, 2006. (28) De Bruijn, J.; Busser, F.; Seinen, W.; Hermens, J. Determination of octanol/water partition coefficients for hydrophobic organic 11758

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759

Environmental Science & Technology

Critical Review

(49) Atapattu, S. N.; Poole, C. F. Determination of descriptors for semivolatile organosilicon compounds by gas chromatography and non-aqueous liquid-liquid partition. J. Chromatogr., A 2009, 1216, 7882−7888. (50) Poole, S. K.; Poole, C. F. Chromatographic models for the sorption of neutral organic compounds by soil from water and air. J. Chromatogr., B 1999, 845, 381−400. (51) Dobrev, I. D.; Reddy, M. B.; Plotzke, K. P.; Varaprath, S.; McNett, D. A.; Durham, J.; Andersen, M. E. Closed-chamber inhalation pharmacokinetic studies with hexamethyldisiloxane in the rat. Inhal. Toxicol. 2003, 15, 589−617. (52) Anderson, M. E.; Sarangapani, R.; Reitz, R. H.; Gallavan, R. H.; Dobrev, I. D.; Plotzke, K. P. Physiological modeling reveals novel pharmacokinetic behavior for inhaled octamethylcyclotetrasiloxane in rats. Toxicol. Sci. 2001, 60, 214−231. (53) Dobrev, I. D.; Nong, A.; Liao, K. H.; Reddy, M. B.; Plotzke, K. P.; Andersen, M. E. Assessing kinetic determinants for metabolism and oral uptake of octamethylcyclotetrasiloxane (D4) from inhalation chamber studies. Inhal. Toxicol. 2008, 20, 589−617. (54) Thrall, K. D.; Soelberg, J. J.; Powell, T.; Corley, R. A. Physiologically based pharmacokinetic modeling of the disposition of octamethylcyclotetrasiloxane (D4) migration from implants to humans. J. Long Term Eff. Med. Implants 2008, 18, 133−144. (55) Reddy, M. B.; Dobrev, I. D.; McNett, D. A.; Tobin, J. T.; Utell, M. J.; Morrow, P. E.; Domoradzki, J. Y.; Plotzke, K. P.; Andersen, M. E. Inhalation dosimetry modeling with decamethylcyclopentasiloxane in rats and humans. Toxicol. Sci. 2008, 105, 275−285. (56) Abraham, M. H.; Ibrahim, A. Gas to olive oil partition coefficients: A linear free energy analysis. J. Chem. Inf. Model. 2006, 46, 1735−1741. (57) Abraham, M. H.; Ibrahim, A. Air to fat and blood to fat distribution of volatile organic compounds and drugs: Linear free energy analysis. Eur. J. Med. Chem. 2006, 41, 1430−1438. (58) Abraham, M. H.; Ibrahim, A.; Acree, W. E., Jr. Air to muscle and blood/plasma to muscle distribution of volatile organic compounds and drugs: Linear free energy analysis. Chem. Res. Toxicol. 2006, 19, 801−808. (59) Abraham, M. H.; Ibrahim, A.; Acree, W. E., Jr. Air to liver partition coefficients for volatile organic compounds and blood to liver partition coefficients for volatile organic compounds and drugs. Eur. J. Med. Chem. 2007, 42, 743−751. (60) Abraham, M. H.; Ibrahim, A.; Acree, W. E., Jr. Air to lung partition coefficients for volatile organic compounds and blood to lung partition coefficients for volatile organic compounds and drugs. Eur. J. Med. Chem. 2008, 43, 478−485. (61) Abraham, M. H.; Weathersby, P. K. Hydrogen Bonding. 30. Solubility of gases and vapors in biological liquids and tissues. J. Pharm. Sci. 1994, 83, 1450−1456. (62) Abraham, M. H.; Ibrahim, A.; Acree, W. E., Jr. Air to blood distribution of volatile organic compounds: A linear free energy analysis. Chem. Res. Toxicol. 2005, 18, 904−911. (63) Endo, S.; Brown, T. N.; Goss, K. U. General model for estimating partition coefficients to organisms and their tissues using the biological compositions and polyparameter linear free energy relationships. Environ. Sci. Technol. 2013, 47, 6630−6639. (64) Endo, S.; Escher, B. I.; Goss, K. U. Capacities of membrane lipids to accumulate neutral organic chemicals. Environ. Sci. Technol. 2011, 45 (14), 5912−5921. (65) Seston, R. M.; Powell, D. E.; Woodburn, K. B.; Kozerski, G. E.; Bradley, P. W.; Zwiernik, M. J. Importance of lipid analysis and its implications for metrics of bioaccumulation. Integr. Environ. Assess. Manage. 2014, 10, 142−144. (66) Geisler, A.; Endo, S.; Goss, K. U. Partitioning of organic chemicals to storage lipids: Elucidating the dependence on fatty acid composition and temperature. Environ. Sci. Technol. 2012, 46 (17), 9519−9524. (67) Goss, K. U.; Schwarzenbach, R. P. Linear free energy relationships used to evaluate equilibrium partitioning of organic compounds. Environ. Sci. Technol. 2001, 35, 1−9.

(68) Velicogna, J.; Ritchie, E.; Princz, J.; Lessard, M.-V.; Scroggins, R. Ecotoxicity of siloxane D5 in soil. Chemosphere 2012, 87, 77−83. (69) Weyman, G. S.; Rufli, H.; Weltje, L.; Salinas, E. R.; Hamitou, M. Aquatic toxicity tests with substances that are poorly soluble in water and consequences for environmental risk assessment. Environ. Toxicol. Chem. 2012, 31, 1662−1669. (70) Xu, S.; Chandra, G. Fate of cyclic methylsiloxanes in soils. 2. Rates of degradation and volatilization. Environ. Sci. Technol. 1999, 33, 4034−4039. (71) Xu, S.; Wania, F. Chemical fate, latitudinal distribution and longrange transport of cyclic volatile methylsiloxanes in the global environment: A modeling assessment. Chemosphere 2013, 93, 835− 843. (72) Wania, F. 2006. Potential of degradable organic chemicals for absolute and relative enrichment in the Arctic. Environ. Sci. Technol. 2006, 40, 569−577. (73) Yucuis, R. A.; Stanier, C. O.; Hornbuckle, K. C. Cyclic siloxanes in air, including identification of high levels in Chicago and distinct diurnal variation. Chemosphere 2013, 92, 905−910. (74) Genualdi, S.; Harner, T.; Cheng, Y.; MacLeod, M.; Hansen, K. J.; van Egmond, R.; Shoeib, M.; Lee, S. Global distribution of linear and cyclic volatile methyl siloxanes in air. Environ. Sci. Technol. 2011, 45, 3349−3354. (75) Krogseth, I. S.; Kierkegaard, A.; McLachlan, M. S.; Breivik, K.; Hansen, K. M.; Schlabach, M. . Occurrence and seasonality of cyclic volatile methyl siloxanes in Arctic air. Environ. Sci. Technol. 2013, 47 (1), 502−9, DOI: 10.1021/es3040208. (76) McLachlan, M. S.; Kierkegaard, A.; Hansen, K. M.; van Egmond, R.; Christensen, J. H.; Skjoth, C. A. Concentrations and fate of decamethylcyclopentasiloxane (D5) in the atmosphere. Environ. Sci. Technol. 2010, 44 (14), 5365−5370.

11759

dx.doi.org/10.1021/es503465b | Environ. Sci. Technol. 2014, 48, 11748−11759