Do Persistent Organic Pollutants Reach a Thermodynamic Equilibrium

Article pubs.acs.org/est

Do Persistent Organic Pollutants Reach a Thermodynamic Equilibrium in the Global Environment? Sebastian Schenker, Martin Scheringer,* and Konrad Hungerbühler Institute for Chemical and Bioengineering, ETH Zürich, 8093 Zürich, Switzerland S Supporting Information *

ABSTRACT: Equilibrium partitioning between different environmental media is one of the main driving forces that govern the environmental fate of organic chemicals. In the global environment, equilibrium partitioning is in competition with long-range transport, advective phase transfer processes such as wet deposition, and degradation. Here we investigate under what conditions equilibrium partitioning is strong enough to control the global distribution of organic chemicals. We use a global multimedia mass-balance model to calculate the Globally Balanced State (GBS) of organic chemicals. The GBS is the state where equilibrium partitioning is in balance with long-range transport; it represents the maximum influence of thermodynamic driving forces on the global distribution of a chemical. Next, we compare the GBS with the Temporal Remote State, which represents the long-term distribution of a chemical in the global environment when the chemical’s distribution is influenced by all transport and degradation processes in combination. This comparison allows us to identify the chemical properties required for a substance to reach the GBS as a stable global distribution. We find that thermodynamically controlled distributions are rare and do not occur for most Persistent Organic Pollutants. They are only found for highly volatile and persistent substances, such as chlorofluorocarbons. Furthermore, we find that the thermodynamic cold-trap effect (i.e., accumulation of pollutants at the poles because of reduced vapor pressure at low temperatures) is often strongly attenuated by atmospheric and oceanic long-range transport.

■

INTRODUCTION Persistent Organic Pollutants (POPs) are of particular concern for the environment and their environmental fate has been extensively studied. POPs were observed in the Arctic and Antarctic already in the 1960s.1,2 The problem of persisting organic pollution was early recognized as a global issue1,3 and the question of what processes control the global and particularly the Arctic contamination by POPs was raised.4 Not only do POPs have the potential to reach the Arctic, but they also tend to accumulate there due to reduced vapor pressure and Henry’s law constant at low temperatures,5 reduced degradation rate constants,6 and in some cases increased scavenging by snow.7,8 However, it is still not clear what the main driving forces of the long-term distribution of POPs in the global environment are.9,10 In particular, the question of whether the long-term and long-range distribution of chemicals will eventually be determined by equilibrium partitioning has not yet been clarified.9,11 Mackay and Wania provided an important starting point for this discussion when they showed that equilibrium partitioning, a thermodynamically controlled process, can increase the “condensation” of pollutants at the poles5,12 and concluded, “it is this thermodynamic or equilibrium effect that drives the condensation phenomenon in cold climates.”5 But how likely is it that this thermodynamic effect determines the large-scale and long-term distribution of organic chemicals? Two main factors that may prevent chemicals from reaching thermodynamic equilibrium in an environmental multimedia © 2014 American Chemical Society

system are (i) advective phase-transfer processes such as wet deposition from air to surface media and (ii) degradation. For persistent volatile substances such as chlorofluorocarbons (CFCs), carbon tetrachloride, and even hexachlorobenzene (HCB), these two factors are not very influential because these substances are long-lived and do not sorb strongly to atmospheric aerosols and particulate matter in ocean water (no strong advective phase transfer with settling particles). These substances can therefore reach a global distribution that is controlled by airborne long-range transport in combination with equilibrium partitioning between air and surface media. This leads to nearly uniform spatial concentration distributions in air and concentration distributions in surface media controlled by equilibrium partitioning between air and surface media such as water13 or vegetation.14,15 We denote such a situation in which long-range transport and equilibrium partitioning are balanced as the “Globally Balanced State” (GBS). Importantly, in systems with a temperature gradient between different regions, a spatially uniform concentration in air and water and an air−water distribution according to equilibrium partitioning cannot be reached at the same time. Thus, the GBS reflects the optimal balance between two counteracting driving forces, long-range transport and equiliReceived: Revised: Accepted: Published: 5017

December 13, 2013 February 27, 2014 March 21, 2014 March 21, 2014 dx.doi.org/10.1021/es405545w | Environ. Sci. Technol. 2014, 48, 5017−5024

Environmental Science & Technology

Article

The coefficients, ai, are derived from the initial condition, c(⃗ t = 0) = Σi ai · vi⃗ . A requirement for the GBS is that the concentrations, cj, in all compartments, j, remain constant with time and the system, therefore, is at steady-state and dc(⃗ t)/dt = 0. The steady-state solution, cstst ⃗ , for the mass-balance equation is:

brium partitioning, and it is the global distribution with the strongest influence of equilibrium partitioning. For semivolatile organic chemicals (SOCs) such as polychlorinated biphenyls (PCBs), the situation is different. Advective phase transfer and also degradation may strongly affect the long-term distribution of SOCs in the global environment. The south−north distribution of PCBs in Europe, for example, seems to be controlled by airborne transport and removal from the air by deposition and/or degradation and, thus, distance from emission sources rather than by temperature-dependent equilibrium partitioning between air and surface media.16 In ocean water, SOCs are transferred from the surface layer to the deep ocean by settling particles.11,17 This “biological pump”17 creates a downward flux of SOCs that may prevent substances with strong sorption to settling particles from reaching air−water equilibrium.11 Here, we systematically investigate whether POPs and other SOCs can reach a GBS as it has been observed in the global environment for persistent volatile chemicals such as CFCs. To this end, we employ a global environmental fate model for organic chemicals; we use this model to calculate (i) a global distribution of chemicals that is only affected by long-range transport and equilibrium partitioning (the GBS) and (ii) a state that reflects the actual long-term fate of chemicals in the global environment, i.e., the combination of long-range transport, phase partitioning, and degradation (the Temporal Remote State, TRS18), and we then compare these two states for a range of organic chemicals.

c ⃗ stst = A−1·q ⃗

However, in the special case of the GBS, the matrix A is constructed without any degradation processes and is, therefore, singular (see Supporting Information, SI, for a proof). Under these conditions, a steady-state solution does not exist, because A is not invertible and no continuous emissions take place. In this situation, we apply a trick to calculate a solution that fulfills the steady-state conditions; we introduce an arbitrary removal process in an arbitrarily selected compartment of the model, denoted by p. This is done by adding a removal matrix, B, to the matrix A; all entries of B are zero except one entry on the main diagonal, B pp = −1. Correspondingly, the emission vector, q⃗, has elements equal to zero except for the element at position p, which is qp = 1. Now the equation dc (⃗ t ) = −(A + B) · c (⃗ t ) + q ⃗ dt

METHODS Globally Balanced State. We define the Globally Balanced State, GBS, as the state in which equilibrium partitioning and long-range transport are in balance everywhere on a global scale. Processes that counteract equilibrium partitioning such as degradation and advective (i.e., one-way) phase-exchange processes such as settling to the deep ocean or wet deposition from air to surface media are not considered in the GBS. We use the multimedia environmental fate model CliMoChem to calculate this state. The mass-balance equations in CliMoChem form a system of first-order differential equations. In a system with n compartments, all n concentrations can be combined in a column vector, c(⃗ t). For c(⃗ t), the following equation holds: (1)

A is an n × n matrix containing all rate constants of the processes that determine the amount of chemical in each compartment. q⃗ is a column vector of the rates of continuous emissions of chemical into the different compartments. The matrix A is invertible and has positive eigenvalues, γi, and eigenvectors, vi⃗ . For the case of an initial emission of chemical without further emissions (q⃗ = 0), the mass-balance equation is as follows:

dc (⃗ t ) = −A· c (⃗ t ) dt

(2)

and the time-dependent solution of this equation is as follows: n

c (⃗ t ) =

∑ i=1

ai ·vi⃗ ·exp( −γi·t )

(5)

can be solved according to eq 4, which yields the concentrations in the Globally Balanced State, cGBS ⃗ . We also calculate the amounts of chemical in each compartment, j, and = 1. a vector of normalized amounts, m⃗ GBS, with ∑j mGBS j In the same way as described above for the GBS, we also calculate a state where, in addition to equilibrium partitioning and long-range transport, also advective phase transfer processes are active, such as wet deposition from air to surface media. We call this state the “Disturbed Globally Balanced State”, DGBS. The spatial distribution of normalized chemical mass in the DGBS is denoted by m⃗ DGBS, with ∑j mDGBS = 1. j Temporal Remote State. A second characteristic solution of the mass-balance equation (eq 2) is the Temporal Remote State, TRS.18 The TRS is defined as the state where the timedependent solution shown in eq 3 is dominated by the exponential with the smallest eigenvalue, γmin; all other exponentials are virtually zero in this state because they decrease faster and have already (numerically) disappeared.18 In contrast to the GBS, advective phase exchange processes and degradation processes are active in the calculation of the TRS. The TRS describes the distribution of a chemical at times sufficiently long after the end of the emissions. In the TRS, the amounts of chemical in all compartments decrease at the same rate, which is defined by the smallest eigenvalue of the model matrix, γmin. Accordingly, the relative fractions of chemical in all compartments, j, remain constant in the TRS; importantly, they are independent of the initial condition, i.e. the concentrations at the time when the emissions stop. The TRS is calculated by solving eq 2 in a stepwise fashion with an arbitrary initial concentration vector, c(⃗ t = 0). Each time step, Δt, yields a new concentration vector, c(⃗ t + Δt), and a new normalized mass vector, m⃗ (t + Δt). This process is stopped when the new normalized mass vector deviates by less than 0.1% from the preceding one, and this last normalized mass vector is used as the global mass distribution of the chemical in the TRS, m⃗ TRS.

■

dc (⃗ t ) = −A· c (⃗ t ) + q ⃗ dt

(4)

(3) 5018

dx.doi.org/10.1021/es405545w | Environ. Sci. Technol. 2014, 48, 5017−5024


Article

Figure 1. Mass fractions (bars) and concentrations (lines) of CCl4 (A) and α-HCH (B) in the Globally Balanced State calculated with the 18-zone CliMoChem model. Each zone represents 10° of latitude.

Comparison of GBS and TRS. Our initial question of whether equilibrium partitioning controls the global distribution of chemicals in the long run can now be posed as follows: Is the long-term distribution of a chemical in the global environment, i.e., the TRS, similar to the GBS, which is the state with the strongest possible influence of equilibrium partitioning? To quantify the similarity of the TRS and the GBS, we calculate the correlation coefficient, R2GBS‑TRS, of the normalized mass vectors, m⃗ GBS and m⃗ TRS, of the two states. We then analyze how R2GBS‑TRS changes when the physicochemical properties of the chemicals are varied. The same analysis is also performed for the DGBS, but the focus is on the GBS as the state with the strongest possible influence of equilibrium partitioning. The CliMoChem Model. In this work, the global distribution of chemicals is modeled with CliMoChem, a zonally averaged global multimedia mass-balance model.19 We use a version of CliMoChem that consists of 18 latitudinal zones; each zone represents 10 degrees latitude and contains up to seven media: bare soil and vegetation-covered soil, oceanic surface water, tropospheric air, vegetation, ice, and snow. The processes included in the current version of CliMoChem have been described elsewhere; they include chemical and/or microbial degradation in all media,19,20 advective phase transfer such as wet and dry particle deposition,21 deposition to deep ocean water,22 long-range transport by wind and water,19 as well as vegetation-specific23 and cryosphere-specific24 processes. All processes are described by first-order kinetics and for all media, yearly averaged parameters for transport, temperature and meteorological data are used. The CliMoChem model is based on extensive empirical information19−24 and has been validated in several earlier studies. The model has been shown to reliably reproduce the environmental distribution of different types of organic chemicals measured in the field.19−21,25,26 Here we use the model to calculate the two conceptually defined reference states, the GBS and the TRS, but not to reproduce current levels of chemicals in the environment. For comparisons of the model results to field data, the readers are referred to our earlier studies.19−21,25,26 Test Chemicals. In this work, we investigate a set of 22 test chemicals that were also used in earlier studies with the CliMoChem model. The chemicals include volatile substances such as carbon tetrachloride, CCl4, and a range of semivolatile organic chemicals such as DDT, selected PCB congeners, and atrazine. The test chemicals and their properties are listed in Tables 1−4 in the SI.

CliMoChem uses 18 chemical-specific input parameters. These are the degradation rate constants in water, soil, and vegetation (ks/w/v), the degradation rate constant of direct photolysis in air and in the top soil and top water layers (kphoto s/w/a), the degradation rate constant of the reaction with hydroxyl radicals in air (kOH a ), activation energies for the degradation rate constants in water, soil and vegetation (Eas/w/a), the partition coefficients for air/water, octanol/water and hexadecane/air (KAW, KOW, KHA), air/water and octanol/water phase transfer enthalpies (ΔUAW/OW), and solvatochromic (Abraham-)parameters for H-bond donor and acceptor strength (α, β). A detailed list of all parameter values is provided in the SI. Chemical Space Plots. For the calculation of chemical space plots, we employ a set of hypothetical chemicals as defined by Fenner et al.27 to cover the properties of organic chemicals that might be of concern for the environment. The hypothetical chemicals are defined by their half-lives in air, tair 1/2, which range from 4 h to 1 year (5 steps); by their half-lives in water, twater 1/2 , ranging from 24 h to 10 years (5 steps), by log KAW ranging from −11 to 2 (27 steps), log KOW ranging from −1 to 8 (19 steps) and log KOA ranging from 2 to 12. We confirmed that these parameters actually determine the properties of chemicals in the studied system by a sensitivity analysis, as shown in the SI. This approach allows us to define 513 hypothetical chemicals with different combinations of log KOW and log KAW, which leads to 513 different GBS and DGBS mass distributions and to 25 times as many different TRS mass distributions, 12 825 in total, due to the different possible degradation half-lives.

■

RESULTS AND DISCUSSION Globally Balanced State: Mass Distributions of the Test Chemicals. As a first step, we consider the spatial distributions of two chemicals, CCl4 and α-HCH, in the GBS. CCl4 has a log KAW of −0.05 and a log KOA of 2.88 and is therefore characterized as a flier according to Gouin and Wania28 and expected to predominantly partition to air in the GBS. α-HCH has a log KAW of −3.85 and a log KOA of 7.39 and is therefore characterized as a swimmer. The mass distribution in the GBS along with concentrations in air and seawater are shown for both chemicals in Figure 1. For CCl4, we find 85.3% of the total mass in the atmosphere; the concentration is uniform across all latitudinal zones and the mass distribution follows the volumes of the air compartments of the 18 latitudinal zones. 13.1% of the mass of CCl4 reside in seawater and only 1.6% in soil. The other media do not contain notable amounts of CCl4. Concentrations of CCl4 in seawater 5019



Article

increase from the tropics toward the poles by a factor of 9. This is consistent with field measurements, which show an increase in CCl4 concentrations in seawater by a factor of 6 between the tropics and the Arctic and constant concentrations in air.13 From the concentrations in air and seawater, we calculate an apparent air−seawater partition coefficient, Kapp air/sea, for each zone and compare it to the temperature-corrected value of the T true air−seawater partition coefficient, Kair/sea , which is employed in the solution of CliMoChem and reflects the effect of temperature on the air−seawater partitioning. This apparent partition coefficient shows the extent to which transport in S−N direction reduces the range of concentrations as they would derive from the temperature-dependent true partition coefficient. T For CCl4, Kapp air/sea is nearly identical to Kair/sea in all latitudinal zones. The distribution of CCl4 is clearly controlled by longrange transport in air and equilibrium partitioning to seawater as it is predicted by KTair/sea. For α-HCH, 96.8% of the mass resides in the soil, see bars in Figure 1B (mass fractions in seawater and air are 2.8% and 0.4%, respectively, and not shown in Figure 1B). The concentration in seawater increases in the Arctic compared to the Tropics, whereas for the concentration in air the opposite trend is observed, see full and dashed lines in Figure 1B. In contrast to CCl4, the concentrations of α-HCH in air and seawater vary only by factors 1.7 and 1.4, respectively. The apparent partition coefficient, log Kapp air/sea, of α-HCH has values between 3.6 and 4.0, whereas log KTair/sea ranges from 3.5 to 4.5. The apparent partition coefficient thus spans a narrower range than what would be expected under local, temperaturedependent equilibrium partitioning. We also calculate the fraction to which long-range transport reduces the effect of the cold-trap effect in the GBS; this fraction is denoted by fΔlog Katm/sea and is given by the following: fΔlog K

= atm/sea

Table 1. Properties of the Globally Balanced State and Correlation Coefficients, R2GBS‑TRS, of the Normalized Global Mass Distribution in the Globally Balanced State (GBS) and the Temporal Remote State (TRS)a

a

Δlog Kapp air/sea

Δlog KTair/sea

fΔlog Katm/sea

R2GBS-TRS

CCl4 α-HCH γ-HCH dieldrin tri-BDE tetra-BDE penta-BDE hexa-BDE hepta-BDE octa-BDE nona-BDE deca-BDE PCB-28 PCB-153 PCB-180 DDD DDE DDT alachlor atrazine terbuthylazine HCB

0.99 0.38 0.37 0.77 0.65 0.38 0.86 1.22 1.30 1.32 1.35 1.39 0.76 0.79 0.51 0.61 0.37 0.44 0.67 0.74 0.43 0.88

1.00 1.17 1.17 1.08 0.98 2.18 2.73 3.26 3.02 3.50 3.54 3.55 0.84 1.51 1.95 1.70 2.23 1.82 1.67 2.08 1.74 0.95

0.98 0.32 0.29 0.71 0.66 0.17 0.31 0.37 0.43 0.38 0.38 0.39 0.90 0.53 0.26 0.36 0.16 0.24 0.40 0.36 0.25 0.93

0.98 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.01 0.01 0.00 0.01 0.01 0.00 0.01 0.01 0.01 0.01 0.02

fΔlog Katm/sea calculated according to eq 6.

GBS. Also the concentration profiles in air and seawater are very similar to those obtained for the GBS. Accordingly, the correlation coefficient, R2GBS‑TRS, of the two mass distributions is 0.99. We conclude that for CCl4 the thermodynamic driving force plays a dominant role in the long term global distribution. For α-HCH, in the TRS 99.3% of the mass is found in the seawater of the Northern hemisphere and 0.6% in the seawater of the Southern hemisphere, which is very different from the GBS of α-HCH, see Figure 1. In contrast to the GBS solution, the concentrations in air and seawater vary by factors of 200 and 20 000, respectively, compared to factors of 1.7 and 1.4 in the GBS. In the case of α-HCH, the mass distributions in the TRS and the GBS are not correlated at all (R2GBS‑TRS = 0.01). We conclude that for α-HCH thermodynamic driving forces only play a minor role for the long-term distribution in the global environment. In the TRS, a chemical forms a reservoir in the compartment where the chemical has its longest degradation half-life. For αHCH, we used a degradation half-life in seawater that is longer than the half-life in soil (Table 3 in the SI) and this leads to a TRS that is dominated by α-HCH in seawater. The GBS of α-HCH, in contrast, is dominated by high fractions in soil, because soil is the thermodynamically favored environmental medium for α-HCH. (Note that the environmental degradation half-lives of persistent organochlorine chemicals such as α-HCH are uncertain and that it is also possible that the actual half-life of α-HCH in soil is longer than that in water. However, this does not imply that TRS and GBS are similar, as will be discussed further below with the example of other persistent organochlorine chemicals.) For the 22 test chemicals, the correlation between the mass distributions in the GBS and the TRS is shown in Table 1. A high correlation of the TRS with the GBS is only found for

app app max(log K atm/sea ) − min(log K atm/sea ) T T max(log K atm/sea ) − min(log K atm/sea )

chemical

(6)

fΔlog Katm/sea indicates how much of the temperature-related spread in log KTair/sea is conserved in a system with long-range transport and equilibrium partitioning in competition. fΔlog Katm/sea and the ranges of log KTair/sea and log Kapp air/sea are listed in Table 1 for the 22 test chemicals. For most of the test chemicals, the T range in log Kapp air/sea is only one-third of the range of log Kair/sea, and this is caused by airborne transport in S−N direction, which reduces the spatial concentration gradient that is built up by differences between phase partitioning in warm and cold regions. Only chemicals that partition substantially to air, such as CCl4, HCB or PCB-28, have log Kapp air/sea values close to the theoretical, temperature-corrected log KTair/sea values. This is consistent with evidence for conditions that are close to air− water equilibrium for HCB and lighter PCBs derived from field data by Galbán-Malagón et al.29,30 Temporal Remote State distributions of the Test Chemicals. In order to evaluate whether the GBS of a chemical in the global environment is stable and a potential final state, the spatial distribution of the chemical mass in the GBS is compared to the mass distribution in the TRS, see Figure 2 for CCl4 and α-HCH. For CCl4, we find a TRS distribution that is very similar to the GBS distribution: 87% in air, 12% in water, and 0.7% in soil, compared to 85% (air), 13% (water), and 1.6% (soil) in the 5020



Article

Figure 2. Mass fractions (bars) and concentrations (lines) of CCl4 (A) and α-HCH (B) in the Temporal Remote State calculated with the 18-zone CliMoChem model. Each zone represents 10° of latitude.

2 Figure 3. Correlation coefficients between mass distributions in the Globally Balanced State and the Temporal Remote State, RGBS‑TRS , obtained from the set of hypothetical chemicals with different air−water partition coefficients, KAW, and octanol−air partition coefficients, KOA, and different water half-lives in air, tair 1/2, and in water, t1/2 . Cross-hairs: approximate position of test chemicals; PBDEs: tetra- to hepta-BDE; PCBs: PCB-153 and 180.

CCl4. All other test chemicals are not controlled by equilibrium driving forces in the long term but rather form a final reservoir that acts as a kinetically controlled long-term secondary emission source. The environmental distribution is determined by the emission rate from that reservoir and subsequent degradation in all other compartments. This also applies to chemicals that have their longest half-life in soil and for which soil also is the thermodynamically favored reservoir, such as PCBs. In the TRS, PCBs form a reservoir in the soil of zone 2 of the model (Arctic soil), whereas in the GBS, PCBs reside in the soil in zones 2−6 (see Figures S7−S8 in the SI). However, the reservoirs in zones 3−6, which are

stable in the GBS, are degraded more rapidly than the reservoir in zone 2 and, therefore, disappear during the transition to the TRS. This is why the GBS is not stable and why, in the long term, a final reservoir remains that acts as a kinetically controlled secondary emission source. Chemical Space Plots. In the next step, we compare the TRS distributions of the set of hypothetical chemicals to the chemicals’ GBS distributions. Figure 3 shows the 25 chemical space plots; the coloring of the plots represents the value of the correlation coefficient, R2GBS‑TRS. There are two types of TRS that are highly correlated with the corresponding GBS. Either the half-life in water is on the order of 10 years or higher, which 5021



Article

Figure 4. (A) Maximum correlation coefficient of mass distributions in the Globally Balanced State (GBS) and the Temporal Remote State (TRS), 2 . (B) maximum correlation coefficient of mass distributions in the Disturbed Globally Balanced State (DGBS) and the Temporal Remote RGBS‑TRS,max State (TRS), R2DGBS‑TRS,max.

allows chemicals with low log KAW and low log KOA to reach equilibrium globally balanced with long-range transport in water, or the half-life in air is two months or higher, which allows chemicals with high log KAW and relatively low log KOA to reach equilibrium globally balanced with long-range transport in air. Only 909 (7.1%) of the 12 825 hypothetical chemicals fall into these groups and have correlation coefficients, R2GBS‑TRS, above 0.9. In Figure 4A, the highest possible correlation coefficient between the GBS and the TRS, R2GBS‑TRS,max, is shown for each combination of log KAW and log KOA independently of the degradation half-lives in water and air, i.e., a superposition of all panels in Figure 3. The regions that include chemicals characterized as fliers, swimmers, or single and multiple hoppers, according to Gouin and Wania,28 are marked in gray. Figure 4A shows that high corrrelation coefficients, R2GBS‑TRS,max, are only found for hypothetical chemicals that are either characterized as fliers with a log KOA below 4.5 or as swimmers with a log KOW below 2. This shows that there are two general conditions under which chemicals have a stable GBS: (i) a chemical has to have its longest half-life in the medium for which it has the strongest thermodynamic affinity (fliers with longest half-lives in air; swimmers with longest half-lives in water); (ii) these media have to be mobile (air, water) so that the spatial distribution of the chemical can be globally balanced by long-range transport within the medium of highest thermodynamic capacity for the chemical. This second condition is not fulfilled for soil. Accordingly, chemicals that have the longest half-life in soil and also the highest thermodynamic affinity for soil, such as many POPs, do not form a stable GBS. In the real environment, it may well be that chemicals with highest affinity for soil and also longest half-lives in soil reach, at some point in time, a state that is close to the GBS. In this sense, their global circulation is indeed driven by thermodynamic controls, i.e., equilibrium partitioning, but only for a certain period of time. A possible case of this type is HCB with increasing concentrations in vegetation in colder regions, which is a thermodynamically driven effect.14,15 However, this state is not stable in the long term; it only forms a transitional phase between an initial phase that is dominated by primary emissions and the final phase, which is defined by the TRS.31 In warmer regions, the thermodynamically stable reservoirs of such a

chemical are degraded more rapidly than in colder regions and only the most long-lived reservoir in the coldest region will be left after a certain time. Where this reservoir is formed, is determined by the longest half-life, but not by the partition coefficients, of a chemical. Finally, Figure 4B shows the maximum correlation coefficient between the mass distribution in the TRS and in the DGBS, i.e., the state where advective phase-transfer processes such as wet and dry particle deposition or transfer to the deep ocean are included in addition to equilibrium partitioning and long-range transport. Interestingly, this correlation coefficient, R2DGBS‑TRS,max, is close to 1 in the same domains where also the GBS has high correlations with the TRS. In other words, there are two domains where all three states, GBS, DGBS, and TRS, are highly correlated (dark brown color in Figure 4A, B), and other domains where either advective phase exchange processes or degradation processes or both prevent chemicals from reaching a thermodynamically controlled long-term distribution in the global environment (medium and light brown colors in Figure 4A, B). Dampening of the Cold-Trap Effect by Long-Range Transport. Finally, we use the GBS to test whether the coldtrap effect can be a driver for air−water distributions observed in the global environment. There are two prerequisites for the distribution of a chemical to be controlled by the cold-trap effect. First, the chemical needs to have properties that allow the chemical to reach its GBS and, second, the GBS of the chemical needs to be controlled by the cold-trap effect. Therefore, the correlation coefficient, R2GBS‑TRS,max, needs to be high and fΔlog Katm/sea from eq 6 above needs to be close to 1. Values of fΔlog Katm/sea are shown in Figure 5, indicating how much of the spread in log KTair/sea is conserved in the GBS. The hatched area in Figure 5 indicates chemicals that have a stable GBS, provided that the half-lives of the chemicals are in a feasible range. In this domain, chemicals with a log KAW above −2 show high values of fΔlog Katm/sea, whereas a log KAW below −2 reduces fΔlog Katm/sea. This finding can be explained by the fact that north−south transport in air is much faster than in water. Long-range transport in water has hardly any effect on the concentration profile in water of a volatile chemical and a stable spatial concentration gradient in water, determined by the temperature-dependent partition coefficient, log KTair/sea, can be formed 5022



Article

(6) Scheringer, M.; Salzmann, M.; Stroebe, M.; Wegmann, F.; Fenner, K.; Hungerbühler, K. Long-range transport and global fractionation of POPs: Insights from multimedia modeling studies. Environ. Pollut. 2004, 128, 177−188. (7) Lei, Y. D.; Wania, F. Is rain or snow a more efficient scavenger of organic chemicals? Atmos. Environ. 2004, 38, 3557−3571. (8) Roth, C. M.; Goss, K.-U.; Schwarzenbach, R. P. Sorption of diverse organic vapors to snow. Environ. Sci. Technol. 2004, 38, 4078− 4084. (9) Dalla Valle, M.; Jurado, E.; Dachs, J.; Sweetman, A. J.; Jones, K. C. The maximum reservoir capacity of soils for persistent organic pollutants: Implications for global cycling. Environ. Pollut. 2005, 134, 153−164. (10) Lohmann, R.; Breivik, K.; Dachs, J.; Muir, D. Global fate of POPs: Current and future research directions. Environ. Pollut. 2007, 150, 150−165. (11) Jurado, E.; Lohmann, R.; Meijer, S.; Jones, K. C.; Dachs, J. Latitudinal and seasonal capacity of the surface oceans as a reservoir of polychlorinated biphenyls. Environ. Pollut. 2004, 128, 149−162. (12) Wania, F.; Mackay, D. Tracking the distribution of persistent organic pollutants. Environ. Sci. Technol. 1996, 30, 390A−396A. (13) Hunter-Smith, R. J.; Balls, P. W.; Liss, P. S. Henry’s Law constants and the air-sea exchange of various low molecular weight halocarbon gases. Tellus B 1983, 35B, 170−176. (14) Calamari, D.; Bacci, E.; Focardi, S.; Gaggi, C.; Morosini, M.; Vighi, M. Role of plant biomass in the global environmental partitioning of chlorinated hydrocarbons. Environ. Sci. Technol. 1991, 25, 1489−1495. (15) Simonich, S. L.; Hites, R. A. Global distribution of persistent organochlorine compounds. Science 1995, 269, 1851−1854. (16) von Waldow, H.; MacLeod, M.; Jones, K.; Scheringer, M.; Hungerbühler, K. Remoteness from emission sources explains the fractionation pattern of polychlorinated biphenyls in the Northern Hemisphere. Environ. Sci. Technol. 2010, 44, 6183−6188. (17) Dachs, J.; Lohmann, R.; Ockenden, W. A.; Méjanelle, L.; Eisenreich, S. J.; Jones, K. C. Oceanic biogeochemical controls on global dynamics of persistent organic pollutants. Environ. Sci. Technol. 2002, 36, 4229−4237. (18) Stroebe, M.; Scheringer, M.; Hungerbühler, K. Measures of overall persistence and the temporal remote state. Environ. Sci. Technol. 2004, 38, 5665−5673. (19) Scheringer, M.; Wegmann, F.; Fenner, K.; Hungerbühler, K. Investigation of the cold condensation of persistent organic pollutants with a global multimedia fate model. Environ. Sci. Technol. 2000, 34, 1842−1850. (20) Becker, L.; Scheringer, M.; Schenker, U.; Hungerbühler, K. Assessment of the environmental persistence and long-range transport of endosulfan. Environ. Pollut. 2011, 159, 1737−1743. (21) Götz, C. W.; Scheringer, M.; MacLeod, M.; Wegmann, F.; Schenker, U.; Hungerbühler, K. Dependence of persistence and longrange transport potential on gas-particle partitioning in multimedia models. Environ. Sci. Technol. 2008, 42, 3690−3696. (22) Scheringer, M.; Stroebe, M.; Wania, F.; Wegmann, F.; Hungerbühler, K. The effect of export to the deep sea on the longrange transport potential of persistent organic pollutants. Environ. Sci. Pollut. Res. 2004, 11, 41−48. (23) Wegmann, F.; Scheringer, M.; Möller, M.; Hungerbühler, K. Influence of vegetation on the environmental partitioning of DDT in two global multimedia models. Environ. Sci. Technol. 2004, 38, 1505− 1512. (24) Stocker, J.; Scheringer, M.; Wegmann, F.; Hungerbühler, K. Modeling the effect of snow and ice on the global environmental fate and long-range transport potential of semivolatile organic compounds. Environ. Sci. Technol. 2007, 41, 6192−6198. (25) Schenker, U.; Scheringer, M.; MacLeod, M.; Martin, J. W.; Cousins, I. T.; Hungerbühler, K. Contribution of volatile precursor substances to the flux of perfluorooctanoate to the arctic. Environ. Sci. Technol. 2008, 42, 3710−3716.

Figure 5. Values of fΔlog Katm/sea for the hypothetical chemicals considered in this study. Green: fΔlog Katm/sea = 1, cold-trap effect is fully visible. Dark red: fΔlog Katm/sea = 0, no cold-trap effect at all. Hatched: chemicals with a stable Globally Balanced State (GBS).

in the GBS, see CCl4 concentration in water in Figure 1. In contrast, the fast advective transport in air has a strong effect on the concentrations of swimmers in air and tends to even out a concentration gradient in air that is caused by temperaturedependent equilibrium partitioning between water and air. Therefore, a stable spatial concentration gradient in air cannot be as steep as a stable concentration gradient in water, which leads to a dampening of the cold-trap effect for swimmers. In conclusion, global distributions controlled by the cold-trap effect can occur for highly volatile and long-lived chemicals only. For all other long-lived organic chemicals, increased concentrations in polar regions are most likely to be explained by direct deposition and conservation, i.e., slow degradation of the chemicals deposited, rather than by increased equilibrium partitioning because of low temperatures.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed information on the GBS solution of the CliMoChem model, physicochemical property data of the test chemicals, and graphs of the mass distributions of the test chemicals in the GBS, DGBS, and TRS. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone +41 44 6323062; e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Risebrough, R. W.; Rieche, P.; Peakall, D. B.; Herman, S. G.; Kirven, M. N. Polychlorinated biphenyls in the global ecosystem. Nature 1968, 220, 1098−1102. (2) Tatton, J. O’G.; Ruzicka, J. H. A. Organochlorine pesticides in Antarctica. Nature 1967, 215, 346−348. (3) Holden, A. V.; Marsden, K. Organochlorine pesticides in seals and porpoises. Nature 1967, 216, 1274−1276. (4) Goldberg, E. D. Synthetic organohalides in the sea. Proc. R. Soc. London. B. 1975, 189, 277−289. (5) Mackay, D.; Wania, F. Transport of contaminants to the Arctic: Partitioning, processes and models. Sci. Total Environ. 1995, 160/161, 25−38. 5023



Article

(26) Schenker, U.; Soltermann, F.; Scheringer, M.; Hungerbühler, K. Modeling the environmental fate of polybrominated diphenyl ethers (PBDEs): The importance of photolysis for the formation of lighter PBDEs. Environ. Sci. Technol. 2008, 42, 9244−9249. (27) Fenner, K.; Scheringer, M.; MacLeod, M.; Matthies, M.; McKone, T.; Stroebe, M.; et al. Comparing estimates of persistence and long-range transport potential among multimedia models. Environ. Sci. Technol. 2005, 39, 1932−1942. (28) Gouin, T.; Wania, F. Time trends of arctic contamination in relation to emission history and chemical persistence and partitioning properties. Environ. Sci. Technol. 2007, 41, 5986−5992. (29) Galbán-Malagón, C. J.; Berrojalbiz, N.; Ojeda, M. J.; Dachs, J. The oceanic biological pump modulates the atmospheric transport of persistent organic pollutants to the Arctic. Nat. Commun. 2012, 3, 862 DOI: 10.1038/ncomms1858. (30) Galbán-Malagón, C. J.; Berrojalbiz, N.; Gioia, R.; Dachs, J. The “degradative” and “biological” pumps controls on the atmospheric deposition and sequestration of hexachlorocyclohexanes and hexachlorobenzene in the North Atlantic and Arctic Oceans. Environ. Sci. Technol. 2013, 47, 7195−7203. (31) Nizzetto, L.; MacLeod, M.; Borgå, K.; Cabrerizo, A.; Dachs, J.; Di Guardo, A.; et al. Past, present, and future controls on levels of persistent organic pollutants in the global environment. Environ. Sci. Technol. 2010, 44, 6526−6531.

5024


Do Persistent Organic Pollutants Reach a Thermodynamic Equilibrium

Recommend Documents