Activity, aqueous concentration and toxicity

Ferguson (1939) demonstrated that chemical activity could be used as a metric of toxicity, the inherent assumption being that at equilibrium the activity in the organism will approach the activity in the exposure medium. Fundamentally, equilibrium partitioning of a substance between two phases occurs when the criterion of equilibrium chemical potential of the substance is equal in both phases, Schwarzenbach et al, (2003). More convenient criteria of equilibrium are the related quantities of chemical activity and fugacity that are linearly related to concentrations, at least at low concentrations, and can also be applied to air, water, soils, sediments and biota. Fugacity is essentially the chemical’s partial pressure and can range from zero to a maximum of the substance’s liquid state vapour pressure. Activity is the dimensionless ratio of fugacity to that vapour pressure and can thus range from zero to 1.0. Activity is essentially the fraction of saturation. The activity concept is also used for ions but with a different definition from that used here.

Activity thus serves as a direct link between external exposure and delivered dose. Further, for a series of chemicals, if it is hypothesised that narcotic toxicity occurs at relatively similar concentrations (and hence activities) in membrane lipids and in whole organisms, then activities in the exposure medium of water will also be similar, however, the corresponding lethal concentrations in the exposure medium (LCs) can be widely different. The test of the hypothesis is that the highly variable LCs for a diverse set of chemical substances will correspond to a relatively narrow range of activities. Rather than calculate the activities corresponding to the LCs and ECs, it is more convenient to plot these metrics of toxicity against solubility of the liquid state chemical. Since activity is the ratio of concentration and solubility, points corresponding to equal activity will fall on a 45 degree diagonal on a log-log plot and a cluster of points will fall around a 45 degree diagonal with a slope of 1.0. In reality, the slope observed by Mackay et al (2009) was lower and about 0.8.

When the chemical is a solid, i.e. the melting point (TM; units K) exceeds ambient temperature, it is necessary to use the sub-cooled liquid state properties to estimate chemical activity. In a solution at low concentration the chemical behaves as if its saturation condition or reference state is that of the sub-cooled liquid state vapour pressure or solubility, not the solid state that is additionally influenced by crystalline interactions in the solid. The vapour pressure and solubility of the solid substance are thus lower than that of the hypothetical sub-cooled liquid by a factor termed the fugacity ratio (F). The fugacity ratio can be estimated at the ambient temperature (T; units K) from the substance’s TM (also units K) and the entropy of fusion at the melting point (ΔS; units J/mol K). A value of 56.5 J/mol K can be assumed in some cases to estimate ΔS and thus F can be calculated as exp (-6.79(TM/T-1)), where 6.79 = 56.5/8.314, i.e. the estimate for ΔS divided by the gas constant (R; units 8.314 J/mol K).

An example is solid naphthalene with a molar mass of 128 g/mol, melting point of 80OC (353K) a solid vapour pressure of 10.9 Pa and solubility of 33 mg/L. At 25OC, F is 0.286, thus the corresponding liquid state values are 38.1 Pa and 115.4 mg/L or 0.90 mol/m3 and 0.00090 mol/L. At a low concentration in air and water the effective reference or saturation state is that of the liquid, thus at 1% of saturation the fugacity or partial pressure of naphthalene is 0.381 Pa, the concentration in water is 1.154 mg/L and the activity is 0.01. The activity corresponding to the solid state vapour pressure and solubility is 0.286, the fugacity ratio. An implication is that naphthalene cannot exist in solution in air or water at conditions exceeding an activity of 0.286 because at higher activities solid naphthalene will phase separate or ‘precipitate’ from solution. High melting point solids such as hexachlorobenzene may be unable to achieve concentrations and activities necessary to cause toxic effects (Di Toro et al, 1991). This constraint does not necessarily apply to liquid mixtures of high melting point solids such as commercial polychlorinated biphenyls (PCBs), crude oils and petroleum products (Kipka and Di Toro, 2009).

It is apparent from the work of Mackay et al (2009, 2011) that Ferguson’s hypothesis appears to be valid and that the use of chemical activity provides an estimate of toxic potency for narcotic chemicals within an order of magnitude. As such, chemical activity is a preferred metric over the use of chemical concentrations, which can span several orders of magnitude in environmental and toxicity testing media. It should be noted that the relatively simplistic chemical activity / toxicity concept cannot be applied to non-narcotics because the toxicity of chemicals with specific mode(s) of action do not have a simple relationship between toxicity and hydrophobicity. The potency of such chemicals is greater than baseline (narcotic toxicity) because these chemicals have a tendency and/or ability to interact with biological processes in organisms through non-hydrophobic and more specific modes of action / binding mechanisms (e.g. hydrogen bonding, ionic interactions or covalent bonding). Thus, screening out chemicals with toxicity exceeding baseline toxicity is seen as one of the advantages of using the chemical activity approach (Mackay et al, 2009). More recently, the Target Lipid Model (TLM) has been successful in expressing the toxicity of narcotic chemicals to aquatic organisms (Kipka and Di Toro, 2009; McGrath and Di Toro, 2009). The TLM is consistent with these concepts of narcosis in that chemical toxicity is induced by a relatively constant concentration of the chemical (e.g. hydrocarbons) in lipid membranes causing loss of essential function. For structurally similar substances, the lipid concentration is proportional to the chemical activity because their activity coefficients in octanol and probably in lipids, are similar (Xiao and Wania, 2003). Mayer et al (2009) also observed similar activities for a range of PAHs in several lipid types. The TLM has been successfully applied within the CONCAWE PETROTOX model, which has been used to predict the aquatic toxicity of various petroleum distillate substances as part of EU REACH registration requirements (McGrath et al, 2005; Redman et al, 2007; 2012).

An ECETOC Task Force was set up to consider the value of CBB related strategies such as activities as defined by phase equilibrium thermodynamics. The aim was to evaluate the potential for the activity framework to contribute to more effective risk assessment by integrating information on chemical structure and properties, MoA, acute and chronic effects for a range of aquatic organisms. In doing so the observed variability in activity levels corresponding to toxicity and time to steady state and equilibrium, and how activity may assist in the assignment of toxic MoAs was addressed. If successful, the activity concept or hypothesis could be applied in the regulatory process as a ‘weight of evidence’ component for toxicity evaluation and eventually applied predictively to reduce the number and cost of acute and chronic toxicity studies and animal usage in a regulatory context.