Much of the confusion surrounding the chemistry of chloro alkanes can be remedied by a better understanding of how we test for them in environmental/ product samples. One of thekey interests of the chlorinated alkane (CA) sector group is to educate people on this so that decision makers, academics and the general public are informed on these essential, versatile chemicals.
Key points of caution to understand when analysing materials for chlorinated alkanes:
- It is not currently possible to accurately quantify individual CA components;
- State-of-the-art CA analysis techniques can qualitatively identify groups of CA isomers by carbon chain length and chlorination level, although this remains difficult;
- Iinstrument detector response is affected by both chlorine content and distribution of chlorine atoms on the carbon chain;
- Reference substances matching commercial products are currently unavailable. These are necessary for more accurate individual congener quantification;
- Caution should be exercised when interpreting published data on environmental/ product CA content;
- OECD tests involving CAs, often used in regulatory assessments, should be carefully conducted as their incorrect usage may lead to incorrect assumptions.
The chemical composition of chlorinated alkanes.
A commercial CA product contains thousands of individual CA congeners (chemical components). These vary in carbon chain length (depending on the feedstock used) and number of chlorine atoms (which depends on the degree of chlorination). Substances displaying such a complex composition are defined under REACh as UVCB substances (substances of unknown or variable composition, complex reaction products or biological materials). As CAs are manufactured by the chlorination of linear alkanes, a 'normal distribution' of congeners with different numbers and distribution patterns of bound chlorine atoms is present (which defines the product properties). This chlorination distribution proceeds according to specific thermodynamic rules. These include chlorination preferentially taking place on a –CH2- group rather than a terminal CH3 group, chlorination being less likely to occur on carbon atoms that have already been chlorinated, or that are adjacent to chlorinated carbon atoms.
The result of these general, expected, ‘rules of thumb’ is that the chlorine atoms spread along the alkane chain. As more chlorine is added, fewer 'options' for chlorination according to this pattern remain leading to the formation of more densely chlorinated clusters. The most abundant congener in the normal distribution curve is the one that is labelled on the product.
Understanding state-of-the-art detection, analytical methods and their limitations.
State-of-the-art CA detection uses 2-dimensional gas chromatography combined with electron capture detection (GCxGC-ECD). The GCxGC separation method is able to qualitatively identify groups of CA isomers by carbon chain length and chlorination level, although this is very difficult due to the complex nature of CAs. Advantages of this technique include the detection of lower chlorinated congeners, the high separation power of congeners with different chlorination levels and the ability to detect groups of congeners with equal chlorine levels.
However, the most commonly used method of detection and quantification is either high or low definition gas chromatography followed by electron capture negative ion mass spectrometry (GC-ECNI-MS). Whilst popular, this method has difficulty in accurately separating different congeners with the same chlorine number, and the detection of congeners containing low numbers of chlorine atoms (≤ Cl5). Whilst limited, it can provide valuable information on higher chlorinated congeners to complement analysis by GCxGC-ECD.
There has been a particularly comprehensive review of the current analytical situation in Chemosphere 136, published by van Mourik et al. (2015).
Further limitations in CA analytical methods.
As implied above, there are difficulties with the accurate quantification of CA components. This mostly comes from a need to correct any results obtained, caused by different ammounts and distribution patterns of chlorine atoms (as described above). Existing individual congener standards are not suitable for quantification as they do not match those congeners found in commercially available CA products. Without reference substances that match commercial products, only semi-quantitative analysis of CA components is possible. For both of the detection methods described above, the response of the CA congeners increases as chlorination level goes up. This leads to an over-estimation during the detection of higher chlorinated congeners (i.e. greater numbers of chlorine enhance the signal out of proportion with the increasing molecular weight).
Other limitations include:
- low singal caused by CAs in detection systems;
- CAs large range of octanol-water coefficients (larger than any other organochlorine substance);
- interference by other compounds such as PCBs, insecticides and pesticides;
- lack of consistency in inter-laboratory studies (one article reports over 137% variation in results during an analysis of just one environmental sample!)
These limitations create a source of errors when people analyse CA data. They often incorrectly state that a certain sample contains SCCPs, when actually they mean that the sample contains shorter chain CA constituents of a defined length (e.g. C10 or C11 or C12). This error is often amplified by international differences in the definition of the range of lengths for CAs (SCCPs and MCCPs). In China, for example, CA products are mainly identified by their level of chlorination whilst the carbon chain length may vary from C10 to C20.
Other problems surrounding CAs
CA persistence and bioaccumulation in the environment is often tested using OECD tests. In a paper by Honti and Fenner (ES&T, 2015), the problems surrounding the current OECD 308 tests were described. This very expensive (over 100,000 euros per compound) but necessary (required for many regulatory assessments) test is often criticised for its lack of environmental relevance, non-specific rules on experiment vessel 'shape' (which can affect the ratio between the sediment and water and thus can bias by upto 40% those chemicals which are transformed faster in one media or the other) and lack of clarity on how to analyse any data produced. In order to make this test more applicable, Honti and Fenner argue that by involving additional data sets and by combining any obtained data with results from stirred systems, the OECD 308 data can be given additional ‘value’. As OECD testing also allows for minor modification for highly hydrophobic chemicals (like CAs), all concerned parties should be made aware of this essential information.
In light of regulatory pressure, the CASG is supporting efforts to help clarify the understanding of this test so that any data produced from such methods can be applied with scientific rigour during regulatory processes.