Image credit: EU Today.
The Salisbury novichok poisonings are a real can of worms. Media coverage is obviously politically, rather than scientifically, driven. Social and mass media reporting is highly partisan and the scientific components and reports (which are mostly classified) can become slaves to the particular political masters. I find the whole drama a mystery and certainly do not want to tie myself to any of the conspiracy theories, official or otherwise, that are floating around. It’s probably a subject to keep well away from.
However, one aspect intrigues me – the claimed identification of novichock residues in the London hotel room used by the Russian duo, Alexander Petrov and Ruslan Boshirov. In particular, is the identification of the material reliable and, further, is the reporting of this identification factual and reliable?
Media reporting: This generally assumes a positive identification, although at trace levels. The Sun, for example, reported:
“Petrov and Boshirov stayed in the City Stay Hotel in Bow, East London, during their time in the UK.
Cops searching their room two months later on May 4 are said to have discovered minute traces of Novichok, a high-grade military nerve agent created by Soviet scientists.”
“police found traces of Novichok in the hotel room in which the pair stayed for two nights.”
Similarly, the Independent reported:
“Investigators later found traces of novichok in their room at the City Stay Hotel.
They said the amount was too low to present a health risk but are appealing for any hotel guests who stayed there between 4 March and 4 May to contact investigators.”
Since Petrov and Boshirov surfaced and were interviewed the media coverage has become even more partisan and the discovery of these traces of novichock is being portrayed as even more definite.
The police report: In the absence of an official scientific report of the analyses this is the best we have to go on:
“On 4 May 2018, tests were carried out in the hotel room where the suspects had stayed. A number of samples were tested at DSTL at Porton Down. Two swabs showed contamination of Novichok at levels below that which would cause concern for public health. A decision was made to take further samples from the room as a precautionary measure, including in the same areas originally tested, and all results came back negative. We believe the first process of taking swabs removed the contamination, so low were the traces of Novichok in the room.
Following these tests, experts deemed the room was safe and that it posed no risk to the public.”
This raises more questions, for the scientifically inclined, than the answers, seemingly, provided:
- How many samples were taken – 2 positives is probably a low proportion of the total measurements?
- Where were the sample sites located in the room
- How do the low levels reported compare with the detection limits for the methods used?
- Was the decision to take further samples based on lack of confidence in the results form the first sampling?
- Again, how many further samples were taken and from what sites in the room?
I suspect that the two positive detections were probably false positives which the analyst had low confidence in. It is likely many samples were taken from the room so that two positives near, or at, the level of detection is not a good result. I suspect experts would challenge this evidence in court.
Absence of evidence is not proof of innocence
I should stress that in questioning the results I am not trying to argue for the innocence of the two guys. After all, a true professional would not have contaminated the hotel room. If the evidence is genuine, though, it may be more suggestive of a non-professional or non-state actor than a professional hitman.
The problem, though, at this stage is that all the other evidence made public is circumstantial and unlikely to stand up in court. The claimed positive detection of novichock-type compounds in the hotel room could be the key to a successful conviction so any doubts should be removed.
The following presents my views on the problems of detecting novichock compounds at low levels and why I think we should not accept the current media reports as positive evidence. A court would have to look very critically at the actual data and detection methods used. At the moment the political and police statements could be expressing far more confidence in the reported findings than is actually warranted by the real evidence.
An Iranian paper from two years ago, Hosseini et al., (2016) provides information on the synthesis, structure and detection of novichock-type compounds. It is probably the most up-to-date information publicly available and its citation is
Hosseini, S. E., Saeidian, H., Amozadeh, A., Naseri, M. T., & Babri, M. (2016). Fragmentation pathways and structural characterization of organophosphorus compounds related to the Chemical Weapons Convention by electron ionization and electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 30(24), 2585–2593.
The paper describes the micro-synthesis of two compounds that are listed under Schedule 2.B.04 of the Chemical Weapons Convention. These are:
- Compound 3: N-[Bis(dimethylamino)methylidene]-P-methylphosphonamidic fluoride, and
- Compound 4: O-alkyl N-[bis(dimethylamino)-
methylidene]-P-methylphosphonamidate Novichok derivatives
The figure shows the chemical structures of these compounds.
The F atom in compound 3 is replaced by an organic group (R) to form the novichok derivative. As this can be either of a wide range of organic groups (the authors list nine different groups for derivatives they synthesised) the novichock-type compounds include a range of different chemicals with differing levels of toxicity.
This is why more official reports on the Salisbury poisonings refer to novichock-type nerve agents and not just novichock.
Before any clever reader decides to use this paper to synthesize their own samples of these or similar compounds I must stress the warning provided by the authors:
“It should be noted that, due to the extreme toxicity of these materials, the separation and purification of CWC-related chemical are very difficult and therefore should be carried out only by a trained professional in an efficient fume cupboard equipped with an active charcoal filtration system.”
Detection of novichok-type compounds
Mass spectrometry methods are used for detection. This involves breaking up the molecules into fragments using an electron ionizer (EI). These molecular fragments are then separated according to mass and charge and the amounts of each detected in a mass spectrometer (MS) to produce an EI-MS spectrum.
Each compound has its own “fingerprint” – a pattern of peaks defined by the mass/charge (m/z) of each molecular fragment and the relative intensity of each peak. The figure below shows the EI-MS “fingerprints” for compound 3 and the O-ethyl derivative of compound 4.
We can see why the detection of a compound relies not only on a single peak but also other characteristic peaks and their relative sizes.
For example, the largest peak (H) at m/z = 71 occurs in both compounds. This is because the molecular fragment (see the chemical structure to the right) responsible for it is produced by ionization of both compounds. So that peak cannot be used alone to differentiate between the two compounds. Identification of a specific compound requires locating all the major characteristic peaks and ensuring their relative intensities are correct.
This is straightforward where the compounds are available at relatively high concentrations and the combination of mass spectroscopy with gas or liquid chromatography helps to remove some of the background chemicals. The ability of UK experts to conclude that the type of novichok used to poison the Skripals is the same as that in the fake scent bottle used by the second victims (Dawn Sturgess and Charlie Rowley) means that they were able to recover samples containing the nerve agent at sufficiently high concentrations.
But, at low concentrations one may simply not be able to find all the characteristic peaks, and identification using just the most intense peaks is not so reliable. For example, compounds 3 and 4 could not have been differentiated at low concentrations if all that could be detected were very small peaks at m/z = 71, 135 and 150. Yet that is the situation when searching for trace levels and one is always conscious that the peaks that are detected could be due to low levels of a completely different compound.
I suspect the description of the two possibly positive samples in the London Hotel as trace levels or “at levels below that which would cause concern for public health” were interpretations driven by “wishful thinking” and exaggerated confidence and not surety. After all, scientists often face such pressures when their political masters are looking for results to fit a preconceived narrative. It is easy to be persuaded in such situations. And it is tempting for both scientists and police to describe their findings in a more confident way when presenting to the media than they would during peer discussions in the laboratory or office.
My suspicions are supported by the fact that the total number samples taken from this hotel room must have been quite large so that makes the reliability of the positive values at such low levels for only two samples quite suspect (although information on locations of sampling sites would help this interpretation).