The Chemistry of Treblinka

A few months ago, British forensic archeologist Dr. Caroline Sturdy Colls finally published the long-awaited results of her detailed forensic studies on and around the areas of the former German wartime camps Treblinka I (labor camp) and Treblinka II (alleged extermination camp). This issue of Inconvenient History features two reviews[1] of the book she published on this, titled Finding Treblinka, which has some 350 pages of text.[2] On pages 261 and 263 of her book, we read the following passage:

“Although it seemed highly likely at this point that the discovered masonry belonged to the Old Gas Chambers, samples of them were sent to the Nicolaus Copernicus University in Toruń for further tests. Thermal Desorption Spectroscopy (TDS), which measures the absorption of molecules to a surface, revealed that the bricks and mortar had a higher emission rate of carbon monoxide compared to the reference samples tested, which ‘may indicate the presence of chemically bound CO [carbon monoxide],’ as opposed to carbon monoxide naturally absorbed from the environment.⁷⁸ Therefore, it may indicate exposure to carbon monoxide, con­sistent with the descriptions provided by witnesses about how the gas chambers functioned. The researchers noted that a larger sample size was needed to offer a more detailed profile of the exposure. Excavations being undertaken by the Warsaw University of Science and Technology in 2024-25 may offer the possibility for such examinations.”

Her footnote 78 on page 426 states:

“FN 78: Uniwersytet Mikołaja Kopernika w Toruniu, “Badania Konserwatorskie Ceglanych i Betonowych Obiektów z Dawnego Obozu Pracy I Zagłady w Treblince,” unpublished report, 2013; Caroline Sturdy Colls and Kevin Colls, Finding Treblinka: Archaeological Evaluation, Project no. P13-02 (Centre of Archaeology, Staffordshire University, 2013), 22 and appendix 3.”

Neither of these two sources are publicly accessible, since the Collses’ paper titled “Finding Treblinka: Archaeological Evaluation” is only an internal archaeological site evaluation and project report. It is therefore impossible to determine what exactly was analyzed by the unnamed team at the Nicolaus Copernicus University in Thorn/Torun, West Prussia, now Poland, and how.

Dr. Sturdy Colls referred to the results of this analytical study already in earlier years.[3] A short while after these earlier brief references were published, Italian historian Carlo Mattogno approached me, pointed out these sources, and asked whether there is any known chemical process binding carbon monoxide firmly in brick and/or mortar, making it detectable after some 70 years of exposure to the elements. My gut reaction to this was: impossible. Not wanting to rely on gut reactions, I did some investigation into the chemistry of carbon monoxide, especially with iron, as I remembered from my university time that carbon monoxide has quite a strong affinity to bind with iron ions. While this turned out to be true, there is no known mechanism to form any long-term-stable compounds similar to those that iron forms with cyanide ions (hexacyanoferrates), which can survive decades.

Transition-metal carbonyls – such as iron pentacarbonyl (Fe(CO)₅) – are notoriously volatile, liquid or gaseous at room temperature, and highly unstable in the presence of air, moisture and light. They rapidly decompose and cannot persist for decades in an open environment or buried soil. While carbon monoxide can form transient compounds upon interacting with iron oxide under pristine laboratory vacuum conditions, these species are highly reactive intermediates. Exposed to humidity, oxygen and soil microbes over 70 years,[4] these loose complexes break down quickly or oxidize almost instantly, forming carbon dioxide in the process.

Because carbon monoxide does not form stable chemical compounds with iron or other components in building materials, the elevated carbon monoxide peaks recorded by Thermal Programmed Desorption (TPD) in forensic archaeological reports cannot be explained by ancient exposure to carbon monoxide in exhaust gas. I even doubt that this is what the researchers at the Nicolaus Copernicus University claim in their unpublished report. Dr. Sturdy Colls may have misunderstood what they did.

When the Torun researchers detected elevated carbon monoxide during TPD, they most certainly were not liberating intact, ancient gas molecules that had been sitting trapped in a film on the surface. Instead, the TPD process heated the brick samples to high temperatures, potentially triggering chemical reactions from permanent, stable carbon sources, such as soot and polycyclic aromatic hydrocarbons (PAHs), partially oxidizing them to carbon monoxide and carbon dioxide.[5] In our case, soot and PAHs in the recovered building material could originate from two sources:

1. A fire that burned down the specific building, containing construction wood. Since the Treblinka Camp experienced an inmate uprising on the 2nd of August 1943,[6] which led to some if not most of the camp’s buildings being burned to the ground, that could explain the deposition of significant amounts of soot.
2. Extended use of engine exhaust, especially if derived from a Diesel engine running under heavy load and/or with severe restrictions to its air intake, as would have been required for diesel executions to work at all.[7] The use of gasoline-engine exhaust, which tends less to the formation of soot and PAHs, would have led to less carbon deposits.

In fact, there is a possible third source: If the brick analyzed was fired in a kiln furnace heated by coal/coke, which is rather likely, then this would have deposited deceptive fossil soot from coal/coke uniformly throughout the brick. Getting a carbon baseline for such a brick would require a brick from the same or a similar kiln with a history clear of any other exposure to carbon contamination.

One test to verify whether such TPD-exuded carbon monoxide in excess of any baseline is from engine-exhaust soot/PAHs or from soot/PAHs stemming from the building burning down, would be the detection of any residual radioactive Carbon 14 (¹⁴C). ¹⁴C has a half-life of roughly 5,730 years. It is constantly created in the upper atmosphere, induced by violent nuclear reactions caused by high-energy cosmic radiation. As a result, the concentration of ¹⁴C in the biosphere is rather stable (roughly one part per trillion).

Therefore, lumber from recently felled trees used for a building has exactly that amount of ¹⁴C. Being radioactive, the amount can be detected rather reliably. Soot and PAHs derived from fossil fuel, on the other hand, have been excluded from the biological carbon cycle for millions of years, enough to make all ¹⁴C decay. Hence, soot/PAHs derived from fossil fuel has no ¹⁴C at all.

The trick is to test the carbon gases (monoxide and dioxide) coming out of those samples during TPD-testing for any ¹⁴C-induced radioactivity (beta radiation of 156 KeV). If there is none (“dead carbon”), it is from fossil fuel. If there is any, it is at least partially from recently alive carbon sources (burned lumber). The exact amount of ¹⁴C could even determine the ratio of the two (radiocarbon-based source apportionment of black carbon). If it sits at one part per trillion, or the radioactivity level expected for living material (biomass), then there is no engine-exhaust-derived carbon deposit detectable. However, this probably requires more sample material than is available.

Since TPD is a destructive sledge-hammer method of releasing any carbon deposit in porous building materials, other better analytical methods would be as follows:

1. Gas Chromatography & Mass Spectroscopy

It would have been smarter to grind up the samples and subject them to a combined gas-chromatography mass-spectroscopy analysis (GC-MS), which can detect the type of carbon present in the samples, making it possible to distinguish between soot/PAHs deposits from wood fires or fossil-fuel fires:

• Certain monosaccharide derivatives (levoglucosan, mannosan, and galactosan) are created only via the high-temperature pyrolysis and incomplete combustion of cellulose and hemicellulose (wood and plants). If GC-MS detects robust levels of them, the soot is definitively biomass-derived (e.g., from a structural wood fire). It cannot be produced by burning gasoline or diesel fuels, which lack cellular plant walls.[8]
• On the other hand, certain highly durable, complex molecular hydrocarbons (“chemical fossils” such as hopanes and steranes) present in crude oil and refined petroleum products are an unambiguous signature of fossil-fuel combustion.[9]
• Alternative, a combination of GC-MS and ¹⁴C detection can be used as well.[10]

2. Inductively Coupled Plasma & Mass Spectrometry

In contrast to wood fires of brick kilns, engines always deposit metallic byproducts alongside soot. Inductively coupled plasma analysis with subsequent mass spectrometry (ICP-MS) can detect inorganic trace elements mixed into the carbon:

• Engine exhaust soot typically contains elevated markers from engine wear, spark plug erosion, and fuel/lubricant additives utilized in 1940s motor mechanics, such as lead, zinc, phosphorus, barium and copper. Especially indicative would be any lead traces, as lead additives were strictly limited to gasoline fuels, but never added to diesel fuels. Hence, elevated lead concentrations would indicate exposure to gasoline-engine exhaust, whereas the lack of it would exclude them.
• Wood-fire soot is characterized by the complete lack of elevated heavy-metal traces, but would contain elevated levels of alkaline and earth-alkaline metals, such as potassium, calcium and magnesium.

Other methods detecting structural and optical features that are different for soot from engine exhaust versus that originating from wood fires could be added to the list, such as Raman Spectroscopy and Optical Absorption Fingerprinting. I will not go into more details here, as I probably have already overstretched my reader’s patience with the details provided.

While I have a PhD-level background as a chemist, I am not an expert on any of the issues addressed here. However, applying artificial intelligence on this particular problem quickly yields some of the results laid out here. Caution needs to be practiced when using AI, however. I always chase down any sources given by the AI answers to make sure they back up the claims made, which they sometimes do not do. In the present case, AI initially mixed up carbon monoxide and carbon dioxide, which is an egregious error. It also made some untenable claims to make Sturdy Colls’s study look good, which it later retracted, after I had cornered it step by step with its own answers backed up by reliable sources.

While it is true that those AI resources were not at Dr. Sturdy Colls’s disposal in 2013 when submitting her samples to the unnamed research team at Nicolaus Copernicus University, such an excuse is no longer valid. Anyone can obtain the above results from any major AI engine, exposing the technique used by the Torun researchers as amateurish.

After chasing Google AI around for some 6 hours, I finally had it compile a list of possible analytical methods, together with sample requirement and possible limitations, that can be applied to the challenge at hand. They can be used as a “methodological critique” of Sturdy Colls’s flawed analytical approach, as Google AI put it:

I had an exchange with another Polish team of scientists in 1994/95, that had done some forensic research on Auschwitz.[11] They, too, exhibited a rather amateurish, if not to say fraudulent attitude. I have no idea what the average educational level of Polish forensic scientists is, but this track record of incompetence combined with an apparent attempt at deceiving the public is deeply concerning.

Anyone attempting in the future to analyze samples of building material recovered from the area of the former camps at Treblinka, Bełżec and/or Sobibór should consider doing more professional, much-more revealing analyses of the type suggested above, so that we can finally get some definitive answers – presuming that they are honestly seeking the truth and not merely trying to confirm some preordained result.

Endnotes