Markiewicz vs Epstein vs Rudolf - Detection Limits
Posted: Fri Nov 07, 2025 4:35 pm
In his 1994 paper, linked below, Markiewicz tells us:
Citing Joseph Epstein's paper: Epstein J., Estimation of Micro quantities of Cyanide, Analytical Chemistry
1947, Vol. 19, p. 272.
He then claims
Lets analyze the original 1947 Epstein paper as published, which was cited and utilized by Markiewicz in his 1994 paper. Link below, along with DOI code for those who wish to retain a copy in their archives:
https://pubs.acs.org/doi/abs/10.1021/ac60004a018
https://doi.org/10.1021/ac60004a018
Epstein starts with "exactly 10 micrograms of cyanide ion per milliliter of water"
This equates to 10 μg/ mL = 10mg/L = 10 ppm
From this stock solution, serial dilutions were prepared a range containing:
0.2 μg/mL = 0.2ppm
0.4 μg/mL = 0.4ppm
0.6 μg/mL = 0.6ppm
0.8 μg/mL = 0.8ppm
1.0 μg/mL = 1.0ppm
1.2 μg/mL = 1.2ppm
Minimum detectable threshold: The Epstein paper explicitly states that "as little as 0.2 micrograms of cyanide can be estimated with an accuracy of 99% ±4%."
This equates to 0.2 μg/mL = 0.2ppm, with the range given as lying at 0.2 - 1.2ppm
Conclusion: Rudolf clearly is right to call into question the detection limits as specified by Markiewicz.
So what gives?
The unfortunate answer is: we don't know. We do not know exactly what Markiewicz did to "enhance" the Epstein method to obtain a degree of sensitivity orders of magnitude greater than demonstrated by the known method, or whether he is simply misapplying, misreading, or misunderstanding the methodology and literature.
What we do know, is that whist somewhat similar, the Markiewicz methodology requires one further preparation step in the sample handling over the Epstein method, this is the step of liberating the cyanide from the carrier material via acidification (this step is not needed under the Epstein method, as the cyanide is already present in its simplest form in the solution, by design).
Per Markiewicz:
Final note: Rudolf reproduces Markiewicz' 1994 results in section 8.3.2 and denotes readings below the established threshold as ND, and represents numerically those readings at or around the established threshold range (0.2ppm - 1.2ppm) using ppm notation, mg/kg
https://codoh.com/media/files/downloads ... 94-ocr.pdfWe decided therefore to determine the cyanide ions using a method that does not induce the breakdown of the composed ferrum cyanide complex (this is the blue under discussion) .....carried out by Epstein's method.
Citing Joseph Epstein's paper: Epstein J., Estimation of Micro quantities of Cyanide, Analytical Chemistry
1947, Vol. 19, p. 272.
He then claims
Germar Rudolf challenges this in his work "The Chemistry of Auschwitz", as follows:Under present circumstances we established the lower limit of determinability of cyanide ions at a level of 3-4 μg CN- in 1 kg of the sample.
So, who is right?The Polish scientist used the microdiffusion-chamber procedure, which does not permit the detection of iron-
cyanide compounds like Iron Blue. The Poles claim that the detection limit for other cyanides lies at a concentration of 3-4 μg per kg sample material, yet the paper they cite clearly gives 0.2 mg/L (200 μg/kg) as a detection limit for aqueous solutions. The samples of the second series were analyzed three times. For the resulting differences see Table 30.
Section 8.2.1
Lets analyze the original 1947 Epstein paper as published, which was cited and utilized by Markiewicz in his 1994 paper. Link below, along with DOI code for those who wish to retain a copy in their archives:
https://pubs.acs.org/doi/abs/10.1021/ac60004a018
https://doi.org/10.1021/ac60004a018
Epstein starts with "exactly 10 micrograms of cyanide ion per milliliter of water"
This equates to 10 μg/ mL = 10mg/L = 10 ppm
From this stock solution, serial dilutions were prepared a range containing:
0.2 μg/mL = 0.2ppm
0.4 μg/mL = 0.4ppm
0.6 μg/mL = 0.6ppm
0.8 μg/mL = 0.8ppm
1.0 μg/mL = 1.0ppm
1.2 μg/mL = 1.2ppm
Minimum detectable threshold: The Epstein paper explicitly states that "as little as 0.2 micrograms of cyanide can be estimated with an accuracy of 99% ±4%."
This equates to 0.2 μg/mL = 0.2ppm, with the range given as lying at 0.2 - 1.2ppm
Conclusion: Rudolf clearly is right to call into question the detection limits as specified by Markiewicz.
So what gives?
The unfortunate answer is: we don't know. We do not know exactly what Markiewicz did to "enhance" the Epstein method to obtain a degree of sensitivity orders of magnitude greater than demonstrated by the known method, or whether he is simply misapplying, misreading, or misunderstanding the methodology and literature.
What we do know, is that whist somewhat similar, the Markiewicz methodology requires one further preparation step in the sample handling over the Epstein method, this is the step of liberating the cyanide from the carrier material via acidification (this step is not needed under the Epstein method, as the cyanide is already present in its simplest form in the solution, by design).
Per Markiewicz:
What we are left with is, not only does Markiewicz misstate (or overstate) his sensitivity, and not only does he fail to explain how this was achieved by orders of magnitude, he introduces additional complexity to the methodology requiring the samples to undergo extraction, with an unknown impact on the qualitative and quantitative results, along with the reliability of the chosen method.The sample under examination was placed in the internal part of the chamber and next acidified with 10% sulfuric acid solution and allowed to remain at room temperature (about 20°C) for 24 hrs. The separated hydrogen cyanide underwent a quantitative absorption by the lye solution present in the outer part of the chamber.
Final note: Rudolf reproduces Markiewicz' 1994 results in section 8.3.2 and denotes readings below the established threshold as ND, and represents numerically those readings at or around the established threshold range (0.2ppm - 1.2ppm) using ppm notation, mg/kg