I'm a high school student who decided to investigate the relationship between caffeine concentration and brewing temperature of coffee out of their love for the substance. For my method, I opted for iodometric back titration, as detailed in this paper and shown in the image below. I've verified the method is viable with a known concentration of caffeine obtained by using caffeine tablets, but I am struggling to repeat the results on brewed coffee.

How the titration should look like


When performing iodometric back titration on brewed coffee (as opposed to pure caffeine solution) with starch as an indicator, the end point isn't clearly defined because it is no longer colorless due to the natural brown pigments in brewed coffee.

I've looked into several potential routes that could help me mitigate the issue (detailed below), but I'm looking for advice on the most convenient (requiring the least time, resources, and skills) way to proceed. Of course, I'd like to maintain a certain level of quality, but it isn't as vital.

Paths I've considered

1. Titration with spectroscopy
Carrying on with titration as previously, but using a colorimeter to detect the end point. From my limited understanding of the technology, I presume that I would be able to detect when the concentration of a certain compound (e.g. iodine-starch complex) would reach zero. A limitation I can foresee is that I would need to test a sample of the solution every time I add a drop of the titrant, which would become very time-consuming on many trials.

2. Extracting caffeine
Using an organic solvent to isolate caffeine and then perform the titration on that substance. The problem is that I'm not sure if the titration reaction (described in the paper linked above) will still hold. Furthermore, I am limited by the availability of organic solvents (can't use chloroform and likely not DCM either). Some other solvents are addressed in this question.

3. Filtrating the color of coffee away
Practically the inverse of the above, using a filter (activated carbon) to produce colorless liquid by removing pigments (melanoidins?). The problem is that information on this is quite limited when it comes to coffee and from my research both caffeine and iodine would be adsorbed by activated carbon. Perhaps I could mitigate this by pre-saturating the activated carbon with caffeine or iodine?

4. Switching to green tea
My supervisor actually advised this in the beginning, but I insisted on coffee (I know, I know). I wouldn't prefer this path because of the costs already sunk into preparing coffee samples, but if there is no other alternative...

All in all, these are just some of the possibilities I've thought of, but I am sure there are some missing. What would be your advice on how to proceed with iodometric back titration, considering the limitations in time, resources and skills?

  • $\begingroup$ Maybe take an aliquot after the precipitation reaction and dilute several times before titration to make the solution translucent enough to be able to see the blue-colourless change. $\endgroup$
    – Tom P
    May 19, 2023 at 12:38

1 Answer 1


Classical volumetric titrations are not selective. This means if there are multiple components in the mixture, they will react with the reagents used in the titration. Since coffee and tea are real-world samples with hundreds of other compounds besides caffeine, there will be other oxidizable materials, which may/will react with iodine. One would suspect that tannins would also be oxidized by iodine. Your results will always be overestimated. What is the solution? A real sample titration requires sample preparation. Ask your teacher to help with extracting caffeine from tea or coffee with a suitable solvent in a separatory funnel. After extraction, you will have to evaporate the solvent. Those colors would not be extracted in a non-polar solvents. I will let you research more on what solvents starting from here Caffeine extraction. After that one may proceed normally as suggested in the Olympiad manual.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.