Greetings, coffee enthusiasts!

In this post, I'm going to go into details about the "Mini Caffeine" project, including the goals, objectives, experimental design and the outcome of this little study. I want to keep it as transparent as I can, so that you can decide if you want to participate in this endeavor! I'll try as hard as I can to pitch it to you professionally, but I will not deny that I'm very inexperienced in doing this kind of thing and there is a personal burning enthusiasm regarding this topic. I also want it to be as clear and understandable for general audience as I can, so I will simplify things where I can.

It is intended to be a small crowdfunded coffee study with a primary goal of developing a quick and fast method of estimating caffeine concentration in brewed coffee samples with UV-spectrophotometry. I want these brewed coffee samples to be reflective of what people drink in day-to-day life, and for this study I want to compare caffeine concentrations from pourover coffee vs soup-style coffee, i.e. percolation vs puck percolation.

This project/study is NOT about precision, but about estimation. This is a small study that will NOT be HPLC-validated (compared to something like a Lighttells CA-700 caffeine analyzer), rather, the funding will allow me to buy pure caffeine samples and use it as a reference for estimation in this method.

I would like to plot estimated caffeine concentration vs estimated extraction yield of coffee for pourover (brewed on V60) and soup-style coffee (brewed on Oxo Rapid Brewer) for three different coffee beans (representing America, Africa and Asia, for example, Colombia, Ethiopia and Indonesia, preferably washed process) at two roast levels (filter/espresso or light/medium). This will allow us to get more evidence towards making a coffee consumption guideline in the future with more rigorous testings and replications from fellow researchers and coffee enthusiasts.

My goal is to collect 100000 rubles (or ~1333 USD) to start this project.

50% of these funds will go towards taxes, service commissions and into my personal account to pay the cost of my time and effort;

35% will go towards buying coffee beans, lab consumables, reagents and just about anything related to this project.

15% will go towards lab consumables for my laboratory and colleagues (because I'm not the one in charge of my workplace, and it would be very rude and unethical to use my workplace to research something that is outside the scope of the laboratory without giving anything back) 

I will track all expenses and spendings and will share it publicly to keep it all transparent.

Due to this project being a small study with small number of objectives, I expect this study to be done in 1-2 months, starting from the moment when we reach our funding goal, with a regular update each week. There might be some delays due to order delivery of coffee, reagents or lab consumables, and I will tell you about it if it happens.

I will be brewing coffee with a V60 and Oxo Rapid Brewer and comparing estimated caffeine in pourover brews vs brews at the soup regime (10-30 seconds brews depending on the ratio). Coffee doses will be kept at constant 15g dose. Grind size will be kept constant as well for every brew and the extraction yield will only be tuned by adjusting the coffee to brewing water ratio from 1:3 up to 1:20 ratio. The brews will be done either to a final cup with different extraction yields or in a salami-shot style with a subsequent mixing in correct proportions to emulate the different extraction yields. 

All of the brews will be done with the water straight off-boil.

For pourovers I will do 30g bloom, wait until 01:00, and pour the rest of water in circling motion in the centre.

For soup-style I will pour the water, pump until soaked and finish the brew as fast as I can.

I intend to record every brew to keep track of added coffee, added water, bloom time, total brew time etc; all of the recordings and data will be available when the project finishes. If I do include the data from unrecorded brews, I will make sure to note that. Every brew will be at least three times for preliminary statistics. If everything goes well - I will repeat much more to have a more robust evidence.

TDS will be estimated with an analog Brix refractometer (RIP my eyes), using a formula from known linear regression: %TDS = 0.85 x Brix.

Extraction yield will be calculated using the formula for percolation brews: EY% = TDS% x Mass of the brew / Coffee dose

Caffeine concentration will be estimated using a UV/Vis spectrophotometer. Caffeine absorbs UV light at a wavelength of ~272 nm, which allows you to measure the absorbance at this wavelength and estimate its concentration. HOWEVER, there are many other compounds, including chlorogenic acids, trigonelline, which can absorb light at this wavelength as well. It is then necessary to develop a method to exclude these confounding signals for a more accurate analysis. One of which is a post-measurement analysis using a derivative spectrophotometry method, which I will be using in my caffeine analysis. In my preliminary measurements with the instant coffee (traditional vs decaf from the same commercial company), this method works under some conditions, which are to be optimized. The funding will allow me to buy pure caffeine samples and spend time on optimizing these conditions before going into coffee samples.

Lots of it. To start off, the caffeine concentrations will not be HPLC validated, which means that it might not represent the actual content in the coffee samples. However, the most important part of this study are the comparisons: we are going to see the trends and differences between the group.

Then, I'm using an analog Brix refractometer, which is prone to human error. The measurements can only be as accurate as my eyes and hands are, so there's that. Human error also include my brewing technique, errors in sample handling etc. 

This study is limited to the coffee beans, brewers and the recipes that's in the experimental design, so no conclusion about whether longer steeping time or any other variables affect the results can be derived from this. All of the data will be available publicly when the project finishes, so maybe some secondary conclusions can be made by experienced statisticians and researchers.

If this project succeeds, I hope to try to reach a bigger goal and collect more fundings for the next caffeine project, in which I hope to upgrade my refractometer, collaborate with my colleagues to validate the measurements with their HPLC and work with the roasters to get more diverse samples. But it can start with smaller steps, e.g. studying the effect of different other variables on caffeine content.

All of the data will be available for anyone on YouTube or Boosty or Google Documents/Sheets. 

I will NOT use or publish any data obtained from this project for my academic purposes or for my job, as it is intended to be for the public without involving publishing in the journals etc. (also the small scope of this project is most likely unpublishable, and the open access publishing costs your bank account and your soul, so nope-nope-nope).

Hopefully, we will be able to collect the funds needed for this "Mini Caffeine" project. I will be collecting the funding from June 5th, 2026 to July 5th, 2026. Even if we don't reach the goal of 100000 rubles (~1333 USD), I will still proceed with the project in a smaller format, if at least 33% is reached. If less than 33% is reached, this money will go towards my YouTube channel and coffee expenses.

I'm very excited for this project and hope it will become a reality! Have a nice day!