Coffee Data Science
Dirty Refractometers and Coffee
Smudging the sensor
Smudge is weird and strange. Face smudge is organic and degrades over time. Coffee smudge dries but also will dissolve again in water. So how does smudge and dirt affect refractive index?
This is an important question as refractometers are used widely in the coffee industry to evaluate extraction efficiency. Refractometers can be used to measure the Total Dissolved Solids (TDS) of coffee, and this can be used to determine the efficiency of an extraction technique.
Some suggest a protocol of using an alcohol wipe to clean the sensor between uses. I prefer wiping with a paper towel or microfiber cloth to where I can’t see anything visually. However, I don’t have data, so let’s create some smudge and see how the sensor does!
Data Types
Finger print smudge is pretty easy to replicate and see.
Then you can add a sample. I used sugar water because that is the cleanest in terms of measuring refractive index. If dirt doesn’t affect the sugar water reading much, it should not affect coffee much because coffee has more noise.
I also tried olive oil, which is an extreme example of what could be in coffee that modifies measurement. However, this was too much olive oil. There is almost no usable reading with too much oil.
I also used espresso, and I dabbed it to leave a messy stain that would not be possible with a good wipe. So then I also wiped this, but not as cleanly to make sure it would leave some residue.
TDS Measurements
I took 10 consecutive samples for each, and there was little movement except for the dried coffee stain which changed probably as the solids on the sensor dissolved in the liquid.
All of these methods showed little movement in the TDS reading.
In terms of refractive index, there was a slight change, but you have to make sure you look at the scale. The scale has a range of 0.00050.
When looking at filtered water instead of sugar water, and the effect was a bit larger.
The key in any protocol is consistency. So if you wipe the sensor or use an alcohol wipe, be consistent. I would also suggest to visually inspect the sensor. If you can see a finger print, then the sensor isn’t clean enough. I actually thought the impact would be larger, but for espresso strength, the impact was much less.
It is interesting that the impact was larger for filtered water in terms of absolut, but the raw difference was similar to the sugar test with about a 0.1% difference.
If you like, follow me on Twitter, YouTube, and Instagram where I post videos of espresso shots on different machines and espresso related stuff. You can also find me on LinkedIn. You can also follow me on Medium and Subscribe.
