After coming up with a hypothesis on heat resistance, I needed to find if and how it changes. In this post, I will describe the details and methods as well as present the results. If you've not read the first part, I encourage you to read that first.

Is it possible that coffee lipids affect resistance?
There are many factors influencing bed resistance. The most obvious is a grind size - the burr aperture defining average particle size. Finer grind setting increases resistance.

The temperature of the beans affects the Particle Size Distribution (PSD). For a given burr aperture you get different results with different bean temperatures. Cold beans when ground result in smaller fines, but there's more of them (Uman, 2016). The fluctuations in PSD are very likely to occur in the cafe throughout the day.

Freshly roasted coffee contains more Carbon Dioxide. During the extraction, the CO2 is released from the coffee acting as a counterforce to pressure. This results in increased resistance, forcing the barista to use a courser grind size to offset the effect. Well-degassed coffee allows you to grind finer and thus increase your extraction.

Interestingly, a similar effect occurs when the brewing water already contains Carbon Dioxide (Wellinger, 2017). It will depend on your filtration system. I've dedicated an entire article on the subject - read it here.

It's important to understand the equipment and its settings as well. Flow restrictors, pre-infusion chambers and filter basket design will affect the resistance. 

Pump pressure could influence it too. Very high pressures will compress the coffee cake to the point it's almost impenetrable to water. Thus lower pressures allow finer grind settings.

There are more factors in bed resistance, such as the loss of volatile compounds (mentioned in the previous post) or environmental factors. However, there's little or no research on that.

This article is limited to the heat (temperature) of ground coffee and its impact on bed resistance. My research has been motivated by finding the thermal role of lipids in ground coffee texture. I've also formed a hypothesis linking it directly to resistance:

As the temperature of the grinds increases and the lipids transform into a liquid state the resistance increases due to that transition.

Equipment and Coffee
The espresso machine used for the testing was La Marzocco GS3. It was set to 9 bar of pressure and 93°C brewing water temperature. It was equipped with a standard 17g La Marzocco basket.

The water was filtered using an ion-exchange cartridge resulting in 160ppm (using conductivity meter). 

The grinder was Victoria Arduino Mythos 1. For some part of the testing, the fans were covered to induce overheating. The hopper was filled with at least 700g and no more than 1000g of coffee during the testing for consistent pressure in the feeding chute. 

For the most part, the grinder was switched on with heating element plugged in at least 20 min before using (to allow it to preheat). However, part of the data was obtained using a cold grinder with the element unplugged.

Prior to data collection, the grinder was dialled in using 16.5g dose to 35g yield and 35s of contact time for each dataset. The grind setting and grinding time were not adjusted during the tests.

3 different coffees were used:

1. Guatemala, Bourbon variety, Natural process
Days after roast: 171
Roast: medium for espresso, 12 min
Agtron: Whole bean - 40.9, Ground - 47.2

2. Blend (post-roast): 
- Brazil, Yellow Bourbon, Natural
- El Salvador, Bourbon, Natural
Days after roast: 196
Roast: medium for espresso, 12 min (both components), Agtron: Whole bean - 41, Ground - 47.3

3. Brazil, Yellow Bourbon, Natural
Days after roast: 12
Roast: medium for espresso, fast start, 11:30 min
Agtron: Whole bean - 42.9, Ground - 47.1

I've used very fresh and relatively old coffee to test whether CO2 influences the path of resistance. Analysing the data I didn't notice a difference.

How did I measure the heat and why?
With this particular setup, I've noticed some interesting things.
The frictional heat per 16.5g dose adds about 10°C of heat to coffee. When the grinder is cold and the burrs are at room temperature, beans at 25°C will result in 35°C grounds. Also, the temperature will be uniform across the ground dose - that changes as you grind more coffee.

As the grinder is used, the heat will be also absorbed by the burrs. These, in turn, will heat the grounds as they pass the grinding chamber. This plus the frictional heat result in much higher ground coffee T than bean T. Also, the difference between the first few grams and the last few will increase. I call this a thermal gradient.

The gradient averages at 10°C across many different temperature points. However, it gets closer to 15°C if the grinder is not used for more than 2 min - I suppose the coffee inside the grinding chamber absorbs more heat.

Therefore I decided to use the highest reading (the very last gram ground) from the top of the grounds pile in the basket as a Temperature reading. I think it's safe to assume the average dose T is 5°C lower (based on a gradient of 10°C).

The measurements were taken using an infrared thermometer. It's important to remember that temperature readings are difficult to compare. That is due to the instruments, their placement, technique or calibration. Rather than focusing on strict values, I'd encourage you to look at the intensity and direction of change.

Test: how temperature (T) affects resistance

First, I wanted to find the point of increasing resistance. I define the resistance increase as an increase in contact time since all other variables stay constant. In my dataset, it happens around 45°C. This corresponds with the research quoted in the first part (Illy, 2005). The dataset could be statistically described using metrics. For temperature (T):

The T values range from 36 to 75, with the vast majority falling into the 50-55 range. This temperature spectrum is easy to maintain with the Mythos. The cooling system is highly effective when pulling 1 shot every 2 minutes. To reach higher T, I had to inhibit the fans and grind large amounts of coffee prior to data collection.

And contact time (t):

Secondly, I wanted to track the change across the temperatures. Including very low values of T - when the grinder is cold, as well as very high values when the grinder is overheated. You can see the data-based trend line below:

In the background - average values for each T value. In the front - trend line as a polynomial function.

This trend line is extremely important. Michael Cameron has drawn the first part up to the first local maximum (Cameron, 2016). So that only confirms his findings but also shows what happens above 60°C.

Initially, the resistance is relatively high. As the grinder warms up and the grounds temperature increases, the resistance goes down. It plateaus between 40°C and 45°C and then increases, reaching a peak around 60°C. Temperatures above 60°C result in decreasing resistance. Note that this temperature is only possible to reach with Mythos 1 when pushing it above its capacity.

Each coffee produced a very similar path of resistance. The graph above combines all 3 for the maximum range representation.

Grinder's thermal energy
Why grinders produce so much heat and why you should manage it? Fast grinding for espresso requires a lot of energy. Much of that energy is transformed into heat. Resistance and extraction are linked to thermal energy. If not for the continuous change in grind profile, espresso would be super-consistent. 2 espressos can taste differently even at the same contact time. Usually though, controlling that variable is challenging and never stops.

Frictional heat is correlated with the grinding time. I've explored the effect of grinding time on the temperature in this post from 2017. In short, higher doses take more time to grind and this contributes greatly to overheating. Finer grind setting requires longer grinding time, which produces more heat. But I'm not sure if the aperture of the burrs has a thermal impact in itself.

As more coffee is ground, the grinder is progressively getting hotter. The temperature in the grinding chamber has a big impact on the final ground coffee temperature. The heat is transferred to coffee as it passes the chamber.

It's not only friction but also the thermal energy of the grinder causing grounds temperature to increase.

If you use Mythos, the chamber is kept at around 35°C. That could mean that you never get grounds below 40 degrees. I'd argue that's a good thing. It helps avoid the initial decrease of the resistance, allowing the barista to make fewer grind changes. Also, the temperature of extraction is kept within a smaller range.

You could keep the heating element unplugged. It will slow the approach towards increasing resistance and overheating.
But even then it will quickly climb up to 50°C and above due to frictional heat. Then fans will try to keep it below 55°C.

Bean temperature and volatiles
I used only room temperature beans. On hotter days you might approach the increasing resistance point sooner. But the effect of bean temperature on final grounds temperature is minimal. The grinder's heat and frictional heat have much more influence.

What's noticeable though is the change in Particle Size Distribution. Through PSD, resistance depends on bean temperature. Different bean temperature could produce a similar path of resistance, but the changes in the trend might occur at different temperature points.

I suppose part of the decrease in the resistance after 60°C mark can be attributed to beans warming up from the overheating grinder. It would be worth to do further testing isolating that as a variable or putting the probe in the hopper to measure the effect.

Other than that I could only think of the loss of volatiles. Which can occur already in the grinding chamber, where the heat would catalyse the process.

Conclusions and limitations
I managed to establish how the resistance changes across the temperatures. I couldn't confirm whether the lipids are responsible for the increase in resistance around 40°C. But in line with the previous research on lipids becoming fluid around 40°C, there's a strong correlation.

Why would lipids influence that is still unclear to me. 

Aside from surface texture and PSD, escaping volatiles need consideration too.

This research is only limited to particular equipment with particular settings. The path of resistance is likely to vary with different setups, but I'm certain it would follow similar trends. Isolating the bean temperature seems to be a natural progression in order to confirm the shape of the resistance curve past 60°C mark.

The coffees used are very similar in many regards, therefore I'd be cautious applying the findings to very different beans.

Nevertheless, we learned what impact the temperature has on bed resistance allowing us to maintain consistency.

If you've managed to read both parts in full - Thank you! It's not an easy read and I appreciate your effort. I would also like to thank Alchemy Coffee for allowing me to use the space, equipment and coffee for the research.

1. Illy A., Viani R, (2005) Espresso Coffee: The Science Of Quality, Second Edition; Grinding p.228
3. Uman, E., Colonna-Dashwood, M., Colonna-Dashwood, L. et al. The effect of bean origin and temperature on grinding roasted coffee.Sci Rep 6, 24483 (2016).
4. Wellinger M., Smrke S., Yeretzian C., Water for Extraction – Composition, Recommendations, and Treatment, The Craft and Science Of Coffee, p. 395

Title Photo by Timo Müller on Unsplash
Mythos Photo by Yan Khanafi on Unsplash Mythos


Post a Comment