Some thoughts on brew temperature

I am fully expecting to have this picked to pieces or just generally rubbished, but I thought I would jot down a few thoughts having read a post over at Coffee Aspirations

Let’s treat something going into solution as a chemical reaction (because it is), and let us also assume that this means that increased heat will result in more chemical reactions, or more compounds going into solution. (This comes under the banner of Arrenhius’ Equation I believe, and no I am not going to try and explain it right now!).

So – the hotter my brew water the more I will dissolve from the surface of the ground coffee particles, and this seems to manifest itself to me in the cup. Typically we see it as a bell curve wth cup quality increasing quite suddenly around a certain temperature and dropping off quite radically either side.

Here is a hypothetical espresso extraction, with brew temperature and also exit temperature measured:

(please note this is a very approximate graph – it is just for illustration)

We would expect to see an exit temp lower than the brew temp – energy having been lost through heat loss (radiation, conduction etc) as well as having been spent dissolving lots of wonderful things for us to drink. The shaded section’s surface area should correspond to energy spent.

So if we increase the brew temperature by a small amount (be it 0.1F or 1C) then we are inputting more energy into our little graph:

Question 1 – Do we expect the exit temp to rise by that amount also?

If so – then surely the amount of energy spent (the difference between input and output) would be the same, so therefore would there have been the same reactions occuring regardless? Or would more reactions have occured in the top section of the puck, where the coffee is exposed to the most heat?

Question 2 – Can we compensate? If we are stuck with a lower brew temp can we come close to achieving a similar result by creating a brew time that gives a matching surface area or would the different dose or grind necessary influence the quantity of solubles going into the cup too much?

Other thoughts:

Occaisonally it seems ridiculous that a 0.1F difference in brew temp can change an espresso’s cup quality. Yet if you read about the aromatics present in coffee (and we, as an industry, often talk about the incredible numbers of volatiles in roasted coffee, as if the cells of the bean can barely contain them) you realise that whilst there may be several hundred different ones there are not present in very great quantities. They are capable of having being a character impact odorant (one of the ones that give a product a distinctive taste) at concentrations of parts per million or part per billion.

This being the case then suddenly a subtle shift in chemistry from a subtle shift in temperature could easily mean a big chance in perception. Though perhaps this begs the question of – if I get red fruit notes at 91C but chocolate tones at 93C, why do I not also get a stronger red fruit note at the higher temperature?

Perhaps it is down to the complex relationship between what we taste and the flavours we percieve. Paul Songer briefly spoke to me about how the level of percieved acidity can influence the flavours/aromas you are better able to pick out crediting the brain’s expectation for the difference.

And yet if espresso really is this picky, the cup quality balanced on such a knife edge – how then do people have similar taste experiences with a blend? Perhaps they don’t – I know this is something that is bugging Stephen Morrissey at the moment. Or perhaps the cleverest blends are ones that contain character impact odorants in great abundance, so wherever you hit the shot (within a certain sensible tolerance) you get strong tastes of caramel or whatever they intend.

I also think that the taste of a coffee – its balance of sweetness, bitterness, acidity and on occaison saltiness – is less temperamental. I often find that a coffee will exhibit the same feel and taste in the mouth from brew to brew whilst the flavour/particular aromatics may vary. (This could easily end up spiralling back into earlier thoughts though….)

I seem to have posted more questions that musings, but that is inevitable with any post about espresso these days….

[tags]coffee, espresso, food science, aroma chemistry, taste perception, barista[/tags]

18 Comments

  1. Pingback: michaelbuxton
  2. Pingback: kimrynio
  3. Disclaimer: it’s been a few years since my last chemistry class:
    The relationship between input and output temperature should not be a linear relationship. By raising the input temperature say, 1 degree, the output will not necessarily be 1 degree higher as well, as you must consider the effect of cooling by the outside environment. 92 input, 21 degree environment vs 93 degree input, 21 degree environment. The difference would be small, but it most likely noticeable by the precision methods you tend to use.

    Although you would expect an increase in reactions speed, some reactions (exothermic?) are hindered by heat, and you may find that some of your critical flavor characteristics are coming from the bottom of the puck, thanks to the lower temperature.

    As for question 2: I sense an experiment in the works.

  4. James, thanks for continuing my thought process with more details concerning variables I had not talked about.

    The question now is not only what is the ceiling for total dissolved solids to saturate a droplet of water, but also in timing, and how long this saturation takes.

    Heat can only speed up the process, and given the same contact time, more solubles will be extracted.. but heat can also change volatile components chemically.

    Does the temperature of water also impact the (pardon the expression) “cooking” of the flavor components being extracted? I think so, but have no clue about the specifics.

    A higher temperature, for me, anyway, tends to not lend itself well to highlighting acidic flavor components.. but for some blends, this higher temperature unleaches an amazing sensory experience that one never knew existed.

    Farthermore, what is the corelation between roast profile, roast level, brew temperature, and espresso flavor?

    What of temperature and mouthfeel or texture of the crema?

    Too many variable to wrap one’s head around in one sitting.

    I hope this keeps going, and someone else continues this focusing on another factor.

  5. Barrett – I agree that the output temp is unlikely to have a linear relationship.
    I know that some compounds are less able to dissolve at higher temperature, and often these are surfactants – which begs the question – does a cooler brew give a better crema?

    Jason – I do not think that heat is likely to change a volatile aromatic since it was created at much higher temperatures than brewing, so is pretty thermally stable. You will not change chemically the makeup of any of the odorants in the brewing process (I am fairly confident not anyway!)

    A higher temperature will reduce acidity in the cup, but this is down to how the tongue percieves acidity and how the increased solubles affect the detection of free H ions. Which is something I should probably write an article about.
    As for the point of saturation in water – I don’t know. It is clear from the crema that the brew pressure allows the water to become supersaturated with CO2, releasing it in the environmental pressure creating bubbles and hence our lovely coffee foam. I have read nothing that implies that the increased pressure allows more solids to dissolve though. Also we should probably look at the saturation of our non-polar solvent – the oils – and how brewing affects that.

    Brew temperature is ultimately the search for balance between as full an extraction as possible and try not to go too far and start to take things we don’t really want. Do we really need to know more than that? :)

  6. First, we should consider the solubility of these flavor components. It may change dramatically in few degrees. And some components may be dissolved in coffee more easily then water (they might not be water-soluble, but liposoluble). Since water flow through the cake, the oil content will increase and the temp will decrease. So the tiny themperature change may change the flavor component concentration in the coffee.

    The second thing is the sensitivity thresholds of the flavor and taste chemicals. If 0.1F increase can rise the concentration of one chemical just across the threshold, then we can sense it.

    Third, the higher temp also increase the degradation rates of some chemicals. That’s why sometimes you cannot get more flavors at higer temp. And because the sense of smell and taste usually are not caused by a single cemical but by a set of chemical mix. When the ratio between these chemicals changed, the flavor changed.

    The last things is about the scale. If we have to adjust the temp by 0.1F and claim that change is sensible. How could we control the other variable factors under the same parameter?

    I must apologize for my poor english, it’s not my native language. If there is any words unpolite, I really do not mean it.

  7. Jim – I do think that heat is likely to change a volatile aromatic. Though the volatile aromatics might be thermally stable in vacuum, but with oxygen and water together, some are extremely reactive, such as methanethiol, Strecker aldehydes… The higher temperature will speed up the oxydation rate. But by 1F change will change the sense impact due to the higher oxidation rate? I do not know.

  8. Great comments, thanks you!

    You are not at unpolite at all, and thank you for correcting me on the changes to the aromatics. If you can point me in the direction of more reading on how solubility of different compounds change, and more on the break down of aromatics I would be very grateful indeed.

    Thank you for reading, and do feel free to post more comments here!

  9. “I know that some compounds are less able to dissolve at higher temperature, and often these are surfactants – which begs the question – does a cooler brew give a better crema?”

    Jim – Have been doing experimentation with exactly this lately, and almost always, a cooler brew gives more stable crema, but higher temps often give more voluminous crema (but highly unstable). This is especially true of coffees that are tasting better at higher temps.

    Also I’ve been wondering, given comments about higher temps hiding acidity. When roasting coffees that are particularly acidic in nature, they have a far greater tendancy towards bitterness if over-roasted. I don’t know where I’m going with this, just generally exciting my brain. I know there are many relationships between roasting and brewing, over/under-roasting have similar effects on coffee tastes to over/under-extraction, coffees that are roasted faster de-gas and go stale faster, and also have sharper profiles, i could go on but i shouldn’t

  10. I think “Yuan-Cheng” has said most of it (embarassingly in better English than I probably would have managed!).

    It is unfortunately impracticle to consider the coffee extraction process with pure formulae such as ‘Arrenhius’ equation – as these hold only for single reactions under laboratory conditions of pressure, vacuum etc.

    The overall coffee flavour is made up of a huge number of individual reactions, all with differing activation energies (temperature/pressure), inhibitors (oil, temperature, water quality) and catalysts (you get the idea!).

    Instead of imagining one Bell curve, the reality is 1000′s of simultaneous Bell curves, all of differing amplitudes and standard deviations – all reacting to different factors. While the reactions which offer that acidic flavour may peak at 93.4C, so the reactions which deliver delicate floral flavours may peak at 92.3C – thus it is impossible to achieve a peak of both under a single temperature reaction.

    When you know that commercial espresso blends are roasted a tonne at a time, in open warehouses, with a loss of roughly 12% in weight, often with 5 or 6 different green beans using gas burners with high /- tolerances – the temperature of the group head is a tiny step in the huge journey the bean takes from earth to cup.

    If you taste a well prepared espresso, change the temperature of the group head by 1 degree and repeat – you will clearly taste a difference. This is because the only changed factor is the group head temperature. If you taste an espresso, change the temperature, roast another tonne of the identical blend and try again you will taste a difference – but then you would have anyway…

    I think the only common denominators which can be taken from this are ‘different’, ‘variables’ etc. There are just too many factors to attempt to break it down in to it’s constituent parts. As with old world wine – which has been trying in vain to nail down how to guarantee the perfect vintage for 100′s of years – you try and control the big factors as best you can, but make sure you make the most of it when you get that exceptional result!

  11. Sorry for not reply immediately, cause I am preparing for moving to yale for two years and for the 2006 taiwan barista championship judges certification workshop to get the certification of judgement.

    To Jim- though my major is biochemistry, but coffee chemistry is still too complex for me to fully understand. My coffee chemistry knowledge is mostly from two books: coffee recent developments, edited by R. J. Clarke and O. G. Vitzthum; Espresso Coffee The science of Quality, first and second edition, Edited by A. Illy and R. Viani.
    These two books provide some information and some clues to this topic, but there are not direct evidence to show us how one F degree difference changed the aroma impact or the taste. Maybe the references in these books can give us more, but most of them are not reachable by me. I also hope anyone can suggest me more materials to read.

    To Owain- Yes, there are huge extraction curves of different chemicals during espresso brewing. But maybe we can consider the some major chemicals which take the major responsibility for certain flavor. In this way, might simplify the analysis for finding out the best brewing temp for all the flavor we want. I think you all have read this article in Scientific American:The complexity of coffee.
    If not, here is the link:
    http://www.illyusa.com/pr/coffee.pdf
    In this article,at page 5, there is a figure which shows the cumulative chemical composition of espresso with increasing extraction time.
    ( I do not know how to paste the figure here, maybe someone can help.)
    In this fig. we can see how to decide the optimal brewing time, and we can also find out the optimal ratio between the bad and good flavor. If these data are reliable, we can find out the optimal brewing temp by fixing all the other brewing parameters but temperature.
    I must admit, this is all my fantacy. Please do not take it too serious. And I am dreaming one day I am rich enough to have a odor lab like illy….

  12. At the temperatures you mentioned the 1 degree higher feed temperature, will give approximately a 1 degree higher outlet temperature (slightly less than 1 degree). Environmental influences like a small draught can have a bigger influence…

    From a chemical engineers perspective I can’t shoot any holes in the arguments of Yuan-Cheng, it looks very solid. Extraction is very temperature sensitive, and there are more factors playing a role, like thermal decomposition of the flavours, oxidation etc. It is far more complex than one can imagine and if we understand it completely it can be questioned if we can influence it. There is chemistry, biochemistry, psychology and many more professions involved.

    By being able to play with temperature, one gets “different” tasting espressos from one batch. Different temperatures don’t mean IMHO that either of the temperatures produces wrong espresso, it is just different. I made two successive shots for a friend. We were both surprised about the difference, the first (hottest group) was strong, a bit bitter and with a complex taste, the second one was surprisingly soft.

    I like the article and the way and passion Jim (and many others) are searching to make espresso better. Keep up the good work.

  13. The graph within the Illy article is interesting. I would love to see a comparative qualatitive graph showing human sensory responses vs extraction time for a large demographic. I would wager they would struggle to correlate for the factors mentioned by Chemistone (“chemistry, biochemistry, psychology and many more professions”)

    Yuan-Cheng – picking out the signature sensory tastes to use as sort of ‘trackers’ sounds like an excellent idea. The only problem may be that the “optimum” extraction time as determined by Illy – which you wish to keep constant – was calculated for a given temperature. As this temperature changes, so will the optimum extraction time.

    To keep all variables constant except temperature is therefore almost impossible and would require a 3D Bell curve of “Extraction time” vs “Temperature” vs “Chemical concentration”, which (as it is a 3-dimensional problem) would most likely offer more than one result.

    Either way, I ultimately agree with Chemistone that once you get to this level of “correctness”, all of the espressos are great, just different – which means that personal preference must therefore hold more sway than the tolerance of the group temperature.

    I find this very interesting, especially since I am viewing this perhaps from more of a Physics background than others. I don’t, however, fully appreciate the finer details of aromatics, like Yuan-Cheng, so it is great to have such a well balanced discussion.

    Let me know when you get that odor lab!

  14. Owain- The 3D plot idea is great and certainly correct. All the parameters are somewhat linked. Like everybody knows, if you got a perfect cup in every aspect but a weak body, then you want to get a cup with more heavy body, without adjusting the grinder and the brewing temp, you can dose up a little. In the same time, to keep the flow rate unchanged, you have to pack the cake with less force. So we have to adjust two parameters in same time to move one espresso character toward one certain direction.
    (the sentences above are very fuzzy. It is difficult for me to express my thought clearly. Please somebody help me!)
    Anyway, if you just dosing up only and keep others unchanged, you will slow down the flow rate and get a over-extracted cup.

    And yes, it is really great to have these discussions. You pulled my head out from the chemical mud, thanks.
    I will let you know when I got it. But a GC mass costs about 600000 USD at least!

  15. Yuan-Cheng – I know it was just an example, but if you up-dose, and tamp lighter, a heavier body won’t be the only change in flavour. Surely you’ll be extracting less from each particle of coffee, but you’ll have a greater quantity of coffee that you’re extracting from, so as well as having greater body, you’ll probably have more defined flavours (due to lack of complexity), but a far less complex cup.
    I think that kind of says what I mean, but in a simplistic way. Perhaps someone can further clarify (and probably disprove) me.

    Basically I don’t think, regardless of how many variables we change to balance things out, that we can ever have an effect on just a single aspect (such as body), there’s always going to be other aspects that get effected.

  16. Maybe the references in these books can give us more, but most of them are not reachable by me. I also hope anyone can suggest me more materials to read.

  17. “A higher temperature will reduce acidity in the cup, but this is down to how the tongue percieves acidity and how the increased solubles affect the detection of free H ions. Which is something I should probably write an article about”

    Hi, have you written such an article ? I’m really interested in understanding better the effect temperature has on perceived acidity. Thank you !

Submit a comment