TGRWT #1: Garlic, coffe and chocolate

Reading the comments on how to prepare a dish using garlic, coffee and chocolate, I figured it could actually be a good idea to make this into a food blogging event. Inspired by Is My Blog Burning (IMBB), Sugar High Friday (SHF) and the like, I hereby launch a new food blogging event called They Go Really Well Together (TGRWT).

The name refers to flavour pairing of ingredients based on their content of volatile aroma compounds. The idea behind flavour pairing is that if two (or more) foods have one or more volatile compounds in common, chances are good that they might taste well together. Click for a list of other flavour pairings and to read previous blog posts on the topic. The molecule shown in the logo is of 2-methylfuran-3-thiol, a very potent aroma chemical found in coffee, chicken, meat, fish and popcorn – to mention a few.


Many flavour pairings seem strange at first, especially when the combination is not found in any recipes. To illustrate the flavour pairing one can always just eat the two ingredients together. But it would be so much nicer to actually make a proper dish out of it. Therefore I’m quite excited to see what the creative minds of all the food loving bloggers can come up with!

This is how the first round of the blogging event works (hereafter referred to as TGRWT #1):

  1. Prepare a dish that combines garlic, coffee and chocolate. You can either use an existing recipe (if there is any) or come up with your own.
  2. Write a entry in your blog by May 1st with TGRWT #1 in the subject and make sure to include a link to the header of this post for trackback links. Readers will probably be particularily interested in how the flavour pairing worked out, so make an attempt at describing it.
  3. Deadline for submissions is May 1st. A round-up will be posted by me here some days later with pictures.
  4. Please send me an email at webmaster (at) khymos (dot) org with the following details: Your name, URL of blog and URL of the TGRWT #1 post and a picture for your entry in the round-up.
  5. If you don’t have a blog, email me your recipe, name and location and I’ll be glad to include it in the final round-up.
  6. In due time I will ask one of the participants to host the next round on their blog (and provide an updated logo).


  1. the original chocolate drunk by Maya was a quite potent brew. It had no milk but contained maize (preferably a mouth-chewed fermented maize) and jalapeno peppers. It was served whipped (the frothy part was supposedly the best). I bet it could accomodate to some garlic and coffee.

  2. I had a wonderful idea for a raw meat dessert a few months ago which I’ve not yet tried. Take some beef, the best quality steak you can find, and cut into lumps about 4 by 1.5 by 1.5cm. Slice a “pocket” into the long side of each lump, and stuff with candied ginger. Dip in molten dark chocolate. Chill in the fridge, and serve.

  3. Hi Martin! (do you remember me? We met in Erice a couple of years ago; I am the bald guy from Spain)

    I am not so sure about the food pairing theory. I have been studying and doing research about volatile compounds in foods for more than 15 years, and it is strange that two different foods do not share some volatiles. Therefore, following the flavour pairing theory, everything should go well together. Just to check, could you (or someone else) give me some examples of foodstuffs that don’t go well together, just to have a look to their profile of volatile compounds (Leffingwel database is ok, but there is much more information when considering the whole information contained in each scientific paper, in which no volatile compound is left behind). It should be also considered that the profile of volatile compounds of each foodstuff is strongly influenced by the method used for analysing volatiles (i.e.: using SPME some compounds are not extracted at all, while doing distillation, some of the lowest MW volatile compounds are lost).

    Best regards from Spain.

  4. Thank you for all comments! I’m looking forward to see what kind of garlic-coffee-chocolate concoctions you’ll come up with!

    Jorge: I certainly remember you and thank you for your comment! You (and others) question the validity of the flavour pairing hypothesis, and you make a very good point when pointing to the fact that when you pick two random foods, chances are good that they share one or more volatile compounds. However, the volatile compounds do not all have the same impact on the overall flavour. I think one should take into account both the concentration at which the volatile compounds are present and their respective impact. In the litterature I see that the term odor activity value (OAV) is used (defined as odor concentration divided by odor threshold). So to be more precise, the I think flavour pairings like this are limited to the volatile compounds with the highest OAV in each food.

    Next challenge of course is – are there databases or books that list let’s say the volatile compounds with the highest OAV for foods?

    Although it’s easy to find examples that contradict the flavour pairing hypothesis, what really fascinates me is that by applying this principle, a number of combinations have come up which really work well together! But even so I do agree that when considering all the combinations of flavours found in recipes, probably only a fraction can be attributed to flavour pairing on a molecular level as described above.

  5. Yes, this makes sense. There are many volatiles in huge amounts that do not contribute at all to the overall flavour, whereas some almost undetectable ones can be key volatiles in the flavour of a foodstuff. There is a database called Flavornet (, by Prof. Acree from Cornell University (great researcher) in which I think I remember there are some information about dilution factors. Other databases concerning flavour are , the LRI and odour database ( , there is also some useful information in by the FAO, and also the Aldrich flavour catalog (I haven’t found the web page, the book is really good).

    In addition, there is a lot of information in the scientific literature, specially in the Journal of Agricultural and Food Chemistry and in the Flavour and Fragance journal (also some in Food Chemistry). I will search flavour dilution factors of some of the foodstuffs in the flavour pairing contest to check the hypothesis.

    I think flavour pairing is an interesting initial hypothesis that, perhaps, has to be more concisely defined and needs some (a lot, always a lot if it has to be confirmed; research is hard) additional research.


  6. Martin, I side with Jorge on this (as I already stated elsewhere on this blog) You mention that it is easy to falsify the flavour pairing hypothesis, which of course is a mojor problem.
    I would like to suggest a heavily oaked Spanish red wine and pair it with vanilla ice cream. Obvious pair is the vanilla. Have not tried it but think it will not work.
    The theory stems from backward reasoning I think, a surprising flavour pair is found and then retrospectively matching pairs of flavour compounds are sought. That is not scientific.
    We must nevertheless not be conservative and search for new combinations, but that is a different matter. Will we find many? One could consider modern cookbooks the result of at least 2000 years of research on flavour pairings…

  7. Jorge: Thank you for links to the different flavour databases that are available. I agree that this needs more research!

    Jurgen: I understand you skepticism, and appreciate your comments, but I think you miss the point if you expect that every pair of foods that share volatile compounds will be a perfect pairing. Rather, the hypothesis suggests that some combinations might work.

    I think the reason why red wine and vanilla ice cream would not work is that red wine in general does not go well with sweet food because of the tannins and the acid. A sweet red wine like Recioto however should work well with ice cream.

  8. Martin, As a fellow scientist you should appreciate the fact that if a hypothesis is proven only in some instances and in other instances not, the hypothesis is wrong. That is what I wanted to say.
    I think you have one of the best molecular gastronomy weblogs on the net, so please consider my critisism a scientific discussion. I believe firmly however that if we want to apply science to cooking we must adhere to basic scientific principles. We might find that cooking is not science, or that we do not understand taste perception well enough (yet…).

  9. Jurgen, I agree with you about a hypothesis being wrong if it can be disproven in only a single instance. But if the hypothesis is that “some of the combinations might work”, I’m not quite sure if there is much to prove or disprove. I guess it rather suggests that the “statement” doesn’t really qualify as a hypothesis in a scientific sense. But I can’t find a better name, because after all it is a “suggested explanation for a phenomenon”. And even if we could prove or disprove the hypothesis, one could ask – is there in an objective sense such a thing as “good” or “bad” pairings? Are they only a result of subjective preferences and cultural learning? I think that would be a very interesting discussion (yet a totally different one)!

    Nevertheless, I’m still fascinated by the flavour pairing concept. Take the combination of cocoa and baked cauliflower for instance. I think (but correct me if I’m wrong) our olfactory system is not very good at decomposing and tracking back a smell sensation into the individual compounds that caused it. Therefore, eating baked cauliflower, it had never occured to me that it’s smell is reminiscent of cocoa. That is until I read about Heston Blumenthal combining the two and actually tried the combination myself. I guess the reason I’m so fascinated is that the combinations have (mostly) been delightful surprises!

    If I had time, access to the requried databases and access to the analytical equipment needed, this is what I would like to do to gather more data on flavour pairings:

    I would like to find the most important odours of each food in the pairings listed, based on the odour activity values (= concentration/threshold of each individual compound) and check if there really is a significant overlap. This should then be compared with similar data for a couple of random combinations. It would also be interesting to know how many of the volatile compounds are generally needed to recreate the characteristic smell of each food? Based on this one could hopefully formulate the hypothesis more precisely so that it could be tested.

    I would then be interested in testing the predictive power of the hypothesis. Heston Blumenthal used the VCF database to come up with several of the combinations he listed, but how did he end up with that list? Were there any false positives and if so how many? If it turns out that the hypothesis has some predictive power, it would be interesting to do some number crunching on a database to extract lists of combinations that should be tested in a kitchen. If not – well – then we should just leave it there.

    But anyhow – at the present I think the concept needs to be explored in much greater detail. And I do hope some flavour research group will catch up on this.

  10. HI, Just stumbled across your blog as I search for inspiration to do a presentation on coffee and cheese pairings. Have given your challenge some thought and although I know I’m late and am not offering recipes or photos, I’d like to offer my suggestion….a hearty pot roast braised in coffee, finished with a Mole sauce and served over garlic mashed potatoes. Flavours should work beautifully. On a side note I have a piece of Pecorino Al Tartufo which I find has this familiar garlic taste but none of the lingering after bite that garlic has. I’m going to sample it with an earthy Indonesian coffee, Sumatra or Sulawesi. Camembert worked well with them. There must be similar structure between garlic and truffles?Cheers! Barb

  11. Texas chili-already has chili and garlic and in Ohio they always add chocolate-btw-in texas (and elsewhere) coffee is often added as well.

  12. Just came across this site courtesy of James Hoffmann. thought I’d share that I created an aioli for a roast beef panini we sell that employs not only coffee (espresso) and garlic, but also horseradish. It’s even good on its own as a dip for breadsticks. And here i find out why!

  13. […] They Go Really Well Together is a monthly food blogging event that celebrates flavors paired at the molecular level. For this month’s round, Inge of African Vanielje chose apple and lavender as the magical combination to work with. I’ll be the first to confess that I have never cooked with lavender before, so I had to mull this one over for a little while before I could choose what to make. After much thought, I finally settled on an applesauce cake recipe from my old standby, The Silver Palate Cookbook, but in keeping with the theme, I added lavender to the citrus glaze. […]

  14. I don’t know if anyone here has come across black garlic (garlic that has undergone high temperature fermentation) but it blends very well with dark chocolate, bringing out a fruity, coffee note in the chocolate.

    I tried it after reading TGRWT #1 but interestingly I’d completely forgotten about coffee being the common denominator, so my perception of a coffee flavour should be relatively unbiased.

Leave a Reply

Your email address will not be published. Required fields are marked *

The reCAPTCHA verification period has expired. Please reload the page.