Clarified Coffee

coffee espresso shot comparison

The shot glass on the left is a much stronger coffee than the one on the right. Yes, you did read that right – the one on the left, that looks much weaker is in fact stronger. I could just be lying to you but I will explain.

For a little while now I have had a the goal of creating a cup of coffee that is crystal clear. I think, in the run up to the WBC I rambled on at a few people (sorry Zander) about my grand plans and in the end I did something totally different (about which I am very glad).

The original plan was to try and use a Rotary Evaporator, which looks a bit like this:

rotavapor rotary evaporator

I had planned to use it by loading the bowl on the right with coffee. This is then lowered into the waterbath below and gently turned as it is heated. Whatever evaporates goes up into the cooling column which is kept nice and cool by pumping water through the coil so that it would condense into the bowl on the left. What particularly appealed to me about this is that it is a closed system pressure wise and you can see the large display showing you the millibars of pressure inside. This you could reduce, which would in turn mean that your coffee would boil at a very low temperature meaning you would damage it less and it would all go quicker.

The idea of a crystal clear liquid at the end of it, that tasted of all the volatiles of coffee got me quite excited. I spoke to Chris, back when he was at the Fat Duck, and he had done some tests and from a taste point of view suggested I look for a different route. In the end I dropped the idea completely and went back to the drawing board and went down the tasty/flavour marriage route.

Last week I was hanging out with Chris and we were pulling some shots and the subject came up again. He’d recently been chatting with a few people about ice gelatin filtration and suggested I give it a try. When I got home I googled a few things looking for a precise recipe and I came across Harold McGee’s article in the New York Times. It also links to the rather excellent blog Ideas in Food which has some great ideas for this kind of thing (Hot Dog Stock! I love it!)

So I began my experiments to try and succeed with this method. As you can see from above I have not yet been totally successful but I figure if a few of us try we might get there much faster.

You are going to lose a fair percentage of the liquid each time you do this, so start with quite a lot. I brewed about a litre in a press:

brewing a french press of herbazu

Herbazu was the coffee going through the transformation. The problem with press coffee is that is leaves a lot of solids in the cup and we want rid of those. I chose to filter it again through heavily rinsed filter paper:

filter paper filtering coffee

You could of course just brew it in a Chemex but I can’t brew 1.2 litres in my Chemex so I did it this, rather odd, way around.

For the next stage you need some nice leaf gelatin. Prepare/soak it as per its instructions. I worked through a range from about 0.5% to 0.8% by weight. (So for 1 litre I used between 5g to 8g of leaf gelatin). Melt it in whilst the coffee is nice and hot and allow to cool and then put it in the fridge so it gels. This is very important. VERY. What is happening when it gels is that the gelatin is creating a nice coherent network woven through the liquid and this is going to do all the hard work later. After a few hours in the fridge you can then put it into the freezer. Leave it overnight. Nothing in this process can be hurried.

For the next part you need a bowl and some muslin and maybe a perforated tray. I didn’t have any muslin to hand so just used a sieve, though please understand that using a sieve with holes this tiny is total overkill (I was assured):

Sieve 90 microns

All you need the muslin to do is keep the block away from the bowl and stop any pieces of ice falling into it. I am sure you could just put the muslin on top and secure it there like the skin of a drum and pop the block on top and let it melt. Be patient. Nothing will happen for a few hours. But eventually the water crystals will start to melt and head south to your bowl. What is important is that the fridge keeps the gelatin solid so that its network filters out all of the solids. The whole melting process can take up to 48 hours if you want maximum yield. You will be left with a goopy and unpleasant feeling jelly afterwards. In the bowl below you will have a liquid that is a concentrate of the water soluble coffee flavours.

melting gelatin ice filtration

I have yet to manage a really great clarification in one step and the above shot glass of liquid was the result of a second clarification. Maybe I am doing something wrong, maybe it is unlikely to get any clearer without several repeat efforts. I will probably bug Chris (when he gets back from El Bulli – so jealous!).

One thing I haven’t covered: How does it taste?

As mentioned above this is very much a water extraction. All of th oils we hope to leave behind in the gel above, and this is certainly a goal if you are using this for meat stocks.

A few things surprised me: The strength of the chocolately flavours in the reduction, and there is a certain sweetness though often I was tasting it straight from the fridge which really caused the bitterness to pop out. The other interesting thing is the length of finish – I always associated the long finish of coffee with oil droplets coating the tongue and slowly releasing both tastes and flavours. It seems that some water soluble stuff likes to hang around for quite a long time too.

And finally – what is the point? Good question. I think it would be fun to serve people a cheeky glass of clear coffee, though probably slightly sweetened as part of a signature drink and I guess it would be kind of fun to do some work with gelan and do a vertical split of clear coffee on one side and normal coffee on the other. Perhaps have the clear one hot and the normal one cool and sweetened. I don’t have enough gelan to hand to play with this, and not enough time in the next few weeks but it could be cool.

Let me know if you have a go with this – it is not very labour intensive and just requires a little faith/patience. And if you have a more successful filtration then please, please share the recipe.

Clarified coffee

Hope at least a couple of you have fun. Question, as always, in the comments…

UPDATE:  Seriously – the repercussions and possibilities of this for competition are just starting to dawn on me.  This would have been a much cheaper and easier way to do the liquid donut, and this method opens up a load of foodstuffs for work with coffee.  Prepping this stuff up obviously can’t be done on stage but it does open up a whole range of ingredients and possibilities.  Be interesting to see if people take this where it can go.  Almost makes me wish I could compete again just to put together another signature drink.

17 Comments Clarified Coffee

  1. Dave

    Isn’t Gellan supposed to be add when cold and then heated?
    If so would you cool the coffee and then reheat it?
    What does reheating do to the flavours then?
    Is this what you did with last year signature drink?

    That’s enough questions…

  2. wheretherearewolves

    Hey James-
    Yes the quest for clear coffee has been kicking around these parts for a bit-
    If you are ever in NYC I will take you over to WD50 my friends cook there and we will get this straightened out properly. Last year’s sig for me was a sabajon where I reversed the elements of the cappuccino. I used a liquid/warm almond milk as the base, and then sabajon’d the espresso by adding methylcellulose/xantham gum to the shots heated over an induction burner, wisked it till it foamed!
    Glad to see the WC isn’t sitting on his laurels and is more insane than before.
    I am currently up at night thinking about all the possiblities with the pacojet….

  3. jim


    Isn’t Gellan supposed to be add when cold and then heated?
    Yes Gelan should be dispersed cold and then heated til it hydrates but I couldn’t get gelan to work well with the clarification process (though I could well have got the %s wrong on that one). Adding the gelatin to warm coffee didn’t seem to damage the process though I should probably do another batch to check.
    If so would you cool the coffee and then reheat it?
    I guess so – making sure that the liquid is as free of solids that could cause overextraction through heating becomes very very important I guess.
    What does reheating do to the flavours then?
    I don’t really know.
    Is this what you did with last year signature drink?
    I used Gelan in Berne to create the vertical split portion of the drink. I haven’t used the gelatin filtration for anything yet and seeing as I am not competing I doubt I will get a chance. Perhaps it is something staff from the future cafe may take and experiment with.


    My friend Chris (Young) is pretty good friends with Wylie I think. I know that Chris put Wylie onto the gelatin filtration.

    How was the sabayon? I found using both methylcell and xanthum for an espresso mousse (like I did in Berne) meant my percentages had to be bang on or the texture just went a bit funny. I used an ISI whipper to make the foam, loaded with just a single cartridge of nitrogen – any more and the texture got a bit stiff.

  4. RichW

    My question would be the value of using these essences for cupping – or actually, as an aid in consumer education (say, at a press pot tasting of several varietals in the same region, where understand nuance becomes more important).

    After reading the NYT article I’d assumed I could just filter coffee brewed through the method of my choice through the gelatin… but seems that’s not the case. So thanks for doing the work on this and getting it started.

  5. jim

    Rich – I don’t think these essences are valuable in terms of giving a genuine representation of the coffee. We’ve lost all of the oil fraction and the flavours that go with it, as well as a lot of the solids so mouthfeel is pretty much gone too.

    You hit upon my original idea for all of this which was to use the rotavapour to create essences of each of my blend components to give to judges in little vaporizers so they could spray a little in their mouths and get a good picture of the green/blend components without having to drink/ingest any brewed coffee. (This came to me after reading about the chef at Alinea in Chicago and his use of vaporizers full of seafood broth)

    I just don’t think it can be done like this (someone please prove me wrong) because of the many and varied contributing factors to the taste of a cup of coffee.

  6. Adam

    Even though your result isn’t crystal clear, it’s still cool, and I’m quite jealous that this hadn’t occurred to me before.

    And the world of coffee expands a bit more…

  7. wheretherearewolves

    Hey James-
    This is regard to your comments about the sabajon-
    Yes if the meth/gum snot percentage isn’t nailed down it can get way to gummy and interfere with the flavor. I played around a bit with the isi – espuma thing but found nothing that just kicks ass. For me whether be coffee or anything culinary, artistic – it’s finding that right balance between being conceptual and actually tasting f’n good. Sounds like you nailed it this year in Tokyo –

    Here in NYC there is a new restaurant a couple of blocks away from the cafe that I work at called Tailor – they use our coffee and the head chef is Sam Mason, ex pastry chef from WD. If you are in NY sometime I’ll take you by there.
    There is actually an ex barista from Gimme coffee working there as a chef. They whole concept is creating a menu that deals 1/2 with sweet dishes and 1/2 with savory dishes. Combining flavors that are point/counterpoint ie/ chocolate and banana raviolis, mustard ice cream, coffee soil…. check his stuff @

  8. David Craig

    Couldn’t help but interject James. I love the idea of drinking a colourless shot. The colour is originating from the roasting process yeh? I wonder if you could remove this element and ‘roast’ the greens in a different way that will minimise/eliminate the browning? I’m thinking about the ‘invisible crust’ bread. If you ended up with a colourless ‘roasted’ bean, it would make a colourless Espresso? Just an idea…

    You’ll need to tell me again what the food bible was you said I should pick up…

    cheers, Dave

  9. David Pier

    Has anybody tried diatomaceous earth filtration? It has much finer pores than cellulose filter paper and it is more inert than gelatin. It could also be done fast, with fresh hot espresso. The diatomaceous earth could even be preheated. It seems probable to me that gelatin, as a protein, would trap certain types of chemicals that a truly inert filter medium would not.
    And why did you reject the rotary evaporator?

  10. Tristan

    I’m giving the gelatin/freezer method a go at the moment. Will let you know the results, might blog it myself but i’ll credit you etc.

    What’s this diatomaceous earth stuff then? Sounds cool.

  11. Martin

    I just recently discovered your blog thanks to a comment you made to my recent post “Wonders of extraction: Espresso”. And to my great surprise I discovered that you even have a “food chemistry” tag – wonderful!

    Gelatin filtration is an elegant method and your take on coffee was indeed a good idea, especially since we are so used to coffee beeing black. Once you’re at it you might also be interested in trying an agar filtration (as described in v. 2.1 of “Texture – A hydrocolloid recipe collection”.

  12. herbalaire vaporizer

    You are amazing! I will try to experiment on this too. So i can serve my friends a clear coffee and if they love it i will share your recipe to them. And as they say “Patience can make things more better”.

  13. Pingback: my explanation

Leave A Comment