Espresso Vivace Roasteria
David Schomer’s Blog | Cutting-edge espresso technique and equipment review

Roasting for Espresso

roast1The range for a Brazil.  Roasting lighter produces a flavor dominated by citric acid, darker and the flavor is dominated by bitterness. In between is a world of varietal notes shading from sweeter into more bittering.

ROASTING FOR ESPRESSOno-burned-match_high-res2 David Schomer, July 2012

no-lemon_high-res

The key to developing your espresso roast is precision feedback. The better you are at extraction the more accurate, repeatable flavor profiles you have during tasting. Of course you must perform a rigourous cupping for defects, but espresso coffee must also be tasted using the espresso method. A saturation method such as cupping brings out a different range of flavors than a well made espresso.  You need to be able to accurately determine what flavors you have developed in each bean you try to master.  In espresso this is a witchy prospect: consistent flavor preservation through the brewing process.  But the better you are at extraction the better you will be able to detect how dark to roast each bean. (By now you know the drill on extraction: fresh coffee, sharp grinder burrs, machine is clean and tuned, and shot times of 25 to 27 seconds. For review, look at my archives at espressovivace.com or dig up a copy of my book. )

Most roasters use a cupping formula which at its essence is simple and repeatable to minimize variables affecting the result.  Think of French Press without the screen. But soaking your coffee in water, no matter how concentrated you make the final brew, will not work to replicate espresso flavors. Many notes will not be concentrated enough, like being diluted by the longer soak in more water. But some flavor notes, caramelized sugars or dark chocolate for example, simply do not survive in a drip cone or a French Press. Espresso’s unique contribution to brewing coffee is the pressure and very short percolation cycle. Combined with the temperature controlled machines we live in a time when you can pull a shot that tastes as good as the ground coffee smells.  But, you have to be extremely precise technically, and be working with a machine that holds its brewing temperature to within one degree F..

The way it worked for me in 1991 was to scoot down to SanFrancisco and take a roasting lesson from Robert Henley. I had learned from Ernesto Illy that a Brazil was a good variety of coffee to start with for an espresso blend. A friendly broker, Tom Kilty in Oakland was willing to supply small amounts of green coffee for my experimentation and I acquired a small supply.

I installed a probat 2-barrel sample roaster in my basement at home and begin to roast batches every day that I would take in after de-gassing and try at my cart. Robert had taught me that fresh roasted coffee loses 85% of its internal CO2 after 10 hours, but I always let my roasted coffee stand for two days before brewing and tasting.   I brought each batch to my cart at 5th and Union and Jenny Vanderbeck and I would try the espresso.  It was Jenny that remarked on the caramel taste coming out more with each batch as I tweaked and tuned the roast profile. It took about 2 weeks but I zeroed in on the sweetest band and moved on to a Guatemalan.  I thought that if I enjoyed the Brazil at that exact temperature and degree of development maybe the Guatemalan would peak there too.  It was true for awhile….

Even then I did all my sampling using the espresso method.  My espresso machine was a  prototype LaMarzocco Linea that I had tricked-out to hold a 2 degree F. range of brewing accuracy (instead of a six degree range), and my hypothesis seemed true: the color of the Brazil and the color of the Guatemalan were the same at peak flavor development.  Later with the precision PID machines available now that hold to within a degree, my roast profiles are different for the two coffees with the maximum sweetness point emerging at slightly different roasting intonations for each bean.  Precision brewing allows for precision roasting.  So much so that after the development of the Synesso Cyncra in 2004 I wrote a piece that stated that this was the beginning of the era of roasting and blending in artisan espresso operations because a machine finally existed that could offer a repeatable flavor profile in the final cup. Previously we were flyin’ blind.

Vivace developed our Northern Italian roast in 1991.  It is deep mahogany brown with no oils on the surface, a smell like toasty caramels and earth, and at the time people thought it was light.  Lately I have tried coffee roasted much lighter with a sharp lemon taste in the shot.  This article is to suggest, (and defend) limits to roasting experimentation that are unique to the espresso method. Espresso is a concentrated coffee.  Some things that are desirable for drip or press brewing methods might not play well at the concentration levels found in our little dab of crema: especially citric acid.  An eloquent Brazilian coffee taster I met at a show thought that a little acidity in a French press or drip coffee acted like carbonation in a drink and sort of cleansed his tongue, and ultimately assisted him in enjoying the unique aspects of each bean.  I took it to heart and roasted deliberately acidic coffees of various potency for six months or so and came away totally convinced it is not good for espresso.  An acid taste in the cup is composed of sugars or varietal notes are still not developed yet in the roast. It is too light.  But even worse the citric acid has a destructive effect on any flavors that may have been developed-it literally destroys the fragile molecular structure of caramelized sugars or varietal flavors.

This is my attempt to depict the relationship between acidity, sugar development, and carbonization in roasting coffee.

This is my attempt to depict the relationship between acidity, sugar development, and carbonization in roasting coffee.

For example, I recently walked into a new place in Fremont. It has a beautiful design and feel, with LaMarzocco Strada machines promising some artisanship to the preparation. Big name roasters had stickers on the grinders. My friend and I ordered straight shots of a Harrar they were featuring. Lovingly pulled shots were completely dominated by sour lemon the cup. ( My friend takes sugar and cream, no help. At least with burnt coffee you can get a nice flavor out of it with enough milk and sugar). I was speechless. Is this what you really want in your Harrar? I thought. Is this intentional? Later I found it is aggressively intentional. Well I thought, you just go ahead and serve that and I will just be over here with dark chocolate, blueberry, and caramel in our Harrar.

I will state here that any lemony flavors in the final cup indicate a defective espresso coffee. And of course, in Italy, the birthplace of caffe espresso, an acidic espresso is considered defective.  (You may encounter acidic espresso as you approach Switzerland because they primarily have a drip coffee culture featuring very light roasts).   A persistent lemony flavor in the cup through a wide variety of extraction parameters means that the defect resides in the roast. (Remember you can get lemony flavors from a sweet roast by brewing under temperature as well.)  But before you strap on your boxing gloves let’s try for a little common ground.

Choosing a roast for a single origin bean is a very personal choice and should generate heated discussion about the best way to roast it. The individuality is what makes a coffee culture. (Take a look at the great bloom of craft brewers going in the US, making some of the best beers on the planet. Dozens of styles, beautiful). But, you want a Guatemalan Antigua to be recognizable and have some distinctive notes that make it an Antigua…right? Also roasters, particularly single origin espresso fans, want to taste differences in each bean…varietal characteristics imparted to the arabica coffee through unique soil conditions combined with harvest conditions including weather, and processing.  This is a “given” when we approach roasting any coffee: we want to taste that unique coffee. We don’t want lemony flavors or burnt rubber.


Different varieties of coffee are usually not roasted together. We blend them after roasting each one separately to achieve the best development of that coffees varietal notes and caramelized sugars. Speaking of blending we do not sell or serve single origin coffee as espresso. The reason is the same one I laid out 20 years ago…it is boring to have only one flavor profile in an espresso coffee, like one color lights on the Christmas tree… A blend may have three to five coffee varieties in it. Some beans for power, some for sweetness and spice. Now with the precision machines we see that very small changes in the flow rate of the coffee coming out result in different flavors in the blend being prominent in the cup. The beans vie for your attention on a changing basis, different every day.

For example, our Espresso Dolce has caramel, dark chocolate, leather/salt, honey, blueberry and toast notes present in a balanced shot. But when the flow rate varies (as it almost always does), within that 2 second window of 25 to 27 seconds to pull the shot, different notes compete for dominance. If it is a little slow and short, more caramel comes out. Some shots feature more blueberry or dark chocolate. Toast is the most elusive flavor and only comes out in the most balanced shots. It is so cool hunting for the toast….The resulting experience is quite complex and beautiful day after day as espresso reveals it is even more sensitive than I ever imagined. And for me, it would be a shame to limit the symphony of flavors by only brewing one type of coffee. (Maybe “single origin Fridays” or something like that would work for us…)

In our case at Vivace we also want sweet coffee in addition to varietal flavors, and we try to roast to the peak of sweetness for each coffee we use. This is an aesthetic choice made by myself. As I have said many times, “one man’s cough-syrup is another man’s chardonnay” to indicate my attempt at broadmindedness in the face of darker roasts. This self effacing style is stretched pretty thin by the lemon coffee however.

In roasting coffee we are developing flavors in what is classified as a Maillard Reaction due to the development of caramelized sugars, CO2, and heat in the final stages. We are trying to pick off the peak of the sugars without a trace of acidity. For our roast most Arabica coffees are not suitable because they retain acidity at their peak sugar point, and bitterness creeps in before the acidity is roasted out. Most South American, Hawaiian, and many Indonesian coffees are too acidic to consider for our blend because the acidity drops away at a darker color than we use, towards the end of second crack for those coffees. . We choose our beans from coffees classified as “mild arabicas” because of their low acidity at the peak-caramelized sugar concentration. And yes, Brazils are usually good in this regard.

Another point of agreement is that as you roast darker acidity decreases and bitterness increases. (Andrej Illy “ESPRESSO COFFEE:The Chemistry of Quality”)  Each single origin coffee will have a unique temperature and color point where citric acid drops off. (For my roast, Northern Italian, many arabica coffees feature acidity in the cup. So I search and sample for low acid arabicas.  The darker you go the more coffees that will work for your own roasting intonation. Shopping is easier.).  As temperatures rise the acidity disappears, then you have the development of varietal flavor notes and caramelized sugars.  As you roast darker sugars begin to carbonize and bitterness increases as sweetness decreases. (A note about sweetness…The bitter/sweet preference in roasted coffee is perhaps quite different for different people.   People from hot climates with spicy food traditions may want more bittering in the coffee, as evidenced perhaps by the darker roast as you head south in Italy).   At some point, the carbonization of all the sugars dominates the coffee and all varietal flavor notes are incinerated or covered by the burnt- rubber taste.  This is the limit for dark roasting espresso.   When all varietal characteristics and caramelized sugars have been carbonized it is a defective roast. And when all sugars and varietal notes are in the acidic stage of development, the roast is too light and is defective, a waste of the coffee bean. In between is a world of flavor development within each bean.  A world unique to yourself and your brewing acumen and technology combined.

So you have lemon acidity and burnt rubber at each end of the roasting spectrum as limits for coffee intended for espresso, and an entire world of subtle flavors and caramelized sugars in between. Andrej Illy has documented hundreds of fleeting molecular compounds that compose this mercurial shape-shifter known as roasted coffee. Find your own unique favorite intonation or each bean you use and vigorously defend it as the best way to roast that coffee.  Paint your own picture between yellow and black.

no-lemon_high-res1