Why coffee tastes bitter, and why caffeine is barely to blame

Strip the caffeine out of a coffee and the cup is still bitter. Decaf drinkers live with that fact and rarely ask about it, because the popular explanation is so tidy: caffeine is bitter, coffee has caffeine, case closed. A food chemistry group at the Technical University of Munich spent years pulling roasted coffee apart compound by compound and putting the pieces in front of a trained panel. Their answer to why coffee tastes bitter is blunt. The molecule everyone blames is a minor contributor, and the compounds doing the real work are not in a green bean at all.

Caffeine explains about 15% of coffee bitterness

Thomas Hofmann, professor of food chemistry and molecular sensory science at the Technical University of Munich, put a number on it when the work was presented at the American Chemical Society's 234th national meeting in Boston in August 2007. Only 15% of the bitterness a drinker perceives is due to caffeine. The rest comes from two classes of compound that a green bean does not contain.

That moves the question off the bean and onto the roaster. Coffee bitterness is not a property the seed arrives with. It gets manufactured, deliberately or by accident, somewhere between the drop into the drum and the water hitting the grounds.

The roaster builds the bitter taste in coffee

First out of the reaction are the chlorogenic acid lactones, known to chemists as quinides. Green coffee is loaded with chlorogenic acids, and heat rearranges them into lactones whose bitterness most drinkers would file as coffee-like rather than harsh. One of them, an orphan compound the Munich lab isolated and named 3-O-caffeoyl-epi-gamma-quinide, has a bitter recognition threshold of 58 micromoles per litre.

How long the heat runs decides how many survive. A 2010 study from the same lab roasted beans at 260 degrees Celsius for anywhere from 60 to 600 seconds, then held the time at 240 seconds and swept the temperature from 190 to 280 degrees. Those quinides formed best at slight to medium roast degrees and broke down again once the roast was pushed harder, which makes the window for pleasant bitterness narrower than a roast level printed on a bag suggests.

What the heat makes once the quinides go

When those lactones degrade they leave something rougher behind. The Munich chemists heated caffeic acid on its own, found it produced a bitterness the panel called reminiscent of strongly roasted espresso, then screened the reaction products and isolated 10 bitter compounds. Eight were multiply hydroxylated phenylindanes, five of them never described before. Recognition thresholds ran from 23 to 178 micromoles per litre.

Dark roast coffee carries those in quantity, which changes the character of the bitter taste in coffee and not simply the amount of it. A cup pushed too far in the drum hangs on after the swallow instead of clearing. The American Chemical Society announced the finding at the time with roasting, not the bean, named as the deciding factor.

A tongue with 25 bitter taste receptors

No drinker separates a quinide from a phenylindane by taste, and the reason is anatomy. Humans carry roughly 25 bitter taste receptors, the TAS2R family, against thousands of bitter molecules in the diet. A 2010 paper in Chemical Senses challenged all 25 of them with 104 natural and synthetic bitter chemicals and found the tuning wildly uneven. Three receptors between them picked up about 50% of everything tested, while 19 of the compounds set off as many as 15 receptors at once.

Twenty-five sensors reporting on thousands of molecules can report a category, never an identity. Bitterness arrives as one flat signal, much the way a sharp cup gets filed under acidity whatever actually caused it.

Why coffee tastes bitter in one cup and not the next

The same 2010 paper tracked what leaves the grounds and when, and that is where a home brewer has leverage. The caffeoylquinic acids and the mild monocaffeoyl quinides wash out fast. The dicaffeoyl quinides come away slowly. The 4-vinylcatechol oligomers, the harsh end of the range, showed strong retention to the ground coffee and gave themselves up last.

That is over-extraction described in molecules rather than in tasting notes. Give the water more time, more heat or more surface to work on, and the last thing it collects is the roughest thing in the bed.

There is a cheap test worth running on the next bag. Brew it the usual way, then brew it again with the grinder one step coarser and everything else held still. If the harshness falls away, it came out of the water and the fix lives at the counter. If it survives both cups intact, it was built in the roaster, and no adjustment on a kitchen bench will take it back out.