The SCAA and ICO standard for green coffee moisture is 9% to 13%.

The reference method is to bake a sample at 105C/220F in a convection oven for 24 hours and note the weight loss, 1 - (weight after)/(weight before). If you bake exactly 100 grams, then the percentage is 100 - after.bake.weight. This is called the ISO6673 or Karl Fisher method.

The standard field method for measuring moisture is capacitance metering. These cost around $300. An interesting low cost alternative is to measure the RH and temperature of the air in a sealed plastic bag with the green coffee inside after it has stabilized for a few hours. The equilibrium relationship between bean moisture content and RH/Temperature is shown in this graph:



Producers try to keep their coffee as close to the maximum 13% as possible, since its cheaper to produce water than coffee. Coffees that hav dried below 9% are probably past crop. Coffees over 13%, maybe even 12%, may have mold problems.

You may want to adjust the drop in and drying period to mathc moisture when roasting coffees

Great post!

You just gave me two ways of doing what I was wondering how to do without the moisture meter!
The baking method has the disadvantage of using up a part of the beans... Probably more appropriate for producers or coffee traders.

Thank you!

You're welcome. My apartment smells of green coffee, since I'm in the middle of baking a bunch of samples for a coffeecuppers job. So I was searching around for something less smelly. Somebody mentioned this graph in an old thread and I found it in a PDF by the Hawaian Ag Extension. I'm hoping this works more easily for good samples, and I'll only need to bake if the coffee is out of spec.

another_jim wrote:The reference method is to bake a sample at 105C/220F in a convection oven for 24 hours and note the weight loss, 1 - (weight after)/(weight before)

[slightly OT] A while back, you and I were baking roasted coffee to determine its moisture content. We probably got answers that were way too high, since the baking method can't distinguish between water and other volatiles -- and the other volatiles are plentiful in roasted coffee.

Whale wrote:The baking method has the disadvantage of using up a part of the beans...

Of course you don't have to start with 100 grams. With a decent scale that weighs to .01g, a 10 gram sample would be plenty.

AndyS wrote:[slightly OT] A while back, you and I were baking roasted coffee to determine its moisture content. We probably got answers that were way too high, since the baking method can't distinguish between water and other volatiles -- and the other volatiles are plentiful in roasted coffee.

It is the ISO approved calibration standard for instruments measuring moisture in both green and roasted coffee; so I don't think volatiles can be much of a problem.

I'm guessing in roasted coffee, they are already driven off, while in green coffee they are not released until higher temperatures. I have 500 grams cooking now, and there was a mild green coffee aroma for about two hours, now it's gone. The smell is a lot less than one gets early in the roast of 100 grams, and nothing like the aromas late in the roast.

another_jim wrote:It is the ISO approved calibration standard for instruments measuring moisture in both green and roasted coffee; so I don't think volatiles can be much of a problem

Interesting.

The main "volatile" I was thinking about is CO2. According to Illy, it makes up 1.5-1.7% of the mass of freshly roasted beans. If one grinds fresh-roasted beans and then bakes them (as we did), how could moisture weight loss be distinguished from CO2 weight loss?

The baking temperature is probably too low to drive off the CO2. If the CO2 stays in the coffee when it's roasted to 430F, it may release so slowly at 220F so as not to make much of a dent after 24 hours.

Fresh roasted coffee degasses over a week, but I do not think it loses 1.5% of its weight either in whole bean or ground form. Here's an article that claims oven drying does well for green coffee, but less well for roasted (if I understand the abstract correctly). So your thought may be more valid for roasted coffee.

AndyS wrote:Of course you don't have to start with 100 grams. With a decent scale that weighs to .01g, a 10 gram sample would be plenty.

Actually I do...sort of... but I do not think that the second decimal of my scale can be relied on. It is not a gem stone scale or a scientific one.

The lower the weight basis the higher the error... So I do not think that my current material would give me an accurate enough moisture value with 10 grams. This is considering that there is only a 4% range to work with.

Anyway I am not dismissing the technique, but I will try the bag-temperature-moisture method for a while and see how repeatable the results are.

Hi Jim.

There is an online calculator for RH/T here (for wood, they also have the equation). Maybe it is similar.

http://www.csgnetwork.com/emctablecalc.html

Cheers