Tea’s water takeup
…or, if you think you have your brewing under control, you might be wrong

Why does this matter?

Recently I got mad at myself over how much trouble I was having doing multiple steeps of gyokuro. The first steep would be great, but then … ugh. The problem wasn’t that the brewing temperature was too hot, it was something else.

I started paying attention to how much water the leaves sorb¹, which is an attribute I for one haven’t seen discussed, but I’m increasingly convinced it’s really important. One thing is crystal clear: different genres of tea sorb radically different amounts of water.

It’s pretty obvious that when a tea sorbs a huge amount of water — hello gyokuro! — the temperature of the water you pour onto the swollen leaves is not the same as the temperature of the tea soup you’ll make from them. Not only that, but even with the brewing vessel lid off, a tea like gyokuro that swells up with water and normally steeps not too far above room temperature, will essentially continue brewing for a while after the tea soup is decanted, so in my experience subsequent steeps should be as fast as possible to avoid overbrewing.

Probably the toughest multi-steep brewing problem is a tea that sorbs lots of water and calls for brewing at 100℃. You can get closer to the objective by various techniques that help maintain a high temperature in the brewing vessel between steeps, but this means extraction will continue in the wet mass between steeps, which leads to a risk of overbrewing.

I’ve been trying to quantify how sorption-capable various genres of tea actually are. In the trials detailed below, I always began with 4 grams of tea leaves and did 4 steeps in a thin porcelain gaiwan, being careful to pour off all the tea soup in each steep. Not all brewing temperatures and steep timings were the same, because hey, I was making tea, not torturing the leaves. A 55ml gaiwan was always used unless otherwise noted. (The 10-minte simmer column in the table, thanks to a suggestion from Jens Dennig, was added several weeks into the project as a means of finding the absolute maximum takeup for a given tea; we’ll see how many of the teas get an entry in that column.)

It isn’t necessarily true that the leaves continue to get heavier after each successive steep. (See especially the final simmer for the Dooars CTC.) I think the reason for this is simple: as water enters the leaves, soluble solids exit the leaves. Since the total soluble solids in dry tea leaves amount to ~40% of tea leaves’ dry weight, if you exhausted 4g of tea leaves you’d be left with only 2.4g of the original leaves.

Speaking of the final simmer, when its weight is a lot higher than that of the fourth steep, that presumably means that four steeps with the brewing parameters I used didn’t come close to exhausting the tea. (See especially the table row for Lu Shan Yunwu at the “normal” 4g/55ml, where I feared making the green tea undrinkably bitter and astringent. The same weight of the same tea in a 120 ml gaiwan takes up a gram more water as of the fourth steep, not to mention tasting much better brewed less aggressively.)

The data (sorted by maximum leaf weight)

Tentative results

So what determines the amount of steeping water a given tea takes up? It looks as if the way tea leaves are processed during manufacture dominates. Most strikingly, what gyokuro (the champion) shares with the runner-up fukamushi sencha is killgreen by steaming. At the low-sorption end of the scale, lao cha tou stands alone while Fu Cha Ju old shupu is heavier and “normal” shu pu’er (Heng Tong Hao) what you might call lower middle class. As to other possible factors in the rankings of the various teas, I’m reluctant to say more until I have 10-minute simmer measurements for more of them.

What do you think? You can explore ideas of what might account for the variation in weight by interactively highlighting tea attributes that seem promising.

¹Eric Scott has pointed out to me that we’re not just talking about absorption of water, that adsorption is important too. He conjectures that because leaves harvested when they’re young and small, like gyokuro, have a higher surface-to-mass ratio, they might adsorb more water per gram of dry leaf. But it seems to me that this idea is undermined by the fact that the fourth-steep weight of the huangpian-only sample—by definition made from huge leaves—is among the highest of all teas tested.

²Because gyokuro is such an outlier, an astute reader suggested that I try it again. As you can see, the result isn’t materially different.

³Admittedly CTC is an unlikely candidate for gongfu brewing, but we want to know why a given tea holds the amount of water it does, and it occurred to me that lots of small particles might take up water faster than a few intact leaves.

⁴Because we have no simmer column for this tea, and because I suspect Thomas was trying to avoid pushing a dancong too hard to avoid its notorious astringency, the apparent low water uptake should be taken with a grain of salt.

⁵Warning: if a tea lacks an entry in this column, its sort order in the table, which is based on its maximum weight, may be misleading.

<perin@acm.org>     @babelcarp on Mastodon

First published: May 4, 2022

Latest revision: July 8, 2022