Saturday, December 28, 2013

Destructive Myths: The Dissolved Oxygen Hypothesis

Part I: The Argument

There are a lot of myths and unfounded maxims surrounding the proper preparation of coffee and tea. Most are harmless, or, at worst, detrimental to beverage flavor. But one such myth has resulted in a massive waste of energy and water. I refer to this myth as the dissolved oxygen hypothesis.

The dissolved oxygen hypothesis states that, when brewing tea, one should always use freshly drawn water, and never reboil water in the kettle. The justification given is that water that has previously been boiled has less dissolved oxygen (DO). The result is that many tea drinkers are wasting enormous amounts of energy by dumping leftover hot water from their kettles.

The dissolved oxygen hypothesis rests on two premises: (1) that once-boiled water contains more dissolved oxygen than twice-boiled water; (2) that dissolved oxygen improves the flavor of tea. Both premises are demonstrably false.

Boiling itself does not remove dissolved gases. It is the change in temperature or pressure that affects the amount of gas that a liquid can hold (i.e., the solubility of a gas in a liquid). Solubility decreases as temperature increases. Assuming normal atmospheric pressure and composition, water at 0˚C can hold a maximum of ~15ppm DO, while water at 50˚C can only hold ~5ppm. Once the water reaches 100˚C, solubility is zero. Therefore, if you've brought water even close to a boil, you've removed virtually all the DO. What this means is that neither once-boiled nor twice-boiled water contain significant levels of DO, refuting premise (1).

Studies (Faust & Aly 1998, Pangborn & Bertolero 1972) have found that the level of DO in drinking water does not have a significant impact on its taste. It seems likely that the folk belief that DO improves water flavor results from the fact that running water (e.g., streams) is generally preferred to stagnant water (e.g., lakes), and is also higher in DO.

In principle, DO could soften the tannins in tea, just as decanting a bottle of red wine does. However, decanted wine contains much more DO, on account of its lower temperature, and wine is usually allowed to breathe for at least 15 minutes, compared to the 1-5 minutes that tea steeps for. Furthermore, tea drinkers can control the level of tannins in their cup via manipulation of steeping time, water temperature, and water/tea ratio. In short, a well-brewed cup has no need of oxidation.

Therefore, both premises that lead to the dissolved oxygen hypothesis are false. (1) Tea water does not contain a significant amount of DO, and even if it did, (2) there's no evidence that the level of DO has any impact on the flavor of the tea.

All argumentation aside, I simply cannot tell the difference between tea brewed with once-boiled water and tea brewed with twice-boiled water. I've done the tasting blind, more than once. In part, this post is a challenge to any believers in the dissolved oxygen hypothesis: try a blind triangle test. If you succeed in distinguishing tea made from once- and twice-boiled water, let me know.

All this is not to say that water is unimportant. Water is important. Alkalinity is important. Salt content is important. Minimal iron content is super important. Dissolved oxygen is not important.

Part II: Confounding Results

Here's the twist: I can easily distinguish between fresh tap water and water that has been boiled for an extended period of time. In the process of investigating the dissolved oxygen hypothesis, I boiled a small quantity of filtered water for 5 minutes, then refrigerated it until it matched the temperature of my tap water. I then drew some fresh filtered water, and tasted the two, blind.

There was a clear taste difference between the boiled and freshly drawn water. The freshly drawn water had a clean, crisp finish, while the boiled water had an off-putting twang to it. This was not a hard distinction to make.

I repeated the test using a different pot to boil the water, to make sure there weren't any contaminants in the first pot. Same result: the boiled water tasted worse than fresh water.

Given that DO does not have an impact on water flavor, what could explain the flavor impact of the 5 minute boil? Boiling can have a number of effects besides removing dissolved oxygen. It can also remove chlorine as well as concentrate dissolved minerals by reducing the water. But my water filter removes all detectable chlorine, and the short boiling time did not reduce the amount of water by an appreciable amount. Boiling can also remove calcium and bicarbonate ions (by precipitating temporary hardness), but my water is quite soft, and I have never noticed any scaling (which would indicate precipitation of temporary hardness.)

The only plausible explanation I can think of is that boiling the water would also have removed any dissolved CO2. Dissolved CO2 will form a small amount of carbonic acid, which can significantly lower the pH of very soft water (like mine). The fresh water would then have a lower pH than boiled water. Since most consumable liquids are at least slightly acidic, this might explain why the fresh water seemed to have a clean, crisp finish compared to the boiled water. This explanation predicts that less of a difference would be evident if more alkaline water were used, because alkalinity buffers against changes in pH.

Why then wouldn't removing dissolved CO2 from water also affect tea flavor? Well, the amount of dissolved CO2 in tap water is not enough to have a significant impact on the pH of any solution much stronger than pure (soft) water. A full explanation of this phenomenon would require an in-depth discussion of pH buffering, but this is why the pH of a brewer's mash depends much more on the alkalinity than the pH of the source water.

Even if small amounts of dissolved CO2 did affect the flavor of tea, heating water close to boiling will remove virtually all dissolved CO2, just as it removes DO. For brewing tea, coffee, or any other hot beverage, dissolved gases are irrelevant.

Faust & Aly, 1998. Chemistry of Water Treatment, 2nd Edition. p. 114 (
Pangborn & Bertolero, 1972. Influence of Temperature on Taste Intensity and Degree of Liking of Drinking Water. Journal of the American Water Works Association.