# Can you make deuterium-depleted water through a freezing method?

Deuterium-depleted water (DDW) is sold commercially by some companies. It is apparently manufactured by fractional distillation, which leverages differences in boiling points of heavy and light water.

My question is this: can DDW be manufactured through a freezing method? Heavy water has a freezing point of 38.8 °F and normal water of 32 °F. If a jar of water were kept in a water bath of say a constant 35 °F, wouldn't the heavy water freeze and sink to the bottom of jar, allowing the DDW to be siphoned off of the top? I am surprised there appears to be no discussion of the merits or demerits of this idea anywhere on the internet.

• – Mithoron Feb 4 '18 at 22:35
• Here is a link to a clinical trial: ncbi.nlm.nih.gov/m/pubmed/23441611 – user69999 Nov 9 '18 at 8:14
• No, slightly deuterium enriched ice will not sink. Tap water is around 150 part per million (1,000,000) or 99.985 % H2O. At these concentrations deuteriated water will most likely be of the form DHO vs D2O. DHO has a freeze temp between D2O and H2O and it has a very high chance its going to freeze to other H2O molecules forming a water crystal that has a lower density than liquid water, thus it will float. – Keith Reynolds Apr 22 '19 at 17:04

Not easily. For all practical purposes, $\ce{D2O}$, $\ce{DHO}$ and $\ce{H2O}$ behave the same -- including complete miscibility. Though some small fraction of the liquid that freezes first would be enriched in $\ce{D2O}$, and to a smaller extent in $\ce{DHO}$, it would still be composed largely of $\ce{H2O}$.

According to Wikipedia, the Girdler process is the most energy efficient means of separation.

To clarify, the objective is not to obtain heavy water. It is to obtain light water. So it is not an issue if the ice contains only some deuterium, but mostly light water, as the point is that the liquid water now has less deuterium than it did before.

I actually just got a reply from a DDW manufacturer who said the following:

"Deuterated ("heavy") water has a higher freezing point than ordinary water, there is indeed a fractionation slightly above zero Celsius, but it does not mean that all heavy water will be frozen.

By holding the temperature at let’s say 1°C, you can get some ice that contains D2O or DHO in a higher concentration than in liquid because the freezing point for H2O is lower.

This way you can reduce the D-concentration of the water with 8-10 ppm in one step. If you wish to achieve further decrease you have to freeze this water in further steps.

According to our knowledge, fractional distillation is the best way to produce DDW in large scale."

This same manufacturer says that home water distillers can at best reduce deuterium concentration by 1 or 2 PPM per pass, which is much less efficient than commercial evaporative methods. But an 8 to 10 drop in PPM from a freezing method (per pass) is considerable (if true), as the early clinical trials in this field suggest that even drops of say 20ppm are potentially clinically significant.

• "Clinical trials"? This would seem to be pseudoscience aimed at marketing. – DrMoishe Pippik Apr 8 '18 at 4:29
• @DrMoishe Pippik, for most, that statement would mean 150 ppm average tap water to 130 ppm. But I think deuterium depleted water used in various studies used 120 ppm to as low as 20ppm. Anyhow the intent is to off set food preparation enrichment of deuterium. Have those clinical result been used to advance and market deterium depleted water, sure. How important are the clinical results, and their meaning, that is really another discussion. – Keith Reynolds Apr 22 '19 at 17:24

I would say. Diffusion rates play a role. The closer the surface of ice formation is to the freezing point of deuterium water. The purer the deuterium ice will be. The process will be slowed because of this but a purer form of deuterium depleted water would result. Since I have no experience this is speculation on my part. With the formation of ice some heat would be liberated and a natural form of convection would occur.