There have been several questions regarding the safety of RO or DI water
for human consumption. The issue is on some levels simple, on others
quite complex. Since what goes for humans also goes for fish, I suppose,
this isn't completely off-topic. And I do say something at the end about
validating the performance of your membrane.
As many people have noted, RO (or nanofiltration) membranes can be used
in emergency situations to turn non-drinkable water into substantially
safer water. A correctly functioning RO membrane rejects essentially all
pathogens from the input side. Viruses, bacteria, giardia, protozoans,
spores, etc are all far to large to pass through an intact membrane, and
you ought to get 6 logs of rejection at least in membranes that are in
I keep explicitly mentioning the condition of the membrane, because it is
crucial in this regard. If a TFC membrane is compromised, then you have
a rather bad situation, from a human health perspective, which is
dechlorinated water which contains potentially not great bacteria or
other bad guys.
Does that mean that you should or shouldn't drink RO water? I wish that
I didn't have to make that call based on commercial units on the market.
What I'd like to be able to point to is a unit with a recirculating
reservoir and a TOC lamp right before the tap. TOC lamps are like UV
sterilizers with a bad attitude. ;) The jackets are made of high purity
quartz and they pass the ~190 nm Hg emission line. That generates
powerful oxidants (ozone) in water, sterilzing it and burning off all
residual organics. I would drink water from a unit like that, no problem.
Why am I being so wierd about this? Because there is no control of
bacteria growing on the supply side of TFC membranes, little control for
CTA membranes. Some bacteria can grwo and multiply in RO water (which is
not devoid of minerals, by the way, it just has less minerals.) I've
seen cyanobacteria growing in a glass container of triple distilled
water, living off of the minerals leached out of the glass container.
Would I drink RO water out of the hose of my unit? I would use it in
coffee, where it is boiled before consumption. I might be ok with
drinking from a refrigerated jug of it in the refrigerator, but I don't do
that. I think that someone is missing a market for a recirculating RO
reservoir with an integral TOC lamp.
How do you know if your RO membrane is OK? There isn't any easy way to
measure how many logs of rejection you get for bacteria, etc, but you can
certainly measure the rejection ratios for various nutrients or hardness
with hobby test kits. In NYC, the tap water is 1.04(2) mg/L phosphate.
A good TFC membrane gives about 50-fold rejection of phosphate, so I
simply do a phosphate determination occasionally to verify membrane
performance. Depending on your water supply, you can do the same with
nitrate or total hardness test kits. If you have an RO/DI unit, you need
to do this test on water before it goes through the DI cartridge.
You can also use sensitive chlorine test kits to validate the activated
carbon prefilter performance. This might save an RO membrane for
someone. With conventional units, that requires you to disconnect the
line leading to the RO membrane, and do a chlorine test with the lowest
range test kit you can get.