Firstly, it is best to consider what RO is ... Reverse Osmosis. Ok, so somehow by having the opposite of osmosis we get clean water. But what is osmosis? For some reason I still remember the definition from school, it's "the movement of water from a less concentrated solution to a more concentrated solution across a semi-permeable membrane". Semi-permeable means that only some compounds can pass thru. By concentrated I mean that more or less "stuff" (chemicals) is dissolved in the water.
What does that mean ... well on one side of a piece of cellulose we have a
pure-ish water and on the other "dirtier" water. Normally we would
expect the clean water to move across the cellulose membrane and go into the
dirty water.
From: http://www.chemistry.nmsu.edu/Instrumentation/RO1.gif
And to go off at a tangent, osmotic shock. You know we have to spend ages and
ages acclimatising stock from the water their come from to your tank water.
A lot of it is due to osmosis. Cell walls are semi-permeable ... so we have
osmosis wanting to take place (its called osmotic pressure). So pure water will
pass in or out of the animals/plants cells ... trouble is that the cell walls
are only so stretchy and the cell can only cope with having its inside so concentrated,
it does however have an active (i.e. takes energy) process to self regulate
but its not the fastest process .. certainly not as fast as osmosis. Therefore
take one nice happy cell, lets say a particularly nice piece of Caulerpa (hey
we're not going to use any Acropora for this test :) ) and stick in a solution
that is either more concentrated than the cell was used to (aaargh as too much
water is "sucked" out of the cell .. ooops , dead) or less concentrated
than normal ... wheee, too much water rushes into the cell and bang goes the
cell wall ... ooops, dead.
Enough of this aside, back to RO units ....
So if we can find a way of reversing this then we can go from dirty water (tap water) to clean water. Ah ha ... an RO unit. In effect, all an RO membrane is a coffee filter but with really tiny holes (I'll come back to this point). Stick in enough pressure and it will force H20 thru the holes. Now the rate that H20 goes thru the membrane is dependent on the temperature. Lets just think of a tub of lard (hmmmm, lard), ... I want you to try and pass that lard thru a plastic sieve. Not going too well ... ok, heat it up a little ... ooh it starting to melt and a bit goes thru, bit more heat and its now a liquid ... goes straight thru doesn't it, even more heat ... oops melted the sieve, doh! Same story with water, the warmer it is the "thinner" (less viscous) it is, right up to the point you destroy the membrane. And just become I love boring facts, for every degree C the temperature drops, about 3% less water comes thru the membrane … boring fact or wot!
I think what we'll do is build our own RO unit (ok this won't work but it's the right idea). Get a piece of kitchen paper (yes right now!), some 1" tube and a 1" T -piece. Using a rubber band cover one of the outlets of the T-piece. Now attach the T-piece to a water supply and turn on the water. You getting any water come thru the paper? Maybe a drop. Ok, lets get a valve and attach it to the unused outlet of the T-piece. Slowly throttle the valve, and what do you know we have water starting to come thru the paper (we've increased the pressure in the T-piece). But notice how much water is going straight through your T-piece and compare it to how much comes out the paper. Not looking too good huh? Well even with a proper membrane for every drop of clean water you collect, you are going to waste anyway between 3 (I wish!) and 10 drops. Now keep on closing that valve … drat, the paper has just been torn apart, too much pressure.
Right so now we've built our own RO unit, lets see how its done properly, not using kitchen paper that's for sure! An RO membrane is a series of thin films bonded together and wrapped around a plastic tube. Now much like the paper, there is a maximum pressure that they can cope with, this can actually be incredibly high (250psi) but the rest of the unit certainly can't with around 100psi being the maximum, but you want some tolerance so that's around 60psi. For temperature then about 24 or 25C is about right (75 deg F) with 29C being about the upper safe limit (yeah like you are ever going to have high temperature problems!). With respect to water pressure, watch that the water supply to the line doesn't suffer from "hammering" (e.g. you don't want it next to a washing machine which can suddenly open and close its inlet. The sudden change in pressure can quite easily create a short period of very high pressure.
To be totally accurate there are RO systems for industrial use that don't obey these rules at all, but you won't come across those unless you need 1000 gallons per day!
Now for the systems we do see, then as time passes, we do see changes in the
types of film used in making the membrane and slowly but surely the technology
has improved, so ions like silicates are better rejected. Now whether it is
worth paying extra for such a unit really depends upon your tap water. The Gee-Whiz
factor is impressive but if you contact your local water company, they will
supply you with a water analysis. If you haven't got high silicates to begin
with then the only thing your high silicate rejection membrane is doing is making
your retailer very happy.
One point that I should come back to, I described RO membranes as a filter with
tiny holes. Well yes they are, and not they are not. The membranes are good
at filtering contaminants that are ionised. Normally this isn't a particularly
important point for your purposes, but be aware, even though a bacterium is
far larger than a water molecule, to a small extent they still can get through
the membrane. Be careful if you store the RO water and then use it for drinking
as there isn't any chlorine left in there to control the growth of these bugs.
One point though, if you are storing RO water for use in your tank then you
could easily have high (and maybe nasty) levels of bacteria growing in there.
Although calcium hydroxide isn't a superb bactericide, it certainly can make
a significant impact and yet another reason why kalkwasser (limewater) is a
good idea.
When it comes to the hobbyist use of RO, then a whole lot of workarounds are used (actually its just plain bad practise but let's not quibble). The most obvious is the usage and maintenance practices. First off usage. Given time, solids will come out of the water and settle on the membrane. Calcium carbonate and calcium sulphate are obvious ones as well as bacteria. Do we do anything about the reducing the hardness of the water before it gets to the RO? Precisely … if you can, get a water softening unit in front of the RO. We really aren't too concerned about sodium ions as there are billions of them in your tank and the odd few more won't make a different. A blocked membrane (which leads to it failing, which leads to it no longer producing clean water) is definitely a bad thing.
Another "workaround" is the intermittent kit. But I only need 10 gallons a week and this RO unit can produce that in a day, we go. Yeah, but what are you going to do for the other 6 days. Some bright spark knew that you couldn't just turn the unit off since the membrane fails if it dries out so we get some plugs so we can just turn the unit off and keep the water in the unit. Bad move on a couple of fronts. Remember all that bacteria in the water … well now you aren't even flushing them out. Not good. Also, recall that solids slowly fall out of the water and block the membrane although most of the time you are flushing those through as well. You are not now! Not good. If you must turn off the unit then take out the membrane and firstly put it in a bucket of RO water for a few hours, as least most of the solids will end up in the RO water and not on the membrane. Even better is a flushing kit which at low pressure pushes clean water into the "clean side" of the membrane and out via the "dirty" side. Afterwards put it in a Ziploc bag and stick it in the fridge, then at least the rate of bacteria growth is slowed.
There is a whole world of "proper" ways to look after an RO membrane, but for our purposes, these few simple steps and the realisation that we go through membranes faster than needs be (so look at 12-36 months, depending on how hard and how "dirty" your tap water is) is the best approach.