[Sneap] De-ionized water

[Klaus Bahner]

Brett,

> My suspicion  is that there's a sort of positive feedback 
> between loop conductivity and corrosion- as the loop gets a 
> little bit conductive, corrosion starts, adding more ions to 
> the loop, increasing conductivity which increases corrosion. 
> Any ideas on this?

One has to distinguish between the "corrosive" effects of external
electrolysis and ordinary corrosion. Let's first focus on corrosion only,
this is cooling of equipment where there are no voltages involved. In this
case the necessary condition of any corrosion to happen is availability of
solved oxygen in the cooling water. Oxygen is solved in tap water, which
means tap water will always be an corrosive agent. Most materials exposed to
water protect themselves from further corrosion by developing a passivation
layer, the best known is the calcification of galavanized steel tubes. It is
this calcification which for example allows the usage of steel tubes for
transport of oxygen containing water in the first place. The same yields for
copper, brass and aluminum, which protects themselves by calcification and
oxidelayers. 
In cases where those passivation layers cannot be generated or are not
stable, you will see corrosion. Passivation layers are influenced by many
factor, for example temperature of the water, chlorine content and
dezincification in case of (inferior) brass alloys. There are probably many
other factors I don't even know of. 
DI water has got the tag of being corrosive, because DI water tends to
destroy and/or prevent the generation of protective passivation layers. It
is not the case that DI water is corrosive in itself, but it excalerates
corrosion if oxygen is present, because the wall materials cannot protect
themselves.
However any corrosive effect requires oxygen, hence in a closed loop system
you will have no corrosion once the solved oxygen has been removed. However
please note that many plastic tubings allows oxygen to pass through! Hence,
just using a closed loop system will not prevent corrosion if the tubing or
an open reservoir allows oxygen to diffuse into the water. In such a case
using DI water, will cause more servere corrosion than ordinary tap water,
because the oxygen will be able to attack wall materials more easily.

In the above case conductivity of the coolant is not a too important
parameter. It gets a very important parameter though, when the walls of the
cooling water system are exposed to a different voltage (typically heatsinks
in magnet power supplys). Then the conductivity of the water will cause
electrolysis which will gnaw at the wall materials.  As the conductivity of
the coolant is increased by electrolytic process, there will be a positive
feedback. 

In conclusion you should take measures to ensure that no oxygen is added to
the water, i.e. closed loop with diffuson tight materials. Last but not
least refilling water typically adds oxygen rich water again, hence when you
have to refill often a closed loop system, it doesn't prevent corrosion.

In case of electrical voltages involved, DI water is a good choice, but only
in the case that you can guarantuee that no oxygen is admitted to the water.
Otherwise it will act very "corrosive" due to the lacking protective layers.


    
Regards,
Klaus


--------------------------------------------
AMS 14C Dating Laboratory
University of Aarhus
Ny Munkegade, Byg. 1520
DK - 8000 Aarhus C
Phone: +45 8942 3736
Fax:   +45 8612 0740
--------------------------------------------