Date:         Thu, 17 Feb 2000 16:41:38 -0700
Reply-To:     Keith Adams <keith_adams@TRANSCANADA.COM>
Sender:       Vanagon Mailing List <vanagon@gerry.vanagon.com>
From:         Keith Adams <keith_adams@TRANSCANADA.COM>
Organization: TransCanada
Subject:      Re: Corrosion of heads
Content-Type: text/plain; charset=us-ascii

I'm supposedly a metallurgical engineer, so I'll see if I can shed some
light.  This is more about general corrosion, as I don't know much about
corrosion in engines, so take this with a grain (a large, large grain)
of salt.

Corrosion only will take place given 4 components 1) anode (metal that
is corroding - in our case, the aluminum heads) 2) cathode (site where
ion transfer occurs - in this case, the iron in the engine) 3)
electrolyte through which electrons and ions can transfer (a liquid
conductive medium - coolant) and 4) an electrical connection between
anode and cathod (metal to metal contact doesn't get much better than
that).

All metals have an electrochemical potential measured against a standard
electrode, going from gold (most noble or cathodic) to lithium (most
reactive or anodic).  Of importance to us is that Iron is above Aluminum
which is above Zinc which is above Magnesium.  This is in seawater,
which is different than the standard emf series for metals in their own
ions. What this means is that if you put two of these metals in
seawater, and connect them electrically, the higher metal will not
corrode and the lower one will.  The further apart on the table they
are, the more severe the reaction is.

So let's pretend we're looking at the standard emf series (metals in
their own ions, not in seawater).  Now, to get corrosion between two
metals (galvanic corrosion) you have to come out with a negative number
between the anode and the cathode.  If we have Fe2+ going to Fe and Al
going to Al3+ you get standard potentials of -0.447V and +1.662V
respectively.  So Ecell = (-0.447) - (+1.662) = -2.109 V.  This is a
mega-spontaneous reaction.

SIDENOTE: wanna see something cool related to that reaction?  Take Rust
(Fe2O3) and pure aluminum powder (Al) and put them in a ceramic
crucible.  Kinda mix and stick a birthday sparkler in.  What you get is
a mega-spontaneous and very, VERY dangerous uncontrolled reaction as the
aluminum takes all the oxygens away from the rust and produces iron.
This is called "the thermite reaction".  DO NOT TRY THIS!!!!  I have
seen molten iron shot 4 stories into the sky from this.  It just goes to
show how powerful galvanic corrosion can be.

Ok, back to the story.  Now, if we stick Zinc into the system, which in
lower than aluminum in the series, we have two possibilities.  Either
Iron will corrode Zinc or Aluminum will corrode Zinc.  Since Iron and
Zinc are further apart, that reaction has a higher potential (more
spontaeous) and will take place first, thus preventing the aluminum from
corroding *as badly*.

Now will this save our wasserleaker heads?  Maybe.  Why only maybe?  I
thought you just said that iron and zinc will corrode away happily?

Ok, well this works well for exposed areas, like say sacrificial anodes
in a tank.  This is not our exact problem.  I've never taken off a
wrecked head, so I can only speak for what other people have said, but
we see crevice corrosion.  This is small pits which grow under the head
gasket mating surface.  Here is a quick description of what crevice
corrosion is, borrowed from the book "Corrosion Control" by Samuel A.
Bradford (one of my professors at University), p84

"When corrosion begins, the crevice corrodes uniformly just as the metal
outside the crevice does, but in time the O2 within the crevice has all
reacted and no more O2 enters the crevice because the crevice is so
narrow.  The cathode reaction continues outside the crevice, however,
with electrons being supplied from the anode within the crevice.
Corrosion outside actually decreases because of the electron current
coming from the crevice."

Acids and chlorides increase this reaction dramatically.  Now, what we
really have in corrosion happening on a very small scale, and since the
effectiveness of sacrificial anodes decreases with distance, our anodes
(probably) won't do much to stop this highly localized attack.  An
impressed current in the heads may help, but since the engine is already
charged since it is part of the electrical system that is unlikely too.
So reducing any agents in the coolant which will contribute to this
corrosive attack will help some.

In summary, you can do a little to *slow down* the corrosion, but you
can't work around a design flaw.  I hope this helps.  Sorry it's rather
rambling, but I'm trying to run out the door as I write!

Cheers!
Keith in Calgary, Alberta