Info Center

Beyond Prestone - The Chemistry of Cooling

Lately, I've had the opportunity to look into the subject of those magnesium anti-corrosion plugs that are placed in the water jacket. The occasion was the first replacement of such a plug that I had put into service on my GTV in late 1990. At the time I replaced it, I had coincidentally just been asked about these plugs from a couple of members so my interest was running high. What I found when pulling the plug, so to speak, wasn't anything like what I was expecting. Just the opposite, in fact. Finding out just what was going on in my cooling system reads like a detective story and even put my college chemistry to the test!

So, do these plugs work? Let's find out...

Sigh. It used to be so simple. Once a year in the Fall, you'd drain the radiator, pour in a gallon of "antifreeze" and fill it the rest of the way with the garden hose. If you were really meticulous, you'd run a can of "fast flush" around first and maybe even risk second degree burns opening the block drain. That was "cooling system maintenance" in the good old days. When did it get so complicated?

Antifreeze has become "coolant" and is used year 'round. Manufacturers, especially those whose engines contain major pieces made of aluminum, admonish us to use nothing but distilled or de-ionized water with it in the system. Now we're supposed to be putting little plugs of magnesium into the water jacket to prevent corrosion. Is all this really necessary? In a word, yes...and no.

The Setup

Back in the summer of 1990 I began to lose coolant from my GTV at a slow but continuous rate. I did the usual things like checking all the hose clamps, looking for green stains on the block or green puddles on the floor, all to no avail. My detective work ended that October when the consumption rate suddenly jumped and I began to hear a sizzling sound in the exhaust manifold after shutting down. It turns out I had breached the cooling jacket around the #1 exhaust valve and was sucking coolant directly into the exhaust system! This is not an uncommon problem on Alfas whose owners have neglected the cooling system. The castings around the valve seats are some of the thinnest in the engine and are one of the first places to corrode through. But I am the original owner of the car and have flushed and changed the cooling system at the factory recommended intervals using nothing but coolant and de-ionized water the whole time.

Even though there was no sign of corrosion anywhere on the head, it was theorized that an inclusion (bubble) in the head casting may have made the jacket wall even thinner than normal, and after 15 years and 240,000 miles it finally corroded through despite the excellent care the cooling system had received. With this in mind, I decided to add a magnesium anti-corrosion plug as extra protection for the new head. [Just for the record, an autopsy on the old head revealed that it was not a bubble in the casting, but a crack that had propagated back from the left A/C mount that did the damage]

The technical name for these plugs is "sacrificial anode" since the magnesium sacrifices itself to the corrosion gods to protect the aluminum in your engine. Just how it does that we'll get to in a moment, but with a name like that you'd expect that the magnesium cylinder would be slowly dissolved away until nothing was left, right? So would I. That's why it was such a big surprise last May when I decided to replace the plug with a new one. Rather than a little withered stump of magnesium, I found a swollen mass of crusty white and tan deposits! They looked just like the stuff you find forming on the edges of faucets in hard water areas. What was going on here? To find out, I asked a couple of experts in the field of ion chemistry. To understand their answers, though, we've got to do a little refresher course in just what ions are.

Ion Drive

The world is made up of molecules and molecules are made of atoms. Normally, atoms and molecules are electrically neutral with the same number of positive and negative charges. But through some chemical reactions, they can be split apart and wind up with an unbalanced number of electrons. These electrically unbalanced atoms are called ions. If there is a surplus of electrons, the ion is negative; if a deficit, the ion is positive (and probably the only time you'll hear a deficit referred to as "positive"). Generally speaking, atoms don't like to be unbalanced and go about looking for ways to balance their charges. They do this by finding other ions of the opposite charge to bond with, or by finding molecules that have their electrons bound less tightly and grabbing one. Electricity, which can be considered nothing more that a stream of negative ions (electrons), will help these reactions right along. Mixing the right ions in a water solution and adding some electrical stimulus can do all sorts of wonderful things like plate gold out onto common metals. It can also eat holes in your cylinder head.

There is a definite pecking order in the chemical world which chemist call the electrochemical scale. The higher up the scale, the more aggressive the ion is. Those elements towards the bottom of the scale are always getting beat up by elements further up. These lower elements are said to be more "reactive" since they are always being forced to bond with the higher elements. Metals are among the most reactive elements and one of the most reactive metals is, you guessed it, aluminum.

The Source

So, where do these ions come from? Is there any way to protect against them? Well, I wouldn't be writing this if there weren't!

It turns out that water is a great solvent for a lot of materials. Common table salt, sodium chloride, will immediately disassociate into sodium+ and chlorine- ions when placed in water. Chlorine is a very aggressive ion and will form aluminum chloride when it comes in contact with aluminum metal. This is closely related to aluminum chlorhydrate, the underarm deodorant. Try to imagine the material in your Ban extra wide stick holding back the fires of combustion and I think you'll begin to appreciate the problem! There are plenty of minerals in ionic suspension in tap water as well. These are what give water its characteristic "flavor" and are partially responsible for making it taste different from city to city. You can bet that none of these minerals are doing your engine any good. Worse yet, there are plenty of chlorine ions in the water left over from the treatment process to kill bacteria. It's beginning to become clear why Alfa recommends only distilled or de-ionized water in the cooling system!

[This is a good point for an aside on purified water. Distilled water is purified by boiling it to steam, which leaves all of the ions behind to recombine into the crusty mineral deposits found at the bottom of your teakettle. De-ionized water, on the other hand, is purified by pulling the ions out by an electrical process. This leaves the water very neutral electrically, but non-charged components in the water, such as dissolved gasses (which we'll get to later) remain behind. This is a cheaper process than distilling and makes the water pure enough for anything short of laboratory use]

It would seem, then, that the best thing to put in our cooling systems would be pure, distilled or de-ionized water, right? Well, as the Wizard says, "not so fast!"

More Problems

Even ignoring the boiling and freezing problems associated with using only water, there still would be problems using just plain distilled or de-ionized water for cooling. When researching this article, I talked with Dick Hagen of Shrader Scientific who has made a career of making ions do all sorts of useful things. When I mentioned using de-ionized water in the cooling system, he blanched and related the following story:

Years ago, while working for another company, he was in charge of the high power radio frequency tubes used by the company in their processing equipment. These tubes were water cooled and to make sure the water was extra pure, it was supplied by an industrial de-ionizer (the same type that makes the d-i water you buy in the store). Most of the plumbing in the system was plastic except for one galvanized iron elbow and the silver cooling jackets around the tubes themselves. It seems that the iron elbow was always corroding away so finally they replaced it with a plastic one only to find that in one week the silver jackets became swiss cheese and the $10,000 tubes were ruined!

What had happened? Well, in the absence of other types of ions, water, known to some chemists as HOH (that H2O stuff is really an outdated 19th century term), will spontaneously decompose into hydrogen (H+) and hydroxyl (OH-) ions. The hydroxyl ion is almost as aggressive as chlorine and it was attacking the zinc and iron elbow. It left the silver jacket alone since silver is much further up the electrochemical scale, but once they replaced the iron elbow with plastic, the hydroxyl had nothing to go for but the silver.

Yikes! Does this mean that all the times I'd been putting de-ionized water into my cooling system (and recommending to others that they do the same) I'd really been eating my engine away? Fortunately, no. The problem above was that the system had an active de-ionizer in it. As fast as the water produced the hydroxyl ions which attacked the plumbing, the de-ionizer pulled them out (along with the metal they had attacked). If they had just left it alone, the water would have eventually saturated and the corrosion would have stopped.

The Solutions

Did everyone catch the moral of that story? It contains the explanation of how the magnesium plug works. As I said earlier, aluminum is one of the most reactive metals, so for this anti-corrosion scheme to work we have to find a metal even more reactive. We don't have much choice. Sodium is so reactive that it will grab oxygen out of the air and start to burn as soon as you open the lid. Potassium is even worse; it will burn under water! That leaves only magnesium which is reasonably stable in air and water, but still lower on the scale than aluminum.

So, finally, we get to the crusty brown deposits. To find out what was going on, I talked with Dr. Carl Nelson, a chemist with a computer disk manufacturer who holds several patents on the ion plating of disks and other materials.

"Yes, the magnesium metal is being sacrificed" Dr. Nelson said, "but it's not going anywhere. Here, let me show you."

With that he whipped out a pencil and scribbled down the equation:

2Mg + 2H2O + O2 = 2Mg(OH)2

[Actually, there were three equations ahead of this one showing how water splits into hydrogen and hydroxyl ions, and magnesium metal into magnesium ions and free electrons, but this one summed them all up.]

"You see how two atoms of magnesium will combine with two molecules of water to form two molecules of magnesium hydroxide?"

"Uh, yeah," I answered, "but what is that oxygen molecule (O2) doing in there?"

"Oh, that's just the dissolved oxygen in the water. If that weren't there, this reaction would give off hydrogen gas."

"Oh, great." I said, "It's not bad enough that the battery gives off hydrogen while charging, now we have to worry about the cooling system too?"

"No, not at all," he replied. "All water contains some dissolved gasses. Even distilled water, which starts out with very little, will pick it up on exposure to air [such as in a coolant recovery tank - JH]."

"But, back to the equation," I returned, "magnesium hydroxide is..."

"...this crumbly tannish white material on the plug" he completed, turning the lumpy remains of my $20 anti-corrosion plug over in his hands. "It's doing its job!"

The Bottom Line

So now we have professional confirmation that these plugs really work. But are they necessary?

You may have noticed something in our little chemistry lesson above: It was all done assuming that the reaction was taking place in just water. What we all have (or at least should have) in our cooling systems is a 50/50 mix of water and "coolant". How does that affect things?

Pretty marginal, as it turns out. The ethylene glycol in the coolant is pretty stable and doesn't really join into the reactions, not when you have reactants like magnesium and hydroxyl running around! Just to make sure, though, the manufacturers of automotive coolants add "anti-corrosion packages" to the mix. These are simply controlled amounts of different minerals already in ionic suspension, very likely some form of magnesium compound (although I didn't confirm this with any of the manufacturers). You can even buy them separately as a can of "corrosion inhibitor." They should provide enough source material to keep your system safe from the nasty hydroxyls for at least a year.


So, what should we do? The consensus boils down to this. You should change your coolant mix regularly. This means once a year if you live in a place that has winter, or you can go a bit longer in places like the Bay Area (I personally do mine ever 24,000 miles which works out to every 18 months). Between draining and refilling you should use something to clean out the system. If the old coolant is still clear then you could get away with just some "fast flush", but if the coolant is murky or rusty, then one of the "heavy duty" cooling system cleaners would be better. Just make sure it says "Safe for Aluminum" on it.

After flushing, fill with one gallon of coolant and top off with a gallon of distilled or de-ionized water. The latter is probably preferable since it already contains some dissolved oxygen. The four cylinder Alfa cooling system holds almost exactly two gallons, but you might have to run the car a couple of times to work out all the trapped air before you can fit the whole gallon of water. Never use water from the garden hose to do the final fill since you have no idea what's in it besides water, and believe me, there's plenty (remember the chlorine?). It's ok to use it for the intermediate flushes, though.

If you're loosing coolant, don't just top off with water since you'd be diluting the coolant thus reducing both the thermal and chemical protection properties. Try to keep the 50/50 mix up.

Finally, do you need the magnesium plug? Well, that's strictly personal choice. I ran my engine for more than 15 years and 240,000 miles without one and there wasn't a trace of corrosion on the head when I took it off. On the other hand, looking at the amount of magnesium reduced to lumpy white powder in just a year or so, convinces me that something powerful is going on in there. Personally, I wouldn't be without one.

    -  Jack Hagerty

Info Center