304 vs 316? Does Molybdenum play a part?

The peculiarly aggressive corrosion of climbing anchors on certain sea cliffs can be attributed to sulphate reducing bacteria (SRB), but what of the distinction we see between 304 and 316 grades of steel?

Introduction:

Many years ago, when the crags of Tonsai/Railay began eating 304 stainless steel anchors it was felt that swapping over to the marine grade, 316, should certainly solve the problem… after all, it’s a marine environment, right? But no. Anchors still continued to fail.

Titanium seemed to holdup ok, but was often rejected as too expensive, with the result that the circus of installing stainless steel and then watching it fail continued for decades.

At some point, many years on, a few people began to realize that much of the steel labelled as 316 was in fact substandard, and there was a possibility that true 316 may be ok . Maybe 316 wasn’t failing after all. So this brings us to a situation where –

To date, I am yet to identify SRB attack on an anchor that analyses as 316.

Of course, absence of proof does not translate to proof of absence and I won’t be happy until I can provide a theoretical basis for the observed difference. Let’s take a closer look.

Listed below are the key compositional differences between the two alloys.

Here we see that both nickel content and molybdenum content possibly could be involved.

To date I have concentrated on understanding the critical impact of nickel content on the austenite stability of the 300 series alloys. This is a big subject, and I have covered it in this post. In brief, as nickel content falls below 10%, the susceptibility of a cold-worked product to hydrogen embrittlement increases drastically – drastically as in, 10% Ni is ok yet 9% is not. The knock-on effect is that we see acute sulphide stress cracking (SSC) of cold-worked 304 but not 316 in situations where exogenous sulphate sponsors attack by SRB.

Whilst I remain confident of the role played by the nickel content in offering resistance to SRB attack, I have no good answer to folks who cry “But what of the molybdenum? Surely it does something?”. Good question. My instinct is that it is irrelevant, but can such an assertion be justified?

But what of the molybdenum?

Typically, 18-8 stainless steels suffer pitting corrosion in chloride-rich environments such as sea water. This disadvantage can be offset by incorporating a small amount of molybdenum into the alloy. Thus 316 with its 2% molybdenum finds use in marine applications.

Despite extensive research, the underlying nature of this protective mechanism is not well understood, although we do know it involves an increase in the electrochemical barrier to pit initiation, and, given the fact SRB attack is not known to have any obligatory requirement for pit initiation, its mechanism being quite different, I am inclined to disregard its relevance. See for a recent contribution.

Further to this, exogenous sulphate, in its own right, suppresses pitting corrosion, and thus there are strong reasons to suggest that, without cost penalty, we could replace the molybdenum of 316 with further nickel to provide an alloy better suited to sulphate sea cliffs.

I don’t think its likely that the well known suppression of pitting corrosion by molybdenum plays a part in the observed resistance of 316 to SRB attack.

Returning to the issue of austenite stability, which is something I do believe is a critical factor, it is worth noting that 2% of molybdenum stabilizes austenite to the same extent as 1.3% of nickel. You can deduce this from the following recipe for the relationship between elemental composition and martensite transition temperature, .

M_{30/50} = 551 - 462(C+N)-9.2Si-8.1Mn-\\29(Ni+Cu)-13.7Cr-18.5Mo

Molybdenum will play a role in the resistance of 316 to SRB attack by increasing its austenite stability by 1.6 nickel percent equivalents.

To understand what I’m talking about above you, need to read this post.

Molybdate inhibits sulphate reduction

What else might molybdenum do? Well, the natural world always has surprises in store for us. It turns out that the biochemical mechanism used by SRB to “eat” stainless steel is vulnerable to the presence of the molybdate ion.

All sulphate reducers share the same archaic set of genes. These code for the enzymes responsible for the sequence illustrated below – Wikipedia.

It is the first step, the conversion of sulphate to adenosine 5′ phosphosulphate that is of interest. The paper by provides a lot of detail on how molybdate can impact this.

SO₄⁼ is clearly analogous to MoO₄⁼ and thus it is not surprising to discover that the action of molybdate is to compete with sulphate at the binding sites of sulphate adenylyl transferase. Thus, what we see, is competitive inhibition, rather than toxicity per se.

Does this experiment mean that bolt environments containing around 1mM of molybdate are going to suppress the growth of SRB? Normally I’d be skeptical about drawing conclusions given the complexity of such an environment, but, there are some indications that the extrapolation may be justified. Firstly, the authors tell us that the enzyme system in question, sulphate adenylyl transferase / ATP sulphurylase (sat/atps), is cytoplasmic, and secondly, most SRBs have active molybdate transport systems to acquire the ion from the outside environment. Thus, it is reasonable to assume that any molybdate arising from oxidation of the metal of the bolt will make its way into the site of action at the sat/atps enzyme complex.

Is there sufficient molybdenum in 316?

Let’s start with a rough calculation of molybdenum availability based on the small dimensions that are presented by anchor geometry.

1mM contains 96mg of Mo in 1000cc of water.
∴ 1000cc of solution needs to contain 4.8g of the 2% alloy.

The density of 316 is 8g/cc.
∴ 1000cc of solution needs to contain 4.8/8 = 0.6cc of 2% alloy.

Now consider a 1x1x1mm volume of liquid in contact with metal within the environment. - this needs to contain 0.6/1000 = 0.0006cc of dissolved metal to give 1mM of Mo.

Next assume that no more than 0.01cm of metal thickness will be dissolved, in which case, the required metal contact area becomes 0.0006/0.01 = 0.06sqcm = 6sqmm.

Is it feasible to wet 6sqmm of metal with our 1x1x1mm drop of water to give our required concentration of 1mM of Mo? 

To a first approximation... yes. 

Based on this rough calculation, it is clear that even a light 2% alloy like 316 would contain just enough molybdenum to achieve inhibitory levels of molybdate in the aqueous films bathing it.

However, the SRB environment is anoxic, so it is reasonable to question whether the oxidation of metallic molybdenum to molybdate ions can occur to any real extent.

We know that when the SRB is actively attacking the metal surface of, say, a 304 bolt, we can measure the pH as being less than 4.0, and we can also observe that the metal surface remains free of sulphide product, indicating the redox potential must be less than -0.5V SHE as per red asterisk in the Pourbaix diagram on the left.

By reference to the Pourbaix diagram on the right, it is clear that the active SRB environment for iron, shown by the red asterisk, falls well short of the conditions required to oxidise metallic molybdenum to molybdate.

Given that we calculate that inhibitory levels of 1mM of molybdate will arise only if the small amount of metallic molybdenum available is completely oxidized, it seems unlikely such could ever occur while there is metallic iron available. Thus –

Based on redox chemistry, is seems implausible that the attack of 316 by SRB would be inhibited by the molybdenum content.

What evidence do we have from the field?

It must be so that, world-wide, thousands of stainless steel bolts have been installed on sulphate crags, and equally anecdotally, it must be that the large majority have failed after 10 to 20 years of service.

So the hard evidence needed to make a comparison between 304 and 316 is out there, but we lack the means to collate it.

At the very least, I need someone to send me a corroded bolt so that I can establish the following –

1. Nickel content
2. Molybdenum content
3. % Martensite by magnetic susceptibility
4. Evidence of brittle fracture or stress cracking
5. Positive response to Iodine-Azide Reagent at the fracture
6. Presence of greigite within the anoxic zone of the fracture

Only then can I assert, with some surety, that this piece of 304 or 316 or whatever has failed under SRB attack.

Currently, we have a disconnect between the large numbers of failures that seem likely to be SRB attack on 304, compared with the relatively few I have been able to test. I’ve done maybe ten or twenty. So how many more do I need to do before we can say there is no doubt that 316 is more resistant?

Add to this inherent sampling uncertainty, there is a sampling bias in the favour of 316 that needs to be considered. The reasoning is as follows. If we go back say twenty years or more, it is likely all the stainless steel being installed was 304, regardless of what it said on the label. Within the past five years, the market has changed, and we are now seeing mainly 316, and possibly quite good 316 at that, being installed. As the ratio of 316 to 304 increases, then proportionately more 316 is being put to the test, and only then will we really know the truth.

To date, I am yet to identify SRB attack on an anchor that analyses as 316. But, maybe there hasn’t been that much true 316 installed? The next ten years might tell a different story.