HYDROGEN DAMAGE
Scrubbing by Bubbles

Everyone knows concentrated acid eats its way through steel, right? (Think of the alien in the movie Alien.) So they must use some pretty exotic material in a sulfuric acid application to keep from eating up the pipes, right? Well, it turns out plain old ordinary carbon steel can be used for handling concentrated sulfuric acid because the acid reacts with the steel to form a protective film of iron sulfate that reduces the corrosion rate to a tolerable level (approximately 0.020 inch per year).

So it was with interest that we recently took a call from a client with a carbon steel pipe that had failed in service while handling concentrated sulfuric acid.

A preliminary examination revealed a deep, longitudinal groove along the inside of the pipe, with a portion of the groove completely through the pipe wall (top figure above). Our first thought was that this might be just a bad seam weld, but cross sectioning (figure below) showed there hadn't been a weld at that location at all. So what had happened?

Things got much clearer when we studied the pattern of attack in more detail and found not just a single groove, but a whole series of grooves radiating toward a central longitudinal groove (bottom figure above). Then, talking to the client about the service orientation of the pipe, we learned the pipe was horizontal in service, with the groove located at the top of the pipe. How did all this help us? The short answer is it allowed us to identify the cause of the problem as hydrogen grooving.

For a more detailed explanation, you need to first realize that although the overall corrosion rate of carbon steel in concentrated sulfuric acid service is low, some corrosion still occurs, and, in addition to iron sulfate, the other reaction product from the corrosion process is hydrogen gas.

Hydrogen gas can be a problem with carbon steel in sulfuric acid environments because it can literally scrub off the mechanically weak iron sulfate film, which is the only thing protecting the steel from attack.

During periods of normal flow, the gas bubbles are very small and evenly distributed, and they are quickly carried downstream along with the acid, generally without causing a problem to the pipe wall. But in this particular application, the flow was periodically interrupted. When the acid ceased flowing, the gas bubbles accumulated along the pipe wall in the upper half of the pipe. When the flow resumed, the hydrogen gas bubbles were dislodged and rose to the very top of the pipe, scrubbing off the protective film along the way. With the film gone along the top of the pipe, the surface corroded very quickly until the sulfate film reformed.

Repeating this process over and over again resulted in a pattern of curved grooves in the top half of the pipe, all radiating toward a central longitudinal groove at the very top. Eventually, the central groove penetrated completely through the wall, resulting in a leak. (And yes, as you might imagine, a leak in a sulfuric acid line has a way of attracting peoples' attention.)

Although we're usually not privileged to see such an excellent example of hydrogen grooving, it's not an uncommon occurrence in sulfuric acid service, particularly under upset conditions. To guard against it, the usual recommendation is to use very thick steel, in recognition of the possibility of localized corrosion, and to keep the average velocity in pipes below 3 ft/sec to reduce the scrubbing effect of any hydrogen bubbles that do form.

Even under these conditions, however, rapid, localized attack can still occur if the fluid flow is disturbed. Features such as short-radius elbows or gaps at socket welds can cause trouble.

Another application where hydrogen grooving is a serious problem is in the upper half of railroad tank cars transporting concentrated sulfuric acid. Because very thick carbon steel would make the tank cars unreasonably heavy, the "make it extra stout" approach to mitigating the effects of corrosion can't be used; instead, the cars have to be protected from grooving by applying a baked phenolic coating on the inside surface.

Counter intuitively, carbon steel won't work for storing sulfuric acid at lower concentrations—such as the acid in your car battery. Why not? Because at lower acid concentrations, the sulfate isn't able to form a stable protective film, without which, the steel continues to corrode quite rapidly.

Interestingly, this wasn't the only project we worked on recently involving a problem with hydrogen. Although this project involved damage from hydrogen in gaseous form, this isn't the only way that hydrogen causes problems, as our next story illustrates.

Delayed Failure of Socket Set Screws

Recently, a client called us to say the socket set screws he was using to attach the ceiling panels in a microelectronics clean room were failing before the clean room had even been put into service. Puzzled, he reported they weren't breaking during installation; rather, they were failing a few days after installation, with a failure rate of several percent. Adding to his puzzlement, he noted he didn't think his installers were at fault, as they were using calibrated torque wrenches and were torquing the screws to only about 4 percent of the expected tensile strength of the fasteners. Clearly, something was seriously wrong, but what?

The screws were made of low alloy steel, electroplated with zinc for corrosion protection. We began with a detailed analysis of a failed screw. At the same time, we set up a laboratory test on several unused fasteners in which we torqued them to various amounts and then let them sit under load for a while. Most of the fasteners tested in the lab failed, some in less than an hour!

 

The failed screws, both from service and our lab tests, had brittle, intergranular fractures that started at the roots of the threads. When magnified at 1,000X in the scanning electron microscope, they showed the classic "rock candy" appearance characteristic of an intergranular failure. (above left)

Further in from the thread root, the fracture mode on the lab fractures changed to mostly ductile dimples with a few small quasi-cleavage facets (above right). Ductile dimples are what we expect to see on a steel fastener. (Dimples resemble the overlapping pattern of scoop marks you see in your ice cream carton after you've dished out a few helpings.) They result from the initiation and growth of microscopic voids that form at impurities called inclusions (which are always present in commercial steels).

If the fractures had occurred by twisting the fasteners until they broke, the dimples would have had a different shape. So, we could tell the failed screws hadn't been overtorqued. Further, we didn't find anything wrong with the material used for the fasteners; the microstructure was tempered martensite, just as would be expected, and the hardness was 51 on the Rockwell C scale (HRC), consistent with the specified range of 45 to 53 HRC (or an approximate tensile strength in the range of 215,000 to 283,000 psi).

So what had caused the failures? They were the result of delayed fracture due to hydrogen embrittlement, a time-dependent, brittle cracking phenomenon that occurs under sustained load in high strength steels. So, why does hydrogen embrittle steel, and how much hydrogen does it take to cause trouble? Hydrogen is present in monatomic form in steel, dissolved within the matrix of the metal. That is, individual hydrogen atoms sit in between the iron atoms that make up the crystalline structure of a steel.

Now, in order to insert a hydrogen atom into a steel matrix, the iron atoms have to be forced apart a little bit, so a hydrogen atom would rather sit at a location where there's a tensile stress, like at the root of a thread or other notch or at the tip of a crack. Why? Because the tensile stress increases the distance between steel atoms, making for just a little bit more room there.

One of the problems with hydrogen atoms in steel is that they're so small that once they enter the steel they can move around easily, even at room temperature, and find their way to the areas of highest stress. So just like a new puppy left alone in your house for the day, the hydrogen goes to exactly where you don't want it to be, and then causes trouble. Other "bigger" interstitial atoms like carbon and nitrogen are so much less mobile than hydrogen that small amounts don't cause the same trouble.

The effects of hydrogen in steel are very complex, to the extent that some materials scientists have spent their entire careers just studying the phenomena, but perhaps the simplest way to state it is to simply say that its presence affects the way the material responds when it tries to deform permanently, thereby making it more susceptible to brittle fracture.

How much hydrogen it takes to cause trouble depends on how strong or hard the steel is. For the type of delayed failure problem reported here, the steel has to be harder than about 35 HRC (which corresponds with a tensile strength of approximately 150,000 psi); at this hardness, embrittlement can result when the hydrogen content is less than ten parts per million by weight; at higher hardness, even less hydrogen is needed.

Typically, high strength components such as these will be subjected to a bake-out treatment, which is intended to remove the hydrogen that is inevitably picked up in the electroplating process. In this case, the bake-out treatment had been either too short or at too low a temperature to effectively remove the hydrogen.

The aerospace industry uses a lot of high strength steel fasteners and thus has a strong interest in being able to test for resistance to hydrogen embrittlement. The standard test method is ASTM F 519, which is used to evaluate not only the plating processes but also the various service environments, which can include cleaning treatments and maintenance chemicals.

The test method uses notched specimens of AISI 4340 alloy steel, heat treated to a hardness of 51 to 53 HRC. A plating process is considered acceptable if notched specimens do not fail after being stressed to 75% of their tensile strength for 200 hours. The socket set screws we tested had the same hardness as those used in the ASTM F 519 test; the laboratory failures at a mere 10 percent of the tensile strength clearly showed that the plating process was not acceptable.

A number of grades of high strength steel fasteners are susceptible to delayed failure due to hydrogen embrittlement, with the susceptibility increasing in proportion to the hardness. Thus, if you don't really need a high strength fastener for a particular application, don't use one under the assumption that stronger is better. And when you really do need a high strength fastener make sure the supplier understands the importance of getting all the processing details right to avoid having delayed failure problems.

Has anything good come from hydrogen embrittlement? About sixty years ago a young metallurgist named Carl Zapffe was studying hydrogen embrittlement of steel. He began examining fracture surfaces at high magnification under the optical microscope, and called this process fractography. Then he enthusiastically began applying this technique to fractures in many other materials. Zapffe's pioneering work sparked a widespread interest in studying fractures at a microscopic level; the use of fractography in analyzing fracture surfaces remains a cornerstone of failure analysis to this day.

And the clean room socket set screw failures? With the problem traced to the screw manufacturer rather than the client's assembly procedure or system design, our client was able to go back to the screw supplier for remedial action.

MEI-C People

Wedding Season

_It's been a season of weddings for us, with both President D. G. Chakrapani's eldest son and Senior Vice President Ralph Hudson's daughter getting married.

Srijayanth Chakrapani and Saileshi Patel were married on December 29 in Vadodra, India. Srijayanth is a vice president with Ovid Corporation in New York. Saileshi is a recent MBA graduate from Duke University.

Serin Hudson and Dave Van Dyke were married on December 15 in Portland. Serin is a mechanical engineer with Epson's ink cartridge manufacturing division in Portland. Dave is a graduate student in mechanical engineering at Portland State University.