§1. King Street Bridge and the George Washington Bridge fatigue cracks [00:02]
This is the King Street Bridge I mentioned yesterday. Here is the weld detail on one of the girders that was holding up the bridge. They developed a fatigue crack at the toe of this weld — lots of restraint as you go around — developed a fatigue crack, and then you had a brittle fracture go up through there. Like I said, no one got hurt or killed. In fact, as you look at the picture, the thing didn't even break all the way through; the top flange didn't break through.
I had a similar failure for the George Washington Bridge in New York. It was a nine-foot crack in a nine-foot web, but it didn't fracture the top flange of the I-beam. It started at the weld. Where did it start? It started at the bottom flange. The brittle fracture ran all the way up through the web, and because it was deflection limited it didn't break all the way through. But they closed off one or two of the lanes from the George Washington Bridge coming into New York City for six months and created a lot of havoc.
§2. Underwater welding: arc instability and depth limits [01:14]
I want to talk about underwater welding. We have a few divers here. I'm not going to spend a lot of time on it because there's not that much that I know about it. One of the problems is, you can strike arcs underwater, or you can strike arcs at high pressure. We actually call welding arcs a high-pressure arc — anything above half an atmosphere in the physics literature is a high-pressure arc, and it basically means the electrons and the ions have the same relative temperature, as opposed to a fluorescent light that doesn't. It turns out the welding arc — shielded metal arc, gas tungsten arc — becomes unstable at 450 psi or greater. That's because as you increase the pressure above one atmosphere, you actually squeeze the arc into a smaller volume; you get it down so it becomes narrow enough that the gas flow in the arc becomes unstable and it'll get blown off to the side from the convection currents that are generated.
That means whether you're wet welding or dry welding, you're limited. If you can only go to 450 psi, and you really maybe can only go to 300 psi — you can strike an arc at 450 but it's not going to be a very stable arc. At 4.44 pounds per foot is the pressure increase of sea water. That means your maximum depth that you could do underwater welding at, whether you're in a dry habitat or a wet habitat, is a thousand feet. Some of the dry habitats actually have gone down deeper, because they change the pressure inside the habitat — you're welding inside a submarine, if you will. It might not be a submarine that can move on its own power, but you're welding inside a dry habitat that can change the pressure so the arc is constricted.
That means the heat intensity is different. If you were to make a bead-on-plate weld at one atmosphere it looks like that; an underwater weld at greater depths might look like that, because the arc up here is much tighter. At one atmosphere it's broad; at 20 or 30 atmospheres it's narrower, and you get a different shape weld pool. If you're only welding at 100 feet deep, you have a hard time telling the difference, but if you get down 500 feet or so you actually start seeing differences. You ought to be able to see it at 100 feet, because that's six atmospheres.
The weld metal chemistry changes — there are a lot of chemical reactions going on in the weld pool. One of them is carbon dissolved in the steel gives you CO, and there's an equilibrium between the CO pressure and the CO2 pressure in the welding. With higher pressure you'll drive that reaction back the other way; you'll retain more carbon. You also retain more manganese and silicon — you'll get a richer weld metal chemistry at higher pressures. And it's noticeable. Your carbon content, instead of 0.1, could be 0.14 — a 40% increase. Your manganese and silicon might go up by 20 or 30 percent. Your weld chemistry changes, and it changes to a more hardenable chemistry. So those people — Barco, or whoever you said it was — have adjusted their weld compositions. They may even sell different electrodes for different depths; I don't know. There's been some work — not a lot — on underwater welding metallurgy, but it does change the chemistry.
§3. Wet welding, the USS Cole, and the cost of dry habitats [05:52]
In addition, if you're wet welding, you're welding and quenching the whole thing in water immediately. The arc itself will generate a pressure — up to 30 atmospheres — to push back the water. But you're welding in a humid atmosphere full of hydrogen. If you look at this old Venn-diagram-style plot: you have more hardenability because the chemistry changes, you have a very high hydrogen environment, and depending on what you're welding on, the amount of restraint, you've got a fair amount of stress if you're doing this as an emergency repair weld.
For example, was it the Cole that had the big hole in it? We didn't want the terrorists out there in the Gulf to think they had destroyed a U.S. capital ship, so they actually went in and they had to do underwater welding. It wasn't very deep — just down to the keel beam. And they towed it back and repaired it. They could have bought a new ship for less. They wanted to prove that some terrorists with a motor launch didn't destroy a U.S. capital ship. So they repaired it, but it was an expensive repair. They couldn't have even gotten it back — just the seas would have torn it apart; the damage was severe enough they had to have some weld repairs before they could even tow it back to the U.S.
With dry habitats, you can get around the quenching and you can get around the hydrogen problem. Most of the time the Navy goes to a dry habitat. A dry habitat is a lot more expensive — the barge costs for doing underwater welding are on the order of $100,000 to $1,000,000 a day, and dry habitat is closer to the million than the hundred thousand. Just to do a wet weld a couple hundred feet deep, you're going to be paying $100,000 a day for the barge. Of course at 200 feet deep a diver can go down and up in a day. At 800 feet deep, you can't just go up and down in a day. You actually have to be pressurized in a chamber before you go down, and when you come back up you have to be depressurized in that chamber. Sometimes those chambers can take up to a week.
Student: Saturation diving.
The Navy really doesn't do it — they essentially did it as a proof of concept. But commercial divers do it every day.
Student: I think the deepest saturation dives [inaudible].
Typically 800, a thousand is probably about the deepest most people go. The chamber is still at 30 atmospheres and they have to change the atmosphere. The Navy did all the basic research on the metabolism — they have to have mixtures of helium and oxygen, because they have to change the gas chemistry. Your blood chemistry is working for some equilibrium at approximately one atmosphere. All the diving is based on research.
Student: [inaudible]
You don't want to put people in harm's way — contract out some sucker, whatever you want to call the person. It's very hazardous work, but there are some people who just love it. Somebody yesterday told me — I was mentioning I was looking for my underwater welding notes — and a person said, oh, his cousin wanted to be an underwater welder. It is sort of neat, but it's a specialized technology. The oil companies with the offshore oil rigs do it all the time, and there are big companies down in Louisiana and Texas that have these barges.
§4. The bends and the Brooklyn Bridge caissons [10:41]
Saturation diving means your blood is saturated with oxygen and other gases. Most of you have seen from the movies about the bends, where you can get too much nitrogen in your blood. As you come up, it's just like the CO2 coming out of the Sprite — the nitrogen comes out, and all of a sudden you have a two-phase blood, nitrogen and liquid. It doesn't flow the same and you get the bends and you die. Does anybody know where we first discovered the bends?
Student: Brooklyn Bridge caissons.
Brooklyn — good. How'd you know that?
To build the Brooklyn Bridge, which was a major engineering feat at the time, they basically floated a structure in the river — is the Hudson on that side, or is that the East River? Whichever river. It was a diving-bell type of structure that they put down on the river floor. They had some hatches, people would go down, they'd pump the water out, they'd have compressed air coming in, and people would dig down. This very heavy caisson would slowly sink into the mud. These things went down five, six hundred feet or so — I don't remember exactly. Just a bunch of men in there with shovels, with little miners' helmets and air blowing in, a pretty miserable environment. But they noticed, as they were getting deeper, that some of them were starting to get cramps as they came up. That's when people learned about the bends. They were breathing regular old air, and they were getting deep enough to get the bends. It was back in the 1880s.
§5. Altitude sickness and PIGMA welding at Rocky Flats [12:53]
Has anyone had altitude sickness? You've had altitude sickness — what's it like? Where were you?
Student: [inaudible] just like it feels like you can't breathe completely.
You get worn out if you try to exert yourself. I have some children who went to school in Utah and I go out there regularly. At seven thousand feet it doesn't usually bother me. We went up to Sundance once to watch a play one summer. Your blood chemistry changes in about three days, so usually your body will adjust to the lower atmospheric pressure. But at 8,000 feet it's about 0.75 atmospheres — you're not getting as much oxygen, you feel tired. If you try to exert yourself you're not getting the oxygen your body was used to. I don't know if it's genetic, but certain people get it more than others. I don't notice it at all at six thousand feet. At seven thousand feet, when I'm out of shape — which I am all the time now — I can notice it. But when I got up to eight or nine thousand, I just laid down; I didn't watch that outdoor play. My children and my wife were watching, and I just laid out on the grass. We almost had to get a cart to take me back down. We parked at about 7,000 feet and we walked up to 8,500.
By the same token, I've been over passes in the Rockies at 10,000 feet and I didn't notice anything, although if I'd stayed there for a little while — because it doesn't hit you right away, it sort of grows on you the second day. By the fourth or fifth day you get used to it in general. So that's altitude sickness.
Well, welding has altitude sickness too. Denver is the Mile High City, and they developed a process to make nuclear warheads at Oak Ridge in Tennessee, which is less than a thousand feet — probably five or six hundred feet elevation. After the development process, they took it to Rocky Flats in Denver, which is now the Denver airport. They wanted to put it into production. It turns out the atmospheric pressure there is about 0.82 atmospheres. So it didn't work the same — the shape of the weld pools wasn't the same when you lost 20 percent of your pressure. They actually had to put the whole thing in a pressure vessel and pressurize it. They called it PIGMA welding. If you look in the literature back in the early '80s you'll see papers in the Welding Journal: pressurized inert gas metal arc welding, or something like that. They didn't exactly explain why they were doing it. But the scuttlebutt is that the Department of Energy developed it at Oak Ridge and moved it up to 5,000 feet, and the arc characteristics changed enough — pressure does affect arcs. But you have much higher pressures in underwater welding.
§6. Why wet welds crack: hydrogen, not the cooling rate [16:22]
Student: Do you mind if we talk about the heat-affected zone a little bit, with the quick quenching? I read a lot about that because we were concerned with changes in water temperature and their effect on welds. It appeared, over the years, that if you pull a weld in really cold water, it tended to crack a lot more than one made in warmer. Looking at some of the research, they focus on the time between 800 and 500, right?
That's not the reason, but go ahead.
Student: They found that between 80 degrees, some of the warmer waters, and 30 degrees, some of the colder water, there wasn't a significant difference between those two times. So it really didn't matter — either way you're going to produce martensite, which is not what we want, because of the lattice structure.
Student: People had kind of hypothesized that it wasn't a delayed cracking issue [inaudible].
Well, I think what's happening is hydrogen is getting forced into the weld from the dissociation of water.
Student: My thought was that at lower water temperature hydrogen wasn't absorbing back into the water [inaudible] — it's getting into the weld pool at a very high saturation, but it's not escaping. That's the problem.
Let me expound. If I look at a welding heat cycle and I go up to 1600 degrees centigrade — actually T 800-500 is in Fahrenheit. No, it's centigrade. So this is my CCT curve in general. What temperature should we be at — 800 to 500? In centigrade, this is where I'd be transforming, 800 to 500. You're right: welders like to talk about the time to cooling, ΔT = T 800-500 in the welding literature. That's because that's how long it takes for you to get through this transformation region. If it's a simple carbon steel you've got to get through there in less than a second.
Student: Do they put time scales on here?
This one is actually slightly alloyed — it's 8630 steel. If this was a carbon steel you'd have to get through there in half a second to not get martensite, or a second to get any depth of hardenability. But this one's got some depth of hardenability, so you're going to get a higher-hardness heat-affected zone. So 800 to 500 is what welding metallurgists like to talk about, and that gives you some idea of how long your welding heat is applied. With the whole thing quenched you get through there a lot quicker, because you basically go up about the same in a wet weld, but you come down a lot faster. You have a shorter T8-5. Metallurgists are trained to worry about that T8-5 cooling time.
Well, you're breaking down water up here, so you're in a 60% hydrogen atmosphere. If you're at 800 feet you're in 60% hydrogen times 25 atmospheres, which is a very high hydrogen potential. You can pump more hydrogen in at higher pressure, so it depends on what depth you're at. But you're going to cool the same — the steel, once you start throwing water on hot steel, this cooling curve doesn't change because of depth. You're trapping in lots of hydrogen, more than you would at one atmosphere because of the higher pressure. However, the problem really is at lower pressures — what happened to my Sprite when it was pressurized before I popped the cap? There were no bubbles, no CO2 coming out. At higher pressures you're going to keep the hydrogen in. So it was hydrogen cracking, first of all. Was it delayed? Probably. It may not have been delayed as long — it might have been tens of minutes rather than a couple of hours, because you had a higher hydrogen content.
Student: Right, so it's delayed.
First of all, it was delayed cracking — that's hydrogen, I don't care what anyone else says. If it's delayed cracking, it's hydrogen induced. Why is it hydrogen induced? Has nothing to do with the high-temperature stuff, other than in wet welding I just have lots of hydrogen present — more than enough for virtually any steel I quench, and I'm going to get martensite. So if I look back at my Venn diagram, which I don't have here anymore, it's a problem.
Now why cold water? Well, I actually told you this. This is for stress corrosion cracking, but I showed you a similar schematic. Cold water — I hold the hydrogen in, the time to failure goes up by a factor of two and a half from warm water. There you go. It might take less time because you have more hydrogen, but it takes longer because of cold water — so maybe it still takes an hour or two before it cracks. But it's delayed cracking. You were right, it's hydrogen. Go back and rub their noses in it. There is a difference between cold water and hot water, but that's because of the hydrogen diffusing out — not something going on at high temperature, but something happening out at much longer time. The high-temperature stuff is like 10 seconds; hydrogen cracking is out at tens of minutes or hours. Does that make sense? You learn the principles and you can figure it out for yourself.
§7. Teaching by first principles, and a sermon about feeding the cow [24:16]
I haven't given you that little lecture of mine. I don't really care whether I'm talking about welding or brazing or Doc Edgerton or anything else. I look at teaching as trying to help you learn to think through an answer, with the idea that most of you actually have all the information you already need. In welding, I might be giving you new information, but hopefully you'll be able to leave here actually having enough information and figuring out how to put it together. The world is not that complex if you can ask the right question. Then you often can figure out things yourself from first principles. So the first principles here are: it's delayed, it's hydrogen. Does temperature have an effect? Yes — I showed you the V-shape, hydrogen diffusion out, so that makes sense. If I went through the heat flow stuff, you can figure out that changes up at the high-temperature end are really not that significant, other than depending on the hardenability of the steel. But in general this thing is quenching so fast in water that no matter what type of steel you have, you're going to get martensite in wet welding. There are lots of problems with wet welding, but sometimes you have to do it; you don't have a choice.
There's a story — it's actually a joke, but it does have a moral. There was a new preacher in town, and he went to church that first day. He'd worked very hard on his sermon, and he's waiting for the parishioners to come in. The only person who shows up that Sunday, because they hadn't had a preacher for a while, is one old farmer who comes in and sits in the back. The preacher waits, no one else shows up, and it's time to start. He walks to the back of the chapel and says, well, what do you think I should do — you're the only one here? And the farmer says, well, if I had a cow and she was hungry I'd feed her. The preacher thinks, well, that's a good philosophy. So he goes up and he gives an hour-and-a-half sermon that he had prepared. After, he comes back down and says to the old farmer, what'd you think? The farmer says, I said I'd feed her — I didn't say I'd give her the whole load. So you don't have to give them everything all the time.
I told you I had a lousy attitude about the way we teach. Most people try to make you think that science, or whatever they're teaching, is complex. I think it's better to teach you that it's actually simple. Maybe it's just because I don't understand it myself.
§8. Critical pitting temperature and the Nautilus air conditioning story [27:17]
[Tom retrieves his pointer.] I did leave something off before I went to nickel-based alloys — this is stainless steels, called critical pitting temperature. 316 stainless steel has a critical pitting temperature around 10 degrees centigrade. This is a function of molybdenum content. 20Cb3 is a steel made by Carpenter Technologies; this is welded and unwelded. 20Cb3 is 30 degrees unwelded. When you weld it, because of segregation in the welding zone and changes in the microstructure, you'll get pitting at a lower temperature. If you add more molybdenum you get 316, 317 — should have been a bigger change there. The best of these — and how hot does it get in the tropics in the sea water? About 37 degrees centigrade, something like that.
I remember a story that when they were designing the Nautilus, their heat-flow calculations of how much heat you could dump in the tropics for the air conditioning system varied by a factor of five. They didn't have fancy computers, and they didn't have all the boundary-layer theory that was going to tell them how well the sea water could cool things on the sub. So they went to Rickover and said, well, what do we do — we've got a variation of five, and the heat exchangers are going to be huge at five, and it's going to change the ship design. Being conservative, he says: allow for five. And it was just barely enough air conditioning. So the Nautilus could operate in the tropics. If they hadn't allowed for five, they would have been limited to certain latitudes, northern or very southern.
I told you the Navy was interested in making an all-stainless-steel submarine, and the alloy they were looking at was AL-6X, which is six percent molybdenum. Molybdenum is a pretty pricey alloy, so this was going to be a very pricey stainless steel. The reason: unless you stay in the tropics for a long time, you wouldn't have pitting problems with a very high molybdenum content, because you're going into a salt solution. So there is some corrosion on stainless steels that relates to the Navy.
§9. Nickel-based alloys: Monel, Inconel, and the Sudbury ore [30:10]
Now let's talk about nickel-based alloys. [Tom retrieves notes.] I showed you Cedric's little genealogy of iron-chrome alloys. If you go to higher chrome, you eventually go to what we call the Inconels, which are nickel-chrome-iron rather than iron-chrome-nickel. Nickel becomes like 60% of the alloy. You can think of that as an extension of stainless steels. Some people do call the nickel superalloys a class of stainless steel; sometimes they're called high alloy, or high-alloy steels. Some people say if it doesn't have 50% iron it's not a steel, because steel is an iron-based alloy. But some people throw the Inconels — with 60% nickel and 20% iron — in there as a steel. Most people just call them high alloys, whether they're steel-based or nickel-based.
There's also cobalt base, but most of the world's cobalt comes from Zaire — used to be the Belgian Congo — because they have very good cobalt deposits. Cobalt is expensive anyway, but the supply out of Zaire has been very erratic for the last 80 years because of the fighting going on. People have been trying to get cobalt out of alloys for a long time. HY-180 is a 10% cobalt alloy. Landing-gear steels that the Air Force developed — AF-1410 — is 10% cobalt, similar to HY-180; you can get very high strength with interesting hardening. Cobalt is the only element that can change the martensite-start temperature in the opposite direction. Everything else, as you alloy, lowers the temperature for forming martensite; cobalt actually increases the tendency to form martensite. So it's unique — but it's also expensive and very unreliable in supply.
If I look at nickel and nickel-based alloys: there's commercially pure nickel, which we use for some corrosion resistance when we have hydrochloric acid. It's not a long-term solution, but if you have hydrogen chloride gas or not-too-concentrated hydrochloric acid you can use pure nickel. It's easy to weld — almost like welding copper, easier than copper because it doesn't have copper's high thermal conductivity. Then we have the solid-solution-structure alloys. The first one, nickel-copper, is actually called Monel. Most of you have probably heard of Monel because the Navy still uses a lot of it. It's a 70% nickel, 30% copper alloy.
[Tom shows image.] Here is Ambrose Monell. He was basically the person who founded INCO. His name is spelled M-O-N-E-L-L; the alloy is M-O-N-E-L. That's because the trademark office said you couldn't trademark the name of a living person. He developed the 70-30 copper-nickel alloy — Monel, named after him. The other guy who's famous in the nickel business is Elwood Haynes, who founded Haynes Stellite. Monell was a British subject in Canada, and there was a particular mine in Sudbury, Canada that had seventy percent nickel and thirty percent copper in the ore. It is no coincidence that Monell discovered an alloy that came directly from the concentrations they had in the Sudbury mine. It had excellent corrosion resistance.
Later, people developed nickel-aluminum alloys — these are the Hastelloy Bs — nickel-iron, nickel-chrome-iron. These are sometimes used for toaster oven heating filaments. Very good high-temperature strength, corrosion resistance, oxidation resistance. Nickel-chrome-moly-tungsten, nickel-chrome-iron-moly, nickel-chrome-cobalt-moly — these are all solid-solution-strengthened, and we still use them particularly in high-temperature or corrosion-resistant situations. Precipitation-strengthened: we start adding aluminum and titanium, or sometimes niobium. These have very high-temperature strength. Your turbine blade alloys are typically made out of these. Many of the Inconels are precipitation-hardening Inconels, where you can heat them up to high temperatures and then do other things.
It turns out nickel and aluminum form a Ni₃Al that melts at 2,000 degrees centigrade. Nickel melts at like 1650 centigrade — higher than steel, but not much higher. The Ni₃Al melts at 2,000 centigrade. Same thing with Ni₃Ti. Nickel and titanium love each other, nickel and aluminum love each other, and they form this ordered intermetallic. I was once asked, when I consulted for a gold company down there back in the early '80s — they said, well, can you develop a harder white gold? People were choosing white gold for their class rings, and they wanted a hard white gold that wouldn't wear out. I looked — it turns out white gold is just gold-copper-silver with some nickel in it, maybe 10% nickel. I was teaching thermo at the time, so I took my phase diagram books home and looked. I said: add two-tenths of a percent nickel to your white gold, heat it up to this temperature to solutionize it, quench it, and then heat it up to an intermediate temperature, and see if you can get precipitation hardening. Precipitation hardening: 450 Vickers.