WM_Su2014_26

Welding Quality Summer 2014 Session · 6 sections 10 cases · Watch on YouTube ↗ all files
Layer 3 — readable edition

§1. Precipitation hardening in precious-metal alloys [00:03]

§1.p1

The old stuff was about 300 hardness. This was 50 harder on the first try. It worked because I knew nickel aluminum and nickel titanium — in this case nickel aluminum — had a very strong affinity for each other. Both of them dissolve in gold-copper alloys at limited concentration, so if you heat them up at high temperatures they should go into solution, and at lower temperatures they should join up and bond to each other as nickel-3-aluminum. They did, and we got high hardness and precipitation hardening.

§1.p2

I went off in 1984 to Tokyo for the Navy. When I came back I asked, well, have you patented it? They said, we decided there wasn't a real application. Then they asked, can you make us a partner system with sterling silver? So I tried the same thing with sterling silver. The first attempt was a failure. Sterling silver is a silver-copper alloy. On the second attempt I added lithium and tin, and it worked. Lithium and tin work on the same principle — they form a high-temperature intermetallic, they both dissolve in silver.

§1.p3

I asked, what's the application? They said lightweight sterling silver dinnerware. So we went over to Reed and Barton — this is around 1985 — and we had our hard sterling silver. Reed and Barton said, no one wants lightweight sterling silver dinnerware. People want everyone to know they're rich and can afford a lot of silver. So lightweight dinnerware was not the application. Obviously this gold company — they're out of business now — didn't do a lot of market research. But they did have an application: baby cups. Sterling silver baby cups. So if you go buy a baby cup today it's twenty percent lighter because of my alloy. How about that. You can still buy it for fifty bucks, but they make an extra ten bucks profit, okay, because they got the strength up.

§2. Nickel superalloy metallurgy [01:59]

§2.p1

Anyway, the alloys that go into these things — the solid solution strengtheners are cobalt, chrome, iron, aluminum, tungsten, tantalum. The precipitation hardening, which they call gamma prime, is the phase based on aluminum and titanium. Niobium also does that, but it's a gamma double prime, slightly different crystal structure. Then they have all these carbides — chrome will form a carbide, tungsten will form a carbide, molybdenum forms carbides, niobium forms carbides — so they throw in different carbide formers to harden things.

§2.p2

Then they have what they call topologically close-packed phases. They tend to be hexagonal close-packed or more complex phases — sigma, P phase, mu phase, and Laves phase. Very brittle. If you hold some of these alloys at temperature for a long period of time, you'll end up with that structure. This also happens in some of the stainless steels. The times we're talking about are tens of thousands of hours. Okay.

§2.p3

That may have been one of the problems we had with my chrome-moly tube over at the Boston Navy Yard. They wanted me to say, yeah, you can use stainless steel to weld it, but the Navy said you had to use Inconel. It turns out the stainless steel probably, after ten thousand hours, formed this brittle phase, and the things would crack. So it's a general problem.

§2.p4

As far as surface oxidation: when these things are used at high temperature, chromium and aluminum are the primary elements that form chromium oxide and aluminum oxide scales. If you analyze the surface oxide on these high-temperature nickel-based superalloys in jet engines, they will be chromium oxide or aluminum oxide depending on the ratio of the two, nearly pure. Just like stainless steel, it's nearly pure chrome oxide. On these nickel-based alloys, the scale will be chrome oxide or aluminum oxide depending on the composition.

§3. The economics of submarine hull materials [04:12]

§3.p1

Okay, that's an overview of the metallurgy. We're talking about alloys that are going to cost thirty to a hundred dollars a pound, as opposed to steel at a dollar a pound for an alloy steel like an HSLA steel. Considerably more expensive. Why doesn't the Navy go make an Inconel submarine? Well, you can make one out of AL-6XN and probably pay only ten bucks a pound for that steel. But why are you paying ten bucks a pound rather than a dollar a pound? If you're going to build a two-billion-dollar submarine and the hull cost is, let's say, five hundred million of that two billion —

§3.p2

If you have to make it out of all stainless steel, now you're talking about a five-billion-dollar hull. Ten times the price. You just ordered nine submarines for eighteen billion. Would you rather purchase three submarines that were non-magnetic? And why don't we build titanium submarines? Because the price of titanium fabricated is probably a hundred dollars a pound. So with eighteen billion dollars you could have purchased one submarine. Now you have to ask, would you rather have a nuclear carrier or a nuclear submarine?

§4. The Soviet Alpha-class titanium submarine [06:02]

§4.p1

In 1980 I'm coming back from my first trip to Europe, I'm on the plane, and they hand out a copy of the International Herald Tribune — or maybe it was the international Wall Street Journal, whatever it was — and on the front page it announces that the Soviets now have a titanium submarine. This was the Alpha sub. The U.S. Navy may have known a little bit earlier, but it was now on the front page of an international newspaper.

§4.p2

I had been working on welding heavy-section titanium since 1977, three years earlier. So all of a sudden I come back and a number of people want to talk to me, because ONR had been funding me to weld heavy-section titanium. They had some conferences down at David Taylor, in Carderock — I'll tell you more of those stories when we get to titanium. But I remember the guys from the Naval Research Lab said, Tom, how do they solve the creep-fatigue interaction in titanium? I kind of shrugged my shoulders, because I hardly even knew what the creep-fatigue interaction was. But the Office of the Naval Research Laboratory knew.

§4.p3

It turns out that if you compress titanium and hold it under compression, fatigue cracks grow really fast. Well, guess what a submarine hull is — it's held under compression when you're down deep. So there were several problems with the titanium subs of the Soviets. Certain things weren't a problem: they could go faster underwater than our destroyers could go on top of the water. They could dive deeper than the collapse depth of the depth charges. Some of these things concerned people, although one person who did a lot of work for the Rocky-class said, well, if you use the right type of depth charge you only have to get within a few miles. I thought, yeah, and as soon as you start doing that we have bigger problems. But in any case, we could have gotten rid of them.

§4.p4

What happened is they developed these cracks, and they were noisy as could be. The real answer that I didn't know — that the people from NRL were asking me about, how did the Soviets solve the creep-fatigue interaction — is that they didn't. They just built the ships and then they found out about the creep-fatigue interaction, which NRL had studied and knew about. That was one of the barriers to our building a titanium submarine. Although the big barrier was cost. It really is a question of, do you want an eighteen-billion-dollar submarine, or do you want nine steel ones, or three stainless steel ones? Unless you want small ones — you can buy small ones cheaper. But that's the real problem with these things.

§5. Liquation cracking in nickel-based alloys [09:13]

§5.p1

Okay, one of the things you need to know about nickel-based alloys — I told you a little bit about this the other day, when I mentioned the one welding code, out of Westinghouse Bettis, where they make Navy nuclear parts. They were welding some Inconels with electron beam welds, and that's the only code I ever knew of that would allow you to have cracks in your welds and pass the spec. There was a minimum size and a maximum size — I think it couldn't be more than about a millimeter in length. When you heat up these nickel-base alloys you get something called liquation cracking, which means it forms a liquid. I've talked about a Slurpee — you have solid crystals surrounded by liquid, and this stuff flows like a snow cone or Slurpee at high temperatures.

§5.p2

So here's your welding temperature-time curve. And on C you have the liquation. When you get up into the heat-affected zone in particular, it's not a uniform structure — you have segregation in the metal when you melt it and reform it. This occurs in the Monels as well; in fact it might have been a Monel rather than an Inconel that I was looking at the welds for, for Bettis. You have your solidification grain boundary, and you have your fusion boundary here, so you're in your heat-affected zone. You have particles that will dissolve, liquate, turn to liquid during the heat of welding, and then some of those will segregate to the grain boundary, particularly in multi-pass welds. They had to make an electron beam weld with a thirty percent overlap in this particular case, and you get a grain boundary film that melts.

§5.p3

If you want to see what it looks like — here's a picture of the microstructure of a nickel-based alloy. They call it a solidification crack, but this is the base metal structure, and it partially melted along some grain boundaries, and some of the liquid in between these grain boundaries migrated. They call this a self-healing crack. Well, this one might be self-healing here, but these two weren't self-healing, they're still voids. So it's not a great situation.

§5.p4

Here's another picture, another alloy, and you've got the liquid coming in and forming this. You have other cracks — here's one — you can actually see, this is the dendritic structure, this is probably a Monel, niobium-enriched. Let's say it's alloy B. What's the B? That's from John Lippold. John, aren't you going to tell me what you got? Not going to tell me what the alloy is, okay. But you can see these little nodules — those are called dendrites in the steel structure. This is probably a cast nickel-based alloy. The region in between is high in alloy content, it got rejected. These are high in nickel in the center of the dendrites, and out at the edge you're rejecting alloy content, so you have a more highly alloyed structure. You heat it up to weld it, and the stresses from cooling of the weld crack things.

§6. Weld geometry and solidification cracking [12:41]

§6.p1

Nickel alloys, whether Monels or Inconels, are very susceptible to liquation cracking — high-temperature cracking due to the stresses of shrinkage. If I make a weld like this and I get all my grains going this way, it pushes all my alloy content there. If I make an electron beam weld, then I can end up solidifying grains from the edges in, and have all my alloy content in the center, and have a crack right down the center of the weld. So the shape of the weld makes a difference.

§6.p2

Here's a fillet weld. This is good because you have the columnar solidification grains growing from the edge out and pushing the liquid out here. If you have something like this, you actually collect it in a seam right here. So the weld profile makes a difference — not just for fatigue. This would be a good profile for fatigue; this is a worse profile for fatigue, with a higher stress concentration at that corner than this one. But this is lousy for solidification cracking. You can get solidification cracking in steels, but you've really got to try. Throw a lot of phosphorus in there and you can crack a steel. If you go look at the literature, they'll talk about solidification cracking of steels. I don't know if I should say I've never seen a solidification crack in steel, but I haven't seen very many, okay.

§6.p3

I talked about this the other day when we were talking about stainless steels — the Rensselaer Polytechnic Varestraint test, where you weld on a stainless steel or a nickel-based alloy, some high alloy typically. You make the weld, and as you're welding you bend the plate along the weld. It's called a transverse-strain Varestraint test. You can see, when you apply the pressure, you just open up the cracks, because the cracks were partially liquid at the time. You open it up and make more space.

§6.p4

Any questions on that? Let's take our break.

Cases referenced

Layer 2 — cleanup edit
p1 00:03

I mean the old stuff was like 300. So it's 50 harder. First try. Well it was only because I knew that nickel aluminum and nickel titanium, in this case nickel aluminum, they had a very strong affinity for each other. Both of them dissolved in gold copper alloys at a limited concentration, so if you heat them up at high temperatures they should go into solution, at lower temperatures they should join up and bond to each other in this nickel-3 aluminum, and they did, and we got high hardness and we got precipitation hardening.

p2 00:30

I went off in 1984 to Tokyo for the Navy, came back afterwards and said, well what happened, have we, have you patented it? And they said, well, uh, we decided there wasn't a real application for it. I said oh. So they said, well, but can you make us a partner with sterling silver? So I went and tried the same thing with sterling silver. First attempt was failure. The second attempt — sterling silver is silver copper alloys — second attempt I added lithium and tin, and it worked. And the lithium and aluminum work same type of principle. Lithium and tin form a high temperature intermetallic, they both dissolve in silver, and that...

p3 01:10

I said well what's the application for this? They said lightweight dinnerware, you know, sterling silver dinnerware. And so we went over to Reed and Barton, and uh it's like 1985, and we had our hike, we had our hard sterling silver, and Reed [and] Barton said, well no one wants lightweight sterling silver dinnerware. They want people to know that they are rich and they can afford a lot of silver. And so lightweight sterling silver is not dinnerware, is not. So obviously this gold company, they're out of business now, they didn't do a lot of market research. But they did have an application. Okay. Baby cups, you know, sterling silver baby cups. So if you go buy a baby cup today it's twenty percent lighter because of my alloy. How about that. You can still buy it for fifty bucks but they make an extra ten bucks profit, okay, because they got the strength up.

p4 01:59

Anyway, so these different alloys that go into these things, the solid solution strengtheners are cobalt, chrome, iron, aluminum, tungsten, panel [tantalum?]. The precipitation hardening, which they call gamma prime is the phase, aluminum and titanium — turns out niobium also does that, but it's a gamma double prime, slightly different crystal structure. Solid solution strengthener is these alloys. And then they have all these carbides — some things, chrome will form a carbide, tungsten will form a carbide, molybdenum form carbides, niobium — so they throw in different carbide formers to harden things.

p5 02:39

And then they have what they call techn— uh, topologically close-packed phases. It turns out they tend to be hexagonal close-packed or more complex phases. Sigma, p phase, mu phase, and Laves phase. Very brittle. And if you hold these alloys, some of these alloys, at temperature for a long period of time, you'll end up with structure. This also happens in some of the stainless steels, and the times we're talking about are tens of thousands of hours. Okay.

p6 03:16

And that may have been one of the problems we had — I told you about my chrome-moly tube over here at the Boston Navy Yard, and they wanted me to say, oh well yeah, you can use stainless steel to weld it. But the Navy said you had to use Inconel. Well it turns out the stainless steel probably after ten thousand hours formed this brittle phase and the things would crack. Okay. So it's a general problem.

p7 03:39

So far as surface oxidation, when these things are used at high temperature, it's chromium and aluminum are the primary things which form chromium oxide and aluminum oxide scales. If you analyze the surface oxide on these high temperature nickel-based super alloys in jet engines, they will be chromium oxide or aluminum oxide depending on the ratio of the two, nearly pure. Just like stainless steel, it's nearly pure chrome oxide. On these nickel-based alloys that see, the chrome oxide or aluminum oxide depending on the composition.

p8 04:12

Okay so that's sort of an overview of the, with the metallurgy of these things. Um, we're talking about alloys that are going to cost thirty to a hundred dollars a pound, as opposed to steel at a dollar a pound roughly for an alloy steel, like an HSLA steel. It's considerably more expensive. Why doesn't the Navy go to make an Inconel submarine? Well you can make one out of AL-6XN and probably only pay ten bucks a pound for this steel, okay. But why are you paying ten bucks a pound rather than a dollar a pound? And you really get down to, if you're gonna build a two-billion-dollar submarine, would you like to build out of steel, uh, and the whole cost is let's say five hundred million of that two billion, okay.

p9 05:10

So if you have to make it out of all stainless steel, now you're going to talk about a five-billion hull, right. Ten times the price. And so, you just ordered nine submarines for eighteen billion, right. Would you rather purchase three submarines that were non-magnetic, okay. And why don't we build titanium submarines? Because the price of titanium has fabricated is probably, or the price of the titanium is probably a hundred dollars a pound. So, you could have purchased with eighteen billion dollars one submarine, okay. Now you have to ask, would you rather have a nuclear carrier or a nuclear submarine?

p10 06:02

I'm not sure if it's, uh, radiation from the nuclear, it could be, I mean that's probably beyond my classification, okay. But what it was, in fact, and what we can talk about, this — well, I can't, I might as well tell you the story now. So in 1980 I'm coming back from my first trip to Europe, I'm on the plane, and I, they hand out a copy of the International Herald Tribune, and I'm reading it there on the front page, or maybe it's the international Wall Street Journal, whatever it was, on the front page it announced that the Soviets now have a titanium submarine. This was the Alpha sub, okay. Uh, maybe the Navy, U.S. Navy, knew it a little bit earlier, but it was now on the front page of an international journal or an international newspaper.

p11 06:45

I had been working on welding heavy section titanium since 1977, three years earlier. And so all of a sudden I come back and every, a number of people want to talk to me, okay, because ONR had been funding me to weld heavy section titanium. Um, well, they had some conferences down at David Taylor, uh, in Carderock, and I'll tell you some more of those stories when we get to the titanium. But I remember the guys from the Naval Research Lab said, Tom, how do they solve the creep-fatigue interaction in titanium? And I kind of shrugged my shoulders because I hardly even knew what the [creep-]fatigue interaction was. But the Office [of] the Naval Research Laboratory knew.

p12 07:27

It turns out that if you compress titanium and hold it under compression, fatigue cracks grow really fast. Well guess what a submarine hull is — it's held under compression when you're down deep. And so there were several problems with the titanium subs of the Soviets. First of all, there were certain things that weren't a problem. They could go faster underwater than our destroyers could go on top of the water. They could dive deeper than the collapse depth of the depth charges, okay. Some of these things concern people, although one person from Rocky [class submarines?], did a lot of work for Rocky class, says, well if you use the right type of depth charge you only have to get within a few miles. I thought, yeah, and as soon as you start doing that we have bigger problems, okay. But in any case, so we could have gotten rid of them.

p13 08:17

But one is they developed these cracks and they were noisy as could be. But they did develop the cracks. And so the real answer that I didn't know, that the people from NRL was asking me, how did the Soviets solve the creep-fatigue interaction — and the answer is, they didn't. They just built the ships and then they found out about the creep-fatigue interaction, which NRL had studied and knew about. And that was one of the barriers to our building a, uh, a titanium submarine. Although the big barrier was cost, okay. And it really is a question of, do you want an eighteen-billion-dollar submarine, or do you want a, you know, a hundred-and-eighteen-billion-dollar submarine, or which like nine steel ones or three stainless steel ones, okay. Or three — no, three would be, yeah, three would be stainless steel. So, uh, unless you want small ones, you can buy small ones cheaper, okay, then all of that. But that's the real problem with these things.

p14 09:13

Okay, one of the things you need to know about nickel-based alloys, and I already told you a little bit about it the other day, when I told you that I only knew of one welding code, out of Westinghouse Bettis, where they make new Navy nuclear parts. They were welding some Inconels and they were doing electron beam welds, and that's the only code I ever knew that would allow you to have cracks in your welds and pass the spec. They had a minimum size, it had a maximum size of the crack — it, I think it couldn't be more than about a millimeter in length. But what happens when you go up and heat up these nickel-base alloys, you have something called liquation cracking, which means it forms a liquid. And I've talked about having a Slurpee — you have solid crystals surrounded by liquid, and this stuff flows like snow cone or Slurpee, okay, at high temperatures.

p15 10:06

So here's your welding, uh, temperature-time, uh, curve. And on C you have the equation [liquation]. When you get up in the heat-affected zone in particular, it's not a uniform structure, you have segregation in the metal when you melt it and reform it. But even in the nickel-base alloys — and this occurs in the Monels, in fact it might have been a Monel rather than Inconel that I was welding for, uh, I was looking at the welds for Bettis. And so you have your solidification grain boundary, and you have your fusion boundary here, so you're in your heat-affected zone. You have particles that will dissolve, the [liquate], turn to liquid during the heat of welding, and then some of those will segregate to the grain boundary, particularly in multi-pass welds. And they had to make an electron beam weld with a thirty percent overlap in this particular thing, and you get a grain boundary film that melts.

p16 11:06

If you want to see what it looks like, here's a picture of, microstructure of a nickel-based alloy. Solidification — they call it solidification crack — but this is the base metal structure, and it partially melted along some grain boundaries, and some of the liquid in between these grain boundaries migrated, and they call this a self-healing crack. Well, this is, might be self-healing here, but these two weren't self-healing, they're still voids, okay. So it's not a great situation.

p17 11:37

Here's another picture of another alloy, and you've got the liquid coming in and forming this. You have other cracks, here's one, you actually can see — this is the dendritic structure, this is probably a Monel, uh, niobium-enriched, let's say a b. What's the b? That's from John Lippold. Uh, John, aren't you going to tell me what you got? Not going to tell me what the alloy is, okay. But you can see these type of little nodules, those are called dendrites in the steel structure. Uh, this is probably a cast nickel-based alloy. And the region in between is high in alloy content, it got rejected. These are high in nickel in the center of the dendrites, and out at the edge you're rejecting alloy content, so you have a highly, more highly alloyed structure. And you can see, you heat it up to weld it, and the cooling, the stresses from cooling of the weld, crack things, okay.

p18 12:41

So nickel alloys, whether they be Monels, whether they be Inconels, are very susceptible to liquation cracking, high temperature cracking due to the stresses of shrinkage. So if I can make a weld like this and I get all my grains going this way, it pushes all my alloy content there. If I make an electron beam weld, then I can end up solidifying grains from the edges in, and have all my alloy content in the center, and just have a crack right down the center of the weld, okay. So the shape of the weld makes a difference.

p19 13:15

Here's a fillet weld. This is good because you have the columnar solidification grains growing from the edge out and pushing the liquid out here. If you have something like this, you actually will collect it in a seam right here. And so the weld profile makes a difference, okay. Not just for fatigue — this would be a good profile for fatigue, this is a worse profile for fatigue, a higher stress concentration at that corner than this corner. But this is lousy for solidification cracking. You can get solidification cracking in steels, but you've got to really try. Okay, just throw a lot of phosphorus in there and you can crack a steel. But, you know, they, if you go look at the literature, they'll talk about solidification cracking of steels. I haven't, I don't know if I should say I've never seen a solidification crack in steel, but I haven't seen very many, okay.

p20 14:08

And I talked about this the other day when we're talking about stainless steels. This is the Rensselaer Polytechnic uh Varestraint test, where you basically weld on a stainless steel or a nickel-based alloy, some high alloy typically. You make the weld and as you're welding you just squ— just bend the plate along the weld. It's called a transverse strength [Varestraint] test. And you basically, you can see, when you applied the pressure, you just open up the cracks, okay, because the cracks were partially liquid, okay, at the time. And you open it up, make more space.

p21 14:45

Any questions on that? Let me, let's take our break, okay. So why don't we, this is the three—