FW_Su2013_01

Okay, so I'd explained that there's about twelve or thirteen lectures, hours of lectures on solid state joining. We're going to start with fusion welding. But to start with fusion welding, I like to, in the MIT spirit of things, go back to the fundamentals. If I were to take a piece of metal, any piece of metal, and heat the surface of it, what I'd like to know is how many watts per square centimeter I put into the surface. There's lots of different heat sources, and in fact at the beginning of the course, that first module when I talked about the importance of welding, I pointed out that virtually any new heat source that someone discovers, one of the first things they do is they try to use it to make a weld.

When Sir Humphry Davy discovered the electric arc in 1805 or thereabouts, one of the first things he did was try to weld with it, okay. One of the first applications, the first commercial application that I've ever heard of with a laser, was repaired— General Electric repairing vacuum tubes in the mid-1950s. I mean the laser just sort of had been invented, but back then they didn't have high power transistors and stuff, they used vacuum tubes. Some of these things were very expensive, costing hundreds of dollars apiece back in 1955, '56 dollars, and sometimes they'd get a bad weld in those little components in the vacuum tube and they'd have to break the seal, fix it, and then put it back together. They found with the laser they could go in there and they could hit that little post of tungsten or molybdenum or whatever it was, and they could cause this joint that is separated, they could cause one of them to get soft, and if they did with gravity as their aid they could cause the two to touch one another. Then they go in with a higher power pulse and they could weld them together without breaking the vacuum seal in the glass, because the lasers would go right through the glass right if they have the right kind of laser.

I taught a class a few years ago in the early '90s, like '92 or '93. The Naval Surface Warfare Center had spent some time, about a quarter billion dollars worth of time, trying to develop a particle beam weapon. And this was going to be the new Phalanx system, that you shoot down the incoming Aegis missile or whatever, right, on the surface— for a surface ship protection. And the idea is you could shoot relativistic electron beams out there thirty miles, and they go at close to the speed of light, and the incoming missile you'd hit with this high power pulsed electron beam. And they built these units up in White Oak and stuff, and at some of the National Labs. And then peace broke out with the former Soviet Union, and all of a sudden they have this technology, what do we do with these pulsed power relativistic electron beams? And so they had a three day workshop, they invited me to come down and see, help them understand whether they could weld nuclear submarines. Okay, because they saw Navy work, right, let's apply it to welding submarines. So I got down there, I looked at what they had, megawatts of power, and I said well you might be able to melt a submarine but I'm not sure you could weld. You only need about ten to a hundred kilowatts to melt, to weld, and you've got so much more energy, so I'm not sure this is really going to be useful. Plus it gives off radiation, x-rays to beat the band, you have to have like two feet of insulation, okay, radiation protection, and you can't just clear out the whole shipyard just because you're going to start welding a joint, okay. So I had to explain some sort of practical things to the scientists.

But nonetheless, what I did is I actually used something that I learned in 1977 when I was a young faculty member, and that is you can put all— [Tom adjusts something] move, what's going on here, upper base off, that's okay. You can put all fusion welding processes on a one dimensional graph. So here's one dimension, I told you I keep things simple, one dimensional graph. You get to the management school you're going to learn about two-by-two matrices, this is complicated complex as they get. But anyway, this goes from 10 to the 2nd watts per square centimeter up to a million watts per square centimeter. And it turns out a regular old flame— [Tom lights a flame] oops, I'm going to turn on cool— that's about a hundred watts per square centimeter. And I really don't want to leave my finger in there too long at 100 watts per square centimeter, my skin will start to melt, okay.

Now it turns out at less than about 300 watts per square centimeter— does everyone know, anybody know on a log scale where 300 is between 10 to the 2 and 10 to the 3rd on the log scale? Is your remedial math. Halfway right? Because 3 squared is 9, this is an order of magnitude on a linear scale. On a logarithmic scale, 3 is halfway between one and 10, right? Actually 3.1 is halfway, but nonetheless, square root of 10 is 3.1 something. Anyway, at above 300 watts per square centimeter, if I took a flame and held it on a piece of steel and just held it in that spot stationary long enough, I should just start to melt the steel. Below that I'll be able to carry the heat away, and I'm going to demonstrate that for you in a little bit, okay, either today or tomorrow.

At a thousand watts per square centimeter you're getting into an oxy-acetylene flame, and I didn't bring the oxy-acetylene torch, but people use oxy-acetylene torches for welding all the time, and that's about a thousand watts per square centimeter. Signor note somewhere— so you guys really don't have to take notes. Get out of your old, stay in that mode for your other MIT courses, okay, taking notes, but just sit there and sit back and learn to absorb, okay. Don't be distracted by I'm going to be quizzed and so therefore I have to take notes. That's what that handout is for. In fact you have in that handout, this about halfway through, which is a little packet that has this exact same graph on it, okay.

So in any case, a thousand watts per square centimeter, oxy-acetylene flames, it's not too hard to melt steel but it takes a few seconds. I get to 10 to the 4th and that's an electric arc, okay, and that's what we do most of our welding in the world. Fusion welding is done with electric arcs at 10 to the 4th watts per square centimeter. 10 to the 5th we actually have something called resistance welding, I'll talk about a little bit later. And 10 to the 6th and above we have electron beam and laser welding, okay. And that's what we're going to talk about, the things that are on this graph, for the next ten or twelve hours. And we're going to go through this, but from a fundamental heat flow.

And what do I mean by fundamental heat flow? Anybody know what Forte's [Fourier's] law of heat conduction is? You've never heard of it before? Yes? And you don't remember, that's okay. The heat flux is equal to minus K, which is thermal conductivity of the material, times the gradient of temperature with distance. This happens to be units of watts per square centimeter is equal to minus the thermal conductivity, delta T over delta X if you want to get out of calculus and go to algebra, that's watts per square centimeter. If I have less than 300 watts per square centimeter, I can't get a delta T hot enough to melt steel, okay. It doesn't really matter what X is. The steel has enough thermal conductivity to carry the heat away. And in fact if I tried to melt copper, which has got about four or five times the thermal conductivity of steel, I'd probably have, I'd have a hard time doing it with an oxy-acetylene flame. Copper, aluminum, be hard to do, I really need two or three thousand watts per square centimeter. But an electric arc I can melt them within a fraction of, within a second or even a fraction of a second depending on how many amps I have in my arc, okay. With a laser and electron beam at the other end of the spectrum.

So down here I'm not putting the heat in fast enough, the material can just carry the heat away before I'll reach the melting temperature, okay. And somewhere in there there's an article that I wrote for the welding handbook that discusses some of this. But at the other end of the spectrum, with a laser, electron beam, when I get to a million watts per square centimeter, in Forte's [Fourier's] first law of heat conduction, I'm putting the heat in so fast that the material can't carry it away. And not only do I reach the melting temperature, I reach the vaporization temperature. I start boiling the metal, if you will. Not really boiling it, but it's vaporizing it so fast that the reactive force of the vapor is coming off so fast, it's like a little mini jet engine in terms of the vapor coming off. It actually digs a hole in the material and gives me a very narrow— I can drill a hole with a laser or an electron beam at 10 to the 7th, and I'll show you that in a little bit, okay. But on this one dimensional graph I can explain everything about melting by a surface heat source, okay. That's neat. You go back to the fundamentals and you can explain everything.

Now it turns out, if we want to just look at the flames at the beginning here, the temperature of a flame is based on a number of things. One is the enthalpy of the reaction, okay. Certain chemicals burn hotter than others, okay. It turns out if I look at that handout that I'd shown you before— [Tom puts up a slide] this is a plot— one, if I need to— Steve, you want to, there's a— on the wall over there there's a panel and it will— you can have room darkness, aid probably on the right, white button down there. No, put the room darkening down. We've got, there you go, this way. That's fine, it'll keep coming down.

Okay, so this is a lousy plot but you have, but you have it in your handout. It is temperatures, degrees centigrade, 2800, 3000, 3200, really 2800, 2900, 3000 degrees centigrade for oxy-acetylene flames. And up here you've got, this is the carbon to hydrogen ratio. Up here you have acetylene, propadiene, you have propylene, butene, okay, a number of things. And the formula for some of these things, which if you want to make notes you can, but on some of this that I'm writing down— acetylene is C2H2, it has a ratio of carbon to hydrogen of one. I did that in my head right? Two over two. Propadiene is C3H4, also known as MAPP gas, which no longer is sold— was sold until about two years ago in this country. Propylene is— seeing, propadiene— propylene is C2H6 [C3H6] if I remember.

It turns out, just straight thermodynamics, any of you ever take chemistry? Some of you are nodding, okay. In chemistry it turns out carbon will burn hotter than hydrogen, okay. Hydrogen actually burns with an adiabatic flame temperature, if I just put the best ratio of hydrogen-oxygen, two to one, and ignited it, the highest temperature I could get is almost 2800 centigrade for pure hydrogen, pure oxygen. You may have done that in high school, okay, make a little explosion in a jar. Acetylene, which has a higher ratio of carbon, actually burns at about 3600. But we're not burning pure carbon, we're burning a one-to-one ratio, 50-50, and you'll get about 3000 degrees Fahrenheit— centigrade— with acetylene. And so we're going to talk about acetylene later, and why we use it, and when we use propylene and MAPP for cutting, flame cutting and things.

But in fact the first thing we need to know is, if we want to know the temperature of a flame, what's the enthalpy of the reaction, how much heat, how much chemical heat is released when we burn this fuel? We then also need to know the stoichiometry. Do we have the right ratio of hydrogen and oxygen? And we need to know if we have any inerts left over, such as nitrogen, okay. If I have nitrogen, the nitrogen is just carried along as baggage. And I gave you a copy of— actually I told you to pick one up, but it turns out there's one in the packet too in black and white. But this is my World Trade Center paper, okay. About three weeks after the World Trade Center, an editor, metallurgical editor, asked me if I'd write an article. And I was so sick of hearing people say that the steel melted in the fire, you know, it's all over the news, and I knew it wasn't true, okay. Because I teach combustion. If I could melt steel that easily, then we didn't need Sir Henry Bessemer in 1856 to teach us how to melt steel and bring us into the steel age, okay. You can't melt— some, melt steel, even, you can barely melt it with an oxy-acetylene flame, and you certainly can't melt it in the air.

And this article actually goes through the combustion and the enthalpy of reaction and the fact I have nitrogen in the air which lowers the 3000 degrees centigrade temperature of acetylene. And if I was talking about fuel gases at 2800 centigrade like gasoline, aviation fuel, in pure oxygen it might burn at 2800 centigrade, but in air with four moles of nitrogen for every mole of oxygen, it's going to be burning at about 2000 degrees. And I can prove that to you right now, because I have some MAPP gas. They don't sell it anymore, but I have— so I have a number of bottles in my office.

[Tom sets up two torches.] Now this is probably not approved by the MIT Safety Office, me doing this in a classroom. But this is a steel welding wire. This is propane, C3H8. This is MAPP, C3H4. I'm going to put a piece of steel in here at the same time— which one goes red hottest first? C3H8, C3H4, the MAPP gas. And I can hold it here all day long and I will not melt the steel, okay. I might oxidize it away to nothing, okay, but, and I might soften it, okay, but I won't melt it. I'll hold that till it cools off long enough, I don't want to put on the table and start a fire, thanks, then they really would get upset with me.

So the enthalpy of the reaction makes a difference. This is a hotter flame, I can melt steel much faster with a MAPP gas torch. I can solder faster with a MAPP gas torch than I can with a propane torch. Ninety percent of the MAPP gas cylinders are used by professional plumbers, but they no longer sell it. There was one firm in Chicago that made it and they got shut down by the EPA about two years ago. And so now if you go to Home Depot or Lowe's, you'll buy, still buy a yellow cylinder, but it'll be propylene, C3H6, okay, because no one makes C3H4 anymore. I got about ten, twelve cans to stockpile, okay. But plumbers use it because they can solder the joints twice as fast with higher heat intensity. And you saw, I just demonstrated, your heat intensity based on enthalpy of the reaction.

I didn't really necessarily have great stoichiometry control. The stoichiometry control, if I take this apart— see all these little holes here, it's not for light weighting, there actually is a little hole in here, a little orifice, and the gas comes shooting out there at near the speed of sound. It's called choked flow of the gas, and that's how you meter the gas coming out of here. These things will run for about three hours if you turn one on and just let it go. It's got about ten cubic feet gas equivalent as a liquid, and it'll go for about three hours before it burns itself out, okay. And if you hold it upside down when it's almost burned itself out, you can get liquid come out, and then you get flares and people get burnt, and the way it tells you not to hold it upside down. So the inert gas, if you're burning in air or whether you're burning in pure oxygen, makes a difference.

Now the combustion turns out— before I put that up— the surface heating of a gas by flames is sometimes called the combustion intensity. Now you have to be careful, if you're working a boiler and these guys start talking about combustion intensity, they're really talking about watts per cubic meter or cubic centimeter, because there's, they got a boiler of so many cubic meters and they want to know how many watts I'm generating per cubic meter inside my great big boiler. But if you're talking surface heating, which is only about ten percent of the literature on flames, it's watts per square centimeter, just like Forte's [Fourier's] first law. And the heating value of the gas in joules per cubic meter times the gas velocity is what's called the combustion intensity. The higher the temperature of the gas and the faster I bring it to the surface, the greater the heat on the surface, okay.

And that's going to— this heating value is proportional to the temperature, burning temperature of the gas, the adiabatic flame temperature. And the gas velocity is also proportional to the temperature, because the gas velocity— I'm essentially shooting this gas out and expanding it, the velocity is created by the PdV work of the burning gas. So it's temperature times temperature. This is proportional to temperature, this is proportional to temperature, if I have proportional to temperature times temperature, the combustion temperature goes as T squared. And we'll talk about that later.

But it turns out that— maybe I got a lot of order here— well, surface heating by the flame is controlled by gas boundary layer at the interface, okay. If I got some gas coming through a tube here, like this tube right here, and the gas hits the surface, it goes off to the side. And you'll learn, how many of you have boundary layers? You're all going to have it before you graduate. You're going to be sick of boundary layers, okay, but causes boundary layers on flow. Anytime you're flowing something like a ship through water, there's got to be boundary layers, creates drag and things like that, okay. Anyway, there's a gas boundary layer, and there's a free stream velocity, and then the velocity has to drop to zero, that what they call the no-slip condition on the solid boundary. And so you have a gas boundary layer, and the heat has to diffuse across this boundary layer. In order to do that I lose some of the efficiency. And it turns out that the combustion intensity, this is a theoretical amount of heat that I've got coming out of that tube, but only two to ten percent of that heat actually gets into the steel. Ninety to ninety-eight percent of the heat is just blown off to the side, is used to heat up the welder, the guy doing the welding, okay. Welding is hot and dirty work, okay, so far as that goes.

So the theoretical heat flux, which is the combustion intensity if I look at the heating value of the gas times gas velocity, the rate at which I'm bringing it to the surface, most of that is wasted heat. It's very inefficient, so far as that goes, okay. Now what do I do to try to get better efficiency? And this is one of the reasons I handed out the World Trade Center paper, because there's several different types of flames and I actually go through that. I sometimes describe the World Trade Center paper as what can you say about a subject when you know nothing at all about it. And this article then, I wrote three weeks after the World Trade Center, became for a while the most cited article on the World Trade Center. I think it's now number two. If you Google "WTC collapse" it might be number three now. But it became very popular and infamous at the same time, in part because I wrote it so a good high school science student could understand what's going on, okay. I tried to keep it simple.

But I talked about different types of flames and what the temperatures are, same types of things I'm talking about here, maybe a little more compactly than I'm doing it right now. But there's three types of flames. One's a diffuse flame, one's a pre-mixed flame, the other's a jet burner. In the diffuse flame you have fuel and oxygen, and they both come out, but they come out and they just sort of mixed together. Or you just have fuel shooting out. If we just had fuel shooting out without the— maybe I haven't tried this before— if I cover up these holes— [Tom tries to demonstrate] not working very well, okay. Okay, well I have to do something to try it. Oh, I got holes down here, I can't hold them all, okay. [Sets it aside.] Close. But if I kept the oxygen from mixing, I would end up with a very yellow flame, incomplete combustion, soot particles all through the flame.

But if I have a diffuse flame and the fuel and the oxygen have to mix together, I'll get— actually no, I'll get a nice yellow flame. How many of you have gas stoves at home? Okay, have you ever seen, it comes out as a nice blue flame because that happens to be a pre-mixed flame. If things are messed up you can see a yellow flame. Or even if you got a propane grill at home, if things are messed up and it doesn't mix the air coming, the air with the fuel coming in before it burns, it won't be a nice blue flame, it'll be a nice yellow flame because you got a lot of soot in there. So with a pre-mixed flame you have the fuel and the oxidant and you mix them together and shoot them out, and you have constant pressure burning, and you get one to a hundred times the velocity of the diffuse flame, okay. I mean if I were to light a match, you see a yellow flame, okay, this is a pre-mixed flame, it's coming out as a jet. Whereas the— the match is a diffuse flame, your fireplace is a diffuse flame, your stove is a pre-mixed flame, this is a pre-mixed flame, okay. And so you get one to a hundred times the velocity, which means you get a higher combustion intensity.

And the last type of flame is a jet burner. In a jet burner— the pre-mixed flame was a constant pressure flame, you let everything out in the air and you burn it in the air at constant pressure. A jet burner, you mix and ignite inside a closed volume. This is a constant volume flame. For the, you thermodynamics people, you do things at constant pressure or you do them at constant volume. A jet burner, you mix and ignite in a constant volume and it comes shooting out with five or ten times the velocity of a pre-mixed flame, which is five hundred to a thousand times the velocity of a diffuse flame. So you get lots more velocity, okay. And because of constant volume burning, because you're getting all your thermal expansion in here and shooting out the hole at a very high rate.

Now it turns out that we're going to learn that arcs are nothing more than electrically augmented flames. One way that we can go even better than a jet burner is to use an electric arc and actually pass current through that boundary layer, okay. So while we take a ten-minute break here— or, what do you guys, yeah, you got about an hour right? While we take a ten-minute break, because you get an hour before your 9:20 or 9:15, right? So the restrooms are out here and I'll get some water.

[After break.] The World Trade Center is sort of an aside. I wrote the article, it didn't take long to write the article, and it has drawn more interest than anything else I've ever written. It's still wrong, you know, their theory, but okay. You ready? Okay, so we were talking about pre-mixed flames and such things. And somewhere in here— find it again— is a, let's go back to this heat intensity plot that I had earlier, but I'm going to do it a little bit differently. So this is the same one dimensional heat intensity plot, let me open this up. It's a little dark. Okay, watts per square centimeter, 10 to the 3rd, 10 to the 6th. Those three orders of magnitude will have all fusion welding processes. So we have oxy-acetylene, we have friction welding, we have arc welding, we have resistance welding which I haven't talked about, laser and electron beam over here. If you go to the 10 to the 7th, you're just going to drill holes. You won't fill the holes back in with molten metal, you just vaporize everything away. If you're less than 10 to the 3rd, you won't even melt the metal. So everything has to fall in that area.

Now it turns out with that simple concept you can come up with a number of other things, such as: as you go from your left to right across here, you will have decreasing heat affected zone size as you go across the heat— and that's why I brought— I don't have a good laser weld but this is actually a piece of HY-80 welded about thirty years ago at Electric Boat, and you can actually see the heat affected zone. You can sort of see it here, I'll pass it around. But here's the weld in here, in this little black area on the side is the heat affected zone. So you can pass that around. Here's another weld which I think we did for an oilfield blowout, and it's an arc weld as well. And I guess that was nice, if this was been polished, and so you didn't show up right as well, but you'll see the heat affected zone around the molten metal there.

It turns out there is a heat intensity or a process that produces about 10 to the 5th watts per square centimeter, and that's resistance spot welding. And I'll talk about that a little bit. But basically it's used in automobiles extensively for putting together the steel sheet metal. So these two pieces of galvanized steel, you put them together, you bring in two copper electrodes at about 10,000 psi pressure, you pass about 10,000 amps for about a third of a second, and you just melt the inside. Okay, you form a little button. The guy who invented this was Elihu Thomson, a professor of electrical engineering at MIT. He was president of MIT for one year. He started a company with a guy named Thomas Edison right up here in Lynn, Massachusetts, called General Electric. Okay, Thomas Edison had about 400 patents, he had more patents than anybody else. Elihu Thomson had about 380, he was number two in patents at the time, around 1890s or so.

Anyway, so Elihu Thomson sort of invented resistance spot welding. I don't know if we're going to talk about it, but I have a lecture on it that I sometimes give. And I point out that they put about 3,000 spot welds in the average automobile because you need 2,000 good ones, okay. They're not, they're easy to make. Costs about a nickel to make a spot weld in an automobile. If there's 3,000 spot welds and a nickel apiece, what's that, a hundred and fifty dollars or something? So it's not all that expensive. In the aerospace industry, they spot weld the combustor cans on the jet engine. Those welds probably cost about three dollars apiece to make because there's lots of quality control that goes into the combustor cans on a jet engine. It's a little more critical than just having a rattle in your door after a few years, and so they have a lot more quality control that goes with it. But in any case, I figured one time that General Motors probably makes 50 billion spot welds a year, so far as that goes.

So there's decreasing heat affected zone size, and there's a paper, I mentioned it briefly before, in your handout, an article that I wrote that basically tells you what types of weld— the heat affected zone, one is right over here. So this is the range of width of heat affected zone as a function of heat source intensity, 10 to the 5th to 10 to the 6th. And a lot of people will say that the size of the heat affected zone in a laser, electron beam weld goes down to zero, but that's not true, okay. In fact you'll find that it says that in the welding handbook, but I've sort of made a career out of going to welding conferences and knocking what it says, the welding handbook is being wrong.

If you look at the temperature versus time for a welding thermal cycle on a traveling weld, it looks something like this. And this might be the melting point. And it takes a certain amount of time to rise up to the melting point, you'll be melted for a certain amount of time, then it'll solidify. You're melted in this region in here. And it doesn't really matter, you can prove that in a dimensional analysis, you'll always have this general shape whether you're at 10 to the 3rd or 10 to the 6th watts per square centimeter. You have heating and cooling. But obviously at higher heat intensities, this thing shrinks down in time to a very narrow value, something that might look like this, okay. And the difference is the heat affected zone size. An oxy-acetylene weld grows while the material is heating, okay. And it also does that in an arc weld. But then a laser, electron beam weld, you actually melt the material so quickly that it's the time that it takes the heat to diffuse into the material that the heat affected zone grows. And so the heat affected zone grows on cooling in lasers and electron beams.

So in an oxy-acetylene weld or an arc weld, the heat affected zone gets smaller and smaller as you go to higher heat intensity. But then once you kind of cross the peak and the heat affected zone is growing over the time that it cools, you don't get a smaller heat affected zone, because the heat affected zone is roughly equal to the width of the weld zone, okay. You put all this molten metal there instantaneously almost, and then it has to soak into the surrounding material. So you don't get rid of heat affected zones.

Now I once saw, was at Lawrence Livermore National Lab, and they showed me a picture of a weld. We were talking about convection in weld pools and stirring of the metal and stuff, and they showed me a cross-section of a piece of metal in the weld there was no heat affected zone. And I said, what type of material is this? I don't see the heat affected zone. They just kept on talking. And I asked the question four or five times, they kept on talking, and finally I pinned them down, we faced most of, what they questions they had for me. So what material is this? And they said, well, it's plutonium, okay. Turns out plutonium has very low thermal conductivity, and so the heat doesn't soak into the plutonium like it does into steel or aluminum. So the heat affected zone will be deeper in aluminum or copper, which has good thermal conductivity. And something that has almost no thermal conductivity, that's an insulator, will have very small heat affected zone. So next time you're welding plutonium, okay, remember you don't have to worry as much about the heat affected zone.

But in any case, the heat affected zone is controlled by the heat flux. It's just a question of when, so on heating or on cooling. You have, as you go across from your left to right, you have increasing travel speed. And that's what this plot is, heat intensity versus the— whoops— interaction time, seconds is one of it. Okay, the travel speed would be one over the interaction time. There's a certain amount of time, as you're sweeping this traveling heat source over the surface of the material, a certain amount of time that it takes to go past its spot size. And this would be the travel speed, this would be the interaction time. In lasers and electron beams I may be going at a hundred inches a second, 100 inches a minute, okay, whereas oxy-acetylene I might be going at one or two inches. So I have to, I go slower with oxy-acetylene than I do with lasers and electron beams.

Now it turns out about twenty-five years ago I actually was one of the people who pointed out in the welding community why we often trained people with oxy-acetylene welding rather than arc welding. Any of you ever taking any welding classes in high school? They train you on oxy-acetylene or arc welding? Oxy-acetylene? A lot of times they start with oxy-acetylene. Anybody else? Yeah, you started with arc welding in high school. Okay, did someone else raise their hand? Okay. In any case, about a half to a third of the time they start you with oxy-acetylene, other times they start you with an arc. Typically if it's going to be something that you have to learn within about a week, they'll start you right off with arc. But if it's something where you have a whole semester to learn how to do it, they will start you off with oxy-acetylene. It all has to do with your reaction time.

Anybody ever played a video game? You don't start out the video game on the highest speed, you start out at the slower speeds. And the interaction time for an oxy-acetylene weld, the time for that weld pool to heat up, is seconds before you melt, okay. And in seconds you can— your reaction time is pretty good. Anyone ever played the bar game of trying to catch George Washington's nose in a bar? This is basically, you take a dollar bill and you bet someone that if you drop the dollar bill and they have their fingers as close as they can to George's nose, that they can't catch it when you drop it. If they do, they get to keep it, okay. You know, so you can do this betting game. And now I can do it every time because my right hand knows what my left hand is doing, right? But if we're actually doing it with another person, it turns out it's about— I calculated one time— it's about 0.12 seconds to drop that distance. What's the typical reaction time? You go over here to the Museum of Science, the heavy little thing, the light goes on, you hit a button. Typical reaction time is 0.2. For most people, stop. 0.12, it might be— [Tom demonstrates with a student.] point we're going now, you owe me a hundred if you get it, if you don't get it, okay. So this is Ben's nose. Oh, sorry, let's try that again. Anyway, you can't do it unless you cheat, okay.

Now I've done it numerous times with students, and if it's a hot day my fingers are sweating, the students have caught it because there's a little adhesive bonding of the moisture. That's the adhesive bonding lecture. But if you have a little moisture on your hands it might stick, okay. So I've had students catch the hundred. Okay, they always gave it back. Anyway, they hadn't gotten their grade yet. In any case, this heat intensity controls the interaction time, the spot size, the weld pool size, a number of different things, okay.

So increasing heat efficiency. At 10 to the 3rd you— oops, there, go back down— well, at 10 to the 3rd you have, we already said, somewhere between 90 and 98 percent of your heat is rejected as hot gases. You only have two to ten percent efficiency in terms of melting of metal with the heat you're generating in the flame, okay. Arc welding, there's lots of ways to define it, but you're typically at thirty to fifty percent efficiency. You either, here over at laser and electron beam, you can be at 99 percent heat efficient in terms of the heat you're generating going into melting metal, okay. In arc welding a lot of your heat is just heating up the heat affected zone. But in laser and electron beam you don't have that wasted heat to preheat the heat affected zone, because you just melt everything almost all at once, and then the heat affected zone grows on cooling. That's not wasted heat, okay. So there's a difference in heat efficiency from about ten percent to a hundred percent going across here.

There's an increasing need to automate, okay. The reason you started with oxy-acetylene in high school, you had weeks to play with this, and you could eventually get to arc welding. Did you ever get to arc welding? Okay. It turns out as you're watching that puddle, you learn, just like in a video game, what to look for and what to respond to. And if the reaction time is on the order of a second or two seconds, that's the slow speed on the video game, when you're doing oxy-acetylene. You get to arc welding and you can prove that the interaction time is on the order of a few tenths of a second, which means you've got, you're getting on the higher skill level of the other video game, and you've got to really pay attention. And it's harder to learn, start out learning in the middle level of the video game with arc welding. But if they only got a week to teach you, they don't have time, and they'll just allow you to make lousy welds, without— no, sorry, but you know, to a certain extent how much time you will spend.

But increasing need to automate— when I get over here to laser, electron beam, no one is taking a hand torch and laser welding that I know of. You have interaction times down here on the less than the tenth of a second, maybe ten millisecond timescale, that if you don't move that beam every ten milliseconds to the next spot, you're going to burn a hole, melt through, too much heat in one spot. No person can do all that. So it turns out you have to automate when you get over here to laser and electron beam. It's just, there's nothing else to it, you have to do it. And that leads to increasing equipment costs. It's not so much that the laser, the electron beam or the laser, is so expensive, it's all the automation and material handling that goes with it.

So it turns out I can set up to do oxy-acetylene welding for about a thousand dollars. I mean I can go down here to welding supply place, buy a couple of tanks— tank of oxygen, I mean buy the tank, not just rent the tank, buy the cylinder, buy the acetylene cylinder, buy a torch. For a thousand dollars I can be in the oxy-acetylene business. For ten thousand I can be in the arc welding business. For a million I can be in the laser and electron beam business. Not that the laser, electron beam cost me a million— they only may only cost me a hundred thousand— but in fact all the automation equipment that goes with it ups the price.

So it turns out instead of watts per square centimeter, I could change that with dollars per capital cost of equipment, which is something I did about 20-25 years ago. Costs of welding capital equipment cost in dollars, 10 to the 3rd. Productivity, centimeters of weld per second, this is the maximum speed. So I may only be able to do two inches or two and a half, let's see, a minute. Okay, I may only be able to do two or three inches a minute with an oxy-acetylene torch. With an arc I can do ten centimeters or 60 centimeters a minute, that's kind of the highest speed with an arc. With laser, electron beam I can do a hundred centimeters per second, okay. And we might talk about some of those things.

Resistance welding on the other hand has an effective speed and cost which is significantly less. So why does General Motors use resistance welding to put automobiles together? Well, there's a lot of reasons, but one is cost and productivity, making lots of welds. It's cheaper than laser and electron beam, okay, on a per unit length of weld basis, significantly cheaper, okay. And the capital cost of the equipment, you can just replace the watts per square centimeter with dollars per capital equipment.

There's increasing depth to width ratio. If I look at an oxy-acetylene weld at 10 to the 3rd, I'll get a very shallow— did they ever take a cross-section of the welds and polish them so you can etch them so you can see how deep the penetration was, do you ever do that? Okay, sometimes in teaching students will have them do that. But an oxy-acetylene weld, you might have a depth to a width ratio of three to five, okay. Or actually it'd be one-third to one-fifth, if it's depth-width right, the depth is not very great compared to the width, okay. With arc welding, you can look at those arc welds I just passed around, it's about a depth to width ratio of one, okay, because it's more of a point heat source. The flame is a distributed heat source, the arc is really concentrated there in the center, and so ideally it's a semicircle. It's not really a semicircle, there's convection going on in the pool and stuff. But a laser, electron beam, I can get weld pools that have depth to width of 10 to 50.

Now it turns out we're going to, when we talk about, we never want to get more than about 10. But you can get 50 in the laboratory, okay. You just can't fit things together reliably in the field with a ratio of 10 to 50. All of a sudden your alignment problems get to be horrendous, okay. And the last one is increasing production volume requirements. Let's move it up. Okay, if I'm going to be making a hundred centimeters a second of weld, okay, I gotta have a lot of material to weld, right?

And as a result, electron beam is used in two industries basically, and laser to a certain extent. Lasers are used in a few more industries, but electron beam is used in two industries. Anybody know which ones use electron beam fairly— I mean ninety-eight percent of equipment is sold into two industries. Automotive, where you have tremendous volume. The catalytic converters a few years ago on Fords, where you've got two pieces of stainless steel to come together like a clamshell, and you've got your ceramic filter on the inside, that seam between the clamshell, between the two pieces of stainless steel coming together, that seam was electron beam welded. They had six electron beam welders at Ford that were sealing up these catalytic converters. It may have been 100 centimeters a second, but it was probably 60 centimeters a second. Because they got millions of these things to make, and you start figuring out how many shifts a year and how many inches and how many, you know, of weld you've got to make. If they did that in gas tungsten arc welding at 10 centimeters, they'd have to have 60 machines, okay. Or with two or three centimeters they may have to have two hundred machines. Now the capital cost is in favor of the very high productivity electron beam welder.

Now they— we might talk about when we get to electron beam— they actually used an out-of-vacuum electron beam, which was a maintenance nightmare, but we'll talk about that later. What was I going to say about— oh, production. The other industry is aerospace industry. The other industry is aerospace— GE [I think you can do this anymore because they're closing it down now], but they used to weld together two turbine disks. I was called up there in the mid '80s by a guy I had worked with. Actually I had a resistance spot welding contract with General Electric. I spent the mid '80s over in Tokyo, Japan with the US Office of Naval Research as a liaison scientist. But right before I left, when General Electric came to me and said we'd like to give you a research contract on spot welding of these combustor cans for our aircraft engines.

And I didn't know it at the time, but I was the whipping boy for General Electric Research in Schenectady, New York. They had been working for five or six years on SWAC, S-W-A-C, spot welding adaptive control. And General Electric Research had spent seven million dollars developing this spot welding adaptive control technology. While I was off in Japan, I had a postdoc and a graduate student who were working on this project, and they showed that this process was not working at all, okay. General Electric Research in Schenectady had blown seven million dollars, had nothing to show for it, okay. So what had happened is, they were supposed to spend their money with General Electric Research, rather than going off and giving it to some stupid university. But they decided, the managers up here at Lynn decided to cut out General Electric Research, and give a portion of the money they would have given to them to me. They were giving over a million dollars a year to them, and they gave me like forty thousand dollars to do some work, fifty thousand, I remember it was a lot less. But we're cheaper, right.

But we actually showed their control algorithm, which had to control within a tenth of a second, actually didn't even start measuring until about two tenths of a second. So the weld was already over before it was adaptively controlling the weld, okay. Kind of hard to adaptively control when the process is already over. But nonetheless— but because, when I got back and we were working on this, and I was hoping to get the next year's funding from General Electric, and I sort of learned about, I was being the whipping boy for General Electric Research to get them in line. The engineer called me up and he says, Tom, we got a problem with electron beam welding of our titanium compressor cases. So these are like thirty thousand dollar, you know, rings of titanium, okay. And they want to weld them together with an electron beam circumferential weld about a half an inch thick. And they do it in this big electron beam chamber about half the size of this room. And they put the two in there and they would run the electron beam weld and they were getting pores in their titanium welds, pore, porosity. And you would, you're supposed to take two of these thirty thousand dollar units and weld them together, now it's a seventy or eighty thousand dollar unit. There's a lot of value added here. And if you make a bad weld you now have scrap. So the loss is about sixty thousand dollar loss or twenty thousand dollar gain, you know, if you actually make a good product.

And I told the engineer at General Electric, I said, what are they using to clean it? He said acetone. I said, oh, okay, why don't you go and get some reagent grade acetone and have him start cleaning it with reagent grade acetone? He says, okay. Well, about five, six weeks later he calls me back up, and says, Tom, we're still getting porosity, and I was wondering if you could come up and look at our process. And I said, yeah, what happened when you used the reagent grade acetone? Oh, we haven't tried that yet. Okay. Oh, okay. And by the way Tom, we don't, this is an old engine project, we don't have an account number to charge your time to, so I was kind of hoping you'd come up for free, since we have this other spot welding contract that's at, oh, okay.

So I agreed, I went up there. Some o'clock one morning they walk me through the whole cleaning process, and then we went into the room where the electron beam was, and they had a couple of these they were going to put together. At that point I counted them, 19 engineers or technicians from General Electric showed up. I don't know what account number they were charging their time to, okay, but I know that no one does anything up there unless they have an account number to charge to. But they didn't have any money to pay me, but these other 19 people were all very interested in this.

And so they got these two things, and the two surfaces before they put them together and put them in the machine, and I said, can I look at it? And I took a little 10x loupe magnifier, and I'm looking at the surface to see how clean it is. And as I'm looking at it I go back to another area, and I noticed there's little white spots that were not there when I looked at it 15 seconds ago. And I kind of look around and I realize it's raining dust in this room, okay. That's all you need is a little dust spot or a fingerprint. Okay, originally I was looking to see if there were fingerprints, but people were using white gloves and everything, and the guy just wiped it down with the acetone and stuff. And I asked him, where you getting the acetone? We got a 55-gallon drum out back. Okay. And I'm looking at it, hey, dust is falling out of the air. And the technician who's supposed to be putting this together, he starts fussing with, well, I would have had them together by now if you hadn't been, if you hadn't interrupted what I was doing, you know. Oh, so it's my fault that they're getting dusty.

Hey, so at that point everybody decided I was just a klutz and creating more problems. So they put it together. In the meantime I found a little glass slide over on the table and I cleaned it off with my handkerchief, so it was nice and clean. And I took some of that squirt bottle of acetone and I squirted it on the glass slide and I blew on it to evaporate the acetone. Everybody else is looking at the guy loading the thing and ignoring me at this point. And I dried off the acetone, you could look, and you could see Newton's rings. People know what Newton's rings are? If you have oil slick on water, you actually see these rainbow of colors, those are Newton's rings, okay. If you have a thin film of grease on water, or in this case on a glass slide, then you actually see these colors, okay. This was a clean glass slide, I had evaporated the acetone. This 55-gallon drum of acetone was industrial grade acetone. It wasn't reagent grade acetone, and in fact it had oil in it. It's not uncommon. I don't know, most industrial grade acetone has one or two percent oils and grease in it. It says it's a grease solvent, okay, that's what you use it for. And if you just buy industrial grade, it's not really very clean.

So they get it all put together, and it's going to take half an hour for the vacuum system to pump down the chamber before they can make the weld, it's now 11:00. And Lyle says, well, why don't we go to the cafeteria and get some lunch? So he and I go over the cafeteria, everybody else decided, like I said, decided I was a klutz, and they went back to work because they couldn't keep charging on this account. We have lunch, we come back, and they had a leak in the vacuum system so it still hadn't pumped down. And I said, well, let me know how it comes out, I'm not hanging around anymore today. So after lunch I went back, came back to MIT.

And about two months later— well, before I left I said, Lyle, get rid of this industrial grade acetone, get some reagent grade acetone and clean it off with the reagent grade acetone. So this time he did. And two months later I get a phone call from him, he's all upset because he had used the reagent grade acetone and the porosity went away, because they weren't covering it with grease, okay. And he was upset, cause all he got— they had welded six of these in the last two months and they had projected cost savings of half a million dollars a year, and all he got was a letter of commendation. I said, well, I didn't even get a letter, okay.

Anyway, so there's a couple of morals to that story, I tell you. I like to tell stories. The fix is often a simple fix. You actually have to try the fix to see if it will work. If you don't try it, it doesn't work, it's sort of a given, okay. But I don't even know what all the others are, you can figure out your other morals of that story. But I could give you other stories about General Electric and other companies about they just don't, I mean, then we want to try it, okay, and when keep on doing it the same way. And you know the definition of stupidity, right? Doing the same thing over and expecting a different result, okay.

Anyway, so it turns out in the aerospace industry we use electron beam welding and laser welding because they can give us faster, deeper, smaller heat affected zone, less residual stress welds. In the automotive industry we do it because we need, we have enough material to feed the beast, okay. In aerospace welding that electron beam unit may only be on two percent of the week, okay, whether you need it because there's no other way to make that joint. And the value of the parts you're putting together, you can easily pay for that capital cost of that equipment because of the high value-added of the parts, okay.

So let's take the last couple of minutes here. Since you're in the Navy and you're interested in welding ships, among other things, you are interested in turbine blades and other things in different cases, so why don't we use laser and electron beam in the shipyard for welding? You don't have the production volume requirements, right? You don't have the value-added of an aerospace component most of the time. Sometimes the Navy does, and sometimes they use laser welding and stuff. The Navy has an applied physics research lab at Penn State University, okay, where they do all kinds of work on laser processing. The Navy had the first 25 kilowatt non-classified laser welder in the country, back in the mid '70s at the Naval Research Laboratory. Because everybody wanted to use laser, high-power lasers, which were coming into their own in the early '70s, for improved productivity.

But in fact, if you don't have either a very high value-added part like an aerospace part, or high volume that you can just keep feeding in there, then it turns out not to be economical. And it's been proved over and over again, but you're still going to have some manager come, so why don't we use laser welding? Okay, one of my first students was hired by Avco Everett to go over here— this is Avco over in Everett, across from there, there's, BNY Mellon Bank, this big building, that building used to be the Avco Everett research facility, and they used to build these 25 kilowatt lasers like the one that they sold the Naval Research Lab. One of my students went to work there back in the early '80s, and his job was to see if he could have the laser working when the customer came by for the demonstration, because ninety percent of the time it wasn't working. It was really a research machine.

Now today, you know, thirty years later, lasers are more reliable, but it's still a problem of the capital cost of the equipment versus do you have enough material to put in it. I can give you another example of high-power electron beam, laser— well, high-power electron beam welding, and that's bandsaw blades.

Anybody ever go to the Home Depot? More expensive bandsaw blades or hacksaw blades are by-metal blades. Anybody know what that means? Not by metal, not B-U-Y, but bi-metal, okay, bimetallic blades. So bimetallic blade, and this was actually one of the first— I think this goes back to the late 1940s or 1950s for electron beam technology, which was really new at that time. A bi-metal blade, you got a bandsaw blade and it's got teeth on it that looks sort of like this, right? And you'd like the base steel to be spring steel which is hardened and very springy, okay, got a lot of stiffness to it. You'd like the tips to be a high speed steel with molybdenum or tungsten or other very expensive alloying elements in it, because these maintain their strength at high temperatures.

So they actually used to be a firm out here in western— I think in Springfield, Massachusetts, and they had a vacuum system with electron beam, and they would feed in a little strip about two millimeters wide of this high speed steel with a piece of spring steel, and they would weld this together right here, and then they would grind the teeth on the high-speed steel. And so when you buy a bimetallic blade, it means the tips are the expensive steel, and the backing, which is just there to give you stiffness and to hold the tips if you will, is the cheap steel. So it's like getting a very expensive steel for the cost of a low-cost steel. And in fact this steel on the tips has very good high temperature strength, but it doesn't have the same type of fracture and fatigue strength as this steel. So you end up with a composite. And they're whipping that stuff through it about a hundred centimeters, well, I don't know, 10 to the 7th meters a second, but probably 10 centimeters a second, just comes through a vacuum air feed-through, and goes out a vacuum air feed-through, just making reels and reels of this stuff to make bimetallic blades.

And you can now buy hacksaw blades that are bimetallic, okay. So there's an example, there are niche applications for these things. So I'll see you tomorrow, and then we'll do a couple hours a day, if that's alright, if you can survive it. I'll try to survive it too.