SSW_S2013_12

Okay so they didn't miss much, only my ranting about some of my colleagues. Um, but anyway, so this is the old packaging technique which is still used in many things, but these are J-leads coming out of an aluminum oxide package. And this is basically showing you the types of solder flow you can get, many of which are bad, okay, because of variations in the solder wettability at different locations on these things. You'll get different contours, and people used to do computer models of the surface tension of how these things would flow across here. This is a gull wing — this is J-leads and these are called gull wings. Can't read it very well but it says goldwy — gull wing, right down there. That says gull wing. But again you run into problems with how the solder flows.

Is area bonding — we talked about that briefly. And so you have, you can have a PC board that might — well, I showed you this before, but I forgot to bring in the piece. I showed you this Cisco Systems, and you put a bunch of uh packages on each one of these slots, and this is a composite that's probably about 13 layers thick. This other one I passed around was I think 21 or 22. This is one of the layers. And they metallize the layers, and they have to register very closely when they bond them together. But that's a uh resin, um, plastic that's been metallized with copper, and then that's been tin plated, and they're basically gonna form a transient liquid phase diffusion joint, okay — tin into copper. And they'll get the vias between the layers. They may put it in a plating bath later and plate that to get connections going through the thickness of the board. But then you'll take components that may have solder balls on them, you'll flip it over and you'll make the solder bond.

Now it turns out that — and so, we've talked about flip chip and the problem of thermal expansion. Coefficient of 2.6 for the integrated circuit, thermal expansion for glass aluminum epoxy, glass polyester resin which is basically what that sheet of plastic is, of 6 to 27 — you got a real problem. I mean the plastics have large coefficients of thermal expansion. Silicon is a ceramic, fairly low. Sometimes people are bonding from silicon to a copper heat sink. When you get to a powerful chip like a Pentium or one of their Itaniums now, or whatever the Intel chips are, these little chips will put out 30 watts when they're humming along, okay. 30 watts per square centimeter is roughly equivalent to taking a plumber's propane torch and putting it right on the surface of something, okay, just holding it there. That's the heat transfer rate of that flame against the surface. When these chips are operating, you've got to cool them down.

Now, in your laptop and stuff you're running at two percent duty cycle or something, but when you're in a high duty cycle application, you're number-crunching in a supercomputer or something, uh, we basically are limited by the heat we generate. We're limited by lots of things — we're limited by the frequency, you know, how fast you can operate the computer. But one of the things you're limited by is the amount of heat you can generate and pull out of there. IBM, for their 360 — there's a, this is a mainframe computer — but that had, they had gaseous helium going in among the chips. So the whole thing was sealed and they basically had a heat exchanger with gaseous helium.

Why helium? The lighter the element, the better the heat conductivity in a gas. It goes as the inverse square root of the mass of the gas, okay. The kinetic theory of gases tells you that its thermal conductivity of the gas is proportional to one over the mass. And so, if I have uh, if I have air, and you've got diatomic oxygen at 32 or nitrogen at 28 or whatever, but you've got helium at four — one over four, you take the square root of it is one half. One over thirty is about one fifth. So you can get two and a half times the thermal conductivity with helium that you can with air, um, and even better than about one — uh, three times the conductivity you get with argon. This actually gets to be important when we do arc welding, uh, when we like to use helium. But anyway, they now have to actually go to water cooling in many computers, okay, because gas is just not enough molecules in the gas to carry the heat away fast enough.

One of the things about these solder balls — well actually let me back up on the side of solder balls. I told you about the Bond number, and that something by surface tension, if you melt something properly, it should form a sphere if it's less than about three millimeters in diameter. So one of the ways they make these little solder balls on these flip chips, or C4 connections, is they'll electroplate lead-tin alloy. It'll be a little thing like this. And they remove the resist, the polymer that protects the surface and only lets you plate where you want. They end up having this little pad, and then they'll just reflow it. They just melt the lead-tin and it balls up naturally.

And once you get that ball, it turns out, fortunately, again because of surface tension, it helps you if you mess up. And you, if this is your pad right here, and you put — you misalign the chip so the balls are way over to one side of the pad you're trying to bond to, it will actually create a restoring force. The surface tension wants to bring those misaligned balls into registry, so it's self-aligned. So it has a little torque, there's a little translation, XY translation, and it's self-aligning, which is one of the advantages of flip chip and other things.

So, well, there are probably better things than this bigger one I'm showing, but in fact, when you start looking at your silicon chip, and you start looking through your circuit board, and your balls — you may have several layers of these things, but you have all kinds of layers in between with, you know, wires routed. You know, it's just a rat's nest of wiring. And remember, well, back in the old days, being when you were in elementary school, they had a problem: they only put 10 million transistors on a chip and most of them had to be good or you had a bad chip. Nowadays they build redundancy into the chips with 100 million transistors. You can afford to lose 5 or 10 million transistors as redundant, so that you can reprogram your circuits and have other things, other areas. If you have a bad area on the chip you can reprogram it in software, uh, so the hardware takes rerouting. Or you may go in there with a little focused ion beam, which is — anybody know what a focused ion [iron] beam is? It is — you know what it is? What is it?

Gallium atoms, usually, okay. Focused ion beams kind of came around in the 1980s. They've been around since the 1950s, but you take a tungsten tip that you can essentially — they've been making tons and tips very small down to an atomic scale, for atomic probe microscopes, since the 1950s and 1960s. So they have different etching techniques to take a thin tungsten wire and make it essentially atomically smooth. You can have a radius at the tip of 50 atoms, okay. But then they would put gallium on here, and gallium sort of is second to mercury in terms of low melting point metals. Mercury's liquid at room temperature, gallium melts at around 85 degrees centigrade. So it doesn't take much. You put gallium on here and you get a little thin by surface tension, a little thin layer of gallium. And then you put 20,000 volts across here and you can strip off gallium ions, and you get a focused ion beam, okay.

They now have these things in scanning electron microscopes. They first put them in transmission electron microscopes, and you could machine your sample on like a 10 or 20 angstrom scale, okay. Pretty neat. You're just hitting it with 20 kilo — kilovolt, not electrons, gallium atoms. And the gallium atoms got, what, 20,000 times the weight of an electron or something. So you can be pretty impressive, the holes that you can drill on a nano scale.

But in any case, where was I going with that? Um, oh — well, I was talking about one over the square root of the mass and the gas thermal conductivity. I don't know where I was going with that. But anyway, focused ion beams are a technology — oh, because they can repair, they can use a focused ion beam to go on that chip and they can machine things down to 10 angstroms, which means they can break circuits, and they can actually deposit gallium as a solder joint, basically, to make — to bridge new circuits, so you have extra transistors on your chip.

And the largest single use of scanning electron microscopes in the world is quality control for the semiconductor industry. All of your four-, $500 chips, you know, your big computer chips, expensive ones, they all get an inspection in the scanning electron — a scanning electron microscope. Now I haven't seen it, but I'm told there are rooms at Intel with a hundred scanning electron microscopes, just because it can take an hour — hey, it's a $500 chip, right? It can take an hour in the SEM to check it out, program it, find where the good spots are, the bad spots, and everything else, okay. Because one bad joint and the thing's no good. But now they basically have redundancy so they can reprogram it, use focused ion beams to repair them and things like that, okay. So it's been a lot of technology that goes into that.

Unless you've got more questions on micro-joining and electronic packaging, I'm now going to go into brazing and do it fairly quickly. Any questions? Okay. We've already talked a fair amount about brazing. Uh, there are some things about brazing that — now you guys can remember, ordinarily I might be running two or three lectures behind now. It's because you're not asking me questions, so I can digress, okay. What I need is, you know, like the focused ion beams or one over the square root of the mass. I'd much rather digress than go over these ancient lecture notes.

So brazing, we know, is just a higher temperature than soldering. But it's the same process. But because of the higher temperature, you have several advantages. One is more flexibility in fluxes, and the more reactive fluxes become, because the chemical reactions proceed greater at higher temperatures. It's just the Arrhenius equation, right? So while there were some things that couldn't be soldered — beryllium, titanium, you know, tantalum, niobium — because you just can't get rid, you don't have the chemical reactivity at the soldering temperatures below 450 to clean off those very stable oxides. At brazing temperatures there is no oxide that can't be destroyed by a good brazing flux or brazing technique.

So essentially everything — where you can get the right wettability of your filler metal with your substrate, which means your filler metal has to create a lower interfacial energy than whatever your fluxing component has. You want to displace the flux and bring in the filler metal with a lower interfacial energy. And we usually use metals on metals but you might use ceramics on ceramics. You have more flexibility in your choice of fluxes. I have never seen anything that can't be brazed except things like polymers that can't go above 450 C — they just decompose. If it's stable above 450 C as a solid, I have never come across anything. Carbides, okay, in fact I think I passed around a carbide drill tip. Diamond, okay. Actually graphite gets brazed all the time, okay, carbon gets brazed all the time. Works with platinum, works with carbon at very high temperatures. But so, you can join almost anything that's stable at high temperatures.

Uh, room temperature strength is greater. And I told you the other day a very real rough Tom Eagar rule of thumb — you won't find this in the book — for a solder joint, you should never have more than a thousand psi. That's seven meg, seven megapascals, okay. In fact you'd like to be around one or two megapascals as the stress level on a solder joint, because otherwise over a thousand or two thousand or ten thousand hours the thing will just pull apart, okay. With braze joints, a typical strength will be 5000 psi, and creep is not usually a problem — you're at higher temperatures. Now you can go as high as 40,000, but now you really got to get into controlling the coefficients of thermal expansion, because in fact your braze joint should be stronger than even 5000, but you usually have residual stresses in there with this number of metals that lower it.

Yeah, when I say a thousand degrees with solders, I mean that's higher temperature solders to go to a thousand. Lower temperatures it's going to be around 100 psi. Hundred psi the solders, the brazes are stepped up three, four, five hundred degrees in temperature, and so they're over the creep regime. And I'm just telling you, if I had to use a good rule of thumb, 5000 psi is a good number. But I know cases where people braze carbide bits or ceramic bits to machine tools, and they do it — or these machine tools are things where you're tearing up the road or you're mining, okay, you're reclaiming asphalt. I mean I have a student developed a braze alloy for Norton where they wanted to put carbide tips on these things that reclaim asphalt, okay. You know, night — they're tearing up, taking all the asphalt off the road and they leave these grooves in the highway. Those are carbide tips, okay. They might last one evening of tearing up the road. I mean it's pretty aggressive service for something.

Well, if you start losing your tips by brazing, you can't lose very many tips before this whole great big expensive machine goes down and now it's going to take the rest of the evening to re-tip it, and you lose tremendous productivity. So the student actually developed a braze alloy where he got a better matching coefficient of thermal expansion, okay. And what he did is, he ended up using titanium nitride as a paste powder with copper and tin. And I remember he had like six components in there — carbon — and he would actually form titanium carbo— actually this one, I'm sorry, it wasn't titanium nitride, titanium hydride that he used. Titanium hydride decomposes around three, four hundred degrees centigrade, just drives the hydrogen off, so you end up with titanium metal. If you're brazing this in a vacuum, he ended up with a copper-tin titanium carbide composite. Titanium carbide is actually a ceramic metal that melts above the temperature of the brazing alloy. So you're making an in-situ composite brazing alloy. But the coefficient of thermal expansion — he had enough titanium carbide in there, his coefficient of thermal expansion of his brazed alloy was half of what you would get with a copper-base brazed alloy.

With lower coefficients of thermal expansion he got tremendous strengths. I mean, well above 5000. I don't know that we measured the strength — his way for measuring the strength, Norton would go out and tear up roads, okay. I mean, so it was kind of an engineering approach. He wasn't looking at a shear strength or just a simple tensile strength of the braze joint. They just go out and see how long the things lasted. They lasted three times as long. Well that's pretty good. I mean, before they might have to re-tip once a night and lose hours. If you can get the whole evening when the roads are closed and everything, and you know, just run the whole period of time, that's a tremendous advantage in productivity, okay. So essentially he made a composite braze alloy. It was reacting while it was brazing to form solids within the braze alloy, and it was still wetting the alloy. Anyway.

So you have to worry about residual stresses to get the very high strengths, but you can get very high strengths. You have to worry about volatilization — violet, no. It turns out zinc and of course magnesium are big problems. They both have very high vapor pressure. Zinc boils at 900 C. I can't remember what magnesium is, it's like 1100 C or something. You're getting to reasonably high vapor pressures at your brazing temperatures. And in fact, you stay away from zinc in your silver — your vacuums and silver brazes — because you'll just contaminate your whole vacuum furnace. That vacuum furnace just sucks that zinc right out of there and you'll end up depositing it all through your furnace and you'll destroy your furnace.

In fact, I had a situation where a guy had a two-million-dollar brazing furnace down in San Diego, and someone put the wrong braze alloy in there, ran the brazing system, contaminated the whole system with zinc. And then rather than think about how to do it, they decided they would just try to heat the whole thing up even hotter to — just, they were going to vaporize it out of the furnace. Except the furnace has got hot spots and cooler spots. I mean, the outside, you can put — it's water cooled, you can put your hand on the wall of the vacuum furnace when it's running, okay. So all they did was move it into the cooler regions. Anyway, it was just a mess by the end. By the time they screwed up um trying to fix it, uh, the quick and dirty way, they ended up destroying the whole furnace. Anyway.

Um, at higher temperatures you end up with — and I told you about problems like when they were trying to braze the molybdenum to the copper for the cross-field amplifier for the radar thing. And they originally had an 82 gold 18 nickel, and they were forming nickel-three-moly intermetallic, and you get like five degrees bend and the whole thing would snap. And then they got to a transient liquid phase — they only had three percent nickel in their braze alloy, a lot less gold, a lot of copper. Formed a TLP bond with no intermetallics, you could bend the thing like a hairpin 180 degrees, okay.

Of erosion — and soldering, you can, but it's not usually a problem. But in brazing, you can have erosion of 50 or 100 microns or even more. So here, actually, looking down here, I have some other pieces. I'll hand these around but don't cut yourself. This was a heat exchanger about the size of this room. It's all brazed together all at once. This consists of stainless steel, and there are little tubes and big holes, and they — in this particular piece they were crossing each other. And this was for an internal cooled recuperator for a naval ship.

But basically the destroyers and the frigates and the cruisers all run on oil, and they get their oil from the aircraft carrier that runs on nuclear energy. But the aircraft carrier is big enough that it carries all the oil for all the sister ships that are protecting it. Now because of the problem of a nuclear strike, the destroyers and frigates need to be about 30 miles away from the carrier when they're on duty station. The problem is, 30 miles is a significant fraction of how much fuel they have to burn, and they would spend 10 to 20 percent of their duty time just going back to refuel and go back out into station, okay. So they wanted to get to the point where, rather than 10 or 20 percent of duty cycle time was going back and forth to the carrier to refuel, is only like five percent. And they wanted to preheat the incoming air that was going to go into the boiler that's burning the oil.

And so they built a heat exchanger about the size of this room, and they stuck it on a turbine at Rolls-Royce in England. And this whole thing, this whole braze system, was supposed to last for a hundred thousand hours. I calculated it lasted for two minutes before it started leaking. And it turns out, um, it's like a 100-million Navy program, about 15, 20 years ago. And they were trying to save money. It was right after 1992 when peace broke out with former Soviet Union, and the defense budgets were going down, and they decided they'd only do a 2D finite element model of the heat transfer in this thing. And no one told the gas it had to go through this thing symmetrically, okay. And so this great big three-dimensional object, some of the gas went this way, so that went that way. And within two minutes they had tie bars that were six-inch thick hardened steel, okay, on this thing holding the whole thing together. And within two minutes — that's, you can calculate that's how long it takes to start heating up some of this metal to high enough temperatures the thermal expansion stress has just started snapping these tie bars. You know, the whole thing just turned into a pretzel, okay.

And so I was on a team that had to go in and tell them what happened. Wasn't too hard to figure out. Had nothing to do with the braze joints. Had to do with the fact that you cannot assume that the gas will always go through symmetrically, okay, and that things will heat up uniformly because you want them to. Wishful thinking is not good engineering, okay. Anyway, that's the story on those braze joints. But those are some complex — you can make some very complex objects with braze joints.

But now this is another braze joint where it gets to the problem of erosion. Inside a jet engine, you have the turbines spinning around inside the barrel, the — what do we call it, the shroud, the stationary shroud. You've got the turbine spinning on the — with the disc going around and the turbine blade. You don't want gas to slip past the turbine blade at the top. And so they used to originally just try to machine things very precisely, but the problem is these things get hot and they creep, and so their length increases. So if you actually look — not on this one, but if I had some others, you'd see they actually in some cases braze on a carbide, and they want the carbide to eat into the shroud as it gets longer, as it gets older, okay. So you get a perfect seal.

Because you can lose 10, 20 percent of your efficiency of your engine — for the gas is just that you're trying to compress. This is the hot section. But in the compressor section you can lose a lot of your efficiency if the compressed gases start leaking back towards the front of the engine rather than being pushed to the back of the engine. You're losing efficiency. So what they do is they make a honeycomb of high-temperature nickel-based superalloy, and they braze it together. But this honeycomb has only got a thickness of about three or four thousandths of an inch of sheet metal. And this is a used one. You actually can see where the turbine blades have been cutting right through this as it spins around at high speed. This is just a section of a shroud.

Oh, no, it's very fine. This stuff is — this is microchips, they just add more fuel to the engine, okay. No, they're going to be microns, maybe 10 micron size, but it's just grinding swarf, okay, very finely divided, and it'll just oxidize. The engine adds more fuel, right? When they first assemble the engine, what's the clearance? Well, when they first assemble it, the clearance is probably slightly negative, like one or two thousandths negative, and it cuts in. The first couple of times you spin around, it doesn't take long because that honeycomb is pretty weak, okay, but you're just eating into it. Sometimes — well, you can't have too much clearance for the thing, you can't get the thing to spin obviously, but that first time you want to spin it, you've got a little inertia from all the friction. But once you get a good spin, a couple of spins just eating it away, and now you've made a good seal, okay.

Well, so, in the old days, because of the erosion of the base material, they used to use this 82 gold, 18 nickel brazed alloy that I told you about. And the reason being is there's no erosion of nickel and gold against a nickel-based superalloy. But the types of braze alloys they'd like to use, and what they tend to like to use and what they tend to use now, are these things that are nickel-phosphorus, nickel-carbon, nickel-boron. And these things actually dissolve part of the base material when they first try to use these things without any inhibitors. And the inhibitors can be things like chromium and palladium, which one time was cheaper than gold. It's now even more expensive sometimes because of catalytic converters and cars.

But in the old days, when I made my wife's wedding ring, I used palladium because I couldn't afford all the gold — because it was cheaper than gold. I was actually using platinum for her engagement ring that I made the engagement ring out of. But I ran out of iridium, it's a long story, okay. I had to lie to Engelhard to get the iridium. Anyway, iridium was — there's only like 100 ounces a year of iridium mined in the world, and I only wanted half an ounce. But they wouldn't sell it to me. I had to go to my thesis advisor, and I was working on superconductivity, niobium-three aluminum, but there was a niobium-three iridium alloy. And I got him to sign a note to Engelhard saying that we were going to use this for scientific research, and so I got my half ounce of iridium. I made the engagement ring. And a year later, Engelhard sent two people by Bob Rose's office to find out how the research went. And he didn't tell him we already had a son, but anyway, nonetheless, okay, he just made something up. But I had to get permission to buy half an ounce of iridium, which I thought was kind of silly. Anyway. So I ended up making the other stuff out of palladium because we had some lying around the lab. Anyway.

The gold-nickel does not erode the base material. These types of alloys could erode two thousand[ths], 50 microns deep. And so your four-thousand[ths] sheet metal, you get eroded from both sides and you got nothing left, okay. You just have a — you don't have a honeycomb, you just have a blob of melted metal on the surface. So they were using — this is in the mid-70s — 82 gold, 18 nickel. By the time the early 80s came by, they actually got rid of this because the price of gold went up to $800 an ounce from about $200 an ounce, and there was enough financial incentive to get rid of it.

But let me tell you what happened here. At an aerospace company located here in New England but will remain nameless, they would get this as a powder in 50-ounce jars, okay, just a brazed powder. And they would take a salt shaker — literally a salt shaker — and they would put the honeycomb inside the shroud, and they would have a person that would take the salt shaker and shake some of this gold-nickel brazed alloy into the pores on the thing. And then they would take some Krylon crystal clear lacquer like you buy at the hardware store, and they would spray it on there to hold the powder in place. And then they would put it into a brazing furnace and they would melt the gold-nickel alloy and make the shroud, okay, with the seals.

Well, it turns out, in the mid-70s, they were having — they're building a lot of this particular military engine, and the guy would go through — that was in charge of the brazing area — would go through about one jar of this 50-ounce stuff a week. But then production scaled up, and so he would get two jars a week. And to get a jar of this stuff, all he had to do was go down to the stores area, you know, the supply area in the plant. He had to have his supervisor's signature, because if you think about it, at $400 an ounce, which was about what the price of gold was at that particular time, if you've got 41 ounces of gold in a 50-ounce container, that's sixteen thousand dollars a jar, okay. And it's not a very big jar, it's only four pounds of gold which doesn't take up much space, right? Um, there are three pounds of gold, three and a half pounds.

Anyway, so production ramped up and he was using a couple of weeks. Then production started ramping down, he thought, "Ah, what the heck, I'll still get two jars — one for the company and one for me," okay. No one caught him until about 16 months later, someone turned him in. Their whole system didn't have controls. I was hired to come in because the whole thing — the whole operation under government contract — it was bonded, which meant an insurance company had to pay for any losses of things like the precious metals. So the aerospace company they really didn't care, they had been paying the insurance and passing the insurance costs on to the company. But the insurance company wanted to know, well, how much do we reimburse? Because all they were doing is putting, you know — where did they put a lot of salt on their other — on their uh partner, they put a little salt on. So I was hired to come in and do an audit to figure out how much this person stole over the previous 16 months. And turns out, I can't remember, it was like 1.6 — somewhere between 1.6 and 2.5 million dollars worth of gold they got, okay, over 16 months.

But by the time I got there, they still were using the salt shaker technique, but they would weigh the salt shaker before they sprinkled the part, and then they would weigh it afterwards. Before that, they didn't even know how much gold they were using per part. Who cares? The government was paying for it, okay. This is a cost-plus contract, okay. So anyway, I came up with a range of numbers, and I'm sure the insurance company paid the government back that amount of money. But the rest of the story is, the guy in Connecticut who was doing this — I told you I wouldn't tell you the company — the guy in Connecticut who was doing this, well, he got a one-year life sentence, okay, for stealing several million dollars. This mid-70s, so it'd be like stealing 10 or 15 million today, right? And you have to ask yourself whether one year in jail is worth 10 or 15 million dollars, I guess. The guy who was running it to New York City got six months suspended. And the guy who was fencing it in New York City to sell it didn't get anything. He got off, okay. Because it's just an insurance company paying the bill, right? No one really cared, okay.

But anyway, that all had to do with erosion of the base metal in brazing. If you think that's true, anyway. Um, see, I got stories. Um, now, there are other things about brazing, um, that we could worry about. Just like I showed you kind of solder alloy families, braze alloys above 450. And remember there really is nothing between 400 and 500 centigrade. You kind of start the braze alloys in these aluminum alloys. Aluminum is difficult to braze. However, we braze a lot of aluminum. Anybody can think of what the largest application of aluminum brazing is in the world?

Automobile radiators. Back when I was your age, most radiators were copper in cars. And a mechanic could go in there and he could repair a copper radiator — if it sprung a leak he'd just get out his little propane torch and he'd go in there, and he just kind of poured braze alloy in there until it held air pressure, and loaded up until you know another leak starts. But what happened is, in the around 1980 or so, Alcan Aluminum developed a brazing flux for aluminum that was not so fussy you had to do it in a plant. It was the same type of thing — the mechanic in the auto shop could just take some of this flux, paint it on with a little acid brush, and braze it just like he was brazing copper. And this was a fluoride-based flux. I mean, don't tell him that he's, you know, strengthening his teeth by breathing this in, but anyway, um, probably doing a lot of other things — hydrofluoric acid and others. But anyway, they developed a brazing technique so that radiators didn't have to go back to the factory. If you looked in early 1970s, the only automobiles that had all aluminum radiators were things like the Chevy Corvette, where weight was critical. And if you could afford to buy a Corvette, if you sprung a radiator leak you just buy a whole new radiator, okay. You could braze them in the factory in the early 70s but you couldn't field-braze them until Alcan developed this brazing flux. So now, with that brazing flux, you opened up a whole new industry, got rid of all that copper, and now aluminum is used for brazed aluminum radiators.

Silver solders, they're actually braze alloys, have different amounts of copper and zinc. Again, you have to keep the zinc out if you're going to use it in a vacuum application. Down here we got cadmium, okay. Well, cadmium's a problem. Why is cadmium a problem? Cadmium's toxic. It's just like mercury or lead or other things, and it's very volatile. It'll come off. It turns out, you can get what they call — if you were doing cadmium for torch brazing in a factory, or you're a plumber or something doing silver brazing for a critical plumbing application, you could get metal fume fever. And metal fume fever's when you breathe in certain types of metal. Zinc is one of these. Zinc is not really that harmful, but if you work in a brass foundry and they don't have good ventilation, and you breathe in a bunch of zinc oxide — and zinc oxide is not really that harmful. I mean, you go to the health food store, they'll sell you zinc supplements. You just pop a few zinc pills in your mouth. If you don't want to do that, go eat some oysters — about one and a half percent zinc, okay.

So zinc is not harmful in itself, but if you ingest it through your nasal passages, gets down your lungs, the zinc oxide forms the zinc hydroxide, and you get zinc fume — or metal fume fever. I've never had it. A technician I know said he had it once, and he says first you get so sick you're afraid you're gonna die, then you get so sick you're afraid you're not gonna die, okay. But it only lasts for about 48 hours and you're fine afterwards, and there's no long-term effects. Now, cadmium, that's not true. Cadmium, your liver doesn't get rid of it very well. But there are certain advantages to cadmium brazing, silver brazing. First of all, you use less silver than a lot of your other alloys. This is silver content versus temperature. It also operates at the lowest temperatures. And so for production, cadmium silver brazing is great. But in general, they aren't using anywhere near as much because of the toxicity problem.

But there are other stories about cadmium. Light switches, right here, okay. I just released some cadmium atoms into the room as I turn the light switch on and off. Uh, every light switch about below 15 amps will have a little bit of cadmium in it. Because you need a little bit of cadmium oxide with the silver. You have silver contacts because you'll get an electric arc, and if you don't have cadmium, just plain silver contacts are good for about seven amps. If you put a tenth of a percent cadmium oxide in there, you can get 15 amps. Then you can break a circuit at 15 amps without creating an arc — it will suppress the arc.

Well, so in Sweden, back in the 1970s or 80s, they're very environmentally conscious in Sweden, and they decided to outlaw cadmium in the environment. The Scandinavians also wanted to outlaw chlorine in the environment, okay. I don't know how you do that when you're surrounded by ocean, but nonetheless, um, they decided to outlaw cadmium, and they took the cadmium oxide out of their electrical switches. And within three years, they decided they would rather have cadmium oxide in their electrical switches than have people's houses burn down. Because they were burning down houses left and right.

So there are — sometimes the whole system is fairly complex. There are gold brazes, there's the gold-18 nickel, gold-copper, here's a gold-silver-copper. There are lots of other brazes — copper brazes, nickel brazes. We've talked about the boron and phosphorus brazes, manganese and palladium. This is one that doesn't erode things, so when they replace some of this stuff to make those seals for the jet engines, they were probably working initially on nickel-manganese-palladium. And palladium alloys, now we're getting up 1200 C. These are brazes that are up close to the melting point of steel, okay. Um, so you can get some pretty high temperature brazes.

Um, silver solder or brazing or soldering — when you're soldering, used to be lead-tin solders until 1978. Massachusetts outlawed lead in plumbing. Now it's, uh, typically antimony — five percent silver or three percent silver, antimony ten — I'm sorry, tin-antimony with maybe a little silver for strengthening, um. Is that what you're asking?

Well, they call the silver solder is actually a braze, that's what I just put up, okay. And that's a misnomer. It really is a silver braze. If Handy & Harman still existed — they were the leaders in silver brazes — they would call them silver brazes, not silver solder. Silver solder is just a generic — silver solder is just a generic thing, slang term. They are silver brazes, not solders. Soldering the copper pipes is usually done at lower temperatures, soldering temperatures. There are cases like in copper heat exchangers where you are — boilers or something — you might get some higher temperatures, you don't want the creep problems of the lower temperature solders, and you will use silver brazes.

But they don't — most places have outlawed cadmium in the brazes. There are still some cadmium brazes, and just like there's still lead-tin — those little solder bumps on the semiconductor chips, most of them are still lead-tin. I told you, I showed you the thing how you don't need much superheat with lead or lead-tin alloys to get them to flow well. You got so much — so little lead in that semiconductor package. And if you want to talk about toxic metals in a semiconductor package, what do you think gallium and arsenic are, okay? So I mean, the whole thing is toxic, okay. So who cares about a little bit of lead, I mean, you know, um. But because lead-tin has such great soldering characteristics, they still use it in that microscopic amount in those solder bumps. They still use cadmium — you can still buy cadmium-based brazed alloys, but they'll have all kinds of toxic labels on them and everything else, and you really should have ventilation or do it in a hood or something like that, even if you're torch brazing, because you don't want to be breathing cadmium vapors. It's just like lead vapors.

I mean, shoot, when I was a student and you break a mercury thermometer in the lab, you just go get some sulfur and sprinkle it on the floor to make mercury sulfide to tie it up, you know, or you just swept it into the corner and let it vaporize into the room for the next six months. Well, I'm not kidding, okay. Nowadays you gotta call out the EPA and they're gonna come in with their hazmat suits, cordon off the whole room, and declare it a disaster area for the next six months while they vacuum it out for only half a million dollars, okay, I'm not kidding.

Um, they would [hot short] if they weren't so thin, okay. If you had a big bulk thing and it was solidifying in a thick joint, they'd be really hot short, because of the big — the hot shortness comes from the big freezing range, okay. It's phosphorus — phosphorus with copper, phosphorus with nickel, phosphorus-iron — give you big freezing ranges. Silicon with aluminum gives you a big freezing range, okay. And those things cause hot shortness. So it's the big freezing range. If you want to know about hot shortness — that looks like this. Anything over here or here is not going to be hot short. Anything that has a big freezing range is going to be a hot short, okay. Nickel-copper, that's your prototype hot short range. Now, that phase diagram happens to be isomorphous, so it looks like this. That's not hot short, that's not hot short, but everything in between, big freezing range, hot short, okay.

Hot short means as it's solidifying and shrinking, it will just rip itself apart under its own thermal contraction. But the delta T is huge. You know what, if delta T is large, lots of strain. If delta T is small, not so much strain, not so much strength.

Yeah, yeah. So aluminum, when you're welding — arc welding aluminum — you, if you're, depending on your aluminum alloy, if it's heavily alloyed you use a heavily alloyed filler. You don't mix low-alloy and high-alloy to end up with something in the weld metal that's the worst of all worlds. If it's low alloy, like nearly pure aluminum, use a very low alloy filler metal, okay. You have to worry about that with aluminum, you don't have to worry about that with steel, but with aluminum you do, okay. See, these are the types of questions you should have been asking. Also best.

So, class next Tuesday, enjoy the snow tomorrow. I would have lectured a little bit tomorrow, but I decided — yeah, I don't know what it's going to be like. I may have —