§1. What makes something expensive? [02:30]
Student: What makes something expensive?
Expensive is a relative thing. For some people a five-hundred-dollar watch is expensive — you can buy a watch for ten dollars. I wear a Seiko I got for $150 at Costco. You can buy one for $500 at Costco and they might consider that an expensive watch. Then you can start putting diamonds on it, and you can start having a name, which has no functional value — it's all pride — and that could cost fifty thousand dollars.
At the lower end of the materials scale, when we're talking about steel and aluminum and copper, it's the energy content. It used to be labor and then materials, but today it's materials and energy. The labor cost has sort of gone out of these things in the last fifty years. You're seeing a revolution right now equal to the industrial revolution of 150 years ago, in manufacturing productivity. I'm going to put up a slide today: in 1990, steel was one person-hour per ton. Today it's 0.3. In twenty-five years it's improved by a factor of three.
So energy cost is one thing. Once you get there, if you want to go to the really high-value-added materials — fancy composites and stuff — it's all processing. For a pipeline, thirty percent of the cost is the metal that goes into it; the rest is digging the hole in the ground, the inspection, rights-of-way. For an automobile, ninety percent is the processing, because it's a more complex product. For an aircraft or a spacecraft, you're getting to the point where ninety-eight, ninety-nine percent of the cost is all the processing — manufacturing, machining, joining, inspection.
So you really have to talk about the market you're in. Are you down here in the ships and automobiles, where we're talking about less than two dollars a pound fabricated? Are we talking about the twenty-thousand-dollar-a-pound spacecraft? That's where it's going with the watches — you can have fifty-thousand-dollar watches, you can have ten-dollar watches. What makes something expensive in different industries really depends on the industry. The lower-end industries, it's mostly energy today. The higher end, it's all engineering design and marketing.
What makes beer expensive? It's all the ads on TV and other places. My first big consulting job was in a brewery in upstate New York, and at that time it cost them, as I remember, thirteen cents a gallon to make beer. They were never selling it, even in 1975, for thirteen cents a gallon. There's a big markup. Today beer might cost fifty or sixty cents a gallon to make, but they get a big markup because there's an awful lot of advertising in that.
[Tom produces a pair of surgical scissors.] Here are my surgical scissors. What makes these expensive? They're just stainless steel scissors. You could buy stainless steel scissors for twenty bucks. Twenty-five years ago these cost three hundred; they're probably seven or eight hundred now. Why? Because you're buying about five hundred dollars worth of medical malpractice insurance along with them. The big cost is the medical application it goes into.
So again, the answer is, it depends on the industry. Any other questions?
§2. Chaparral Steel and net-shape casting [07:01]
Last time the theme was that productivity is nearly everything if you want to live well in a nation. In the steel industry, the productivity since the time of Henry Bessemer has all been in melting and casting technology, and we've had tremendous growth over the last 170 years.
Here's the ingot that comes off the continuous caster at Chaparral Steel. Back in the early 1990s they used to get a rectangular ingot, but then they changed their molds, and within about five years they would get something that starts to look like an I-beam. Here's just another picture of a bunch of them — they cut them to length here, here's a strand coming off the continuous caster, in the background some have just been cut.
[Tom shows a continuous-cast near-net-shape section that Gordon Forward gave him about twenty years ago.] Here's the continuous cast part. Now it actually does look like an I-beam — you can see the end of it there.
And someone says, well, is the quality good? It turns out — and these slides are from almost fifteen years ago — when we were doing the evaluation of Chaparral Steel, 1990, one person-hour per ton. Net-shape casting of structural beams drove down the cost by one-third. As Gordon Forward used to say, he didn't care if the Japanese or the Koreans or the Chinese were paying their employees nothing. Everybody said the advantage is low-cost labor. He said it costs about fifty dollars a ton to ship the product across the ocean, and he was here in the market. The United States is sort of the world's market — everybody wants to dump their product here because we consume so much of everything. So he said if he could make his steel within fifty dollars of the cost they were making it for — and that's raw materials and energy cost, excluding labor — then he could beat them in the marketplace no matter what, even if they paid people nothing. He drove the cost down by a third by going to net-shape manufacturing.
There's been a competition — I talked about plastic water bottles and aluminum cans. Well, there's a competition between steel and concrete. Concrete still has steel in it. [Tom passes a galvanized reinforcing bar.] Here's a reinforcing bar if you want to pass it around — it's galvanized. We talked about reinforcing bar, and how they were going slower to roll four times as much: if you went half the speed but rolled four strands rather than one, then you just doubled your productivity on the rolling mill.
Concrete and steel compete for buildings. If you go around the country and look at parking garages — that's the example I like to give — you can look at parking garages and see if they're made out of steel or out of concrete. And if you know anything about the relative cost at different points in time of reinforced concrete versus steel frame structures, you can almost pinpoint within about a one-decade period when that parking garage was built, just on straight economics of which one was more competitive at the time.
Chaparral had a new facility, and they had $450,000 in sales per employee. What does that mean? That's one of the metrics I used to use twenty years ago. If you look at manufacturing companies, you'll find that most of them today have sales per employee in the $150,000 to $200,000 per year range. If you think about it, you'd better sell two hundred thousand dollars worth of product in order to pay employees a hundred thousand dollars a year, because with overhead and everything else it's $130,000, and that only leaves you seventy thousand dollars to buy the materials and other stuff. So when I would look at a company, if I looked at a balance sheet I would say, okay, what are their gross sales, divide by the number of employees, see what the sales are per employee. Back when I made this slide, the typical value was like a hundred thousand dollars per employee. So $450,000 per employee of sales is pretty good — you're way out there in the bell curve, and you can be profitable at that rate.
§3. The Boeing door product center [12:53]
I had a student back in the '90s working at Boeing at the door product center. It's a completely deterministic manufacturing business. It takes three years to deliver an airplane from laying the keel beam to flying it away. The door product center takes about one year to produce the doors that go in the fuselage — you walk through the door when you get on an airplane. They know exactly how many doors they have to produce, because they know they've got these airplanes up in the Everett facility, the fuselages being built, and they know how many doors. There are exit doors, there are cargo doors — doors not much bigger than someone to crawl through, doors someone can walk through, and doors where they can take pallets, like a double garage door, in big 747s. There's all kinds of doors, but they have a door product center south of Seattle, I think it's in Tacoma.
I was getting the tour of the plant, and I asked, well, how much does a door cost? The guy said, I don't know, because it's all internal sales. They only sell to Boeing, and they know exactly how many they have to make that year. I said, well, how many employees do you have? He said 750. I said, how many doors do you make a year? He said about 250. I said, then your average door costs about fifty thousand dollars — because at that time I was figuring Boeing's at maybe $150,000 sales per employee.
He said, well, I don't know if that's true or not. So a little bit later I asked his boss, and I told him how I'd estimated about fifty thousand dollars a door. So that door you walk through is about a fifty-thousand-dollar door. Today it's probably a hundred thousand, because that was almost twenty years ago, but nonetheless, that's one way to estimate things on a very high level, and it's actually pretty accurate. His boss said, yeah, that's probably about right — but they didn't even know. The accountants hadn't figured out for them what it cost to make a door. But I can tell you, they had to put a drill by hand for every rivet hole in that door — a mechanic, and these mechanics are making about forty or fifty bucks an hour at Boeing, today probably sixty or seventy. With Boeing overhead that means like $150 an hour. A mechanic has to take a drill and put it through that hole about four times: rough drill, size, ream — I can't remember all the things they had to do, but for every hole they had to put a different drill through about four times. And there are lots of rivets on the door. So what's expensive? Compared to my door at home, that's an expensive door, but there's a lot more labor that goes into that aircraft door.
§4. Chaparral wide-flange beams and the four-strand rebar mill [16:13]
[Tom hands around a fiber wear part with a neodymium magnet.] This is another thing we can pass around. I talked about transformation of austenitic steel to martensite. This is a fiber wear part — there's a whole story I'll tell on another occasion — but you can feel by taking this neodymium magnet that it's almost non-magnetic down here, and as you get up to the top it gets very magnetic. It's got cracks in it; these are hydrogen cracks, because it transformed to the type of steel that's susceptible to hydrogen cracking.
The other slide is on Chaparral rolled steel beams. They had 41 percent imports to the U.S. prior to their new mill in the mid-'90s. After Chaparral, they dropped the imports of wide-flange beams to the United States to zero, because they could beat them in the marketplace, and they exported twenty percent of the product outside North America. The new wide-flange mill, because you're starting out with net-shape casting, gave 75 percent energy savings.
In-line reheat — it used to be that in the steel mills you'd have this continuous caster and you'd have this red-hot band of steel, and you'd take an oxygen jet and flame-cut it with pure oxygen, because it's already hot enough to oxidize, and then you'd let it sit down there and cool. People said that's silly to waste all that energy, so now they have things that go directly from the casting machine into a reheat furnace to get it uniform in temperature, and then directly into the rolling mill. There's no setting aside and reheating. Twenty-eight minutes total — the processing time goes from days down to twenty-eight minutes.
Because you're doing all this stuff faster, you can roll the steel at lower temperatures, and you get finer grain. All chill cast, higher tensile and toughness. If you're a civil engineer: steel bridges and buildings used to be 36 ksi yield, A36 steel. A higher grade was 50 ksi. With the fine grain size and the good toughness — fine grain size gives you both strength and toughness — they were actually rolling compositions that were a 36 composition but got 50 ksi yield. So they didn't have to inventory the product. By 1995 or 1998, you order A36 from Chaparral, they'd ship you out of this pile. You order 50 ksi, they'd ship you out of the same pile. Equivalent weldability, just as easy to weld as the lower-strength A36. They got 50 ksi as a premium. Actually you couldn't simply ship one as the other, because you have to have certifications for the building people, but in fact it was just as easy to make 50 ksi material as it was to make 36 ksi material, and the toughness was off the charts.
§5. Gordon Forward and the Chaparral culture [19:44]
Student: Who was the guy behind all of this?
Well, that's another story in and of itself. Gordon Forward was an MIT PhD. He's a Canadian — I think he was working for one of the big Canadian steel companies, Stelco or Dofasco. He and other people kind of looked at it — and I told you the story, you could buy scrap iron for a hundred dollars a ton, but it cost you two hundred dollars a ton to make cast iron. If you wanted the electric furnace route as opposed to the basic oxygen furnace to melt your steel — a basic oxygen furnace required 70 percent virgin iron ore, blast-furnace hot metal. Electric furnace, you could take all scrap and melt it, but you'd always have impurities, residual elements, and people thought you could never make anything better than rebar. So back in the early '70s, people like Bethlehem Steel, when I was working for it, considered that garbage — slow, low premium, low markup.
But Gordon Forward, and Ken Iverson, who started a company called Nucor Steel which became the biggest of the mini mills — and it wasn't just one mill. We did build new steel mills after Bethlehem Steel's Burns Harbor plant, which was built in 1965, but by the 1970s it wasn't the steel companies, it was people like Ken Iverson, who started out in a strip mall in Charlotte, North Carolina, to build a steel mill in South Carolina just over the border. And Gordon Forward, who found some investors and built a mini mill in Midlothian, Texas, just outside of Dallas. They built mini mills that had one-twentieth the capacity. They just had a big electric furnace, and at that time electric furnaces could only melt 50-ton heats. Today those same electric furnaces get up to like 165-ton heats, almost as big as the BOF shops that'll cost you a billion dollars. These guys were paying twenty-five million or fifty million dollars for an electric furnace shop. They started water-cooling everything and putting in bigger transformers, and they could melt steel almost as fast as the big guys in their billion-dollar facility, but they were doing it with a 50 or 100 million dollar facility. Tremendous capital cost savings.
But they also had a different philosophy. I wouldn't plan on spending this much time on steel, but — Steve Wheelwright and I went down there. We were doing an evaluation of why good product development and bad product development happens at different companies, and we were doing Chaparral. They had some failures, which I won't go through, but on the first day, they were telling us about one of the failures, and I said, well, who gets the blame? And the people acted like, what do you mean, blame? I didn't realize until the second day: you don't get blamed if you try something different at Chaparral. You get blamed if you don't try something new. So there is blame, but the blame is not for trying something different and failing. It's for not doing something new this year. If you just keep doing the same old thing at Chaparral, you're a loser.
Also, if you go into every one of those shops, on the bulletin board they will have the weekly profit of the company. It's a profit-sharing company. Every employee there could tell you exactly how much money they made the week before, because they posted the weekly profits and the employees knew how many shares they had. So people had an incentive, and they would jury-rig things and figure out cheaper ways to do things. So it wasn't Gordon Forward, it was the culture that Gordon Forward put forward.
At Bethlehem Steel, at the Taj Mahal research labs that we had, they had an executive dining room. I told you about the strawberries the size of — bigger than golf balls — that I had never seen in 1975. You can get them now, but I'd never seen them before. And you weren't allowed in the executive dining room. We had a beautiful dining room that looked over the mountain, but it wasn't as nice as the executive dining room.
There are great stories about these Taj Mahal dining rooms. There's a story about Lee Iacocca wanting to know once, when he was president of Ford, why — one of the Fords, chairman of the company, I don't remember which Ford it was at the time — said he loved hamburgers, and he said, I never get any hamburger as good as the one here at the Ford executive dining room. So Lee Iacocca goes in to see the chef and says, why are your hamburgers so good? The chef says, I'll show you. And he goes and gets a sirloin steak out of the freezer, prime, puts it in the grinder. Hey, Henry Ford can afford it.
At Chaparral, the "executive dining room," as Gordon Forward used to call it, was a folding table, and they would bring in barbecue, and we ate with plastic knives and forks out of foam containers. That was the executive dining room at Chaparral. As opposed to Bethlehem or Ford Motor Company or Kodak or all these other places where they treated the managers like royalty. At Chaparral, everyone was equal. If you go back and read my leadership-management education article at MIT, there's that same theme in there — leadership is treating people equally. My little sound bite is: leaders seek to help others; managers seek to control others.
At Bethlehem Steel, they just wanted to control others. We had a manual that thick that told me who I could write a letter to within the company. I could write one level up or one level down within my own research organization, or I could write at the same level to some person in another organization at Bethlehem Steel, but the manual was that thick on who I could write a letter to. And they had rules about who could have carpeting on the floor, who could have drapes on the windows. One of my colleagues in his office was looking at x-rays, and he took some black cloth and covered up the light from the windows. About a month later some manager came by, saw the cloth on the window, and said, what gives you the right to have drapes? That's what the managers were worried about — this guy had drapes and he was just an engineer. It's a good thing he didn't put a towel down to mop up something on the floor. Anyway, this is one of the reasons I work at a university. Not that we don't have the same type of stupid administrative bureaucracy, but with tenure you can thumb your nose at them. You've got to get tenure first, but I got it, so I can.
Okay, so that's all on steel. Anybody have any questions? I didn't plan on going through that, but hey — anybody have another question?
§6. Bullseye glass and the history of plate glass [27:25]
What I'd planned on doing today is take another industry and just show you that I'm not negative about everything. I want to talk about glass technology. I told you all the bad things about glass — it's brittle, it's not a structural material — but in fact it is a structural material, and it has a history. If you go to the web you can read about Sleepy Hollow Glass Company; they make nice artistic glass. They will actually sell you what's called bullseye glass. Anybody know what bullseye glass is? Yeah — that little thing in the corner, that's right. There's a history of bullseye glass in the United States. Remember "no taxation without representation" type stuff, and the things the British were trying to tax us.
Glass used to be a very labor-intensive industry. The way they made glass at least into the 18th century in England was, just like we do down in the glass lab: you take a blowpipe — just a steel tube — take some glass out of the furnace, stick it on there, get it to the right temperature, and you blow through the pipe. You blow it up and make a little glass beaker, you can shape it, you can form it. Eventually you can cut a hole in it and keep spinning it in the furnace, and you can make a disc of glass. If you're clever about it, you can make relatively flat squares and rectangles of glass — not very big.
At the center of this disc you're going to end up with something that's got all kinds of distortion. You actually break it off the blowpipe by taking a little file, putting a file mark on it, then hitting it quickly so it breaks off — and you end up with a piece of bullseye glass. People would put this over the lintels of the doors. The reason they would do it is, glass was expensive. They didn't have electric lights back in the 18th century, and people wanted some light in their house. So over the door they would put glass to let the light in. But bullseye glass didn't have a tax, because it was considered scrap. It distorted everything, but what do you care — you just want light to come through, you're not trying to look out the window. Looking out the window required one of the pieces of flat glass that was taxed, and people didn't put a lot of windows in their houses because they were drafty and they cost a lot of money.
Student: They cast the plate glass —
I'm going to talk about that. So this is the way they made glass — they just blow it, make a big circle, cut rectangles out of the circle, throw away the center, so the cheapest stuff was the bullseye glass. However, if you go today to Sleepy Hollow Glass, they'll sell you one of these little bullseye things, six inches square, for fifty-five dollars, or eight inches square for like ninety dollars. That's pretty pricey, but now it's art. People put this over the doors of their home still, so they can look like they're in an 80- or 200-year-old home. Be careful of the sharp edges — that's why you keep it up top.
Plate glass manufacturing, on the other hand — which is what you were asking about — was invented about 1688 by Abraham Thévart, and that became the Saint-Gobain company. Anybody ever heard of Saint-Gobain? Saint-Gobain's got a lot of contacts with this department, in part because one of my former doctoral students, Kapoor, was head of their research center in Northborough, and he started the MadMec and all this other stuff — they fund those things. But Saint-Gobain, in 1688, Abraham Thévart learned to pour glass onto a metal table, spreading it with rollers, annealing, grinding and polishing. That was the beginning of Saint-Gobain.
If you want to see a picture, probably from the 19th century — here's a picture of making plate glass. You start with a metal table — slightly inclined, that was an invention in the 1800s, to use an inclined table — and here's your steel roll, and they take this molten stuff and roll it out flat. The problem is, there are inhomogeneities in the glass, and that type of plate glass, unless you ground it and polished it, and even then, still had a lot of distortion. This is an old window in Jena, Germany. Here's a piece of modern glass — which is not plate glass, it's actually what we call float glass, I'm going to talk about that — and you can see the reflection of a tree in that glass. Here is some of the old plate glass in the rest of this window, and you can sort of see the black outlines of the tree, but it's been distorted. Someone broke the glass up here and they had to replace it.
§7. Pittsburgh Plate Glass, Libbey-Owens-Ford, and the Pilkington float process [33:22]
Steel and glass don't like each other. As it cools down, thermal expansion causes them to separate at the interface. It's a problem to try to bond glass to steel — we can do it, but we have to put an inner layer like titanium or something reactive in there.
So — continuous ribbon production. Plate glass they would roll out on these big plates. A company like Pittsburgh Plate Glass — you ever heard of PPG? — as I'm told, around the turn of the 20th century, they had a facility in Pittsburgh that was a mile-long building. Nice and straight: at one end they would cast the plate glass, and they would move it straight, no turning, because glass is brittle. They would grind it and polish it to where it was fairly smooth. That's what we call plate glass, and typically it might be a half an inch thick or even thicker. We call it plate glass because it's thick — if you roll it out and make it too thin, you have these great big sheets of it and you're going to break them. So it tends to be thicker.
You ever heard of Libbey-Owens-Ford? Glass used to have a little LOF on most of the Ford cars. Continuous ribbon production of glass was invented at the Ford Motor Company in 1920, and by 1926 the production was achieved by flowing glass from a tank furnace over a weir. A weir is just a little ceramic brickwork — and just like going over a waterfall, this is a controlled waterfall. You just have this thick viscous glass going over the weir, you pull it over that weir onto a table, let it cool, and then you do the plate glass process. Why did Henry Ford need that? Because he was making lots of windshields for Model Ts and Model As. Pilkington Brothers put that into production, and they developed a continuous grinding and polishing process by 1925.
By probably the 1940s, Pilkington developed their own process called the Pilkington process, which is the float glass process. You have a blending and melting furnace — this is called the batch house — and then the furnace in which you melt it. Let me tell you, I have been through a float glass plant in Ohio, and you're walking by this furnace at 1550°C that's about half the size of a football field, and it's hot. A lot of radiant heat even though it's insulated and everything. They pull it over the weir and then they float it on a bath of molten tin. Not a very deep bath — molten tin probably only a couple of inches deep, but it's probably 15 feet wide. Now you can pull off stuff that's an eighth of an inch thick, three-sixteenths, a quarter of an inch thick. You take that and you anneal it, because glass can have serious problems if it's got residual stresses in it, and you cut it to size and ship it. It's not a mile-long building — might be a couple hundred yards, but it's not a mile.
Student: So do you still have any inhomogeneity in this process?
The inhomogeneities originally came from taking the sand and the limestone and all these different components and putting them into a furnace. It was in the 1800s that they learned to stir the bath. Glass is very viscous, and the reason — that Jena window, a lot of those inhomogeneities were not surface irregularities, they were chemical inhomogeneity. That's why they had to go to Jena, Germany to find that window — that's 200-year-old technology. In fact, mechanical stirring of molten glass was not introduced until 1798 by Guinand, and was perfected in 1805.
France kept this as a trade secret. The British didn't get it until the late 1700s or early 1800s. So for the first hundred years France sort of controlled the world's glass manufacturing, and it was mostly for mirrors. Saint-Gobain made all the windows and mirrors in the Versailles Palace. If you look up Saint-Gobain you'll find they're very proud of the fact that they go back to Versailles. They had the technology to make big glass plates, the British didn't, for 100 years. But the British learned to stir the bath and get more uniform mixtures. Then Pittsburgh Plate Glass comes along — mile-long building, very expensive, lots of hand labor, grinding and polishing. And then Pilkington comes along, and the night after World War II comes up with the float glass process. That Jena window is actually plate glass, probably — if it's the original from this building. If it's been replaced, which is a fair chance, it's going to be float glass. You don't usually see a half-inch-thick float glass. So plate glass processing is still around.
§8. Glass compositions and the LCD stack [39:51]
In the float glass process you want to keep it relatively thin, to move stuff across that bath fairly quickly. Most of these glasses are what we call soda-lime glasses, meaning they're made with sodium oxide, calcium oxide, and of course all glass has SiO₂ in it. Some has some Al₂O₃, but you see calcium, limestone, and sodium — 74, 11, and 12. Most of this is soda-lime glass, and that's float glass for windows. If you go to borosilicate glasses, you can drop the temperature considerably by adding borax, six percent. It has more aluminum in it, more MgO in it, and this could be made into textile fibers — fiberglass. Why do you like to use borosilicate? It drops the melting temperature several hundred degrees, and therefore spinning the fibers is a lot easier at lower temperatures.
If I look at other glass compositions — TV panel glass: 72 percent silica, so high silica, some barium, some strontium, some potassium. Some of these things start adding the right transmissivity. You may not notice the color of that glass, but if you did a spectral analysis you'd find that it's not letting all the wavelengths through equally. On a TV you'd like to have something a little better. And fiber wool borosilicate glass — so we're still making fibers, typical. There's lots of different glass compositions. Structural glasses are tens of millions of tons a year — not as big as aluminum, but we use glass in a number of applications, in buildings like the Versailles Palace, or you look at the PPG building, it has lots of windows. Probably plate glass-type windows, but PPG is doing float glass now.
Student: Is there a Fraunhofer glass company?
I thought Fraunhofer was a bunch of physicists who were looking at Fraunhofer diffraction and things like that. There are some great glassworks in Europe — Swarovski, and Schott glass, S-C-H-O-T-T. Fraunhofer was looking at optics, and improving optics. I'm sort of talking about structural glass. There is a whole different thing of what they call crown glass, which is making optical glass. Who was the guy who made the first eyeglasses? It was like the 1600s, I can't remember — was it Leeuwenhoek? Leeuwenhoek, okay. So there's a whole group of people who were making small little discs for eyeglasses, polishing them, and they were mixing and getting homogeneity. They couldn't tolerate any inhomogeneity that would change the index of refraction, because then you'd be looking through a Coke bottle — literally — when it wasn't a prescription Coke bottle.
But there are some very high-value-added glasses. Corning used to have very close ties to this department, but has even closer ties to Alfred University, which is the New York State College of Ceramics — the best glass technology program. Corning had a whole business in television sets. The old cathode ray tubes — making this blown glass with a slightly curved surface into a steel mold was a technology that Corning perfected, the technology of blowing that glass into a mold and doing it cheaper, more uniformly, than anyone else in the world. Corning is sort of a trade secret company — they don't patent a lot, they just make it. Let me tell you, I can't think of any other company I've ever been in where it was harder to get through the door than the security at Corning. Not only in terms of signing non-disclosures, but you can't bring a camera. If you needed pictures, they would take the pictures, they would review the pictures, and then they would send you the pictures afterwards.
Corning had this huge business in the 1970s in two products. One was light bulbs — just a little blown glass, incandescent light bulbs. The other was TV tubes. They were losing the TV tube business when plasma displays started coming around in the '80s, then LEDs and liquid crystal displays. A liquid crystal display is actually a composite of glass, and it's got a number of layers. This is right off Wikipedia. Layer one is not glass, it's actually a polarizing filter film — plastic — to polarize the light because this is going to be a liquid crystal display. Then you've got a glass substrate, which is layer two, which has indium tin oxide on it. Why indium tin oxide? It's conductive and transparent — a very interesting set of properties. Most things that are conductive will absorb electrons and therefore no light — they're impervious to light, like metals. You've never seen — only on Star Trek do we have transparent aluminum. There are no transparent metals, almost by definition. So they have glass here, then another layer, which is the liquid crystal — twisted nematic liquid crystal — then another glass substrate, number four. Two and four are glass substrates, five is another polarizing plastic filter, and six is a reflective back surface, or a light surface if you're doing a transmissive white backlit display.
§9. Corning ultra-thin glass and water-vapor corrosion [46:38]
[Tom holds up a sheet of ultra-thin Corning glass.] If you go to the Corning website, you'll find that they will make glass for you that is 0.3 millimeters — twelve thousandths of an inch — thick, and you can buy it from them in sizes of one-and-a-half meters by 1.8 meters. So if you want to have a big TV screen, but it's only twelve thousandths thick. Be careful how you hold this — because it's brittle, believe me, it's brittle.
On the Corning website for slim glass and the product information sheet, this is now a big business and proprietary for them. So it's an alkaline-earth boroaluminosilicate, just like the others, except it's the way you make it. It's a fusion-drawn-down sheet. Even the melting furnaces are sort of interesting, because you put the cullet in and you melt it and you hold it, and then you draw it out of the furnace. But the parts that have to be ductile — like the tubes and the nozzle — they're all made out of platinum. A typical melting furnace at Corning could have 10 million dollars worth of platinum in it.
Student: Where did the microcracks in a very smooth glass come from?
I'm going to talk about that — hold on. The next sheet of this goes through and gives you properties: mechanical, density, modulus, shear modulus, Poisson's ratio, hardness. Didn't give you tensile strength, did it? Didn't give you bending strength. Hey, if it's only twelve thousandths thick, it does have a little bending strength — just don't twist it too much. But thermal expansion, thermal conductivity, optical loss, and all this other stuff. Now to answer your question — then comes chemical resistance. Everybody knows you use glass, glassware, for the chemistry lab. Weathering is defined as corrosion. Glass doesn't grow, does it? Corrosion by atmospheric-borne gases and vapor such as water and carbon dioxide. Glass is corroded by water vapor. You didn't know that, did you?
Glasses rated one will almost never show weathering effects. Those rated two will occasionally be troublesome, and those rated three require more careful consideration of durability. How do we prove this? Well, if I pull a glass fiber — a brand-new glass fiber, aluminum borosilicate glass — and I'm going to make it into fiberglass with the best strength, you can buy this from Corning, it's called E-glass, but they have different letters and different compositions. I had better coat that with the epoxy within 24 hours, because the humidity in the air is going to degrade the strength. When I pull it and measure the strength right away as a fresh new surface of that fiber, it may have six or seven hundred thousand pounds per square inch tensile strength. An hour later it may be five hundred thousand. The next day, four hundred thousand. A month later it's going to be — it is a piece of glass — but it'll be so brittle that you want to use it as a strengthening element in fiberglass. So when you make fiberglass, you epoxy-coat it real quick to protect it from the moisture in the air. On the ligands on the end of this you've got these silica tetrahedra, and then you've got sodium, and the sodium wants to form sodium hydroxides, it'll form silicon hydroxides. Those things will actually form little etched pits on the glass, microscopic — you can't see them.
But if you go through the fracture toughness of this glass, go back to the Ashby plot and look where the glasses are in terms of fracture toughness and calculate the minimum flaw size — it's like 10⁻⁴ millimeters. We're talking angstroms of critical flaw size, or tens of angstroms. You don't have to have very much corrosion. Just the humidity in there. Now, I've never seen this, but my thesis advisor, Bob Rose, used to do it. He still lives up on the North Shore. Sylvania up in Danvers has a glass manufacturing plant — a light bulb manufacturing plant, part of their lighting division — they used to make light bulbs, I'm sure under license from Corning technology. Bob would stop by before he would lecture 3.091 and pick up some freshly made light bulbs, bring them into class, and an hour later he would throw them against the wall, and they'd just bounce off the wall. Try that the next day, and they shatter. He didn't do it the next day, but freshly made glass is extremely strong, because it has no surface flaws — it hasn't been corroded by the moisture in the air yet. Does that answer your question?
So glass is a structural material if you protect it from the corrosion of moisture in the air. But most places have a lot of moist air. Not Arizona — except when it rains, they have monsoons.
§10. Optical fiber drawing and wrap-up [52:28]
Student: [inaudible question about optical fibers]
The next lecture I give is on Tuesday, and I didn't get all the way through the glass stuff I had thought about this morning. If you want to go to YouTube or Corning, they have some videos on making optical glass fibers. I decided optical glass fibers aren't a structural material, but it's a wonderful story. They take silicon tetrachloride gas, because it's so pure, and they flame-spray a starting rod, and they build up something that's like an inch or two in diameter. They grade the composition as they do it, so they get a graded index of refraction. Then they draw it — it's called drawing the glass — they take this precursor and they draw miles and miles of optical fiber.
The water content — they actually have a drying step before they do everything. They have to dry it in an oven to get the moisture out. The glass can pick up moisture — sodium hydroxide, silicon hydroxide, potassium hydroxide — you can get OH ions in there and it will screw up some of the optical qualities of optical glass.
So if you want to learn more about glass, it's a very interesting material. Corning really owns the technology in the world. They hold it as a trade secret. Although they've hired a lot of ceramists out of MIT who rose very high in the corporation, it's a very inbred corporation, and after they spend enough time in Corning, New York, they usually prefer to go to Alfred University. We don't even have a glass person in the department anymore. The last real glass person we had was Don Uhlmann; he went to Arizona to retire. So, thanks. Simone will be here the next two days, I'll be here on Tuesday.