Arthur Squires: And I probably did not appreciate, during the war itself, the extent to which this was such a remarkable effort. Kellex – I am sure some of this you have already heard. Kellex was put together by M.W. Kellogg Company pretty much on a command basis. They just went to the top people all over the industry and got the top instrument man, the top man in the power field, and top people in compressors and all the various phases of the project.
Stephane Groueff: Who contacted you first, Percival Keith or Manson Benedict?
Squires: No, I was still in graduate school working for Professor John Gambell Kirkwood.
Groueff: How old were you at that time?
Squires: Let me see. This was 1942. So I was 26.
Groueff: I see, so you were quite a young.
Squires: Oh, no, no, no, I was just finishing up my graduate school at Cornell. He came to me in March and said they had this—
Groueff: Who was that?
Squires: Manson Benedict.
Groueff: Oh, Manson Benedict.
Squires: He had known Kirkwood. They had both been at Massachusetts Institute of Technology at various times. At some time or other, their careers had overlapped. And Manson asked Professor Kirkwood if he did not have a student that he could recommend. And Kirkwood put him in touch with me. That was pretty much it. I was available for a war job.
Groueff: So who contacted you personally, Kirkwood or Benedict?
Squires: I think Kirkwood told me that Benedict was coming. I cannot imagine that Kirkwood knew what this was about, other than in a general way.
Groueff: And you did not know?
Squires: No. I speculated. I knew it was the M. W. Kellogg Company. I speculated that it was something to do with aviation, gasoline, or something like that. But I did not really speculate very much because I was too busy finishing up.
Groueff: But you knew that it was something important and secret?
Squires: Yes. I did not have the foggiest notion how important or even how secret it would be. After I signed in, June 15, 1942, Manson and I walked across the cinders of the yard there of the Jersey City M. W. Kellogg plant, a very grubby, sort of an industrial atmosphere. They had a works manufacturing high-pressure piping. And he told me what it was. Of course, I had read all about the uranium discoveries, and the whole scientific world knew that something like this was possible. But all talk of it had ceased in about 1940. And so he told me this was what it was.
Groueff: When would that have been, 1942?
Squires: I went to work for the M. W. Kellogg Company June 15, 1942, yeah.
Groueff: And you started working in New Jersey first?
Squires: Yes, yes. I vividly recall. He showed me a list. I think it had about eight people that I was allowed to talk to. And it was called “Project X” then. This was before the Kellex Corporation was even formed. Kellogg – you probably have some of this kind of story before. Kellogg was approached by Columbia and the NSRD [National Security Research Division] or whoever it was at that time to study the engineering possibilities of a large-scale plant of this kind.
Project X was put together under Percival Keith in February or so, I think, of ’42 with Manson Benedict as sort of certainly the chief scientific brain and chief process brain thinker in charge of this really quite small team. He had a handful of mechanical engineers and maybe one or two process men and a secretary. And that was about it. The activity was quite small. And at that stage at Kellogg, completely paper. But we were located over in Jersey City.
Groueff: Jersey City, not in Woolworth?
Squires: No, no, this was long before the Woolworth building. Because all through the summer of ’42, we gradually picked up people and I imagine – I am guessing. I imagine that on December 1, 1942, there might have been 40 or 50 of us, somewhere in that neighborhood. But shortly after, I vividly recall it. Shortly after December 1 and, of course, hindsight, one correlates this with the December 2 experiment at Stagg Field in Chicago, we got a green light.
Groueff: That was ’43?
Squires: No, it was ’42. I would say about December 15 of ’42. We got the green light that a plant would be built. And things started to happen very fast. The Kellex Corporation was organized as a subsidiary to handle this. Keith cut himself off entirely from his other activities with the Kellogg Company. And we started to hire people like crazy. We moved to the Woolworth Building in February of ’43. And the actual authorization to procure—start making deals with manufacturers for equipment—I think came through at about February or March of ’43.
We had to do the process design for this plant between about December 15 and March 1.
Groueff: In four months?
Squires: Oh, less than that.
Groueff: Less than that.
Groueff: But until December, you were still working on the design for the plant?
Squires: Well you see what had happened through the summer and fall of ’42, we were playing. Playing is maybe not quite the right word. But anything like this that gets built, you make 10, 20, 30 designs, then tear them up.
Squires: Because you have to explore all the different – or at least as many as you have time for – as many of the various possible ways of putting this equipment together. At the outset, you do not know what pressures to work at and what temperatures, and whether to work with a lot of little pieces of equipment, a few big ones, or all the various – are you building a Cadillac or are you building a Volkswagen? It is that kind of a question. And just simply been given a requirement that we produce so and so much of a particular material at such and such a rate per day, then the designer maybe will make 10 or 20 designs, which on paper will do this, before he finally finds the one that he likes the best.
There is a horrible word that is used with describing all of these activities. It is called “optimizing.” Yeah, and I do not like the word because it implies it’s the best. It implies, however, an omniscience that the design engineer never has. There is a tremendous amount of judgment that goes into selecting the final design as well as any optimization calculations based on the, let us say, horsepower or cost of equipment or various actual physical things. You might design a plant. Which according to some – this is, I think, a fault with a lot of the work that is being done today with computers in design work. You can design a plant according to a very elaborate set of rules, which then will come out okay; this is the cheapest plant according to these rules.
Squires: But you take one look at it and you know it is silly. You just would not build it because maybe it has complications in it or there has not been a leavening of judgment.
Groueff: So it is good only according to one set of rules?
Squires: To the rules that you started out with.
Squires: But by the time you are through the study, you should know enough by now that you should go back. And you realize that your rules may be were not the best.
Groueff: I see.
Squires: So there is a feedback in any of these design studies. And after you have done—
Groueff: That is what you did in the summer of—
Squires: This is what we were doing all through the summer and fall of ’42.
Groueff: How did you work in this period with material? I mean, you and Benedict in a room, or each one making designs? Or how many people were involved? What was the atmosphere of this kind of work?
Squires: It was done under tremendous pressure. It was done with the idea that each study – at the time we were making the study, each study was it.
Groueff: I see.
Squires: Now this, I am sure, is partly a hoax that was perpetrated upon us by Mr. Keith. This is the kind of a man he is and the kind of a driver he is. At that time, he was in his prime. He was born in Austin. He was 42 years old. And he was a driver. He always attempted to create a sense of urgency. He would call us up at ten o’clock in the morning and want an impossible amount of work done and on his desk. Manson Benedict was supposed to be there with his briefcase by three o’clock that afternoon. We would just sit there aghast. It is just not possible. But you would do something.
Squires: You would do the best you could. So you asked if we were working individually. Of course, everybody works at his own pace and with his own paper and pencil. So in that sense, it was individual. But it was very much of a group effort. And as I say, the group got larger.
I recall one period. It must have been roughly September when I must have had fifteen or twenty calculators [people] that I was sort of directly supervising. And I would calculate some myself. But mostly, I was running around gathering up papers from everybody and putting all of this together into some kind of sensible form.
Groueff: Were they mostly young men?
Squires: Yeah, it is mostly—
Groueff: Graduate students.
Squires: No, mostly I would say just out of chemical engineering degree, chemists even, or mathematicians.
Groueff: Most of them undergrads?
Squires: Oh, yes. Most of these were just younger than I was. I was an old man in this.
Groueff: But why were you chosen out of everyone?
Squires: As much as anything, because I was there first. Why I was chosen?
Squires: Well, I was a physical chemist. I was trained as a physical chemist. Well actually, I leaned very heavily on the math and physics. And this was the kind of man that Manson conceived that he needed. He probably felt that he needed this type of man a little more than it actually turned out that he did need, because I very quickly became an engineer. I cannot say that the work I did in graduate school was of enormous help to me during these years. I became a new fellow as this was going along.
Now, obviously, the discipline of graduate school made it possible for me to do this faster than someone who had been on the streets.
Groueff: But was it because of some pieces of your role?
Squires: No, no connection.
Groueff: You had no connection to diffusion, particularly?
Squires: No, no, absolutely not.
Groueff: Like Manson Benedict, he was involved with the diffusion of the foreign and—
Squires: No, no.
Squires: No, no, he was also a physical chemist. But he had made this shift to becoming a chemical engineer earlier. I started to say, he was a pupil of George Skatchard. But I am not positive of that.
Groueff: No, he told me about—
Squires: I know he did some work with Skatchard. He probably told you.
Groueff: I do not remember this name. But he was a Massachusetts Institute of Technology man.
Squires: Yes, he was at Massachusetts Institute of Technology. And Skatchard is at Massachusetts Institute of Technology. And I am sure that Manson did some work with Skatchard perhaps as a post-doctoral thing. But Manson was a brilliant physical chemist and thermo-dynamicist. And Keith hired him in about ’37 primarily for a big API [American Petroleum Institute] project on the physical properties of hydrocarbons and hydrocarbon mixtures.
Up until that time, there was a great amount of empiricism in distillation/separations of complicated hydrocarbon mixtures, which has to be done in petroleum refineries. And petroleum refineries around 1930 were put up in units of maybe one thousand, fifteen hundred a day. But toward the end of the ‘30s, they were going up in units of twenty thousand-odd barrels a day, maybe thirty. Now, they are going 100 thousand barrels a day.
Well as this happens, you are less and less willing to rely on empiricism. I think the petroleum industry really was first in recognizing the enormous power of a really scientific approach to the design of some of this large-scale processing equipment. The chemical industry, dealing with much smaller quantities of material, were just generations behind them at the end of the war. And the whole profession of chemical engineering grew out of the petroleum industry.
Groueff: But I think Keith himself had some experience.
Squires: Keith was one of the first MIT. Right. He was in the first graduating class that called themselves the chemical engineers.
Groueff: He talks about a job.
Squires: That is right. Well this whole profession got its real start in the petroleum industry. And it has constantly been upgraded by pulling in people like Manson, who was trained as a physical chemist and brought to the profession greater scientific exactitude. In the 1920s, chemical engineering was very empirical in many, many respects. It still is. There are some things that will always be empirical.
But towards the middle of the ‘30s, there was felt this great need to really understand the thermodynamics of hydrocarbon mixtures on a thoroughly sound scientific basis. And Manson was hired to do this. And out of it came what is known as the Benedict-Webb-Rubin equation of state, which is still the best—
Groueff: He is now one of the world’s authorities on this problem.
Squires: Yes, although he has left it long since behind him. He has not worked on problems like this for at least twenty-odd years. But this is how he got into engineering. He was not trained as an engineer—just exactly my experience. And this is, I suppose of all of the engineering professions, this is probably truer with chemical engineering and perhaps electrical engineering certainly than civil or mechanical.
Groueff: Who are the other members of the – the top members of the team for the design? Benedict only? There were the mathematicians, Montroll. Montroll was on the—
Squires: Yes, Elliot Montroll and Joe Lehner were probably the two chief mathematicians. And in the process area, I think Charlie King was number one. He brought him over, Charlie, having had excellent process experience with petroleum refining type of installations.
Groueff: He was a scientist?
Squires: No, he was trained as a chemical engineer and grew up more closely to the chemical engineering type of problems.
Groueff: Actually, there were two scientists. You and Benedict developed more and more engineering experience. And the engineer, King—
Squires: Had to learn from us, because the gaseous diffusion plant is a peculiar thing. I have never really come across anything quite like it since in chemical engineering. It led itself to relatively precise analysis in terms of quite sophisticated mathematics, differential equations and so on, partial differential equations with time. And most processes, you like to be able to do this. But your analysis is much, much more approximate.
Squires: So we were able to predict things that the plant would do from our equations much more precisely than one would expect to be able to predict the behavior of an ordinary plant when one had so limited experience to go on. I do not want to give you a false impression, though. Our process design, which was finished by about March 1, 1943, was really out in the wild blue. At that time, the largest piece of barrier that anybody had ever made and tested with uranium hexafluoride was not much bigger than a dime.
Groueff: That was Columbia University.
Squires: That is when we worked down at the Columbia University. It was a very nice man who did a lot of this early separation work with uranium hexafluoride. And it was a bitch. It was the kind of thing he had to just struggle, and struggle, and struggle. And then after a couple of months, he had one answer. It was that kind of thing. What was his name?
Groueff: Was he a Columbia man?
Squires: I do not know his ridge origins. But he was a physicist. And he was with the SAM [Substitute Alloy Materials] group.
Groueff: E.T. Booth or A. von Grosse?
Squires: No, a younger man. It may come to me.
The plant design, in order to have the group optimize it, we needed to know how the barrier would perform as a function of pressure level and also how it would function as a function of back pressure. The material goes across the barrier from a high side pressure to a low side pressure. And you need to know how the separation varies as a function of both of those variables. And we did not know. There was no theory to guide us. There was practically no data, and no barrier that was commercially usable. It was not developed yet.
Groueff: No pumps.
Squires: Oh, no pumps. But that you had confidence in. One had real confidence that the pumps could be designed and built to do whatever we, the process, required.
Groueff: You were not quite sure.
Squires: But barrier, this was a gamble. But there was also a big element of gamble. There were two gambles. It was just the basic gamble of, could a barrier be built? But then second: after it was built, maybe it would be a very good barrier, but would it make any sense in relation in relation to our process design?
Squires: It might be a perfectly good barrier for another plant.
Groueff: Another plant, yeah.
Squires. And another design. But could it be made to function in this design? So actually, having made our design, we put sharp constraints upon what the barrier people could do and have it be called satisfactory. It not only had to be mechanically strong, so that it could be formed in the shape of tubes. And that was a violently opposed decision. The people up in Columbia were just horrorstruck at this decision because at that time, the only barrier they had, if you bent it about like that, it broke.
Groueff: A flat barrier?
Squires: It was flat. And that decision was made in, oh, I suppose mid-February of ’42 – ’43. And I vividly remember the long faces coming out of that meeting. Keith made the decision on a very arbitrary basis that—he just did not know how to build thousands of very large pieces of industrial equipment with the barriers that supported these flat sheets.
At the time, I was horrified. But with the kind of experience in chemical engineering and mechanicals that I had picked up by now, I see the reason for his decision. It was the right one. The plant would never have gotten off the ground for a long list of reasons, if the mechanical people, if the people at Chrysler had been saddled with the job of turning out a piece of equipment to accommodate the flat barrier.
So when the barrier people were told they had to come up with round ones, this put a constraint on what they could do. But we also put a process constraint by saying that we were going to work between this pressure and that pressure and that we wanted a porosity of this. And so we were only going to give them so and so many square feet of barrier in a particular piece of equipment. And it had to hand so and so much gas between these two pressures.
So suddenly, their job was narrowed down to meeting a specification that had been set with really very little to go on. I think there was an element of luck. We had luck – yeah, yeah. We were lucky all right.
Groueff: This is a gamble, yeah.
Squires: Yeah, but we were lucky on that design. I do not want to take credit away from Manson because he was certainly 98-99 percent responsible for the crucial decisions, the pressures, the flows, the porosities. The final judgments on all these things usually came right from him. And he was the one who had to look off in the corner of the room and say, “That is it. Now we have to live with it.” I do not want to take credit away from that. But gosh, it could have been wrong.
Groueff: And you all knew it. And everybody knew it, actually.
Squires: Yeah, oh yes, everyone knew that this was a gamble, from this point on.
Groueff: And actually, I hear that Professor Harold Urey said formally that it was wrong, the whole diffusion thing. And Urey was very pessimistic about the whole project.
Squires: Well that more or less checked, because we saw very little of Urey all through ’43 and in ’44. We saw a lot of the SAM people who continued to collaborate on this project. But Urey personally was very little involved. And a number of his people, such as Karl Cohen, I can understand their point of view. I understood it probably better having come so briefly from an academic atmosphere myself.
They sort of felt that we were taking their baby and running away with it. But a mathematician like Karl Cohen did not have the breadth of knowledge that was needed to specify the kinds of things, which Manson had to specify, in order to get this process design frozen. He also was not used to working at the pace.
Groueff: And no industrial experience.
Squires: —and with no industrial experience. And this was the case where if the thing was going to be finished on the time schedule that had been set for us, and it was. The time schedules were met right down the line.
Groueff: Do you think that with another man than Keith and involving General Groves—the two of them were kind of slave drivers. Do you think the job could have been done, or done in the same period of time?
Squires: That is a hard question to answer with anything very positive. But I think it is unlikely.
Squires: The project badly needed a man like Keith who could stick his neck way out in February of ’43 and say, “We are going to use round barrier and to hell with all this flat stuff. Stop making it. Stop working on it.”
This is what is wrong with our missile program today. I read the headlines in the papers. It is less wrong – less true today than it was four or five years ago. But four or five years ago, it was so transparently clear. Now, I have sat in on meetings like this, too. I have done work with government people in the last ten or fifteen years since the war just on a very, very minor scale.
But you get these big meetings with fifteen people there and you get three opinions as to what should be done. And the resolution of the dilemma is to do all three, to set up one group here, another group over there, and a third group somewhere else. And all three get done badly because each one is saying, “Well if we flop, there are these other two.”
Groueff: Yes, yes.
Squires: There is that insurance policy. And I am sure that in many of our early rockets, even the ones that we are sending up now, have safety factors built into them, safety devices. And then safety devices on the safety devices sparing this, that, and the other, having two of everything or three of the most crucial ones. And these complications lead to failures. The Russian rockets bear the relationship to our rockets that a Model T Ford does to a racing car.
Groueff: Yeah, they’re simpler. But either they work or do not.
Squires: That is right.
Groueff: We do not if they really do not work.
Squires: And we do not know it if they do not work.
Groueff: And they do not care.
Squires: And they do not care, and I am sure the cheapest way to have rockets, because after all, these are expendable items. I am not saying that maybe it is the best way to have rockets if you are going to send men to the moon. But if you just want to hurl something, the Model T approach would be the one I would take. And this is the approach that Keith would take on the gaseous diffusion plant time and again, rather than complicating the thing by throwing in bells on top of suspenders. No, it had to be this way. And it had to be right.
Groueff: And he was the kind of man to take risks.
Squires: To take the risks.
Groueff: Ready to be blamed now.
Squires: Ready to be blamed afterward if this was the wrong decision and if the whole thing went wrong. Yeah, there was a big, big element of luck. I am sure. But the thing would never have – well, I am now being categorical and I should not.
Having had this experience, all I have to do is read the headlines and I know what is wrong with our missile program or what has been wrong at times. I think there is still a large element of this in our missile program, tremendous waste, because one has different approaches being tried simultaneously, I think sometimes in the same shoot. And this just is not good engineering. And it is largely because you do not have a technical man. You need both a technical man and a driver. And gosh, I know so few. And I think they are getting scarcer because there is much more work done by committee.
Squires: The big management approach, the big companies are run by committees. The one-man company is a thing of the past. This is really harping back to Henry Ford’s time.
Squires: When he was the president of the Ford Motor Company in 1920 or ’21, whenever, this was true. He could walk into the shop himself and change it, and that was it.
Groueff: It was lucky that several such men were found and worked together during the Manhattan Project.
Squires: Yeah. I do not know to what extent you come across this. And I am not sure I would want to be quoted on some of this. But I came in contact with the people over at Y-12 in about September of ’44.
Groueff: The electromagnetic—
Squires: The electromagnetic project. This project lacked this.
Groueff: So they did not have a boss there like Ernest Lawrence?
Squires: Lawrence did not have it. I never met the man. I have never talked to him. I know nothing about him. But all I had to do is look at that plant, walk through it, and hear some of the problems that they had had. And they kept shaking their heads.
We kept telling them, “Oh, we are going to have a building. And we can send you material on such and such a date, and this and so on such and such a date, such and so on such and such a date.”
They would write it all down. And they would shake their heads. And sometimes they would even say, “Look, you are going to have troubles. We cannot count on these deliveries.” We are always ahead of our dates.
Squires: Yeah, always ahead. I wanted to remind me to tell you why we were ahead, because we got a big surprise out of our plant. The electromagnetic people were plagued with mechanical problems. Maybe I am not being quite fair because I do not, obviously, command this technology. I do not know the problems that they faced at the outset. Again, on absolutely unassailable grounds. I had the impression that they just took a piece of laboratory apparatus and said, “Okay, let us make it yay much bigger,” and did not really call in really first-class electrical engineers. But it was much more a scientist’s show. It was as if the Columbia people had tried to build the gaseous diffusion plant, and they would have had the same kind of problems and the same kind of experience that the white glove people had.
Groueff: I think maybe it was that Keith was an engineer and not a theoretical scientist.
Squires: Yeah, it would just have been a very different kind of a show.
Groueff: I was told also that the barrier, once invented, when they went to production, acres of the stuff was done not like a small laboratory experiment in labs so many times, but it was done like a mass production.
Squires: Oh yes, at Houdaille-Hershey.
Groueff: One by one by workmen who were making every piece of barrier, one by one, like you would do it in a workshop, and not transpose a laboratory experiment.
Squires: This was true, I think, all through the project. There was a mass production approach in the thinking right from the start. The assembly line. I did not see the barriers being made. But I saw the assembly line for the turbines. And I saw the assembly lines for the barriers. The turbines were made by Allis-Chalmers in Milwaukee, and the assembly line for the barrier assemblies at Chrysler.
And here, you could just see hundreds of these things lined up moving along just as if they were automobiles or anything else. People who know how to set up an assembly line and not only do it efficiently and at low cost, but do it at all in a time span. Actually, the coin was time rather than money. We spent 500 million dollars on K-25 between March of ’43 and August of ’45, and that is good going. That is a lot of money in that space of time.
But the thing that was, I think, even more spectacular, our first building operated February – I think February 20, 1945. I am not absolutely sure of that date. But it was within a week of that. It might have been a few days earlier.
Early in March, like within two weeks, we had an authorization to spend 80 million more on what was called K-27, an extension of the K-25 plant, which increased its output very considerably. It was really put in at the bottom of the plant, at the feed-in. And it was much, much bigger so as to really be able to push more material into this plant to have more come out. And we spent that 80 million dollars between March and, I think, November of ’45. I believe K-27 was on stream in November, or certainly December at the latest. That was spectacular for industrial spending/industrial building.
Your lead time on an ordinary plant, if you were to go to the M. W. Kellogg Company today and order an ammonia plant or a polystyrene plant or some kind of a plant to make a chemical, you would be talking fifteen or eighteen months, or you would be paying enormous premiums in order to cut that time down.
I am not being quite fair. The comparison is not quite fair because, after all, Houdaille-Hershey were all tooled up, ready to go with more barrier manufacturer, and Allis could turn out the pumps. They had their assembly line and Chrysler had their assembly line.
But just to get this rolling, get the site prepared, foundations poured, and get these pumps delivered, installed, and running in eight or nine months was a fantastic show, really.
Groueff: As far as the design—
Squires: We did the design. We had no inkling that anything like K-27 was coming along. And we were told it was going to happen. And we did the process design oh, I think in something like three or four days, because that was precious time. You just did not have any time to scratch your head. You had to get on with it because that held up everything else, procurement.
Groueff: No pilot plants before you built the K-25?
Squires: None at all. Of course, this order for K-27 in March, there was a lovely pilot plan.
Squires: Our first building was running.
Groueff: But for the first one, K-25, you had nothing to model from?
Groueff: When you started the design of K-25, did you start it from scratch? Benedict, you, and your team? Or did you use what Columbia people were already working on?
Squires: We used what Columbia people were already working on. And in some respects, the work during the summer and fall of ’42 was kind of a race and kind of a competition. We worked on many of the same problems. But we worked on them often from more of an engineering approach, whereas they would take a more—
Groueff: It was more of a competition than cooperation. And it was not the same.
Squires: No, there was enormous cooperation. Do not misunderstand me. There was competition at SAM.
Groueff: As one thing, or as two separate things?
Squires: It was in many ways separate and in many ways a competition. But there was the rapid exchange of reports. We never did anything without writing it up and getting it to them almost at once. And there were frequent meetings so that the criticisms could fly back and forth.
Groueff: There were two designs, their design and your design.
Squires: They were not really making a design. That would be a false way to think of the thing. Their approach, their point of view was more fundamental. It was, again, more of the scientists approach. And I am not using this in an invidious sense.
They asked themselves the questions of the following kind to give them the existence of a piece of equipment of this general type. “What are the mathematics, which will describe its performance? And how can this mathematics be quickly solved? In a sense, it was almost kind of like tooling up. Given the plant, how will it behave?” That would be one kind of question that they would ask. I want to come back to that question in a second because this is one I worked on myself very, very hard.
Another type of question related to the problem of the efficiency of the barrier assembly, given a barrier of certain characteristics. It is impossible to put that together in a practical piece of equipment with thousands of square feet inside of one pressure shell and not have inefficiencies come in, which are a correction upon this fundamental data on the little piece of barrier. Are you with me?
Squires: Now that factor could be very important because, after all, if the big piece of equipment only does seventy percent as good as the small one, you might be in serious trouble, where if the multiplying factor was ninety-five percent or ninety-eight percent, you might just not hardly even worry about it. Oh sure, you would take it into account in your calculations. But gracious sakes, the fundamental data on the barrier was not that good at that time. Well, how big was this correction? That is a particular problem.
This was an aerodynamic problem. And this is an interesting contrast in the approach that we took and that they took. We solved this problem using empirical chemical engineering methods, but applying them to this particular problem in really quite a straightforward way. And in a sense, the solution was the kind of solution that you could practically write down on an envelope, an engineer’s approach.
Groueff: Yeah, simple right.
Squires: A very simple approach. Karl Cohen, in the meanwhile, took some of the very latest, most sophisticated hydrodynamic theory, boundary layer theory from Kármán’s stuff, and solved the same problem. And was tremendously proud of his solution, very, very proud of this report and I think a bit miffed when our answers turned out to give practical—
Groueff: Better results.
Squires: No, practically precisely agreement. What told us was that the chemical engineering formulas, which we used, had a fundamental basis, which we had not known. Let us admit that. Which was maybe more secure than the empirical work, which had led to this. So everyone was happy.
Do not misunderstand me. But I think Karl was a little bit upset that we had gotten pretty much the same answer in such a simple and direct final design.
Groueff: But the final design for the big plant on the big industrial scale was prepared by your team?
Groueff: And not by the Dunning and Cohen team?
Squires: Not at all. They had review privileges. They could criticize. But the real responsibility for setting physical dimensions and turning it over to the mechanical engineers—“Okay, this is the piece of equipment, we want now you design it.” And then the blueprints that came back out of that, we checked it to see if it conformed to what we had asked for. This was our responsibility. It had to be.
Groueff: The men who actually did it then were Benedict, you?
Squires: I would say that Benedict, Charlie King, and I carried the main burden of this work. But heavens, we must have had ten or fifteen people working for me. And I could not even remember all of their names now.
Groueff: No, but that was the main team making the decisions and consulting each other every day?
Squires: That is right. That is right. I do not want to disabuse you of my role. In February – December or January – December of ’42 and January, February, and March of ’43 when this process design was being frozen and when there was this element of luck, that is when – do not forget, I was not even out of graduate school yet a year. I was a hand. I was working awful hard. And sometimes, I did not quite know what I was doing.
Groueff: The spokesman of the group was Benedict?
Squires: Benedict was really making the responsible decisions. I mean, he might ask me sometimes what I thought. And if I had thought he was making some horrible mistake, I might scream and he might pay attention, and sometimes he might not. I worked so hard during those – say roughly those first nine months.
It was a completely new discipline. I did a lot of work on the differential equations, the mathematical side of the story. Again, paralleling work that was done at Columbia. And we got into some controversies over some of our answers. I think we caught them in a few mistakes. And they may have caught us in a few. All of this was done very quickly. And it was very good that there were two different groups approaching it from slightly different points of view.
But during this period, I made myself an expert on the unsteady state behavior of this plant, the time behavior. You have to realize that this plant was so large and involved so much inventory of enriched material that if the whole plant existed, all standing there ready to run, you push the button, it would be thirty – forty – fifty days later before any usable product came out.
Groueff: I see.
Squires: At least at the top purity, at the final purity that the plant was capable of making. Partly, I suppose, this is just a little piece of luck for me. In the latter part of ’42, I spent perhaps all total six or eight weeks on this problem and got my feet under me on it. And then later, when the mathematical team was put together, Leon Henkin, Elliott Montroll, Joe Lehner, and some of the others worked on this problem from time to time.
Leon Henkin worked directly for me. You do not mind aggression? This is an interesting personality. I do not know if you want to look him up. He is on the West Coast — unless you are anxious.
Groueff: What is his name?
Squires: Leon Henkin, H-E-N-K-I-N.
Groueff: I have not heard that.
Squires: Have you heard that name before?
Groueff: Not yet, no.
Squires: Well, remember it because if you are on the West Coast, he is a personality. You would enjoy an hour with him. He is a professor of mathematics at the University of California at Berkeley.
Groueff: At Berkeley.
Groueff: I intend to go to Berkeley later because there are many—
Squires: Of course, many of the people are there.
Groueff: So I will make a point to stop there.
Squires: I think it might amuse you to talk to him. Leon Henkin is a Brooklyn Jewish boy. His father was born in Russia. I am sure his mother was too. He had his undergraduate degree in mathematics and then went to work for us. So we caught him before graduate school. But he was a very bright boy and very hard working. So we used him as a tool.
Through ’43 and ’44, Leon and I became, I suppose, for that time, the world’s experts on the time dependent behavior of this huge monster. And this knowledge became a kind of an aristocracy of brainpower late in ’44 when the plant was about to start up. So Leon and I – I think I had the prime responsibility. But I often used him as the errand boy. We practically – one or the other of us— made a trip at least once a week to Oak Ridge starting in about September or October of ’44.
I would often go down twice in one week. Oh, and that was a murderous flight in those days on the DC-3’s. I remember one night getting as far as Tri-Cities and spending the rest of the night there because the—
Squires: Tri Cities in the far northeastern corner of Tennessee.
Groueff: I see.
Squires: A few miles from the go, but the Knoxville airport was—
Squires: —closed in with fog. And so we just sat there in the plane the rest of the night. But we did. Leon and I made this trip I do not know how many times – probably thirty or forty times apiece. And these trips were involved first with discussions with the Y-12 people as to what they could use, and also to notify them of what they would be receiving. But then later in ’45, it became a complicated logistics problem. You had questions of the following kind. Okay, we got six buildings running. There are two more next week, two more the week after that, and two more the week after – or five more the week after that, say.
So the plant is going to grow according to this schedule. It is now producing at such and such a rate. Question – I will use numbers so you will understand. Let us say we were now delivering to Y-12 two percent enriched material – material at two percent purity at such and such a rate. As these new materials or buildings come on, we could increase the rate of which we ship two percent materials. Or we could stop shipping material and wait a week or wait ten days and send them four percent material. Now which was the better thing to do? You see, that kind of thing.
Squires: And this called for endless discussions and calculations both by the Y-12 people and by us, who understood the unsteady state behavior of the plant. We would be able to tell them, “Okay, if you go from three percent purity to seven percent purity, it is going to cost you one week or it is going to cost you four days or whatever for the plant to pull itself up to this new higher level.”
I can remember being furious and very angry with Manson Benedict over this because it was not that I felt like shirking responsibility. But I felt that this is something that he should have been doing. But he was completely and thoroughly engrossed in completion reports at this time. The contract required that we send them literally a five-foot shelf of these completion reports. And he set himself up as the main editor. And he even asked me to do some writing on them. I just point blank refused because I felt this other was so much more important. I was just scurrying back and forth. And to be perfectly honest, it was a tremendously exciting time for me.
But I said that our plant always did better than we expected. This was a real mystery that developed. You are probably not going to hear this story from anyone else because I do not think anybody else looked at it from my point of view and had quite the same attitude toward it.
Leon and I were experts on what the plant should do, both when it reached a steady state and how long it should take to go from one state to another, and it always did better than our predictions. It always did about 10 or 20 percent better. We could not understand it, because we were working with differential equations. We were working with differential equations that ought to give the answers. It always did better. This remained a mystery all through the summer of ’45.
I went to Oak Ridge. I think it was September 1st.
Groueff: So after the bomb?
Squires: Yeah, after the bomb. I went there September 1st full-time to organize a process analysis department. I put together a team of about twenty-five people, many of them the Army boys, the technical service boys who had been called in. A lot of boys were drafted right out from under our nose and then came back to us a few weeks later. I do not know whether you are aware of this.
Groueff: Yeah, I was told.
Squires: A lot of people, Fred Zinn, who is my business partner now, worked for me first as a civilian. And then he went away a few weeks and he came back in uniform. Well he worked for me in this department that I organized in Oak Ridge. And we hired George Garrett, whose name you may have come across. He is post-war. He would not be perhaps part of your story. But I hired him as the head of the process development department – process analysis department for Carbide at Oak Ridge. And he has headed up the effort for designing all the future gaseous diffusion plants.
So I feel like I am sort of grandfather to these plants even though I had nothing to do with them. But I trained George. I was in charge of the department, in direct charge, I think, oh, through about February. And then he took over perhaps maybe March 1. And I stayed on until about July 1. My status was that I was still a Kellogg’s employee, but I was on loan to Union Carbide. I did not want to work for Carbide.
But all through September and October, this plant kept doing better than anything that we could calculate that it ought to. And I screamed and screamed that something was wrong with the—
Squires: No, not with the calculations. The calculations had to be right. With the mass spectrometer. Actually, I am not casting any asparagus at Alfred Nier or anybody. This was just one of those things—there just had to be something wrong with the mass spectrometer analyses for these ratios of U-235 to U-238 on the fundamental data for what the diffusion barrier would do.
Squires: Do you follow me?
Squires: So finally, we got some completely pure U-238 – one hundred percent U-238 and one hundred percent U-235 from the Y-12 people, which they had made on an experimental basis. You see, they could do this by fine-tuning and fine adjustment of their machine. They could make small quantities of absolutely pure isotope with either of the two isotopes. Then on an analytical balance, you can weigh and mix the thing and make up synthetic mixtures. And sure enough, the mass spectrometer had a bias. It had a memory effect. It always remembered the previous sample.
You were always comparing. You were always comparing a sample, which was not enriched.
You were comparing an unenriched sample with an enriched sample. And then going back and checking the unenriched one again. And then comparing the enriched one. And that always pulled your result together. And the bias was just right to make our calculations correct, or to make the plant be doing what it should do, so to speak, rather than doing better than it should do.
Groueff: Correctly, it worked on this one, not the opposite one, because the spectrometer—
Squires: Had had a bias in the other direction, we would have been—
Groueff: You would have been in trouble.
Squires: Yeah, we would have been in serious trouble because this particular piece of fundamental data, we called it the enrichment factor. The plant performance goes as the square of that number. So if you are 10 percent low in your estimate of the thing and the true value is really 10 percent higher than is coming out of the mass spectrometer laboratory, then the plant will do 20 percent better. On the other hand, if the laboratory data had been 10 percent off in the other direction, the plant would have done 20 percent worse, which would have been very serious.
Groueff: I see.
Squires: It would have been serious not only from the standpoint of how much material the plant could have made, but it also would have been very serious in terms of these times. It could have easily delayed the July 15 and the August 6 date because those were our bombs, the U-235 ones. It could have conceivably delayed them several weeks or a month, simply because the stretch out from the gaseous diffusion plant—
Groueff: But from what I understand from you now is that the final product, which went to Los Alamos for the bomb actually came from Y—
Squires: Oh, it came from Y-12.