The Manhattan Project

In partnership with the National Museum of Nuclear Science & HistoryNational Museum of Nuclear Science & History

John Marshall's Interview

Printer-friendly version
After receiving his doctorate, John Marshall was hired to assist Leo Szilard with his experiments at Columbia. Afterwards, Marshall traveled to Chicago to work on Chicago Pile-1, and finally to Richland to work on the B Reactor at the Hanford site. Marshall was on duty when the reactor shut down due to xenon poisoning. He discusses his experience working for Szilard and alongside Fermi, as well as the steps taken when the B Reactor shut down on his watch.
Manhattan Project Location(s): 
Date of Interview: 
1986
Location of the Interview: 
Los Alamos
Collections: 
Transcript: 

[At top is the edited version of the interview published by S. L. Sanger in Working on the Bomb: An Oral History of WWII Hanford, Portland State University, 1995.

For the full transcript that matches the audio of the interview, please scroll down.]

Book Version:

I got my doctorate in physics in 1941 at the University of Rochester. At that point, the group under Fermi and Szilard was at Columbia University. They were trying to answer some questions, to carry the investigation further. Szilard was a very bright guy, but not very good with his hands. He wasn't an experi­mentalist in any sense. Well, he wasn't a hands-on experimentalist. He was looking for somebody to be his hands and he must have asked Victor Weisskopf who was at Rochester if there were anybody around who would be available and Weisskopf gave him my name. I visited Szilard at Columbia and he gave me a long rambling account of what the problems were. Next thing I knew, I was hired.

The project was in full swing at that point. They were interested in find­ing out whether it was possible to make a chain reaction in a pile. There had been a series of experiments and new experiments were coming up. There was an effort being made to find a source of graphite which was pure, had less boron in it or whatever other neutron-absorbing impurities there might be. Also, they were looking for pure uranium and making it suitable for putting in a reactor. There was a system worked out by I guess mostly Fermi, called an exponential pile, which generally consisted of a stack of graphite in the shape of a prism, higher than it was wide and deep.

I did whatever needed doing at Columbia. For instance, Szilard was doing some experiments on spontaneous fission, photo-neutrons, photofission, gamma ray induced fission. All of which would be relevant to a chain reac­tion. I was still Szilard's hands, to some extent. Szilard tended to do things he was interested in, then he would switch over and be interested in something else with great enthusiasm.

Somehow, I found myself measuring neutron distribution in exponential piles, then in the construction of them. Getting the graphite machined, pressing uranium lumps that went into it. All this was in the fall of 1941. Along about Christmas, well, there was a group at Princeton, there was Sam Allison's group at Chicago, and isotope separation at Berkeley, us at Columbia, and they wanted to bring the thing together. They wanted someone to head it up and that was Arthur Compton, so the chain reaction group was moved to Chicago shortly after Christmas, 1941.

At Chicago, again I got mixed up with one of Szilard's enthusiasms which was to try to get uranium metal for fuel instead of uranium oxide. For quite a while I was busy trying to figure out how to cast uranium, making vacuum furnaces and trying to melt some small samples that Szilard had managed to scrounge up someplace. Metal was somewhat better in the pile, it was denser and tended to enhance the lumpiness of the geometry of the system. You just don't mix the uranium or the heavy water and the uranium, whatever it is, but you put the uranium in lumps. That was probably Fermi's idea.

Fermi was a marvelous guy, I liked him very much. He was an extremely productive person. He had an active imagination and he didn't waste time. If he was doing something, he would do it, instead of wandering around think­ing of what he ought to be doing. He was very much a hands-on physicist, but he also was a very competent theorist. When he built experimental apparatus, it was no better than it had to be, in fact, it was downright sloppy at times. But it would get the work done. If a new idea came into his mind, he would work on it very rapidly until he got sufficient stuff to come to some sort of conclu­sion. Then he would publish immediately. If somebody else got the same idea he did from the same source at the same time, Fermi would be there a month ahead.

I was in rather close contact with Szilard until the end of 1942. Szilard was one of these guys who is a little bit too bright. He had the right conclu­sion as to what should be done but it would turn out in practice to be something that couldn't be done for 20 years. Like he was insisting that the Hanford production reactors should be liquid metal-cooled. Now that is the way you do high-performance reactors nowadays but to try to have done that in the time available then was not possible. He was irritated by what he called quote engineers sneer unquote, particularly the Du Pont engineers. At one point, I'm told he went to Arthur Compton and said the engineers were getting intolerable. Either they had to be fired or he was going to quit. Compton said, "You have just resigned." After half an hour Compton cooled off and came back and said, "I didn't mean that, Leo."

For a while, I was at Indiana University, with Louis Slotin, to run a cyclo­tron experiment, measuring temperature co-efficients, to find what effect temperature of the uranium lumps would have on reactivity. I went back to Chicago, and they were moving the original pile out to the Argonne site. We were happily doing experiments of various sorts, Fermi, Wally Zinn, Leona, measuring neutron cross-sections and god knows what. We used that reactor as a source of neutrons. By then, it was 1943.

Well, I was interested in staying with the reactors, to see them going, and the Du Pont Company was in charge at Hanford, so I got a job with Du Pont in January, 1944. I stayed at Argonne for six more months, mostly working on dull stuff figuring out all sorts of procedures of how to start up a reactor and how to operate it. In June Leona and I went to Hanford.

We came by train from Chicago. We had a car but the tires were not fit to go that distance. We were assigned a house, right off. I started in the 300 Area, a technical and laboratory area, and we worried about startup procedures, get­ting geared up for starting the reactors. There was a bunch of us known as babysitters. Our job was to be on hand 24 hours a day, some of us, during startup and sort of monitor what was going on, and assist the chief operator in checking things out. We also were there when the reactor was loaded. Some-body was making a joke once, probably John Miles, a Du Pont physicist who headed our technical group. He said we should be called proctologists because we were specialists in piles.

I was there for the poisoning. That's an amusing story. What happened was that the brass was out there for the startup day, I was on the four to midnight shift. The thing was started without a hitch, then everybody went home, Fermi, Crawford Greenewalt, John 'Wheeler and various others. I wasn't there during that time. I came on just as- the reactor died. They had started pulling the control rods out and out and out, finally they couldn't come out any farther and it turned off. Well, of course, there was consternation. The chief operator and I were trying to figure out what in hell had gone wrong, and chased around looking for what seemed probable to us. For instance, a water leak in the reactor, or a loss of helium and a replacement with air. There was something like a percent of reactivity that came from getting rid of the air and putting in helium. The reactors were slightly pressurized with helium, so we checked for helium leaks. There weren't any.

The shift ended and I went home. It was still dead. Somebody came on, I think it was George Weil. He decided he would find out how far below criti­cal the reactor was. He put in a foil that measured the radioactivity and told what level of neutrons there was if all the control rods were pulled out. By that you could tell how far below critical it was. He put the foil in and had the rods pulled out and the thing started. Of course, it shut down again by the time the next shift came on. During one shift it killed itself, then started again. Then it shut itself down again.

The next day Fermi and Leona and John Wheeler and somebody work­ing with him, a Du Pont theoretical type with an Irish name [Dale Babcock], decided something radioactive was doing the job. Both teams were analyzing reactivity as a function of time, as given by the control rod position, and from that deriving the radioactive periods that were involved. Then, all they had to do was look in Seaborg's table of fission products and they could pinpoint that it was iodine and xenon-135. By that afternoon Fermi had worked out a set of formulas so you could predict the reactivity as a function of time. Shortly thereafter, the fuel load in the reactor was increased. The reactor had been taken just barely over critical in the initial loading.

At Hanford there wasn't any organized opposition to the use of the bomb. We were concerned, and hoped it wouldn't be used against cities. There were relatively few people there who knew there was a bomb, just the scientific staff and some management. There was discussion among a few of us about what would be done with the bomb but we didn't try to exert any pressure on anybody. Since most people there didn't know what we were doing, all sorts of stories popped up. I remember one was that we were making the front end of horses to send to Washington, D.C. to assemble with the other end.

We didn't have much time for a social life, except some with the other physicists. Once we went by bus to Seaside, Oregon, and another time to Chelan and took a boat up Lake Chelan. Recreation was hiking or working around the house figuring out ways to get a swamp cooler to work or putting up a fence to keep the baby from wandering. The only thing rigorous about the weather was the dust storms. They were spectacular. Richland is laid on river pebbles, which come in any size from microscopic to the size of your head. They bulldozed the place flat, got rid of whatever top soil there was and brought in silt from the Yakima River flats. They put six inches of this stuff over the town. There's all of, I think, six inches of rain there a year and when the wind blew you wouldn't be able to see across the street. Every morning there was a sand dune under the front door, a crescent dune.

 

Full Transcript:

S. L. Sanger:  This is an interview on March 5, 1986 with John Marshall at his residence near Los Alamos. Say how you got started in the project originally.

John Marshall:  I was a graduate student at the University of Rochester and finished and got my degree in 1941. At that point, the group under Fermi and Szilard was at Columbia University, and they were trying to answer some questions to carry the investigation further. I guess Szilard, who was a very bright guy but not very good with his hands – he wasn’t an experimentalist in any sense really. Well he wasn’t a hands-on experimentalist. I’d say he probably was an experimentalist in the sense that he inspired and supervised experiments. 

Anyway, he was looking for somebody to be his hands, and I think he must have asked Victor Weisskopf, who was on the faculty at Rochester at the time, if there was anybody around that would be available. Weisskopf gave him my name. So I went and he asked me to talk to him. I think it may have been at a meeting of the Physical Society or something, I don’t remember exactly. Anyway, I visited him at Columbia. He gave me a long rambling account of what the problems were. The next thing I knew, I was hired there. So I went there in August 1941.

Sanger:  To Columbia?

Marshall:  To Columbia.

Sanger:  Now you had a doctorate in physics then?

Marshall:  Yes.

Sanger:  In physics.

Marshall:  Right. I had been working on with the cyclotron at Rochester. Anyway, the project was in full swing at that point. They were interested in finding out whether it was possible to make a chain reaction in a pile – in a reactor – and there had been a series of experiments, and new experiments were coming up.

There was an effort being made in finding sources of graphite which was pure, had less boron in it or whatever neutron absorbing impurities there might be. Also finding pure uranium in means of making it suitable for putting in a reactor. There was a system worked out by mostly Femi, which was called exponential pile. It generally consisted of a stack of graphite in the shape of a prism – square prism – higher than it was wide and deep. It would have uranium in it.

Sanger:  What was your particular duty when you were at Columbia?

Marshall:  I was just helping out with whatever was going on. To begin with, there were some experiments that Szilard was trying to do on things, like spontaneous fission and photo neutrons from photofisison. It’s from gamma ray induced fission, all of which would be relevant to chain reactions. I was figuring out how to do experiments and operating them to make that kind of measurement. I then rather rapidly got into the exponential pile business, and I was just joining in with the other people there. 

Sanger:  Were you still sort of being Szilard’s hands or did you pass that?

Marshall:  Well, I passed it to some extent. I think Szilard tended to have things that he was interested in, then he’d switch over to being interested in something else with great enthusiasm.

Anyway, somehow I found myself in the business of measuring neutron distributions and the exponential piles, and then just in the construction of getting the graphite machine and pressing the uranium pellet and the uranium lumps that went into it.

Sanger:  This would have been when about?

Marshall:  It would have been in the fall of 1941.

Sanger:  Is that when you started or?

Marshall:  Yes.

Sanger:  Okay.

Marshall:  I was at Columbia from August 1941 to about April 1942. Anyway, during that time, around about Christmas time – well there were other people involved in it besides the people at Columbia. There was a group at Princeton and there was Sam Allison’s group at Chicago. They wanted to bring the thing together. Also of course, there was the other business of isotope separation going on at Berkeley, but that didn’t really concern us. Also of course, Columbia had the isotope separation business that did the uranium hexafluoride business.

Sanger:  You were always listed in the reactor end of it?

Marshall:  Yes. I was with the reactor end of it. Well, except for minor excursions here and there. Anyway, the idea was to bring the whole project together, and they wanted a name, somebody to head it up. They picked Arthur Compton. And that moved it to Chicago. There were arguments as to where it would be and Chicago was picked.

Sanger:  You’re talking about bringing the chain reaction—

Marshall:  Yes, the chain reaction research development that was going on moved to Chicago. People started moving shortly after Christmas 1941 and I was still finishing up some experiments there, so I didn’t get to Chicago until 1942 about April.

Sanger:  Were you still with Fermi then?

Marshall:  Yes.

Sanger:  Did he go right away?

Marshall:  Yes. Pretty much.

Sanger:  Okay, and I suppose you got really involved with Chicago Pile?

Marshall:  Well, yes. At Chicago, again I got mixed up with one of Szilard’s enthusiasms which was to try get uranium metal going so it could be used for fuel rather than oxide. And so for quite a while in Chicago, I was busy not working on the pile directly, but trying to figure out how to cast uranium, making vacuum furnaces and trying to melt some small samples that somebody had gotten, that Szilard had managed to scrounge up from some place.

Sanger:  Metal is better? Is it?

Marshall:  Well, it would be somewhat better being denser. It would tend to enhance the lumpiness of the geometry of the system. These reactors like to be inhomogeneous in the sense that you don’t just mix the graphite and the uranium or the heavy water in the uranium or whatever it is, but put the uranium in lumps.

There are good scientific reasons why that should work. I think again that was probably something that was Femi’s idea. But if you put the uranium in lumps, the neutrons – the fast neutrons – get a chance to get slowed down between the lumps and come down to thermal energies without being absorbed by resonance absorption in uranium-238 on the way down, as they diffused down in energy.

Anyway, there were other people around the country working on making uranium. In particular, there was an outfit called metal hydrides that was on the north shore in Massachusetts. I forget the name of the town, but someplace there. And they were making uranium metal by first making uranium hydrides and then decomposing that, and it made a powdered uranium. Which was fine, except that if you just looked at it, it would catch fire. Pyrophoric was the word. So they’d made quite a lot of that and there were attempts being made to turn that uranium into something that would be useful in the first test pile, in the west stands pile in Chicago.

MIT volunteered its facilities, and there were two groups working on trying to make the uranium suitable – make suitably metallic uranium. I went with one kind, I mean with the idea of doing vacuum casing uranium so as to make solid cast lumps. Bill Jesse was the man who came from Chicago to work on that. I forget who else was on that particular thing.

Anyway, Bob Walter and I went from Chicago to the facilities that belonged to Professor John Chipman at MIT. He had facilities for doing induction heating, induction melting, making induction furnaces at frequency. So we used his apparatus and developed a means of casting uranium in graphite crucibles – vacuum casting. So while we were there we produced several tons out of the more or less unsuitable stuff. While we were doing the casting stuff, Jesse and company were trying to do it by pressing.

Sanger:  To make it into metal.

Marshall:  Yes. Take this powdered stuff. Turned out even if you pressed the stuff and tried to cinder it afterwards that the metal was still quite likely to catch fire. If you just dropped one on the floor it would catch fire and would sort of glow like a cigarette butt.

Sanger:  Where, incidentally, did the uranium come from in those days?

Marshall:  I think it was coming from the Eldorado radium mines of Great Slave Lake in Canada. There was a Canadian bunch that were interested in – that were helping develop the uranium. Getting pure uranium.

Sanger:  This metal hydrides company, what were they making it for?

Marshall:  I think they were making it for this project. I’m not sure, but I think it was something that Szilard arranged. He and the Eldorado people may have been together somehow. He had contacts with people in France and Canada, and God knows where.

Sanger:  There was also a great store from the Congo. The Belgians.

Marshall:  Well, I know of some that appeared in Holland. It stayed there during the war, which was picked up later by the Norwegians and the Dutch to put together the Keller reactor. That was some that just sort of stayed in the back lot of some industrial outfit. I forget which one. Maybe Phillips or somebody.

Sanger:  I think there is a story that the Belgians had shipped some to – had to have been a big large warehouse – and later it was obtained by the United States.

Marshall:  Could well be. I don’t remember that. I don’t know the details.

Sanger:  And that was from their mines.

Marshall:  Honestly I was worrying about hardware.

Sanger:  Yeah. So then what?

Marshall:  Anyway we made several tons of the stuff. That started in August 1942, and we were done. The reactor had run at Chicago before Christmas 1942.

Sanger:  So you used this in the first one?

Marshall:  I think it was the major source of metallic uranium that went into that. There was some also being made at Ames, Iowa by [Frank] Spedding and company. But they were making it by another chemical process. Not the hydride business, but a fluoride business.

Sanger:  How much uranium went into that first one, do you know?

Marshall:  I don’t know.

Sanger:  Would have been more than several tons.

Marshall:  I would say tens of tons. Most of it was oxide. U3O8 and U3O2.

Sanger:  And that was in lumps?

Marshall:  That was in lumps. Pressed in a dye and the lumps were sort of cylindrical, but nearly spherical.

Sanger:  I saw Norman Hilberry, I think he had a couple. They’re black.

Marshall:  U3O8 was black, yes. It would be about this big around and about this tall. They were cylinders about the same height as their diameter.

Sanger:  And you could rub it off?

Marshall:  Yeah.

Sanger:  He brought one out. He had one scrolled away someplace.

Marshall:  Anyway, the metal went just into the central core, where it would do the most good. The idea was to try to get this thing to operate as soon as possible. It did operate, and then I found about it when somebody came visiting from Chicago and said, “Well I guess you ought to know,” and told me. It was going.

Sanger:  You weren’t there the day it started?

Marshall:  No, I was at MIT still making uranium to go into it.

Sanger:  I see. Did things change a lot after that happened or not?

Marshall:  Yes. The next thing that happened to me was that I got back to Chicago at Christmas time, and they were looking for somebody to run an experiment with the cyclotron at Indiana University in Bloomington.

What’s his name, the guy that got killed in the chain reaction accident later at Los Alamos? Louis Slotin. Louis Slotin and I went down there to do some experiments using the cyclotron as the source of neutrons. We were measuring a temperature coefficient of the temperature of the lumps themselves and what effect that would have on reactivity. It was a sort of nasty experiment because it was sort of a fraction of a percent effect and took a lot of fiddling around to get the experiment done. A lot of data.

We did that and then went back to Chicago. When I got back to Chicago, the bunch were moving the west stands pile up to the Argonne site. Palace Park. So I joined in on that. And that was people like – well, Fermi was in charge of it, and Wally Zinn. You know who he is?

Sanger:  Yeah.

Marshall:  And a lot of other people. Leona [Libby] was there working on it. Herb Anderson, Darragh Nagle, and a lot more people. I can’t remember all the names. They don’t come to mind right at the moment.

Anyway, we were happily doing experiments there of various sorts, measuring neutron cross sections and God knows what. We operated that reactor once it got going. Used it as a source for neutrons and a means of measuring things. You could measure things, you could measure cross sections and purity of materials by measuring the criticality of the reactor. See that was 1943. I was interested in getting and staying with the reactors, to see them going.

About that time, the DuPont Company was put in charge of doing the Hanford reactor. And since I was interested in that, I got a job with the DuPont Company to work on that reactor – on those reactors. I started working with them formally in January of 1944 after having spent about seven or eight, nine months at the Argonne Lab at the Palace Park site, doing various experiments.

Sanger:  Did you ever go to Oak Ridge?

Marshall:  Only on a visit.

Sanger:  You didn’t go there to work on that reactor there?

Marshall:  No.

Sanger:  So, then you joined DuPont formally in the first part of ’44?

Marshall:  Yes, and worked there doing dull sort of stuff figuring out, writing out all sorts of stuff about how to operate a reactor, start up a reactor, and operate it for about six months. Until June 1944, at which point we went to Hanford, went to Richland.

Sanger:  What’s your first memory of that?

Marshall:  Well, somehow I don’t remember the exact moment when we arrived. We came by train. We had a car, but the tires weren’t fit to go that distance so it was shipped out there by the DuPont Company. We arrived a month and a half or two months later, and we were met, I guess, by the company at Pasco Station.

We were assigned a house right off – had one for us. I think it was – I can’t remember the exact address – I think it was 208 Armitage [misspoke: Armistead] Avenue in Richland. All the streets were named after generals. I don’t know if they still are or not.

Sanger:  Yeah, they are. The older ones are.

Marshall:  Yeah. Anyway, we had a new baby at that time.

Sanger:  Was that boy or girl?

Marshall:  Boy.

Sanger:  So he was born in Chicago?

Marshall:  Yes, he was born in Chicago at the end of March, and Leona took all of the week off to have the baby.

Sanger:  Yeah. He was about three months old when you got to Richland?

Marshall:  Yes. And Leona’s mother came with us and lived with us while we were in Richland.

Sanger:  The whole time?

Marshall:  Yes.

Sanger:  How long were you there then? [00:27:00] Before I forget.

Marshall:  There from June or July, I guess. I’m not sure of the exact time when we got there. From July 1944 until December 1945. So we were there for a year and a half during which time the reactors were constructed, mostly. Started up, operated, and produced quite a lot of plutonium I guess.

Sanger:  What did you do then when you were at Hanford?

Marshall:  Well, we started off at what they called the 300 Area, which was a sort of technical – it was a laboratory area that was down close to Richland. Just barely inside the reservation. We were inside the excluded part doing exponential piles again and worrying about startup procedures. That was essentially all it was. Just still getting geared up for the start of the reactors.

Sanger:  Did you go out to the first one quite a bit?

Marshall:  Yes, I was there for the start-up.

Sanger:  What was that like?

Marshall:  Well that was interesting. There were sort of a bunch of babysitters for the pile – for the reactor. These were myself, George Weil, Don Hughes, Rudolph Kanne, and Henry Newson. I guess that’s five. And our job was to be on hand 24 hours a day – someone of us – during start-up and to sort of monitor what was going on, and assist the chief operator in getting things done and checking things out.

Of course before that, we were there while the thing was being loaded and mostly worrying about the start-up procedures and operating procedures and writing a bunch of stuff. Once the thing was started, we kept on somebody there for babysitting it for many, many months.

Sanger:  You were probably there for the poisoning.

Marshall:  Yes. I was there for the poisoning. That’s an amusing story, of course which you’ve heard before. What happened was that the brass was out there for the start-up day. I was on the four to midnight shift, and I think it was George Weil but I’m not sure, who was on midnight to eight a.m. It might have been Henry Newson.

The thing was started without a hitch. Taken up to ten megawatts – just a nice, reasonable level. A low level of power. Then everybody went home. Fermi and Crawford Greenewalt and John Wheeler and the various people who were helping them. Maybe some others I’ve forgotten. I wasn’t there during that time.

They went home at quitting time, four in the afternoon. I came on just as the reactor died. It started having to pull the control rod out and out and out. Finally, it couldn’t come out any farther. The thing just plain turned off. Well of course, there was consternation.

The chief operator whose name I can’t remember and I were trying to figure out what in the hell had gone wrong. We chased around the site looking for what seemed probable to us that something had gone wrong. The kind of things that could’ve gone wrong would have been, for instance, a water leak into the reactor, or a loss of helium from the reactor and replacement with nitrogen – air.

Sanger:  The reason they used helium because it didn’t absorb as many neutrons as—

Marshall:  Right. I believe was something like a percent of reactivity that came from getting rid of the air, putting in helium instead. So the things were slightly pressurized with helium, and what we did was go around and look at all the possible places to see if helium was coming out. There wasn’t any coming out that we could find. Finally, the shift ended and I went home.

Sanger:  And it was dead by then?

Marshall:  It was still dead, yes. After having shut itself down, the thing was put in a formal shutdown state with all the rods put in, so it wouldn’t have gone under any circumstances.

Sanger:  And what was your position when you were out there? You were kind of the babysitter on hand there or what?

Marshall:  Yes.

Sanger:  What was your title? If you had one.

Marshall:  I don’t know.

Sanger:  You were sort of the physicist in residence at the reactor when that went on?

Marshall:  Yes, there was a technical group that was a group under John Miles, who was a physicist working for the DuPont Company, who were the babysitters. I think somebody was making jokes, probably Miles. We were called proctologists. Because we specialist in piles.

Sanger:  So anyway, the thing was shut down.

Marshall:  The thing was shut down. I think it was George Weil who came on and he decided he would find out how far below critical the reactor was. So we put a foil to measure slow neutrons and measure the radioactivity of it, and used that to tell what level of neutrons there was in the thing before all of the rods were pulled out. And by that, you could tell how far below critical it was. Because when the thing is below critical, the spontaneous neutrons that are produced in the thing are amplified perhaps by a factor of a 100 or something like that, depending on how far down you are. It can tell how far down you are by the level. So he put the foil in – indium foil – and had the rods pulled out, and the thing started up again.

Sanger:  But it was ready to go?

Marshall:  Yeah. It started up again. Well of course they operated it immediately through the midnight shift, and it shut down again by the time for the next shift to come on. It shut itself down.

Sanger:  So in the course of one shift it was down, then the course of the next one it came back up?

Marshall:  During one shift it had killed itself, and then it was put in a state where it couldn’t go. Then it started again, and then it shut itself off again. I’m not exactly sure what the details were during the day the next day, except that Fermi, and Leona working with him, and John Wheeler and the fellow working with him – he was a DuPont chemical engineer who was sort of a theoretical type, with an Irish name. I’ve forgotten what it was.

Sanger:  I know who you mean but I can’t remember either.

Marshall:  Anyway, they formed another team, and it was perfectly obvious at this point that there was something radioactive doing the job. So both of them were analyzing the reactivities as a function of time, which was given by the control rod positions, and from that deriving the radioactive periods that were involved.

Then all I had to do was to look at Seaborg’s tables of fission products, and they could pinpoint that it was iodine and xenon-135. So they did, and by that afternoon, Fermi at least had worked out a set of formulas so that you could predict the radioactivity as a function of time depending on the power level and so forth.

I don’t remember exactly how long it took people to figure it out, I mean to decide, but the load in the reactor was increased. It had been taken just too critical really – or just barely over critical – in the initial loading, because it’s just conservatism.

Sanger:  Is that when they were just using 900 or so tubes?

Marshall:  I’d forgotten how many there were. There were 1500 tubes available in the original design, and that’s what they had in mind for a load. That was a cylindrical load. But there were 2004 tubes available because of the DuPont Company, in their conservatism, decided to put as many tubes in as they could. And lucky it was. It was good management. Anyway, the thing was brought up rather rapidly to 2004 tubes. And this allowed the thing to be operated at considerably higher power levels then what had been in the original design.

Sanger:  And that had the effect of overcoming the poisoning phenomenon?

Marshall:  Yes. By having a whole bunch of access reactivity, the thing could be operated in spite of the poisoning. The poisoning amounted, perhaps I’ve forgotten to what percentage, something in the order of the percent of reactivity.

Sanger:  And that was just enough to bring it below it.

Marshall:  Well it was such that if you took it just to a starting-up load that you’d be barely over critical, and then just a little bit of poison would kill you. So by loading the thing right up to the top, it was possible not only to overcome that, but also to do a thing which was the flattening of the neutron distribution and the power distribution in the pile.

This was done by putting some poison into the center of it. This depressed the local central neutron level and gave it a flat top. Normally, you would expect that such a thing would have a cosine distribution – that is, the thing would fall off like a cosine function. We were limited in how much power we could have in the hottest tube, just because of engineering conservatism. Didn’t want some of the uranium to break loose or something in there.

By depressing it, there was more excess reactivity than was required to take care of the poison. So it was possible to do this flattening. This then meant that since the distribution without any flattening would have been like that, the distribution with flattening now could be like this.

This could have a considerably larger integral under the distribution, so that at the same maximum you could have more power. What it amounted to was that the thing could now produce more plutonium than what had been planned.

Sanger:  How long was it, do you remember, before the people like Fermi and so on got out there when they heard about the shut down?

Marshall:  Well, they got out there at eight o’clock the next morning.

Sanger:  And originally it was thought it was something mechanical, a leak or—

Marshall:  That’s what we thought. Fermi was mad that we hadn’t called him up and got him out there in the night, because then he could have gotten it all done before John Wheeler got there.

Sanger:  Who did you work for then? This guy, Miles?

Marshall:  Miles, yes.

Sanger:  Were you still closely in touch with Fermi by then or not?

Marshall:  Yes, pretty close.

Sanger:  Is that because – scientifically? Or just because you knew him better than most people did?

Marshall:  Well my wife was working with him, and on that account we saw each other frequently in our leisure time – which there wasn’t much of but, I mean we’d take a hike or something like that. And then I’d just been in contact with him over the years so that we discussed things with him if it were any problem. Mostly it wasn’t anything much that he needed to do.