Schreiber's memo to Froman on the Slotin Accident. 28 May 1946 (Schreiber Collection)
‍
D. K. Froman
R. E. Schreiber
‍
Report on May 21, Accident at Pajarito LaboratoryThe following presents the material you requested last Friday.Perlman and I had gone to Pajarito Tuesday after lunch to carry out the field checking of the initiators and active material for the Crossroads tests. We carried 5 initiators in the three special shipping cases. Slotin had arranged for the three 49 (2) cores to be brought down that morning.
‍
We arrived about 1:30 P.M. and found the multiplication measurements on the composite core, started that morning, still in progress. Since the source being used in this test interfered with our field-checking measurements we waited for about one hour to start our work. (It turned out that one of the 25 (2) shells had been left out by accident during the morning measurements so part of the multiplication measurement had to be repeated.)
At about 2:30 PM, Perlman and I started the counts on initiator Au-18. Meanwhile, Graves came in. A half-scale shot was being fired by M-4 at the lower Pajarito Site.

We suspended operations during the shot but finished our neutron and gamma counts about 3:15 PM. Meanwhile Graves and Slotin discussed the critical assembly work. In the course of the conversation, Graves asked Slotin to demonstrate a critical assembly since he (Graves) was planning to do some experiments during Slotin's absence and wanted some indoctrination. At first Slotin said that he didn't have the proper materials for one. Then he remembered that we had the 49 cores there so he said he would do one "in about two minutes" in a beryllium tamper after we (Schrieber and Perlman) had finished our counts. I remarked that if he were going to do it in two minutes I was going to leave but would stick around if he took a half hour for it. This was not intended seriously since we all had confidence in Slotin's ability and judgment.
‍
We finished our counting at about 3:15 PM and I put on rubber gloves and started to take the initiator from its case to check the counting. The neutrons from the driving source to be used by Slotin during the assembly would not affect this check so Slotin started his experiment. I was working on the initiator so did not pay close attention to Slotin's operations since I assumed he had done this before and knew how far he could go safely. I did, however, glance around at intervals so the following is my observation and is not based on subsequent conversation. I cannot swear to all the details but state what I believe to be true.
‍
The beryllium tamper was set up on a low steel table near the north end of the room. Slotin stood before the table facing south and Graves stood behind him to his right. Perlman was working with me at a bench along the east side of the room about 15 ft. away. Kline, Young, Cielslicki and Cleary (SPO Security Guard) were around the assembly table at various distances but I do not know their precise positions.
‍
The table on which Slotin was working was fitted with various detectors, some operating counters and others driving Esterline-Angus recorders. I believe that these were tested before the assembly was started. Kline and Cieslicki were running these so could give definite information. A strong Pu-Be source was used as a driving source. I do not know where it was put during the assembly.
‍
I had assumed that the approach to critical would be rather slow so continued to work on the initiator, thinking that when the multiplication got to an interesting point I would turn and watch. It could not have been more than two or three minutes after the start that I turned because of some noise or sudden movement. I saw a blue flash around the Be tamper and felt a heat wave simultaneously. At the same instant, Slotin flipped the outer top tamper shell off (The tamper was made of two concentric shells, 9" and 13" O.D, I believe). This stopped the reaction.
‍
The blue flash was clearly visible in the room although it (the room) was well illuminated from the windows and possibly from the overhead lights. I believe that the flash appeared only around the hole in the upper tamper hemisphere and around the equatorial gap. Slotin's left hand, which was holding the top hemisphere, was definitely in the glowing region. The total duration of the flash could not have been more than a few tenths of a second. Slotin reacted very quickly in flipping the tamper piece off. The time was about 3:20 PM.
‍
A few seconds after the accident, only Slotin, Graves, and myself were left in the room. Perlman had run up the corridor a few steps and was waiting, the other four had gone out the east door or up the corridor. The rest of us left immediately, going up the corridor.
Slotin called for an ambulance and then prepared a sketch showing our positions at the time of the accident.
‍
Slotin had picked up a Watts ionization chamber meter which must have been near the assembly. We attempted to measure the radiation from the various articles with this but it was so active itself (about 15 div. on the 1010 scale) that nothing conclusive could be found out.
‍
After about 5 minutes, at Slotin's suggestion, I took a handful of film badges and put them on the tamper. I carried the Watts meter. The room was quite hot, the meter going over full-scale near the assembly, so I did not linger.
‍
After about 12 or 15 minutes, I went down again. The meter had meanwhile cooled down to about 5 divisions on the 1010 scale so was fairly usable. The reading about 10 ft. from the assembly was about 20 divisions on the 1010 scale but again off-scale nearer. I tried the G-M counter in its Columbia shield which was under the bench near the east door. It had been turned on but with the "Counts" off during the accident. It was working and clicked at about 4 or 5 scales per second. The Super-Zoute on this same bench had been left on and was jammed at full-scale. I did not explore further because the room was too "warm" to stay long safely. I took Slotin's and my jackets back up the corridor as I left. I do not believe that I was in the room for more than a minute.
‍
No one else entered the room after the accident before we left. Werner (machinist at Pajarito) was going to close the outside doors after we left. He was warned not to go inside nor to stay in the vicinity long.
‍
P.F.C. Cleary (Security Guard) phoned for an emergency relief after the ambulance had arrived (about 15 minutes). We suggested that he ask for two men to come and he relayed by telephone our instructions that one should stay outside the building where he could watch the laboratory outside door and the other stay at the end of the corridor furthest from the hot lab. They were told to stay out of the room under all circumstances.
Incidentally, the boys leaving via the east door had warned the MP's so the gates were opened and all personnel had gone perhaps 50 yds. up the exit road. They remained there until we called them back about 15 minutes after the accident.
‍
The foregoing is my recollection of what happened. I repeat that my observations may not have been correct as to what happened during the few seconds following the accident.The following section is an attempt to give an impersonal analysis of the accident with the sole purpose of trying to analyze the causes and so help prevent its re-occurrence.1. Slotin was not, by any possible interpretation, guilty of what legal minds like to call "criminal negligence". The monitors were set up and running, an adequate number of observers were present who, by their silence, agreed to the procedure, and he had provided a safety device in the form of wedges to keep the tamper from dropping if it slipped. The fact that this safety device failed does not alter the situation as far as this point is concerned.
‍
I feel emphatically that there should never be an attempt to establish legal responsibility for the accident. If this should be attempted, it is my opinion that, excepting Perlman and Cleary, all should be held equally responsible, since the rest of us knew enough about critical assemblies to voice a protest if we objected to the procedure. No such protest was made.
‍
2. The assembly was made too rapidly and without adequate consideration of the details of the method. A "dry run" without the active material to check the details of the mechanical operations should have been made.
‍
3. Only those persons actually concerned with the assembly should have remained in the room.
‍
4. No conventional safety devices operating from a neutron monitor could have prevented the accident. The neutron rise was too rapid to have the reaction stopped by any of the devices I have seen used at Los Alamos.
‍
5. While operation of such an assembly by remote control would have eliminated the hazard from this particular "burst", it might well have become a real explosion with equal of greater damage to personnel in spite of shielding walls unless a positive and fast-acting safety device were a part of the assembly. In this case, Slotin was that safety device.I do not know what safety recommendations for future assemblies will evolve as a result of this accident. There are several points I should like to make.1. The formulation, approval and publishing of a new set of rules will not prevent more accidents. As stated in item (1) of the preceding section, the existing rules were nominally complied with. In addition to the restatement of rules, there must be a continuous and vigorous campaign to keep the people participating in this work aware of the potential danger in every assembly. Perhaps the work should be rotated among a number of people. As soon as a person ceases to be nervous about the work he should be transferred to another job.
‍
6. Any new assembly should be planned in detail a considerable time in advance. The plan should be sent to several responsible people, any one of which could veto the plan or ask for clarification.
‍
7. Every new assembly should be attended by one or more observers whose job would be to stop any procedure which they consider hazardous.
‍
8. Whenever feasible, an assembly should be done by remote control provided that this control incorporates safety devices guaranteed to stop the reaction in the shortest possible time.
‍
9. A complete account of each assembly should be kept, possibly with a running commentary fed into a wire recorder and either a movie camera or an automatic still camera.
‍
10. New critical assemblies should never be reduced to a routine matter to be "run through before lunch."
‍
11. A detailed file of all critical assemblies should be kept up to date. This file would be valuable as a guide in making future assemblies and could also be evidence for removing assemblies known to be safe from the rather severe restrictions which will undoubtedly be imposed on all untried assemblies.

Cc: Froman
File