Construction Classics

Residual radiation and personnel dose

Nearly every cyclotron report describes a machine while the beam is on. These two describe what is left when it is off — which nuclides the structure has become, how fast they decay, and what it costs in dose to walk up to the target and change it. Both are measurements, not estimates.

Why these documents can be hosted

ORNL-3158 is an Oak Ridge National Laboratory report prepared by Union Carbide Corporation for the US Atomic Energy Commission; UCRL-8276 is a Crocker Laboratory report printed for the AEC and carrying the standard AEC legal notice. AEC contractor reports are treated as US Government works and carry no copyright; pre-1978 distribution without a copyright notice is an independent public-domain basis. Neither scan carries a copyright notice, publisher imprint, or journal-reprint marking, and neither carries classification markings. Verified against the scans and the OSTI records; determinations made August 2026.

What the machine becomes

ORNL-3158 exists because Oak Ridge was considering building an 850 MeV proton cyclotron with beams of order 100 µA — the programme whose design volume this collection hosts as ORNL-3540 — and wanted to know what it would be signing up for. Rather than calculate, a group went to Berkeley and measured an existing machine during a five-day shutdown in November 1960, reasoning that a factor of about 100 in residual activity separated the two machines.

The method is reproducible at any scale and is the most transferable thing in the report. Three lines of attack: a radiation monitor surveying the vault and the machine, from which a decay curve was built; foils of iron, stainless steel, copper and aluminium affixed near the cyclotron and counted afterwards; and a 3 × 3 inch NaI crystal in a large lead collimator, pointed at the magnet gap near the probe position, feeding a 256-channel pulse-height analyser, with the spectra followed for five days. Dose survey, foil activation, and gamma spectroscopy — one tells you how much, one tells you what, and one tells you where.

The results are material-specific, which is the point. Copper foils showed Cu-64 and Co-58; iron foils showed Mn-56 (2.6-hour) and Mn-52 (5.7-day); aluminium showed Na-24. Longer term the report is precise about the alloys: stainless steel produces two long-lived isotopes, Cr-51 (27-day) and Mn-54 (300-day), where plain iron produces only the Mn-54 — so the chromium is what distinguishes them, and the nickel in stainless brings 71-day Co-58 as well. The marker nuclide follows the alloying element, which is the whole argument for treating material choice as a radiological decision.

Decay rates matter as much as identities. Aluminium’s only product is 15-hour Na-24 — effectively gone in days — while the Mn-54 in stainless runs 300 days. A material can be the right choice not because it activates less but because what it becomes dies sooner.

Two techniques in the foil work are worth lifting whole. Samples in the median plane carried more Mn-52 than those above it, as expected from where the beam is — so foil position is data, and positions must be mapped before the run. And wrapping one of a matched foil pair in cadmium splits the induced activity into a slow-neutron capture component and a fast-particle component, which is a cheap way to learn not just what activated but by what. The gamma spectra from the gap region — dee, dee support, liner and pole pieces all contributing — held two intense long-lived peaks at 510 and 810 keV, which the abstract attributes principally to Co-58, though the report itself weighs a contribution to the upper peak from the 830 keV line of Mn-54. Initial radiation levels were under 8 r/hr and fell to about 10 mr/hr after 48 hours, which is the single most operationally useful number in the report: the short-lived component dominates, and waiting two days is most of the protection.

The conclusions section (p. 17) names three ways to keep the problem manageable, and the second is a design rule rather than an operating one: control the beam so that high-energy particles are not lost in the first place; choose construction materials to minimise reactions leading to long-lived gamma emitters; and design for remote maintenance and gamma shielding from the outset. Material selection as a radiological decision, made at the drawing stage, is a habit worth forming even where the specific nuclides here do not apply — the corollary being that the right material list is the one derived from your own machine’s energies.

The shielding argument in the same section is the one most often got backwards elsewhere: size the shield from the measured line energies rather than from a worst case. For the 510–810 keV residual lines the lead half-thickness is about 0.6 cm, so the required thickness is modest once the spectrum is known — and unknowable, hence over-specified, when it is not. Measure first, then shield.

What it costs the people

UCRL-8276 is a different kind of document: five years of film-badge readings, 1953 to 1957, for the men who actually worked on the 60-inch, plus a survey of the radiation environment in the target area. It is the only document in this collection that treats operating a cyclotron as an occupational-health problem with data attached.

The dose-rate picture is built around a wide-range beta-gamma meter placed about three feet from the target plate, on the same level as the target, standing in for the crew member’s position. Alpha-particle bombardments accounted for 73% of all bombardments over the period, and during an average one the meter read 10–20 roentgens per hour. One minute after the beam stopped it had fallen to about 1.0 r/hr; five minutes after, with the target removed, about 0.3 r/hr; after ten minutes, still slightly over 0.2 r/hr.

The targets themselves are the hazard. Targets and their assemblies normally emitted 100 to 500 r/hr five minutes after bombardment, and an internal target — one inserted into the tank on a probe — may emit more than 10,000 r/hr. That last figure is the reason the operational discipline in this report matters more than its dosimetry: the object you most want to pick up and look at is the most active thing in the room.

So the procedures are timed to the minute. Setting up a target takes about three minutes, removing the assembly about one, dismantling less than one — done behind a shield with a 2-inch lead-glass front and a 2-inch lead base, sides deliberately left open for tool access. Transuranic work uses a dedicated cart supplying the target with cooling water, air and helium, about five minutes to set up and under a minute to remove. Targets go into shielded containers immediately. The whole design is time-and-distance discipline made routine, which is the part that transfers to any machine that activates anything at all.

One finding deserves separate billing because it changes what you monitor: in target-handling work the hands take roughly ten times the whole-body dose. Film badges on the chest measure the wrong thing for this task. Extremity monitoring, and tools that add a few inches of reach, do more than any amount of shielding at the console. The report also quarantines the operations that activate the machine hardest into scheduled windows — at Crocker, deuteron runs at the end of the week, with a mandated wait before anyone approached the target.

Table I (p. 7) is worth a look for a different reason: it compares badge readings across eight accelerator sites of the era — Carnegie Institute of Technology, Columbia, MIT, Pittsburgh, Princeton, Rochester, Yale and California — with particle, energy, badge count and exposure distribution. It is a rare cross-institution dataset, and it shows the spread between laboratories doing nominally similar work.

The report’s dose limits are those of 1958 — a maximum permissible accumulated dose of 5 rem times the number of years past age eighteen, with no more than 3 rem in any thirteen consecutive weeks. Those are historical figures, superseded many times over, and are quoted here only because the report’s conclusions are stated against them. Current limits are far lower; use your jurisdiction’s, not these.

Design rules extracted from these documents

11 Design Guide rules come from these two reports — ORNL-3158 and UCRL-8276 — each carrying its energy scope with it so that no rule reads as a prediction for a machine that cannot drive the reactions behind it. Representative examples:

• dg-864 — treat construction-material choice as a radiological decision made at the drawing board
• dg-865 — different structural metals leave different residual inventories under identical irradiation
• dg-866 — hang cheap witness foils of candidate materials at mapped positions before a run
• dg-867 — wrap one of a matched foil pair in cadmium to split slow-neutron from fast-particle activation
• dg-868 — size gamma shielding from the measured line energies, not from worst case
• dg-871 — treat handling time as the primary dose control, and choreograph it
• dg-872 — in target handling the hands take roughly ten times the whole-body dose

Used by the shielding and safety deep dives.