Construction Classics

Targets and the target plane

A cyclotron exists to put a beam into a target, and these three reports cover both halves of that sentence. One is a bench recipe for making self-supporting films a few tens of micrograms per square centimetre thick. The other two ask what the beam actually is when it arrives — its energy, its energy spread, and where on the target it lands — and answer with aluminium foil, photographic film and activated carbon strips.

Why these documents can be hosted

All three are Oak Ridge National Laboratory reports under AEC contract W-7405-eng-26 — ORNL-3021 operated by Union Carbide Corporation, the two 1952 reports by Carbide and Carbon Chemicals Company — and AEC contractor reports are treated as US Government works, which carry no copyright. Pre-1978 distribution without a copyright notice is an independent basis. ORNL-3021 additionally states Printed in USA. Price $0.50. Available from the Office of Technical Services, Department of Commerce, making it a priced US Government publication. ORNL-1347 and ORNL-1348 are marked Unclassified throughout with their AEC distribution lists bound in.

One check worth recording: ORNL-1348’s reference list cites an article by its own author submitted to Review of Scientific Instruments. That is a citation to a separate paper, not a marking on this report — the hosted artifact is the ORNL-numbered laboratory report, with its own cover, contract number and distribution list. Any journal version that was subsequently published is separately copyrighted and is not offered here. Verified against the scans and the OSTI records; determinations made August 2026.

Making a target that is only its own material

ORNL-3021 is eleven pages and reads like a workshop notebook, which is exactly its value. The problem it solves is that a backing foil ruins a measurement: a very thin, backing-free target improves the probability that all reaction products escape the target at all, reduces the energy spread caused by straggling inside it, and cuts plural and multiple scattering. The authors had already learned to make self-supporting carbon films at 25–100 µg/cm² and set out to do the same for boron-10 and magnesium-24.

Boron (p. 6) is evaporated from a covered boat machined from 5/16-inch spectroscopic-grade carbon rod — covered specifically to reduce radiation cooling and to stop the charge scattering as it heats, with a dimensioned drawing given. The cavity takes about 250 mg of amorphous boron; pressure is held below 10⁻⁵ mm Hg while the boat is brought up; about 280 A at 8 V takes it to white heat. The visual cue is precise and useful: the boron stays darker than the boat until just before vaporisation, then brightens and disappears quickly. A boat is good for at most two runs before the slot clogs with boron carbide.

Recovery is the fiddly part and gets the most attention. Films are collected on 3¼ × 4-inch glass plates washed in Calgonite, then floated off with warm tap water with the plate at about 20° to horizontal. Single-thickness films are fragile and difficult; double films are much easier, made by bringing the target frame up through the middle of the floating film so it folds over and adheres to itself, on square aluminium frames 0.030 inch thick. Breakage drops if a pinch of detergent is added to cut surface tension just before pickup, and if the frame is paused just below the surface to let trapped water drain. Double-thickness enriched B-10 targets came out at 50–100 µg/cm².

Magnesium needs more care because the source material is an oxide and the metal reacts with water. MgO is reduced in situ on a tantalum boat — a shiny metallic film appears on the collector at about 1200 °C, well below the melting point, as the MgO reduces to tantalum oxide and metallic magnesium — with the honest caveat that this risks contaminating the target with oxygen and with tantalum from Ta₂O₅ decomposition at 1470 °C. Evaporation uses 3-mil tantalum boats about 3/4 inch wide, with the current raised slowly over 10–20 minutes so that outgassing happens quietly rather than scattering the charge. The glass is coated with weak Calgonite, dried, then given 5–10 µg/cm² of arc-evaporated carbon, which both keeps the magnesium off the water — magnesium is slightly soluble and decomposes to Mg(OH)₂ — and holds the film together during and after removal. Floating starts warm and is then cooled to slow any reaction through pinholes in the carbon. Single films of 30–80 µg/cm² resulted, and because enriched Mg-24 is expensive a hood built from lantern slide covers caught and returned the material that missed.

Boron targetry was a named gap in this site’s source literature until this document was mined, and it is a complete, costed, reproducible answer to it — down to the detergent and the angle of the plate. Note what it is not: it is evaporation, not sputtering, and it produces microgram-per-square-centimetre research targets, not the thick durable targets a production irradiation would need.

That other half is now covered too, though not from this shelf. The International Nuclear Target Development Society published its conference proceedings as national-laboratory reports between 1975 and 1983, and those volumes carry what evaporation recipes cannot: rolling, sputtering, electrodeposition, chemical vapour deposition, target cooling, and the beam-heating calculation that decides how much current a foil will take before it fails. They are cited throughout the Design Guide rather than hosted here — they are international conferences, and the foreign-authored contributions are not cleared for republication — but all four are free from OSTI, and between them they took the guide’s targets domain from twelve rules to more than a hundred.

Finding out what the beam is when it lands

The two 1952 reports are a matched pair by the same author on the ORNL 86-inch, and they open with an admission that should stop any builder who has calculated a beam energy and believed it.

Because most fixed-frequency cyclotrons had deflection apparatus and worked with an external beam, nobody had much needed to know an internal beam’s energy precisely; experimenters were content with a value calculated from the magnetic field and the radius of curvature, with the centre of curvature assumed to be at the ion source. Measurements on the 86-inch showed early that the proton energy might vary as much as ±10% from that calculation. Worse, the energy of maximum intensity moved by several hundred kilovolts under minor adjustments of ion-source position, dee voltage, magnetic-field tuning and oscillator frequency. A cyclotron’s beam energy is not a property you can compute once and write on the machine.

ORNL-1347 gives two absorption methods for measuring it. The photographic one is the cheaper and is well within amateur reach: a film is covered with a stepped absorber made by folding an aluminium foil, wrapped in aluminium, and exposed directly in the beam, with an uncovered strip left at the top to correct for intensity variation across the film. Beam current is monitored by a neutron counter near the target and can be brought down to 0.1 µA with the ion-source arc off — though holding currents that low takes considerable magnetic detuning — and an exposure of about 0.1 µA‑second suits the film stock used. Development density behind each step is read on a densitometer and plotted as the antilog of density against absorber thickness, with a second abscissa converting range to energy; the theoretical curve for any other energy is the same curve slid horizontally. The second method measures the excitation function for the (p,n) reaction on copper.

The report is candid about what these methods cannot do. Both are absorption methods, so both are limited by range-energy data and straggling, which caps accuracy on the most probable energy at a few hundred kilovolts. The high-energy side of the distribution comes out with nearly the same accuracy; the low-energy side is much less certain, and where the high-energy side falls slowly the curve overstates the true spectrum. Stated limits of method, in 1952, on the first page of the results.

ORNL-1348 answers the companion question — not how energetic the beam is, but where it lands. A target was built to hold seventeen carbon foils, 1/32 inch thick and 5½ inches long; after a fifteen-minute bombardment at 15 µA each foil was broken along pre-scored lines into half-inch sections and counted, mapping the beam across the target. The interpretation is the interesting part: the measured distributions are read back as RF phase and radial grouping, giving the phase of the main part of the accelerated group under different tuning conditions, in agreement with orbit theory. An activated carbon strip, broken into pieces and counted, tells you what phase your beam is at.

Why these three belong together

The target plane is where a cyclotron stops being an engineering project and starts being an instrument, and it is the part amateur builds reach last and document least. These reports cover it end to end at bench scale: how to make a target thin enough not to corrupt the measurement, how to find the energy and spread of the beam hitting it, and how to map that beam across its face. The techniques need aluminium foil, film, carbon strips, an evaporator and patience — no part of this requires a national laboratory, which is why it transfers.

For the beam measured on the way in rather than at the target, see beam measurement and the probe-versus-calorimeter cross-check in UCRL-3187.

Design rules extracted from these documents

15 Design Guide rules come from these three reports — ORNL-3021, ORNL-1347 and ORNL-1348. They are the first substantial entries the targets domain has had, and the 1952 pair supplies the caution that a calculated beam energy is not a measured one. Representative examples:

• dg-881 — never quote an internal-target beam energy from the Bρ calculation alone
• dg-882 — measure internal-beam energy with film behind a stepped absorber folded from aluminium foil
• dg-883 — know the accuracy floor of any absorber-based energy measurement
• dg-885 — distrust diagnostics taken with the machine deliberately detuned to reach a workable intensity
• dg-886 — map the beam by sectioning the target itself into pre-scored strips
• dg-887 — the turn-to-turn radial step at the target edge is a direct RF-phase meter
• dg-877 — budget evaporation boats as consumables — two runs before the slot clogs

Used by the beam measurement deep dive.