How To Find Neutrons

Detecting Neutrons Actually Looks Like This

Neutrons don't ionize directly. That's the whole problem. They have no charge, so they pass through most detector material without leaving a trace. The trick is making them collide with something that does respond to their presence. I spent two weeks last year trying to set up a simple neutron flux measurement in a university basement lab using a homemade BF3 proportional counter. The manual said the tube was "potted and ready to use." It was not. The potted end had a micro-fracture from shipping, and I was measuring background counts that looked like signal for three days straight before I noticed the decay curve was wrong. Once I swapped in a fresh tube from a different batch, the baseline dropped from 42 counts per minute to about 3. The lesson here is that neutron counting always has noise, and sometimes the noise is the detector itself lying to you.

How To Find Neutrons in Practice

Start with the interaction you want to exploit. The two most common conversion reactions are the boron-10(n,alpha) lithium-7 reaction and the helium-3(n,p) tritium reaction. B-10 has a thermal neutron cross-section of about 3840 barns. He-3 is around 5330 barns at thermal energies. Both reactions produce charged particles that ionize the gas in your detector volume, and that ionization is what you actually measure. Fast neutrons are harder because they scatter elastically off protons or light nuclei before they thermalize enough for these cross-sections to become useful. If you're working with a reactor or an accelerator beam, you need moderation — usually a polyethylene or water surround — to slow the neutrons down before they hit your sensitive volume. Without moderation, your He-3 tube might see almost nothing from a modest source. Scintillation detectors are the other major path. Li-6 or B-10 loaded glass or crystal converts the neutron into a charged particle, and that particle excites the scintillator, which emits photons. A photomultiplier or silicon photomultiplier picks up the flash. These are faster than gas proportional counters and work well for time-of-flight measurements, but they suffer from gamma-ray pulse pileup. Gamma rays also produce signals in these detectors, and if your experiment has any X-ray or gamma background, you'll need pulse shape discrimination to separate the neutron events from the photon noise. PSD buys you maybe 100 to 1000 times better gamma rejection, depending on the scintillator and the electronics. Don't skip it. A detail people miss: neutron detection efficiency is energy-dependent in a way that's easy to underestimate. A He-3 tube that's 80 percent efficient for thermal neutrons might be less than 1 percent efficient for 1 MeV neutrons, even with moderation. The moderation geometry matters more than the tube size. A 2-inch tube buried in 10 cm of polyethylene around a Cf-252 source will often outperform a 4-inch tube sitting in the open, because the surrounding moderator captures neutrons that would otherwise leak past your detector. I learned this the hard way when I swapped to a bigger tube and my count rate actually went down.

If you need to build something from scratch, the simplest credible setup is a He-3 or BF3 proportional tube, a high-voltage supply in the 1000 to 1500 volt range, a preamp, a shaping amplifier, and a multichannel analyzer or simple counter. Wire the tube through a feedthrough insulator, keep the cable short between tube and preamp to reduce capacitive pickup, and shield everything with lead if gammas are present. The gamma sensitivity of a properly designed He-3 system is roughly 1 in 10,000 of the neutron sensitivity, but if your gamma dose rate is high enough, that 0.01 percent still adds up to measurable counts. Factor that in before you claim a detection limit. Alternative approaches exist when He-3 is unavailable, which it mostly is now because of global supply constraints after the US bought most of the world's stockpile. Boron-lined proportional counters are the closest substitute. B-10 enrichment at 96 percent gives you a thermal cross-section within 10 percent of He-3 per unit pressure, but BF3 gas is corrosive and requires careful material selection for the anode wire and tube wall. Aluminum or stainless steel walls work fine, but copper or brass components will degrade over time from fluorine attack. Solid-state alternatives like Li-6 glass scintillators paired with SiPM readout are becoming more common in commercial systems. They don't need high voltage and they're compact, but they have lower energy resolution than gas detectors and they drift with temperature unless you actively stabilize the gain. There's no magic solution that works in every scenario. If you're measuring reactor flux, use fission chambers — they're absolute detectors calibrated against known standards, and they don't need the same kind of empirical cross-checking that proportional counters do. If you're hunting stray neutrons from an am-Be source in a warehouse, a simple moderated GM tube or a commercial search meter will get you to within an order of magnitude, which is usually enough for radiation safety work. If you need spectral information — what energies the neutrons have — then time-of-flight with a fast scintillator and a pulsed source is the only real option, and it requires a pulsed beam and a flight path of at least several meters to resolve the energy range you care about.

The fundamental constraint of neutron detection is that you're always indirect. You detect the charged particle or the gamma ray produced by the neutron interaction, never the neutron itself. Every conversion reaction has a threshold, a branching ratio, and an energy-dependent cross-section that you need to convolve with your source spectrum to get an absolute rate. If you skip that calculation and just report counts per minute, you're reporting something real, but it's not the quantity anyone outside your specific setup cares about. Converting to fluence rate or dose equivalent requires knowing the neutron energy distribution, and if your source is unmoderated and broad-spectrum, that distribution is exactly what you're trying to measure in the first place. It's a loop, and the only way out is a Monte Carlo simulation or a benchmark measurement with a standard source of known spectrum. I ended up running my Cf-252 calibration against a NIST-traceable standard source and found my homemade system was reading 22 percent high. The correction came down to the polyethylene moderator being slightly denser than I'd assumed because of moisture absorption over the winter. Drying the blocks and re-running brought it into agreement within the stated uncertainty of the standard. That's the thing about neutron work — the environment is always part of the measurement, and sometimes it's the biggest part.