Seven Days Before the Charcoal Tells Me What the Soil Exhaled

You know how sometimes you own equipment that you’ve never quite understood what it measures? The RDF setup that failed to recover any meteorites has a Geiger counter as backup—basic pancake detector, clicks when it sees ionizing radiation, mostly useless for direction-finding but good for confirming you’re near something radioactive. Never paid it much attention beyond the calibration checks.

Started wondering about that drawer of mineral specimens—some of them are uranium minerals, weakly radioactive but detectable. Carnotite specimens from Utah, a piece of pitchblende that was probably a terrible purchase given how little it cost. Put the Geiger counter near them. Click-click-click-click, 40 counts per minute. Background radiation here is about 15. The difference is radon.

Radon-222, specifically. It’s a noble gas—chemically inert, doesn’t bond to anything—that comes from radium decay, which comes from uranium decay, which exists in trace amounts in pretty much all rock. Most of it stays trapped underground, but some fraction percolates through soil and escapes into air. The half-life is 3.8215 days. Not 3.8 days, not four days—3.8215. That tenth of a day matters when you’re trying to account for decay during a week-long measurement.

Radon is what home inspectors test for before you buy a house, because concentrations above 4 picocuries per liter increase lung cancer risk. But outdoor concentrations vary with geology—granite areas produce more, sedimentary basins produce less, fault zones sometimes create linear anomalies where radon migrates from depth. You can map it. Deploy detectors on a grid, retrieve them after 90 days, send them to a lab, get back concentration values, plot contours. Invisible landscape features made visible through systematic measurement. That’s exactly the satisfaction from theodolite triangulation—you’re creating data where none existed, revealing structure through precision.

So I ordered passive detectors. Charcoal canisters—$22 each, you expose them for 2-7 days and the activated charcoal adsorbs radon which then decays and leaves alpha particle tracks the lab counts. Eight detectors, weatherproofed in ziplock bags with desiccant, deployed in a 200-metre grid around the property. GPS coordinates logged for each one. Retrieval date calculated accounting for the half-life: if I deploy on Day 0 and retrieve on Day 7, the detector is measuring radon that’s been decaying the entire time. The concentration I calculate has to correct for that, or the numbers are wrong by nearly a factor of four.

There’s also a Lucas cell sitting on the desk. Bought it used from a university surplus auction—1965 vintage, brass cylinder about the size of a thermos, interior coated with zinc sulfide doped with silver. When radon’s alpha particles hit that coating, it produces tiny flashes of light. A photomultiplier tube at the top counts the photons. You’re measuring radioactive decay by watching for light bursts in a sealed, darkened chamber. The connection to amateur radio isn’t obvious until you realize both are about detecting electromagnetic emissions—RF waves in one case, visible photons from particle collisions in the other. Same core skill: discriminating signal from noise, calibrating sensitivity, logging measurements over time.

Here’s the strange thing about radon nomenclature. Until 1957, “radon” only referred to Rn-222. The other naturally occurring isotopes had different names: Rn-220 from thorium was “thoron,” Rn-219 from actinium was “actinon.” IUPAC decided to make “radon” the element name instead of just one isotope, which made sense chemically but erased historical distinctions. It was controversial—people thought it gave undue credit to Friedrich Dorn’s identification of Rn-222 over Rutherford’s identification of Rn-220. Reading pre-1957 papers about “radon emanation” means specifically the 3.82-day isotope, not the element broadly. That kind of terminological archaeology matters when you’re trying to understand measurement techniques from the 1920s.

The unit is picocuries per liter—trillionths of the radioactivity Marie Curie isolated from pitchblende in 1898. A picocurie is 10⁻¹² curies. You’re measuring almost nothing, but that almost-nothing accumulates over years of exposure. The EPA action level for indoor air is 4 pCi/L. WHO recommends 2.7 pCi/L. Outdoor soil gas concentrations can hit 20-50 pCi/L over granite. The conversion to SI units is 1 pCi/L = 37 Bq/m³, becquerels per cubic metre. Both measure decay events, just different denominators.

Weather ruins short-term measurements. Radon concentration fluctuates hourly based on barometric pressure (low pressure draws more from soil), wind (disperses it), temperature gradients (affects diffusion rates), soil moisture (blocks gas migration). A two-day test might catch a weather system that gives you a false high reading. EPA recommends 90+ days for accurate assessment. The challenge is that you’re characterizing something that changes continuously using a substance that decays in less than four days. It’s like trying to measure ocean currents with drifting buoys that dissolve in 96 hours—the measurement window is narrow, the signal is noisy, and timing precision becomes critical.

The first deployment is running now. Eight canisters, buried 15 cm deep in PVC tubes with screened caps to let soil gas in but keep precipitation out. Coordinates logged. Retrieval scheduled for 7 days from now, which gives enough accumulation without excessive decay loss. The Lucas cell is for spot-checking individual locations—pump soil gas through it for 20 minutes, seal it, wait 3 hours for equilibrium, then count for 5 minutes. Faster turnaround but less spatial coverage.

What makes this interesting is that the mineral specimens are producing measurable radon right now, continuously, predictably. Uranium-238 decays to radium-226 with a 1600-year half-life, radium decays to radon-222 with a 3.82-day half-life, radon decays to polonium-218 and continues down the chain until it hits stable lead-206. The rocks are laboratories—controlled sources with known decay rates. Measure them first, verify the instruments work, then deploy for terrain mapping where geology varies and concentrations are unknown.

Field deployment protocol is borrowed from benchmark documentation: GPS coordinates, photographs showing context, notebook entries with exact timing. But instead of photographing a brass disk, it’s a white PVC cap flush with soil surface next to a surveyor’s flag. Instead of vernier readings from a theodolite, it’s decay counts from charcoal canisters. The underlying discipline is identical—systematic measurement, precise logging, accounting for error sources, building spatial datasets that reveal structure.

The grid fills in over the next week. Then lab analysis. Then a map showing something completely invisible: radioactive contours across terrain, granite showing high, alluvial valleys showing low, maybe fault zones creating linear features no one could see without 90 days of passive detection. Making the invisible visible. That’s the whole hobby.