Understanding radiation units
Gray, sievert, rad, rem — what each one means.
The four units of radiation exposure, the absorbed-dose vs effective-dose distinction, and the typical background levels.
Absorbed dose vs effective dose.
Two different things. Absorbed dose is the energy deposited per kg of tissue (joules per kilogram). SI unit gray (Gy); legacy unit rad (1 Gy = 100 rad). Effective dose weights the absorbed dose by the biological harm potential — alpha particles damage tissue 20× more per gray than X-rays do, so the weighting matters. SI unit sievert (Sv); legacy unit rem (1 Sv = 100 rem). For photons and beta radiation, gray and sievert are numerically the same.
The scale.
Background radiation (cosmic + terrestrial + radon): 2-3 mSv/year typical, 5-10 mSv in some areas. Chest X-ray: 0.1 mSv. CT scan: 5-15 mSv depending on body area. Long-distance flight: ~0.05 mSv. Banana (potassium-40): 0.0001 mSv. Annual occupational limit for nuclear workers: 50 mSv. Acute radiation syndrome threshold: ~1 Sv. Almost always lethal within weeks: 6 Sv. The range from background to lethal is roughly 1000×.
Dose rate vs total dose.
The rate at which dose is delivered matters as much as the total. 1 Sv delivered over a year is dramatically less harmful than 1 Sv in a minute — the body has repair mechanisms that work on the slower timescale. Geiger counters measure dose rate (μSv/h); occupational monitoring tracks accumulated dose over time. For most contexts the two together tell the story.
Becquerel and curie — radioactivity.
A different question: how active is a source? Becquerel (Bq) = 1 disintegration per second. Curie (Ci, the older unit, used in medicine) = 3.7 × 10¹⁰ Bq. These measure the source; dose units measure what reaches the receiver. A source's activity in Bq combined with distance and shielding determines the dose rate at a given point.
The linear no-threshold model.
Regulatory radiation protection assumes the risk of cancer scales linearly with dose — there's no "safe" threshold below which dose is zero risk. The model is debated for low doses (some researchers argue for hormesis; others for a steeper response than linear), but it's the conservative basis of every dose limit. Practically: every medical-imaging decision balances the diagnostic benefit against the dose-driven probabilistic risk.
Conversion in practice.
The most-needed conversions: rad → Gy ÷ 100; rem → Sv ÷ 100; mrem → μSv × 10. US medical contexts still use mrem and rad frequently; international contexts use SI. A dosimeter calibrated in mrem can be read in μSv with multiplication by 10; a CT scan reported as 1500 mrem is 15 mSv. Convert once, work in either.