Guides / Radiation safety math

Half-value layers compared: lead, steel and concrete for Cs-137, Co-60 and Ir-192

Half-value and tenth-value layers of lead, steel and concrete for the common industrial gamma sources, by thickness and by weight, with a worked shield.

Practical guide

By Steven Michael Barber · Published October 17, 2026 · Updated October 25, 2026 · 6 min read

Our dose-rate guide introduces half-value and tenth-value layers with one isotope and one worked shield. This guide puts the three common industrial gamma sources and the three common shielding materials side by side, so you can see which material does the job in the least thickness, which does it in the least weight, and why the answer changes with the energy of the photons.

The layers, side by side

A half-value layer (HVL) is the thickness that cuts the dose rate in half; a tenth-value layer (TVL) cuts it to a tenth. The values below are the broad-beam figures from NCRP Report No. 49 as reproduced in health physics texts, the same table our calculator uses. Broad-beam means they include the photons that scatter inside a wide shield and still come out the far side, for the geometry and photon energies of those tables. Other geometries and spectra can differ.

Two patterns stand out. Reading down a column, cobalt-60 needs about twice the lead of cesium-137 for the same layer-based reduction, but only about 30 percent more steel or concrete (half-value layer ratios 1.85, 1.31 and 1.29 in this table). Higher-energy photons penetrate further, but by how much depends on the material. Reading across a row, the ratio between materials is not fixed: for cesium-137 concrete needs about 7.4 times the thickness of lead, for cobalt-60 only 5.2 times. Energy changes not just how much shield you need but which material is the better deal.

Thickness for a tenfold, hundredfold and thousandfold reduction

In the constant-layer model each extra tenth-value layer removes another factor of ten, so the thickness for a given reduction is just a multiple of the TVL:

A thousandfold reduction of cobalt-60 takes 12 centimeters of lead or two feet of concrete. That is why cobalt-60 irradiators live inside concrete vaults with walls measured in feet, while a cesium-137 gauge source travels inside a few centimeters of shielding in a box you can lift.

Thickness is not weight

Multiply each layer by the material's density and you get the mass per unit area, in grams per square centimeter, that the shield puts between you and the source. This number tells a different story from the thickness:

For cobalt-60 the three materials need almost the same mass per area: within about 20 percent, a kilogram of concrete shields as well as a kilogram of lead. At those energies photons lose energy mainly by Compton scattering off electrons, and every material has roughly the same number of electrons per gram, so what matters is how much matter is in the way, not what kind. Lead wins only on thickness.

For cesium-137 lead needs about 40 percent less mass than steel or concrete. At 662 keV the photoelectric effect, which grows steeply with atomic number, is already doing part of the work in lead and almost none in concrete. Below a few hundred kilo-electron-volts that advantage becomes large, which is why x-ray rooms are lined with thin lead rather than thick concrete.

In practical units, a one-square-meter panel one tenth-value layer thick for cesium-137 weighs about 238 kilograms in lead, 413 in steel and 369 in concrete, which is roughly 49, 85 and 76 pounds per square foot. None of them is light. Lead is the thinnest and, for this isotope, the lightest, but it is soft, needs a frame or a backing to hold it, and must be handled with care because it is toxic. Steel is structural and can be bolted and welded. Concrete is cheapest by the kilogram and holds itself up, which is why it is the material of walls.

Half-value or tenth-value: why the two models disagree

If attenuation were perfectly exponential, one tenth-value layer would be exactly log₂ 10, or 3.32, half-value layers. The tabulated ratios run from 3.23 (cesium-137 in lead) to 3.42 (iridium-192 in concrete), some below 3.32 and some above. Buildup, changes in the spectrum, the geometry and rounding in the table all contribute, so the gap alone does not show which layers are thicker. The two models therefore give slightly different answers for thin shields. Three half-value layers of lead against cesium-137 is 1.95 cm; the half-value model says that passes 12.5 percent, the tenth-value model 11.8 percent. The calculator defaults to the tenth-value model. Both apply one constant layer to the whole thickness, and their agreement is not a check on either. For iridium-192, with several photon lines, the spectrum hardens as it passes through the shield, so a constant layer can misstate the attenuation, most of all in a thin shield.

A second check comes from the narrow-beam data in the NIST attenuation tables, which exclude scatter entirely. From the mass attenuation coefficients, the narrow-beam half-value layer for cesium-137 photons is about 0.56 cm in lead and 3.8 cm in concrete, against the broad-beam 0.65 and 4.8 cm in the table above. The broad-beam figures are 15 to 25 percent thicker, consistent with scattered photons adding to what comes through a wide shield. Do not use narrow-beam values in place of broad-beam ones. A reference with a thinner half-value layer than ours may be a narrow-beam value or may assume a different geometry or spectrum, so check its conditions before using it.

Worked example: a cobalt-60 check source

A 1 curie cobalt-60 calibration source has to be stored so that the rate at 1 meter from it is no more than 2 mR/h. Unshielded, the rate at 1 meter is Γ × A = 1.32 × 1 = 1.32 R/h, or 1,320 mR/h. The shield has to pass 2 ÷ 1,320 = 0.0015 of the radiation, which is 2.82 tenth-value layers:

The half-value model gives 11.2 cm of lead, the same within a millimeter. Whichever material is chosen, the shield weighs well over a tonne per square meter of face, so a source of this size is stored in a pit or a purpose-built cask, not a cabinet. The same target for a 100 millicurie cesium-137 source, with an unshielded 33 mR/h at 1 meter, needs 2.56 cm of lead, 6.45 cm of steel or 19.1 cm of concrete, which is a cabinet-sized problem.

What the tables leave out

Iridium-192 emits a spread of photon energies, and the layers for it depend on how much of the soft component the source capsule has already removed; the steel values here are the ones used in industrial radiography practice, and references differ by 10 to 20 percent. Tungsten and depleted uranium, the materials of radiography cameras and transport containers, are denser than lead and stop photons in less thickness still; they are not in these tables because the calculator does not model them. Layers are for a shield that fully covers the line between source and person; gaps, seams, cable penetrations and the open top of a lead-brick enclosure let radiation stream through at the unshielded rate. And the dose rate behind a shield in a room includes scatter from the walls that no layer table accounts for. Use the tables to size a shield and to understand a survey reading; confirm the result with a calibrated instrument, and have a qualified person sign off on any shield that people depend on.