A radiation-therapy vault is a building inside a building: a heavily reinforced concrete shell whose walls, roof, and floor are sized not by structure alone but by a qualified medical physicist's shielding calculation. This article covers the physics and construction of the shielding envelope itself — concrete mass and density control, labyrinth (maze) attenuation, and the disciplined detailing of every penetration that breaks the barrier — for linear-accelerator (LINAC) vaults and related high-energy treatment rooms.

The companion Article on overall room layout, equipment program, and the patient-facing aspects of the vault is handled separately; here the focus is the shielding barrier as an engineered system and the constructability problems it creates.

The shielding design is owned by a qualified medical physicist, not the architect

Vault shielding is one of the few building elements in healthcare where the governing calculation is performed by a licensed clinical professional rather than the design engineer of record. A qualified medical physicist (QMP) prepares a stamped shielding report that drives wall thicknesses, barrier locations, maze geometry, and door requirements. The architect and structural engineer then build the physicist's numbers into the documents.

• The methodology follows the National Council on Radiation Protection and Measurements (NCRP) structural-shielding reports — NCRP Report No. 151 for megavoltage external-beam therapy is the standard reference, with NCRP Report No. 49 still cited for older/lower-energy contexts and brachytherapy treated under separate guidance.
• Inputs the physicist needs early: machine make/model and maximum beam energy (commonly 6 MV through 18–20+ MV; energies at and above roughly 10 MV introduce neutron production), workload (W, integral dose delivered at isocenter per week), use factor (U, fraction of time the beam points at a given barrier), and occupancy factor (T, how continuously the adjacent space is occupied).
• Design dose limits derive from regulatory and consensus criteria — typically a controlled-area limit for staff and a far lower public/uncontrolled-area limit, with the physicist targeting a conservative weekly design goal at each barrier.
• The report distinguishes primary barriers (directly struck by the useful beam) from secondary barriers (exposed only to leakage and scatter). Primary barriers are thickest; the primary beam can be aimed at the floor, the side walls, and the ceiling, so the primary zone often wraps a large portion of the enclosure.

Because the report is the legal basis for the build, late changes to machine energy, vendor, or room use force a re-run of the calculation — a common and expensive source of rework. Lock the machine selection before the shielding documents are finalized.

Concrete mass — density, thickness, and pour discipline — is the primary shield

For most vaults the shield is cast-in-place normal-weight concrete, valued for being structural, monolithic, durable, and relatively economical per unit of attenuation. The physics is straightforward: attenuation scales with mass per unit area (areal density), so thickness and density are interchangeable trade-offs the physicist manages.

• Density is a specified, verified property, not an assumption. Shielding calculations assume a minimum concrete density (normal-weight concrete is typically taken near 2.35 g/cm³, often expressed as ~147 lb/ft³). The structural specifications must call out and the field must verify the as-placed density, because a lighter mix silently under-shields the barrier. Density confirmation (e.g., unit-weight testing of the delivered mix) belongs in the QA plan.
• Primary-barrier thicknesses for high-energy LINACs are commonly on the order of several feet of normal-weight concrete (rule-of-thumb figures of roughly 5–8 ft are typical for primary walls at the highest energies; secondary barriers are thinner). Treat all perspectives as physicist-derived, never assumed.
• High-density alternatives trade cost for space. High-density (heavyweight) concrete using heavy aggregate (e.g., magnetite, barite, or steel-aggregate mixes) reaches densities well above normal-weight, shrinking wall thickness where floor area or structural reach is constrained. Steel and lead are used as space-saving laminates or for localized barriers, but high-energy beams make pure-metal shields problematic for neutrons (see below), so concrete usually remains the bulk medium.
• Pour continuity matters. Voids, honeycombing, cold joints, and segregation create low-density paths that leak radiation. Best practice favors monolithic pours with controlled lift heights, consolidated placement, and staggered/offset construction joints so no joint forms a straight radiation path through the barrier. Form-tie holes and any through-form hardware must be detailed so they do not become a void path.
• Reinforcing congestion is severe. Thick barriers carry heavy rebar mats; coordination of rebar with embedded conduits, blockouts, and waveguide/sleeve penetrations must be resolved before the pour, because there is no fixing it afterward.

The structural engineer must also reconcile the shield's enormous dead load with the building — vaults are frequently slab-on-grade or in below-grade levels precisely so the floor and primary-down barrier sit on soil rather than on a suspended structure. (Structural load coordination is treated in the companion systems Article.)

The labyrinth (maze) trades door mass for geometry

A direct doorway through a primary or high-secondary barrier would require an impractically massive shielded door. The labyrinth, or maze, is the standard solution: an angled corridor that lets people and equipment enter while forcing radiation to scatter off multiple surfaces before it can reach the entrance. Each scatter event sheds a large fraction of the energy, so the maze does most of the shielding work and the entrance door can be far lighter — sometimes eliminated entirely at lower energies.

• Geometry is engineered, not aesthetic. The number of legs (bends), the length of each leg, the cross-sectional area of the maze, and the angle of the inner maze wall relative to the beam are all calculated. Longer, more-turned mazes attenuate more but consume floor area; the physicist balances maze length against door weight.