Therapy Maze Design, Photoneutrons & Shielded Doors

Therapy Maze Design, Neutron Shielding & PET/CT Facilities

Designing radiation barriers for modern medical imaging and radiotherapy requires engineering solutions for two demanding radiation environments: high-energy megavoltage radiotherapy vaults ($> 10\text{ MV}$) with secondary photoneutrons and capture gammas, and nuclear medicine Positron Emission Tomography/Computed Tomography (PET/CT) suites emitting penetrating $511\text{ keV}$ annihilation radiation. In external beam radiotherapy vaults, entry maze configurations drastically reduce the radiation burden on access doors by exploiting Compton scattering kinematics. However, above $10\text{ MV}$, photonuclear interactions introduce fast and thermal neutrons alongside capture gamma rays, mandating composite hydrogenous and high-$Z$ door construction. In PET/CT suites, mobile patient sources emit monoenergetic annihilation photons requiring dedicated structural barriers governed by AAPM Task Group 108.

---

Maze Geometry & Multiple Compton Scattering Dynamics

A radiotherapy vault access portal can be designed as either a direct-shielded door or a maze-entry layout:

• Direct-Shielded Vault: The door is positioned directly in line of sight of the treatment room, exposed to direct head leakage and primary/patient scatter. Such doors must stop much more penetrating radiation than a maze door and can weigh many tonnes, which brings slow motorized operation, maintenance demands, and entrapment concerns; NCRP 151 also notes the problem of neutron capture gamma rays produced in the door itself.
• Maze-Entry Vault: A baffled hallway (single-turn or double-turn maze) connects the treatment room to the door. Radiation reaching the door must undergo at least two or three successive scatters from concrete walls.

Kinematics of Compton Energy Degradation

The physical principle underlying maze design is the fundamental angular dependence of Compton scattering:

$E' = \frac{E}{1 + \frac{E}{m_e c^2}(1 - \cos\theta)}$

where $m_e c^2 = 0.511\text{ MeV}$ is the electron rest mass and $\theta$ is the scattering angle.

At a scattering angle of $\theta = 90^\circ$ ($1 - \cos 90^\circ = 1$):

$E' = \frac{E}{1 + \frac{E}{0.511\text{ MeV}}}$

As primary photon energy $E \to \infty$, the maximum scattered photon energy at $90^\circ$ asymptotically approaches $511\text{ keV}$, regardless of whether the initial beam was $6\text{ MV}$ or $18\text{ MV}$. When these photons undergo a second $90^\circ$ scatter along the maze wall, their maximum energy drops to about $255\text{ keV}$:

$E'' = \frac{0.511\text{ MeV}}{1 + \frac{0.511}{0.511}(1)} = \frac{0.511\text{ MeV}}{2} = 0.255\text{ MeV} = 255\text{ keV}$

At these degraded energies the lead TVL falls from $5.7\text{ cm}$ for the primary beam to a few millimetres. Consequently, a vault for a low-energy ($\le 10\text{ MV}$) linac with an adequate maze usually needs only a modest lead-lined door, sized from the maze calculation below.

Total Photon Dose at the Maze Door

For low-energy accelerators NCRP Report 151 evaluates the photon dose equivalent at the door with the beam pointed at the wall that faces the maze (gantry orientation $G$):

$H_G = f\,H_S + H_{LS} + H_{ps} + H_{LT}$

1. $H_S$ (wall-scattered primary): the primary beam strikes the wall facing the maze and scatters down the corridor; $f$ is the fraction of the primary beam transmitted through the patient.
2. $H_{LS}$ (leakage scatter): head leakage scattered from room surfaces into the maze.
3. $H_{ps}$ (patient scatter): radiation scattered by the patient toward the maze and then off the maze walls.
4. $H_{LT}$ (leakage transmitted): head leakage that passes straight through the maze wall.

Each scatter term uses wall reflection coefficients, the illuminated areas, and inverse-square distances. To cover all gantry angles, NCRP 151 (after McGinley) estimates the weekly door dose as $H_{\text{Tot}} = 2.64\,H_G$, an approximation to be used with caution when the room geometry differs from the reference design.

Photoneutron Production & Capture Gamma Rays in High-Energy Linacs (> 10 MV)

When linear accelerators operate at accelerating potentials exceeding $10\text{ MV}$ (e.g., $15\text{ MV}$, $18\text{ MV}$, or $20\text{ MV}$), the energy of bremsstrahlung photons exceeds the nuclear binding energy of nucleons in heavy elements. This triggers photonuclear interactions via the Giant Dipole Resonance (GDR).

Physics of the Giant Dipole Resonance

In the GDR process, the electric field of a high-energy photon induces a collective, resonant dipole oscillation of all protons against all neutrons within the nucleus. The nucleus de-excites primarily through the emission of one or more neutrons: $(\gamma, n)$ and electro-disintegration $(e, e'n)$ reactions.

• Reaction Thresholds: Photoneutron production thresholds depend strongly on atomic number $Z$:
  • Tungsten ($^{184}\text{W}$, target, jaws, MLC): $E_{\text{th}} \approx 7.4\text{ MeV}$
  • Lead ($^{207}\text{Pb}$, primary collimators): $E_{\text{th}} \approx 6.7\text{ MeV}$
  • Concrete elements (Oxygen $^{16}\text{O}$, Carbon $^{12}\text{C}$): $E_{\text{th}} \approx 15.7\text{ to } 18.7\text{ MeV}$
• Cross-Section Peak: The GDR cross section peaks at photon energies between $13\text{ and } 15\text{ MeV}$ for heavy nuclei ($Z > 73$).

Because a clinical $18\text{ MV}$ linac produces a continuous bremsstrahlung spectrum with maximum photon energy of $18\text{ MeV}$, a substantial fraction of the photon fluence lies above the GDR threshold. At $\le 10\text{ MV}$, photoneutron fluence is low enough that it does not normally pose a radiological hazard, although NCRP 151 cautions that a beam labeled 10 MV can be tuned closer to 12 MV and produce measurable neutrons.

Neutron Source Strength ($Q_n$)

The total neutron production rate is characterized by the neutron source strength ($Q_n$), defined as the number of neutrons emitted from the accelerator head per Gray of X-ray absorbed dose delivered at isocenter ($\text{neutrons}/\text{Gy}$):

• $10\text{ MV}$ (stated): about $0.06 \times 10^{12}\text{ n/Gy}$
• $15\text{ MV}$ (stated): about $0.47$ to $0.76 \times 10^{12}\text{ n/Gy}$, depending on the model
• $18\text{ MV}$ (stated): about $1.2$ to $1.5 \times 10^{12}\text{ n/Gy}$

(values from NCRP 151 Table B.9, which also shows how much machines with the same nominal energy differ).

Neutrons are produced isotropically with an evaporation spectrum peaking around $1.0\text{ to } 2.0\text{ MeV}$, accompanied by a direct knock-out component extending to several MeV.

Neutron Transport and Kersey's Method

As fast neutrons escape the linac head, they undergo multiple elastic and inelastic collisions with concrete room surfaces, rapidly thermalizing down to room temperature ($E_{\text{th}} \approx 0.025\text{ eV}$). The neutron dose equivalent at the outer maze entrance can be estimated with Kersey's method (NCRP 151 Eq. 2.18):

$H_{n,D} = H_0 \left(\frac{S_0}{S_1}\right) \left(\frac{d_0}{d_1}\right)^2 10^{-d_2/5}$

where $H_0$ is the total neutron dose equivalent per gray of x rays at isocenter measured at $d_0 = 1.41\text{ m}$ from the target, $S_0/S_1$ is the ratio of the inner maze-entrance cross-sectional area to the maze cross-sectional area, $d_1$ is the distance from isocenter to the point on the maze centerline where the isocenter is just visible, and $d_2$ is the maze length beyond that point in metres. The maze attenuates neutrons with a tenth-value distance of 5 m (measured values are about 16% shorter, so 5 m is conservative). A second bend in the maze reduces neutrons at least threefold more, and McGinley and Huffman's modified Kersey method (with a TVD proportional to the square root of the maze cross-section) gives more realistic results.

The Capture Gamma Hazard ($^{56}\text{Fe}(n,\gamma)^{57}\text{Fe}$)

When thermal neutrons diffuse into vault materials, they undergo radiative capture $(n,\gamma)$. Thermal neutron capture in structural steel, linac metal components, concrete rebar, and door frames involves iron-56:

$^{56}\text{Fe} + n_{\text{th}} \to {^{57}\text{Fe}} + \sum \gamma$

Iron produces prompt capture gamma rays as energetic as about $7.6\text{ MeV}$, and capture on hydrogen in concrete ($^{1}\text{H}(n,\gamma)^{2}\text{H}$) gives $2.22\text{ MeV}$. NCRP 151 cites an average capture-gamma energy of about 3.6 MeV at the maze entrance, ranging up to about $10\text{ MeV}$ for very short mazes, so a lead TVL of about $6.1\text{ cm}$ may be needed. For mazes serving machines above 10 MV, the scattered and leakage photon dose is typically an order of magnitude smaller than the sum of capture gammas and neutrons, so capture gammas and neutrons usually dominate the door dose.

Heavy Shielded Door Construction: Borated Polyethylene & Laminates

To shield the combined field of scattered X-rays, fast/thermal neutrons, and high-energy capture gammas at the entrance of a high-energy ($> 10\text{ MV}$) vault, a specialized multi-layer composite door is required.

Composite Layer Architecture & Sequencing

A high-energy vault door incorporates three essential materials arranged in a specific functional sequence:

NCRP 151 reports that the average neutron energy at the maze entrance is about $100\text{ keV}$, recommends a BPE TVL of $4.5\text{ cm}$ for door calculations (thermal neutrons have a TVL near $1.2\text{ cm}$ in BPE), and notes that many rooms with mazes of about $8\text{ m}$ or longer need only $0.6$–$1.2\text{ cm}$ of lead and $2$–$4\text{ cm}$ of BPE in the door. The outer lead may be unnecessary when the maze is long enough to attenuate the neutrons before they reach the door.

1. Lead Layer 1 (Room Face): Lowers the energy of incoming neutrons through nonelastic scattering and attenuates photons arriving from the vault.
2. Borated Polyethylene (BPE Core, 5% Boron by weight):
   • Fast Neutron Thermalization: Pure polyethylene ($\{\text{CH}_2\}_n$) possesses a high concentration of hydrogen nuclei (protons, $A = 1$). In elastic neutron-proton collisions, maximum fractional kinetic energy transfer occurs ($Q_{\text{max}} = 1.0$), rapidly moderating fast neutrons ($1\text{ MeV}$) down to thermal energies ($0.025\text{ eV}$).
   • Thermal Neutron Capture via $^{10}\text{B}$: Natural boron contains $\sim 20\%$ $^{10}\text{B}$, which possesses a massive thermal neutron capture cross section of $\sigma_{\text{th}} \approx 3,840\text{ barns}$: $^{10}\text{B} + n_{\text{th}} \to {^7\text{Li}} + \alpha + \gamma\,(478\text{ keV}) \quad [94\% \text{ branch}]$
   • Why Boron is Essential: If thermal neutrons were allowed to strike steel or unborated materials, they would capture on iron, emitting penetrating $7.64\text{ MeV}$ gammas. By incorporating boron, thermal neutrons are captured into $^{10}\text{B}$, which emits only a soft, easily shielded $478\text{ keV}$ gamma ray and a non-penetrating alpha particle.
3. Lead Layer 2 (Outer Face): Attenuates the secondary $478\text{ keV}$ boron capture gammas and residual scattered X-rays before radiation escapes into the public corridor.

Mechanical Engineering & Safety Systems

• Weight and Drive Systems: Laminated doors are heavy enough to need motorized drives with obstruction-sensing safety edges.
• Emergency Egress: A manual opening method must work during a power failure so no one can be trapped in the vault.
• Interlocks & Overlaps: The door interlock must stop irradiation when the door opens, and the door must overlap the jambs and threshold so radiation cannot stream through gaps; NCRP 151 suggests alternatives such as reducing the inner maze opening or adding a light borated door at the inner maze entrance to keep neutrons out of the maze.

---

Health Physics Pitfalls

• Inverting the door laminate: NCRP 151 suggests lead, then BPE, then lead. The inner lead reduces neutron energy through nonelastic scattering so the BPE works better, and the outer lead attenuates the 478 keV boron capture gammas; leaving out the outer layer is acceptable only when the maze is long enough that few neutrons reach the door.
• Door gaps and undercuts: An ordinary floor undercut lets neutrons and capture gammas stream under a high-energy vault door; doors need overlaps at the jambs and a sill or threshold detail designed by the qualified expert.
• Trusting a "nominal 10 MV" label: NCRP 151 describes a nominal 10 MV beam whose measured accelerating potential was 11.6 MV and that produced photoneutrons in lead; verify the actual beam energy before assuming no neutron shielding is needed.