Array. GRETA (Gamma-Ray Energy Tracking Array): 120 HPGe (high-purity germanium) crystals in 30 quad modules (2 type-A + 2 type-B irregular-hexagon crystals each), two hemispheres of 15 quads, ≈ 80% of 4π. Close-packed on a geodesic shell of 120 hexagons + 12 pentagons (here the chiral Goldberg tiling GP(3,1), T = 13, with the beam through two opposite pentagons). Quads are grouped on the 30 icosahedron edges — an approximation of the real module placement. Completed at Berkeley Lab Aug 2025; commissioned at FRIB (Facility for Rare Isotope Beams) and declared ready for science July 2026 (FRIB, ANS).
Crystals: 90 mm long, 80 mm Ø cylindrical rear tapering to an irregular hexagonal front, 36-fold segmented (6 azimuthal wedges × 6 longitudinal slices; slice depths 8/14/24/38/54/90 mm here, approximate) plus the core contact (GRETA design; arXiv:2512.12484). Front faces at R ≈ 18.5 cm from the target; dead gaps of 2.5 mm per crystal face inside a quad (capsules) and 4.5 mm towards neighbouring quads (capsule + aluminium end-cap walls). Each quad's end-cap is drawn as a faceted housing around its 4-crystal diamond; like all passive material it is not included in the γ transport. Downstream the S800 spectrograph (quadrupole doublet + two dipoles, drawn schematically: first quads only) identifies the outgoing nucleus and measures its momentum and angle event by event. Entry/exit rings (5 quads around each beam pentagon) are removable: 20-quad configuration as in front of the S800 before the HRS (High Rigidity Spectrometer).
Response: γ transport in Ge only (no Al/cryostat material), photoabsorption μpe/ρ ≈ 0.40 (E/100 keV)−3.1 cm²/g, Klein–Nishina Compton, pair production above 1022 keV with both 511 keV photons tracked (attenuation approximated to NIST XCOM at the 10–20% level). Electron ranges ignored. Interactions in one segment merged (energy-weighted), as in signal decomposition; positions smeared by σ = 2 mm per coordinate (GRETINA: σ ≲ 2 mm). Crystal energies smeared with FWHM² = 1.16 + 0.0027·E[keV] keV² (≈ 1.2 keV at 122 keV, 2.2 keV at 1332 keV, GRETA acceptance-test means).
Tracking: cone clustering (20° about the target), then all orderings of each cluster (≤ 7 points) scored by the Compton figure of merit χ² = Σ[(E′geo − E′meas)/σ]²; with an effective 6 mm position error (segment merging + resolution); clusters with χ²/ν > 5 rejected, single-point clusters kept if E < 1.5 MeV (after Deleplanque et al., NIM A 430, 292, 1999; Lee, Deleplanque, Vetter, Rep. Prog. Phys. 66, 1095, 2003). Doppler correction E₀ = Elab·γ(1 − β cos θ) with θ to the first point (tracked) or to the crystal centre (non-tracking reference).
Sources: fast beams at β ≈ 0.4 on a ≈1 mm ⁹Be target (β spread ±0.02 through the target, nominal β used for the correction); level energies: ³²Mg 885 / 1438 keV (Motobayashi et al., PLB 346, 9, 1995), ⁴²Si 742 / 1431 keV (Takeuchi et al., PRL 109, 182501, 2012), ⁵⁴Ca 2043 keV (Steppenbeck et al., Nature 502, 207, 2013); lifetimes rough estimates. ReA rotational band: generic E(I→I−2) = 11·(4I − 2) keV plus a few statistical γ, β = 0.03. ⁶⁰Co 1173/1332 keV. No passive material, so the efficiencies shown are upper-limit toy values. Toy model — not GEANT4 / not the GRETA simulation.