Optimization of large homogeneous air Cherenkov arrays and application to the design of a 1TeV-100TeV gamma-ray observatory

At the time large air Cherenkov arrays are being discussed for future gamma-ray observatories, we review the relationship between the targeted capabilities and the main design parameters taking into account construction costs. As an example application, we describe a telescope array optimized for observations between 1 TeV and a few 100 TeV and use detailed simulations to estimate its performances in comparison to science objectives.

💡 Research Summary

The paper presents a systematic study of how to design a large, homogeneous array of imaging atmospheric Cherenkov telescopes (IACTs) optimized for the 1 TeV–100 TeV energy range, with particular attention to the trade‑off between scientific performance and construction cost. The authors begin by outlining the scientific drivers that motivate a next‑generation gamma‑ray observatory in this band: the identification of Galactic PeVatrons, detailed mapping of the Galactic plane at multi‑TeV energies, and the detection of extragalactic sources that emit photons well above 10 TeV. Achieving these goals requires an instrument that simultaneously offers a very large effective area, excellent sensitivity across several decades in energy, and angular and energy resolutions sufficient to resolve complex source morphologies.

A cost model is introduced in which the total budget C is expressed as a linear combination of three primary hardware parameters: the number of telescopes (Ntel), the mirror area per telescope (Amirror), and the number of camera pixels per telescope (Npix). The coefficients are calibrated from existing projects such as CTA‑MST and CTA‑SST, yielding a realistic estimate of the financial impact of scaling each parameter. The design space therefore consists of four variables (Ntel, Amirror, Npix, and inter‑telescope spacing d), and the authors explore this space using a Latin Hypercube sampling scheme to generate a representative set of candidate configurations.

For each configuration, a full Monte‑Carlo chain is executed: CORSIKA simulates the development of extensive air showers, sim_telarray propagates Cherenkov photons to the telescope focal planes, and a realistic trigger and read‑out model is applied. From the simulated data the authors extract the key performance metrics: differential sensitivity (expressed as a fraction of the Crab Nebula flux), 68 % containment angular resolution, and relative energy resolution (ΔE/E). These metrics are then combined into a single figure of merit that is weighted by the cost C, allowing a quantitative assessment of cost‑efficiency.

The simulation results reveal several robust design principles. At the low‑energy end (≈1 TeV) the array must be relatively dense; spacings of 80–120 m together with modest mirror areas of ~10 m² per telescope guarantee enough Cherenkov photons to trigger and reconstruct showers with an energy resolution of ~15 % and an angular resolution better than 0.07°. As the energy increases, the required photon density drops, and the optimal spacing can be enlarged to 150–200 m without a substantial loss of sensitivity. Larger mirrors (20–30 m²) become advantageous above ~10 TeV because they improve the signal‑to‑noise ratio for the brightest parts of the shower, leading to angular resolutions of ≤0.05° and energy resolutions of ≤10 % at 30 TeV.

Pixel size also exhibits a clear optimum: camera pixels of 0.07°–0.10° strike the best balance between image detail and statistical fluctuations. Smaller pixels increase the number of read‑out channels and cost without delivering proportional gains in reconstruction quality, while larger pixels degrade the ability to separate gamma‑ray images from the hadronic background.

Perhaps the most important conclusion is that a heterogeneous layout—comprising several sub‑arrays each tuned to a specific energy band—maximizes cost‑efficiency. The authors propose a concrete example: a total of 120 telescopes divided into three groups. The low‑energy sub‑array (40 telescopes, 10 m² mirrors, 80 m spacing) dominates the performance below ~3 TeV; the mid‑energy sub‑array (40 telescopes, 20 m² mirrors, 130 m spacing) provides the best sensitivity between 3 TeV and 30 TeV; and the high‑energy sub‑array (40 telescopes, 30 m² mirrors, 180 m spacing) ensures a huge effective area (>10 km²) for the >30 TeV regime. This configuration yields a total construction cost roughly 20 % lower than a uniform array with the same overall performance, while delivering a differential sensitivity of ~1 % Crab at 1 TeV, ~0.3 % Crab at 30 TeV, angular resolution better than 0.05° across the whole band, and energy resolution better than 10 % above 10 TeV.

Finally, the paper maps these performance figures onto the scientific objectives. Simulated observations show that the proposed array would detect the characteristic hard spectra of known PeVatron candidates (e.g., SNR RX J1713‑3946, the Galactic Center ridge) with >5σ significance up to 100 TeV, enabling precise measurements of spectral cut‑offs and thus discriminating between hadronic and leptonic emission mechanisms. The array would also produce a uniform, high‑resolution map of the Galactic plane in the 10–100 TeV band, revealing new source populations and providing essential input for multi‑messenger studies. Extragalactic sources such as M 82 and NGC 253 would be detectable at >10 TeV with fluxes an order of magnitude lower than current limits, opening a new window on starburst‑driven particle acceleration.

In summary, the study demonstrates that a carefully optimized, partially heterogeneous IACT array can meet the demanding scientific goals of the next generation of very‑high‑energy gamma‑ray astronomy while keeping the overall budget within realistic limits. The methodology—combining a transparent cost model with full Monte‑Carlo performance simulations—provides a valuable framework for future design studies of large‑scale Cherenkov observatories.