The Ultra-High Energy Cosmic Ray Spectrum Measured by the Telescope Arrays Middle Drum Detector

The Telescope Array’s Middle Drum fluorescence detector was constructed using refurbished telescopes from the High Resolution Fly’s Eye (HiRes) experiment. As such, there is a direct comparison between these two experiments’ fluorescence energy spectra. An energy spectrum has been calculated based on one year of collected data by the Middle Drum site of Telescope Array and agrees well with the HiRes monocular spectra. The quality of the Middle Drum results has also been determined to show good agreement.

💡 Research Summary

The paper presents the first‑year results of the Telescope Array (TA) Middle Drum (MD) fluorescence detector, which was built from 14 refurbished telescopes originally used in the High Resolution Fly’s Eye (HiRes) experiment. Operating from December 16 2007 to December 8 2008, the MD detector collected data on nights with at least three hours of darkness, achieving a total of approximately 835 hours of scheduled time, of which about 754 hours (≈90 % of the scheduled time) were classified as “good” based on minimal cloud cover. This corresponds to an overall duty cycle of roughly 9 %, i.e., about one‑fifth of the running time of HiRes‑1.

The MD telescopes view a 120° azimuthal sector and elevations from 3° to 31°, each camera containing 256 photomultiplier tubes (PMTs) read out with a 5.6 µs sample‑and‑hold gate. Monte Carlo (MC) simulations using the same hardware response as HiRes‑1 were employed to determine the detector’s aperture. The resulting aperture is roughly half that of HiRes‑1, and when combined with the reduced live time the total exposure of MD is about one‑tenth of HiRes‑1’s exposure.

During the first year, 1 156 events with reconstructed energies above 10¹⁸ eV were recorded; no events were observed above 10²⁰ eV, which is consistent with the limited exposure. The energy spectrum was derived by binning events in decade‑wide logarithmic energy bins and dividing by the exposure per bin. The resulting spectrum agrees, within statistical uncertainties, with the HiRes‑1 monocular spectrum, thereby providing a direct cross‑calibration between the two experiments.

The MC sample was generated with an isotropic arrival direction distribution and a broken power‑law spectrum (spectral index 3.2 below 10¹⁸·⁵ eV and 2.8 above), deliberately omitting any Greisen‑Zatsepin‑Kuzmin (GZK) cutoff. Showers were simulated using the Gaisser‑Hillas parameterization to model the longitudinal development of charged particles, and the corresponding fluorescence light was propagated to the detector. Reconstruction algorithms then produced estimates of the primary energy, impact parameter (Rₚ), and in‑plane angle (Ψ).

Resolution studies show that for the central energy range (10¹⁸·⁵–10¹⁹ eV) the reconstructed energy deviates from the true value by roughly ±0.5 dex, while Rₚ and Ψ are reconstructed with comparable precision. Data‑MC comparisons of Rₚ and zenith angle (θ) distributions demonstrate good agreement; despite large statistical error bars in the data, the MC histograms overlay the measured points closely, confirming that the simulated aperture accurately represents the real detector response.

In conclusion, the MD detector’s first‑year monocular analysis yields an energy spectrum consistent with HiRes, validates the MC‑based aperture calculation, and demonstrates satisfactory energy and geometry reconstruction. These results establish MD as a reliable bridge between the HiRes and the broader TA experiment, enabling future combined analyses that will incorporate other fluorescence stations and the surface detector array to refine the energy scale and improve statistics at the highest energies.