The low-x structure of protons and in particular nuclei is not well constrained experimentally, while the knowledge about it is a crucial ingredient for the interpretation of many measurements at high-energy hadron colliders. It is widely expected that the growth of parton densities at low x predicted from linear QCD evolution cannot continue indefinitely, and that non-linear effects will lead to gluon saturation. Many experimental observations are compatible with saturation, however no real proof has been found yet. We propose the measurement of forward direct photons in proton-nucleus collisions at the LHC as a decisive probe of gluon saturation. Due to the harsh environment of such a measurement, existing detectors are not suitable. In particular an extremely high-granularity electromagnetic calorimeter is required, which we propose as a detector upgrade to the ALICE experiment, the Forward Calorimeter (FoCal). To facilitate the design of the upgrade and to perform generic R&D necessary for such a novel calorimeter, a compact full high-granularity electromagnetic calorimeter prototype has been built. This prototype is a Si/W sampling calorimeter using CMOS sensors of the MIMOSA type with a pixel pitch of 30 micron and binary readout with a total of ~39 million pixels. About 85% of the total volume is W absorber, a small Molière radius of 10.5 mm is expected.
The raw data shows the properties and limitations of this prototype. To obtain a good performance, several selection procedures and an additional analysis methodology based on hit density is applied. In simulation, to make a valid comparison, misalignment, noise, and charge diffusion are considered. A realistic detector is simulated, a realistic number of noise pixels has also been added. To study the effects of dead areas, both the `ideal detector' and `real detector' are simulated. We present on performance studies of the prototype with test beams at DESY and CERN in a broad energy range. The results of the measurements demonstrate that a very small Molière radius and good linearity of the response. Unique results on the detailed lateral shower shape, which are crucial for the two-shower separation capabilities.
The studies demonstrate the feasibility of this high-granularity technology for use in the proposed detector upgrade. Furthermore, they show the extremely high potential of this technology for future calorimeter development.