Design and performance of a silicon–tungsten calorimeter prototype module and the associated readout

https://doi.org/10.1016/j.nima.2020.164796Get rights and content

Abstract

We describe the details of a silicon–tungsten prototype electromagnetic calorimeter module and associated readout electronics. Detector performance for this prototype has been measured in test beam experiments at the CERN PS and SPS accelerator facilities in 2015/16. The results are compared to those in Monte Carlo Geant4 simulations. This is the first real-world demonstration of the performance of a custom ASIC designed for fast, lower-power, high-granularity applications.

Introduction

The parton structure of protons and nuclei is typically characterized in terms of parton distribution functions (PDFs) which parameterize the non-perturbative physics that cannot at present be calculated from Quantum Chromodynamics (QCD) [1], [2]. The PDFs are determined from global fits to experimental measurements, in particular from deep inelastic scattering (DIS) experiments such as H1 and ZEUS at HERA [3]. The gluon PDF is found to rise dramatically in the small-x region (x=102106), where the Bjorken variable x is the longitudinal fraction of the momentum of the nucleon carried by the partons (quarks and gluons). At small x one expects non-linear behavior of QCD with competing processes of gluon splitting and fusion, which should eventually lead to saturation of the gluon density. Despite extensive experimental studies, there is no direct evidence of gluon saturation, nor the creation of the Color Glass Condensate [4]. By measuring direct photons in the forward direction, which are produced in the quark–gluon Compton process, one directly probes the gluon PDF, which allows exploring QCD in the non-linear regime, and placing stringent constraints on the gluon nuclear PDFs [5].

Motivated by the considerations above, a proposal was developed to build a forward calorimeter system in the ALICE experiment at the LHC, called FoCal [6], [7]. The FoCal is composed of an electromagnetic part (FoCal-E) and hadronic part (FoCal-H), with a planned pseudo-rapidity coverage of about 3.4η5.8. The FoCal-E is designed to have an excellent two-shower separation to identify and reject decay photons from neutral mesons with transverse momenta in the range from a few GeV/c to above 20 GeV/c, i.e. total momenta up to about 1 TeV/c (pT=5 GeV/c corresponds to p370 GeV/c at η=5) to enable the measurement of direct photons at the LHC in the small-x region. It uses tungsten converter layers with a thickness of 1 radiation length each and silicon sensor layers, with a compact design to minimize the shower size. The sensor layers consists of silicon pad sensors with a granularity of 1 × 1 cm2 (low granularity) that will primarily measure the energy of the shower, and a few high-granularity pixel layers that will be used in the future detector to split contributions from overlapping showers.1

A Si-W calorimeter prototype using the Si-Pad technology of similar granularity (1 × 1 cm2 and more recently 0.5×0.5 cm2) has previously been built and tested by the CALICE collaboration [8], [9], [10]. The prototype discussed here uses a different approach for readout electronics, summing signals from several layers to reduce the necessary number of channels without compromising the performance needed for measurements of electromagnetic showers. In addition, the use of pixel layers with their much higher granularity is a unique feature. The analysis of events with simultaneous read out of the high-granularity and low-granularity layers shows that sub-millimeter precision on the shower position is obtained with this setup.

In this paper, we present the results of test beam measurements at the CERN Proton Synchrotron (PS) and the Super Proton Synchrotron (SPS) in 2015/2016 for a prototype using pad and pixel layers. The pad readout uses a new custom ASIC developed for very fast (fast enough for low-level trigger decision), lower power, fine granularity (reduced cost per channel) applications such as the FoCal. The ASIC provides differential drive to downstream electronics which can be located farther from the acceptance with reduce requirements for radiation tolerance. For the test setup with only pad layers, the energy resolution, energy linearity, and shower profiles are presented and compared to simulations using the Geant4 simulation took kit [11], version 10.5.01 with physics lists FTFP and BERT. We also show the performance of the integrated system, i.e. combined pad and pixel layers together, as a first-principle design of the FoCal-E prototype from the 2016 test beam data.

Section snippets

Detector design

The FoCal-E design is designed to provide very high lateral segmentation to discriminate between decay photons and direct photons. The electromagnetic calorimeter requires compact shower size to minimize the effect of shower leakage and optimize shower separation. Therefore, the FoCal-E prototype has been designed as a silicon and tungsten (Si+W) sampling calorimeter, because tungsten, which is used as an absorber material, has a small Molière radius of 9 mm and a radiation length of 3.5 mm.

Readout electronics

Traditional charge-sensitive preamplifiers (CSP) are commonly used for readout of “capacitive” detectors (detectors such as silicon pads, strips, etc. for a which a simplified model consists of a capacitor in parallel with a current generator) for two reasons. First, all the charge generated in a detector due to a radiation event is ultimately collected by the preamplifier irrespective of the detector capacitance. Higher detector capacitance tends to slow the preamplifier bandwidth such that it

Detector setup for test beam experiments

The test beam performance for the prototype detector was carried out at PS and SPS facilities at CERN in 2015 and 2016. In 2015, we used the T9 test beam facility at the PS accelerator providing a secondary beam (with mixed electrons and hadrons) with a beam energy from 0.5 to 10 GeV. In the same year we used the SPS T4-H6 beam line with the energy setup of 5–180 GeV. In 2016, the SPS H6B beam line was used with a very similar setup as the previous year.

The purity of the electrons or positrons

PAD detector performance

Fig. 8 shows the distribution of measured charge (ADC) for single cells (i.e. longitudinally summed groups of four consecutive pads) as a function of time for the four segments (LGL0, LGL1, LGL2 and LGL3) for 40 GeV positrons. The horizontal axis is time (25 ns per bin), and the vertical axis is the charge collected at the summing board on each LGL, i.e. average amplitude of the ADC values over all events in each time bin. Of the four segments, LGL1 shows the largest signal, which corresponds

Pixel detector performance

One of the main purposes of the pixel layers is to facilitate the reconstruction of pairs of photons from neutral pion decays. To do this, they need to provide good two-shape separation and provide good estimates of the position and energy of the single showers in the pair. Measurements with a full digital pixel calorimeter prototype [12] have shown that the energy linearity and resolution are good and that the position resolution is excellent. With the present measurements we can verify

Correlations between pad and pixel layers

Although pad and pixel layers were integrated to be a common system, their detector designs are quite different from each other, and the readout systems were independent. Therefore to integrate hit data from pad and pixel layers, a common beam trigger given by the coincidence of two scintillation counters was stored in each data stream. For the 2016 test beam data, we successfully obtained correlated data sets by using the common trigger info for both pad and pixel-layer data stream. The data

Summary

We have constructed a prototype of a silicon–tungsten sampling electromagnetic calorimeter module and the associated readout electronics, and studied performance with a test beam experiment at the CERN PS and SPS accelerator facilities in 2015/2016. To facilitate the readout, the front-end electronics was equipped with a new custom ASIC developed for this and other application where fast, fine-granularity imaging is required. The energy resolution obtained is consistent with the results

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank H. Muller and the RD51 collaboration at CERN for their kind support on the pad readout and frontend system using APV25 hybrid boards and SRS. We would also like to thank the staff members of the CERN accelerator complex for providing stable beams at PS and SPS beam tests. This work was supported by JSPS KAKENHI Grant Numbers JP25287047, JP17H01122, JP26220707. Authors affiliated with ORNL are supported by the U.S. Department of Energy, Office of Science, Office of Nuclear Physics ,

References (15)

There are more references available in the full text version of this article.

Cited by (0)

View full text