To prepare for the higher collision energy and luminosity of the subsequent running period of the LHC significant improvements are ongoing in the CMS detector. One of the planned upgrades is the installation of large-area GEM detectors in the forward muon region. Following an extensive R&D program since 2009, Triple GEM (Gas Electron CMS GEM Upgrade  Multiplier) chambers have been developed as the optimal solution for three detector stations of the CMS Endcap. GEM technology has been successfully used in the past years in high-energy experiments including STAR, TOTEM and LHCb. The CMS GEM Upgrade project represents the next major step in the evolution GEM detector systems, both in terms of detector size and quantity, from a small number of medium-sized detectors to a large number of large-sized detectors. The sensitive area covered by GEMs in these three upgrades would be ~ 350m², totaling around 720 detector modules and about 1000m² of GEM foils.

The GEM detector can operate in the high-rate environments of Run 3 and HL-LHC and provide precise tracking thereby improving muon momentum resolution. This is done by exploiting the measurement of the bending angle of muons that emerge at an angle of around 10° relative to the beam axis. The adopted triple-GEM technology consists of a stack of three GEM foils interspersed at a distance of few mm enveloped in an argon carbon dioxide gas mixture delimited by drift and readout electrodes. The GEM foils are made of a metal-clad polymer with a thickness of 50 μm, chemically etched with millions of holes, typically 50 to 100 per mm.

Figure 1. (a): Scanning Electron Microscope (SEM) picture of a GEM foil (left). and schematic view of the electric field lines (b): Principle of operation of a generic triple-GEM chamber and definition of drift, transfer, and signal induction gaps. Stack of three foils arranged in tandem for sharing maximum operational gain within the drift and readout electrodes separated by mechanical frames.

When muons pass through a GEM detector, gas molecules within the detector are ionized and release electrons. These electrons drift towards the holes where they experience the very intense electric field inside the holes thereby acquiring enough kinetic energy to produce secondary ionization in the gas. This produces an electron avalanche process, which induces an electrical signal of the readout strips as shown in Figure 1. With three foils in tandem the multiplication process is shared, thereby improving stability against any possible discharges in the detector.

To meet the requirements of the CMS experiment, an early Phase 2 upgrade with 144 one-meter GEM detectors have been produced and are scheduled to be installed in the CMS endcap during 2019-20. Given the large number of detectors and the complexity of this technology, a very comprehensive and stringent quality control process has been established to ensure good performance and timely production of all the components.

A 5-year R&D programme resulted in five generations of prototype detectors that were built and tested between 2010-2014. Thanks to the knowhow and technical experience accumulated, the R&D programmes demonstrated that large-area GEM foils can be reliably manufactured and that using these foils for building triple-GEM detectors can satisfy the required performance. For CMS, an innovative  technique of stretching foils to build detectors without glue was elaborated and validated over several versions of prototypes. The details of these approaches is documented in a technical design report submitted and subsequently approved by CMS and LHCC : CERN-LHCC-2015-012.

Figure 2.  Top: Main steps of the GE1/1 chamber construction after several iterations 2010-2015. Bottom Left: Final chambers after closing. Middle: 10 chambers ready to go for installation for Slice Test in CMS (2015-2017). Right: Cosmic tests before final installation.

Once the procedures of construction, assembly and validation tests with rigorous quality control procedures had been established several production sites from all over the world participated in the production. Extensive training workshops were held at CERN in the CMS GEM laboratory to train all the teams at the respective production and satellite sites. Apart from CERN the production sites include Gent in Belgium with Aachen in Germany, INFN Bari and LNF Frascati in Italy, Florida Institute of Technology in the USA, Bhabha Atomic Research Centre (BARC), Panjab and Delhi Universities, and Saha Institute of Nuclear Physics in India, and National Centre for Physics in Pakistan. The flow of detector components from and to the production sites was streamlined and well documented in a Database. Once the detectors were complete at the sites, they were sent to CERN for final certification. In a period of about two years 144 chambers were completed and are all at CERN presently undergoing the last stages of the quality control chain. An interesting observation is that the uniformity of all these detectors has been measured to within 15% as specified and expected at the technical design. Kudos to the meticulous work of all the production site managers and their teams.