Vertex detectors are of growing importance in particle physics experiments as the knowledge of the event flavour is becoming an issue for several aspects of the most fundamental questions to investigate in the coming decades. Existing technologies (i.e. CCDs and Hybrid Pixels) exhibit high performances, but those needed to take full advantage of future experiments may be too demanding for these techniques. This triggered the development of a new type of detector, Monolithic Active Pixel Sensors (MAPS), in our laboratory (in collaboration with LEPSI-Strasbourg).

These devices were used previously for visible light detection (e.g. commercial camcorders). The principle of operation for charged particle detection is illustrated on the figure. The impinging particle produces excess carriers in the (low-resistivity) epitaxial layer at a rate of about 80 electron-hole pairs per micron. The electrons liberated diffuse thermally inside the layer, which lies inbetween two highly dopped zones: the substrate and the p-wells. The dopping levels of the latter are three orders of magnitude higher than the epitaxial one, translating in potential barriers at the region boundaries. As a consequence, the excess electrons remain inside the epitaxial layer. Regularly implanted n-wells collect the electrons passing in their neighbourhood. The density of the n-wells is the leading parameter for the sensor spatial resolution.

Figure : Principle of operation of a CMOS sensor designed for charged particle tracking

.

Prototype matrices, MIMOSA-1 and MIMOSA-2, (standing for Minimum Ionising particle MOS Active pixel sensor) of 64x64, 20 µm wide, pixels were fabricated in 1999 and 2000, using standard CMOS technology. They were calibrated with a ⁵⁵Fe X-ray source. The charge collection speed was evaluated with laser shots. Mounted on a beam telescope, they were tested at CERN with 120 GeV/c pion beams. The data analysis demonstrated that the sensors had very low irreducible noise, large signal-to-noise ratio (> 40), high detection efficiency (> 99 %) as well as excellent spatial resolution ( ~ 1.5 µm). The adequacy of this new technology for charged particle tracking was thus demonstrated at the microstructure level. The performances suit in particular the requests for a vertex detector at the Next Electron-Positron Collider (e.g. TESLA).

The present phase of developpement consists in adapting the sensors to their various potential applications, some of them departing substantialy from particle physics. Major goals include exploration of the technology intrinsic performances, fast read-out electronics and sparsification allowing efficient data flow compression. Another research line deals with ever better radiation tolerance. Finally, large and thin detectors meant to equip tracking detectors need to be fabricated, with performances as good as those observed with microstructures. These goals have triggered the fabrication of new microstructures (MIMOSA-3 and MIMOSA-4) in the first half of 2001, and of the first real size detector ladder ( MIMOSA-5) in August 2001.