The Lepix goal is to design novel monolithic pixel sensors with enhanced speed and radiation hardness through the use of deep submicron CMOS technologies fabricated on moderate-to-high resistivity wafers.
The first submissions have targeted a 90 nm CMOS process, in cooperation with a major silicon foundry.
Owing to the availability of low-k dielectrics, a CMOS process of the 90 nm generation and beyond offers an extremely high interconnection density . This can be exploited to achieve a massive parallel read-out, which allows to process independently the signals of every pixel. On the one hand, a fast read-out minimizes the deterioration of the signal quality due to radiation-induced leakage current. On the other hand, it makes the sensors suitable for applications in which a prompt time tagging of the events is of primary importance. The use of uniform substrates with moderate or high resistivity will lead to significant depletion regions even with moderate reverse bias voltages. In this manner, a bigger signal is obtained. Furthermore, since the charge is collected by drift, a significant suppression of radiation damage due to non ionizing radiation is expected.
While the wafers substrate of interest for Lepix have a resistivity higher than what is normally found in standard microelectronics circuits, the CMOS technology employed is a fully standard and commercial process. This offers in addition the perspective of high volume sensor production at a moderate cost.
From the start of the project, two submissions have been done in the 90 nm process, one on standard substrate and one on high resistivity one. The prototypes have been tested in the lab (though laser tests and X-ray sources) and in beam test with particles. Estensive characterization of the radiation tolerance has also been carried-out.
The test results of the base material, performed on simple diodes, have been very promising, indicating the possibility of applying reverse bias voltage of at leat 60 V and achieving depletion region in excess of 50 um.
The results of the diodes could not be fully replicated on more realistic sensors,
due to the side-effect of the filling structures that could not be adequately modelled
In these first prototypes a full efficiency of the sensors could not be achieved, but the test results have provided a comprehensive information on the criteria to be followed in the future in the design of such devices.
Due to a change in the policy of the foundry, the possibility of accessing the 90 nm process at a price affordable for a small R&D activity was discontinued, making impossible a further run with a more optimized design. In the course of 2012, the Lepix sensor concept was thus experimented on a less aggressive 180 nm process provided by TowerJazz, which has however the plus of allowing an optimal isolation between the sensing diode and the nearby front-end electronics. The submission was done at the end of 2012, using the standard option of the foundry which offers epitaxial layers with a maximum depth of 20 um. This new prototypes will be completely tested in the course of 2013. The preliminary results have however been very encouraging and, combined with the ones obtained in the 90 nm submission, have stimulated the foundry to offer also wafers with uniform resistivity, similar to the ones available in the 90 nm process.
Several concepts explored by Lepix are now being considered for the design of monolithic sensors for the upgrade of the ALICE tracker.