ALICE is carrying out the R&D to replace during LS3 its innermost tracking layers by a new device (ITS3) that approaches the "golden detector". The aim is to reduce the material budget to a minimum to less than 0.05% X0, get closer to the interaction point with the first layer at a radial distance of 18 mm while the beam pipe has a radius of 16 mm, and keep a high intrinsic spatial resolution of less than 5 μm in both longitudinal and transverse directions.
These figures of merit are obtained by exploiting the mechanical properties of thin (<50 μm) silicon chips which are flexible enough to be bent onto truly cylindrical surfaces of radii well below 18 mm. This not only minimises the achievable distance from the beam pipe (no staggering needed), but also gives the structure an intrinsic stability, removing the need of considerable support structures. Using stitching, full half-layers are integrated into single silicon pieces (up to ~28⨉9 cm² in size cut from a 300 mm wafer), removing the need of printed circuit boards for interconnection in the active area. Finally, the low power consumption (<40 mW/cm²) that can be achieved with MAPS makes air cooling sufficient to operate the detector. All this together will allow to build the whole detector from only six large sensors and barely anything else within the active volume.
In this seminar, the motivation of building such a device from a physics and detector performance point of view will be shortly reviewed, the conceived detector layout and specifications will be explained, and a number of obtained R&D highlights will be illustrated. The latter will contain results of beams tests with bent CMOS sensors, mechanical prototypes of detector half-layers, as well as an overview of the CMOS sensor developments. I will conclude with an outlook of the remaining R&D plans and a timeline towards the Technical Design Report as well as construction and installation.