Experiments in subatomic physics rely on multi-measurements to identify
precisely the final quantum state under study. This results in rather large
detection systems involving various technologies dedicated to specific tasks, like
tracking and calorimetry to name a few. The evolution of science demands for
increasing event rates and thus drives detectors towards higher granularity in
space and time. But these specifications are conflicting in a single sensing
technology, typically because more channels leads to longer readout time.
Reaching the same specifications by combining heterogeneous technologies with
small sensing elements also meets some practical obstacles related to
We propose to overcome these limitations with very compact detection systems
built by stacking CMOS monolithic active pixel sensors (CMOS-MAPS) in direct
contact. The overall volume of the apparatus will be continuously sensitive but
for the thin (few micrometres) electronic layers and the necessary
interconnections. The best analogy describing such systems is the one of an
electronic nuclear emulsion.
The main strength brought by multi-point measurement is to maximize the
information extracted for each radiation penetrating the stack. The system
registers the propagation through sensitive materials with potentially 3+1+1D
granularity: space, time and energy-loss. The particle history is traced through
the stack both in a tracker and a calorimeter ways. The complete set of
measurements achieved includes: initial impact position, direction and time,
energy loss, particle range and potential decays.
Regarding the sensitivity to the various particle types (especially charged versus
neutral) and the range of energies, stacking sensor layers present decisive
benefits. While first layers stop low range radiations, the more penetrating ones
are still measured further away in the stack. Potential inserts of non-active
materials can also be considered, like thin scintillator layers to the light of which
CMOS-MAPS are sensitive.
Without being exhaustive, we underline two additional benefits of proximity
measurement redundancy. First, multi-measurement enhances radiation
tolerance, since missing information from a damaged cell can be replaced with a
subsequent adjacent cell without severe loss for the overall information
extraction. Second, such a system is more capable to handle large particle flux;
because either radiation are measured by different layers or the historical
information on each particle helps disentangle each of them.
Types of information and stack depth needed depend naturally on the
application. This is the reason CMOS-MAPS are particularly well suited to the
task, since they nowadays exist with a variety of detection performances, with
respect to sensitive depth, pixel size, readout speed and signal amplitude
resolution. Additionally, a number of these features can be adjusted after the
chip fabrication, enhancing the plasticity of the apparatus. For instance, the
sensitive depth obtained over a highly resistive substrate is controlled through
the biasing voltage in a range from 10 to 100 micrometres. In the near future,
sensors will offer re-configurable embedded signal treatment algorithm, like
neural networks, to fit the same chip to various situations.
The proposed system presents both unprecedented performances and
versatility. It is hence expected to impact a wide variety of domains, from which
we take a few illustrative examples.
In particle physics, a few layers stack will allow reconstructing directly helices
corresponding to particle trajectories within a magnetic field, complemented
with finer timing and more radiation tolerance. A tracker based on such stacks
would feature improved counting rate, signal-background separation capabilities
and an overall low material budget since a low number of such stacks would be
needed to complete a full tracker.
A deep stack about a centimetre depth, would detect simultaneously charged
ions species, as a perfect Delta-E / E telescope used in nuclear physics, and
neutrals as a segmented scintillator block, with the addition of correlated timing
information. The system would match perfectly proton radiography
Considering X-ray detection, a thick sensor stack would be effective over a wide
range of energies and increase the counting rate with respect to a single sensing
layer. The evolution of counts with depth would also directly provide
spectroscopic analysis of the energy distribution.
In the general scientific images domain, the proposed stack would bring
advantages like multi-modality (X-rays, visible light, electrons, neutrons) and
analysis power (since more information is collected). Non-destructive tests
widely spread in industry would benefit from these progresses at various scales,
from production good or material controls to aging building checks (in nuclear
installations for instance).
While CMOS-MAPS fulfilling the needs exist or almost exist, integrating a large
number of these chips over a useful area (at least 100 cm2) and thickness (from
few millimetres to centimetres) set the real challenge. Interconnection,
mechanical stability, power dissipation and data extraction count among the
pivotal issues to solve. A number of processes from the semi-conductor industry
already provide potential solutions. Nevertheless, a dedicated large scope effort
is required within the next ten years to properly optimise and
The trend in subatomic physics experiments is to increase the granularity of
measurements, in space and time. Practical difficulties limit the achievable
performances, since current experiments mostly rely on the integration of
heterogeneous technologies. In contrast, a continuous pixelated sensitive volume
could replace a complete complex setup and provide unprecedented
performances, if the material can detect various particle types.
CMOS monolithic active pixel sensors (CMOS-MAPS) benefit nowadays from a
high sensitivity and a thickness almost entirely sensitive. A stack of CMOS-MAPS
in direct contact would act as the volume dreamed for, providing tracking,
calorimetric and timing information. The number of layers in the stack, their
thickness and the specifications of their pixel sensors would be adapted to
optimise the overall performances depending on the type of particles (charged
particles, ions, X-rays, gamma-rays), their energy and flux.
The potential applications of this new type of instruments span a vast range of
domains, from scientific to industrial measurements. The plasticity of the stack
configuration and versatility offered by CMOS-MAPS will grant cross-fertilisation
between these fields.
The realisation of such stacks of CMOS-MAPS will combine and optimise
processes from the semi-conductor industry to solve the main issues, among
which are interconnections, mechanical stability, power dissipations and data