ATTRACT featured stories: Integrated signal processing for a new generation of active hybrid single photon sensors with ps time resolution (FastICpix)

ATTRACT featured stories: Integrated signal processing for a new generation of active hybrid single photon sensors with ps time resolution (FastICpix)
What has the ATTRACT seed funding enabled you to do so far?

With ATTRACT FastICpix seed funding, University of Barcelona and CERN are studying the feasibility of a new generation of active hybrid single photon sensors with picosecond time resolution. We are developing novel architectures of reconfigurable front-ends in a 65 nm CMOS technology. The deliverables of the project are IP blocks (including layouts) for the analogue front-end, Time-to-Digital Converter (TDC) and clock distribution network. This involves very complex microelectronic design at different levels.  The final goal for ATTRACT phase I is a Monte Carlo simulation including the integrated readout described at transistor level and the sensor model and interconnects in order to validate the FastICPix concept: 10 ps time resolution for a large (> 10 mm2/ch) detector. With the ATTRACT funding we have hired dedicated personnel to work in microelectronics design and Monte Carlo system level simulations and we have funded travels for internal meetings and dissemination.

What challenges have you faced so far?

With FastICpix we want to improve by one order of magnitude the performance of the current technology and this implies that we have to face challenges at all levels:

  • We are proposing a new regime of detector segmentation for solid state single photon sensors. Therefore, to define the architecture we need to explore a large space of solutions involving the simultaneous optimisation of many parameters: sensor technology, sensor segmentation, interconnects and characteristics of the front end amplifier. For that we have defined a figure of merit that involves time resolution and power. We are approaching the problem both from analytics and numeric (optimisation) point of views.
  • The design of the analogue front end electronics is challenging by itself, as we need to push the 65 nm CMOS technology to the limit. On top of that, as said before, the design parameters of the input stage are not independent of the complete sensor optimisation. And finally, this should be a power aware optimisation, otherwise the sensor won’t be feasible.
  • Again, we are beyond the state-of-the-art of TDC and clock distribution networks in 65 nm CMOS technology. For instance, we need to develop a TDC with 20 ps time bin (< 10 ps time resolution) with a power budget lower than 1 mW.
  • One final challenge brings a new level of complexity. Our sensor must be reconfigurable in order to adapt to different sensors and applications and still achieve the optimal figure of merit. The implications are twofold: the architectural study is even more intricate and the design of the IP blocks has additional constraints.
Where does your ATTRACT journey go from here?

In the first phase of ATTRACT we will try to proof the principle of the FastICPix hybrid device.  As said before, the final deliverable for ATTRACT phase I is a Monte Carlo simulation to validate the FastPix concept: 10 ps time resolution for a large area detector. Positive results will justify its implementation in the second phase of ATTRACT.

Once we achieve that, we plan to build hybrid sensor prototypes and demonstrate the vast potential of such sensor in different applications. The idea will be to develop a set of demonstrators for different applications which will be made available to potential industrial partners. The second phase of ATTRACT would provide a perfect funding scheme to achieve this final goal.

This project has a huge transformational impact on Society, from medical imaging to transport or homeland security. The transformational impact on Medical Science & Technology is at the core of the proposal. Together with advances in light generation, this will allow to have an on-line 3D PET image of the dose concentration without having to run complex and lengthy reconstruction algorithms. This will make come true the personalised (or precision) medicine, i.e. the right prevention and treatment for the right patient at the right time. Personalised medicine is at the heart of the health care priorities of the 21st century. A 10 ps CTR also implies a 10-fold increase in sensitivity, which translates in the reduction of radiation dose, scan time, and cost by an order of magnitude. A low dose and a cost of 100 € per scan will transform in-vivo molecular imaging into a standard tool for personalised medicine, thus bringing obvious societal and economic benefits, such as screening and early diagnosis of cancer, cardiovascular and neurodegenerative diseases.  In the short term, even with current scintillating crystal technology, FastICPix will bring a huge improvement on PET images, leading to a new generation of PET scanners.  In summary, the social impact is out of question and the economic impact will be at the level of thousands of millions per year in the PET industry (a market above 500 M$/year only in the US) for a new generation of PET scanners.

But the impact of FastICPix is not limited to medical imaging. It will open a new era in Light Detection and Ranging (LIDAR) by achieving millimetric spatial resolution requiring no averaging, and the development of diffuse optics systems for Fluorescence Lifetime Imaging (FLIM) which drastically reduces the system complexity and cost. An example in the latter field is DNA sequencing, which has been successfully performed with SiPMs. Currently available SiPMs have large active area (typically > 1mm2), which is not necessarily required for these applications. Hence, the dark count rate can be significantly reduced if the SiPMs with smaller active area are designed. This will be achieved by FastICPix thanks to its reconfigurable granularity and would make it usable as photon counter in many more fluorescence detection applications, for e.g. spectrography.

Future high luminosity experiments in High Energy Physics are relying in precise time stamping in order to deal with increased levels of pile up. Fast timing allows to distinguish overlapping events by adding the time dimension.  Many detectors are planning to include time of flight layers (e.g. CMS timing layer). LHCb TORCH is a novel time-of-flight detector in which timing is required to provide charged-particle identification. The FastICpix readout chip would be suitable for reading out the MCP used in this detector. FastICPix will find many applications in time-of-flight detectors for High Energy Physics.

As said before, we will seek for industrial collaborations in different domains of application and we may  envisage the creation of a spin-off company to bring our new hybrid sensor to the market. We plan to explore different industrialisation options and it is obvious that ATTRACT ecosystem fits extremely well to do that.

Sum up in two sentences the advantages of the ATTRACT Programme over other research funding schemes.

ATTRACT has managed to bring together different areas of expertise, both at scientific and industrial level and to create a network that can have an important impact in EU social and economical challenges. Furthermore, it is an agile funding scheme as the procedures required for funding application and supervision are simplified.

For more information

Visit the FastICpix project site.

Get latest news from ATTRACT