SPAD-Camera

SPAD-Camera

Many applications require the detection of two-dimensional images with very high sensitivity and a sufficient number of pixels. For the most demanding applications, e.g. Fluorescence Lifetime IMaging (FLIM), micro-array-based biological analysis and confocal microscopy, it is compulsory to have single-photon sensitivity in the visible and near infrared (400 nm – 850 nm) wavelength range. The number of pixels should be sufficiently high, though not as high as in digital cameras: usually some hundreds of pixels could suffice.
The standard CCDs (Charge Coupled Devices) can have many pixels (some millions), but they suffer of a notoriously slow read-out. CMOS Active Pixel Sensors (APS) albeit being fast are limited to poor sensitivity since they do not employ any internal amplification process. Electron-Multiplying EM-CCDs can reach single photon sensitivity, but they are not fast enough for many applications. Instead arrays of Single Photon Avalanche Diodes (SPADs) prove to be ideal candidates when very high sensitivity is required together with high-frame rate. Unlike Photo-Multiplier Tubes (PMTs) and Micro-Channel Plates (MCPs), SPAD arrays are insensitive to magnetic fields, require lower bias voltages, and can be easily integrated with complex electronics using standard CMOS processes. Some applications have already been published, e.g. in confocal microscopy for studying intracellular process or adaptive optics for Extremely Large Telescopes.
SPAD arrays so far reported in literature are either composed by a low number of pixels operated in parallel or by a large number of pixels that are multiplexed. Both solutions are unsuitable for the aforementioned applications. In particular, multiplexed acquisition does not exploit high frame rate capability and inherent parallelism of SPAD arrays.

SPAD-Camera sketch
We have developed a two-dimensional (2-D) imaging array of single-photon detectors, that provides a total of 1024 pixels, laid out in 32 rows by 32 columns array, integrated within a monolithic silicon chip with dimensions of 3.5 mm x 3.5 mm. We employed a standard high-voltage 0.35 μm CMOS fabrication technology. The pixels are the same as in the linear array so that every pixel is a completely independent photon-counter. The imager global electronics employ a two-level pipeline architecture for minimizing the inter-frame dead-time to as low as 20 ns. The integration time is limited by the time necessary to perform the complete array readout. Using a convenient system clock of 100 MHz every pixel is read in 10-ns thus resulting in a maximum acquisition speed of about 100 kframe/s for all the 1024 pixels. The system logic allows the selection of a user programmable array sub-area, the less are the selected pixels the higher can be the maximum frame-rate. Finally, for applications exploiting synchronous excitation, we equipped the imager with a gating input pin GATE, able to quickly gate on/off the integration (sub-nanosecond transitions).
The devised gating mechanism makes it possible the accumulation of photon counts in different gating time windows, though belonging to the same frame window. In applications requiring gated measurements, such a feature reduces post-processing complexity as the accumulation is performed directly at the detector-level; this can be a important advantage in very large single-photon counting array systems, when huge amount of data must be managed, e.g. in some adaptive optics applications. Eventually the imager is a general-purpose variable-area imager for singlephoton, high-speed applications both for free-running and gated-mode acquisitions. It can operate as a stand-alone imager chip (with 1 kpixels) or used as the building block of much larger arrays (e.g. 1 Mpixels), by simply placing chips side by side (e.g. 10 x 10 self standing imager chips) and delivering data to the external processor through just one time-multiplexed bus.

The software and the Camera

We already developed a “SPAD camera” based on the 2D imager. The camera can be used in three different modalities: the single shot high-speed mode, the free-running mode and finally the live mode.

Pulications

  • F. Guerrieri, L. Maccone, F. N. C. Wong, J. H. Shapiro, S. Tisa, F. Zappa
    "Sub-Rayleigh Imaging via N-Photon Detection"
    Quantum Electronics and Laser Science Conference (QELS), San Jose (USA), 2010.
  • X. Michalet et al.
    "High-throughput single-molecule fluorescence spectroscopy using parallel detection"
    SPIE Photonics West 2010 conference proceedings, San Jose (USA), 2010. Invited Paper.
  • F. Guerrieri, S. Tisa, A. Tosi, F. Zappa
    "Single-Photon Camera for high-sensitivity high-speed applications"
    IS&T/SPIE Electronic Imaging 2010 conference proceeding, San Jose (USA), 2010
  • F. Guerrieri, S. Tisa, F. Zappa, F. N. C. Wong
    "Single-Photon Imager with Number-Resolving Capability"
    Single-Photon Workshop 2009, Boulder (Colorado, USA), 2009
  • F. Guerrieri, S. Tisa, and F. Zappa:
    "Fast Single-Photon Imager acquires 1024 pixels at 100 kframe/s"
    IS&T/SPIE Electronic Imaging 2010 conference proceeding, San Jose (USA), 2009

Contacts

Franco Zappa - zappa@elet.polimi.it
Simone Tisa - tisa@elet.polimi.it