Functional near-infrared spectroscopy (fNIRS) and diffuse correlation spectroscopy (DCS) have shown promise as non-invasive optical methods for cerebral functional imaging. DCS approaches currently have limited sensitivity in adults. fNIRS sensitivity is also limited, particularly in high-detector-density applications. Sensitivity can be improved using temporal discrimination (TD), where the laser excitation is of short (~400ps) duration and the detector rejects early photons that have not penetrated into the brain while maintain high sensitivity to those that have. We report here on the development of a novel 32x32 Single-Photon Avalanche photo-Detector (SPAD) array and Read-Out Integrated Circuit (ROIC) that can operate in either the visible or NIR enabling high-channel-count TD-fNIRS or TD-DCS systems.
Integrated technologies represent a key enabling capability for future compact and portable atomic physics systems, including optical clocks and other sensors. In this talk, I will discuss our recent demonstration of high-fidelity detection of the state of a trapped Sr+ ion with a single-photon avalanche detector (SPAD) integrated into a microfabricated surface-electrode trap. Using an adaptive technique, we achieve ion state detection in 450 us with 99.92(1)% average fidelity. I will also discuss ongoing efforts to combine integrated detectors with integrated photonics to enable ion traps that completely eliminate the need for free-space optics for light delivery and collection.
Arrays of Geiger-mode avalanche photodiodes (GmAPDs) are fabricated on a new type of engineered substrates with an epitaxial layer grown on silicon-on-insulator (SOI) wafers. The SOI-based structure facilitates rapid die-level bump bonding of the GmAPD array to a CMOS readout integrated circuit (ROIC) followed by substrate removal to make a backilluminated image sensor. To fabricate the engineered substrate, a commercial substrate with a 70-nm-thick SOI layer is implanted with BF2 ions to create a p+-doped passivation layer on the light illumination surface. Subsequently, a lightly p-doped silicon layer on which the GmAPD will be fabricated is grown using a homoepitaxy process. This approach allows for the use of chip-level hybridization to CMOS, avoiding the high cost and demanding wafer flatness and smoothness requirements of wafer-scale 3D integration processes. The new process yields cleaner wafers and allows for tighter control of detector layer thickness compared to the previous process. GmAPDs fabricated on 5-μm-thick epitaxial silicon have over 70% photon detection efficiency (PDE) when 532 nm light is focused into the center 3 μm of the device with an oxide layer that remains after substrate removal. With an anti-reflective coating, the PDE can be improved.
We have developed 2048x32 arrays of silicon Geiger-mode avalanche photodiodes (GmAPDs) for a terrain mapping lidar to be used in planetary exploration missions. These devices support single photon detection with sub-ns timing. A key performance-limiting issue with arrays is optical crosstalk, in which hot-carrier light emission produced by an avalanche triggers spurious detection events in nearby pixels. To address this challenge, we have demonstrated a deep trench isolation process. This paper will report measurements of crosstalk between pixels in silicon GmAPD arrays, measurements both before and after the arrays are hybridized to a readout integrated circuit. Initial wafer probe measurements before hybridization show order-of-magnitude crosstalk reduction in a pair of test GmAPDs separated by 25 µm. These measurements use a novel technique based on analysis of the statistics of the time difference between the first detection events during a bias pulse. The results of our measurements are consistent with simulations of optical crosstalk modeled using optical ray tracing software.
We present the design of an innovative time-gated 32×32 InP/InGaAs-based Single Photon Avalanche Diode (SPAD) array with sub-nanosecond gating capabilities operating up to 10MHz repetition rate specifically designed for time-domain diffuse correlation spectroscopy at 1064nm. We present the detector design, experimental characterization and preliminary validations on a liquid phantom. This testing is informing us for a revision of the photodetector which will allow to reach up to 192 optical channels towards functional blood flow changes measurements with full head coverage with improved brain sensitivity thanks to early-photons rejection.
Jet Propulsion Laboratory is developing a Europa Lander astrobiology mission concept to search for biosignatures within Europa’s subsurface. However, Europa’s rugged terrain presents a number of physical hazards for landing. MIT Lincoln Laboratory is designing a radiation-hardened real-time direct-detection LIDAR system at 532nm to aid with autonomous hazard avoidance and landing site selection for this Europa Lander concept. The detector for this system is a 2048x32 array of silicon Geiger-mode APDs and covers the required field-of-view in one dimension, removing the need for 2D stitching and enabling real-time data processing. Detector design, improvements for radiation tolerance and component characterization results are presented.
We present the design of an innovative instrument for time-gated diffuse correlation spectroscopy. It features a 1064nm pulsed sub-ns long coherence-length laser operating up to 75MHz, a 100-channel in-FPGA correlator and a novel time-gated 32×32 InP/InGaAs-based Single Photon Avalanche Diode (SPAD) array with sub-nanosecond gating capabilities operating up to 10MHz repetition rate. We present components experimental characterization and preliminary validations on a liquid phantom. This testing is informing us for a revision of the photodetector which will allow to reach up to 192 optical channels towards functional blood flow changes measurements with full head coverage.
A system-level performance evaluation of Geiger-mode avalanche photodiode (GmAPD) arrays requires accurate measurement and prediction of the background rate of the device due to dark counts and other spurious detection events. Since a GmAPD detector reports only a binary value and timestamp associated with an avalanche event, dark count rates are typically measured by averaging thousands of frames to support a statistically significant measurement. For both synchronous and asynchronous detector, the Poisson distributed background rates are referenced to the time each pixel is armed. Unlike for synchronous GmAPD imagers where all the pixels are armed to an array-wide arm signal, an asynchronous pixel operates independently from its neighboring pixels; requiring the background rates to be calculated using an interarrival histogram. For both types of imagers, the background rate is typically evaluated by fitting an exponential distribution to a fixed window within a measured histogram of time intervals between detection events However, if the statistics of the background rate are insufficient – whether that is due to low population sizes, saturation, or a large dynamic range of population size across the array, the pixel, or array-wide, performance metrics may report results with varying accuracy. This paper reports on an implementation of an algorithm that evaluates GmAPD background rates based on statistical metrics rather than fixed windows. The algorithm functions by determining the appropriate integration window within the interarrival time histogram based on a per-pixel count rate set by a predetermined tolerable measurement error. The implementation of the algorithm allows us to characterize GmAPD arrays with orders of magnitude spread in background rates across the detector using common statistical parameters.
Our team has recently shown the SNR and depth-sensitivity advantages of using 1064 nm light for diffuse correlation spectroscopy as well as the challenges of commercially available single-photon detectors at this wavelength. We will review two strategies for custom readout integrated circuit designs that simultaneously target lower pixel dead times and lower afterpulsing probabilities. Both designs use macropixels comprising many detectors, each having a programmable hold-off time. We will compare simulated autocorrelations for our detector models and compare predicted performance against commercial InGaAs/InP detectors.
We have developed a new approach for rapid die-level hybridization of backside-illuminated silicon avalanche photodiode (APD) arrays to CMOS readout integrated circuits (ROICs). APD arrays are fabricated on a custom silicon-on-insulator (SOI) wafer engineered with a built-in backside contact and passivation layer. The engineered APD substrate structure facilitates uniform APD substrate removal by selective etching at the die level after bump bonding. The new integration process has the following advantages over wafer-level 3D integration: 1) reduced cost per development cycle since a dedicated full-wafer ROIC fabrication is not needed, 2) compatibility with existing ROICs that are in chip-format from previous fabrication runs, and 3) accelerated schedule. The new approach is applied to produce 32×32 100-μm-pitch silicon GmAPD arrays. Electrical performance of the APD arrays show 100% pixel connectivity and excellent yield before and after substrate removal.
Over the past 20 years, we have developed arrays of custom-fabricated silicon and InP Geiger-mode avalanche photodiode arrays, CMOS readout circuits to digitally count or time stamp single-photon detection events, and techniques to integrate these two components to make back-illuminated solid-state image sensors for lidar, optical communications, and passive imaging. Starting with 4 × 4 arrays, we have recently demonstrated 256 × 256 arrays, and are working to scale to megapixel-class imagers. In this paper, we review this progress and discuss key technical challenges to scaling to large format.
Geiger-mode avalanche photodiodes (GMAPDs) are capable of detecting single photons. They can be operated to directly trigger all-digital circuits, so that detection events are digitally counted or time stamped in each pixel. An imager based on an array of GMAPDs therefore has zero readout noise, enabling quantum-limited sensitivity for photon-starved imaging applications. In this review, we discuss devices developed for 3D imaging, wavefront sensing, and passive imaging.
KEYWORDS: Sensors, Avalanche photodetectors, Image sensors, Signal to noise ratio, Signal detection, Electrons, Silicon, Charge-coupled devices, Single photon, Photodetectors
Single-photon imaging detectors promise the ultimate in sensitivity by eliminating read noise. These devices could
provide extraordinary benefits for photon-starved applications, e.g., imaging exoplanets, fast wavefront sensing, and
probing the human body through transluminescence. Recent implementations are often in the form of sparse arrays that
have less-than-unity fill factor. For imaging, fill factor is typically enhanced by using microlenses, at the expense of
photometric and spatial information loss near the edges and corners of the pixels. Other challenges include afterpulsing
and the potential for photon self-retriggering. Both effects produce spurious signal that can degrade the signal-to-noise
ratio. This paper reviews development and potential application of single-photon-counting detectors, including highlights
of initiatives in the Center for Detectors at the Rochester Institute of Technology and MIT Lincoln Laboratory.
Current projects include single-photon-counting imaging detectors for the Thirty Meter Telescope, a future NASA
terrestrial exoplanet mission, and imaging LIDAR detectors for planetary and Earth science space missions.
This paper summarizes progress of a project to develop and advance the maturity of photon-counting detectors for
NASA exoplanet missions. The project, funded by NASA ROSES TDEM program, uses a 256×256 pixel silicon Geigermode
avalanche photodiode (GM-APD) array, bump-bonded to a silicon readout circuit. Each pixel independently
registers the arrival of a photon and can be reset and ready for another photon within 100 ns. The pixel has built-in
circuitry for counting photo-generated events. The readout circuit is multiplexed to read out the photon arrival events.
The signal chain is inherently digital, allowing for noiseless transmission over long distances. The detector always
operates in photon counting mode and is thus not susceptible to excess noise factor that afflicts other technologies. The
architecture should be able to operate with shot-noise-limited performance up to extremely high flux levels,
>106 photons/second/pixel, and deliver maximum signal-to-noise ratios on the order of thousands for higher fluxes. Its
performance is expected to be maintained at a high level throughout mission lifetime in the presence of the expected
radiation dose.
We have demonstrated a wafer-scale back-illumination process for silicon Geiger-mode avalanche photodiode arrays
using Molecular Beam Epitaxy (MBE) for backside passivation. Critical to this fabrication process is support of the thin
(< 10 μm) detector during the MBE growth by oxide-bonding to a full-thickness silicon wafer. This back-illumination
process makes it possible to build low-dark-count-rate single-photon detectors with high quantum efficiency extending
to deep ultraviolet wavelengths. This paper reviews our process for fabricating MBE back-illuminated silicon Geigermode
avalanche photodiode arrays and presents characterization of initial test devices.
For adaptive optics systems, there is a growing demand for wavefront sensors that operate at higher frame rates and with
more pixels while maintaining low readout noise. Lincoln Laboratory has been investigating Geiger-mode avalanche
photodiode arrays integrated with CMOS readout circuits as a potential solution. This type of sensor counts photons
digitally within the pixel, enabling data to be read out at high rates without the penalty of readout noise. After a brief
overview of adaptive optics sensor development at Lincoln Laboratory, we will present the status of silicon Geigermode-
APD technology along with future plans to improve performance.
We present a unique hybridization process that permits high-performance back-illuminated silicon Geiger-mode
avalanche photodiodes (GM-APDs) to be bonded to custom CMOS readout integrated circuits (ROICs) - a hybridization
approach that enables independent optimization of the GM-APD arrays and the ROICs. The process includes oxide
bonding of silicon GM-APD arrays to a transparent support substrate followed by indium bump bonding of this layer to
a signal-processing ROIC. This hybrid detector approach can be used to fabricate imagers with high-fill-factor pixels and
enhanced quantum efficiency in the near infrared as well as large-pixel-count, small-pixel-pitch arrays with pixel-level
signal processing. In addition, the oxide bonding is compatible with high-temperature processing steps that can be used
to lower dark current and improve optical response in the ultraviolet.
Mohan Vaidyanathan, Steven Blask, Thomas Higgins, William Clifton, Daniel Davidsohn, Ryan Carson, Van Reynolds, Joanne Pfannenstiel, Richard Cannata, Richard Marino, John Drover, Robert Hatch, David Schue, Robert Freehart, Greg Rowe, James Mooney, Carl Hart, Byron Stanley, Joseph McLaughlin, Eui-In Lee, Jack Berenholtz, Brian Aull, John Zayhowski, Alex Vasile, Prem Ramaswami, Kevin Ingersoll, Thomas Amoruso, Imran Khan, William Davis, Richard Heinrichs
KEYWORDS: Sensors, LIDAR, 3D image processing, 3D acquisition, Target detection, Imaging systems, Image processing, Control systems, Image sensors, Data processing
Jigsaw three-dimensional (3D) imaging laser radar is a compact, light-weight system for imaging
highly obscured targets through dense foliage semi-autonomously from an unmanned aircraft. The
Jigsaw system uses a gimbaled sensor operating in a spot light mode to laser illuminate a cued
target, and autonomously capture and produce the 3D image of hidden targets under trees at high 3D
voxel resolution. With our MIT Lincoln Laboratory team members, the sensor system has been
integrated into a geo-referenced 12-inch gimbal, and used in airborne data collections from a UH-1
manned helicopter, which served as a surrogate platform for the purpose of data collection and
system validation. In this paper, we discuss the results from the ground integration and testing of the
system, and the results from UH-1 flight data collections. We also discuss the performance results
of the system obtained using ladar calibration targets.
Lincoln Laboratory has developed 32×32-pixel ladar focal planes comprising silicon Geiger-mode avalanche photodiodes and high-speed all-digital CMOS timing circuitry in each pixel. In Geiger-mode operation, the APD can detect as little as a single photon, producing a digital CMOS-compatible voltage pulse. This pulse is used to stop a high-speed counter in the pixel circuit, thus digitizing the time of arrival of the optical pulse. This "photon-to-digital conversion" simultaneously achieves single-photon sensitivity and 0.25-ns timing precision. We discuss the development of these focal planes and present imagery from ladar systems that use them.
Geiger-mode avalanche photodiodes (APDs) can convert the arrival of a single photon into a digital logic pulse. Arrays of APDs can be directly interfaced to arrays of per-pixel digital electronics fabricated in silicon CMOS, providing the capability to time the arrival of photons in each pixel. These arrays are of interest for "flash" LADAR systems, where multiple target pixels are simultaneously illuminated by the laser during a single laser pulse, and the imaging array is used to measure range to each of the illuminated pixels. Since many laser radar systems use Nd:YAG lasers operating at 1.06 um, we have extended our earlier work with silicon-based APDs by developing arrays of InGaAsP/InP APDs, which are efficient detectors for near-IR radiation. 32x32 pixel arrays, with 100-um pixel pitches, are currently being successfully used in demonstration systems.
Brian Aull, Andrew Loomis, Douglas Young, Alvin Stern, Bradley Felton, Peter Daniels, Debbie Landers, Larry Retherford, Dennis Rathman, Richard Heinrichs, Richard Marino, Daniel Fouche, Marius Albota, Robert Hatch, Gregory Rowe, David Kocher, James Mooney, Michael O'Brien, Brian Player, Berton Willard, Zong-Long Liau, John Zayhowski
Lincoln Laboratory has developed 32 x 32-pixel ladar focal planes comprising silicon geiger-mode avalanche photodiodes and high-speed all-digital CMOS timing circuitry in each pixel. In Geiger mode operation, the APD can detect as little as a single photon, producing a digital CMOS-compatible voltage pulse. This pulse is used to stop a high-speed counter in the pixel circuit, thus digitizing the time of arrival of the optical pulse. This "photon-to-digital conversion" simultaneously achieves single-photon sensitivity and 0.5-ns timing. We discuss the development of these focal planes and present imagery from ladar systems that use them.
MIT Lincoln Laboratory is actively developing laser and detector technologies that make it possible to build a 3D laser radar with several attractive features, including capture of an entire 3D image on a single laser pulse, tens of thousands of pixels, few-centimeter range resolution, and small size, weight, and power requirements. The laser technology is base don diode-pumped solid-state microchip lasers that are passively Q-switched. The detector technology is based on Lincoln-built arrays of avalanche photodiodes operating in the Geiger mode, with integrated timing circuitry for each pixel. The advantage of these technologies is that they offer the potential for small, compact, rugged, high-performance systems which are critical for many applications.
An integrated optoelectronic device, the monolithic optoelectronic transistor (MOET), has been demonstrated. The MOET functions as an optical sum-and-threshold device with large- signal optical gain. It can be electrically biased to achieve either abrupt switching thresholds or quasi-sigmoidal optical transfer characteristics, and excitatory and inhibitory inputs can be incorporated through a simple modification of the single-input device. Initial MOET devices displayed an optical gain greater than 10 and an output contrast ratio exceeding 50. The MOET has promising characteristics as a building block of optoelectronically implemented neural networks and image preprocessing systems.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.