Single-photon Avalanche Diode Sensor Research Articles

A Gradient-Gated SPAD Array for Non-Line-of-Sight Imaging

Time-resolved non-line-of-sight (NLOS) imaging based on single-photon avalanche diode (SPAD) detectors have demonstrated impressive results in recent years. To acquire adequate number of indirect photons from a hidden scene in the presence of overwhelming amount of early-arrival photons, a single-gated SPAD is widely employed to mitigate pile-up. However, additional prior knowledge of the hidden range and relay surface profile are required to preset the gating position and implement the reconstruction. With this work, we propose a gradient-gated technique to alleviate these shortcomings (pile-up and prior knowledge) in NLOS imaging with the development of a 6 × 6 SPAD array fabricated in a standard 55-nm Bipolar-CMOS-DMOS (BCD) technology. The SPAD sensor includes a delay-locked loop (DLL) block, a gating generator and a pixel array, which can be flexibly configured to free-running, single-gating and gradient-gating modes. The rise time of the gates is less than 450 ps (from 10% to 90%). In gradient-gating mode, the interval time between adjacent gates can be configured from 300 ps to 2 ns. In single-gating mode, the minimum gate window is measured below 5 ns. We demonstrated the sensor in a confocal NLOS imaging system that reconstructs hidden scenes with both retroreflective and diffuse objects. The presented scheme enables the development of calibration-free NLOS imaging modality for practical applications.

Multi-Scale Histogram-Based Probabilistic Deep Neural Network for Super-Resolution 3D LiDAR Imaging

LiDAR (Light Detection and Ranging) imaging based on SPAD (Single-Photon Avalanche Diode) technology suffers from severe area penalty for large on-chip histogram peak detection circuits required by the high precision of measured depth values. In this work, a probabilistic estimation-based super-resolution neural network for SPAD imaging that firstly uses temporal multi-scale histograms as inputs is proposed. To reduce the area and cost of on-chip histogram computation, only part of the histogram hardware for calculating the reflected photons is implemented on a chip. On account of the distribution rule of returned photons, a probabilistic encoder as a part of the network is first proposed to solve the depth estimation problem of SPADs. By jointly using this neural network with a super-resolution network, 16× up-sampling depth estimation is realized using 32 × 32 multi-scale histogram outputs. Finally, the effectiveness of this neural network was verified in the laboratory with a 32 × 32 SPAD sensor system.

Single Photon Kilohertz Frame Rate Imaging of Neural Activity.

Establishing the biological basis of cognition and its disorders will require high precision spatiotemporal measurements of neural activity. Recently developed genetically encoded voltage indicators (GEVIs) report both spiking and subthreshold activity of identified neurons. However, maximally capitalizing on the potential of GEVIs will require imaging at millisecond time scales, which remains challenging with standard camera systems. Here, application of single photon avalanche diode (SPAD) sensors is reported to image neural activity at kilohertz frame rates. SPADs are electronic devices that when activated by a single photon cause an avalanche of electrons and a large electric current. An array of SPAD sensors is used to image individual neurons expressing the GEVI Voltron-JF525-HTL. It is shown that subthreshold and spiking activity can be resolved with shot noise limited signals at frame rates of up to 10kHz. SPAD imaging is able to reveal millisecond scale synchronization of neural activity in an ex vivo seizure model. SPAD sensors may have widespread applications for investigation of millisecond timescale neural dynamics.

Smart Wide-field Fluorescence Lifetime Imaging System with CMOS Single-photon Avalanche Diode Arrays.

Wide-field fluorescence lifetime imaging (FLIM) is a promising technique for biomedical and clinic applications. Integrating with CMOS single-photon avalanche diode (SPAD) sensor arrays can lead to cheaper and portable real-time FLIM systems. However, the FLIM data obtained by such sensor systems often have sophisticated noise features. There is still a lack of fast tools to recover lifetime parameters from highly noise-corrupted fluorescence signals efficiently. This paper proposes a smart wide-field FLIM system containing a 192×128 COMS SPAD sensor and a field-programmable gate array (FPGA) embedded deep learning (DL) FLIM processor. The processor adopts a hardware-friendly and light-weighted neural network for fluorescence lifetime analysis, showing the advantages of high accuracy against noise, fast speed, and low power consumption. Experimental results demonstrate the proposed system's superior and robust performances, promising for many FLIM applications such as FLIM-guided clinical surgeries, cancer diagnosis, and biomedical imaging.

Time-resolved spectral-domain optical coherence tomography with CMOS SPAD sensors.

We present a first spectral-domain optical coherence tomography (SD-OCT) system deploying a complementary metal-oxide-semiconductor (CMOS) single-photon avalanche diode (SPAD) based, time-resolved line sensor. The sensor with 1024 pixels achieves a sensitivity of 87 dB at an A-scan rate of 1 kHz using a supercontinuum laser source with a repetition rate of 20 MHz, 38 nm bandwidth, and 2 mW power at 850 nm centre wavelength. In the time-resolved mode of the sensor, the system combines low-coherence interferometry (LCI) and massively parallel time-resolved single-photon counting to control the detection of interference spectra on the single-photon level based on the time-of-arrival of photons. For proof of concept demonstration of the combined detection scheme we show the acquisition of time-resolved interference spectra and the reconstruction of OCT images from selected time bins. Then, we exemplify the temporal discrimination feature with 50 ps time resolution and 249 ps timing uncertainty by removing unwanted reflections from along the optical path at a 30 mm distance from the sample. The current limitations of the proposed technique in terms of sensor parameters are analysed and potential improvements are identified for advanced photonic applications.

Dynamic fluorescence lifetime sensing with CMOS single-photon avalanche diode arrays and deep learning processors.

Measuring fluorescence lifetimes of fast-moving cells or particles have broad applications in biomedical sciences. This paper presents a dynamic fluorescence lifetime sensing (DFLS) system based on the time-correlated single-photon counting (TCSPC) principle. It integrates a CMOS 192 × 128 single-photon avalanche diode (SPAD) array, offering an enormous photon-counting throughput without pile-up effects. We also proposed a quantized convolutional neural network (QCNN) algorithm and designed a field-programmable gate array embedded processor for fluorescence lifetime determinations. The processor uses a simple architecture, showing unparallel advantages in accuracy, analysis speed, and power consumption. It can resolve fluorescence lifetimes against disturbing noise. We evaluated the DFLS system using fluorescence dyes and fluorophore-tagged microspheres. The system can effectively measure fluorescence lifetimes within a single exposure period of the SPAD sensor, paving the way for portable time-resolved devices and shows potential in various applications.

Compact SPAD pixels with fast and accurate photon counting in the analog domain

A compact pixel for single-photon detection in the analog domain is presented. The pixel integrates a single-photon avalanche diode (SPAD), a passive quenching & active recharging circuit (PQARC), and an analog counter for fast and accurate sensing and counting of photons. Fabricated in a standard 0.18 µm CMOS technology, the simulated and experimental results reveal that the dead time of the PQARC is about 8 ns and the maximum photon-counting rate can reach 125 Mcps (counting per second). The analog counter can achieve an 8-bit counting range with a voltage step of 6.9 mV. The differential nonlinearity (DNL) and integral nonlinearity (INL) of the analog counter are within the ± 0.6 and ± 1.2 LSB, respectively, indicating high linearity of photon counting. Due to its simple circuit structure and compact layout configuration, the total area occupation of the presented pixel is about 1500 μm2, leading to a high fill factor of 9.2%. The presented in-pixel front-end circuit is very suitable for the high-density array integration of SPAD sensors.

Drug diffusivities in nanofibrillar cellulose hydrogel by combined time-resolved Raman and fluorescence spectroscopy

Hydrogels, natural and synthetic origin, are actively studied for their use for implants and payload carriers. These biomaterials for delivery systems have enormous potential in basic biomedical research, drug development, and long-term delivery of biologics. Nanofibrillated cellulose (NFC) hydrogels, both natural and anionic (ANFC) ones, allow drug loading for immediate and controlled release via the slow drug dissolution of solid drug crystals into hydrogel and its subsequent release. This property makes NFC originated hydrogels an interesting non-toxic and non-human origin material as drug reservoir for long-term controlled release formulation or implant for patient care. A compelling tool for studying NFC hydrogels is Raman spectroscopy, which enables to resolve the chemical structures of different molecules in a high-water content like hydrogels, since Raman spectroscopy is insensitive to water molecules. That offers real time investigation of label-free drugs and their release in high-water-content media. Despite the huge potential of Raman spectroscopy in bio-pharmaceutical applications, the strong fluorescence background of many drug samples masking the faint Raman signal has restricted the widespread use of it. In this study we used a Raman spectrometer capable of suppressing the unpleasant fluorescence background by combining a pulsed laser and time-resolved complementary metal-oxide-semiconductor (CMOS) single-photon avalanche diode (SPAD) line sensor for the label-free investigation of Metronidazole and Vitamin C diffusivities in ANFC. The results show the possibility to modulate the ANFC-based implants and drug delivery systems, when the release rate needs to be set to a desired value. More importantly, the now developed label free real-time method is universal and can be adapted to any hydrogel/drug combination for producing reliable drug diffusion coefficient data in complex and heterogeneous systems, where traditional sampling-based methods are cumbersome to use. The wide temporal range of the time-resolved CMOS SPAD sensors makes it possible to capture also the fluorescence decay of samples, giving rise to a combined time-resolved Raman and fluorescence spectroscopy, which provides additional information on the chemical, functional and structural changes in samples.

Time-Delay-Integration Imaging Implemented With Single-Photon-Avalanche-Diode Linear Array

In this paper, we propose a novel kind of time delay integration (TDI) image sensor based on single photon avalanche diode (SPAD), i.e. TDI-SPAD. Compared with existing TDI imaging devices, the proposed TDI-SPAD sensor is characterized by low noise, high sensitivity, high modulation transfer function (MTF) and transfer efficiency. The proposed architecture design of multi-stages TDI-SPAD sensor and the digital transfer circuits are discussed in detail, and the full function of TDI-imaging is demonstrated with experiments based upon two SPAD sensors. First, the imaging function of TDI-SPAD is verified through imaging of scanning scene with a $128\times128$ SPAD array. Then, we fabricate a $256\times2$ TDI-SPAD sensor, build a TDI-imaging camera system, and demonstrate the feasibility and SNR improvement of the proposed TDI-imaging sensor.

High-speed 3D sensing via hybrid-mode imaging and guided upsampling

Imaging systems with temporal resolution play a vital role in a diverse range of scientific, industrial, and consumer applications, e.g., fluorescent lifetime imaging in microscopy and time-of-flight (ToF) depth sensing in autonomous vehicles. In recent years, single-photon avalanche diode (SPAD) arrays with picosecond timing capabilities have emerged as a key technology driving these systems forward. Here we report a high-speed 3D imaging system enabled by a state-of-the-art SPAD sensor used in a hybrid imaging mode that can perform multi-event histogramming. The hybrid imaging modality alternates between photon counting and timing frames at rates exceeding 1000 frames per second, enabling guided upscaling of depth data from a native resolution of 64 × 32 to 256 × 128 . The combination of hardware and processing allows us to demonstrate high-speed ToF 3D imaging in outdoor conditions and with low latency. The results indicate potential in a range of applications where real-time, high throughput data are necessary. One such example is improving the accuracy and speed of situational awareness in autonomous systems and robotics.

Event-Based Processing of Single Photon Avalanche Diode Sensors

Single Photon Avalanche Diode sensor arrays operating in direct time of flight mode can perform 3D imaging using pulsed lasers. Operating at high frame rates, SPAD imagers typically generate large volumes of noisy and largely redundant spatio-temporal data. This results in communication bottlenecks and unnecessary data processing. In this work, we propose a neuromorphic processing solution to this problem. By processing the spatio-temporal patterns generated by the SPADs in a local, event-based manner, the proposed $128\times 128$ pixel sensor-processor system reduces the size of output data from the sensor by orders of magnitude while increasing the utility of the output data in the context of challenging recognition tasks. To test the proposed system, the first large scale complex SPAD imaging dataset is captured using an existing $32\times 32$ pixel sensor. The generated dataset consists of 24000 recordings and involves high-speed view-invariant recognition of airplanes with background clutter. The frame-based SPAD imaging dataset is converted via several alternative methods into event-based data streams and processed using the proposed $125\times 125$ receptive field neuromorphic processor as well as a range of feature extractor networks and pooling methods. The output of the proposed event generation methods are then processed by an event-based feature extraction and classification system implemented in FPGA hardware. The event-based processing methods are compared to processing the original frame-based dataset via frame-based but otherwise identical architectures. The results show the event-based methods are superior to the frame-based approach both in terms of classification accuracy and output data-rate.

Dark Count Rate Degradation in CMOS SPADs Exposed to X-Rays and Neutrons

Single-photon avalanche diode (SPAD) arrays fabricated in a 180-nm CMOS technology with a high-voltage option have been exposed to calibrated neutron and X-ray sources to evaluate their radiation tolerance. The technology is being investigated in view of the design of low material budget detectors for charged particle tracking based on the coincidence of the signals coming from two or more overlapping layers of SPAD sensors. Each element in the array is a monolithic detector including the processing electronics together with the diode in the same substrate. Different sensor dimensions and structures have been implemented in the test chip to thoroughly explore the technology features. This paper will present and discuss the results from the characterization, in terms of dark count rate, of SPAD arrays irradiated with X-ray doses reaching 1 Mrad(SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ) and with neutron fluences up to 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sup> 1-MeV neutron equivalent cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup> .

A Point-of-Care Device for Molecular Diagnosis Based on CMOS SPAD Detectors with Integrated Microfluidics.

We describe the integration of techniques and technologies to develop a Point-of-Care for molecular diagnosis PoC-MD, based on a fluorescence lifetime measurement. Our PoC-MD is a low-cost, simple, fast, and easy-to-use general-purpose platform, aimed at carrying out fast diagnostics test through label detection of a variety of biomarkers. It is based on a 1-D array of 10 ultra-sensitive Single-Photon Avalanche Diode (SPAD) detectors made in a 0.18 μm High-Voltage Complementary Metal Oxide Semiconductor (HV-CMOS) technology. A custom microfluidic polydimethylsiloxane cartridge to insert the sample is straightforwardly positioned on top of the SPAD array without any alignment procedure with the SPAD array. Moreover, the proximity between the sample and the gate-operated SPAD sensor makes unnecessary any lens or optical filters to detect the fluorescence for long lifetime fluorescent dyes, such as quantum dots. Additionally, the use of a low-cost laser diode as pulsed excitation source and a Field-Programmable Gate Array (FPGA) to implement the control and processing electronics, makes the device flexible and easy to adapt to the target label molecule by only changing the laser diode. Using this device, reliable and sensitive real-time proof-of-concept fluorescence lifetime measurement of quantum dot QdotTM 605 streptavidin conjugate is demonstrated.

A CMOS SPAD Imager with Collision Detection and 128 Dynamically Reallocating TDCs for Single-Photon Counting and 3D Time-of-Flight Imaging.

Per-pixel time-to-digital converter (TDC) architectures have been exploited by single-photon avalanche diode (SPAD) sensors to achieve high photon throughput, but at the expense of fill factor, pixel pitch and readout efficiency. In contrast, TDC sharing architecture usually features high fill factor at small pixel pitch and energy efficient event-driven readout. While the photon throughput is not necessarily lower than that of per-pixel TDC architectures, since the throughput is not only decided by the TDC number but also the readout bandwidth. In this paper, a SPAD sensor with 32 × 32 pixels fabricated with a 180 nm CMOS image sensor technology is presented, where dynamically reallocating TDCs were implemented to achieve the same photon throughput as that of per-pixel TDCs. Each 4 TDCs are shared by 32 pixels via a collision detection bus, which enables a fill factor of 28% with a pixel pitch of 28.5 μm. The TDCs were characterized, obtaining the peak-to-peak differential and integral non-linearity of −0.07/+0.08 LSB and −0.38/+0.75 LSB, respectively. The sensor was demonstrated in a scanning light-detection-and-ranging (LiDAR) system equipped with an ultra-low power laser, achieving depth imaging up to 10 m at 6 frames/s with a resolution of 64 × 64 with 50 lux background light.

CMOS Single-Photon Avalanche Diode Pixel Design for a Gun Muzzle Flash Detection Camera

We present the architecture and design of a CMOS single-photon avalanche diode (SPAD) pixel that was selected to be the basis for a gun muzzle flash detection camera. The SPAD sensor and auxiliary circuitry are fabricated in a standard $0.18-\mu \text{m}$ CMOS image sensor technology. The pixel integrates a $25-\mu \text{m}$ pitch SPAD, a variable-load quenching circuit implemented with a 1.8-V pMOS, and digital processing electronics providing an 8-bit output bus. The SPAD is low noise (around 100-Hz dark count rate at 1.8-V excess voltage) and has a real peak photon detection efficiency (not averaged on pixel pitch) of 9.2% at 450 nm (1.8-V excess voltage). The pixel delivers intensity information through photon counting, up to 256 counts per frame with down to $5-\mu \text{s}$ integration time for the full dynamic range. The pixel memories enable parallel processing and global-shutter readout, preventing motion artifacts and partial exposure effects. The pixel can acquire very fast optical events at a high frame-rate (up to 200 kilo frames/s) and at a single-photon level. The pixel has an 8-bit parallel output bus.

LinoSPAD: A Compact Linear SPAD Camera System with 64 FPGA-Based TDC Modules for Versatile 50 ps Resolution Time-Resolved Imaging

The LinoSPAD camera system is a modular, compact and versatile time-resolved camera system, combining a linear 256 high fill factor pixel CMOS SPAD (single-photon avalanche diode) sensor with an FPGA (field-programmable gate array) and USB 3.0 transceiver board. This modularization permits the separate optimization or exchange of either the sensor front-end or the processing back-end, depending on the intended application, thus removing the traditional compromise between optimal SPAD technology on the one hand and time-stamping technology on the other hand. The FPGA firmware implements an array of 64 TDCs (time-to-digital converters) with histogram accumulators and a correction module to reduce non-linearities. Each TDC is capable of processing over 80 million photon detections per second and has an average timing resolution better than 50 ps. This article presents a complete and detailed characterization, covering all aspects of the system, from the SPAD array light sensitivity and noise to TDC linearity, from hardware/firmware/software co-design to signal processing, e.g., non-linearity correction, from power consumption to performance non-uniformity.

A 15 × 15 single photon avalanche diode sensor featuring breakdown pixels extraction architecture for efficient data readout

This paper proposes a breakdown pixels extraction architecture for single photon avalanche diode (SPAD) based faint light detection systems. The proposed extraction circuit detects the breakdown pixels and only their addresses are readout. Therefore, under the faint light environment, this SPAD-based sensor significantly improves the data readout efficiency. In addition, since the readout sequence is 4× faster than that of the conventional architecture in the dark condition, the proposed system does not need an independent on-chip event detection circuit that consumes additional area and power. A test-of-concept chip with a 15 × 15 SPAD pixels array was fabricated in a 0.18 µm 1P5M standard CMOS process and pinhole diffraction patterns were successfully captured thanks to the high sensitivity of the SPAD sensor. Under the faint light condition, a high-speed readout is verified by measurement and the robustness of the proposed architecture is successfully demonstrated.

A Monolithic Implementation of Active Inductor VLQC and Jitter Free Photon Current Detection for Bio-Luminescence SPAD Sensor

An active inductor variable load quenching circuit monolithically implemented for sensing chlorophyll fluorescence and other sensor applications like bioluminescence, tomography, and DNA fingerprints is reported for the first time in this paper. An efficient biosignal acquisition with single-photon counting facility and fast quenching and reset for Single-Photon Avalanche Diode (SPAD) is implemented in 120 nm technology. This is the first ever monolithic active inductor implementation reported for quenching and reset of SPAD pixel. It is faster compared to all previously reported AQRC and variable load quenching circuit (VLQC) monolithic implementations. Along with the principle, a detailed design is presented in 120 nm technology. Analog simulations are done using LT Spice version IV and layout is primed using Microwind 3.1. The results reveals that a 50 ps light pulse with 250 ps dead time, i.e., light pulse with repetition rate of 300 ps can be quenched and detected.