1479
High-Performance InP-Based Geiger-Mode Avalanche Photodiodes and Integration Technologies

Monday, May 12, 2014: 13:20
Gilchrist, Ground Level (Hilton Orlando Bonnet Creek)
X. Jiang, M. A. Itzler, M. Entwistle, and K. Slomkowski (Princeton Lightwave Inc.)
In recent years there has been a growing effort to develop alternative solid state detectors to replace photomultiplier tubes (PMTs) which retain a dominant role in low light-level detection. In the short-wave infrared (SWIR) spectral band, InP based single photon avalanche diodes (SPADs), also referred to as Geiger-mode APDs, provide a practical solution for single photon detection due to their high performance, reliability, compactness and low cost.  Significant effort has been made during the last few years to improve the performance of InP based SPADs [1]. To address various applications, different integration technologies have been developed as well to fabricate various discrete and array format Geiger-mode APDs.

In Figure 1, we illustrate the dependence of DCR on photon detection efficiency (PDE) for a fiber-coupled detector operated with 1.0 ns gates at 100 MHz gate repetition rate for optical pulses with a mean photon number of 0.1 at 1.5 μm photons and an operating temperature of -50°C.  At 20% detection efficiency, DCR is less than 0.4 kHz, and at this operating point, the device exhibits an afterpulsing probability of <2% for a 1-ns gate with a hold-off time of 10 ns.  Advances in integrated circuitry for operating the SPADs also allows PDE well beyond 30%.

Discrete InP based SPADs work as a binary switch and do not have the capability of photon number resolution. For certain applications, it is important to resolve the number of incident photons. To address this need, we have developed a novel integration technology to fabricate negative feedback avalanche diode (NFAD) solid state photomultipliers (SSPMs) by monolithically integrating a SPAD with a thin film quenching resistor and connecting multiple elements in parallel which share common cathode and anode connections (i.e., matrix devices) [2]. When a low-light level signal is present and the probability of multiple photons impinging on the same active region is low, the pulse height from an NFAD matrix will be proportional to the number of incident photons. Figure 2 shows the avalanche distributions from a 4x4 NFAD matrix device, and a micrograph of the device is shown in the inset. As can be seen from Figure 2, the 4´4 matrix device can provide quasi-analog output for multi-photon inputs. The operation of discrete NFAD devices and NFAD matrix devices is very simple since only a DC bias voltage is required. The devices developed through this integration technology provide a unique solution to applications where single photon sensitivity and photon number resolution capability in the SWIR region is critical.

To address a large class of 3D imaging and LIDAR applications, over the last few years we have developed large format (32x32, 128x32) Geiger-mode APD-based sensors through hybrid integration technology. A schematic illustration of the 3D GmAPD focal plane array (FPA) architecture is given in Figure 3. The 3D GmAPD FPA consists of the hybridized chip stack with a CMOS readout integrated circuit (ROIC), Geiger-mode APD photodiode array (PDA), and microlens array (MLA) attached to a ceramic interposer for electrical signal routing and cooled by a thermoelectric cooler (TEC). Single photon detection is enabled by a photodiode array (PDA) with GmAPD detectors in each pixel.  The ROIC provides pixel-level electrical interfacing, and the MLA ensures high fill factor. The histogram of PDE for a 1.06 µm 128x32 GmAPD array is shown in Figure 4. The PDE was taken at 50 kHz frame rate with 0.1 photon per pixel per illumination pulse for a total of 25,000 frames.  The average PDE is 25.1% and the average dark count rate (DCR) is 3.0 kHz under this bias condition. The 128x32 array demonstrated >99.9% pixel yield, high PDE, low DCR and narrow distribution of DCR and PDE. These FPAs developed through hybrid integration technology serve as the sensor engine for 3D imaging LIDAR systems [3].

References

[1] M. A. Itzler, et al., J. Modern Optics, 58, 174 (2011).

[2] X. Jiang, et al., Proc. of  SPIE 8375, 83750U (2012).

[3] M. Entwistle, et al., Proc. of the SPIE 8375, 83750D (2012).