AVALANCHE PHOTODIODES: Near-IR APDs expand photon-counting applications

Aug. 1, 2006
A variety of new technologies provide efficient single-photon detection in near-IR wavelengths, enabling new applications in trace-gas analysis, photodynamic therapy, and quantum communications.

KEITH FORSYTH AND NOAH CLAY

Although photon counting is a common technique at optical wavelengths shorter than 1 µm, it is not widely used at longer wavelengths. Among the applications that can benefit from photon-counting techniques in the 1 to 2 µm wavelength band are Raman light detection and ranging (lidar) for fluorescence-free trace-gas detection, laser radar (ladar) imaging with eye-safe laser diodes, singlet oxygen detection for photodynamic therapy dosimetry, microprocessor failure analysis by time-resolved single-photon-emission microscopy, and quantum-key distribution for secure communications.

Photon-counting techniques are not more widely used in the near-IR (NIR) region primarily because of the difficulty in obtaining suitable detectors. However, detectors sensitive over a substantial portion of the 1 to 2 µm NIR spectral band are availableeach with its own advantages and disadvantages. The three most widely used NIR photon-counting technologies are photomultiplier tubes (PMTs) with indium gallium arsenide (InGaAs) photocathodes, superconducting bolometers, and InGaAs/indium phosphide (InGaAs/InP) avalanche photodiodes (APDs).

Performance comparisons

While it is impossible to do full justice to all the design variations, operating modes, and performance parameters encompassed by these three detector technologies, the primary detector parameters can nonetheless be summarized to assist nonexpert users in determining which detector technology or technologies may be suitable for a particular application (see table).

Near-infrared single-photon detectorsOperating temperature (K)Typical photon-detection probability @ 1.55 µmDark counts per secondRatio of detection probability to dark-count rate (x 106)Time jitter (picoseconds FWHM)Detector active-area diameter (µm)Detector Dimensions w/required cooling (cm3)InGaAs PMT2130.005200,0000.025300160030,000Superconducting NbN bolometer30.021,0002010010100,000Telecom InGaAs/InP APD2130.2600,0000.330040500Photon-counting InGaAs/InP APD2130.4400,000115050100

With the exception of the superconducting detectorwhich requires cryogenic cooling-performance comparisons are made at a single temperature. This normalization is important because for all quantum detectors the dark-count rate is an exponential function of the temperature. For photon-counting APDs, as an example, a temperature reduction of 20 K will typically result in an order of magnitude reduction in dark-count rate.

For purposes of this comparison, we give particular attention to the ratio of single-photon detection probability (1 = 100%) to the dark-count rate. We refer to this ratio as the “performance ratio” of the detector, a major factor in determining the signal-to-noise ratio (S/N) of the photon-detection system. The performance ratio is also important because at a given operating temperature it is a more stable parameter than either the detection probability or the dark-count rate alone, especially when detector operating parameters such as bias voltage or pulse-discriminator threshold are varied.

InGaAs photomultipliers

Photomultiplier tubes with InGaAs photocathodes offer the advantages and disadvantages typical of familiar PMT technology. Advantages include large active area, photon-number sensitivity, and wide-count-rate dynamic range. The large active area can be a significant advantage when gathering light from a large-area optical source such as a scintillating crystal or cuvette, or where fine alignment of the coupling optics is difficult. The potential for high pulse-counting rates well in excess of 1 MHz and the ability to determine the photon number of multiphoton pulses with pulse-height analysis can be advantageous in wide-dynamic-range situations such as lidar systems.

The disadvantages of InGaAs PMTs include relatively large size when the cooling system and vacuum pump are included, the need for high-voltage bias, and in particular the low photon-detection probability. Note that the performance ratio is the lowest of the detectors considered, mainly due to the low quantum efficiency that is typical of photocathode technology.

Superconducting bolometers

The superconducting bolometer is typically a meandering wire of thin-film niobium nitride (NbN) cooled to liquid-helium temperatures. The detector is biased with a constant current close to the critical current for the superconducting-normal transition, so that absorption of a photon will heat the wire enough to cause a portion of the metal to leave the superconducting state, generating a measurable voltage pulse. The main advantages of these detectors are the very low dark-count rates, very small timing jitter, and very high maximum count rates in excess of 100 MHz. The performance ratio of these detectors is the highest of the three types compared here, roughly three orders of magnitude higher than that of a PMT. Active areas are typically quite small, but can be well matched to the core of single-mode optical fibers with coupling optics.

The primary disadvantages of these detectors are their large size (at least ten 19-inch rack units) and high operating costs because of the need for cryogenic temperatures below 10 K. Currently, such detectors are available only on a custom basis and have not yet been offered as standard commercial products.

InGaAs/InP avalanche photodiodes

The third detector alternative is the InGaAs/InP APD. Originally developed for use in long-distance optical-fiber communication systems, these devices have been commercially available for about ten years and are acceptable for NIR photon counting in some applications. Readily available at low cost from several manufacturers, and offered with prealigned fiberoptic pigtails in hermetically sealed packages, these APDs are much smaller than PMT or bolometer detectors and can easily achieve photon-detection probabilities more than an order of magnitude greater than a PMT. When suitably cooled, this type of APD is capable of achieving a performance ratio about an order of magnitude higher than a PMT but two orders of magnitude lower than a superconducting bolometer.

One disadvantage of InGaAs/InP APDs as photon counters is the need for more-complex electronics for operation in a photon-counting mode, since the operating bias must typically be more complex than a simple DC voltage. When true single-photon sensitivity is needed, InGaAs/InP APDs must be operated in the Geiger mode. In this mode, the APD is DC-biased to give a modest linear gain of about 10, in combination with a pulsed bias exceeding the breakdown voltage applied to the device for a time period of nanoseconds to microseconds. During this “excess bias” gating pulse, the internal gain of the APD is quasi-infinite, and a single electron-hole pair generated within the InGaAs absorbing layer is capable of generating a large avalanche current within the InP multiplication layer of the device. Avalanche currents exceeding 1 mA in response to absorption of a single photon are typical.

Unfortunately, this avalanche will continue for a very long time unless quenched by external circuitry or terminated by the end of the gating pulse. The good news is that commercial modules incorporating all of the necessary bias, gating, and quenching circuits are now commercially available. Many people also construct their own circuitry based on the extensive published literature on this topic.

A second major disadvantage of InGaAs/InP APDs is known as “after-pulsing”: an exponential increase in the dark-count rate that occurs when the time between successive detection events gets too short. This minimum hold-off time between counts can range from as little as a few hundred nanoseconds to as long as 50 µs. Minimum hold-off time depends on APD quality, operating temperature, and the specific design of the bias and quenching circuits. For applications that require count rates above a few megahertz along with high detection probabilities, Geiger-mode photon counting may not be possible due to the after-pulsing effect. In these cases the APD can be operated in high-gain linear mode with a post-amplifier, which eliminates after-pulsing.

A significant disadvantage of using telecom APDs for photon counting is the difficulty of cooling the packaged device to a sufficiently low temperature, typically -40°C to -80°C. Because of the large mass of the packaged parts and the thermal conductivity of the fiber pigtail, cooling a telecom APD to these temperatures requires nearly as much volume and electrical power as cooling a PMT.

Specially designed InGaAs/InP avalanche photodiodes

The InGaAs/InP APD designed specifically for single-photon counting (see Fig. 1) is based on APD chips that are similar in design to telecommunications APDs, but modified for the differing requirements of Geiger-mode photon counting. They also use a special package that incorporates a multistage thermoelectric cooler and a temperature sensor, as well as high-efficiency standoff coupling optics for either single- or multimode optical fiber. These devices typically achieve higher photon-detection probability, lower dark-count rates, and lower timing jitter than telecommunications APDs. The Sensors Unlimited unit of Goodrich manufactures NIR APDs for Geiger-mode single-photon counting that achieve a performance ratio three to five times larger than the best telecommunications APDs.

One significant difference between dedicated photon-counting APDs and telecom APDs is in the maximum detection probability that can be achieved. It is very difficult to obtain a detection probability greater than 0.2 with a telecommunications APD, regardless of the bias voltage or temperature used. In contrast, an APD properly designed for photon counting can achieve a useful detection probability of at least 0.4 at 213 K, with higher values achieved in selected devices or at lower operating temperatures. 

Another major practical advantage of the dedicated photon-counting APD is the ease of cooling the device to low temperaturesas low as 213 K with a room-temperature passive heat sink and less than 2 W of cooling power. This device is therefore the most compact of the practical near-IR photon counters. It dissipates the least electrical power and it achieves the highest performance ratio of any device that does not use cryogenic cooling (see Fig. 2).

Keith Forsyth is the lead engineer and manager of the Optical Modules group and Noah Clay is a research engineer at Sensors Unlimited, Goodrich, 3490 Route 1, Building 12, Princeton, NJ 08540-5914; email: [email protected]; www.oss.goodrich.com.

Sponsored Recommendations

March 31, 2025
Enhance your remote sensing capabilities with Chroma's precision-engineered optical filters, designed for applications such as environmental monitoring, geospatial mapping, and...
March 31, 2025
Designed for compatibility with a wide range of systems, Chroma's UV filters are engineered to feature high transmission, superior out-of-band blocking, steep edge transitions...
March 31, 2025
Discover strategies to balance component performance and system design, reducing development time and costs while maximizing efficiency.
March 31, 2025
Filter accessories including cubes, sliders, and rings, designed to enhance the performance and versatility of optical systems. These components ensure precise alignment and stability...

Voice your opinion!

To join the conversation, and become an exclusive member of Laser Focus World, create an account today!