SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Defense + Commercial Sensing 2018 | Register Today

SPIE Photonics Europe 2018 | Register Today!

2018 SPIE Optics + Photonics | Call for Papers




Print PageEmail Page

Electronic Imaging & Signal Processing

Ultra-performance in the Ultraviolet

For high-accuracy UV detection, PtSi/n-Si Schottky photodiodes offer many advantages.

From oemagazine September 2003
31 September 2003, SPIE Newsroom. DOI: 10.1117/2.5200309.0005

When selecting a UV photodiode, it is important to match the needs of the application to the performance of the detector. At the heart of the matter is material selection, structure, and how they relate to performance. The different types of UV photodiodes vary in their performance based on spectral responsivity, stability of the responsivity under exposure, spatial non-uniformity of the responsivity, solar blindness (insensitivity through IR and most of the visible spectral region), and size.

A photodiode is a semiconductor structure in which a photon passing into the depletion region creates a charge carrier pair that is eventually separated to be registered as a small electrical current (see oemagazine, August 2001, page 34). The most commonly used standard silicon p-n junction photodiode consists of a p-type layer overlaying an n-type silicon layer with a depletion region sandwiched in between. This sort of device offers the benefits of high spectral responsivity, good spatial uniformity, and virtually no limits to size, but has the disadvantage that the responsivity may rapidly degrade when the device is exposed to UV radiation. This effect is related to the oxide layer used for passivation of the silicon surface. UV exposure of the photodiode damages the interface between the oxide layer and the silicon p-type layer, which results in a change of spectral responsivity.

This radiation damage can be avoided if a semiconductor other than silicon is used. Silicon carbide photodiodes, for example, are very stable under UV exposure and also solar blind. However, these diodes cannot be easily constructed in sizes greater than 1 mm2, so their use is limited to applications like long-term monitoring of UV lamps, in which detector area is not an issue.

Schottky-type photodiodes offer another way to avoid the radiation damage described above. Unlike a p-n device, a Schottky photodiode features a depletion region sandwiched between a metallic top layer and a semiconductor layer that may be different from silicon. The most stable types of Schottky photodiodes are gold on gallium arsenide phosphide (GaAsP) and platinum silicide on n-doped silicon (PtSi/n-Si) photodiodes. These devices can be constructed with sensitive areas of 1 cm2 or larger and, hence, are the logical choice if the goal is absolute measurement in SI units by an accurately calibrated detector.

To make the final choice of which kind of Schottky photodiode to use for a demanding application, we need to gain a better understanding of issues that can affect UV photodiode performance.

instability of spectral responsivity

Although Schottky photodiodes are not prone to the radiation damage described for silicon p-n junction photodiodes, their spectral responsivity is not exactly stable. This instability under UV exposure should not be confused with the degradation in spectral responsivity common to oxide-passivated silicon junction photodiodes. Schottky photodiodes will lose some spectral responsivity under initial exposure, but after that initial drop, the responsivity levels off. Pre-aging pristine samples of Schottky photodiodes by UV exposure cures this initial instability. Continued UV exposure will lead to further changes in spectral responsivity, though at a much slower rate with a magnitude dependent on the type of photodiode; Schottky-type photodiodes are generally more stable than silicon p-n-junction diodes, for example.

Figure 1. Most photodiodes, including GaAsP Schottky photodiodes, suffer variations in spectral responsivity after prolonged exposure at a particular wavelength λexp (red); PtSi/n-Si Schottky photodiodes are in general better behaved (blue). Though slightly dependent on the irradiance level, the change of responsivity primarily is a function of the radiant exposure (inset). Typically, responsivity changes exponentially, but then tends to recover at higher radiant exposure (red). Again, PtSi/n-Si devices show better performance (blue).

Apart from this principal dependence on the photodiode type, the change of spectral responsivity under UV exposure is a complex phenomenon dependent on radiant exposure (absorbed energy per unit area) and wavelength.1 Though slightly dependent on the irradiance level, the change of responsivity is most strongly dependent on radiant exposure—typically, the responsivity decreases exponentially as a function of exposure, but then tends to recover at higher levels (see figure 1). Exposure wavelength also affects the initial rate of change in responsivity, causing responsivity to drop faster as wavelength decreases. As a result of these effects, the change in spectral responsivity leads to an increase in spatial non-uniformity of the photodiode.

The change in spectral responsivity is not limited to the exposure wavelength of the photodiode. In the case of most types of UV photodiodes, including GaAsP Schottky devices, the change in spectral responsivity after prolonged exposure at a particular wavelength λexp varies significantly as wavelength changes. For wavelengths shorter than λexp the spectral responsivity increases, whereas it decreases at longer wavelengths.

Presently, PtSi/n-Si Schottky photodiodes are the only exception to this rule: If their responsivity changes, the change is of the same sign and about the same magnitude for all wavelengths. This difference makes stability testing of PtSi/n-Si photodiodes an easy task, because the test may be restricted to a single wavelength.

It should be noted that while operated under "easy" conditions, in which irradiance is less than 1 mW/cm2 and radiant exposure is less than 1 J/cm2, devices touted as UV-stable meet performance expectations for wavelengths exceeding 250 nm; however, at wavelengths below 250 nm, most photodiodes degrade even under easy conditions. This is certainly true of GaAsP photodiodes, which used to be the most stable large-area photodiodes available even into the vacuum-UV spectrum until PtSi/n-Si photodiodes were developed.

In contrast, PtSi/n-Si photodiodes show virtually no degradation under easy operating conditions down to wavelengths of about 100 nm. These devices have proven to be the most stable type of photodiode even when heavily exposed to vacuum-UV lasers.

spatial non-uniformity

Owing to the short penetration depth of UV radiation in matter, all types of photodiodes exhibit wavelength-dependent, enhanced spatial non-uniformity at UV wavelengths. A convenient way to characterize the spatial non-uniformity of a photodiode is the RMS scatter of the responsivity σrms that is obtained by scanning a square beam 1-mm on a side over an 8 mm x 8 mm area in 1-mm steps.

High-quality silicon p-n-junction photodiodes normally show σrms between 0.2% and 0.4% at 254 nm. For GaAsP photodiodes, the typical value is 5% but experience shows that about one in 10 GaAsP photodiodes is reasonably uniform, with σrms below 1%. When PtSi/n-Si photodiodes became available several years ago, typically 50% were at least as uniform as the best GaAsP photodiodes (1%). Since that time, progress in the fabrication technology has increased this percentage to about 80%, of which 30% of the PtSi/n-Si photodiodes are of almost the same uniformity as silicon p-n-junction photodiodes.

solar blindness

Figure 2. Spectral responsivity of a solar-blind PtSi/n-Si photodiode (blue) as compared to that of a common PtSi/n-Si photodiode (red).

Due to their wide bandgap, GaAsP photodiodes are solar blind—whereas conventional PtSi/n-Si photodiodes are sensitive in this region. This can be a disadvantage for UV applications. PtSi/n-Si photodiodes can be hardened against IR and visible wavelengths, however, by exploiting the short penetration depth of UV photons in silicon. Because of this characteristic, a thin buried-oxide layer placed at shallow depths in the PtSi/n-Si structure can reduce visible and IR responsivity without compromising UV performance. Katalin Solt, then with ETH (Zurich, Switzerland) demonstrated proof of principle by forming PtSi/n-Si Schottky contacts on the thin upper silicon layer of commercially available silicon-on-insulator substrates. These photodiodes have demonstrated an order of magnitude reduction in responsivity in the visible (see figure 2 on page 22).

The primary advantage of the modified PtSi/n-Si diode structure over compound-semiconductor solar blind photodiodes is that the size of the sensitive area is not limited by any technological factor thanks to the prevalence of silicon fabrication technology. This flexibility makes them appealing for a number of applications. oe


1. L. Werner, Metrologia 35, 407 (1998).

Stabilizing your UV detector

Thermal isolation holds detector irradiance variation to 1.5% over an hour of exposure to 27 W/cm2 of UV exposure; after correction for lamp burn-in, actual irradiance change was 0.4%.

Mentioning UV radiation to a detector, optical component, or filter manufacturer will conjure up issues like solarization, high temperatures, and drift in absolute responsivity. Manufacturers spend a lot of engineering resources researching UV-resistant materials and the effects of UV on critical parts as well as devising ways to combat these ill effects.

Since most materials in nature tend to absorb rather than transmit UV radiation (180 to 400 nm), let's consider four methods commonly used to control the harmful effects and changes of UV and heat: phosphor conversion temperature stabilization, thermoelectric cooling, temperature coefficient factor correction, and thermal isolation.

Although phosphors are sometimes used as conversion materials, the approach suffers from instabilities as a function of temperature or humidity; in addition, phosphor quality is inconsistent.

Limiting the amount of UV radiation reaching the detector inhibits cumulative degradation and drift resulting from high signal levels and long-term exposure. Signal attenuation is accomplished by partially transmissive front-end components such as mechanical screens, neutral-density filters, and radiation integrators. Users must balance signal loss against system detection limits to ensure conformance to the dynamic range demands of the application.

Typically, signal attenuation in itself is not enough to protect the detector and ensure accurate measurement results unless the detector is used sparingly in low-level, short-duration UV exposures. Detector positioning relative to a source is critical for accurate radiometry, however, so detectors normally receive higher amounts of UV exposure than desirable.

Direct placement in front of lamp housings or in reactors or chambers will eventually heat up the photoactive element. Since all detectors exhibit a change in absolute sensitivity due to temperature, normally wavelength dependent, some form of correction or control may be necessary. Temperature stabilization involves maintaining a component at a set temperature, thereby locking it into a known specific temperature coefficient or keeping it within a manufacturer's stated operating temperature range.

Detectors used for long-term continuous monitoring of UV from sources like the sun may be thermoelectrically cooled. Since the detector is constantly absorbing heat from sun-up to sundown, the device is typically refrigerated to keep it well below its maximum operating temperature over the course of the day. Both stabilization and cooling require additional electronics and power supplies, making these methods more costly and complex.

A detector's change in sensitivity as a function of temperature may be supplied by the manufacturer as a percent change per degree centigrade or measured by the end user. Since this temperature coefficient may be wavelength dependent, measuring this change in the actual application will improve accuracy. Once the coefficient is known, it can be combined with detector temperature readings to apply a correction factor to the detector reading.

For cost and ease-of-use reasons, thermal isolation is very effective. In one design, a radiation integrator acts as a signal attenuator, optical diffuser, and "first strike" front component in the thermal isolation set-up (see figure). In such a design, the attenuated UV signal is transmitted through the integrator into a fiber light guide at a length determined by the application. The detection capsule lies at the far end of the light guide so that both the detector and any optical filtering housed in the capsule are kept out of the "hot zone."

—Bob Angelo, Gigahertz-Optik

UV and life

Like many young men, Hans Rabus wanted to know the meaning of life. "When I graduated from high school, I got a gift for the grades I got - a popular science book on biochemistry. I read it, and it made me interested in the origins of life," Rabus says. He continued his academic career at Berlin's Free University, pursuing a biochemistry career. One year later, he found his true calling: physics. "I figured out that I was more interested in the basics - the physical properties and basic laws of nature rather than living organisms - and that is physics," he says.

Rabus continued his studies at the Free University of Berlin through his post-graduate work, completing a PhD in physics in 1992 and receiving the Ramsauer Prize from the Daimler-Benz Foundation for his work in surface science.

In 1992 Rabus joined the Physikalisch-Technische Bundesanstalt (PTB; Berlin, Germany) and started to work in UV radiometry. He developed a so-called cryogenic radiometer operating at the Berlin synchrotron facility BESSY I, which boosted the accuracy of detector-based UV radiometry by an order of magnitude.

In 1995, Rabus steered briefly away from the UV spectrum, focusing his talents on x-ray radiometry. After designing an x-ray beam line for PTB at the new synchrotron facility BESSY II, Rabus switched back to the UV radiometry section in 1997, where he became responsible for several PTB collaborations with major national metrology laboratories across the world. In 2000, Rabus assumed his current position of section lead for the detector-based optical radiometry section.

Like many of us, Rabus's search for the meaning of life continues. That search has taught him the meaning of struggle and success. Recently, Rabus became a cancer survivor, but he has not let that slow down his life or his work. Rabus completed this article from a Berlin hospital, where he is currently recuperating from his latest medical victory.

-Winn Hardin

Hans Rabus

Hans Rabus is head of detector-based optical radiometry at Physikalisch-Technische Bundesanstalt (PTB), Berlin, Germany.