Detection of a single photon is a requirement for many research and medical applications, such as fluorescence imaging, single molecule imaging, and luminescence applications. The sensitivity required for single photon detection has traditionally been provided by a photon amplification tube, such as an image intensifier, a microchannel plate photomultiplier tube (MCP PMTs), or a hybrid photodiode. Improvements to both the quantum efficiency (QE) of the photocathode and dark count rate (DCR), as well as some other factors, can improve both sensitivity and performance of any photon amplification tube, in turn improving the sensitivity of medical diagnostic imaging equipment.

Improvements to the sensitivity and performance of photon amplification tubes has largely been focused on the MCP including reducing pore sizes (as small as 2 μm), using a chevron or Z stack configuration, or by using larger surface areas in an effort to improve resolution or acquire more photons.

Yet the properties of the photocathode (QE, DCR, response rate) define the quality of the detector. The photon detection probability is mainly defined by the QE, while the noise contribution is usually dominated by thermionic emission from the photocathode. Indeed, these thermionic electrons are amplified in the same manner as the photoelectrons, which often makes it impossible to separate dark counts from single photon events. Although the photocathode is a critical component to successful extreme low light imaging, significant technological advancements in this area have largely been missing to date.

Traditionally, S20 photocathodes detect spectral ranges spanning from ultraviolet (UV) through green. The QE of these photocathodes is typically around 20 percent and the dark count is generally low as well. When trying to detect low intensity signals, photon counters are used to amplify the output. Because photon detection probability is mainly defined by the QE of the photocathodes, it is critical that this property is maximized as much as possible.

Many researchers in medical diagnostic applications are primarily concerned with very narrow spectral response ranges, focusing on UV, blue, green, red, or IR in order to ensure that they are using the highest QE available for their application. Once the photocathode is chosen, other factors such as resolution, timing, and dark count rates are balanced to select a photon amplification tube to most closely match the requirements of the spectra being examined. In this manner, they can be sure their results are as accurate as possible.

Higher QE Values

Fig. 1 – Quantum efficiency spectra for newly developed Hi-QE photocathodes: Hi-QE UV (dark blue), Hi-QE blue (light blue), and Hi-QE green (green) in comparison with a conventional S-20 photocathode (CNV S- 20, dashed red).

High quantum efficiency (Hi-QE) photocathodes increase the QE of the detector by as much as 50 percent. These Hi-QE photocathodes, developed by Photonis, were grown on a fused silica input window, which has a high-energy cutoff at 170 nm. In addition to the higher QE values, these photocathodes have extremely low dark rates (typically 30 cts/cm2) and a response time below 50 ps. This represents an improvement of more than 10X.

The Hi-QE series of photocathodes are offered in three distinct types including UV, blue, or green options with each specific range showing overall improvements when compared to the broader spectral response range of the S20 photocathode (See Figure 1)

All testing presented here was done using an MCP-PMT with a dual-MCP chevron configuration. Improvements to the photocathode were based on a standard S20 to increase spectral response in specific ranges. The new Hi-QE photocathodes were then assembled into the detector for quantification. Each Hi-QE photocathode was optimized to peak across a narrower spectral range than the traditional S20 photocathode.

Hi-QE UV photocathodes are optimized for the UV range with a maximum QE at 270 nm of typically 31–34 percent. These photocathodes can also be grown on specially prepared sapphire cathode substrates allowing extension of the sensitivity spectral range down to 150 nm (not shown here).

Fig. 2 – After 10 minutes in the dark, the dark count rate drops below 50 Hz/cm2. Tests indicate that all Hi-QE S20 series photocathodes (UV, blue, green) behave similarly and exhibit nearly the same values of dark count rate.

Hi-QE blue photocathodes were designed to provide the highest QE in the 260–410 nm spectral range. The QE spectrum shows a broad plateau in this range, again with a typical QE above 30 percent. The decrease of QE below 260 nm (compared with Hi-QE UV) is the trade-off for high sensitivity in blue spectral range.

Hi-QE green photocathodes demonstrate a very high QE value above 30 percent in the range of 390–480 nm. At 500 nm the QE is still about 25 percent. Comparing with other Hi-QE photocathodes, the sensitivity of Hi- QE green is much higher at a longer wavelength up to 700 nm. It is important to note that despite high sensitivity at high wavelength, the dark rate of these photocathodes stays extremely low (the same as other Hi-QE cathodes), which makes Hi-QE green unique for photon counting in this spectral range.

Low Dark Rate Counts

When monitoring the DCR of low-rate single photons, it is critical that the dark rate is kept as close to zero as possible. High DCRs often interfere with the accurate identification of a single photon event, which can result in false-positive tests in medical imaging applications. In order to test the performance of the Hi-QE photocathode, it was compared to the evolution of the standard S20 cathode. Figure 2 shows the dark rate versus time at room temperature (23 °C) for a standard S20 and the Hi-QE blue S20 photocathode. It was observed that an extremely low dark rate of Figure 2 illustrates that it takes about 2–3 hours to reach the low-rate plateau. The high dark rate measured at the beginning is believed to originate from population by ambient light of long-living surface and bulk states, lying above Fermi level. It takes time to discharge these states, with a decay time also an important criterion of the detector performance. For Hi-QE photocathodes, the growing process was adjusted to keep a low dark rate and to ensure quick discharge of surface and bulk states.

Fig. 3 – Pulse height distribution recorded with Photonis dual MCP-PMT and Hi-QE photocathode with single photon illumination (blue). The Gaussian curve (black) is a fit to experimental results.

The pulse height distribution (PHD) demonstrates the sensitivity and noise of a detector. The PHD was recorded using a charge sensitive preamplifier CSP10 (1.4 V/pC), a shaping amplifier CSA4 (gain = 10, shaping time of 250 ns), and a multichannel analyzer MCA3 (scale = 0.89 mV/chn). The threshold was set at 32 chn. For the dark-rate measurements, as presented above, the tubes were placed into dark conditions, the MCP voltages were set to obtain a gain of 1-2E05, and the count rates versus time were measured. Figure 3 presents a PHD obtained using a dual MCP-PMT with a Hi-QE photocathode, illustrating the ability for single photoelectron detection.

The PHD shown in Figure 3 was measured with low input background light illumination, keeping the count rate below a few hundred hertz. The MCP voltage of 1625 V (for dual MCP) yields electron gain of 1.1 × 105. The DCR originating from the photocathode was approximately 30 Hz/cm2, while the MCP contribution was negligible at In the resulting PHD (blue curve), the peak is well separated, with a low-energy valley and noise below the threshold. The peak is described well by a Gaussian distribution (solid black curve). The gain point (G) corresponds to the mean energy of the PHD and is just slightly above the position of the PHD peak. Both the peak-to-valley (P/V) ratio and the full width half maximum-to-gain (W/G) ratio are typically used to characterize photon counting detectors. Here the measured values are P/V ≈ 6 and W/G ≈ 0.86, which are the best available for dual MCP-PMTs.


As previously discussed, photon amplification tubes (image intensifiers, MCPPMTs, and hybrid photodiodes) are used across a wide variety of scientific applications ranging from physics research to medical imaging. In many cases, these tubes are combined with emerging and evolving technologies such as electron multiplying charge coupled device (EMCCD) or scientific CMOS (sCMOS). While traditional photon amplification tubes have improved over time, much of the improvement focused on the MCP or the external diagnostic processing.

By developing the new Hi-QE photocathode, Photonis has increased the QE by 50 percent while reducing the dark count by a factor of more than 10. It was determined that extremely low DCRs, down to 30 cts/cm2 and a quick response time, well below 100 ps, make these Hi-QE photocathodes ideal for photon counting devices. When these factors are combined, sensitivity is greatly increased, which could mean the difference between a false-positive or a definitive diagnosis.

This article was written by Dmitry Orlov, applications engineer at Photonis Netherlands B.V. For more information, Click Here.