Astronomers, in the early 1950s, were the first to use photomultipliers in the photon counting mode for measurements on stars of very low magnitude. Since then there has been a rapid increase in photon counting applications. Luminescence techniques are extensively used in clinical medical and drug testing and in the food industry inspecting for antibiotics, insecticides and bacteria. Laser scattering is used in Raman Spectroscopy for molecular analysis and in sub-micron particle sizing. These developments have been helped by improvements in photomultiplier performance and compactness and the advent of microelectronics and computers.

The power of the photon counting technique stems from the fundamental principle upon which it is based: electromagnetic radiation is quantised and so counting these quanta (photons) provides the best and most direct method of measuring the intensity of the radiation. Photomultipliers by virtue of their high gain and low noise are detectors ideally suited to providing a discrete pulse for each photon detected. Pulse height discrimination makes photon counting the least sensitive method to sources of noise other than thermionic emission. Additionally the output is presented in digital form ideally suited to processing by computer.

ET Enterprises can offer a complete photon counting system combining the photomultiplier, high voltage power supply, photomultiplier housing, low noise amplifier/discriminator and computer interface. These systems are tested prior to despatch, including setting up of the photomultiplier for optimum photon counting performance.

Apart from being quick to get into operation, these systems offer flexibility, for example the ease with which the photomultiplier can be changed for another type with a different spectral range.

For more information on the principles and benefits of photon counting please click the more information link below.

Instrumentation for photon counting

The elements of a basic photon counting system are shown in figure 1.

Figure 1

Figure 1. the elements of a photon counting system.

The photomultiplier window material and photocathode spectral response should be chosen to match the wavelength range of the radiation to be measured. The photocathode converts each detected photon into a photoelectron that is amplified by secondary emission at a series of dynodes held at progressively higher potentials. The gain (G) is a function of the number of dynodes and the overall voltage. Photon counting requires the photomultiplier output pulse to trigger the discriminator (D). A typical fast discriminator triggers at 50mV but this would require a photomultiplier gain of 108 to 109. Although such photomultipliers are available the corresponding high voltages and current levels involved invariably cause instability and light feedback problems. In practice a fast, 50Ω input impedance electronic amplifier (A) of gain 100 shares part of the amplification process. A photomultiplier with 10 or 11 dynodes operating at 106 – 107 gain is therefore more practical and thus commonplace.

To avoid electrical interference, the amplifier and discriminator are best integrated into one electronic module and should be connected to the photomultiplier output by a short length of co-axial cable. In a light- detector package we improve on this by connecting the electronics directly to the anode of the photomultiplier. The photomultiplier and electronics are installed in a screened housing (usually incorporating RF and magnetic shielding).

The operating voltage will typically be in the range 800 to 1500V, depending on the choice of photomultiplier. How to determine the operating voltage will be covered in the next section. Suitable power supplies are available in a range of forms: laboratory benchtop units; rack mounting units; DC to DC converter modules; Cockcroft-Walton power bases and active voltage dividers. The power supply can be integrated with the housing or detector package.

Amplifier/ Discriminators are available with a choice of signal logic outputs - TTL, ECL. The logic output of the discriminator is connected to a counter. This can be a benchtop frequency counter, a board installed in a computer with accompanying analysis software, or a counter module with its own processor and computer interface. An important advantage of a counter with application software is that the user is guided through the setting up procedure.

Optimisation of a photon counting system

There are two parameters that must be optimised for photon counting: the discriminator setting and the photomultiplier operating voltage.

Amplifier/Discriminators come in both fixed threshold and variable threshold versions. The former is optimised by the manufacturer and for most applications the flexibility of the latter is unnecessary. In practice a threshold of 2mV offers optimal performance. A lower threshold would render the counter susceptible to interference and a higher threshold would demand higher performance of the amplifier section, previously mentioned.
With regard to setting the operating voltage: in an integrated package this is usually set by the manufacturer. The following procedure should be used for a system comprising discrete units.

Photomultiplier characteristics

To give a theoretical basis to the procedure we must consider the size distribution of photomultiplier anode pulses, when viewing a source of single photons. This can be appreciated by looking at the anode pulses on an oscilloscope.

The gain process in a photomultiplier derives from secondary electron emission. It is a quantised process and ideally should obey Poisson statistics. Although all events start with a single photoelectron, the statistical nature of photomultiplier gain means that the anode pulses will vary in magnitude from one pulse to another. That is, the gain is noisy and the degree of fluctuation is related to the mean gain of the first dynode (the higher the better). Inspection of the pulse height distribution reveals an excess of small pulses which depart from the Poisson distribution. These are the result of other physical processes, such as photoelectrons landing on the edge of d1. The measured pulse height distribution is called the Single Electron Response (SER) of the photomultiplier shown in figure 2 and consists of a peak at p and a valley at v. One way to measure the SER is with a multi-channel analyser. Since most of the signal lies to the right of v then v is a good place to set the threshold of the Amplifier/Discriminator as this is the most stable point on the entire distribution. Knowledge and understanding of the shape of the SER will help appreciate the significance of the plateau characteristic that will be discussed below.

Figure 2

Figure 2. the pulse height distribution of a photomultiplier in response to a low light level is called SIngle Electron Response (SER).

For a photomultiplier with anode pulse width, t, the mean peak pulse voltage at the input of a 50Ω impedance Amplifier/Discriminator is: ~50Ge/t
where e is the electronic charge = 1.6 x 10-19C.
t is typically 5 ns and for G = 107 we have
vo ~ 50 x 107 x 1.6 x 10-19 / 5 x 10-9 = 16mV

As the pulse height is proportional to the gain the selection of G will determine the proportion of the anode pulse height distribution that will exceed the threshold voltage and hence the measured count rate. The gain is related to the HV by a relationship of the form G= aVm , the value of m is about 8 for an 11 stage photomultiplier and consequently G varies rapidly with V.

Starting at a low HV (A in figure 3), the SER has few pulses above the discriminator threshold. By increasing the HV to (B) about 5% of the pulses are now above the threshold. At (C) approximately 50% of pulses are above the threshold, and at (D) most are counted. Further increase in HV has little effect on the number of pulses counted. Note that because the light level is kept constant the areas under the curves of figure 3 are the same.

Figure 3 SER pulse height distribution

Figure 3. SER pulse height distributions showing the effect of increasing the photomultiplier operating voltage resulting in progressively larger output charge distributions from left to right. The areas under the curves are all equal.

A plateau curve is measured by noting the count rate for a series of HV settings, incremented by 25 volts. The resulting curve is shown in figure 4. This curve is known as a plateau characteristic because of the central flat region. The curve is often drawn using a logarithmic vertical scale for the count rate.

Figure 4 the plateau characteristic

Figure 4. the plateau characteristic is the count rate above a fixed threshold hO as a function of HV. A good photon counting photomultiplier is one which attains a long, flat plateau as illustrated.

Setting up procedure

When adjusting the photomultiplier voltage take care not to exceed the voltage for the maximum rated gain, stated on the manufacturer’s test ticket. The photomultiplier may become unstable above this voltage leading to a permanent loss of performance. Note that the voltage given will be less than the maximum rated voltage for the photomultiplier in the manufacturer’s literature.

Start with the light source off and with the HV at 500V. Increase the HV in steps of 25V until there is sufficient gain at this HV setting (V 1) to produce greater than ~ 10 counts / second. This corresponds to region AB of figure 4 for dark counts. Switch on the single photon light source and increase the intensity until the counts are about 100 times those of the dark count. With this light setting you can now measure your first plateau characteristic by starting some 100V lower than V 1. It is instructive to plot a second plateau characteristic at 10 times the light level. The curves should have the same shape and this is a good check that the system is functioning correctly. Repeat measuring dark counts only.

Figure 5

Figure 5. the counting characteristics obtained from the setting up procedure described on this page.

The point of turnover on the plateau characteristic is called the knee voltage and a sensible operating point is to the right of this but its actual location is a matter of choice. For repeatability and consistency we locate the operating point where the slope of the curve first reaches a critical value such as.

An advantage that comes from buying a photon counting photomultiplier that has been selected by ET Enterprises Ltd. for photon counting applications is that the test data supplied with the photomultiplier includes the knee voltage (measured with calibrated test equipment). The user can compare his own results to the test data for confirmation of correct operation. Additional useful information includes the plateau slope, the dark count at the knee voltage and the dark count 100V above the knee voltage. Also, we have a wide range of housings and accessories to allow you to assemble a photon counting system with your choices guided by our technical staff.

Further reading; download our technical papers, in PDF format, relating to photon counting:

RP/066 - A comparison of current measurement with photon counting in the use of photomultipliers.
RP/081 - The determination of photomultiplier temperature coefficients for gain and spectral sensitivity using the photon counting technique.
RP/090 - Optimal use of photomultipliers for chemiluminescence and bioluminescence applications.
RP/096 - Practical photon counting.

© ET Enterprises 2011
ET Enterprises is a trading name of ET Enterprises Limited. Registered in England & Wales, company number 06081468.