Choosing a Photomultiplier Tube for Your Application
Since the first commercial photomultiplier tube (PMT) was developed in the early 1940s, it has remained the detector of choice for experiments requiring fast response times and high sensitivity. Today, the PMT is a staple for research in many fields including analytical chemistry, particle physics, medical imaging, industrial process control, astronomy, and atomic and molecular physics. This tutorial provides introductory material for the principle of operation and key specifications to consider when choosing a PMT for a given application.
Basic Principle of Operation
Photomultiplier Tubes (PMTs) are sensitive, high-gain devices that provide a current output that is proportional to the incident light. The PMT consists of a glass vacuum tube that houses a photoemissive material called a photocathode, 8 – 14 secondary emitting electrodes called dynodes, and a collection electrode called an anode. If a photon with sufficiently high energy (i.e. more energy than the binding energy of the photocathode material) is incident on the photocathode, it is absorbed, and an electron is released in accordance with the photoelectric effect. Since the first dynode is maintained at a higher potential than the cathode (thereby creating a potential difference between these two elements), the ejected electron will accelerate toward the dynode and crash into it, releasing secondary electrons. Typically, 3 – 5 secondary electrons are released during this process. Each of these 3-5 electrons is then in turn accelerated toward and crashes into the second dynode, thereby releasing 3 – 5 more electrons. This process continues through the entire dynode chain providing an electron gain of 3 – 5. Typically, each dynode is maintained at a potential that is 100 – 200 V higher than the previous one. At the end of the dynode chain, the electrons are collected by the anode and a current pulse is outputted. However, to read that pulse, the current usually needs to be converted to a voltage; the simplest way to do this is to connect a low load resistance across the anode and ground. The two PMTs offered by Thorlabs use a transimpedance amplifier (TIA) to convert the nanoamp or microamp current outputted by the anode to a voltage in the millivolt or volt range, respectively.
For example, if a PMT consists of 8 dynodes as shown in the figure below and each electron is able to produce 4 secondary electrons, the total current amplification after traveling through the dynode chain will be 48 ≈ 66,000. Each photoelectron for this example PMT produces a charge avalanche at the anode of Q = 48e. The corresponding voltage pulse is V = Q/C = 48e /C where C is the capacitance of the anode (including connections). If the capacitance is 5 pF, the output voltage pulse will be 2.1 mV.
When choosing a PMT for a given application, the photocathode material should be matched to the intended application. Generally, the long-wavelength cutoff is determined by the photocathode, while the window material determines the short-wavelength cutoff. PMTs are manufactured for wavelengths from the deep UV through the infrared. However, since the photocathode is responsible for converting incident photons into electrons, the efficiency with which it does this for the wavelength of interest is of utmost importance. There are a variety of materials used for photocathodes, each with a different work function and each intended for use in a different spectral range.
Quantum Efficiency (QE) is a specification that is usually expressed as a percentage and is associated with the PMTs ability to convert incident photons into detectable electrons. For instance, a QE value of 20% means that one in every five photons that strike the photocathode will produce a photoelectron. For photon counting, it is desirable to have a PMT with a high QE value. Since QE is dependent upon wavelength, it is important to choose a PMT with the best quantum efficiency over the wavelength range of interest. It should be noted that photocathodes for the visible portion of the electromagnetic spectrum typically have QE values that are less than 30%.
The QE of a PMT can be quickly calculated from its spectral response plot (see the Graphs Tab for the Spectral Response Plots for the PMM01 and PMM02) by using the following equation:
where S is the radiant sensitivity in units of A/W and λ is the wavelength in nm.
PMTs are available primarily with two different geometries: head-on (i.e. the photocathode is located at either end of the vacuum tube) and side-on (i.e. the photocathode element is located on the side of the vacuum tube). Head-on PMTs have semitransparent photocathodes and are characterized by large collection surfaces, better spatial uniformity, and better performance in the blue and green spectral regions. For applications requiring a wide spectral response, such as spectroscopy, the head-on geometry is preferable. In contrast, side-on PMTs have opaque photocathodes and are preferable for applications in the UV and IR. This configuration tends to be less expensive than head-on and is widely used in spectrometers and for applications requiring efficient optical coupling and high QE such as scintillation counting.
The 8 – 14 secondary emitting electrodes (i.e. dynodes) are often arranged in one of two configurations: linear or circular. Linear dynode arrays (such as the one shown in the figure above) are popular due to their fast time response, good time resolution, and excellent pulse linearity. The circular cage-type array is found on all side-on PMTs and some head-on PMTs. This configuration is compact and offers fast response times.
PMTs are unique because they are capable of amplifying very weak signals produced by photocathodes to detectable levels above the readout circuitry noise without introducing substantial noise. In a PMT, the dynodes are responsible for producing this amplification, which is referred to as gain. Gain is highly dependent on the voltage being applied. PMTs can operate well above the manufacturer’s stated high voltage recommendation, yielding gains that are 10 – 100 times above spec; this generally has no detrimental affects to the PMT if the anode current is kept well below the rated value. The two PMTs offered by Thorlabs combine the head-on photocathode geometry with a cage-like circular dynode chain.
Ideally, all of the signal produced by a photocathode would be due to current generated by light incident on the tube. However, in reality, PMTs will produce currents regardless of whether light is present. The signal that results in the absence of light is known as dark current, and it effectively degrades the signal to noise ratio of the PMT. Dark current is due mainly to the thermionic emission of electrons from the photocathode and first few dynodes but with far smaller contributions from cosmic rays and radioactive decay. In general, tubes designed for use in the red part of the spectrum will exhibit more dark current than others due to the lower binding energy of red-sensitive photocathodes. If it is assumed that the primary source of dark current is thermionic emission from the photocathode, the dark count rate is given by
Since thermionic emission depends highly on the photocathode’s temperature and work function, cooling a PMT will greatly reduce dark current counts. By purchasing a PMT equipped with a thermoelectric cooler and using it to cool the PMT from 20oC to 0oC, the dark current will be reduced by a factor of ~10. When using a thermoelectric cooler, care should be taken to avoid condensation at the window since this moisture will reduce the amount of light incident on the photocathode. In addition, excessive cooling should be avoided as it can actually have adverse effects, which include signal reduction or voltage drops across the cathode since the resistance of the cathode film is inversely proportional to the temperature.
For experiments demanding high time resolution, short rise times are a must. Anode pulse rise time is the most commonly specified time response characteristic for a PMT and is defined as the time required for the output of the PMT to rise from 10% to 90% of its peak amplitude when the photocathode is fully illuminated. Typical anode rise times range from 0.5 to 20 ns. Ultimately, the pulse rise time is determined by the spread in transit times for the different electrons. These times vary for several reasons. First, the initial velocities of secondary electrons will vary because they are released from different depths within the dynode material. Some electrons will have no initial energy when leaving the dynode whereas other will have a nonzero initial energy; hence, the latter arrive at the next dynode in a shorter time period. In addition to the variation in initial ejected electron speed, transit time spread is also caused by electron path length variations. Due to these effects, the rise time of an anode pulse will decrease with increasing voltage as V-1/2.
There are several other important considerations. First, choose the electronics that will be used with the PMT carefully. Small changes in the high voltage applied across the cathode and anode can dramatically change the output. Second, the lab environment can also affect the performance of the PMT. Changes in temperature and humidity as well as the presence of vibrations all negatively affect tube operation. Finally, the tube’s housing is of importance; not only does it shield the tube from external and extraneous light, but it can also reduce the effects of external magnetic fields. Magnetic fields of a few gauss can greatly reduce the gain, but these adverse affects can be minimized by creating a magnetic shield from a high permeability material.