The Detector
An external voltage divider is used to apply 100 volts to the acceleration optics (for electron detection), each MCP, the gap between the MCPs, and the backside of the last MCP and the collector (anode). The last voltage dictates the time of flight of the electrons and in this way the pulse-width. The anode is a 0.4 mm thick plate with an edge of 0.2 mm radius to avoid high field strengths. It is just large enough to cover the active area of the MCP, because the backside of the last MCP and the anode act as a capacitor with 2 mm separation and large capacitance slows down the signal. The positive charge in the MCP influences positive charge in the backside metalization. A hollow torus conducts this around the edge of the anode plate. A torus is the optimum compromise between low capacitance and short path and for similar reasons usually no dielectric (Markor) is placed into this region. After a 90° turn of the torus it is possible to attach a large coaxial waveguide. A taper permits minimizing the radius so that an SMA connector can be used. To save space and make the impedance match less critical, the taper is often reduced to a small 45° cone on the backside of the anode plate.
The typical 500 volts between the backside of the last MCP and the anode cannot be fed into the preamplifier. Therefore the inner or the outer conductor needs a DC-block, that is, a capacitor. Often it is chosen to only have 10-fold capacitance compared to the MCP-anode capacitance and is implemented as a plate capacitor. Rounded, electro-polished metal plates and the ultra high vacuum allow very high field strengths and high capacitance without a dielectric. The bias for the center conductor is applied via resistors hanging through the waveguide (see bias tee). If the DC block is used in the outer conductor, it is in parallel with the larger capacitor in the power supply. Assuming good screening, the only noise is due to current noise from the linear power regulator. Because the current is low in this application and space for large capacitors is available, and because the DC-block capacitor is fast, it is possible to have very low voltage noise, so that even weak MCP signals can be detected. Sometimes the preamplifier is on a potential (off ground) and gets its power through a low-power isolation transformer and outputs its signal optically.
The gain of a MCP is very noisy, especially for single particles. With two thick MCPs (>1 mm) and small channels (< 10 µm), saturation occurs, especially at the ends of the channels after many electron multiplications have taken place. The last stages of the following semiconductor amplifier chain also go into saturation. A pulse of varying length, but stable height and a low jitter leading edge is sent to the time to digital converter. The jitter can be further reduced by means of a constant fraction discriminator. That means that MCP and the preamplifier are used in the linear region (space charge negligible) and the pulse shape is assumed to be due to an impulse response with variable height but fixed shape from a single particle.
Because MCPs have a fixed charge that they can amplify in their life, especially the second MCP has a lifetime problem. It is important to use thin MCPs, low voltage and instead more sensitive and fast semiconductor amplifiers after the anode. (see: Secondary emission#Special amplifying tubes, .).
With high count rates or slow detectors (MCPs with phosphor screen or discrete photomultipliers) pulses overlap. In this case a high impedance (slow, but less noisy) amplifier and an ADC is used. Since the output signal from the MCP is generally small, the presence of the thermal noise limits the measurement of the time structure of MCP signal. However with the fast amplification schemes, is possible to have valuable information on the signal amplitude, even at very low signal values. But yet, not successful on the time structure information of the wideband signals.
Read more about this topic: Microchannel Plate Detector