Microchannel Plates
Microchannel plates (MCPs) are compact electron multipliers of high gain.
They have been used in a wider range of particle and photon detection
systems perhaps more than any other kind of detector.
A typical MCP consists of about 10,000,000 closely packed channels of
common diameter which are formed by drawing, etching, or firing in hydrogen,
a lead glass matrix. Typically, the diameter of each channel is ~ 10
microns. Each channel acts as an independent, continuous dinode
photomultiplier. In astronomy, and in the many other fields that use MCPs,
the detectors are generally used for distortionless imaging with very high
spatial
resolution.
Physical principles of MCP operation
The fundamental physical principals of MCP detectors are gain, efficiency,
energy
resolution, spatial resolution, time resolution, and
dark noise.
X-rays interact with the channel plate glass and electrodes (and with the
associated photocathode material) via the
photoelectric effect. For X-ray
energies below about 5 keV, detection proceeds in a 'single channel mode'.
That is, no significant fraction of the X-ray beam entering a given channel
penetrates the channel wall to illuminate the neighboring channels. At
higher energies, this 'channel crossing' phenomena becomes important.
Another important property of MCPs is their relative immunity to
magnetic
fields. A single plate with typical operating parameters is completely
unaffected by being immersed in a 0.5 Tesla magnetic field. Stacks of plates
in certain orientations are immune to much higher fields. This property has
only just begun to be exploited in the space astronomy world.
Gain
For single photon or charged particle detection, MCPs are typically used in
one of several "high-gain" configurations which produce a saturated
(or peaked) output pulse height distribution. Gains of
106 - 108 are achievable. The governing physical
parameter which determines gain is the L/D ratio (length to diameter of the
individual channels). The higher the ratio, the higher the gain. Typical
values are in the range 75:1 - 175:1. Also, the most commonly used
configurations-- the chevron (or 'V' shaped arrangement) and the Z-plate
configuration -- will not only produce 107 gain factors, but reduce
the ion feedback.
The front plate of a chevron pair for use in X-ray astronomy usually has
channels perpendicular to its front surface, since the X-rays emerging from
a grazing incidence telescope do so along the surface of a cone. The
channels of the rear plate are then at an angle of about 15 degrees. This
arrangement, as far as we know, has never been proven to be optimal. It is
just what is typically used. The gap between the plates is usually about 0.
1 mm. Often an intergap voltage field is also used. These two factors stop
the electron cascade from spreading out, and thus reducing the spatial
resolution of the detector system, as it crosses between the plates.
Recently, other technologies (such as transparent metal meshes) have
started to be used rather than the intergap field to produce the same
result.
Theoretical and experimental evidence agree that the pulse height FWHM
should decrease with decreasing channel diameter. Irrespective of geometry,
however, minimal FWHM is achieved when (1) individual stages of the
multiplier are independently operated in 'hard' saturation (the plate bias
voltages independently exceed a level Vo where
Vo = (8.94(L/D) + 450)V and (2) the interplate potential difference
is well chosen. The best FWHM from the curved single plate, the V
configuration, and the Z configuration is around 30.
Quantum Detection Efficiency
The quantum detection efficiency in X-rays for a single "bare" MCP is a low
1-10%. It is strongly correlated with photon energy (the higher the E, the less
the eff) and with the angle of incidence. The efficiency curve has strong
peaks associated with angles of incidence which correspond with the
critical angle of X-ray reflection from the glass substrate and tends to be
zero at both normal (0 degrees) and grazing incidence.
To enhance the efficiency, a material of high photoelectric yield is
deposited on the front surface and the channel walls of the MCP. This can
increase the overall efficiency to over 30%. Other, more complicated
techniques, have been found to push the efficiency to over 60% for the
optimum angle of incidence.
Energy Resolution
Until recently, this section would have stated simply "has none".
However, CsI-coated chevron MCP detectors have now been shown to possess a limited
degree of energy resolution (at least in the soft X-ray region). This
occurs, however, only if the bias voltage is well below the saturation
voltage. Ultimately, the resolution achieved will be determined by
the properties of the coating used on the channel walls.
Spatial Resolution
For any multistage MCP detector at the focus of a grazing incidence
telescope, the FWHM spatial resolution is the sum of terms related to the
geometry of the X-ray interactions and terms related to the readout
element/signal processing chain. Since the functions of the
detection/amplification and of position encoding are separable in MCP detectors, a
wide range of detector geometries has evolved, each with its
resolution dominated by a different term in the sum. However, in all
cases, the fundamental resolution is the channel diameter.
Temporal Resolution
In general, the time resolution of a satellite-borne MCP detector is
determined by the telemetry rate. MCPs are intrinsically very fast
detectors. The pulse transit time through the intense electric field is of
order 10-10 seconds. The transit time for a single plate with a
length to diameter ratio of 40:1 operating under typical voltages is about 50
picoseconds.
Dark Noise
Usually, the internal background count, or dark noise, in the current
generation of MCPs is uniformly distributed across the plate with a
value of 0.2 cts/sec/sq-cm. This is rather high compared to rates seen in
the most commonly used proportional counters. However, it is more
indicative of the sophistication of scintillator rejection techniques and
the ignorance of MCP noise than any intrinsic behavior. Also, contamination by
potassium and rubidium cause the background to be higher in MCPs. Better
manufacturing will therefore lead to reductions in the dark noise.
|
