Saturation Effects in Linear Gasdetectors

Detlef Smilgies, CHESS

Detection Principle

A linear gasdetector is a position-sensitive proportional counter. A strong electric field between counting wire and anode plate (2000V over 5-10 mm) leads to a charge avalanche when an x-ray photon ionizes an atom of the counter gas (typically Ar or Xe at 1 to 5 bar). Because of a quenching gas (often 10% methan; the mix with Ar is called P10) the discharge dies after a brief time and after creating ample charge ("charge amplification"). The charge pulse is detected at either end of the wire. Either the charge ratio (resistive current division - MBraun detector) or the delay time between the arrival times of both pulses is used to determine the position of the primary ionization event along the wire.

Saturation Effects

- integral effect

The saturation of the integral counts along the wire is often a property of the detection electronics. While the time resolution of charge dividing detection is given by the proper integration times of a few microsec, typical delay lines have time constants around several 100 ns. The randomness of the Poisson distributed counts will limit linear detection rates to 1/(100*time constant). G2's Time-to-Amplitude-Converter - Pulse-Height-Analyzer combination (TAC-PHA) will saturate at about 10 kcps, mostly given by the limited speed of the ORTEC PCI board. A good Time-to-Digital Converter (TDC) can go up to 100 kcps. The dead time correction follows the usual formula, and spectra can be corrected.

- local effect

If the detector sweeps over a strong peak, so much space charge can build up locally that the peak saturates or even turns into an inverted peak. On the scope this behavior is associated in a reduction of the peak amplitude from the sum output (the energy-proportional signal). The means, that the gas amplification becomes reduced which causes problems with threshold values in the electronics. Moreover the overall signal rate arounf the peak seems to be affected. I am not aware of any correction procedure other than remeasuring a range of spectra with a suitable attenuator. The overall signal along the wire may also be affected, so spectra with local saturation can not be trusted much. By carefully tuning the gas amplification the detector can be optimized using low-noise preamps.

- wire degradation

If the detector is hit hard at the same spot over and over again, the detection sensitivity at this spot may be strongly reduced. This is due to carbon deposits from cracking methane in the discharge. This problen often occurs at the direct beam position used for line-up. Hence care must be used to line-up in the direct beam with a sufficint attenuation yielding integral count rates below 1kcps. Only remedy is to renew the counting wire.

Conclusion

Gas detectors are elegant, relatively simple, and cost efficient x-ray detectors, as long as they are operated in a good working range specified by a maximum integral and a maximum peak countrate. Unfortunately, these numbers are often not specified by the manufacturer, and it usually means some painful, hard-learning exercise for the experimenter. Moreover, the experimenter usually deals with a "black box", and it is not very obvious how to adjust detector parameters such as the gas amplification, in order to set up a convenient working range of the detector.

For G2 we recently ordered a linear diode array from the BNL detector group which can go up to 100 kcps per diode element. This monstrous performance is achieved in a custom high-integrated circuit providing a preamp and discriminator for each of the 1000 elements. The new detector, expected to materialize sometime in 2008, will be much better adapted to the flux and dynamic range of signals at G2.