Gaseous Detector Laboratory

Gas Electron Multipliers

The Gas Electron Multiplier (GEM), developed by the Gas Detector Group (GDD) at CERN, offers great potential as a high resolution tracking detector for a variety of applications. One such application is particle tracking in the high multiplicity environment of Relativistic Heavy Ion Collisions at RHIC, where one needs not only excellent spatial resolution, but also large solid angle coverage and high rate capabilities. A possible approach to meet these needs would be to incorporate the features of the GEM into a MicroTPC (a small, fast, high resolution TPC), which would provide tracking coverage starting at a relatively close distance to the collision vertex. It would also have a sufficiently fast drift time to operate at the highest rates envisioned for RHIC.

A study of the suitability of GEMs as an amplification stage for a MicroTPC is made. The measurements focused on the gas gain uniformity, stability, ion feedback and interpolating readout of the detector. Several 10cmx10cm framed GEM foils are supplied by the GDD group at CERN. They are the standard 80/60 double conical type. All results have been obtained with a gas mixture of Ar+20%CO2

Gas Gain Uniformity

The gas gain uniformity of a double GEM detector was measured using a collimated x-ray beam of about 1mm2 in size and a flux of 2kcps. The beam was moved in 1mm steps to form a raster scan over a 9cmx9cm area of the detector.  The gain map is shown in Fig.1. The overall variation in the gas gain is about ±20%.

Figure 1. A gain map of a double GEM.  The relative gas gain varies from 91 (darkest) to 146 (brightest) on an arbitrary scale.

Gas Gain Stability

The dependence of gas gain on the x-ray flux has also been studied. Using the same 1mm2 pencil beam, the x-ray flux was changed from time to time and the pulse height of the photo peak was recorded. The results, shown in Fig. 2, clearly show the GEM gas gain is influenced by photon flux. This is inconsistent with published results from the CERN group 

Figure 2. Gas gain variation of the double GEM under different x-ray flux. The relative pulse height, upper curve, represents the most probable position of the photo peak from 1000 events (rate under 1kcps) or 10000 events (rate over 1kcps). 5.4keV x-ray flux, lower curve, varies from 80cps to 10kcps over a 1mm2 area. The effective gas gain is about 600.

Positive Ion Feedback

Space charge distortion in the drift volume is a major factor limiting the performance of a TPC operating under high flux. While the space charge buildup from the primary ionization is inevitable, those positive ions coming from the amplification region should be minimized. The ion feedback fraction is used in this work to quantify the unwanted positive ions relative to the signal forming electrons. It is defined as the ratio of the currents flowing into the cathode window and the anode plane: fi = - Iw /Ia.

Single GEM

In Fig. 3, a large collection of ion feedback measurements with a single GEM is plotted as a function of the field ratio below (induction field, Ei) and above (drift field, Ed) the GEM foil. Although operating parameters in these measurements vary over a large range, the data points congregate along a narrow band. The general form of the data points can be described by:

-Iw /Ia = (Ed /Ei)0.7

Figure 3. Ion feedback as a function of the ratio of the induction field and the drift field. Ar + 20% CO2, 8mm drift gap, 2mm induction gap

Double GEM

There have been a large number of studies on current measurements with double GEMs. Our measurements show similar results: in general, the ion feedback exhibits
1. a linear function with drift field, until the drift field value approaches that of the transfer field;
2. strong dependence on the induction field;
3. moderate dependence on the gas gain;
4. weak dependence on the transfer field.
In addition, the measurements in this work indicate a weak dependence of ion feedback on the distribution of gain between the two GEM stages, for a constant total gain. A larger gain in the second GEM gives a slightly lower ion feedback fraction and visa versa.

At 1kV/cm drift field, the best ion feedback fraction is about 15%.

Triple GEM

The additional GEM stage does not reduce the ion feedback fraction if the fields at each gap are progressively increasing for efficient electron transfer, i.e.: Ed < Et1 < Et2 < Ei. However, if the second transfer field is set lower than the first transfer field, (Ed < Et1 > Et2 < Ei), the ion feedback fraction can be further reduced.

Interpolating Pad Readout

Another test chamber was constructed with several different types of interpolating pad arrays on the anode plane. The size of the pads are 2mmx10mm, a baseline choice for the TPC. The interpolation is along the 2mm direction. The linearity of these anode pad arrays was determined by measuring their uniform irradiation responses (UIRs), which are histograms of reconstructed positions from a large number of radiation events uniformly distributed over the detector. A perfect detector should exhibit a flat response.

Figure 4. Several interpolating pad patterns and their corresponding UIRs.

Several intermediate strip patterns were tested. Figure 4a shows a zigzag pattern and its UIR. Figures 4a-c show designs that use one or two “intermediate” strip(s) that are “floating” between two adjacent readout strips. The charge induced on these floating strips is capacitively coupled to their neighboring readout strips. In practice, the “floating” strips are held to the correct bias through high value resistors. A key point in designing these patterns is that the inter-strip capacitance should be much higher than the strip capacitance to ground. The zigzag pattern is ideal for this purpose. The zigzag periods of all the patterns are 0.5mm, fine enough to give good interpolation. The absolute systematic errors are less than ±80µm.

It has been demonstrated that simple geometrical and capacitive charge division schemes such as zigzag strips and intermediate strips can be used with GEM to achieve moderate interpolating ratio, i.e. the ratio of the readout pitch to the position resolution (~20). In general, compared to the single zigzag pattern, the intermediate strip patterns have lower capacitive load to the preamplifiers; there is more room for plated through hole connections. However, they do require additional resistive connections between the “floating” strips and their neighboring readout strips. A small percentage of charge induced on the floating strips are lost to the ground, potentially broadening the energy resolution of the detector. Resistive charge division should perform well for one dimensional/projective readout. Two dimensional pad readout with resistive charge division may be difficult to realize because of the precision resistive connection required between rows of pads.


B. Yu, V. Radeka, G. C. Smith, C. L. Woody, and N. N. Smirnoff, "Study of GEM Characteristics for Application in a MicroTPC," pres. 2002 Nuclear Science Symposium, Nov. 2002, Norfolk, Va; to be published in IEEE Transactions on Nuclear Science, NS-50 (2003)

Last Modified: Wednesday, 06-Feb-2013 22:33:56 EST