IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 50, NO. 5, OCTOBER 2003
1297
Measurement of GEM Parameters With X-Rays G. Bencivenni, W. Bonivento, A. Cardini, C. Deplano, P. De Simone, F. Murtas, D. Pinci, M. Poli-Lener, and D. Raspino
Abstract—The gas electron multiplier (GEM)-based detectors have been widely developed in past years and have been proposed for many different applications. In this paper, we report on a method able to provide information on the characteristic parameters of a GEM. A single-GEM detector is illuminated with a high-intensity flux of low energy (5.9 keV) photons and all the electrode currents are measured simultaneously. From the analysis of these measurements we extracted a phenomenological and analytical model able to describe the currents induced on the electrodes as a function of electric fields and GEM voltages when the detector is exposed to a continuous ionizing radiation. This model provides information on the characteristic GEM parameters. In conclusion we briefly describe other methods able to extract in a more direct way GEM parameters. Index Terms—Gas electron multiplier (GEM).
I. SINGLE-GEM DETECTORS
by the GEM electrode facing the cathode. In this way, part of the primary ionization is lost. Morover after amplification some electrons (ions) could be collected by the GEM electrode facing the anode (cathode). This effect reduces the signal induced on the anode (cathode). II. CHARACTERISTIC PARAMETERS In a single-GEM detector the charge production and transfer processes can be expressed by four parameters: the electron collection efficiency, the gain, the electron and ion extraction efficiencies. is defined • The ELECTRON COLLECTION EFFICIENCY as the ratio between the number of electrons collected into and the number of electrons produced in the holes the drift gap
A
GEM [1]is made by a thin (50 m) kapton foil, 5 m copper clad on each side. This multilayer is uniformly perforated, and the holes act as electron multiplication channels. The holes have a bi-conical shape with external (internal) diameter of 70 m (50 m). The distance between the centers of the holes is 140 m. By applying an appropriate voltage to a GEM immersed in a gas mixture an electric field as high as 100 kV/cm is created into the holes, allowing electron multiplication in this region. A single-GEM detector consists of a GEM foil inserted between two planar electrodes. The four detector electrodes (cathode, the two GEM contacts and the anode) are supplied with appropriate voltages, to guarantee that electrons created in the gas above the GEM by ionizing particles are drifted toward the GEM to be amplified and that after amplification electrons are extracted from the GEM holes and collected by the anode. The electric field present between the cathode and the GEM . The electric field present (drift gap) is called drift field between the GEM and the anode (induction gap) is called . Ions produced by primary ionization and induction field by electron multiplication are collected by the cathode. (In our setup we neglect the primary ionization in the induction gap, which is about ten times smaller than the one in the drift gap). The GEM voltage and the drift and induction fields have to be chosen carefully to optimize detector performances. In fact, due to the electric field lines defocusing (i.e., some field lines do not enter into the holes) some primary electrons could be collected Manuscript received December 10, 2002; revised May 5, 2003. G. Bencivenni, P. De Simone, F. Murtas, and M. Poli-Lener are with Laboratori Nazionali di Frascati, 00044 Frascati (RM), Italy. W. Bonivento, A. Cardini, C. Deplano, D. Pinci, and D. Raspino are with INFN Sezione di Cagliari, 09042 Monserrato (CA), Italy (e-mail:
[email protected]). Digital Object Identifier 10.1109/TNS.2003.818234
(1) is defined as the ratio between the number • The GAIN by multipliof electrons produced into the holes cation in gas and the number of electrons entering into the holes (2) is defined • The ELECTRON EXTRACTION EFFICIENCY as the ratio between the number of electrons extracted –but only the ones that will reach the from the holes anode after drifting in the induction gap–and the number of electrons available into the holes after multiplication
(3) is defined as the • The ION EXTRACTION EFFICIENCY ratio between the number of ions extracted from the holes –but only the ones that will reach the cathode after drifting in the drift gap–and the number of ions available into the holes after electron multiplication (4) It should be noted here that for the single-GEM detector the is not considered because we asion collection efficiency sume that all ionization processes occur in the drift gap (and therefore there are no ions in the induction gap). The ion collection efficiency would have to be included when considering a GEM-based detector with two or more GEMs.
0018-9499/03$17.00 © 2003 IEEE
1298
Fig. 1.
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 50, NO. 5, OCTOBER 2003
Schematic diagram of the measurements setup.
In a ionization chamber the absolute value of the current inby the electron-ion pairs produced duced on both electrodes by ionization is (5) is the number of pairs produced in the gap for each where ionizing interaction of the radiation with the gas molecules, is the rate of interactions in the gas volume, and is the electron charge. The sign of is positive on the cathode (the current is sourced by the cathode) and negative on the anode. When a GEM foil is inserted between the two electrodes of a parallel-planes ionizing chamber (thus making the “single-GEM detector”) the currents induced on the electrodes due to ionization processes in the drift gap become (6) (7) (8) (9) , and represents respectively the currents on where the cathode, on the GEM electrode facing the cathode, on the GEM electrode facing the anode and on the anode. The single-GEM detector functional parameters defined in (1)–(4) cannot be expressed directly as a functions of the currents only, with the exception of the electron extraction efficiency which is (10)
III. MEASUREMENTS Measurements were taken with a single-GEM detector of 10 10 cm active area, a drift gap of 3 mm and an induction gap of 1 mm. The detector was irradiated, through the cathode, keV) emitted with a high intensity flux of soft X-rays ( by a water-cooled continuous X-ray tube, resulting in a conMHz in the detector (Fig. 1). The X-ray version rate spot on the detector active area was approximately 1 cm . The detector volume was filled with an Ar/CO /CF (60/20/20) K) and pressure gas-mixture, at ambient temperature ( hPa). The gas flux was regulated to have a GEM ( gain independent on the X-ray intensity. The currents induced
on the electrodes were measured with a custom multichannel nanoampere meter connected between the detector and the voltage supply. The primary ionization current was estimated by measuring the cathode current while switching off the potential across the GEM and while changing the sign of the electric field in the induction gap, to avoid including in the measurement the primary ionization in the induction gap. The result was nA. Measurements were taken with various configuration of electric fields and GEM-voltage, usually by performing a scan in one of the parameters while keeping fixed the other two. • GEM-voltage scan in the range 300 V–500 V, with a drift field of 3 kV/cm and an induction field of 5 kV/cm. • Scan in the induction field between 0 and 8 kV/cm, while keeping the drift field at 3 kV/cm and for three different values of the GEM-voltage: 340, 390, and 440 V. • Scan in the drift field in the range 1 kV/cm–8 kV/cm, with an induction field of 5 kV/cm and for three different values of the GEM-voltage: 340, 390, and 440 V. IV. ANALYSIS An analytical model was developed in order to extract the four characteristic parameters of the single-GEM detector. Each detector parameter was expressed by an analytical function of the drift and induction fields and the GEM voltage. In these functions some numerical coefficients were included. Their values were obtained by means of a global fit of the currents of all electrodes in all measurements. of all the squared differences Mathematically, the sum ) between the mea(weighted by the experimental errors and the ones given by the model sured currents (11) was minimized by means of MINUIT [5]. In this expression represents the j-th measurement and takes the values , and , one for each of the four electrodes. The analytical functions were defined as empirical parametrization of the electric fields and the voltage applied to the GEM, based on educated guesses arising from studies on detector behavior [2], [3] and from a detailed microscopic simulation [4]: (12) (13) (14) (15) (16) , and . where the free parameters of the fit are represents the effective electric field present into the holes. has the value of 1 cm/V and was included only to have a dimensionless exponent in (13). The two external field with the same weight because of the complete contribute to geometrical symmetry of the GEM holes with respect to a plane containing all the hole centers.
BENCIVENNI et al.: MEASUREMENT OF GEM PARAMETERS WITH X-RAYS
Fig. 2. Electrode currents on the cathode and on the GEM electrode facing the cathode (a), on the GEM electrode facing the anode and on the anode (b) from measurements (dots) compared to the ones calculated by the analytical model , for E = 3 kV/cm and E = 5 kV/cm. (full lines) as a function of V
1299
Fig. 3. Electrode currents from measurements (dots) compared to the ones calculated by the analytical model (full lines) as a function of E , for E = = 440 V. 3 kV/cm and V
The free parameters have been determined by substituting expressions (12)–(16) in (6)–(9). Equation (11) was then minimized with MINUIT. The additional “scale” parameter—the current of primary ionization —was left free in the fit, and its comparison with the measured value was also used as a further check on the validity of the model. V. RESULTS The global fit procedure converges and gives the following values for the parameters:
By replacing the fitted values in (6)–(9) the values of the currents given by the model can be compared with experimental data. Figs. 2 and 3 show this comparison respectively as a funcand of . A good agreement is generally found, tion of however some deviations from the model are visible for small and for large . Another comparison can be performed between some parameters that can be derived directly from curis defined as the rent measurement. The effective gain mean number of electrons that reach the anode for each electron produced in the drift gap (17)
Fig. 4. Effective gain G from the analytical model as a function of V compared with experimental data, for E = 3 kV/cm and E = 5 kV/cm.
Fig. 4 shows the comparison between the effective gain measured experimentally and the one given by the model, as a func. The current induced by the primary ionization in tion of was estimated by setting to zero and by the drift gap measuring the cathode current. A good agreement was found, the model slightly overestimates the measurements and we attribute this to a systematic error in the measurement of . Fig. 5 shows the comparison between the electron extraction efficiency estimated experimentally using (10) and the one given by the model. The shape of the two curves differs slightly
1300
Fig. 5. Electron extraction efficiency ( ) from the analytical model as a function of the induction field (E ) compared with experimental data, for E = = 440 V. 3 kV/cm and V
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 50, NO. 5, OCTOBER 2003
Fig. 7. Gain (G) as a function of the GEM-voltage (V analytical model.
Fig. 8. Electron extraction efficiency ( (E ) from the analytical model. Fig. 6. Electron collection efficiency ( ) as a function of the drift field (E ) from the analytical model.
but the measured values are always within 10% from the analytical function. The behavior of the model parameters are summarized in Figs. 6 –9. Fig. 6 shows the electron collection efficiency given by the model as a function of the drift field, for three values of and for kV/cm. The increase in electron collection effiis clearly reproduced by the model. ciency with Fig. 7 shows the GEM gain given by the model as a function , for kV/cm and kV/cm. of
) from the
) as a function of the induction field
Fig. 8, (9) shows the electron (ion) extraction efficiency as a function of the induction (drift) field, for three and for kV/cm ( kV/cm). The values of decrease in the electron and ion extraction efficiencies with is clearly reproduced by the model. VI. OTHER METHODS While no obvious method was found to directly measure all GEM parameters simultaneously, small changes in the detector configuration and of the measurement setup allow to directly measure ion collection and extraction efficiencies.
BENCIVENNI et al.: MEASUREMENT OF GEM PARAMETERS WITH X-RAYS
Fig. 9. Ion extraction efficiency ( from the analytical model.
Fig. 10.
) as a function of the drift field (E )
Schematic diagram of the measurements setup with the Fe
1301
Fig. 11. Ion collection efficiency ( ) in Ar/CO (70/30) from direct measurement as a function of the drift field (E ), for various voltages applied to the GEM.
source.
It is obvious, from electron extraction efficiency measurements, that ion extraction efficiencies could be measured in an identical way, provided that in the region where ions are extracted there should be no electron contribution to the electrode currents. This can be obtained by illuminating with X-rays only the induction gap. In this way only ions are collected by GEM holes and are subsequently extracted toward the cathode. To illuminate with X-rays only the induction gap a 1 mCi pointlike Fe source was inserted inside the detector box, between the anode and the GEM frame (Fig. 10). However, in order to increase the small ion currents at the cathode a triple-GEM structure was inserted between the GEM under measurement and the anode to provide electron amplification, and by consequence an increase of the ion current. With this setup ion collection and extraction efficiencies can be measured directly from current ratios. It should be noticed here that, as opposed to what happens during electron measurements, these two parameters can be measured together because there is no ion amplification into the GEM holes, or, rephrasing it differently, because there is no contribution from electrons to the measured currents. If one labels and the currents mea-
Fig. 12. Ion collection efficiency ( ) in Ar/CO (70/30) from direct measurement as a function of the voltages applied to the GEM, for various E values.
sured respectively on the GEM bottom and top electrodes, and if represents the cathode current, we have (18) (19) Preliminary results obtained with an Ar/CO (70/30) gasmixture are shown in Figs. 11 and 12, where the ion collection efficiencies as a function of respectively the ion collection field and the voltage applied to the GEM are shown. One can immediately notice that the ion collection efficiency approches 100%
1302
Fig. 13. Ion extraction efficiency ( ) in Ar/CO (70/30) from direct measurement as a function of the drift field (E ), for various voltages applied to the GEM.
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 50, NO. 5, OCTOBER 2003
Fig. 15. Electron and ion collection efficiency in Ar/CO (70/30) as a function of the collecting field, for V = 100 V.
because there is no charge multiplication into the GEM holes, it is immediate to understand that electron extraction and collection efficiencies can be measured directly only if there is no multiplication inside the GEM holes. For this reason we attempted to measure these efficiencies for low voltages applied to the GEM. In Fig. 15 preliminary measurement of the electron and the ion collection efficiencies as a function of the collecting V are shown. As expected, the ion field and for efficiencies are always larger than the electron ones. VII. CONCLUSION
Fig. 14. Ion extraction efficiency ( ) in Ar/CO (70/30) from direct measurement as a function of the voltages applied to the GEM, for various E values.
for any reasonable field (2–3 kV/cm) and voltage applied to the V). GEM ( Preliminary results for the ion extraction efficiencies in Ar/CO (70/30) are shown in Figs. 13 and 14. The ion extraction efficiency always exceeds 70% for extracting field values in the range 3–5 kV/cm and for voltages applied to the GEM in the range 250–450 V. It can also be noticed that for extracting field values in the range of interest for time projection chambers V/cm) the ion extraction efficiency increases linearly ( with the electric field and does not exceed 10%–15%. If one considers the fact that ion extraction and collection efficiencies can be measured directly from current measurements
A simple analytical model able to describe the currents on the electrodes of a single-GEM detector exposed to a continuous ionizing radiation has been studied. By using empirical parametrization of the detector functional parameters and a global fit to the data a good agreement has been found between the measured currents and the ones given by the model. This work, easily extendable to multi-GEM detectors, can be used to optimize the detector performances. It should be noted that results reported in this work depend on the GEM geometry and on the particular gas mixture used, but not on the gap thicknesses. Other methods capable of evaluating GEM parameters directly from current measurements were also proposed, and some preliminary results have been presented. REFERENCES [1] F. Sauli, “GEM: A new concept for electron amplification in gas detectors,” Nucl. Instrum. Methods, vol. A 386, pp. 531–534, 1997. [2] G. Bencivenni et al., “A fast triple-GEM detector for high-rate chargedparticle triggering,” Nucl. Instrum. Methods, vol. A 488, pp. 493–502, 2002. [3] S. Bachmann, A. Bressan, L. Ropelewski, and F. Sauli, “Operating properties of detectors based on GEM,” in Proc. Int. Workshop Micro-Pattern Gas Detectors, Orsay, France, June 1999, pp. 28–30. [4] G. Bencivenni, W. Bonivento, A. Cardini, F. Murtas, and D. Pinci, “A complete simulation of a GEM detector,” IEEE Trans. Nucl. Sci., vol. 49, pp. 1638–1643, Aug. 2002. [5] F. James, “MINUIT,”, 1998. CERN Library Long Writeup D506.
