- Email: [email protected]
A fast gas ionization calorimeter filled with C3F8 for operation at high counting rates and hard radiation environment
Nuclear Instruments and Methods in Physics Research A 419 (1998) 590—595
A fast gas ionization calorimeter filled with C F for operation at high counting rates and hard radiation environment S. Denisov , A. Dushkin , N. Fedyakin *, Yu. Gilitsky , M. Ljudmirsky, A. Spiridonov , V. Sytnik Institute for High Energy Physics, 142284, Protvino, Moscow region, Russian Federation Moscow State University, 119899, Moscow, Russian Federation
Abstract The performance of a gas ionization EM calorimeter with planar electrodes and steel absorbers has been studied with a 26.6 GeV/c electron beam at the 70 GeV IHEP accelerator. The design of the calorimeter is optimized for the operation at high counting rates by minimizing the coupling inductance and by choosing rather fast and heavy perfluoroalkane C F (v "0.07 mm/ns at a reduced field E/N"1.0;10\ V cm). This gas has been used for the first time in calorimetry applications. The total calorimeter thickness is +21X . The signal readout has been done by remote 25 ) low-noise preamplifiers coupled to towers via 25 ) cable of 4 m length. The choice of a 25 ) input impedance provides a complete matching between preamplifier, cable and tower. The studies of the calorimeter consisted in measuring the signal and noise spectra at different values of HV, ADC gate width and gas pressure. The electron attachment rate in C F with a stated purity of 99.99% is quite low (at a given E/N the mean free path of electrons is j"2.2 cm at 1 atm). The intrinsic energy resolution of the calorimeter after noise subtraction is found to be independent of the gas pressure and equal to +7% at E"26.6 GeV/c. 1998 Elsevier Science B.V. All rights reserved.
1. Introduction The main purpose for the design of gas ionization calorimeters is the operation at very high counting rates and in a high-level radiation environment. Extensive studies in gas ionization calorimetry have been performed in the last five years [1—6]. It has been experimentally proved that hadron/electromagnetic calorimeters with a planar electrode geometry filled with a 90%Ar#10%CF * Corresponding author.
gas mixture possess a number of attractive features like a good energy resolution, high uniformity and stability, simple calibration and high intrinsic radiation hardness at low cost [5,6]. The only disadvantage of these calorimeters is a high gas pressure necessary to obtain a reasonable signal/noise ratio. As a possible solution to reduce gas pressure we propose to use gases with higher densities. In the paper the results of the first study of a gas ionization calorimeter filled with a heavy perfluoroalkane C F are reported. The choice of C F is optimal for calorimetry applications for the following reasons:
0168-9002/98/$19.00 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 9 8 ) 0 0 8 3 6 - 5
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
591
Fig. 1. Equivalent electrical circuit of the calorimeter readout. The front tower (section of 6 bigaps) is equipped with hybrid 0—¹ preamplifiers, followed by monolithic EL2075 amplifiers. The back tower was readout by the circuit consisting of two discrete CB configurations. Charge-sensitive ADCs with a variable gate width (30—145 ns) were used to digitize signals of the front and back towers. The average EM shower profile is shown by arrows.
Fig. 2. The total noise of two towers summed up over the depth as a function of ADC gate width with HV on and off, correspondingly. The measured noise is quite close to SPICE simulations.
VIII. CALORIMETRY
592
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
E high density (4.4 times higher than that of a 90%Ar#10%CF mixture) [5,6]; E high drift velocity v "0.07(0.104) mm/ns at E/N"1.0(4.0);10\ V cm [7]; E relatively low electron attachment rate (the mean free path is j"2.2 cm at 1 atm pressure and at E/N"1.0;10\ V cm for 99.99% purity); E the next perfuoroalkanes in the series of n-C F exhibit an electron attachment rate a hundred times higher.
absorbers were grounded and HV was applied to the signal electrodes. The ratio of HV to pressure in the case of C F was +830 V/atm (v + 0.07 mm/ns) [7], while tests with a 90%Ar# 10%CF mixture have been performed at HV/P+185 V/atm (v +0.12 mm/ns) [8]. The calorimeter was designed to operate at a pressure of up to 16 atm.
3. Readout system 2. Calorimeter design The calorimeter tested consists of 32 longitudinal towers of ionization chambers with a planar geometry interleaved with 30 mm thick steel absorbers. Signal pads made of 1 mm steel were placed between absorbers forming two 3 mm gaps (or a single bigap), which guarantees a full drift time of t "42 ns at a given E/N. There are 4;4 pads in the signal plane and the pad size is 76;76 mm. The total calorimeter thickness is about 21X . The
Each tower consisting of six consecutive chambers can be modelled by a distributed LC-line with a characteristic impedance Z "25 ) (C"33 pF *! and ¸"20 nH). To place the electronics away from the radiation hot area and to provide complete matching between preamplifier and tower, remote low-noise 25 ) preamplifiers have been coupled to the corresponding front and back tower via cables of 4 m length with a 25 ) impedance. This condition is extremely important if one wants to operate at very high counting rates. The front section
Fig. 3. Noise and signal spectra measured at P"5 atm of C F .
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
593
Fig. 4. The pressure dependence of the signal for the mixture 90%Ar#10%CF (opened circles) and freon C F (closed markers), at E/N being kept at 2;10\ V cm for the mixture and 1.0;10\ V cm for C F , correspondingly. Industrial freon with different stated purity (from 99% to 99.99%) has been studied. The values of the electron mean free path j are referred to 1 atm pressure.
Fig. 5. Signal variation with HV at different pressures of C F and 90%Ar#10%CF .
VIII. CALORIMETRY
594
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
was equipped with hybrid 0—¹ preamplifiers (e "0.38 nV/(Hz), followed by monolithic L EL2075 amplifiers with a stated noise referred to the input e "2.3 nV/(Hz. The back sections were L readout by a couple of common-base configurations, connected in series (the details can be found elsewhere [1]). The schematic diagram of the signal readout from the front and the back sections is shown in Fig. 1. The front solution allowed to reduce the noise by 20%. The total noise of the front and back towers, which approximately corresponds to the volume occupied by the EM shower, is shown in Fig. 2 as a function of the ADC gate width.
4. Measurements and results All measurements of the calorimeter response and its resolution have been performed with 26.6 GeV/c e\ beam at the 70 GeV IHEP (Protvino) accelerator. The studies were carried out at different values of HV, ADC gate width and gas
pressure. The events satisfying the following criteria were selected for analysis: E signals in any tower around the beam tower should be 410% of the signal in the beam tower; E energy deposition in the front tower is higher than that in the back one, while exceeding the noise level. These criteria allow to reject the background events due to muons, hadrons and beam halo. In the selected events more than 98% of the energy was released in two beam towers, while the signals in the neighboring towers have been essentially less than the noise level. That is why only beam tower signals have been used in the data analysis. Typical distributions of noise and shower signals taken at t "59 ns are presented in Fig. 3. The dependencies of the signals on the gas pressure are shown in Fig. 4 for the mixture 90%Ar#10%CF (open circles) and for different C F samples taken from sources with different stated chemical purity (from 99% to 99.99%). Fig. 5 demonstrates that the
Fig. 6. The transversal nonuniformity of the detector. Signal variation is less than 2.6%.
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
595
Fig. 7. The dependence of the energy resolution on C F pressure with (closed) and without (open) noise subtraction for t "30 ns. The same results have been obtained for t "59—145 ns.
calorimeter response weakly depends on the HV. A beam scan done in the transversal direction shows that the calorimeter exhibits a quite good transversal uniformity (see Fig. 6). We proved that the energy resolution (after noise subtraction) does not depend on the C F pressure (Fig. 7), which is consistent with our earlier measurements, made with 90%Ar#10%CF [6], and it scales with absorber thickness as (t.
Acknowledgements We express our gratitude to A. Soldatov for the design of the calorimeter and to I. Belyakov, A. Godov, O. Romashov, I. Shein, S. Zvyagintsev
for their help in assembling and testing the calorimeter. References [1] S. Denisov et al., Nucl. Instr. and Meth. A 335 (1993) 106. [2] S. Denisov et al., Pribori i Technika Experimenta 5 (1993) 34. [3] N.D. Giokaris et al., Nucl. Instr. and Meth. A 333 (1993) 364. [4] D. Khazins et al., Nucl. Instr. and Meth. A 233 (1993) 372. [5] S. Denisov et al., Proc. 4th Int. Conf. on Calorimetry in High Energy Physics, Elba, Italy, 1993, p. 49. [6] S. Denisov et al., Proc. 5th Int. Conf. on Calorimetry in High Energy Physics, BNL, USA, 1994, p. 380. [7] S. Hunter et al., Phys. Rev. A 38 (1988) 58. [8] A. Peisert, F. Sauli, Preprint CERN-EP/84-08, 1984.
VIII. CALORIMETRY
A fast gas ionization calorimeter filled with C F for operation at high counting rates and hard radiation environment S. Denisov , A. Dushkin , N. Fedyakin *, Yu. Gilitsky , M. Ljudmirsky, A. Spiridonov , V. Sytnik Institute for High Energy Physics, 142284, Protvino, Moscow region, Russian Federation Moscow State University, 119899, Moscow, Russian Federation
Abstract The performance of a gas ionization EM calorimeter with planar electrodes and steel absorbers has been studied with a 26.6 GeV/c electron beam at the 70 GeV IHEP accelerator. The design of the calorimeter is optimized for the operation at high counting rates by minimizing the coupling inductance and by choosing rather fast and heavy perfluoroalkane C F (v "0.07 mm/ns at a reduced field E/N"1.0;10\ V cm). This gas has been used for the first time in calorimetry applications. The total calorimeter thickness is +21X . The signal readout has been done by remote 25 ) low-noise preamplifiers coupled to towers via 25 ) cable of 4 m length. The choice of a 25 ) input impedance provides a complete matching between preamplifier, cable and tower. The studies of the calorimeter consisted in measuring the signal and noise spectra at different values of HV, ADC gate width and gas pressure. The electron attachment rate in C F with a stated purity of 99.99% is quite low (at a given E/N the mean free path of electrons is j"2.2 cm at 1 atm). The intrinsic energy resolution of the calorimeter after noise subtraction is found to be independent of the gas pressure and equal to +7% at E"26.6 GeV/c. 1998 Elsevier Science B.V. All rights reserved.
1. Introduction The main purpose for the design of gas ionization calorimeters is the operation at very high counting rates and in a high-level radiation environment. Extensive studies in gas ionization calorimetry have been performed in the last five years [1—6]. It has been experimentally proved that hadron/electromagnetic calorimeters with a planar electrode geometry filled with a 90%Ar#10%CF * Corresponding author.
gas mixture possess a number of attractive features like a good energy resolution, high uniformity and stability, simple calibration and high intrinsic radiation hardness at low cost [5,6]. The only disadvantage of these calorimeters is a high gas pressure necessary to obtain a reasonable signal/noise ratio. As a possible solution to reduce gas pressure we propose to use gases with higher densities. In the paper the results of the first study of a gas ionization calorimeter filled with a heavy perfluoroalkane C F are reported. The choice of C F is optimal for calorimetry applications for the following reasons:
0168-9002/98/$19.00 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 9 8 ) 0 0 8 3 6 - 5
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
591
Fig. 1. Equivalent electrical circuit of the calorimeter readout. The front tower (section of 6 bigaps) is equipped with hybrid 0—¹ preamplifiers, followed by monolithic EL2075 amplifiers. The back tower was readout by the circuit consisting of two discrete CB configurations. Charge-sensitive ADCs with a variable gate width (30—145 ns) were used to digitize signals of the front and back towers. The average EM shower profile is shown by arrows.
Fig. 2. The total noise of two towers summed up over the depth as a function of ADC gate width with HV on and off, correspondingly. The measured noise is quite close to SPICE simulations.
VIII. CALORIMETRY
592
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
E high density (4.4 times higher than that of a 90%Ar#10%CF mixture) [5,6]; E high drift velocity v "0.07(0.104) mm/ns at E/N"1.0(4.0);10\ V cm [7]; E relatively low electron attachment rate (the mean free path is j"2.2 cm at 1 atm pressure and at E/N"1.0;10\ V cm for 99.99% purity); E the next perfuoroalkanes in the series of n-C F exhibit an electron attachment rate a hundred times higher.
absorbers were grounded and HV was applied to the signal electrodes. The ratio of HV to pressure in the case of C F was +830 V/atm (v + 0.07 mm/ns) [7], while tests with a 90%Ar# 10%CF mixture have been performed at HV/P+185 V/atm (v +0.12 mm/ns) [8]. The calorimeter was designed to operate at a pressure of up to 16 atm.
3. Readout system 2. Calorimeter design The calorimeter tested consists of 32 longitudinal towers of ionization chambers with a planar geometry interleaved with 30 mm thick steel absorbers. Signal pads made of 1 mm steel were placed between absorbers forming two 3 mm gaps (or a single bigap), which guarantees a full drift time of t "42 ns at a given E/N. There are 4;4 pads in the signal plane and the pad size is 76;76 mm. The total calorimeter thickness is about 21X . The
Each tower consisting of six consecutive chambers can be modelled by a distributed LC-line with a characteristic impedance Z "25 ) (C"33 pF *! and ¸"20 nH). To place the electronics away from the radiation hot area and to provide complete matching between preamplifier and tower, remote low-noise 25 ) preamplifiers have been coupled to the corresponding front and back tower via cables of 4 m length with a 25 ) impedance. This condition is extremely important if one wants to operate at very high counting rates. The front section
Fig. 3. Noise and signal spectra measured at P"5 atm of C F .
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
593
Fig. 4. The pressure dependence of the signal for the mixture 90%Ar#10%CF (opened circles) and freon C F (closed markers), at E/N being kept at 2;10\ V cm for the mixture and 1.0;10\ V cm for C F , correspondingly. Industrial freon with different stated purity (from 99% to 99.99%) has been studied. The values of the electron mean free path j are referred to 1 atm pressure.
Fig. 5. Signal variation with HV at different pressures of C F and 90%Ar#10%CF .
VIII. CALORIMETRY
594
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
was equipped with hybrid 0—¹ preamplifiers (e "0.38 nV/(Hz), followed by monolithic L EL2075 amplifiers with a stated noise referred to the input e "2.3 nV/(Hz. The back sections were L readout by a couple of common-base configurations, connected in series (the details can be found elsewhere [1]). The schematic diagram of the signal readout from the front and the back sections is shown in Fig. 1. The front solution allowed to reduce the noise by 20%. The total noise of the front and back towers, which approximately corresponds to the volume occupied by the EM shower, is shown in Fig. 2 as a function of the ADC gate width.
4. Measurements and results All measurements of the calorimeter response and its resolution have been performed with 26.6 GeV/c e\ beam at the 70 GeV IHEP (Protvino) accelerator. The studies were carried out at different values of HV, ADC gate width and gas
pressure. The events satisfying the following criteria were selected for analysis: E signals in any tower around the beam tower should be 410% of the signal in the beam tower; E energy deposition in the front tower is higher than that in the back one, while exceeding the noise level. These criteria allow to reject the background events due to muons, hadrons and beam halo. In the selected events more than 98% of the energy was released in two beam towers, while the signals in the neighboring towers have been essentially less than the noise level. That is why only beam tower signals have been used in the data analysis. Typical distributions of noise and shower signals taken at t "59 ns are presented in Fig. 3. The dependencies of the signals on the gas pressure are shown in Fig. 4 for the mixture 90%Ar#10%CF (open circles) and for different C F samples taken from sources with different stated chemical purity (from 99% to 99.99%). Fig. 5 demonstrates that the
Fig. 6. The transversal nonuniformity of the detector. Signal variation is less than 2.6%.
S. Denisov et al. /Nucl. Instr. and Meth. in Phys. Res. A 419 (1998) 590—595
595
Fig. 7. The dependence of the energy resolution on C F pressure with (closed) and without (open) noise subtraction for t "30 ns. The same results have been obtained for t "59—145 ns.
calorimeter response weakly depends on the HV. A beam scan done in the transversal direction shows that the calorimeter exhibits a quite good transversal uniformity (see Fig. 6). We proved that the energy resolution (after noise subtraction) does not depend on the C F pressure (Fig. 7), which is consistent with our earlier measurements, made with 90%Ar#10%CF [6], and it scales with absorber thickness as (t.
Acknowledgements We express our gratitude to A. Soldatov for the design of the calorimeter and to I. Belyakov, A. Godov, O. Romashov, I. Shein, S. Zvyagintsev
for their help in assembling and testing the calorimeter. References [1] S. Denisov et al., Nucl. Instr. and Meth. A 335 (1993) 106. [2] S. Denisov et al., Pribori i Technika Experimenta 5 (1993) 34. [3] N.D. Giokaris et al., Nucl. Instr. and Meth. A 333 (1993) 364. [4] D. Khazins et al., Nucl. Instr. and Meth. A 233 (1993) 372. [5] S. Denisov et al., Proc. 4th Int. Conf. on Calorimetry in High Energy Physics, Elba, Italy, 1993, p. 49. [6] S. Denisov et al., Proc. 5th Int. Conf. on Calorimetry in High Energy Physics, BNL, USA, 1994, p. 380. [7] S. Hunter et al., Phys. Rev. A 38 (1988) 58. [8] A. Peisert, F. Sauli, Preprint CERN-EP/84-08, 1984.
VIII. CALORIMETRY