Next: CONCLUSION Up: FIRST RESULTS ALREADY PUBLISHED Previous: ELECTROSTATIC SIMULATION

MGWC tests

For the prototype tests, LeCroy TRA403 current preamplifiers used for the DELPHI Barrel RICH have been used. These preamplifiers have an input resistor of 1000  , a capacitance of 5 pF, an intrinsic rise-time of 6 ns, an amplification factor of 95 mV/ µ A and they saturate near 3  µ A. An extra connection card was necessary for the preamplifier installation, increasing considerably the anode and cathode capacitance compared to the chamber capacitance itself (integrated electronics would drastically reduce the external electronic noise and the signal dispersion).

To create primary electrons in the 3 mm wide drift volume, two different ionization sources were used. First, ⁹⁰Sr (beta) and ⁵⁵Fe (X-ray) radioactive sources were used. In this case, an aluminized mylar 6  µm foil has been used as drift electrode to minimize the energy loss in this window. Second, an adjustable number of primary electrons inside the chamber was produced with a nitrogen pulsed laser. The electrons were created by the ``two-photon'' process [9] inside the drift gas, on the cathode surface (gold) or on the drift electrode surface. In this case, to allow the U.V photons to penetrate in the sensitive area, a quartz window was used. On the inner quartz window face a 50 Å tungsten layer has been evaporated in order to apply a drift voltage (the U.V transmission was about 45%). An optical fiber having 500  µ m diameter has been used in order to bring the U.V light from the laser up to the chamber. Moving the fiber, photoelectrons were created at any desired position inside the chamber. The number of produced electrons was varied by adjusting the laser intensity. This allowed to study the chamber reaction in single electron mode.

Fig. 9 shows an anode signal and the corresponding cathode response using a 90% Ar + 10% CH₄ gas mixture and the ⁹⁰Sr source. For this mixture the anode-cathode voltage was fixed to -300 V. The rise-time of the anode signal is less than 8 ns, which exceeds the preamplifier's rise-time itself only by 2 ns. The signal does not last more than 40 ns. Since the cathode receives the signal of 64 anode wires, its own signal is strongly saturated.

Figure 9 : Signals observed using the gas mixture 90% Ar + 10% CH₄. The anode rise-time is 7.6 ns.

A second gas mixture was also used, 90% He + 10% CH₄ , with an anode-cathode voltage of -400 V. Fig. 10 shows the anode and cathode signals obtained with several electrons produced by the ⁹⁰Sr source (in order to show the correlation between the anode and cathode signals a case where the cathode signal was not saturated has been selected). Here, the drifting electrons are better separated than for the 90 % Ar + 10 % CH₄ mixture, due to the lower drift velocity.

Figure 10: Signals observed using the gas mixture 90 % He + 10 % CH₄ .

Fig. 11 shows the anode signal produced by the laser with the 75 % CH₄ + 25 % C₂ H₆ gas mixture and an anode-cathode voltage of -575 V. The two peaks observed on the anode, separated by 40 ns, come from electrons created on the cathode surface (almost no drift path) and from those created on the surface of the tungsten layer of the quartz window (3 mm drift path).

Figure 11: Anode signal observed using the gas mixture 75 % CH₄ + 25 % C₂ H₆ and the laser light.

To measure the gain of the chamber, the laser intensity has been considerably reduced, in order that the probability to have two and more detected electrons to be negligible compared to the probability to have only one electron (the laser intensity stability has been measured to be below 3%). Fig. 12 shows the induced mean charge value of the observed pulses, for pulses overpassing a relatively low threshold ( ~1 fC), as a function of the laser intensity and for an anode-cathode voltage of -575 V and the gas mixture 75 % CH₄ + 25 % C H₆ . For a high laser intensity, several primary photoelectrons are produced inside the chamber. Their number decreases when the laser intensity decreases reducing thus the observed mean charge per pulse. Below a certain value of the laser intensity, the mean charge does not vary any more because, when a pulse overpasses the electronic threshold, the probability that this signal comes for only one photoelectron is very high compared to the probability to observe more than one photoelectron. Thus, the observed charge spectrum is the one produced by only one primary photoelectron. The flat part of the curve at low intensity indicates the ``single photoelectron mode''. For each laser intensity a Polya distribution has been fitted on the charge distribution in order to extract the mean charge value avoiding the low charge region affected by the electronic threshold and noise. The electronic channels have been calibrated using the same method as in [8]. A known charge is injected at the entry of the preamplifier using a 1 pF capacitor and a fast pulse generator (rise-time  ~1 ns). Thus, the measured gain for an anode-cathode voltage of -575 V was 15 fC, i.e 10 ⁵ electrons.

Figure 12: Gain versus the laser relative intensity for two different anode wires ( 75 % CH₄ + 25 % C₂ H₆ ).

Fig. 13 shows the gain as a function of the anode-cathode voltage. The fitted exponential distribution has a slope of 15.5 V -1 . In the future, wires with a diameter of 5  µ m will be used. In this configuration and for the same voltage, the gain is expected to be about 2 times higher than the one obtained with 10  µ m wires (GARFIELD).

Figure 13: Gain versus anode-cathode voltage.

To study the intrinsic rise-time of the detector and the signal duration without being limited by the rise-time and the RC differentiation constant of the preamplifier, a unity-gain buffer amplifier was used. The amplifier rise-time was 0.7 ns while the input capacitance and resistance amounted for 4 pF and 1 M respectively. Fig. 14 displays the average signals observed with an anode-cathode voltage of -550 V, -560 V and -575 V. The total signal duration, corresponding to the total ion drift time, was less than 2  µ s, proving that almost all the positive ions were collected by the cathode, as shown in fig. 8. The rise-time was of the order of 2 ns, indicating that the intrinsic signal rise-time was less than 1.5 ns.

Figure 14: Anode signal observed using a fast and high input resistance amplifier.

Although a large number of ageing studies on chambers using wires have already been performed, a study on this type of detector has only been started. In the future, other substrates with lower dielectric constant and better resistance to radiations, like kapton, will be used. A segmentation of the cathode in strips in order to have a x-y readout will also be performed.

  ...amplifier       BUF601 high-speed buffer amplifier, Burr-Brown Corporation.

Marcos Dracos Sat Apr 4 18:31:19 METDST 1998

[Accueil] [MGWC]