RICH Prototype

A prototype of the downstream RICH detector was tested in the T9 test beam at the CERN SPS during the Spring and Summer of 1997. A planar array of seven 61-pixel HPD's from DEP were used to detect the Cherenkov photons produced in aerogel, air and C₄F₁₀ radiators. In configuration-1, fig. 3, of the prototype the light is focussed by a 240 mm focal length mirror which corresponds to a 1/4 scale of the RICH1 detector. A full scale prototype (configuration 2) was also used which has a 1143 mm mirror to focus rings from C₄F₁₀ onto an array of six 61-pixel HPD's. This was achieved by adding extension arms to configuration 1.

The 61 pixel HPD has a silicon diode detector segmented as a hexagonal array with pad dimensions of 2 mm face-to-face. The HPD was operated at a high voltage of 12kV. Using a pulsed light emitting diode the complete readout and data acquisition chain was tested. The pedestal, the single, double and triple photoelectrons peaks were clearly visible with a signal/noise ratio of $\approx$ 5.7. Most of this noise is associated with the input capacitance of the feedthrough and printed circuit boards.

The test beam provides charged particles of either polarity and the momentum can be tuned in the range 2 - 15.5 GeV/c. The particle type is identified by measuring the signal pulse height from a CO₂ threshold Cherenkov counter installed 30 m upstream from the prototype. The prototype vessel was aligned with the beam axis. Charged particles which provide the trigger are selected using using scintillation counters, two upstream and two downstream of the vessel. A photoelectron hit is defined to be a HPD pixel with a signal pulse height 4$\sigma$ above the pedestal mean, where $\sigma$ is the rms width of the pedestal peak.

Using RICH configuration 1, data were taken with a 10 GeV/c negatively charged beam with 18 mm thickness of aerogel. Fig. 4 shows an arc of a ring on the central HPD, whose radius is compatible with that expected from C₄F₁₀. The outer HPD's clearly exhibit a ring which originates from the aerogel radiator.

  

Figure: An event display from aerogel and $\rm C_4F_{10}$ radiators in RICH configuration 1, integrated over run 487.

The number of photoelectrons per triggered event was measured for all three radiators in the vessel. For this analysis a threshold of 3$\sigma$ was set for individual pixels and multiple photoelectrons were taken into account. The mean number of photoelectrons are shown in table 1. The partial geometrical coverage of the aerogel and gas rings was calculated from simulation with $\approx$ 5% uncertainity. The expected photoelectron yields was calculated from simulation which included the properties of the aerogel, mirror and photocathode efficiencies. The overall precision in the expected yield is estimated to be 15%. The comparison between observed and expected yields are given in table 1. The numbers from this preliminary analysis are compatible within 30%.

 

Table 1: Observed number of photoelectrons per event for air, aerogel and C₄F₁₀ radiators. The columns give the number of raw hits, the numbers after correction for background and for geometrical efficiency, and the comparison with the expected yield.

Radiator	Raw	Bkg.	Eff.	Ratio
	hits	corr.	corr.
Air	4.92	4.56	4.80	0.99
C4H10	7.85	7.49	33.55	1.07
Aerogel	1.79	1.31	10.71	0.72

 

The full scale RICH1 prototype was studied using configuration-2. The longer focal length of the mirror means the C₄F₁₀ ring now spans the outer 6 HPDs. The event display shown in fig. 5 is obtained from negatively charged 15.5 GeV/c momentum beam. The K : $\pi$ ratio of the triggering particles has been enhanced to 1 : 2 using the threshold Cherenkov counter. Fig. 5 shows segments of two rings; an inner ring from the incident kaons and an outer ring from the pions. It can be seen that the number of hits observed in HPD 3 is lower than in HPD 4. (Similarly HPD 5 has fewer hits than HPD 2.) This is because HPD's 3 and 5 have mylar windows in front of their photocathodes which absorb the UV photons.

  

Figure: An event display showing $\pi$/K separation, using $\rm C_4F_{10}$ radiator in RICH configuration 2, integrated over run 587.