5.5.3.8  Photomultipliers

Five photomultiplier tubes (PMT) are fixed of the two side walls. Each one is surrounded by a high magnetic permeability shielding (mu-metal). The fixing provides high voltage insulation between the PMT and the stell vessel. A set of optic fibers provides light pulses to each PMT for their calibration.

Characteristics

The type of the PMT is a Burle 8854 Photomultiplier. This is a 129 mm (5 inch) diameter, 14 stage head on QUANTACON photomultiplier having a bialkali photocathode of high quantum efficiency and an extremely high gain cesiated gallium phosphide first dynode followed by high stability copper beryllium dynodes in the succeeding stages. The high gain first dynode permits the direct observation of peaks corresponding to one two and three photoelectrons.

The 8854 features high quantum efficiency (22.5% at 385 nm), ultraviolet response (UV response to 220 nm). The number of photon from the Cerenkov light is proportionnal to so it's important to have a extended response in UV light. Figure show a typical curve for the quantum efficiency as a function of the wavelenght.

  

Figure: Quantum efficiency of the photocathode as a function of the

Figure shows a typical ADC pluse height spectrum (with High Voltage around 2000V) where the single photoelectron peak is very well separated from other sources of noise. Thus they allow us to clearly define the electronic threshold to eliminate events below this peak. The number of photoelectron produced in the Cerenkov is expected to be large enough () to put a threshold around few photoelectrons which garantee to get rid of the noise and have a very good efficiency.

The same PMTs have been used for the aerogel Cerenkov detector for the TJNAF Hall A and different configurations of the voltage dividers have been tested (see ref NIM A309(1991)552-554). The optimized configuration with high photocathode to first dynode voltage has been adopted.

  

Figure: Typical ADC spectrum. The single peak is clearly separated form the noise and the peak for 2 is also seen.

Numbering scheme

Figure and figure show the correspondance between the serial number of the PMT and the number from 1 to 10 in the Cerenkov (which correspond to a position in the box) in the hadron and electron arm.

  

Figure: Numerotation for the PMT of the Cerenkov counter of the electron arm.

  

Figure: Numerotation for the PMT of the Cerenkov counter of the hadron arm.

High Voltage adjustment

If an electron is above the Cerenkov threshold, it will produce Cerenkov light in the gas. This light will be translate into an electric signal via a photomultiplicator. Each photomutiplicator will receive the light transmitted by his mirror. As a group of mirror can be touched by a good event, we expect to have sometimes a group of touched PMT. Before the discrimination of the signal we have to sum the signals over the touched PMT. In order to be able to make the summ, we have to know a relative calibration between the ten PMT of a counter.

For the further analysis of experiment which will use these Cerenkov counters, we need to know the efficiency of these counters. From the spectrum of charge, we can find this efficiency from the number of mean photoelectrons in this spectrum (see formula ). To obtain this mean number of photoelectrons from the ADC spectrum, we need to make an absolute calibration in energy of each PMT.

The high Voltage has to be adjust in order to have for each PMT the position of the photoelectron peak at the same position which is around few 20 channels above the pedestal. This adjustment has to verify some conditions:

-

to correspond to a High Voltage where the PMT noise is not high (below 2kHz)

-

to correspond to the maximum value of the quantum efficiency of the photocathode.

If the two conditions are not possible at the same time, a compromise has to be found.

Noise measurement

Figure shows the PMT noise as a function of the High Voltage. The analogic signal of each PMT was dicriminated with a low threshold (60 mV) and the result of this discrimination was going to a scaler. The scaler rate represent the noise in this case because it was done with the PMT in their final position, in the Cerenkov box which has no leak of light, and without beam or any source light except cosmic of course...

We can see from the measurement that below a High Voltage of 2500V, all PMT have a noise rate below 2 kHz except the number 3 (in electron Cerenkov) which has to be perhaps change for the future. This rate is relatively high because the threshold is very low. As we are working with threshold Cerenkov counters which give only photoelectrons (theoretical value which has to be confirm with electron beam) we have to work with a low threshold.

  

Figure: PMT noise as a function of the HV. This was obtained with a low threshold of discrimination ( 60 mV)

Relative gain measurement

From figure you can extract a gain (in arbitrary unit) from the position of the single photoelectron peak. This was done for all the 20 PMT (10 for each Cerenkov counter) at three different High Voltage, 1800, 2000 and 2200 Volts. The PMT was placed in a black box (no leak of light) and an optical fiber was coming just in front of the photocathode. A source of green light (diode) with adjustable amplitude was used and once we suceed to tune this amplitude in order to see the photoelectron peak in a given PMT, the same amplitude was used for each PMT. The results are reported in the table and are also plotted in figurepmga.

 

PMT n⁰ HV₁=1800 V HV₂=2000 V HV₃=2200 V

1 20 75 231

2 45 130 390

3 31 123 saturation

4 21 78 233

5 40 132 410

6 21 75 243

7 18 30 90

8 24 93 283

9 21 74 255

10 16 56 178

11 30 95 280

12 15 45 133

13 20 75 230

14 17 65 208

15 20 52 150

16 21 76 237

17 25 80 250

18 15 62 170

19 15 53 171

20 no peak 15 51

 

We can see that number 3 has a relative gain higher than the others. The noise measurement shows already that this PMT get a lot of noise and at 2200 Volts it is on saturation. The PMT number 20 has a low gain and no single photoelectron peak was seen at 1800 V.

  

Figure: Relative gain in arbitrary units for all the PMTs at three different High Voltage 1800,2000 and 2200 Volts.

Total radiation lenght

When a particle cross the Cerenkov detector, it will go through:

-

the entrance and exit window

-

the gas medium

-

one mirror

The total radiation lengh for the whole Cerenkov detector is wich is equivalent of 5 mm of scintillator.