Beam purity

 In the analysis of particle identification it was crucial to have a good identification of electrons and pions given by detectors separate from the TRT prototype. The beam entering the H8 beamline was a secondary beam produced by directing the 450 GeV beam extracted from the SPS into a target which produced a secondary mixed beam of mainly charged pions and photons (from $\pi^{0}_{}$ decays).

A pion beam was obtained by inserting a lead plate to absorb the photons followed by magnetic optics to select particles of equal energy. As will be shown below the beam had an electron contamination of below 1.4%.

An electron beam was obtained by inserting a copper plate into the beamline where e ⁺e ^- pairs are formed in photon conversions. The magnetic optics following the converter again selected particles of equal energy. The optics was optimised for particles with production point at the copper plate, but still the electron beam had a pion contamination of nearly 50%.

For both the electron and the pion beam external particle identification was required to obtain cleaner electron and pion samples. Three devices were present for this in the testbeam: far upstream a Cherenkov detector suitable for electron/pion separation up to energies of 50 GeV; behind the TRT prototype a presampler; and just behind that a lead glass calorimeter. The presampler was simply a scintillator with a lead plate placed just in front of it. The signals in the Cherenkov counter, the lead glass calorimeter and the preshower are shown in fig. 5.9 for 20 GeV electrons and pions.

  

Figure 5.9: The signals of 20 GeV pions and electrons in the ADC's from the Cherenkov counter (a), the lead glass calorimeter (b) and the preshower (c).

In fig. 5.10 is shown a scatter plot of the signal in the Cherenkov counter and the lead glass calorimeter for a mixed beam of 20 GeV electrons and pions. The regions used for the selection of electron and pion samples are indicated.

  

Figure 5.10: A scatter plot of the signal in the Cherenkov counter and the lead glass calorimeter for a mixed beam of 20 GeV electrons and pions. Electrons and pions for the electron identification analysis were taken from the shaded regions.

The efficiency of finding electrons using the particle identification cuts for pions $\varepsilon_{e}^{}$ and the efficiency for finding pions using the particle identification cuts for electrons $\varepsilon_{\pi}^{}$ are the two important numbers to determine.

$\varepsilon_{e}^{}$ is best measured in a beam of dominantly electrons and an upper limit can be calculated assuming the signals in the Cherenkov counter and the calorimeter give independent measurements. The efficiency for finding electrons when making cuts for pions is

 

$$\varepsilon_{e}^{} \varepsilon_{Ch}^{} \varepsilon_{Cal}^{}$$ (77)

with $\varepsilon_{Ch}^{}$ and $\varepsilon_{Cal}^{}$ the efficiencies for the Cherenkov counter and the calorimeter. Measuring the efficiencies is easy in a beam containing only electrons but unfortunately such a beam was not available.

Four different selection criteria can be defined:

A

Pion identification cuts in both the Cherenkov counter and the calorimeter.

B

Pion identification cut in the Cherenkov counter and electron identification cut in the calorimeter.

C

Electron identification cut in the Cherenkov counter and pion identification cut in the calorimeter.

D

Electron identification cuts in both the Cherenkov counter and the calorimeter.

With X_e and X_$\pi$ defined as the number of electrons and pions after the selection criteria X the total number of particles X_total is given as

 

$$\pi $x \gt 1$$ (78)

Since the cuts on the Cherenkov counter and the calorimeter are independent

 

$$\varepsilon_{Cal}^{} {\frac{{C_e}}{{C_e} + {D_e}}} {\textstyle\frac{1}{1 + \frac{{D_e}}{{C_e}}}}$$ (79)

The fraction of pions in region D will be lower than in region C, so ${\frac{c}{d}}$ > 1 using the notation from (5.3). Inserting this in (5.4) results in an upper limit on the efficiency

 

$$\varepsilon_{Cal}^{} {\textstyle\frac{1}{1 + \frac{c{D}}{d{C}}}} {\textstyle\frac{1}{1 + \frac{{D}}{{C}}}}$$ (80)

However, the large number of particles in region A on fig. 5.10 indicate a large contamination of pions in the electron beam and (5.5) will be a pessimistic estimate.

The analysis of $\varepsilon_{Ch}^{}$ and $\varepsilon_{\pi}^{}$ < $\varepsilon_{Cal}^{}$($\pi$)$\varepsilon_{Ch}^{}$($\pi$) are similar to the calculation of $\varepsilon_{Cal}^{}$ and the results are
 
An upper limit on the amount of electrons in the pion beam before the particle identification, is given by the fraction of events passing selection criteria D in the pion beam scaled with the fraction of electrons actually passing this selection criteria (estimated from the electron beam). The results are that the pion beam at the most contains 1.4% electrons and the electron beam at the most contains 60% pions.

After the external cuts both beams will be pure to below the level of 10^{- 4} which is sufficient for the electron identification studies where rejections below 5 $\cdot$ 10^{- 3} are never reached. The results are summarised in table 5.2

 

Maximum fraction Maximum fraction

before selection after selection

$$\cdot$$ electrons in pion beam 0.014

$$\cdot$$ pions in electron beam 0.603