_a_ -. -

I -*

THE UA2 CENTRAL CALORIMETER

UA2 Collaboration (Bern-CERN-Copenhagen-Orsay-Pavia-Saclay) Presented by A. Clark CERN, Geneva, Switzerland

Summary Experimental Apparatus The UA2 experiment' was recently installed at the CERW Super Synchrotron (SPS) proton- General collider2, to study collisions at an energy de540 GeV. A major objective of this experiment is to identify Figure 1 is a plan view of the apparatus. At the the weak intermediate bosons (Z',W") via their elec- centre is the vertex detector. It consists of four tronic decay modes: multiwire proportional chambers (MWPC) with cathode strip readout, a cylindrical hodoscope, pp + zo + x0 ; z0 + e+e' and two JADE-type drift chambers' with charge division and multi-hit readout. The detector is surrounded by ? + pp * Wf + x ; W' * e-v . 1.5 radiation lengths tungsten and a fifth proportional chamber (PROPS) to provide an accurate position measure- Current theoretical models3 predict a production cross- ment of e.m. showers. This chamber allows improved had- section s 3 x 1O-33 cm2 and a leptonic fraction s 3% ron rejection, and rejection against overlap background for Z", Q 8% for W*. (a low-momentum near a IT', simulating an elec- The low expected Z", W' production rate implies tron) in the following calorimeter. the need for good electron identification and energy Covering +l rapidity units about 0, the vertex de- measurement over a maximal solid angle. The UA2 detec- tector is surrounded by electromagnetic and hadron tor is instrumented over s 80% of the solid angle by calorimeters. segmented lead/scintillator sandwich counters, provid- The forward and backward directions (20" to 37.5', ing a Z" * e+e- acceptance of Q, 63%. At luminosities 142.5O to 160') are each instrumented by twelve L 'L 1030 cm-2 s-1, s 0.15 events/h should be detected toroidal-magnet sectors (0.38 Tern) and associated spec- with a mass resolution at the Z" peak s 1.5%. trometry. Following each sector, nine drift chamber Another major objective of this experiment is to planes allow a momentum measurement on charged tracks. study high-pT hadron jets. For this reason, and to en- After this, a 6 mm lead converter and two proportional- hance electron identification, segmented iron/scintil- tube planes define the position of e.m. showers in a lator sandwich counters are installed in the central calorimeter. region. This talk describes the electromagnetic and had- ronic calorimetry in the central region of the UA2 detector.

p-6 experiment UA2

DRIFT CHAMBERS

FORWARD-BACKWARD CAKIRIMETER

Fig. 1. Plan view of the UA2 detector. '

- 169 - e.-tn. calorimeter: 17 mdiation lengths

total calorimeter: 6.5 absorption lengths

Fig. 2. Typical cell of the UA2 central calorimeter.

The 12 forward and 12 backward calorimeter sectors To monitor phototube stability, a xenon light- are each divided into ten cells (15' in +, 3.5' in 8). flasher is associated with each azimuthal slice of ten Each cell is a lead/scintillator sandwich in two longi- cells. Light (diffused and filtered to approximate the tudinal sections (24 r.1. and 6 r.1.). The light of light spectrum reaching the phototube) is passed via each section is collected in two phototubes via BBQ- plastic fibres to the light-guide of each phototube. doped wavelength-shifting light-guides. The calorimeter The net flasher stability and pulse-to-pulse variations performance is similar to that, of the central region. in light output are monitored by a box containing three vacuum photodiodes. The relative stability of different Central Calorimeter azimuthal slices is monitored by sending light from each flash-tube to a box (J-BOX) containing a scintillator The lead/scintillator electromagnetic and iron/ slab in front of three selected phototubes (XP2012). scintillator hadronic counters cover from 40' to 140° The stability of these phototubes is monitored by d.c. in polar angle, and all azimuthal angles. A spherical current measurements from 6oCo and "Sr sources. structure, the calorimeter is segmented into 240 indi- An identical but independent flasher on each slice vidual cells (towers) pointing to the i;p interaction sends light to two scintillator plates of each e.m. com- vertex. Each cell has three longitudinal sections (plus partment. The same photodiodes monitor flash stability. PROP5). In addition, the last 0.35 attenuation lengths are separately measured to provide a tag on late had- Calibration Stability ronic showers. Each cell has an e.m. length s 17.5 r.1. and a hadronic length % 4 a.1. (Fig. 2). An initial calibration of each cell was made using The light of each section is transferred via 2 mm 10 CeV/c electrons (e.m. compartment) and 10 CeV/c thick BBQ-doped light-guides to a total of seven photo- (hadronic compartments). Since installation in tubes per cell (XP2008, XP2012). The scintillator is November 1981, phototube gains have been monitored using 4 mm NE104B (e.m.), and 5 mm PMMA doped with 1% PBD, the flash-tube of each slice, normalized with respect 0.1% POPOP, and 10% naphthalene (hadron). Wavelength- to the response of: shifting techniques' minimize the dead-space between adjacent cells (Table 1). In practice the polar dead- space is negligible except for particles of normal i) vacuum photodiodes (discarded because of gain incidence. changes); ii) mean e.m. phototube response: the r.m.s. spread, Table 1. Maximum separation between calorimeter cells for individual e.m. phototubes with respect to the mean, is +2.6X; the mean hadron response is un- changed with respect to the e.m. phototubes, with Compartment Polar Azimuth an r.m.s. spread for individual tubes of 3%; (d b) iii) J-BOX response: this indicates a mean change in e.m. light response of 0.5X, with an r.m.s. spread EM 4.6 (light guide) 1 (Fe) + 1 mn (air) of 2%. A mean change of < 1% (r.m.8. of Q 0.6%) has been RADRON 1 9 (light guide) 10 (Fe) measured in e.m. response, from periodic 6oCo d.c. cur- BADRON 2 13.6 (light guide) 10 (Fe) rent measurements on each cell. In the extreme forward (proton) direction, the mean change reaches 1.7X, sug- gesting minor radiation damage. The stability of the phototubes, between their initial calibration and installation at the Fp collider, is being analysed.

- 170 - Electron Response Measurements

In addition to the calibration of each e.m. cell with 10 CeV/c electrons, data were collected between 1 and 70 GeV/c for all 5 e.m. cell types. Figure 3 illu- strates the nomenclature of this section.

Response Linearity and Resolution

Figure 4 shows the normalized light response to electrons of between 1 CeV/c and 70 GeV/c passing through the centre of an e.m. cell. Following a non- linearity correction 0.977 cl.0 + 0.01 In (E+l)), the light response can be estimated to <, *l%. Figure 5 shows the variation of U/v'?? (0.14) with electron momen- tum. The beam momentum spread (s 1%) has not been sub- tracted. In each figure, error bars represent the r.m.6. spread of alZ measurements on all cell types.

Fig. 5. Resolution of light response BBQSL, as a function of incident electron energy. (0) show data collected with tungsten converter 90 cm from the e.m. calorimeter. (0) show data collected with tungsten converter in final UA2 Fig. 3. Nomenclature used in following section. position. 0 and x are measured with respect to the centre of the cell I. BBQS = light response of small BBQ Response Variation with Position (Normal Incidence) BBQL = light response of large BBQ BBQsL = JBBQs-BBQL Spatial scans were made for several examples of each cell type. The r.m.s. spread of measurements in different examples of the same cell type is < +l%. Figure 6 shows the uncorrected response for cell 2 (the second smallest) as a function of position. Similar variations exist in other cell types. Figure 7 shows the same data after correction according as:

BBQ(corr) = BBQ(raw)*exp (corr)

corr = ( A1.x + A2.x2 + A3.x3) +(Bl + B2.x + B3.x2 +(cl + c2.x + c3.x* ;I:I 2 .

Constants Air Bi, Ci have been determined for each BBQ of the five cell types. Data from muons provide a con- sistent parametrization. For beam impacts > 5 mm from a cell interface, the r.m.s. spread of corrected light response for individual measurements of a cell type is < 1.1%. The resolution affi is unaffected. The ratio BBQR - BBQ,/BBQL provides a measure of the beam impact position in the cell (0 < 5 mm). However, because of differing light collection efficiencies for BBQS and BBQL, a variation of light response exists along each BBQR contour. For that small class of e.m. showers having no associated track or PROP5 signal, this varia- tion defines the effective resolution of the e.m. calorimeter.

Response Near a Cell Interface (Normal Incidence)'

The azimuthal separation between two cells is small, and a maximum response correction of 10% is required within -+2 mn of the cell interface. Few normal-incidence electrons pass through the BBQ (smeared Ep vertex). At Fig. 4. Deviation from linearity of the light the BBQ interface, 2, 20% of electrons are within 3 st. response BBQSL as a function of incident elec- dev. of the peak, and o/E 2 0.25/&. The light response tron energy. The superimposed curve is of remaining events is distributed bela, the peak a [l+O.Olln (E+l)]. (longitudinal escape and eerenkov light).

- 171 - The UA2 calorimeter is short (4 a.1.). so significant longitudinal escape is expected. However, h:gh-energy . 0 = 0 011 degrees hadron jets should be better contained since their be: haviour is similar to that of a single particle inter- o @ ~2601 .* acting at the calorimeter entrance. Data exist both a Q z&676 .* for single particles and for simulated jets. x@=BL64 .. Event Selection

To ensure good energy containment, cuts were ap- plied: E(hadron section Z)/E(calorimeter) < O.B,'and E(e.m.) 2 1 GeV or E(hadron section 1) 2 1.5 GeV. The resultant event efficiency is shown in Fig. 8.

Q Q Q Q 0.9 + + i I + + , OBL I I I I + -LO -20 0 20 LO x (mm)

Fig. 6. Variation of light response BBQSL with o trigger position, in cell type 2 of the e.m. calorimeter. . Single trigger I

Beam energy (Gel/) . 0 .OOll degrees 0 @ .2601 ,, Fig. 8. Efficiency of detection in hadron a@;.&676 calorimeter as a function of incident IT- energy. . 0.646L ” Applied cuts are described in text.

0.16,

09259-LO -20 0 20 LO Beam energy (GeV) x (mm) Fig. 9. Deviation from linearity of the light Fig. 7. Data of Fig. 6, after correction for response BBQSL as a function of incident TI- the position of beam impact in calorimeter. energy.

Non-Normal Beam Impact

At the pi; collider, the interaction vertex is Light Response smeared along the beam direction with u s 10 cm, and data were collected to simulate vertices in the range A fit was made to normal-incidence JET and SINGLE -20 < z < 20 cm. For IzI > 2 cm, no deterioration of data to find parameters, relating the response of each the efficiency or resolution is measured near the BBQ compartment, that optimize the response linearity interface, and no additional response corrections are (Fig. 9). The resultant resolution is shown in Fig. 10. required. Using the correction formulae above, the average response at each vertex position changes by Response Uniformity < 1.4% with respect to normal incidence. The r.m.6. spread of individual measurements is < *1X for each The above light-response analysis used available cell type. data in the median plane of each cell ($ " O), with linear attenuation corrections. The resultant response Hadron Response Measurements was uniform to < +4X for each cell type. For normal- incidence data near a BBO interface. the acceptance is Data have been collected at x- momenta of 1 to reduced for SINGLE triggers, but not for JET triggers 70 GeV/c and v+ momenta of 1 and 2 GeV/c. Preliminary (Fig. 11). The resolution is not significantly affected, results are shown for IT' data of momenta above 6 &V/c. as shown in Fig. 12.

- 172 - . . .- . *-i. -.. . . Y__/. . c -1 ..,.. .* u*-.y ., L ; 1: :. :i. 75’. 1.‘;’

0.8

e - 0.6 b 0 20 LO 60 60 100 * l Single (mm) o Jet 04 : I I I I 1 I Fig. 12. Variation of IT- resolution across 6 6 10 1 20 40 70 cell interface of central calorimeter. Beam energy (GeV)

Fig. 10. Resolution of light response BBQSL as Non-Normal Incidence a function of incident TT- energy. Data collected at an effective vertex position of k10.4 cm shows an average change in light response of < +1x. The resolution is unchanged. The variation in aperture across a BBQ interface is reduced to < f5X for t SINGLE triggers. . .-

References

. 1. UA2 Collaboration (M. Banner et al.), CERN/SPS/78-08 . (1978) and CERN/SPS/78-54 (1978). . 2. C. Rubbia, P. McIntyre and D. Cline, in Proc. Int. Neutrino Conference, Aachen, 1976 (eds. H. Faissner, I H. Reithler and P. Zerwas) (Vieweg, Braunschweig, BBQ 1977), p. 683. S. Van der Meer, CERN/SPS/423 (1978). Design study of a pp colliding beam facility, CERN/PS/AA 78-3 (1978). 3. See, for example: C. Quigg, Rev. Mod. Phys. 6, Fig. 11. Variation of 71- acceptance across 297 (1977). cell interface of the central calorimeter. 4. JADE Collaboration (W. Farr et al.), Nucl. Instrum. Methods 156, 283 (1977). 5. R.L. Ga&, Rev. Sci. Instrum. 2, 1010 (1960). W.B. Atwood, C.Y. Prescott, L.S. Rochester and B.C. Barieh, SLAC-TM-76-7 (1976).

- 173 -