Download High Mass Dilepton Pairs in Proton-Nucleus Collisions and more Thesis Particle Physics in PDF only on Docsity!
Nevis Laboratories ColumbiaPhysics DepartmentUniversity Irvington-on-Hudson New York
$ and ¢' Production in Hadron Collisions
DAVID CHU HOM
Reproduction in whole or in part is permitted for any purpose of the United States Government
Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Faculty of Pure Science, Columbia University 1977 National Science Foundation NSF PHY 76-
R-
CU-
NEVIS-
-ii-
is found to be 0.018 + 0.006 at 400 GeV and 0.014 + 0.
at 300 GeV. This result was not expected in current theories which try to understand w production as a direct res~lt of quark collisions.
TABLE OF CONTENTS
I. Introduction II. Experimental Apparatus A. Design Requirements B. Overview c. Apparatus Description III. Data Acquisition A. Fast Logic B. DC Logic c. Readout System D. On-Line Computer System E. Data Taking IV. Data Analysis and Results A. Description of Analysis B. Results c. Studies v. Summary and Conclusions
Page 1
7 8 8
14 15 17 18 19
21 25 30
A. Summary of Results 35 B. Theoretical Significance 35 VI. Acknowledgments 39 Appendix I: Lead Glass Calorimeter 41 Appendix II: Digitizers and Digitizer Readout System 46 Appendix III: Lead Glass Calibration 49 Appendix IV: Proportional Wire Chamber Reconstruction 52
I. INTRODUCTION
This thesis is on an experiment proposed to search for high mass dilepton pairs in proton nucleus collisions.
p + Be .... "'Y" + anything
Lepton pair production is a second generation experiment of a series of lepton experiments by the Columbia-Fermilab Collaboration. The first generation lepton experiment was single lepton production. 1 The motivations for this experiment are at least twofold: probe the nucleon structure and search for possible resonances. In this experiment we probe the electromagnetic structure of the nucleon by studying the continuum dilepton spectrum in an unexplored mass range. We are looking at the process shown in Fig. 1. This is analogous to deep inelastic lepton-nucleon scattering (Fig. 2), where the virtual photon is spacelike. 2 We are looking at a timelike virtual photon. In both cases a known probe, the photon, is used to study the nucleon electro- magnetic structure. The lepton-photon vertex is totally understood by QED. It is then the photon-hadron vertex we are studying. Measurements of the hadronic vertex function give us information about the nucleon structure. Another related process is hadron production in colliding e+e- beam experiments 3 (see Fig. 3). Again the photon is timelike. There are of course theoretical models of how a photon interacts with hadrons. One of the most studied of these
models is the parton model as deployed by Bjorken. 4 In this model the hadron is composed of pointlike constituents. This decomposition allows calculations to be done exactly_ (except for sums over parton distributions) as the hadronic vertex is reduced to QED. This picture of the hadron can easily be shown to predict Bjerken scaling in deep inelastic lepton nucleon scattering. That is the cross section is a function of x = q 2 / 2Mv (see Fig. 2), instead of both q 2 and v· Drell and Yan^5 showed how massive lepton pairs could be produced in hadron- hadron collisions using the parton model. Their model is shown in Fig. 4. Their differential cross section also exhibits scaling: f (T)
where f (,-)
4~a 2 /3Q 2 is the total cross section for e+e- _,^ μ+ -^ μ^.^ is
essentially a sum over parton distributions. A· l is the charge of the ith parton type. T is the scaling variable: T = Q^2 /s where Q^2 is the lepton pair mass squared and s is the pp c.m. energy squared. x.l is the parton fractional momentum. The F 2 i and F 2 i are vw 2 as measured in deep inelastic scattering. 2 The ~, v measurements for the antiparton distributions are at the moment very uncertain. 6 This makes comparison of the parton model with lepton pair production very difficult. 7 We can turn
the theory was not renormalizable. It is hoped, that at high energies and high transverse momentum (pt> 1 GeV/c), the electromagnetic and weak interactions would be competitive with the strong interactions. If that is the case there is the possibility of studying and improving our knowledge of weak and electromagnetic interactions at high energy. Using the full energy of the primary beam at Fermilab, we explore the highest accessible masses. Predictions of intermediate bosons have been made for a 1 ong t .ime.^ So f^ ar t h^ ey h^ ave not been found. Yamaguc h.15l , using eve, has been able to relate the electromagnetic process of ~ + - t^ production to the production of w-.+ A measurement of lepton pair production can be used to set limits on the W mass l'f i't i's not found. 16 A (^) prac t'ica (^1) use o f d'li ep t (^) on measuremen ts with the use of scaling is to determine the parameters of a w-^ + search experiment or of an accelerator designed to produce w-.^ +
Lee and Wick 17 have introduced a theory of QED with finite renormalization. The photon propagator is modified by the introduction of a new particle, the B^0 • B^0 , being massive, can decay into lepton pairs directly. Recently there has been a great deal of interest in gauge theories. It started when Weinberg and Salam^18 independently succeeded in unifying weak and electromagnetic interactions into one theory. The w-^ + and the photon became gauge bosons in their model. The disparity between the weak
and electromagnetic interactions was attributed^ / to a spon- taneous breaking of the gauge symmetry. The symmetry breaking gave the w± its mass. It was suggested that the theory might be renormalizable as the gauge bosons are massless at the start. In addition to the charged vector bosons w±, a neutral boson, was predicted: z^0 • The z^0 could be detected by its
R,^ + - R, decay mode.
Georgi and Glashow 19 proposed an alternative model without
neutral currents (z 0 ). Instead they predicted heavy lepton
counterparts of the electron and muon (E,^0 E,+^ M,^0 M+ ). As of
this writing neutral currents have been observed 20 , but the existence of heavy leptons is an open question. 21 Th~re are models which have both neutral bosons as well as heavy leptons.^.^22 Before the advent of renormalizable gauge theories, Bjorken and Glashow 23 motivated the introduction of a fourth "charmed" quark. Their argument was based upon the leptons being a pair of doublets and arguing for a similar symmetry for hadrons. \le c (^ u e d μ (^) s Glashow, Iliopoulos and Maiani 24 provided a better basis for such a conjecture when they found that strangeness changing weak currents were suppressed to first order in G (Fermi coupling constant) in accord with experiments.
II. EXPERIMENTAL APPARATUS A. Design Requirements Since we are looking at a process of the Drell-Yan gender, we expect a very small cross section. To be able to pick up any unusual pearls in this debris of hadronic sand, we would require a hadron pair rejection of order 10 7 • There are two approaches to this problem. The first approach is to observe electron-positron pairs. Here we must be able to distinguish between hadrons and electrons. The problem with this approach is that we would be limited in incident beam intensity by the large hadronic flux. The second approach is to absorb the hadrons before they can reach the apparatus. Here we must observe muon pairs. This approach suffers from loss of resolution due to energy loss and multiple scattering in a hadron absorber. Our decision was to perform the experiment with the first method. A followup experiment will try to increase the sensitivity in the high mass region using.^ muon pairs..^31 To observe electron-positron pairs we require a hadron rejection of 10 3 to 10 4 per arm of the spectrometer. Since there is a large flux of hadrons we must be able to take high counting rates to explore small cross sections. A problem created by high rates is accidentals. To alleviate part of this problem we must have good time resolution. Due to the beam structure (18.9 ns), a resolution of 15 ns was sufficient.
As we are one of the first experiments to look at such a high mass region, we also want to be sensitive to resonances. Resonances would appear as peaks sitting on a continuum. Very good mass resolution is required so that any peaks would not be smeared out (a= 1%). Since there is a large 7f^0 and 11 flux, care must he taken to avoid conversion of ~ rays from their decay. B. Overview The apparatus is a double arm spectrometer with each arm on the opposite side of 90° in the pp c.m. for 300 and 400 GeV proton beams. Each arm covered a solid angle of 0.06 srad-in the pp c.m. system. Each arm consisted of a 9-plane wire chamber-hodoscope magnetic spectrometer. No detectors were placed before the magnet so that we could take high beam flux. Charged particles were deflected in the vertical plane and the target is viewed in the horizontal nonbending plane. Electron identification was provided by a lead glass spectra- meter sitting behind the magnetic spectrometer. A detailed description of the apparatus follows. c. Apparatus Description Our description of the apparatus will proceed in the same order as the beam would see it. The experiment was performed in the Proton Center Laboratory of the Fermi National Accelerator Laboratory. We will describe only one arm of the spectrometer since the arms were symmetric. The two arms were identical
into the secondary beam. About 1 ft after the filters, a 2 in. tungsten collimator defined the actual vertical apertures of+ 3.5 mrad. Next came the shielding anq 7-ft collimator. The collimated beam saw only vacuum from the upstream end of the collimator until the exit of the analyzing magnet. The 10-ft long analyzing magnet was a dipole magnet with the fields in the horizontal plane perpendicular to the spectrometer axis. It sat 35 ft from the target. At a setting of 1300 A the integral field was 35 kG-meters or equivalently a pt kick of 1.05 GeV/c. The major part of the data taking was at four settings of the spectrometer: 600, 800, 1100, 1300 A. In the left arm of the spectrometer, charged particles were bent upward. This arm was referred to as the up arm (up). The right arm had charged particles bent down into it. This arm was referred to as the down arm (dn). The reason for this bending up and down was that we gained 30% in acceptance. The two magnets were always set to symmetric fields strengths although the polarities were varied. Helium bags were placed between the magnets and the first detectors to minimize conversions. The first detector station occurred after the magnet 80 ft from the target. This allowed the low momentum charged particles to be swept out. A separation between charged and neutral particles allowed us to place the detectors in the charged beam only.
The neutral beam which views the target directly consisted of a tremendous flux of ~·s and neutrons. The first counters were a set of three planes of multiwire proportional chambers (MWPC) with 2 mm wire spacing (Y 1 , P 1 , Q 1 ). One chamber had horizontal wires. The other two were 'small angle stereo
chamber (+ arctan(l/8) with respect to horizontal). Right
behind the wire chambers was a plane of 38 strips of vertical scintillation counters (denoted by v 1 ). v 1 was then followed by a plane of trigger counters called T 0. Two meters behind the 80 ft station was a set of counters used to monitor our targeting efficiency. These counters sat in the neutral beam. They should not have seen any charged particles therefore they should have had no magnet setting dependence. They did in fact have a magnet dependence of < 10%. This dependence was attributed to scattering of soft charged particles off the magnet polefaces. These counters will be denoted by NDN (or NDNBY). The so-called neutral beam blockers were between the 80 ft and 100 ft stations. Their purpose was to stop the neutral beam before it reached that part of a lead glass spectrometer which would have intercepted the neutral beam envelope. They consisted of slabs of steel backed by concrete blocks stacked outside of the aperture. A one-plane wire chamber (Y 2 ) was placed at the 100 ft station. The chamber had horizontal wires of 3 mm spacing. A plane of trigger counters called T 1 was placed immediately behind the chamber.
The other function of the lead glass spectrometer was to measure the energy of the electron. Its resolution increased with energy whereas the magnetic spectrometer resolution decreased. Thus they complemented each other very well 32 (see Fig. 8). The lead glass was 25.8 radiation lengths and 1. absorption lengths deep. Table I gives the various properties of the lead glass.
III. DATA ACQUISITION
A. Fast Logic We will follow the data acquisition process in a natural order: from triggering until it is recorded onto magnetic tape. The basic fast trigger required a track through the apparatus and an associated electromagnetic shower in the lead glass. A track was defined by a coincidence of three trigger counters: T = TO • T 1 · (^8 ) A typical T rate was about 1 MHz. The uncorrelated counting rates in the separate planes (singles rates) were typically four times higher. Note that s 2 counts single minimum ionizing particles here. A coincidence was then made with T 2 • We required that the pulse height in T 2 correspond to greater than 15 minimum ionizing particles. This signaled a shower of some sort after the first layer of lead glass. E = T • T (^2) A typical E rate was 100 kHz. E was defined as the fast trigger for a single electron. An electron-positron pair trigger was formed by the coincidence of an electron in one arm and a positron in the other: Eun = EUP • EDN (where EUP = E for up arm, EDN = E for down arm) • The Eun rate was about 1 kHz. The fast logic is schematically
