ISBN 91-628-3661-7 LUNFD6/(NFFL-7173) 1999
Potential for B-physics measurements with a fixed-target proton-collision experiment
Thesis submitted for the degree of Doctor of Philosophy in Physics
by
Jenny Ivarsson
Department of Physics Lund University Professorsgatan 1 Box 118 SE-221 00 Lund Sweden
Abstract
HERA-B is a high-energy physics experiment at HERA. It is a fixed-target experi- ment with a forward spectrometer to benefit from the strong boost of beauty hadrons. The main goal of HERA-B is to detect and measure the degree of CP violation in 0 0 exclusive B-decays, with emphasis on the golden decay, B J/ψKS. A computer simulated study on this decay is presented and the accura→cy of the measurement is estimated. Necessary conditions for a detection of CP violation are investigated. Various other channels are reviewed, which can give a sign of CP violation within and beyond the Standard Model. Very high rates are required for measurements of the typically very much suppressed signals. The thesis includes a comparison of the situation with the B-factories and the hadron colliders. HERA-B has a very tight time schedule and physics measurements will be performed even before the detector is completed. In particular, the order of installation of detector parts is determined not only by technical factors but also by the feasibility to measure the b¯b cross- section, which is a crucial parameter in the precision of any B-physics measurements at HERA-B. A thorough study has been performed on the possibilities to measure the b¯b cross-section with a partially equipped detector. The emphasis is placed on the inclusive decay B J/ψX. Studies on double semileptonic decays and Υ decays are also presented. →
Contents
Preface 3
1 B Physics and CP Violation 7 1.1 Heavy-quarksymmetries...... 8 1.2 Openheavyflavourproductionanddecay ...... 11 1.3 Heavy-quarkonium production and decay ...... 17 1.4 CPviolationintheStandardModel ...... 18 1.5 Mixingofneutralmesons ...... 22 1.6 ExperimentalobservationofCPviolation ...... 24 1.7 CPviolationbeyondtheStandardModel ...... 32
2 The HERA-B experiment 36 2.1 Requirements...... 37 2.2 VertexDetector...... 38 2.3 Trackingsystem ...... 41 2.4 Particleidentificationdetectors ...... 45 2.5 Trigger...... 48 2.6 OfflineSoftware...... 52 2.7 ComparisonwithotherB-physicsexperiments...... 63
3 Optimization for σb¯b measurements 68 3.1 Experimentalandtheoreticalpredictions...... 69 3.2 Theoptimalgeometry ...... 70 3.3 Semileptonicdecays ...... 91 3.4 Υreconstruction ...... 106
4 The golden decay 112 4.1 ExtractingaCPviolationsignature ...... 112 4.2 Thesimulatedeventsample...... 116 4.3 Triggersimulation ...... 116 4.4 Analysisandefficiencies ...... 119 4.5 Background...... 125 4.6 B flavourdetermination ...... 127 4.7 The precision of sin 2β ...... 130 iv CONTENTS
5 Summary 135
Acknowledgements 138
A Kalman filter 141
Bibliography 145 Notice the World
The sun shines on a flower stalk On the path where I do walk On the field where startled deer Leap into bush to disappear Each step the scenery’s new I could live for this and this view I could die for this and this vital fear.
Spirited songs of greyish lark Cracklings in the aged bark The humming little flies and bees Whispering wind up in the trees Rustle from leaves on the ground I could live for this and this sound I could die for this and this unified peace.
There’s a scent I can’t locate But it’s clearer near the lake Deep in the wood it’s more intense like moss that clings to a meadow fence And fir growing close by the well I could live for this and this smell I could die for this and this nuance of sense. 2 CONTENTS
The gentle wind that heats and chills Fondles over grassy hills Caresses too my cheek and hair Dabbed by a wing so weak and fair I touch the nature ceiling I could live for this and this feeling I could die for this and this varying air.
Sweet dynamics of death and birth Complexity engulfs the earth With all its souls that need refill Nature itself is rich in it still A superior beauty reigns too I could live for this and I do I could die for this and I know that I will
Jenny Ivarsson Munich 1995 Preface
Have you ever reflected on the endlessly rich complexity in nature? Have you ever walked in a bright friendly forest in early summer, when the wind rustles quietly in the leaves, little creatures run around in the grass and the sun shines on your skin and finds the way to your amygdala? How can a world which exhibits such complexity be built? If you want to describe it, paint it, make a map of it, you can spend your whole lifetime without getting all the details, because once you notice the little needle, you will soon have to accept that it contains a whole world by itself. I pondered this. In physics books, they talked about atoms, protons and even mysterious things like quarks. In biology, the science of the living things, they talked about cells, that cells build up all this. But that dead and living things must be fundamentally the same, I was certain. The burning question I had during my first years as a thinking teenager was, whether cells were made up of atoms or atoms made up of cells. But I was lucky; I asked this question in a time when the wisdom of other people who had already thought about everything was documented in book after book. There was Democritus from ancient Greece with the idea of something undividable, Rutherford who found that the atom has a nucleus, and Gell-Mann postulating the quarks as the constituents of protons and neutrons. And there were Scheiden and Schwann who in 1838 proposed that the cell was the smallest form of life, and Mendel who got the idea of genes. You just need luck to find a book where both atom and cell are mentioned in the same chapter. The cell has a kernel in which there are chromosomes made up of genes, which in turn are made up of bases, and each base consists of five atoms. This sentence gave me my second scientific revelation (the first was when I understood how the earth can be round when it looks flat). No wonder that nature is so complex when it is constructed of such small building blocks. The number of possible combinations is just incredible. Perhaps you also sometimes wonder about the vast empty sky. When you lift your eyes upwards at night, do you then behold infinity? How can the universe be infinite? And if it is not, what lies beyond? And has the universe existed for an eternity? If not, what was there before? These are a couple of very uncomfortable paradoxes. And then along comes Einstein and tells us that time is relative. This concept is very hard to accept, but if you do, you can get a relieving solution to the space–time paradox. If space is bent in an additional dimension, just like the surface 4 CONTENTS of the earth and time is our perception of the extra dimension, then the universe has no boundary but is still not infinite. In my interpretation, space and time were created from a singularity and questions like ‘before’ or ‘outside’ are irrelevant. The history of the universe can tell us a lot, if not all about the world we are living in. In the beginning, when the available space was very small, everything was very concentrated and the energy available for physics processes enormous. In particle physics we try to reconstruct those processes by building particle accelerators, to achieve high energies concentrated on small spots. Under those conditions a lot of particles are created which do not exist in our normal low temperatures. Early in history, in the Big Bang theory, there existed electrons and positrons, quarks and antiquarks in equal proportions. Conservation laws required balance be- tween them. As the universe expanded it cooled down and some processes became impossible. At a certain threshold pair production was no longer possible and an- nihilation of matter and antimatter became an irreversible process. Electrons and positrons annihilated and released energy in the form of photons. Light quarks and antiquarks annihilated after 10−4 seconds, the heavier b-quarks and anti-b quarks after 10−9 seconds. If the universe had not expanded and cooled down, the annihi- lation processes would have continued until there was only energy left, which would have led to a terribly boring world without life or complexity. In the end there was more quarks than antiquarks. How could this happen? All those conservation laws required that there should be equal concentrations of both. I read in a book that some very heavy gauge boson (a force mediator), translated more often to a quark pair than the corresponding antiboson translated to an antiquark pair, although their total decay rates were the same. When the temperature decreased so that the reverse reaction was no longer possible, all of those gauge bosons eventually decayed and we were left with more quarks than antiquarks. I thought this was ingenious, but I was still uneasy. What the author of the popular book had not bothered to tell me was that in fact, some of the symmetry laws must be broken to lead to this decay. In this way, one question brings the other and at this stage you are addicted to knowledge. Then you go to university and learn the mathematical formulas for everything and how all the fantastic world of elementary particles and forces and their interactions can be described with mathematics; and all the time new questions arise, like how the particles get their masses and what happens at even smaller distances, at even higher energies. One of the greatest achievements of the retreating century has been the math- ematical formulation of three out of the four different kinds of interactions under the same concept. Those were the electromagnetic, the weak and the strong forces. The fourth force, gravity, has no effect at energies accessible in particle physics lab- oratories of today. The collected knowledge in particle physics is described by the Standard Model. The asymmetry which is broken and gives an excess of matter over antimatter, is called CP violation. C for charge conjugation and P for parity symmetry. But the Standard Model contribution to CP violation is not enough to account for all the matter–antimatter asymmetry observed in the universe. On the other hand, the size of CP violation is not yet measured, and it will not be known whether the Standard CONTENTS 5
Model is sufficient to explain the CP violation in B decays until it is measured. Eventuallt you reach a point when the world heritage can not tell you more. Then you have to put on your working gloves and contribute yourself. This thesis describes my own attempts to bring the physics knowledge of the world a small unnoticeable step forward. I have been participating in a high-energy physics experiment at DESY, the HERA-B experiment, which uses the 920 GeV proton beam of the HERA accelerator and directs the beam halo particles onto target wires. The high luminosities lead to several interactions per proton bunch crossing and to very high rates. HERA-B is designed to perform b-quark physics, also called beauty physics. Particles with b or anti-b quark constituents are called B hadrons, or beauty hadrons. The energy of the HERA proton beam is enough to produce the heavy B hadrons. A crucial parameter for the HERA-B experiment is the rate of b and anti-b quarks that can be produced, i.e. the b¯b cross-section.The main part of this thesis describes the work and preparation studies I have made for the measurement of the b¯b cross-section at an early stage of HERA-B running and the reconstruction of the so called golden decay of neutral B hadrons. In Chapter 1 I introduce the reader to beauty physics and the role of the HERA-B experiment in this context before describing the experiment in more detail in Chapter 2. Studies on the b¯b cross-section are presented in Chapter 3. A computer simulated study on the golden decay of neutral B mesons is presented in Chapter 4. This study is made to find the possibilities of measuring CP violation in this channel in the HERA-B experiment. The CP reach in the HERA-B experiment depends on some unknown factors: The size of the CP-violating parameter sin 2β. • The b¯b cross-section at HERA-B kinematics. • The hardware and its power for event selection and background reduction. • Reconstruction efficiencies, including pattern recognition, bremsstrahlung cor- • rections and particle identification. The size of sin 2β as it is known today is presented following a study of the recon- struction power of the golden decay using detectors with resolutions and efficiencies of the final technical design. The most crucial point is the b¯b cross-section. This is the motivation to learn about the size of the b¯b cross-section and in turn, the power of the detector. The measurement of the beauty cross-section can be performed during the build- ing phase before the detector is completed. HERA-B has a tight time schedule and needs to achieve as much as possible in as short as possible a time. The computer simulations of the detector are constantly updated in parallel with the evolution of production and installation plans. A chain of scenarios eventually emerges in the real situation and meets with real data. The extensive computer simulated studies on the b¯b cross-section for different detector geometries are treated in Chapter 3. I took part in a σb¯b task force in the spring and early summer of 1997. The goal of the task force was to optimize the detector for the 1998 physics run, with the main emphasis 6 CONTENTS on one of the detector components, the outer tracker. The results were collected in the notes [1] and presented to the advisory committee to the DESY director general (Physics Research Committee) [2]. In the task force I was responsible mainly for vertex reconstruction and computing the final efficiencies. Using the tools developed for the task force, I carried out studies on semileptonic decays of B mesons, which were presented to the Physics Research Committee [3], and together with A. Somov a study on Υ reconstruction. One year later, the conditions for physics in 1998 had changed. This years study on physics options in 1998 concentrated on the calorimeter and the vertex detector. I made the studies on the σb¯b cross-section, assuming that the magnetic field would be switched off during data taking of 1998. The results were collected in notes [4] and presented to the Physics Research Committee. I also made a study on semileptonic decays with this geometry, which became a note [5]. Later I continued the study for the case of a magnetic field and for different configurations of the vertex detector. That work was done to be able to optimize the detector configuration for the early run of 1999 and get another chance of seeing a B. In 1995, before being involved in HERA-B, I was working on the ATLAS Transi- tion Radiation Tracker (TRT). I implemented a modular geometrical structure into a GEANT simulation and studied the electron and pion identifications using polypropy- lene fibre radiator. This work does not fall within the scope of this thesis. A detailed description can be found in my thesis for the degree of philosophie licentiat [6].
Jenny Ivarsson CERN, 1999 Chapter 1
B Physics and CP Violation
The HERA-B experiment will have access to the system of B hadrons. Because of the large mass of the b-quark there is a rich scope of decays to be studied in this hadron family. The mechanisms of open heavy flavour production at fixed-target experiments have been investigated in Refs.[8] and [9] and are reviewed here. When the mass of a quark Q is much larger than the QCD scale the quark is called a heavy quark and simplifications arise due to the asymptotic freedom of the strong interactions. The Non-Relativistic QCD (NRQCD) approach [10], leads to predictions of heavy-quarkonia production and decay, which can explain the data at the pp¯ colliders very well. At fixed-target experiments, where the heavy quarks are produced close to the threshold, the predicted total production cross-section is consistent with data, but there are some discrepancies in the χ1c/χ2c production ratio and the transverse polarization fraction of the J/ψ, which can not be explained by NRQCD alone. For Υ, the corresponding ratios have not yet been measured, but are expected to be more consistent with theoretical predictions. The presentation on NRQCD and comparisons with data is based on the articles in Ref. [11]. For heavy quarks (m ) there is a spin–flavour symmetry because the Comp- Q →∞ ton wavelength of the heavy quark is much smaller than the QCD scale, ΛQCD.A range of measurements to be performed in the B system will be viewed from the Heavy-Quark Effective Theory (HQET) based on the heavy-quark symmetry. A derivation and detailed review can be found in Ref. [12]. Maybe more importantly, CP violation can be studied in a variety of ways in the B system. CP violation is a small effect. It can be studied only at very high luminosities in high-energy environments, where quark–antiquark composites can be produced. The measurement of CP violation in the B system is the main goal of the HERA-B experiment. The theory of CP violation is well described in Ref. [13]. 8 CHAPTER 1. B PHYSICS AND CP VIOLATION
1.1 Heavy-quark symmetries
The mediators of the strong force, gluons, carry themselves the strong charge, colour. As a consequence, gluons couple to other gluons. The emission and absorption of virtual gluons produce an antishielding effect, so that the strength of the field is magnified at larger distances. At distances of the size of a hadron, Rhad 1fm, the quantum numbers of a quark can not be resolved. This phenomenon is∼called con- finement. The coupling constant, which describes the interaction strength, depends on the momentum transfer q:
2 12π αs(q )= , (1.1) q2 (33 2Nf )ln Λ2 − QCD where Nf is the number of interacting quark flavours. In processes with very high momentum transfers, which corresponds to interactions at short distances, the effec- 2 0 tive coupling constant is very small: αs(q ) 0asq . In other words, the constituents of the hadron behave as free par→ticles. This→∞phenomenon is known as asymptotic freedom. For sufficiently large momentum transfers, the strong interaction 2 can be described in terms of a power expansion in αs(q ) according to perturbation theory. Because of confinement, perturbation calculations will always need nonper- turbative corrections of non-free quarks and gluons from long distances, for which the coupling constant is not small. The scale which separates the regions of large and small coupling constant is ΛQCD. Strong interactions of systems containing heavy quarks allow some simplifications due to the large momentum transfers at the heavy-quark scale, mQ λQCD,and due to the small Compton wavelength of the quark compared with t≫he size of the hadron: λ R . Q ≪ had Nonrelativistic QCD
For a heavy quark, mQ ΛQCD and the effective coupling constant αs(mQ)issmall due to asymptotic freedo≫m. At length scales comparable to the Compton wavelength of the heavy quark, the strong interactions are described perturbatively. A heavy- quarkonium system (QQ¯)isofsizeλ /α (m ) R .The velocity of the heavy Q s Q ≪ had quarks in the rest frame of the quarkonium, vQ, is small. Hence heavy quarkonium can be treated as a nonrelativistic bound state. In order to exploit the smallness of αs(mQ) and perform perturbation calculations, the scale mQ must be separated from 2 the smaller momentum scales, mQvQ, mQvQ and ΛQCD, that involve nonperturbative physics. mQvQ is the three-momentum of the constituents in the quarkonium rest 2 frame. mQvQ is the kinetic energy of the heavy quark and the scale of binding energies. The Lagrangian for NRQCD is
= + + δ (v2)+O(v4 ) . (1.2) LNRQCD Llight Lheavy L Q
light is the normal relativistic QCD Lagrangian for light constituents. heavy is the nLon-relativistic approximation of the Lagrangian for a heavy quark andLis invariant 1.1. HEAVY-QUARK SYMMETRIES 9 under heavy-quark spin symmetry. δ includes the relativistic correction and breaks the spin symmetry. The size of the Lhigher order terms depends on the momentum scale, mQ, and can be calculated using perturbation series in αs(mQ). The differential cross-section formula for producing a quarkonium state, H,with four-momentum p~ is factorized as
H dσ(H(p~)) = dˆσQQ¯[n,p~] On , (1.3) n X where the sum runs over all colour and angular-momentum states of the QQ¯ pair. The short-distance factor, σˆ, describes the cross-section to produce the QQ¯ pair. It involves momentum scales of order mQ and larger, and can be calculated perturba- H tively. The long-distance factor, On is proportional to the probability that the QQ¯ pair will form the quarkonium state H plus soft hadrons, whose energies in the
H rest frame are of order mQvQ or smaller. In the non-relativistic approximation the last factor is zero for all states except when the produced pair is already in the colour singlet state H.
Heavy quark effective theory The typical momenta exchanged between heavy and light constituents in a hadron are of order ΛQCD and described nonperturbatively. Distances which can be resolved are of order Rhad. However, the Compton wavelength of the heavy quark is much smaller than that, which means that quantum numbers like flavour and spin of the heavy quark can not be resolved. Only the colour field extends over large distances because of confinement. Relativistic effects vanish as m . Such effects are colour magnetism and Q →∞ spin. The quark is at rest with respect to the hadron as mQ mhad. The effective result is that the heavy quark acts as a static source of colour.→This implies relations ∗ ∗ of heavy mesons like B, D, B , D or heavy baryons like Λb and Λc. The descriptions of these hadrons differ only due to nonperturbative corrections in powers of 1/mQ and perturbative corrections in powers of αs(mQ). For Nh heavy-quark flavours, there is an SU(2Nh) spin–flavour symmetry group, under which the effective strong interactions are invariant. The aim is to rewrite the Lagrangian for a heavy quark, = Q¯(iD m )Q, (1.4) LQ 6 − Q into an expansion of 1/mQ in order to construct a low-energy effective theory, which is called the Heavy-Quark Effective Theory (HQET). A heavy quark has more or less the velocity of the hadron, v, and the momentum can be rewritten as µ µ µ PQ = mQv + k , (1.5) where the residual momentum k is small and is affected by interactions with light degrees of freedom like soft gluons. The quark field can be written as a sum of the large and small component fields hv and Hv:
imQvx imQvx hv(x)=e P+Q(x),Hv(x)=e P−Q(x), (1.6) 10 CHAPTER 1. B PHYSICS AND CP VIOLATION
where P+ and P− are the projection operators: 1 v P = ±6 . (1.7) ± 2
The operator hv annihilates a heavy quark with velocity v. It describes the massless degrees of freedom. Hv creates a heavy antiquark. It corresponds to fluctuations with twice the heavy quark mass, which are the heavy degrees of freedom appearing as corrections to the effective Lagrangian. The derivation of the Lagrangian is well describedinRef.[12].
¯ 1 ¯ 2 gs ¯ µν 2 eff = hviv Dhv + hv(iD⊥) hv + hvσµν G hv + O(1/mQ), (1.8) L · 2mQ 4mQ µ ν µν where [iD ,iD ]=igsG is the gluon field-strength tensor. In the limit mQ , only the first term remains. The second term arises from off-shell residual m→∞otion of the heavy quark. The third term describes the colour-magnetic coupling of the heavy quark spin to the gluon field. Both these terms scale like 1/mQ. This effective Lagrangian reproduces all physics at long distances. It does not account for interactions with hard gluons with virtual momenta of order mQ. Short- distance corrections are calculated in perturbation theory in powers of the running coupling constant, αs(mQ), and thus induce a logarithmic dependence on the heavy- quark mass. One way to test and use HQET is in spectroscopy. The mass of the hadron should be the mass of the heavy quark plus a term which is independent of the flavour (up to corrections). Because of heavy-quark symmetries, hadronic states HQ are classified by quantum numbers of the light degrees of freedom:
2 ¯ ∆m 2 mH = mQ + λlight degrees + + O(1/mQ) . (1.9) 2mQ This predicts that the SU(3) mass splittings should be similar for the B and D systems: m m = λ¯ λ¯ + O(1/m ) Bs − Bd s − d b m m = λ¯ λ¯ + O(1/m ) (1.10) Ds − Dd s − d c The prediction is confirmed experimentally, since m m =(90 3) MeV and Bs − Bd ± mDs mDd =(99 1) [58]. The mass of the P-wave beauty mesons is needed in studies− on same side±tagging (Chapter 4, Section 4.6). The mass has been estimated in Ref. [54] using the measured masses of charmed P-wave mesons. Spin corrections are included in ∆m2 above. They depend only on the spin of the heavy quark and the light degrees of freedom and are the same for D and B as well as for D∗∗ and ∗∗ 2 B . The difference from the ground-state mesons should be equal up to order 1/mQ. In this way, the mass of B∗∗ was estimated to be around 5.8 GeV (or less depending 2 on the size of the O(1/mQ) corrections). The semileptonic branching ratios and the average number of charm hadrons per B decay (known as charm counting) are other challenges for HQET, as well as lifetime predictions. 1.2. OPEN HEAVY FLAVOUR PRODUCTION AND DECAY 11
b b n (ud) n (ud) d d
u u p (ud) p (ud) - b- b
b- n p, ... d u- u p (ud) b
Figure 1.1: Lowest order contributions to open heavy flavour production. Upper diagrams: gluon–gluon fusion. Lower diagram: quark–antiquark annihilation.
1.2 Open heavy flavour production and decay Heavy-quark production Heavy quarks are predominantly produced in the hard collision of one light parton from each colliding hadron. The two basic processes for heavy-quark production are gluon–gluon fusion (gg QQ¯) and quark–antiquark annihilation (qq¯ QQ¯). In perturbative QCD→the heavy-quark cross-section of two partons (→i, j)isaprod- uct of the scale-dependent parton distribution functions (PDF) and a finite short- distance cross-section, σˆ. The total cross-section is a sum over all parton types and an integration over the kinematic region:
S 2 2 σQQ¯ = dx1dx2fi(x1,µF )fj(x2,µF )ˆσ(x1,x2,µF ,µR) . (1.11) 2 i,j x1x2S=4mQ X Z This factorized formula is valid for quarks with large masses compared to the hadron 12 CHAPTER 1. B PHYSICS AND CP VIOLATION
masses. The corrections are suppressed by powers of mQ. The PDF, fi and fj are evaluated at the momentum fraction x and the factor- ization scale µF chosen at the heavy-quark mass. The short-distance cross-section is calculated as a perturbation series in orders of the strong coupling constant, αs(µR). The renormalization scale µR is also chosen at the heavy-quark mass, since sensi- tivities to lower momentum scales are included in the PDF. The leading order (LO) 2 diagrams in Fig. 1.1 are of order αs. The cross-section has been calculated up to or- 3 der αs (NLO), which includes all virtual corrections as well as diagrams with initial- or final-state gluon radiation. The uncertainties in the perturbative QCD description of the heavy-quark pro- duction include the following.
The heavy-quark mass, m . • Q The size of higher-order corrections. This uncertainty is normally estimated by • varying µF and µR between m/2and2m.
The value of the QCD scale Λ which is strongly correlated to the shape of the • parton densities.
Nonperturbative (long-distance) effects. • It is difficult to measure the cc¯ and b¯b differential cross-sections because, in the confinement process, the heavy quarks lose some of their energy and the one-to- one correspondence between the heavy-quark pair and the heavy hadron is lost. In experiments of limited phase space, measurements of the total cross-section depend on extrapolation using fragmentation functions and are thus very uncertain. The energy dependence of the total b¯b cross-section at fixed-target experiments is displayed in Fig. 1.2. The data has a tendency to fall somewhat above the central theoretical values. This can be explained by nonperturbative effects, which have been found to be size- able when the b¯b pair is produced near threshold, as in HERA-B and the presented fixed-target experiments. The b¯b cross-section at HERA-B energies is discussed in more detail in Chapter 3.
B decays B mesons primarily decay via the spectator diagrams in Fig. 1.3. The spectator diagram in Fig. 1.3b is suppressed by colour conservation, which requires gluon exchanges between the initial and final states. The b u transition is Cabbibo suppressed. Charmless decays are therefore rare, with branc→ hing ratios of the order 10−5. Examples of charmless decays are B π+π− and B D X. ∼ → → s The semileptonic decays are of particular interest, since they involve fewer strong- interaction effects and thus fewer hadronic uncertainties. In the heavy-quark limit, there is only one form-factor, which depends only on the momentum transfer to the 1.2. OPEN HEAVY FLAVOUR PRODUCTION AND DECAY 13
1000 bb Production
NLO QCD: πN bb + X 100 pN bb + X
10 / ( nb/nucleon ) σ
1 πN Data: pN Data: E653 E771 E672/E706 E789 WA92 NA10 0.1 200 400 600 800 1000 Beam Momentum / GeV
Figure 1.2: Theoretical calculations at NLO from Ref. [8] of the b¯b cross-section compared with fixed-target data. The dashed curves are the calculated πN cross- section and the solid curves the calculated pN cross-section. The lower limits, central values and upper limits are represented. 14 CHAPTER 1. B PHYSICS AND CP VIOLATION
e µ u c 33xx νe νµ ds b c W u W b c, u d u, d u, d u u a) b)
Figure 1.3: a) External spectator diagram for the decay of B mesons. b) Colour- suppressed spectator diagram.
ℓν¯ system. The rate of the semileptonic decay is described by the hadronic form factor and the amplitude of the electroweak quark transition:
dΓ(B¯ Dℓν¯) → = V 2 2(q2) . (1.12) dq2 | cb| F
Exclusive semileptonic decays of B mesons are studied to extract values of the el- ements Vcb and Vub of the CKM matrix, which describes the weak transitions between| the| quark|s (1|.20). The heavy-quark symmetry implies relations between form-factors of different heavy mesons, because in the limit mQ the form- factor can only depend on the Lorentz boost γ = v v′. Matrix eleme→∞nts of scattering of heavy quarks or decays to other heavy quarks are· all determined by the Isgur-Wise function ξ(v v′),ξ(1) = 1. For example, for the decay of a B¯ meson into a D meson: ·
1 ′ µ ′ ′ µ D(v ) c¯v′ γ bv B¯(v) = ξ(v v )(v + v ) . (1.13) √mB mD | | ·
Additional spin rotation factors appear in the relation between vector mesons and n pseudoscalar mesons. Nonperturbative power corrections (ΛQCD/mQ) are added. n Perturbative corrections of order αs (mQ) are also applied to account for short dis- tance interactions, which do resolve the spin and flavour of the heavy quark. The combined result is that ξ(v v′) is replaced by · (v v′)= (1)[1 ̺ˆ2(v v′ 1) + ...]. (1.14) F · F − · − (1) is calculated using HQET and including power corrections. The slope ̺ˆ2 can be mFeasured experimentally and extrapolations to zero recoil can be made (see Fig. 1.4). In such a way V can be extracted from the semileptonic decay | cb| dΓ(B¯ D∗ℓν¯) → = k 2(v v′) V 2, (1.15) d(v v′) F · | cb| · where k is a kinematic constant. 1.2. OPEN HEAVY FLAVOUR PRODUCTION AND DECAY 15
0.04 ALEPH 0.03
)|Vcb| 0.02 . v v'
F( 0.01
0 1 1.1 1.2 1.3 1.4 1.5 v .v'
′ ′ 2 Figure 1.4: (v v ) Vcb as a function of v v . (1) Vcb corresponds to the intercept of the straightF -lin· e fi| t. The| example is from· theF ALEPH| | Collaboration [14].
Vcb can also be determined from inclusive B decays. HQET applies to the theo|retica| l expressions of the decay widths with the nonperturbative corrections in 2 powers of (ΛQCD/mB) . The determination of Vub depends on form-factors for heavy- to light-meson transitions, where the heavy-qu| | ark symmetry does not help. The ratio V / V can be determined from the inclusive semileptonic decay: | ub| | cb| dΓ(B¯ Xeν¯)= V 2dΓ(ˆ B¯ X eν¯)+ V 2dΓ(ˆ B¯ X eν¯), (1.16) → | cb| → c | ub| → u where dΓ(ˆ B¯ Xqeν¯) denotes the contribution to the total semileptonic rate from the part of the→ weak current where a bottom quark couples to the quark q = c or u. The contribution from dΓ(ˆ B¯ X eν¯) can be isolated by examining the electron → u spectrum dΓ/dEe in the end-point region near the maximum allowed electron energy, since only b u transitions can contribute for E > (m2 m2 )/(2m )[15]. → e B − D B
Bc mesons Among the heavy-quark states, the (¯bc) system takes a particular place. Since top quarks are too heavy to form a stable state, (¯bc) is the only system, composed of two heavy quarks, which can neither decay strongly nor electromagnetically. The study of Bc mesons can increase the understanding of QCD dynamics and important parameters of the electroweak theory. 16 CHAPTER 1. B PHYSICS AND CP VIOLATION
+ Figure 1.5: Some typical Feynman diagrams for the subprocess g + g Bc + b + c¯. The fragmentation-type diagrams a) and b) have cc¯ pairs created on→ the leg of the ¯b. In the recombination-type diagram c) the quark–antiquark pairs are created independently.
The question has been raised whether there is a possibility to observe Bc mesons in HERA-B. The predictions on the total cross-section of Bc production at HERA-B do not give precise information except that it is small [39]. The production of Bc mesons is suppressed relative to other beauty hadrons, because of the hard production of an additional c-quark pair. At HERA-B there is a strong threshold effect because additional pairs of heavy B and D mesons have to be produced. There is an additional suppression in the small probability of the (¯bc) quarkonium state formation. The two first diagrams in Fig. 1.5 describe b¯b production and subsequent frag- mentation of a ¯b-quark. The fragmentation of a c-jet contributes too, but it is less important than that of a b-jet. The contribution of recombination diagrams of the type depicted in Fig. 1.5c is not negligible at pT values below 10 GeV. According to Ref. [39], σ =0.07 0.08 pb. Bc − On the other hand Ref. [40] presents
σ =0.002 0.09 pb. Bc − The interaction frequency in HERA-B could be as much as 50–60 MHz. In a nominal 7 year of 10 seconds about 100–2300 Bc mesons would then be produced in HERA-B. The Bc mesons can decay via three main processes as is shown in Fig. 1.6, leading to a considerable number of different decay products. The HERA-B trigger is a dedicated J/ψ trigger. Therefore it is advantageous to search for Bc in decays to J/ψ. The inclusive decay rate to J/ψ is 20%. The decay mode Bc J/ψπ deserves special attention because all of the three final-state particles can be w→ell reconstructed and allow for a mass reconstruction of the Bc meson. They also have a common secondary vertex, which makes the topology very clean. The invariant mass of the lepton pair from the J/ψ supplies an additional constraint on the secondary vertex. 1.3. HEAVY-QUARKONIUM PRODUCTION AND DECAY 17
u, c, l b d, s,ν b b u, c, l c d, s W W ν W d, s, ν b c c d, s, c c u, l a) b) c)
Figure 1.6: The decay of Bc mesons. a) and b) are the semileptonic and nonleptonic spectator decays and c) illustrates nonleptonic weak annihilation.
The branching ratio of this channel is estimated in [39]:
BR(B+ J/ψπ+) 0.2%. c → ≈ Including the branching ratio of J/ψ ℓ+ℓ−, it follows that a maximum of only → one Bc signal event can be expected every second year in HERA-B. This is of course not possible to detect. There are many other channels, which could improve the statistics. The decay mode Bc J/ψℓν has an expected branching ratio of 3% leading to a maximum of seven e→vents per year, which is obviously also too low≈for an observation.
1.3 Heavy-quarkonium production and decay
The NRQCD factorization approach to the production of heavy quarkonium, J/ψ and Υ and higher states describes well the existing data. The quarkonium states can be produced by gluon fusion (Fig. 1.7) or fragmenta- tion of a single parton (Fig. 1.8). The fragmenting parton is produced in a two-to- 2 one process of order αs. The fragmentation contribution dominates at high pT.The cross-section is a product of the heavy-quark pair production at short distance and the formation of the bound quarkonium state at long distance. Concentration is here laid on the J/ψ production, but the description can be generalized to ψ(nS)andΥ(nS) states. In the NRQCD factorization approach, the nonrelativistic Lagrangian heavy L 3 for the heavy-quark system corresponds to the formation of the J/ψ state, S1, through the production of a cc¯ singlet state with the same quantum numbers, de- 3 noted (cc¯1) S1. This type of production is illustrated in Figs. 1.7a and 1.8a. If the velocity of the quarks, vQ, is sufficiently large, a quark–antiquark pair in a colour-octet state can evolve into a singlet state via electric dipole transition (soft ′ 3 gluon emission). For the J/ψ and J/ψ states, S1, the cross-section is dominated 3 by the production of the colour-octet state (cc¯8) S1 in Fig. 1.8b. This diagram is suppressed at long distance by a factor v4 relative to the colour-singlet-state diagrams Figs. 1.7a and 1.8a. At short distance, on the other hand, the operator for the 18 CHAPTER 1. B PHYSICS AND CP VIOLATION
J/ψ, χ , χ χ 0c χ χ χ χ 1c, 2c 0c, 2c 0c, 2c
a) b) c)
Figure 1.7: Schematic diagrams for the production of J/ψ and Υ in gluon fusion. a) 3 2 and b): Leading-order diagrams of order αs and αs respectively. c): An example of 4 a virtual diagram of order αs. The lowest states can be produced directly as in a) or via the χ1 state produced in a) or the χ0 and χ2 states produced in a), b) and c).
3 2 (cc¯8) S1 state is only of order αs at high pT and αs at low pT, where the production 3 is dominated by gluon fusion. This should be compared with the order αs for the 3 (cc1) S1 state. At high pT the domination of the colour-octet production becomes even more pronounced as the cc¯ pair produced in Fig. 1.8b carries all the momentum of the fragmenting gluon. 3 Also for the χ1c state, P1, the colour-octet production is expected to dominate 3 because of the suppression in αs of (cc¯1) P1 production. Also for the other χic states, the colour-octet production is important. The leading contributions from the 3 nonrelativistic Lagrangian in the NRQCD factorization approach are (cc¯1) Pi and 3 (cc¯8) S1. At fixed target, even after including the colour octet contribution, the predicted ratio χ1c/χ2c is almost one order of magnitude too low, compared to data. Another discrepancy at fixed-target energies, is the fraction of transversely polar- ized J/ψ, which is predicted to be an order of magnitude larger than the measured fraction. The reason could be not yet understood interactions of the colour-octet pair with soft gluons as it traverses the target in combination with spin symmetry breaking. The total production of J/ψ and Υ includes decays of the higher quarkonia states. In summary, the total production of J/ψ and Υ is dominated at high pT by the process gg g, where the final-state gluon fragments into a heavy-quark pair in a colour-octet→state; g (QQ¯ ). At low p , the process gg (QQ¯ ) dominates. → 8 T → 8 The decays of quarkonia can be formulated using the same nonrelativistic QCD. The two principal decay diagrams are the electromagnetic decay into a lepton pair (Fig. 1.9a) and the hadronic decays (Fig. 1.9b).
1.4 CP violation in the Standard Model
CP is the combined effect of charge conjugation (C) and parity transformation (P). Charge conjugation flips the signs of internal charges, such as the electric charge, the 1.4. CP VIOLATION IN THE STANDARD MODEL 19
ψ′ ψ′ ψ′
−(1) − − QQ QQ(8) QQ(8)
a) b) c)
Figure 1.8: Parton fragmentation into J/ψ and Υ via colour-singlet and colour- octet states: a) leading-order gluon fragmentation via a colour-singlet QQ¯ state; b) Leading-order gluon fragmentation via a colour-octet QQ¯ state; c) example of a real 2 fragmentation diagram at order αs.
u c e,++ µ d c γ* d c d d c e, µ u a) b)
Figure 1.9: a) Electromagnetic decay of quarkonium; b) hadronic decay of quarko- nium. 20 CHAPTER 1. B PHYSICS AND CP VIOLATION baryon number and lepton numbers. Under parity operation the space coordinates are reversed so that for example the direction of motion is reversed. The spin is not affected. Parity and charge conjugation are violated by weak interactions. These circum- stances are described with the concept of helicity, ~s p~ h = · , (1.17) ~s p~ | || | which describes the direction of the spin relative to the direction of motion. Particles of positive helicity are called right-handed and those of negative helicity left-handed. Experiments show that only left-handed fermions and right-handed antifermions par- ticipate in weak interactions. Whereas P and C are not conserved in weak interactions, the combined operation CP was long thought to be a good symmetry in all types of interactions. Violation of CP symmetry was first discovered in the long-lived neutral K meson. The effect of CP violation is expected to be larger in the system of B mesons. Several independent measurements are needed to determine the nature of CP violation and the access to the B system offers a wide range of decays with different quark transitions. All field theories are automatically invariant under the succession of C, P and T (time reversal) operations because of the requirement that the Lagrangian be hermitian and invariant under Lorentz transformations. One consequence of this CPT theorem is that particles and antiparticles have the same masses and lifetimes. The consequence of CP violation or equally T violation is in very generalized words the sense of a direction of time in the evolution of history. It is believed that, in the early stages of our universe, CP violation resulted in an excess of matter over antimatter, which lead to the observed baryogenesis. The Standard Model prediction is too small to account for the observed baryon asymmetry in the universe. Therefore, measurements of CP violation are of high interest. Owing to its smallness, CP violation is one of the least tested properties of the Standard Model. Exploration of the B system gives access to new sources of CP violation, which will lead to more precise measurements of the CP-violating parameter in the Standard Model. It might also indicate scenarios for CP violation beyond the Standard Model.
Theoretical description Theories within the Standard Model include CP violation in both strong interactions (QCD) and electroweak interactions. If there is CP violation in strong interactions it has been shown experimentally to be very small (see, for example, Refs. [13] and [16] ). In the Electroweak Model the CP violation is incorporated in the CKM (Cab- bibo, Kobayashi, Maskawa) matrix which describes the weak interaction between the quarks, i.e. the flavour-changing charged current interactions:
d′ g L ′ ′ ¯′ µ ′ + int = (¯uL, c¯L, tL)γ sL Wµ + h.c. (1.18) L −√2 ′ bL 1.4. CP VIOLATION IN THE STANDARD MODEL 21
′ Here qL are the left-handed quark doublets. Because of interaction with the Higgs doublet, these are nonphysical fields, having a non-diagonal mass matrix. The CKM matrix appears when the Lagrangian is written in terms of the mass eigenstates:
dL g µ + int = (¯uL, c¯L, t¯L)γ VCKM sL W + h.c. , (1.19) L −√2 µ bL where Vud Vus Vub V = V V V (1.20) CKM cd cs cb Vtd Vts Vtb is the unitary CKM matrix. It has nine parameters (the square of the dimension). There are three Euler angles in three dimensions: These are the quark mixing angles. The rest of the parameters are phases. The phases of all the quark fields are arbitrary, unmeasurable quantities. Under i(φ(k)−φ(j)) phase transformation of the quark fields, Vjk is replaced by e Vjk. There are accordingly five unmeasurable phases and only one measurable phase, which is the origin of CP violation in the Standard Model. The CKM matrix is often written in the Wolfenstein parametrization, which is an expansion in powers of λ = Vus =cosθC ,whereθC is the Cabbibo angle known from the mixing of two generatio| ns.|
λ2 3 1 2 λAλ(ρ iη) − λ2 − 4 VCKM λ 1 Aλ2 + O(λ ). (1.21) ≃ 2 Aλ3(1− ρ iη) −Aλ2 1 − − − A, ρ and η are real numbers of order unity as indicated by experiments. V and V | ud| | us| are known to better than 1% accuracy. Vcd and Vcs are known to 10–20%. Vcb is known to 5% accuracy. Hence both λ an| d A| are w| ell|determined experimenta|lly.|
V λ =0.2205 0.0018,A= cb =0.81 0.04. (1.22) ± V 2 ± us
Vub and Vtd are known to 30% accuracy, implying a high uncertainty in ρ and η. | | | | 0 0 Vtd is obtained from B B¯ mixing and Vub from charmless decays of B mesons. The| | phase is accessible as−described below.| | The unitarity condition of the CKM matrix yields the following relations:
V V ∗ =0 (j = k). (1.23) ij ik 6 Among these, the relation
∗ ∗ ∗ VudVub + VcdVcb + VtdVtb = 0 (1.24) 22 CHAPTER 1. B PHYSICS AND CP VIOLATION
b t d b d
± ± B0 B0 B0 t t B0
d t b d b
Figure 1.10: B0 B¯0 mixing. − deserves special attention, since the phases between Vub,Vcd and Vtd are large and can be detected in beauty decays. The condition can be presented as a triangle. 2 2 Below, the sides are rescaled by V V ∗ and ρ¯ = 1 λ ρ, η¯ = 1 λ η. cd cb − 2 − 2 (¯ρ, η¯) η α Vtd λ Vcb
γ β 00.20.40.60.81.0 ρ
The lengths of the two sides are
1 V 1 V R ρ¯2 +¯η2 = ub and R (1 ρ¯)2 +¯η2 = td (1.25) u ≡ λ V t ≡ − λ V cb cb p p and the three angles are V V ∗ V V ∗ V V ∗ α = arg td tb ,β= arg cd cb ,γ= arg ud ub . (1.26) −V V ∗ − V V ∗ − V V ∗ ud ub td tb cd cb The angles are physical quantities and can be measured independently by CP asym- metries in B decays. If CP were conserved, the quark mixing matrix would be real and the triangle collapse to a line.
1.5 Mixing of neutral mesons
Neutral mesons mix via an intermediate state. The description for B0 and B¯0 mesons 0 0 given here can be generalised to other neutral mesons like Bs, K and D and their antistates. Any arbitrary state is a superposition of the flavour eigenstates, which 1.5. MIXING OF NEUTRAL MESONS 23
obeys the time-dependent Schr¨odinger equation d a a i a i = H = M Γ . (1.27) dt b b − 2 b The eigenvectors are the mass eigenstates B = p B q B¯0 , | H i | 0i− (1.28)
B = p B + q B¯0 ; p2 + q2 =1. | Li | 0i | | | |
Solving this for the CP eigenstates g ives 0 1 B (0) = ( BH + BL ) , 2p | i | i (1.29) B¯0(0) = 1 ( B B ) . 2q H L | i−| i
The time evolution of the ma ss eigenstates is given by (1.27)
1 −iMH t − ΓH t BH (t) =e e 2 BH , | i −iM t − 1 Γ t | i (1.30) B (t) =e L e 2 L B , | L i | Li i where Mi 2 Γi are the eigenvalues. The width difference between the physical states is negligible− compared with the mass difference: ∆Γ | B | < 10−2. (1.31) ∆mB
From this relation it follows that one can set ΓH =ΓL(= ΓB), while the masses differ 1 1 with a magnitude ∆mB,sothatMH = mB + 2 ∆mB and ML = mB 2 ∆mB .With this notation, −
0 1 −im t − 1 Γ t − i ∆m t i ∆m t B (t) = e B e 2 B e 2 B B +e2 B B 2p | H i | Li 1 (1.32) −imB t − ΓB t 1 0 q 1 0 =e e 2 cosh ∆mB t B + i sin ∆miBt B¯ . 2 p 2 h i The time evolution of B¯0 is determined in a similar way. Studying CP violation, the ratio q/p is of interest. It depends on the off diagonal elements, M12 and Γ12: q M ∗ i Γ∗ = 12 − 2 12 . (1.33) p M i Γ B 12 − 2 12 Using (1.31) it follows that Γ M and the ratio can be approximated as | 12|≪| 12| q M ∗ 1 Γ 12 1 Im 12 . p ≃−M − 2 M B | 12| 12 To order 10−2 the rate of B0 B¯0 mixing is just a phase, i.e − q 1. (1.34) p ≃
24 CHAPTER 1. B PHYSICS AND CP VIOLATION
∗ The vertices in Fig. 1.10 are proportional to Vtb and Vtd. Therefore q/p can be expressed in terms of the CKM parameters. In the B0 system
q M ∗ (V ∗V )2 V ∗V 12 = tb td = tb td =e−2iβ , (1.35) p ≃− M V ∗V 2 V V ∗ B | 12| | tb td| tb td where β is one of the angles in the famous unitary triangle. In the approximation of quark–hadron duality, the box diagram contributions of Fig. 1.10 are responsible for the nondiagonal element M12 of the mass matrix. The contributions of the light charm and up quarks can be neglected. At high loop momentum, k Λ , which is the case in the B and B systems, the box diagram ≫ QCD s is a good approximation to the Standard Model contribution to M12. The matrix element is roughly proportional to the masses of the two internal quark lines, so that contributions from box diagrams with up or charm quarks are negligible. In lighter systems (K or D), the hadronic uncertainties are large due to contributions from light intermediate states (long-distance contributions). In the Standard Model
G2 ∆m = F η m m2 (B f 2 )S (x ) V V ∗ 2, (1.36) B 6π2 B B W B B 0 t | tb tq| where GF is the Fermi constant, ηB is a QCD correction factor calculated to NLO, fB is the B-decay constant, BB parametrizes the value of the hadronic matrix element, S0(xt) is the electroweak contribution of the top quark and xt is the ratio of the top mass and the W mass, mW .Theq in Vtq stands for the light-quark content in the mixing B (d or s). The world average of the mixing frequency is [18]
∆m =0.480 0.016ps−1 (1.37) B ± ∗ From a measurement of ∆mBd the side VtbVtd of the unitary triangle can be determined. The largest uncertainty comes|from t|he matrix element of the four- quark operator between the meson states, fB√BB. The uncertainty is smaller in the ratio between Bd and Bs mixing:
B f 2 2 ∆mBd mBd Bd Bd Vtd = 2 . (1.38) ∆mB mB BB f Vts s s s Bs
In the Standard Model, Γ12 is related to box diagram s with internal u and c quarks. The width difference of B0 and B¯0 is proportional to (V V ∗ )2 (Aλ3)2 and cb cd ∼ may be too small to be detected. The width difference of Bs and B¯s is proportional to (V V ∗ )2 (Aλ2)2, which might be large enough. cb cs ∼ 1.6 Experimental observation of CP violation
Apart from a recent measurement of sin 2β by CDF [17], the only unambiguous measurement of CP violation is in K decays. As explained above, the effect of CP violation is expected to be larger in the B system. The methods of extracting 1.6. EXPERIMENTAL OBSERVATION OF CP VIOLATION 25 the value of the CP-violating parameter in the electroweak model from decays of B mesons will be discussed in the remaining part of this chapter. Three manifestations of CP violation in pseudoscalar mesons can be distinguished: direct CP violation, indirect CP violation and the interference between the two. Direct CP-violation results from interference of decay amplitudes in the weak de- cays. The CP operation is a charge conjugation combined with parity transformation. For pseudoscalar mesons, this means
CP P (p~) =eiϕP P¯( p~) . (1.39) | i | − i P and P¯ are CP-conjugated states. The phase is unmeasurable and arbitrary. Consider the decays of CP-conjugated pseudoscalar meson states (e.g. B+ and B− or the neutral B mesons) to CP-conjugated final states f and f¯. The CP conjugated amplitudes are A = f P = A eiδk eiφk , h |H| i k Xk A¯ = f¯ P¯ =ei(ϕP −ϕf ) A eiδk e−iφk . (1.40) |H| k k X The sum runs over all possible decay diagrams. δi appear in scatterings due to strong interactions. φi are the weak phases that violate CP. The physically meaningful quantity is the absolute value of the ratio between the amplitudes: A¯ =1 Direct CP violation. (1.41) A 6 ⇒
To avoid mixing, it is b est t o observe direct CP violation in the decays of charged mesons. The measurable CP asymmetry is defined as 2 Γ(P + f) Γ(P − f¯) 1 A/¯ A a = → − → = − . (1.42) f + − ¯ 2 Γ(P f)+Γ(P f) 1+ A/¯ A → → If there is only one partial decay amplitude, the pha se is unmeasurable. Inter- ference between at least two diagrams is required. The candidates are found among those nonleptonic decays which receive contributions of the same order from ‘tree’ and ‘penguin’ diagrams, or decays with contributions from penguin diagrams only. The ± ± 0 ∗ −3 contribution of the tree diagram of the decay B K ρ is small, VubVus 10 , which is the same order as the penguin diagram. →Examples of tree-f|orbidden|∼decays are shown in Fig. 1.11. The hadronic uncertainties in direct CP violation are large because of poorly known hadronic matrix elements and strong phase shifts. Direct CP violation has not yet been measured experimentally. Indirect CP violation arises in the mixing of neutral mesons. Because of mixing, P 0 and P¯0 form the mass eigenstates 0 0 ¯0 P1 = p P q P , | i | i− | i (1.43) P 0 = p P 0 + q P¯0 ; p2 + q2 =1. | 2 i | i | i | | | | 26 CHAPTER 1. B PHYSICS AND CP VIOLATION
γ b s, d φ, K 0 q s- b s, d B- s - q * - ρ - - B K , K - -u u
Figure 1.11: Penguin diagrams for some tree-forbidden B decays.
If p = q =1/√2then
CP P 0 = P 0 ,CPP 0 = P 0 , (1.44) | 1 i | 1 i | 2 i −| 2 i which is equal to the CP eigenstates. The physically meaningful probe is
q =1 Indirect CP violation. (1.45) p 6 ⇒
Indirect CP violation ca n b e searched for in semileptonic decays. Decays which are only allowed in the presence of mixing are studied:
Γ(P¯0 ℓ+νX¯ ) Γ(P 0 ℓ−νX) 1 q/p 4 aSL = → − → = −| | (1.46) Γ(P¯0 ℓ+νX¯ )+Γ(P 0 ℓ−νX) 1+ q/p 4 → → | | The value of q/p is to the first approximation only a phase (as explained in Section 1.5). The asymmetry (1.46) is of the order 10−3. Flavour eigenstates different from the CP eigenstates have been detected in the neutral K mesons [19]. The measured asymmetry was
Γ(K e+ν π−) Γ(K e−ν π+) aK = L → e − L → e =(3.27 0.12) 10−3. (1.47) SL Γ(K e+ν π−)+Γ(K e−ν π+) ± × L → e L → e The measurement in the decay of the KL meson is usually presented in terms of the parameter Re ǫ¯ aK /2, such that K ≃ q 1 2Re ǫ¯ , (1.48) p − ≃− K K
so that q/p K =0.99673 0.00012. It is difficult to relate the asymmetry in terms of fundamen| tal| CKM paramet± ers because the calculation of q/p involves poorly known hadronic matrix elements. | | There exists yet another method to detect CP violation. CP violation in the interference between direct decays and decays via mixing gives access to the phases of the quantities A/¯ A and q/p. In the decay of neutral mesons into CP eigenstates,
A = f P 0 , A¯ = f P¯0 , (1.49) CP |H| CP|H|
1.6. EXPERIMENTAL OBSERVATION OF CP VIOLATION 27 the product q A¯ λ = (1.50) p · A is physically meaningful. This is because all phase convention dependence of q/p cancels against A/¯ A:
λ =1 CP violation. (1.51) 6 ⇒ In this case, interference between partial decays is not necessary and hadronic uncertainties can therefore be minimized. For many B decays it is true that A/¯ A = e2iφ as well as q/p =e−2iβ to first approximation. This gives
λ =1, but Im λ = sin(2β 2φ). (1.52) | | − − The actual measurement in an experiment is the time-dependent asymmetry be- tween the number of B0 and B¯0 mesons decaying to the same CP eigenstate:
0 0 Γ(B (tˆ) fCP ) Γ(B¯ (tˆ) fCP ) af (tˆ)= → − → . (1.53) CP Γ(B0(tˆ) f )+Γ(B¯0(tˆ) f ) → CP → CP The asymmetry is only visible after allowing some time for mixing. At an arbitrary time, tˆ, the flavour composition is given by Eq. (1.32). The amplitude of the decay of this B to the final state fCP is
0 −im tˆ − 1 Γ tˆ 1 q 1 f B (tˆ) =e B e 2 B A cos ∆m tˆ + iA¯ sin ∆m tˆ CP 2 B p 2 B (1.54) |H| ˆ 1 ˆ −imB t − 2 ΓB t 1 ˆ 1 ˆ = Ae e h cos 2 ∆mBt + iλ sin 2 ∆mBt i The width is ˆ 1+|λ|2 1−|λ|2 Γ(B0(tˆ) f )= A 2e−ΓB t + cos(∆m tˆ) Imλ sin(∆m tˆ) . → CP | | 2 2 B − B h (1i.55)
In a similar way, the corresponding width for B¯0(tˆ)is
ˆ 1+|λ|2 1−|λ|2 Γ(B¯0(tˆ) f )= A 2e−ΓB t cos(∆m tˆ)+Imλ sin(∆m tˆ) . → CP | | 2 − 2 B B h (1i.56)
Using the last two equations, the measurable asymmetry in (1.53) is calculated as (1 λ 2)cos(∆m tˆ) 2Imλ sin(∆m tˆ) a −| | B − B . (1.57) fCP ≃ 1+ λ 2 | | In the approximation λ =1, | | ∆m a = Imλ sin x t, x = B , (1.58) fCP − d d Γ where ∆mB is the mass difference between the mass eigenstates of the B mesons, Γ is the inverse lifetime and t is measured in units of the lifetime. 28 CHAPTER 1. B PHYSICS AND CP VIOLATION
± ± bc d d
± 0 J/ψ ± 0 B B KS ± s c ± d s b q c ψ ± KS J/ d ± c a) b)
Figure 1.12: The golden decay. a) The tree diagram and b) the penguin diagrams. The quark in the loop can be either t, c or u.
The Golden Decay
0 0 The decay B J/ψKS offers a very clean measurement of CP violation for four main reasons: →
1. CP violation is expected to be large in B0 B¯0 mixing. − 2. The final state is a CP eigenstate.
3. The contamination from penguin diagrams is small.
4. It has a clear signature to pick out from a high combinatorial background.
The last point is explained in Chapter 2. Because of these favourable conditions, the channel is labelled the gold-plated mode for measuring sin 2β. The decay diagrams are depicted in Fig. 1.12. The amplitude of the tree diagram ¯ ∗ 2 4 is proportional to Atree = VcbVcs = Aλ + O(λ ). There are three types of penguin contributions, each with a t-, c-orau-quark in the loop. All penguin diagrams are suppressed due to this loop. The amplitude of ∗ ¯ 4 the t-quark diagram is proportional to VtbVts = Atree + O(λ ). With a c-quark the ¯ − ∗ 4 amplitude is proportional to Atree and with a u-quark to VubVus = O(λ ). Hence all penguin diagrams contribute with the same weak phase as the tree diagram, up to corrections of order λ4 10−3. The hadronic uncertainties are only of that order. ∼0 0 ¯ 0 The CP eigenstate KS is formed from K K mixing. This mixing factor adds to − ∗ the value of λ 0 . The participating amplitudes, Vcs and V have no CP-violating J/ψKS cd phases in the Wolfenstein parametrization up to order λ4 and do not contribute to the imaginary part of λ 0 . The ratio of the decay amplitudes in Fig. 1.12, J/ψKS
¯ ∗ Atree VcbVcs = ∗ , Atree VcbVcs 1.6. EXPERIMENTAL OBSERVATION OF CP VIOLATION 29
b s φ t s- - 0 B s
K S - d
Figure 1.13: Penguin diagram for the decay B¯0 φK0 . → S is also real, giving φ = 0; and using (1.52) and (1.35),
ImλJ/ψK0 = sin(2β 2φ)= sin(2β). (1.59) S − − − Detection of an asymmetry in the golden decay gives a direct measurement of sin 2β of the unitary triangle with hadronic uncertainties of order 10−3.
Other measurements of the angle β Some other B decays which can be used to measure the angles of the CKM unitary 0 0 triangle are B ππ, B DD¯, B φKS and Bs ρKS. →¯0 0→ → → The decay B φKS involves a flavour-changing neutral current in the quark transition b sss¯ .→Since flavour-changing neutral currents are forbidden at tree level in the Standard→ Model, this decay proceeds through the penguin transitions shown in Fig. 1.13. The diagrams with a t and a c quark in the loop contribute with the same phase up to order λ4, whereas the u contribution is of order λ4. Relative to the leading penguin diagram, the penguin with a u quark in the loop contributes with ∗ ∗ the order VubVus / VtbVts 0.03. The deca| y amplitude| | r|≃atio ¯ ∗ Apenguin VtbVts = ∗ Apenguin VtbVts is real and consequently
ImλφK0 = sin(2β 2φ)= sin 2β. (1.60) S − − − The decay B¯0 D∗−D∗+ shown in Fig. 1.14 has a tree diagram contribution of order λ3, which is →the same as for the penguin diagram. The penguin diagram contri- bution is loop suppressed by an order of 2%. On the other hand, the hadronic matrix elements of penguin operators are usually enhanced, leading to a total suppression of (4–10)%. The hadronic uncertainties are reduced to this order. The problem with a final state composed of two vector mesons is that its different helicity states will have different CP parities and hence different signs of the CP asymmetry. Experiments on other decays to vector mesons show that the longitudinal polarization dominates and the same could be true for this decay. Another B decay with a final state composed of vector mesons, which can be used to measure sin 2β is B¯ J/ψρ0. → 30 CHAPTER 1. B PHYSICS AND CP VIOLATION
d D*- d - b c D*- q - b c c - 0 B c - 0 + B D* - d D*+ - d
Figure 1.14: Tree and penguin diagrams for the decay B¯0 D∗+D∗−. →
d - π d - b u π - q - u b u - B0 - + u B0 π - d π + - d
Figure 1.15: Tree and penguin diagrams for the decay B¯0 π+π−. →
Measurements of the angle α The decay B¯0 π+π− proceeds through the quark decay b uud¯ , as shown in Fig. 1.15. The→tree and penguin diagrams are both of order →λ3, but the penguin diagram is suppressed by the order of 10%. The admixture of penguin contributions leads to λ = 1. These hadronic uncertainties of the order of 10% can be reduced by using iso|spin|6 analysis. To good approximation ¯ ∗ A VubVud −2iγ ∗ =e A ≃ VubVud and Imλ = sin(2β +2γ) = sin(2α). (1.61) ππ −
Measurements of the angle γ Of the three angles of the unitary triangle, the γ angle is the most difficult to measure, because there are no decays with high branching ratios and small penguin contribu- tions, which give direct access to this angle. In a hadronic B factory the angle can be measured in the decay of the Bs meson. In the Bs system the oscillations are fast. The time-dependent asymmetry can only be measured if the mixing parameter xs is not too large compared with the z-resolution of the experiment. 4 In the mixing of Bs mesons q/p is real up to order λ :
∗ q VtbVts = ∗ 1 . (1.62) p VtbV ≃ Bs ts 1.6. EXPERIMENTAL OBSERVATION OF CP VIOLATION 31
- - b u s s ρ - 0 - B K0 Bs s S -u d -s b q u d ρ 0 K0 s- S -u
Figure 1.16: Tree and penguin diagrams for the decay B¯ ρ0K0 . s → S u- b u - 0 - 0 - D B b c D B c- u- u- s K- K- s u-
Figure 1.17: Diagrams for the decays B− K−D0 and B− K−D¯ 0. These diagrams interfere in the case in which the →D0 and D¯ 0 mesons →decay to the same final state.
¯ 0 0 The decay Bs ρ KS shown in Fig. 1.16 has a tree contribution of the order of ∗ 3 → VubVud = O(λ ). The hadronic uncertainties are of the order of 10% for the same |reasons|as explained for the B¯0 π+π− decay. Thus → ∗ ∗ ∗ VtbVts VudVub VcsVcd −2iγ λρK0 = = e (1.63) S V V ∗ V V ∗ V V ∗ tb ts ud ub cs cd and ImλρK0 = sin(2γ). (1.64) S − ± ∓ Other examples of decays which measure the γ angle directly are the Bs Ds K decays. → The γ angle can also be measured in direct CP violation using some less straight- forward approaches. A method with small hadronic uncertainties is presented in Ref. [21]. The idea is to measure the asymmetry Γ(B− K−f) Γ(B+ K+f¯) adirect = → − → , (1.65) Γ(B− K−f)+Γ(B+ K+f¯) → → where f is the final state of a D or D¯ decay. Thus the two interfering amplitudes needed for accessing the CP asymmetry are the two quark transitions b cus¯ and b cus¯ . The two diagrams are shown in Fig. 1.17. The B− K−D¯→0 decay is co→lour suppressed. To make up for this, the final state is chosen→such that D¯ 0 f is Cabbibo allowed whereas the D0 f decay is doubly Cabbibo suppressed.→For the modes f = K+π− and f = Kππ→the two interfering amplitudes are of the same order. Another clean approach is to measure the four B πK decays. This can give an upper limit on sin2 γ [20]. → 32 CHAPTER 1. B PHYSICS AND CP VIOLATION
1.7 CP violation beyond the Standard Model
All CP violation arises from a single phase, which makes the Standard Model picture very predictive. From the single measurement in the K system, all other CP-violating observables can be predicted. Most probably, however, the Standard Model does not describe the whole picture of CP violation. There are several reasons to believe that there is CP violation beyond the Standard Model. First of all because almost any extension of the Standard Model includes additional sources of CP-violating effects. Such models are denoted as new physics. Models like those including FCNC (Flavour Changing Neutral Currents) and supersymmetry introduce a host of new CP violating phases, which are all small. In most models, CP is an approximate symmetry, whereas in the Standard Model, CP is almost maximally violated (as found from the asymmetry measured in the K system). Another strong argument for CP violation beyond the Standard Model comes from cosmology. In the initial state of the expanding universe, baryons and an- tibaryons were created with a number density comparable to that of photons. As the universe cooled, baryon pair creation could no longer compensate for the annihilation to photons. If CP had been conserved, the annihilation of baryons and antibaryons would have continued until the expansion rate exceeded the annihilation rate and the final ratio would have been [22]
n b 10−18 . (1.66) n ∼ γ CP
From the present distribution of matter , the baryon-to-photon ratio in the universe is estimated to be n b =(4 7) 10−10 . (1.67) n − × γ Observed
−8 This corresponds to a baryon–an tibaryon asymmetry of 10 in the early universe. A baryon asymmetry requires baryon-number violation, CP violation and a deviation from thermal equilibrium [23]. The baryon asymmetry could be generated by the decay of a super-heavy particle (for example an X or Y boson or a majorana neutrino) at the GUT scale (1015 GeV). In this case it is very unlikely that the responsible CP- violating processes have any consequences in high-energy physics experiments. The baryon asymmetry could also have been created in the electroweak phase transition. Baryon-number violation in the Standard Model of electroweak interactions is rapid at high temperatures but stops after the phase transition, when the electroweak symmetry is broken into electromagnetism. This happens at a scale of 100 GeV, which is accessible in laboratories. This picture can only explain the observed baryon asymmetry if the Standard Model is extended to account for a phase transition of enough strength and enough CP asymmetries. The Standard Model prediction of the CP asymmetry is many ( 10) orders too small to explain the imbalance between matter and antimatter [24∼]. The so-called ‘strong CP problem’ urges also for extensions to the Standard Model. QCD allows a source of CP violation, but extreme fine-tuning is needed 1.7. CP VIOLATION BEYOND THE STANDARD MODEL 33
- - b c s s ψ - - J/ B φ Bs s -c s -s b q c s φ J/ ψ s- -c
Figure 1.18: Tree and penguin diagrams for the decay B¯ J/ψφ. s → in order that its contribution to the electric dipole moment of the neutron does not exceed the experimental bound. Some mechanism beyond the Standard Model is wel- come to explain this situation. The electroweak contribution to the electric dipole moment is strongly suppressed and therefore any detected electric dipole moment of the neutron would be due to strong CP violation. The measurement of the electric dipole moment of the neutron is thus a very interesting probe for CP violation. In summary, CP violation is one of the least tested aspects of the Standard Model and one of the most promising directions in the search for new physics. If the new physics is not too high above the Standard Model scale (i.e. much lower than the GUT scale), deviations from the predictions are expected to be observed in the B system accessible at HERA-B. In the rest of this chapter, various ways to detect non-Standard Model CP viola- tion in HERA-B are investigated. The primary goal in the search for new physics is to look for inconsistencies. Independent measurements of an angle can give different results and the unitarity conditions of the CKM matrix can appear to be broken. An important observation is that the unitary triangle is a natural condition in models with three lepton families. Unless a fourth family is introduced the unitarity of the CKM matrix is still valid. From this assumption it is possible to find information that can tell about the nature of the new physics contribution. Does it for example appear in mixing (∆B = 2) or in decay (∆B = 1) or in both mixing and decay? Are the new sources of CP violation flavour diagonal or due to flavour changing neutral currents, or both? Such information is valuable in order to rule out models and select directions in the construction of theories of new physics.
The Bs system The detection of deviations from the Standard Model prediction requires precision measurements. Attention is drawn to the asymmetry detected in the interference be- tween direct decay and decay via mixing because of the small hadronic uncertainties. A clean method is to search for asymmetries in decay channels where a small CP violation is predicted by the Standard Model. In particular, CP asymmetries in all Bs decays that do not involve direct b u transitions are all the same and approximately zero. → Such a mode is B J/ψφ depicted in Fig. 1.18. It is predicted to have the s → relatively large branching fraction of 0.3% [27]. Combined with the Bs production 34 CHAPTER 1. B PHYSICS AND CP VIOLATION suppression, this decay will occur at roughly the same rate in HERA-B as the golden decay. Also the trigger and reconstruction conditions are similar. The CKM ampli- tudes contributing to this decay are all real up to order λ4:
∗ q VtbVts = ∗ 1 , (1.68) p VtbV ≃ Bs ts ∗ Atree VcbVcs = ∗ 1 . (1.69) A¯tree VcbV ≃ Bs cs The amplitudes of the penguin diagrams are also real to O(λ4). In addition they are loop suppressed. Therefore, this mode is as clean from hadronic uncertainties as the golden decay. Any CP asymmetry detected in this decay will be a sign of new physics, with practically no Standard Model background.
Measurements involving the golden decay
0 0 New physics can also be found in the golden decay B J/ψKS, which is dominated by the tree contribution. This amplitude is very unlik→ ely to be modified by new physics, since the scale of the latter is expected to be large compared to MW .Onthe other hand, the mixing amplitude can easily be modified. The strong suppression of the Standard Model box diagrams in B B¯ mixing by the fourth order of the weak coupling and small CKM angles allows fo−r large contributions from new physics. This would shift the asymmetries in B decays universally (φ φ + θd). In particular for the golden decay: → Imλ 0 =sin2(β + θd) . (1.70) J/ψKS It is only possible to disentangle the new physics contribution from the Standard Model contribution if the unitarity of the triangle β angle can be constrained by independent measurements on the sides and on other angles (for example sin 2(α + θd)). The β angle in the unitary triangle can be measured in numerous decays. The 0 0 measurement in the golden decay B J/ψKS is extremely clean and in the decay 0 → B φKS it can be measured with a precision of 3%, due to the effect of light qua→rks in the penguin loop. The Standard Model thu±s predicts that the asymmetries measured in those two decays should be equal:
aB→J/ψK0 = aB→φK0 3% . (1.71) S S ± A violation of this equality would be a sign of new physics. New physics in B B¯ 0 − mixing introduces a shift of the same size for the two decays. The B φKS involves a flavour changing neutral current. The leading SM contributions are →penguin diagrams with a suppression by αs and a loop factor. Here extensions to the SM are possible which could lead to larger CP asymmetries [20]:
Imλ 0 =sin2(β + θd + θA) (1.72) φKS 1.7. CP VIOLATION BEYOND THE STANDARD MODEL 35
so that aB→J/ψK0 = aB→φK0 . (1.73) S 6 S Relatively small effects, of order 10% can lead to an observation of such an inequality. Another interesting example is the decay KL πνν¯ of the long-lived K-meson, which also measures the β angle [28]. This decay m→easures the relative phase between K0 K¯0 mixing and the quark transition s dνν¯. In the presence of new physics, there− is no reason for a relation between the→asymmetry in this decay and decays of B mesons.
Semileptonic decays 0 0 New physics contributions to the B B¯ and Bs B¯s mixing matrices could also lead to an observable CP asymmetry −in semileptonic− B decays. This is measured in the decay of a BB¯ pair into equal-signed leptons:
Γ(BB¯ ℓ+ℓ+X) Γ(BB¯ ℓ−ℓ−X) Γ Γ a = → − → =Im 12 = 12 sin φ (1.74) SL Γ(BB¯ ℓ+ℓ+X)+Γ(BB¯ ℓ−ℓ−X) M M 12 12 12 → → −3 In the Standard Model this asymmetry is very small (O(10 )) beca use of the small difference in the decay rates and the small phase difference φ12 between Γ12 and M12. New contributions to B B¯ mixing would change the phase φ but not the relation − 12 Γ12 M12 , since the new physics are only expected to contribute at higher order |[25].|≪| | + 0 The final hadronization states can be any combinations of B , B , Bs,Λb and their conjugates. However, if the flavour content can be determined and neutral mesons selected, the asymmetry is larger at the cost of statistics. Chapter 2
The HERA-B experiment
The HERA-B experiment is a forward spectrometer at the HERA electron-proton collider. The proton beam is directed on a fixed nucleus target whereas the electron beam is passing untouched through the detector. The experiment is primarily de- signed to detect CP violation in the asymmetry between the fraction of B0 and B¯0 0 decaying to J/ψKS,wheretheJ/ψ decays to a lepton pair. The rate of this signal is suppressed by a factor of 10−11 per inelastic event. Owing to the smallness of the asymmetry and the total reconstruction efficiency of about 10%, the order of 10 000 such golden events are needed for a CP-violation measurement. The proton bunch-crossing rate at HERA is 10 MHz, which is equal to about 1014 bunch crossings per year. Clearly more than one event per bunch crossing is needed for a CP measurement in one or two years. The solution is to put target wires around the beam core in a manner described in fig 2.1. With this configuration four or five Poisson-distributed interactions per bunch crossing can be extracted from the off-momentum protons that are in the halo. This procedure allows HERA-B to operate without interrupting the other experiments around HERA. The final state of the golden decay is a CP eigenstate, which means that it does
Figure 2.1: Schematic view of the halo target, consisting of eight metal ribbons on indepen- dently movable forks. The ribbons are about 50 µm wide and 500 µm thick. They are sepa- rated along the beam by about 5 cm. 2.1. REQUIREMENTS 37
not determine the flavour of the B meson. The b-quarks can only be produced in pairs (fig 1.1). The B meson under study is therefore always produced together with a B hadron of opposite flavour. This hadron does not normally decay to a CP eigenstate. The most common decay is to the lighter D meson with the subsequent decay to a K meson. Usually those decays are reconstructed and serve as a tag. Exclusive reconstruction has a high quality but poor efficiency. Other possibilities are to determine the charge of the kaon or the lepton accompanying the D meson. Another option is to geometrically reconstruct the decay vertex of the tagging B and count the charges of secondary tracks. Finally the charge of soft pions accompanying the B meson can serve as a tag. Soft pions come from either the decay of an excited B∗∗ meson or from the local charge conservation in the quark fragmentation process. The combination of all these methods will bring the tagging power to the calculation of the precision of the CP measurement. Since the b-quark quantum number is not conserved in weak interactions, the B0 and B¯0 mix to the physical mass eigenstates, which are compounds of B0 and B¯0.If there were no CP violation these compounds would be equal to the CP eigenstates. Because of the mass difference between the mass eigenstates, the asymmetry oscillates with time. The decay time of the B meson has to be measured as the flight distance in the detector. This is the distance between the main vertex at a target wire and a displaced secondary decay vertex. The experimental programme also includes measurements of CP violation in other decay modes like B π+π−, studies on B0 B¯0, B0 B¯0 and D0 D¯ 0 mixing, and → − s − s − properties of the Λb hadron. The matrix element Vcb can be determined from the ∗ semileptonic decay B D ℓνℓ. Rare decays are im→portant to test the Standard Model predictions. The predicted branching ratio of B Kℓ+ℓ− is 5 10−7 and for the decay to a vector meson it is slightly larger: → · BR(B K∗e+e−) 4 10−6 → ≃ · and BR(B K∗µ+µ−) 2 10−6. → ≃ · Because of the small branching ratios of these decays, the trigger must be very selective. Since there are no requirements on the other B in the event, decays with such small branching ratios and two final-state leptons can be detected. Measurements of total and differential cross sections over the entire phase space will be important for testing and developing QCD models. So far, there is only a limited amount of data available on B and Υ production at HERA-B kinematics. Also hadronic decays of B hadrons will be of interest. Charm physics is also included in the experimental programme.
2.1 Requirements
The goal to reconstruct the rare B decays in the environment of a very high track density demands a careful choice on trigger and detector design. In particular a fast readout is required because of the high bunch-crossing rate. 38 CHAPTER 2. THE HERA-B EXPERIMENT
The hardware has to be radiation hard. It will be exposed to 3 107/R2 particles per second, which corresponds to for example 100 krad/yr at a dis·tance R =10cm from the beam. Further downstream the flux is even higher due to secondary in- teractions in the detector material. For cost reasons and to pay regard to the other experiments around HERA (H1, ZEUS and Hermes), the components have to operate for one year at full rate before they can be replaced. The requirements on precise measurements in the full kinematic region and the profile of the signal put the special requirements on the HERA-B detector:
Large geometrical acceptance. • A fine-granularity tracker with good momentum resolution to observe a narrow • peak of a B meson which is reconstructed from four final-state particles.
Reconstruction of multiple secondary vertices at 1 cm from the beam, which • is the mean decay length of the B. At this positio∼n the annual radiation dose is 10 Mrad.
Efficient muon and electron identification. This is particularly important for • the reconstruction of the J/ψ from the golden decay. A J/ψ candidate is selected already at the first triggering stage.
Efficient kaon identification. This concerns in one way the reconstruction of • 0 + − KS π π from the decay of the neutral B in the golden channel, but more pron→ounced the identification and separation of charged kaons of the tagging B in the same event, from the copious production of pions and protons.
2.2 Vertex Detector
The precise reconstruction of multiple displaced vertices relies on the Vertex Detector System (VDS) consisting of seven superlayers of silicon strip detectors, positioned close to the target and starting at 1 cm radius from the beam. The information from the VDS will be used already at the second level trigger. The target wires and the vertex-detector superlayers are placed in a vacuum tank, named the Vertex Vessel in Fig. 2.2. Each superlayer of VDS consists of four quadrants as shown in Fig. 2.3. The quadrants are contained in pot assemblies that can be displaced individually in the radial and lateral directions. Each sector is built up by two double-sided detectors. The principle of the silicon strip detectors is illustrated in Fig. 2.4. The strip pitch is 25 µm and every second strip is read out. Both sides are read out, giving two views per detector module. Each vertex-detector superlayer thus provides four views, 2.5◦, 87.5◦ and 92.5◦,and stand-alone pattern recognition can be performed over most±of the angular range. The resolution of a secondary vertex is about 500 µminthez-direction (along the beam) and 25 µminthexy-projection. The VDS must also determine the impact parameter of tagging particles. The impact-parameter resolution is 25 µm 30 µm/p . ⊕ T 2.2. VERTEX DETECTOR 39
↓↓ ↓↓ ↓ ↓↓
p e+
Vertex Vessel
↑ Magnet TRD ↑ ECAL x RICH Muon System
02468101214161820z[m]
Figure 2.2: The design of the finalized HERA-B detector as simulated in GEANT [61]. Planes used in the first level trigger are indicated by arrows. 40 CHAPTER 2. THE HERA-B EXPERIMENT
Figure 2.3: The geometrical layout and positioning of the silicon vertex-detector su- perlayers. (The position of the silicon detector belonging to the tracking system is also indicated at z = 206 cm.) The polar angle coverage is 10–250 mrad. There are two double-sided detector modules per quadrant in four stereo angles. 2.3. TRACKING SYSTEM 41
+Vbias
Aluminium
SiO2 22X n+ implantation p-compensation Guard Ring Structure n-type silicon 280 µ m
p+ implantation 25 µ m
Figure 2.4: The principle of a silicon strip detector. Two bias rings surround the active detector region. The protective guard ring structure provides a controlled drop of bias voltage. Every second strip is read out both on the p- and the n-side.
2.3 Tracking system
The HERA-B detector (fig 2.2) is a large-acceptance forward spectrometer covering 10 mrad up to 160 mrad. This corresponds to 90% of the coverage in the centre-of- mass system. For momentum measurements, the tracks are bent horizontally in an inhomoge- neous magnetic field. The field integral is 2.2 Tm. The field is produced by a dipole magnet, centered at about 4.5 m from the target. In the bending plane the lateral acceptance reaches out to 250 mrad. The granularity of the tracking system varies with the distance from the beam in order to limit the occupancy and minimize the number of channels. The tracking system is divided in three different technologies depending on the distance from the beam: 1 6 cm. Silicon strip of the same technology as the vertex detector. The spatial • − resolution is 12 µm. The B vertex resolution is σxy =25µm,σz = 500 µm. 6 25 cm. Inner Tracker (IT). Micro-Strip Gaseous Chambers (MSGC) with • Ga−s Electron Multiplier (GEM). The resolution is 80 µm. ∼ > 20 cm. Outer Tracker (OT). Drift tubes in a honeycomb structure. The • spatial resolution is 150–200 µm. ∼ A silicon strip superlayer is placed in front of the magnet, where the track density is very high. This superlayer is referred to as the 8:th and last VDS superlayer. The inner and outer tracker chambers in the magnet will be used for finding the curvature of tracks. The chambers in the field free region between the magnet and 42 CHAPTER 2. THE HERA-B EXPERIMENT
Figure 2.5: The geometry of the MSGC. Two layers are shown forming one superlayer with two stereo angles.
RICH are mainly used for pattern recognition of straight tracks. There are also inner and outer tracker chambers in front of the calorimeter, to extrapolate tracks to clusters in the calorimeter and to the muon system. About half of all tracks in a HERA-B event will be reconstructed in the IT. Four L-shaped MSGC modules with dimension 30 30 cm2 constitute one IT layer around the beam, as shown in Fig. 2.5. Every superl×ayer consists of two or more layers with different strip orientations. The three stereo angles are 0◦ and 5◦. The trigger planes have double layers of each orientation to guarantee full efficiency± . One MSGC is built as a glass box with electrodes on a glass substrate, a 6 mm long drift region and a glass cover coated with a drift cathode. The glass substrate is coated with diamond to prevent accumulation of charges between anode and cathode electrodes. The field becomes stronger near the anodes (as can be seen in Fig. 2.6), causing an avalanche with a gain of 150. The total gain is 3000, separated in two steps to avoid sparks at the anode. For this purpose there is a GEM with a gain of 20 in the middle of the drift region (Fig. 2.6). It is a perforated polyimide foil coated with copper. The coordinate is measured by detecting the centre of gravity of a strip cluster. For radial distances from the beam axis larger than about 20 cm the particle densities are small enough that straw tubes of 5 and 10 mm diameter can be used. The Outer Tracker layers are grouped in thirteen superlayers of which seven belong to 2.3. TRACKING SYSTEM 43
Drift Cathod
3mm
U1 GEM U2
3mm
Anode Cathode
300 µ m 300 µ m
Figure 2.6: The principle of a GEM-MSGC. The particles enter from the glass sub- strate and exit by the drift cathode.
the magnet tracking (MC01–MC06, MC08) and four to the pattern tracking (PC01– PC04). PC01 and PC04 are also included in the first level trigger. The two large trigger chambers (TC01 and TC02) are also useful for extrapolation to the calorimeter and the muon system after the RICH. Inside all superlayers except MC02, MC04 and MC08, there is an IT superlayer. Each superlayer contains one or two sets of the three stereo angles, 0◦ and 5◦. The pattern-recognition chambers and the trigger chambers have two sets. M±C01 has a sequence 0◦, 5◦,+5◦,0◦ to link track segments found in the vertex detector. MC05 consists only−of a single 0◦ layer because of space restrictions inside the magnet yoke. A single layer is assembled from four folded pocalon foils. The pocalon is coated with gold and copper to increase the conductivity. TC1, TC2 and half of PC1 and PC2 consist of double layers as in Fig. 2.8. The diameter of the honeycomb cells is 5 mm for the innermost sections (shaded in Fig. 2.7) to limit the occupancy and 10 mm for the rest. This size provides a sufficient cell granularity for the first level trigger, which uses only wire hit information. At higher trigger levels and offline reconstruction, the wire information is supplemented by drift-time information. The drift time measures the radial distance of the impact point from the hit wire. The momentum resolution of the tracking is very good. For muons from the golden decay, a Gaussian fit gives [68] ∆p/p =(8.1 0.3) 10−3, which leads to a mass resolution of about 8–10 MeV for B0 J/ψK0±. · → S Inside and directly behind the magnet there are also three additional pad cham- bers. They are used for a high-pT hadron extension of the trigger. Its primary purpose is the search for B0 π+π− decays. → 44 CHAPTER 2. THE HERA-B EXPERIMENT
Figure 2.7: The geometrical structure of TC2. The sections in the shaded area have 5 mm cells; the rest have 10 mm cells. The proton and electron beam pipes are indi- cated. An IT superlayer fits in the central hole.
Figure 2.8: The principle of the Honeycomb Drift Chambers: a double layer of the type which are in the trigger chambers. The particles enter from the left.
5/10 mm 25-30 µ m 2.4. PARTICLE IDENTIFICATION DETECTORS 45