Monte Carlo Generation

The main goal of the simulation is to study the behavior of the HERA-B detector and to obtain the acceptance correction function. The generation of Monte Carlo (MC) events is done in two steps: first, particles produced in the primary interactions are generated and then the particles are propagated through the simulated detector.

PYTHIA (version 5.7) and FRITIOF (version 7.02) [#!Ivar!#,#!Fri!#] are the generators used at HERA-B to simulate events. PYTHIA internally makes use of the jet fragmentation utilities of the JETSET [#!Sjoe!#] package. The PYTHIA package simulates the collisions of nucleons (protons and neutrons), nuclear effects are not covered. After the simulation of the primary interaction the remaining energy is passed into the FRITIOF simulation package. FRITIOF is used for simulating nuclear effects.

HBGEAN [#!Novak!#] is the package used in HERA-B for detector simulation, it is a modified version of the GEANT [#!Gea!#] package. At this step of the simulation all geometrical features and materials of detector components of HERA-B are taken into account.

All particles produced by PYTHIA and FRITIOF are passed into HBGEAN and propagated through the simulated detector. All traced particles are leaving Monte Carlo Impact Points (MIMP) in the sensitive volumes of the detector (see Chapter 5). Energy loss and multiple scattering are taken into account at this step according to the initial particle momenta. All important information is stored in ARTE tables: MIMP table for MIMPs, MTRA (MC tracks) table for generated particles and MVER (MC vertices) table for vertices. All these tables are linked with each other in order to provide easy navigation.

Up to this point no information about efficiencies and resolutions of detector components are taken into account. The information at this step can be treated as MC truth.

Simulated information is then digitized in a format so that it can be treated by the reconstruction program in the same way as real data. At this stage realistic detector performance can be simulated by applying efficiencies and dead channel masks. The detector resolution is simulated by smearing the hit positions according to the resolution measured on data.

To extract resolution and efficiencies of the Inner Tracker from Monte Carlo simulation, several runs were selected (one run per minimum bias data taking period) and the efficiencies and masks are produced with the procedure discussed in Chapter 5.

Figure 6.1: Schematic overview over data taking and simulation chains. The picture is based on [#!AGA!#]

After digitization all tables (HITB, HITC (hits in ECAL) etc) are filled. They are identical for MC and data, the identity makes it possible to treat data and MC in the same way (to use the same reconstruction packages in both cases). Figure 6.1 shows the path of data and MC in HERA-B.