488
J. Apostolakis
secondary particles. In the third stage the “hit” information is processed, to estimate
the signals that result.
In detector simulation the secondary particles and their descendant particles are
tracked in turn. Information about the passage of particles through the sensitive
parts of the detector is recorded as ‘hit’ information. For tracking detectors usually
the individual position, momentum and particle type or charge information of each
track is recorded; in calorimeters, the energy deposition within a cell is kept as a sum
over all tracks. A key characteristic of detector simulation is that tracks are treated
independently. Each particle created is tracked in turn, until all have exited the setup,
suffered a destructive interaction, or have been dropped as unimportant according
to a user’s chosen criteria. The dropping can be triggered, for example, when the
energy of a track falls below a threshold or arrives in a particular ‘unimportant’
region. Potential indirect interactions between particles are not treated as part of the
detailed simulation. As such, the creation of space charge in a gaseous detector must
be introduced in an experiment’s ‘user’ code or else treated separately.
11.2.1 Tools for Event Generation and Detector Simulation
The creation of the primary particles by the high-energy interaction is modeled using
specialized event generators. The type of interaction, energy range and applicability
of these generators differ significantly: whether they include hadron–nucleus and/or
nucleus–nucleus interactions, or the type of physics beyond the standard model
they provide. Typically event generators are independent programs: including the
established PYTHIA [8] and FRITIOF [9], which use the Lund fragmentation model
[10], and more recent ones such as HERWIG [11]/HERWIG++ [12]. Most provide
users with tunable parameters and the ability to create sets of parameter values
(‘tunes’) compatible with the most relevant reference data at the energies of interest.
Some Monte Carlo tools include high-energy event generators: e.g. DPMJET is
available in FLUKA, and has been used to simulate ion–ion collisions at RHIC and
the collisions of high-energy cosmic rays in the atmosphere.
Codes for the simulation of the detector must handle geometries of significant
complexity and a large number of volumes and they must model the full set
of hadronic, electromagnetic and weak interactions as accurately as required,
potentially within constraints of CPU time. In the past 20 years different tools have
been used for this purpose, including GEANT 3, FLUKA and GEANT4. Other multiparticle codes for particular applications including the MARS [13, 14] code, and
the SHIELD code which focus on ion–ion interactions. Different codes share some
physics models; for example PHITS and MARS share several models with MCNPX,
an extension of the neutron-gamma gold-standard code MCNP.
GEANT version 3 [15] was utilized by LEP experiments (ALEPH, L3), the
TeVatron experiments at Fermilab, numerous other experiments and also by the
ALICE experiment at LHC as its main simulation engine. It includes detailed
descriptions of electromagnetic interactions down to 10 keV. For hadronic physics it
J. Apostolakis
secondary particles. In the third stage the “hit” information is processed, to estimate
the signals that result.
In detector simulation the secondary particles and their descendant particles are
tracked in turn. Information about the passage of particles through the sensitive
parts of the detector is recorded as ‘hit’ information. For tracking detectors usually
the individual position, momentum and particle type or charge information of each
track is recorded; in calorimeters, the energy deposition within a cell is kept as a sum
over all tracks. A key characteristic of detector simulation is that tracks are treated
independently. Each particle created is tracked in turn, until all have exited the setup,
suffered a destructive interaction, or have been dropped as unimportant according
to a user’s chosen criteria. The dropping can be triggered, for example, when the
energy of a track falls below a threshold or arrives in a particular ‘unimportant’
region. Potential indirect interactions between particles are not treated as part of the
detailed simulation. As such, the creation of space charge in a gaseous detector must
be introduced in an experiment’s ‘user’ code or else treated separately.
11.2.1 Tools for Event Generation and Detector Simulation
The creation of the primary particles by the high-energy interaction is modeled using
specialized event generators. The type of interaction, energy range and applicability
of these generators differ significantly: whether they include hadron–nucleus and/or
nucleus–nucleus interactions, or the type of physics beyond the standard model
they provide. Typically event generators are independent programs: including the
established PYTHIA [8] and FRITIOF [9], which use the Lund fragmentation model
[10], and more recent ones such as HERWIG [11]/HERWIG++ [12]. Most provide
users with tunable parameters and the ability to create sets of parameter values
(‘tunes’) compatible with the most relevant reference data at the energies of interest.
Some Monte Carlo tools include high-energy event generators: e.g. DPMJET is
available in FLUKA, and has been used to simulate ion–ion collisions at RHIC and
the collisions of high-energy cosmic rays in the atmosphere.
Codes for the simulation of the detector must handle geometries of significant
complexity and a large number of volumes and they must model the full set
of hadronic, electromagnetic and weak interactions as accurately as required,
potentially within constraints of CPU time. In the past 20 years different tools have
been used for this purpose, including GEANT 3, FLUKA and GEANT4. Other multiparticle codes for particular applications including the MARS [13, 14] code, and
the SHIELD code which focus on ion–ion interactions. Different codes share some
physics models; for example PHITS and MARS share several models with MCNPX,
an extension of the neutron-gamma gold-standard code MCNP.
GEANT version 3 [15] was utilized by LEP experiments (ALEPH, L3), the
TeVatron experiments at Fermilab, numerous other experiments and also by the
ALICE experiment at LHC as its main simulation engine. It includes detailed
descriptions of electromagnetic interactions down to 10 keV. For hadronic physics it