next up previous contents
Next: Beam Remnants Up: lutp0613man2 Previous: Routines and Common-Block Variables   Contents


Beam Remnants and Underlying Events

Each incoming beam particle may leave behind a beam remnant, which does not take active part in the initial-state radiation or hard-scattering process. If nothing else, these remnants need to be put together and colour connected to the rest of the event. In addition, in hadron-hadron collisions, the composite nature of the two incoming hadrons implies the possibility that several pairs of partons can enter into separate but simultaneous scatterings, `multiple interactions'. In some fraction of events, these additional scatterings can be hard or semi-hard, but due to the infrared peaking of the cross section the bulk of them should normally be fairly soft compared to the primary interaction. This extra component gives a non-negligible contribution to the `underlying event' structure, and thus to the total multiplicity. Finally, in high-luminosity colliders, it is possible to have several collisions between beam particles in one and the same beam crossing, i.e. pile-up events, which further act to build up the general particle production activity that is to be observed by detectors. These three aspects are described in turn, with emphasis on the middle one, that of multiple interactions within a single hadron-hadron collision.

Work is ongoing to improve the theoretical framework for multiple interactions. PYTHIA therefore contains two partly separate approaches to this physics, which also leads to two partly separate descriptions of beam remnants. To make matters worse, from a pedagogical point of view, also the description of initial- and final-state radiation has been improved, with the transition from virtuality-ordered to transverse-momentum-ordered evolution, and with the interleaving of multiple interactions and initial-state emissions in one common $p_{\perp}$-ordered sequence. In total we therefore need to distinguish three main scenarios.

$\bullet$
The `old model', based on [Sjö87a], contains the old beam-remnant and showering machineries. It remains the default while the new one is being developed, for backwards reference and comparisons. For instance, the `Tune A' [Fie02] incarnation of this model is commonly used at the Tevatron, and offers a convenient reference against which other models can be compared, without the need to know how to do a full detector simulation. Furthermore, the newer models below are mainly being developed for $\mathrm{p}\mathrm{p}$ and $\mathrm{p}\overline{\mathrm{p}}$ collisions so far, so have barely been tested with meson and fixed-energy resolved-photon beams, and is not at all integrated into a complete consistent set of photon interactions for a spectum of incoming energies and virtualities.
$\bullet$
The `intermediate model', developed in [Sjö04], attempts a more sophisticated description of correlations in flavour, colour, longitudinal and (primordial) transverse momenta between the beam remnants and the shower initiators than offered by the old one, and also introduces some other improvements. The virtuality-ordered showers are the same as in the old model, but now each interaction can be associated with its own showering activity, while before only the hardest interaction would be showered, for technical reasons. In practice, we do not expect much usage of the intermediate model: either people stay with the old or move right to the new one below. We have kept it mainly so that results from our publications can be reproduced.
$\bullet$
The `new model', with further aspects described in [Sjö04a], where transverse-momentum-ordered showers are introduced and multiple interactions and initial-state radiation are interleaved. The beam-remnant description introduced in the intermediate model stays essentially unchanged.
Several of the key multiple interactions aspects are common between the models, such as the generation of kinematics of the multiple (semi)hard interactions and the impact-parameter picture. Other aspects are addressed in a more or less similar spirit, but in different ways, such as the desire to reduce the string length relative to naïve expectations. Therefore the current section starts with a description of the old model, and thereafter outlines where the intermediate and new models differ.



Subsections
next up previous contents
Next: Beam Remnants Up: lutp0613man2 Previous: Routines and Common-Block Variables   Contents
Stephen Mrenna 2007-10-30