Particle Decays

The ParticleDecays class performs the sequential decays of all unstable hadrons produced in the string fragmentation stage, i.e. up to and including b hadrons and their decay products, such as the tau lepton. It is not to be used for the decay of more massive resonances, such as top, Z^0 or SUSY, where decays must be performed already at the ProcessLevel of the event generation.

The decay description essentially copies the one present in PYTHIA since many years, but with some improvements, e.g. in the decay tables and the number of decay models available. Recently a more sophisticated handling of tau decays has also been introduced. Some issues may need further polishing.

Variables determining whether a particle decays

Before a particle is actually decayed, a number of checks are made.

(i) Decay modes must have been defined for the particle kind; tested by the canDecay() method of Event (and ParticleData).

(ii) The main switch for allowing this particle kind to decay must be on; tested by the mayDecay() method of Event (and ParticleData). By default this is defined as true for all particles with tau0 below 1000 mm, and false for ones above, see the Particle Data Scheme. This means that mu^+-, pi^+-, K^+-, K^0_L and n/nbar always remain stable unless decays are explicity switched on, e.g. 211:mayDecay = true.

(iii) Particles may be requested to have a nominal proper lifetime tau0 below a threshold. When on, only particles with tau0 < tau0Max are decayed. The above tau0Max, expressed in mm/c.

(iv) Particles may be requested to have an actual proper lifetime tau below a threshold. When on, only particles with tau < tauMax are decayed. The above tauMax, expressed in mm/c.
In order for this and the subsequent tests to work, a tau is selected and stored for each particle, whether in the end it decays or not. (If each test would use a different temporary tau it would lead to inconsistencies.)

(v) Particles may be requested to decay within a given distance of the origin. When on, only particles with a decay within a radius r < rMax are decayed. There is assumed to be no magnetic field or other detector effects. The above rMax, expressed in mm.

(vi) Particles may be requested to decay within a given cylindrical volume around the origin. When on, only particles with a decay within a volume limited by rho = sqrt(x^2 + y^2) < xyMax and |z| < zMax are decayed. There is assumed to be no magnetic field or other detector effects. The above xyMax, expressed in mm. The above zMax, expressed in mm.

Mixing

Allow or not B^0 - B^0bar and B_s^0 - B_s^0bar mixing. The mixing parameter x_d = Delta(m_B^0)/Gamma_B^0 in the B^0 - B^0bar system. (Default from RPP2006.) The mixing parameter x_s = Delta(m_B_s^0)/Gamma_B_s^0 in the B_s^0 - B_s^0bar system. (Delta-m from CDF hep-ex-0609040, Gamma from RPP2006.)

Tau decays

Decays of tau leptons can be performed using helicity information from the tau production process and with the hadronic current of the tau decay modelled using form factors fit to data. The tau decay framework is largely based on the corresponding Herwig++ implementation Gre07, with some input from Tauola Jad90. A short summary can be found in Ilt12, while the complete writeup is in Ilt14.

The decays of tau leptons are categorized as correlated, where a tau pair is produced from a single process, or uncorrelated, where only one tau is produced. Currently internally supported tau production mechanisms include correlated decays from gamma, Z^0, Z'^0, gamma^*/Z^0/Z'^0, and Higgs bosons (CP-even, odd, or mixed) and uncorrelated decays from W^+-, W'^+-, B/D hadrons, and charged Higgs bosons. For all mechanisms except B/D hadrons, both the full process, e.g. q qbar → Z^0 → tau^+ tau^-, as well as just the decay of the boson with a given initial polarization, e.g. Z^0 → tau^+ tau^-, can be handled. The axial and vector couplings of the Z'^0 and W'^0 are set from the relevant parameters in New Gauge Boson Processes. Note that the CP of the various Higgs bosons can be set with the options HiggsX:parity, HiggsX:etaParity, and HiggsX:phiParity as described in Higgs Processes where X is either H1, H2, or A3.

The tau polarization and tau decay correlation mechanism can be determined either using internal matrix elements or external SPINUP information provided in the event, e.g. via Les Houches Event Files (LHEF). For internal determination any tau pair or single tau from the processes of the previous list can be handeled. For external determination of a single uncorrelated tau, its polarization is set to its SPINUP information. When the SPINUP for the tau is not valid, e.g. when FSR is applied, the SPINUP for the first copy of that tau is used instead unless also invalid. For the external determination of a correlated tau pair the following options are available. Choice of the external polarization and correlation mechanism for correlated tau pairs. Note: option 1 has limited functionality as SPINUP is intended primarily for particles with 2 spin states. For massive vector bosons SPINUP is interpreted here as the transverse polarization and so the diagonal for the vector boson helicity density matrix is set as [(1 - SPINUP)/3, 1/3, (1 + SPINUP)/3].

A default behaviour is defined when the polarization and decay mechanism cannot be determined using either the internal or external methods. If the tau is known to be produced from a W^+-, gamma, or Z^0, the tau or tau pair is assumed to be produced from an unpolarized boson of this type. If the mediator is unknown but there is a correlated tau pair, the pair is assumed to be produced from an unpolarized photon and a warning is issued. Finally, if the tau is uncorrelated, an unpolarized and uncorrelated decay is performed and a warning is issued.

Both the internal and external determination have advantages and disadvantages. For example, if an LHEF Z^0 → tau^+ tau^- event is passed with SPINUP provided for both taus but without SPINUP for the Z^0 then with TauDecays:externalTau set to 0 the decays of the taus will be uncorrelated. Using 1 instead will result in correlations, assuming an unpolarized Z^0. If using internal determination, then the correlation and polarization will be fully calculated using the correct production mechanism for the Z^0. Consequently, a variety of options on how to determine polarization and correlation are available, with a sensible default in place which should catch most everything. Choice of tau decay model. Warning 1: options 2 and 3, to force a specific tau polarization, only affect the decay of the tau. The angular distribution of the tau itself, given by its production, is not modified by these options. If you want, e.g., a righthanded W, or a SUSY decay chain, the kinematics should be handled by the corresponding cross section class(es), supplemented by the resonance decay one(s). The options here could then still be used to ensure the correct polarization at the tau decay stage. Warning 2: for options 1 through 5, if the polarization and correlation mechanism for the tau cannot be determined (internally or externally) then the default behaviour described above is applied. Polarization of the tau when mode 2 or 3 of TauDecays:mode is selected. Mother of the tau for forced polarization when mode 2 of TauDecays:mode is selected. You should give the positive identity code; to the extent an antiparticle exists it will automatically obtain the inverse polarization.

QED radiation

So far PYTHIA does not have any generic machinery for handling QED radiation in normal particle decays. In order to include this, a program like Photos Bar94, Dav10 could be used as an afterburner. In a few cases, however, the existing shower machinery can be used also here: for two-body decays to a lepton pair (l^+ l^- or l^+- nu_l). Such decays are mediated by gamma^*/Z^0/W^+- exchange, for which PYTHIA does have an existing machinery that can be applied, including first-order matrix-element corrections for the first (hardest) photon emission. Allow or not photon radiations in decays to a lepton pair, see above. Note: The current default is to have radiation switched off, in order to avoid double-counting of emissions if you link to an external QED-radiation program, as is the norm in many collaborations.

Other variables

Minimum mass difference required between the decaying mother mass and the sum of the daughter masses, kept as a safety margin to avoid numerical problems in the decay generation. In semileptonic decays to more than one hadron, such as B → nu l D pi, decay products after the first three are dampened in momentum by an explicit weight factor exp(-p^2/sigmaSoft^2), where p is the three-momentum in the rest frame of the decaying particle. This takes into account that such further particles come from the fragmentation of the spectator parton and thus should be soft.

When a decay mode is defined in terms of a partonic content, a random multiplicity (and a random flavour set) of hadrons is to be picked, especially for some charm and bottom decays. This is done according to a Poissonian distribution, for n_p normal particles and n_q quarks the average value is chosen as n_p/ 2 + n_q/4 + multIncrease * ln ( mDiff / multRefMass) with mDiff the difference between the decaying particle mass and the sum of the normal-particle masses and the constituent quark masses. For gluon systems multGoffset offers and optional additional term to the multiplicity. The lowest possible multiplicity is n_p + n_q/2 (but at least 2) and the highest possible 10. If the picked hadrons have a summed mass above that of the mother a new try is made, including a new multiplicity. These constraints imply that the actual average multiplicity does not quite agree with the formula above. The above multIncrease parameter, except for meMode = 23. The above multIncrease parameter, specifically for meMode = 23. Here the weak decay implies that only the virtual W mass should contribute to the production of new particles, rather than the full meson mass. The above multRefMass parameter. The above multGoffset parameter. When a decay is given as a list of four partons to be turned into hadrons (primarily for modes 41 - 80) it is assumed that they are listed in pairs, as a first and a second colour singlet, which could give rise to separate sets of hadrons. Here colRearrange is the probability that this original assignment is not respected, and default corresponds to no memory of this original colour topology. When a particle decays to q qbar, g g, g g g or gamma g g, with meMode > 90, allow or not a shower to develop from it, before the partonic system is hadronized. (The typical example is Upsilon decay.) In addition, some variables defined for string fragmentation and for flavour production are used also here.

Modes for Matrix Element Processing

Some decays can be treated better than what pure phase space allows, by reweighting with appropriate matrix elements. In others a partonic content has to be converted to a set of hadrons. The presence of such corrections is signaled by a nonvanishing meMode() value for a decay mode in the particle data table. The list of allowed possibilities almost agrees with the PYTHIA 6 ones, but several obsolete choices have been removed, a few new introduced, and most have been moved for better consistency. Here is the list of currently allowed meMode() codes: Three special decay product identity codes are defined.