Hidden Valley Benchmarks


On this page are collected the hidden-valley benchmark models that have been used or could be used in various contexts.

It will be updated quite often during 2006-2007 .





Tentative Classification Scheme for Hidden Valley Models


The input to a Hidden Valley model includes at least

It is useful to distinguish Hidden Valley models by a number of criteria which affect the experimental signatures.

For an experimental study, it is useful to decide what classes of models are desirable before searching for a corresponding benchmark, and to decide whether the goal is a study of trigger efficiency, of vertex reconstruction techniques, or of extraction of a signal from background.

For the majority of the possible values of the criteria listed, and event generators do not exist, and no benchmark models have been selected.

However, current event generators can do Z' and Higgs decays, heavy-flavor weighted final states, with large or small v-particle multiplicities, with or without long-lived v-particles, and with or without missing tranverse momentum.  This allows for a very wide variety of phenomenological signatures to be simulated already.


Back to top:


Tenative Benchmark Models


Higgs production models (PAPER II) with long-lived final states

In PAPER II the possibility of long-lived states in Higgs boson decays, and their use in discovering the Higgs, was explored.  The v-models are only one of a number of classes of models which can produce this scenario.  

Models with similar decays but with short-lived final states have been discussed actively in recent years, see hep-ph/0605162 and the references it contains, such as hep-ph/0502105.


Here are a few models designed to span the space of experimental signatures, without strong theoretical prejudice:

Production Process Higgs mass Br(h ==> SS)


Mass of S


Lifetime of S


S decay channels


All standard model Higgs production

gg ==> h, qq==>qqh, qq==>Wh, Zh

80 GeV; 130 GeV  100% ; 10 % ; 1% 20 GeV; 40 GeV  3 ns; 300 ps; 30 ps; 3 ps bb, tau tau  An 80 GeV higgs decaying to SS with lifetime less than 300 ps may well be ruled out; LEP has not published limits
All standard model Higgs production

gg ==> h, qq==>qqh, qq==>Wh, Zh

170 GeV  1% ; 0.1 % 20 GeV; 40 GeV; 65 GeV  3 ns; 300 ps; 30 ps; 3 ps bb, tau tau A Higgs that can decay to WW, ZZ cannot easily have a large Br to SS

Production of a scalar not coupling to W or Z bosons, such as CP-odd scalars in two-higgs doublet models

gg ==> A

  80 GeV; 130 GeV; 170 GeV; 220 GeV; 400 GeV; 600 GeV


100% ; 10 % ; 1%

20 GeV; 40 GeV; 65 GeV, 100 GeV 3 ns; 300 ps; 30 ps; 3 ps bb, tau tau

A Higgs that cannot decay to WW, ZZ can easily have a large Br to SS even if it is heavy. 

Note that a CP-odd A0 will typically decay to two different scalars S and A' with different CP properties, and possibly different masses and lifetimes

Production of a scalar not coupling to W or Z bosons, such as CP-odd scalars in two-higgs doublet models

gg ==> A

1 TeV 100 % 400 GeV 3 ns; 300 ps; 30 ps; 3 ps top quark pairs Long shot but fun to think about.
All standard model Higgs production

gg ==> h, qq==>qqh, qq==>Wh, Zh

80 GeV; 130 GeV  100% 8 GeV; 4 GeV  3 ns; 300 ps; 30 ps; 3 ps tau tau; rarely mu mu  THIS SET IS A TENTATIVE IDEA; IT NEEDS TO BE CHECKED FOR CONSISTENCY!!!

Other nice models along the same lines:


Back to top:



Z' production models (PAPER I)

At the present, all such models are defined by the inputs to the current event generator, which has fewer parameters than it will have in future.  The parameters currently available in the event generator are only

There are many constraints on Z' models from LEP I and LEP II; these are outlined in PAPER I.   The ratio mZ'/g' (g' the coupling constant of the Z') cannot be too large; more precisely, the mixing angle between Z and Z' cannot be larger than about 10^(-3).  This depends on whether rare Z decays to v-pions are (a) kinematically allowed and (b) visible at LEP I.

Much of the interest in v-models from the experimental community currently has to do with the long-lived particles that can arise.  A cautionary statement is necessary for the Tevatron and LHCb regarding the models discussed in detail in PAPER I:

Therefore it is still worth looking for the rare high-pT, possibly large-MET, possibly-high-multiplicity events characteristic of Z' decays to the v-sector.   In this regard, the current event simulation package may still be useful for preliminary studies, but will not give sufficiently accurate results for detailed studies. A new version of the package will allow for the simulation of a wider variety of models.



These issues should be kept in mind when looking at the the choice of benchmark models given below.


The following benchmarks are designed to stretch the limits of the triggering system at ATLAS; they are preliminary and await comment.

Zprime Mass (GeV) V-pi-zero mass (GeV) V-pi-zero lifetime (ps) V-pi-plus lifetime (ps) comments
2000 30 100 infinity Note the Z' mass is not 3 TeV! The lower Z' mass allows a larger cross section (or order 10 fb) for the same pizero lifetime, and also gives lower energy events with softer jets and less MET, for which the trigger may be more challenged.  I think the issues of non-pointing clusters, muons which do not point back, etc, will already arise in this case.
2000 20 5000 infinity Again the cross-section is of order 10 fb. This will be very challenging!!  The hard vpions escape the detector; the ones in the detector are mostly soft.  Maybe this choice of parameters is too difficult; let's see.
3000 60 5 infinity Here triggering should be straightforward, with everything pointing back to the primary vertex to a good approximation; if not, that's important toknow. Indeed many of the vertices are near or inside the beampipe and should be rather easy to find.
3000 150 0.01 100 Here we will have much less MET, many more jets, and a very busy environment in which to find the vertices, which now have large opening angles.  I have no idea what will happen here.
3000 150 100 infinity This is an interesting experimentally because it is similar to the previous case, but with more MET and less activity, and with large opening angles for the highly-displaced vertices. 


This is simple to understand: the higher mass of the vpizero increases the phase space for its decay, and  the only way to make it long-lived is to increase the mass of the Z', which then means there's no v-quark production rate.

However, one can have this kind of scenario if the vpions are produced  in decays of an LSP; in that case, the Z', which mediates the vpion decays, could be 10-100 TeV in mass.  And there are other scenarios too that I'm sure I can cook up with a little effort. So you might want to study this case at some point, even though the current event generator would be giving you an unrealistic model.

Back to top: