Dark showers meeting
(9 November 2020)
Indico (including recording): https://indico.cern.ch/event/970758/
Discussion today: results of the survey circulated on the mailing list
So far 8 replies and include 17 people together with master students
Three directions
SUEP
Semi-visible jets
“prompt” jets: (means where invisible fraction is very low (= “dark jets”))
Wish to understand commonality between Knapens SUEP plugin and the HV implementation in pythia
Pythia implementation extensions:
Add heavier resonances?
Add dark baryons?
First proposal for working plan is made:
Study pheno and variables for benchmarks used for LHC searches
Identify discriminating variables specifically for semi-visible jets
Connect SUEP to jettier final state structures (connects to 1)
Organisationally
Collect material on common github community
Review resources and organise the work
Kevin: having looked at both SUEP and HV, there is no connection between them. SUEP implements thermal boltzmann style distribution. In pythia for SM alternative fragmentation model called thermal model exists, how close is it to SUEP plugin?
Sarah: Even if the actual implementation is completely different it might be interesting to see how similar or different the particle distributions are? (for extra dimension model)
Marie-Helene: one of the next meeting, we would like to also discuss SUEP (we could invite S. Knappen directly)
Suchita: Isn’t SUEP a limiting case of SU(N) with large N?
Kevin: SUEP works for large t’Hooft coupling, this can be large either when the N is large but even otherwise. There is no reason to expect the overlap between HV and SUEP module.
-- recording of the meeting starts
One of the first papers: Yang Bai, scale of dark QCD.
Collider benchmarks:
S-channel, t-channel mediators
Depending on what the constituents of the jets are doing
Emerging
Semi-visible
Prompt, fully visible jets
What happens in QCD? Is there a dark gluon, and what is the coupling structure (dark QCD or a strong Yukawa coupling)?
The phenomenology of dark gluon radiation would give us 3 dark jets (dark showers). Depending on event-level variables we may also have access to the gluon spin.
Primary proposal: study two and three dark shower/jet events
There are questions in terms of generation / analysis selection:
matching and merging for dark gluon jet not understood
Two dark quark jets events may be too boosted (due to hard dark gluon radiation)
Distinguish dark gluons from QCD partons
Once those are sorted, use the Ellis-Karliner angle (kT jet algorithm) to determine the spin of the gluon.
Will study first lepton collider environment as it’s cleaner.
Q1: Suchita: The issue of matching and merging was discussed in the first meeting therefore it is extremely nice to see it being tackled in your LOI. How do you plan to actually do this?
A1: We have Stefan Prestel on LOI! We will take his help.
Q2 [Caterina]: how much variation is there in the dark shower pheno of the dark gluon? Will it always follow the dark parton? (eg a dark quark leading to an emerging jet = a dark gluon leading to an emerging jet)
A2: Pedro Schwaller says yes.
Q3: matching & merging: why not doing it the same way in MG+Pythia?
A3: the problem is having one extra QCD-like jet
Kevin: the MG models used so far don’t handle the additional radiation
Pedro: try to convert the gluon to the hard gluon, and then do the showering in Pythia. This is “cheating” and may lose the correlations.
Sarah: can we extend MG this way? Useful tool.
Chih-Ting: we tried also to change dark to ordinary QCD, but it wasn’t possible.
Pedro: Feynrules? Maybe Madgraph can do it, but it’s a problem of Feynrules? Don’t know exactly.
Suchita (answering Sarah): would be nice to identify things like these and other showstoppers and then work on those together. Eventually invite an MG/Feynrules author to discuss when we know what is the showstopper.
How was the study developed:
Look at variables for jets closer to the MET in background and signal
Select a slice in PT
Main points and observation of the study:
Signal is more-pronged than background
Model-dependence (turning invisible part off): not much difference in changing alpha_s
Some substructure variables do discriminate at truth level, but they may deteriorate when adding pileup/detector resolution
Questions:
Kevin: why not study the more realistic case of a running coupling?
Deepak: HV alpha_s was taken with no running, but could be studied - but unless you go to higher orders, changing lambda wouldn’t change anything?
Sarah: changing to running would definitely change things eg like the multiplicity in the showers - maybe Kevin could give cards showing how to implement running or not
Kevin: http://home.thep.lu.se/~torbjorn/pythia82html/HiddenValleyProcesses.html
alphaOrder: 0 uses alphaFSR, 1 uses Lambda
Meaning of the Λ value will also depend on the masses of the dark hadrons
Suchita: we would like to see running of alpha_fsr and changing the masses of the dark quarks => effects on the jets. Are you using only pythia and not MG at all?
Deepak: using also MG, model file from Tim Cohen
Suchita: how does your study get affected if you have a different underlying model of the hidden sector?
Deepak: we haven’t tried, but think the conclusions should hold, we would be happy to try if there are other models available
Marie-Hélène: we will contact you offline so that the cards/model files can be put in github for others to try / discuss how they could be modified for interesting studies such as the running coupling
Emerging jets model: Bai, Schwaller, Stolarski (1502.05409, 1803.08080)
Main characteristics:
Confining dark sector with energy scale lambda_dark
Lighter dark pions (among dark hadrons) with mass around lambda_d and lifetime
Decays back to SM → emerging
Different models depending on Yukawa:
Unflavored: Yukawa for only one SM quark flavor, so that the dark mediator can decay only in one dark quark → one lifetime
Flavored: generic Yukawa coupling, can decay into any of the dark quarks → multiple (4) lifetimes
Implementation is done in Pythia using the HV models. See slides for parameters modified by user in cards, and cards are in agenda ( → TODO: move to git)
Note that the flavored, aligned scenario needed Pythia modifications to allow for the flavored dark mesons. Put decays by hand and hardcode the combinations.
Q [Suchita]: I like the flavored model. If you have a non-diagonal scenario, don’t you always generate light jets?
Sarah: you get into a hybrid between semi-visible and emerging jets. The fraction that decays promptly becomes even bigger for the non-aligned models. This is a different search between flavored and unflavored.
Suchita: you get 1 emerging jet or 2?
Yi-Mu: in the current investigation, we look for 2 EJ and 2 regular jets.
Pedro: non-diagonal model are more likely to have a b or a top in the SM jet, so might have even longer lived ‘regular jets’ as they contain b mesons.
Suchita: what if the mediator were produced with a boost, how light can we go for the mediator?
Pedro: colored states, so they can’t be too light, limit is about 1 TeV. If you have a Z’ mediator then you’re more flexible.
Kevin/Sarah: we may be starting to look at the Z’ mediated EJ, already looking at SVJ.
Marie-Hélène: thanks for posting cards and links to indico already. We will propagate this to github.
An email from James Beacham, LLP community:
...we're featuring two working-group sessions dedicated to dark showers on Tuesday 17 November next week as part of LLP8, the LLP Community workshop:
https://indico.cern.ch/event/922632/timetable/
discussions will be about SUEP-to-jets and astrophysics / cosmology meets dark showers, so highly relevant to this group.