LOGO A partridge in a pear tree Simulation of multi-jet processes using the BFKL event generator Rasmus Mackeprang
Jan 17, 2016
LOGOA partridge in a pear tree
Simulation of multi-jet processes using the BFKL event generatorRasmus Mackeprang
Conventional picture of collision
Full matrix element for each final state incalculable
Parton showersParton showers effectively
resums part of the full perturbative series (all orders).
Standard (DGLAP) showering treats collinear part of phase space
Collinearemissions
Matrix element
Normally 22
Two turtle doves
Consequences
Number of hard jets limited by the order to which the matrix element is calculated.
At the LHC there is a non-vanishing phase space for non-collinear emissions
Are we under-estimating our SM background in the multijet channels?
Matrix element
Collinearemissions
Three French hens
Alternative approach
BFKL formalism resums to all orders terms of
Sij is the invariant mass of emissions i and j
ti is a time-like momentum between them
We can investigate to all orders the probability of hard jet emissions.
Large rapidity differences enhance dynamics.
€
αS logSijti
+K ⎛
⎝ ⎜
⎞
⎠ ⎟~ α SΔy ij
i
j
Four calling birds
Jet production
Count “hard” jets in the eventPick the two rapidity-wise extreme jets Fixed order can only give you jets according
to the order of the calculationAt high rapidities BFKL will give more hard
jets
Δy
njets
3
20
4
NLO
BFKL
Five golden rings
Angular decorrelation
Dijet events to LO will have Δφ=0Parton showers will smear this Look at hard jets onlyBFKL should show larger
decorrelation at high rapidity differences
Δφ
Δy
<cos(Δφ)>
1
00
Six geese a-laying
Multijet rates
With fixed order calculations you typically show 3/2 jets rates because you cannot treat higher orders.
Multijet rates at high rapidity differences should show differences between standard approach and BFKL.
Seven swans a-swimming
Parton level results
BFKL MC generator developed by Jeppe Andersen (CERN)
Weighted MCNo hadronizationKt jets with R=0.6Pythia8 vs BFKL (easy to run on a laptop) Looked at dijets and W+jets (We ν)
Eight maids a-milking
Well, Pythia only really
does W+jet…
Jet production
Used pseudo-rapidityHard jet has
Et > 40 GeV
|η| < 4.5ME cut is 20 GeV Little difference in
dijet eventsW+jets an unfair
comparison
Nine ladies dancing
Dijets
W+jets
Angular decorrelation
Low rapidity differences favour Pythia’s collinear emissions
Otherwise compatible for dijets
As for W+jets…
Ten lords a-leaping
W+jets
Dijets
Multijet ratios
Rates are “n or higher”Slightly higher BFKL
multijet rates Effect not stronger at
high eta gaps, though.
Eleven pipers piping
W+jets
Dijets3j/2j
4j/2j
Exclusive rate ratios
Ratios are“n/(2 or higher)”
Largely the same conclusions
W+jets
Dijets3j/2j
4j/2j
Twelve drummers drumming
Step back…
Seems BFKL is rather close to Pythia for dijets
DGLAP in turn seems to do a decent jobATLAS uses Pythia6. This was Pythia8 Taking Kt4H1TopoJets in J-samples we can
make a (very) rough comparison
A dozen and a partridge in a pear tree
Pythia8 vs Pythia6
A dozen and two turtle doves
Jet production in Alpgen
W+2j W+3j
W+4j W+5j
Order by order more jets are produced (well, duh…)
Samples are MLM matched
Can be added by integrated luminosity.
A dozen and three French hens
Accentuating the matrix element
Exclusive rate-ratios order by order
One sees clearly the extra jets entering
W+2j W+3j
W+4j W+5j
3j/2j
4j/2j
A dozen and four calling birds
Grand finale…
Adding Alpgen samples by integrated luminosity
The Alpgen prediction Some agreement
between BFKL and Alpgen
BFKL produces more jets, though
Consistent with missing virtual corrections in Alpgen
An order of magnitude more Alpgen stats after christmas…
A dozen and five golden rings
Last comments
BFKL is fast (1000 times faster than Pythia) It reproduces dijets and agrees with parts
of the pQCD W+jets predictionsW+jets is important background to BSM The harder we kick Jeppe the faster he
works (LSHA interface, unweighted events)So, am I the only one who thinks this is
interesting?What if I say Higgs+jets? LSHA and unweighting done there
A dozen and six geese a-laying
Technicalities and references
BFKL PDF: MRST 2004 NLOOn the BFKL MC Method:
hep-ph/0602182 (Phys.Lett. B639 (2006) 290) hep-ph/0101180 (JHEP 0102:007,2001) hep-ph/9706529 (Phys.Rev. D56 (1997) 5875-5884) hep-ph/0305236 (Phys.Lett.B567:116-124,2003) hep-ph/0309331 (Nucl.Phys.B679:345-362,2004)
On Parton Density Functions: hep-ph/0410230 (Phys.Lett. B604 (2004) 61-68)
A dozen and seven swans a-swimming