Antideuterons at LHCb
Sophie Bakeron behalf of the LHCb collaboration
Imperial College London
Antideuteron 2019March 27 - 29
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 1 / 16
Antideuterons at the LHC
As seen this morning, antideuterons have alreadybeen observed in high energy collisions at the LargeHadron Collider by ALICE.
ALICE acceptance: |η| < 0.8, pT < 5GeV/c
A measurement complementary to this can bemade at LHCb.
LHCb acceptance: 2 < η < 5, 2 < p < 100GeV/c
This measurement will be useful for estimations ofsecondary antideuterons as the background toindirect dark matter searches.
[ALICE Collaboration,Phys. Rev. C 93, 024917 (2016)]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 2 / 16
LHCb detector
Forward single-arm spectrometer,measuring collisions of pp collisionsat the LHC.
Designed for the study ofheavy-hadron decays, precise chargedparticle identification (PID) is vital.
There is also a program of heavy ioncollisions, and fixed-targetmeasurements.
[Int. J. Mod. Phys. A 30 (2015) 1530022]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 3 / 16
LHCb detectorForward geometry of LHCb is chosenfor better observing b-hadron decays.
Identification of particles (PID) isdone by combining outputs from:
• vertex locator around collisionpoint
• ring imaging Cherenkov (RICH)detectors
• electronic and hadroniccalorimeters
• tracking stations
• muon stations [Int. J. Mod. Phys. A 30 (2015) 1530022]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 4 / 16
Particle identification at LHCb
High energy charged tracks areidentified using two RICH detectors.
Cherenkov thresholds (GeV/c):
RICH 1 RICH 2
π 2.3 7.0K 9.3 15.3p 17.7 29.7d 35.4 59.3
[Int. J. Mod. Phys. A 30 (2015) 1530022]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 5 / 16
Particle identification at LHCb
High energy charged tracks areidentified using two RICH detectors.
Cherenkov thresholds (GeV/c):
RICH 1 RICH 2
π 2.3 7.0K 9.3 15.3p 17.7 29.7d 35.4 59.3
K p dπ
[Eur. Phys. J. C 73 (2013) 2431]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 6 / 16
Particle identification at LHCb
Cherenkov angle and trackmomentum are used in the RICHalgorithm to give likelihoods for thetrack to be a charged particle,i.e. e, µ, π, K, p, or d.
Each ring in the detector is comparedto hypotheses for rings from differentparticles, given the knownmomentum of the track.
[CERN-THESIS-2014-102]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 7 / 16
Measurements at LHCb
Antideuterons can be measured at LHCb in prompt production in pp collisions, indecays of heavy-hadrons, and in fixed target collisions.
I will focus on measurement of the prompt antideuteron cross-section in pp collisionswith centre-of-mass energy
√s = 13TeV.
LHCb can be used to make the first measurement of antideuterons in b-hadron decays,and measurements in pHe collisions will provide unique constraints to the productioncross-section in cosmic-rays, similar to the equivalent antiproton measurement,[Phys. Rev. Lett. 121, 222001 (2018)].
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 8 / 16
Deuteron production simulation
The coalescence and cross-section modelsfor deuteron production are included inLHCb simulation.
Following the work by Dal & Raklev, thethreshold for coalescence, and thenormalisation of the cross-sections weretaken from measurements of deuterons atALICE in pp collisions at
√s = 7TeV.
Momentum difference [GeV/c]1−10 1
b]µ d
X)
[→ 2
N 1(Nσ
1
10
210
310Coalescence
γ d →pn 0π d →pn
0π 0π d →pn -π +π d →pn
Based on [Phys. Rev. D 91, 123536 (2015)]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 9 / 16
Coalescence model
[Kapusta (1980)] p and n from each ppcollision are compared to each other; anypair which flies within a momentumseparation cone of 190MeV/c will bind.
Assuming that p and n have the same massand same pT spectra, d production given by
Edd3Nd
dp3d= Bd
(Ep
d3Np
dp3p
)2
,
where Bd =(
4π3mp
p30
), with p0 the radius of
the momentum sphere for coalescence.Momentum difference [GeV/c]
1−10 1
b]µ d
X)
[→ 2
N 1(Nσ
1
10
210
310Coalescence
γ d →pn 0π d →pn
0π 0π d →pn -π +π d →pn
Based on [Phys. Rev. D 91, 123536 (2015)]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 10 / 16
Cross-section model[Dal & Raklev (2015)] pp, pn and nn pairsfrom each event are compared;for each possible production channel, theprobability that they will bind is found, as afunction of momentum separation in theircentre-of-mass frame.
Production cross-section as a function ofmomentum separation in CoM frame, k,
σ(k) =akb
(c− exp(dk))2 + e,
where a, b, c, d, e are best fit parametersgiven in the paper.
Momentum difference [GeV/c]1−10 1
b]µ d
X)
[→ 2
N 1(Nσ
1
10
210
310Coalescence
γ d →pn 0π d →pn
0π 0π d →pn -π +π d →pn
Based on [Phys. Rev. D 91, 123536 (2015)]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 11 / 16
Deuteron production simulation
Momentum [GeV/c]0 20 40 60 80 100
310×
/(4
GeV
/c)
π/N
XN
4−10
3−10
2−10
1−10
1
K
p
d Cross-section
d Coalescence
LHCb preliminary, simulation
The expected deuteron yield in pp collisionsLHCb is low:
−→ one deuteron per 10,000 charged pions
Backgrounds to the measurement will comefrom p, K and π.
To have a significant deuteron signal, the PIDneeds to be effective and known to a very highprecision.
[Link to plots]
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 12 / 16
Neural net for deuteron identification
Improved performance of PID using neuralnets (NN).
NN inputs: RICH likelihoods, kinematicvariables and tracking information
ProbNNd = deuteron neural net variable
Instead of performing a cut-basedselection, a template fit can be used toextract the deuteron yield.
ProbNNd0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
#Eve
nts
[arb
itrar
y un
its]
5−10
4−10
3−10
2−10
1−10
1dpKπghost
LHCb preliminary, simulation
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 13 / 16
Neural net for deuteron identification
Template fit will be performed in bins oftrack momentum.
Neural net is dependent on momentum,due to the RICH thresholds for Cherenkovlight.
Thresholds (GeV/c):
RICH 1 RICH 2
π 2.3 7.0K 9.3 15.3p 17.7 29.7d 35.4 59.3
p [MeV/c]0 10 20 30 40 50 60 70 80 90 100
310×
< P
robN
Nd
>
0.0
0.2
0.4
0.6
0.8
1.0
dpKπghost
LHCb preliminary, simulation
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 14 / 16
Template fit to ProbNNd
Template shapes for charged tracks will betaken from data samples, and thedeuteron and ghost shapes taken fromsimulation.
The best sensitivity is at very high valuesof ProbNNd, so the bin boundaries arechosen such that the template fordeuterons is flat.
ProbNNd(d) percentile0 10 20 30 40 50 60 70 80 90 100
#Eve
nts
210
310
410
510
610
710 toy d
π ghost
K µp e
LHCb preliminary, simulation
Pull
2−1−012
Template fit, 29.7 < p < 35.4GeV/c
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 15 / 16
Summary
Measurement of antideuterons at LHCb will provide information about deuteronproduction at higher energies and more forward pseudorapidities than have been probedat other experiments.
Though LHCb was not designed with this type of measurement in mind, it will bepossible due to its excellent PID and the flexibility of the data taking.
Antideuteron measurements in b-hadron decays and in pHe collisions will complementthe prompt measurement, and provide contraints for DM searches.
Thanks for listening!
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 16 / 16
ProbNNd inputs
Kinematic p, pT
RICH Used RICH 1 gas, used RICH 2 gas, above π, K, p thresholds, log likelihoodfor e, µ, K, p, d hypotheses, log likelihood for being below threshold
Tracking χ2track/ndf , track ndf, ghost probability, track fit match χ2/ndf , track fitχ2V ELO, track fit ndfV ELO, track fit χ2
Tracker, track fit ndfTracker
Muon Muon log likelihood, non-muon log likelihood, #muon shared system hits,binary ‘is muon’ selection, looser ‘is muon’ selection, muon system accep-tance selection
Calorimeter EM calorimeter e and µ ID, hadronic calorimeter e and µ ID, preshower eID, Bremsstrahlung acceptance and e ID
Sophie Baker on behalf of the LHCb collaboration Antideuterons at LHCb Antideuteron 2019 1 / 1