Rivet for Heavy Ions introduction & tutorial Christian Bierlich, - - PowerPoint PPT Presentation

Rivet for Heavy Ions

introduction & tutorial

Christian Bierlich, bierlich@thep.lu.se University of Copenhagen Lund University February 25 2019, COST Workshop Lund

1

Before we start...

• Prepare your laptops for the tutorial while I talk.
• if experienced with rivet:
• 1. Download the latest version of Rivet from

https://rivet.hepforge.org/.

• 2. Remember to also upgrade YODA from

https://yoda.hepforge.org/.

• 3. Run with your favourite generator.
• else:
• 1. Download and install VirtualBox from

https://www.virtualbox.org/.

• 2. Load up the VM distributed on usb-sticks.
• 3. Username: mcnet, password: jetset.
• 4. Rivet 2.7.0 and Pythia 8.240 installed (+ dependencies).
• 5. Also contains small prerun samples in HepMC format.

2

Rivet

• Analysis system for Monte Carlo events. (Buckley et. al.: arXiv:1003.0694.)
• 1. Data preservation.
• 2. Monte Carlo validation.
• Generator independent, HepMC events, many analysis tools.
• C++ library with analyses as ”plugins”, optimally written by

the analyser. The biggger picture

Physics theory Phenomenological model Event generator Nature Collider experiment Detector experiment Analysis and validation Rivet 3

What is a ”rivet analysis”?

• Unfolded data + analysis code.
• Data and code is delivered in a format such that one can

easily compare to a HepMC compatible generator.

• Simple example ALICE 2010 I880049.cc.

4

Rivet for heavy ions

• Heavy Ions have traditionally not been prioritized.
• Lack of common interest (few MCs for HI).
• Lack of specialized functionality → High threshold.

5

Rivet for heavy ions

• Heavy Ions have traditionally not been prioritized.
• Lack of common interest (few MCs for HI).
• Lack of specialized functionality → High threshold.

That has changed! ⋄ Experimental community: pilot project lead by J. F. Grosse-Oetringhaus, P. Karczmarczyk, J. Klein (ALICE: CERN). ⋄ MC community: efforts by C. Bierlich, L. L¨

• nnblad (Pythia, DIPSY: Lund).

⋄ Efforts joined 2018: supported by Rivet core group and University of Copenhagen, resulting in release 2.7.0.

5

New features

• 1. Centrality selection → analysis options.
• 2. Comparing to pp → re-entrant finalize.
• 3. Flow observables → generic framework.
• 4. Several shorthand projections for specific experiments.
• 5. 20 new analyses using these features, pp, pPb, AuAu and

PbPb.

6

Centrality selection

• Centrality is ubiquitous, but not directly measurable.
• Experiment: Forward particle production/energy flow as proxy.

Cannot always be unfolded.

• MC: Not always feasible to fold prediction with ”forward

central” correlation.

7

Centrality selection

• Centrality is ubiquitous, but not directly measurable.
• Experiment: Forward particle production/energy flow as proxy.

Cannot always be unfolded.

• MC: Not always feasible to fold prediction with ”forward

central” correlation. Solution: Users’ choice between several options

• 1. Experimental measure (if existing).
• 2. Generated version of experimental measure.
• 3. Impact parameter distribution.
• 4. MC supplies centrality number.
• Three latter requires a ”calibration run”.

7

Centrality selection, calibration

• Example calibration: ATLAS PBPB CENTRALITY.
• (data points extracted from paper, not unfolded).

bbb b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b

Data MC 10−7 10−6 10−5 10−4 10−3 10−2 Sum EPb

T distribution, Pb–Pb √sNN = 2.76 TeV

(1/Nevt)dN/d ∑ EPb

T

bbb b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b

500

1.0 · 103 1.5 · 103 2.0 · 103 2.5 · 103 3.0 · 103 3.5 · 103

0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 ∑ E⊥ MC/Data MC 5 10 15 20 10−2 10−1 Sum EPb

T distribution, Pb–Pb √sNN = 2.76 TeV

b [fm] (1/Nevt)dN/db

• Generated histograms are preloaded into Rivet: new preload
• ption.

8

Centrality and Rivet options + live demo

• New Rivet functionality: Analysis options, selected at run time.
• Run the same analysis, with different options.
• Example: ALICE 2010 I880049.
• Live demo: ATLAS pPb Calib and ATLAS 2015 I1386475.

b b b b b b b b b b

Data MC [cent=GEN] MC [cent=IMP] 200 400 600 800

1.0 · 103 1.2 · 103 1.4 · 103 1.6 · 103 1.8 · 103

Nch vs. centrality, Pb–Pb √sNN = 2.76 TeV dNch/dη

b b b b b b b b b

10 20 30 40 50 60 70 80 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Centrality [%] MC/Data

b b b b b b b b b b

Data MC [cent=GEN] MC [cent=IMP] 50 100 150 200 250 300 350 400 Npart vs. centrality, Pb–Pb √sNN = 2.76 TeV Npart

b b b b b b b b b

10 20 30 40 50 60 70 80 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Centrality [%] MC/Data

9

Ratios to pp – ”nuclear modification factors”

b b b b b

Data MC 0.5 1 1.5 2 2.5 3 IAA away-side IAA

b b b b

3 4 5 6 7 8 9 10 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 pt,assoc [GeV/c] MC/Data

b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b

Data MC 10−1 1 10 1 10 2 R AA vs. p⊥, Centr = 0 − 5 %, √sNN = 2.76 TeV RAA

b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b

1 10 1 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 p⊥ [GeV / c] MC/Data

ALICE 2012 I930312, ALICE 2012 I1127497. New feature: rivet-merge

• 1. Read in histogram files, and re-generate analysis objects

(must be .yoda streamable).

• 2. Run void finalize() again.

10

Flow observables – generic framework

• Piecewise inclusion of HI observables, first: Flow coefficients

and cumulants.

• Generic framework (the flow equivalent of FastJet!) and

add-ons implemented. (1010.0233, 1312.4572).

• Functionality, calculate any Mm,n.
• Automatic subtraction of lower orders and error calculation.

11

Flow observables – generic framework

• Piecewise inclusion of HI observables, first: Flow coefficients

and cumulants.

• Generic framework (the flow equivalent of FastJet!) and

add-ons implemented. (1010.0233, 1312.4572).

• Functionality, calculate any Mm,n.
• Automatic subtraction of lower orders and error calculation.

1 hc24 = bookScatter2D("c24" ,120 ,0 ,120); 2 ec22 = bookECorrelator <2,2>("ec22",hc22); 3 ec24 = bookECorrelator <2,4>("ec24",hc24); 4 ... 5 ec22 ->fill (...); 6 ec24 ->fill (...); 7 ... 8 // c_n {4} = <<4>>_{n,-n} - 2 * <<2>>_{n,-n} 9 cnFourInt(hc24 , ec22 , ec24);

11

Sample results

• Some HI analyses implemented, here: ALICE 2016 I1419244.
• Correlators and cumulants can be plotted, also without data.
• Data not well reproduced by this MC.

MC (no data) 10 20 30 40 50 60 70 80 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 << 2 >>2,−2, |∆η > 1.| Centrality percentile [%] << 2 >>2,−2

b b b b b b b b b b

Data MC 0.02 0.04 0.06 0.08 0.1 Flow coefficient v2{2} with |∆η| > 1. v2{2, |∆η| > 1.}

b b b b b b b b b

10 20 30 40 50 60 70 80 0.2 0.4 0.6 0.8 1 1.2 1.4 Centrality percentile [%] MC/Data

12

Perspective: HI methods in pp (CMS: Evidence for collectivity in pp collisions at the LHC)

• Heavy ion methods also available for pp analyses.
• Allows for new types pp analyses in Rivet.
• Example: CMS 2017 I1471287.

b b b b b b b b b b b b b

Data MC 0.001 0.002 0.003 0.004 0.005 0.006 c2{2, |∆η > 2|} (0.3GeV < p⊥ < 3GeV) √s = 13 TeV V2∆

b b b b b b b b b b b b

20 40 60 80 100 120 140 160 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Nch (|η| < 2.4, p⊥ > 0.4 GeV) MC/Data

b b b b b b b b b b b b

Data MC 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 v2{2, |∆η > 2|} (Nch < 20) √s = 13 TeV v2{2, |∆η > 2|}

b b b b b b b b b b b

1 2 3 4 5 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 p⊥ [GeV] MC/Data

• (subtraction procedures still unclear – analyser help needed!)

13

Reference: List of analyses

Analyses with data:

ALICE 2010 1880049, PbPb: Multiplicity ALICE 2012 1127497, PbPb: Nuclear mod. factor ALICE 2012 I930312, PbPb: Di-hadron correlations ALICE 2012 I1126966, PbPb: π, K, p ALICE 2013 I1225979, PbPb: Multiplicity ALICE 2014 I1243865, PbPb: Multistrange baryons ALICE 2014 I1244523, pPb: Multistrange baryons ALICE 2015 PBPBCentrality, PbPb: Energy flow ALICE 2016 I1394676, PbPb: Multiplicity ALICE 2016 I1419244, PbPb: Flow ALICE 2016 I1471838, pp: Multistrange baryons ALICE 2016 I1507090, PbPb: Multiplicity ALICE 2016 I1507157, pp: Particle correlations ATLAS 2015 I1386475, pPb: Multiplicity ATLAS PBPB CENTRALITY, PbPb: Energy flow ATLAS 2015 I1360290, PbPb: Mult + spectra ATLAS pPb Calib, pPb: Energy flow BRAHMS 2004 I647076, AuAu: π, K, p CMS 2017 I1471287, pp: Flow LHCF 2016 I1385877, pPb: Forward region p⊥ STAR 2016 I1414638, AuAu: Flow

Analyses without data:

ALICE 2015 PPCentrality, Any: Calibration BRAHMS 2004 CENTRALITY, Any: Calibration STAR BES CALIB, Any: Calibration MC Cent pPb Calib, Any: Calib. example MC Cent pPb Eta, Any: Calib. + mult example MC OPTIONS, Any: Analysis options example MC REENTRANT, Any: Reentrant finalize example

14

What should I do now? (instead of a summary)

• This is a hands–on session, so you should get your hands dirty!
• We can stay here for a long time, pizza dinner provided.
• Use google doc (CLICKME!) to coordinate.

I am new to all this! Use the virtual machine, and run a couple of analyses. Possible goal: Using the demonstrated analysis as a template to write a simple one yourself. I am experienced! Install the newest version of Rivet, and run one or more HI analyses. Take an analysis you know, and implement it! Use existing analyses as a template.

15