Single Diffractive Cross Section - PowerPoint Presentation

Slide1

Single Diffractive Cross Section

Calorimeter Gaps Approach

Soft QCD Meeting

14th Oct 2010

Tim Martin, Paul NewmanUniversity of Birmingham

1

Slide2

Soft Diffraction

2

Diffractive Events

Colour singlet

exchange. Can be Single or Double

proton dissociation. Diffractive mass can be anything from

p+π0 up large systems with hundreds of GeV

invariant mass. Soft PT spectrum. Large forward energy flow.

Less activity in the inner detector.

Pythia

Single Diff.

Transverse

Momentum

Transfer

Slide3

Diffractive Mass Distributions

3

Pythia8

Measure dσ

/dξ over the ξ range accessible at ATLAS

Constrain

the

relative weights of

SD

,

DD

and

ND

At

7 TeV

we should be getting

really big

diffractive systems!

Slide4

Truth Level Energy

Flow

4

Normalised to a single event in ξ.

Differential

in

η

.Expect

rapidity gap devoid of all soft activity

.

Single Diffraction

Linear

relation between

η

of leading particle and log(

ξ

), slope = 1 for

ln

(

ξ

).

Part

hadronisation

, mostly kinematics.

η

max

=

η

of

Leading Truth Particle in X System

Are there any MC diffractive models which do

not

use the lund string hadronisation model?

Slide5

5

Trigger Efficiency

Trigger events with the

MBTS. Trigger is fully within calorimeter acceptance so use loose trigger requirement.

L1_MBTS_2.

`Here be dragons’

Unexploreable

phase space with Calo + MBTS analysis.

Correct for trigger inefficiency in data using MC here

Slide6

6

Experimental Re-

definiton

In

Double Diffraction, if the smaller system (MY

) is too small to get into the ATLAS calorimeters, then the event is

experimentally indistinguishable (exc. ZDC) from a single diffractive event. This occurs

around ξY of 10-6

Reclassify as Single Diffractive.

Name the new sample

Single Diffractive Like

New Definition:

Double Diffractive Like

Systematic Check

: Location of cut

Slide7

7

New Cross Sections

Single

and Double Diffractive Like take into account the migration of events from the Double Diffractive to Single Diffractive

samples. Visible refers to the

subset of the samples which can be triggered with the MBTS.

Slide8

8

Samples & Calo Cells

Current data sample of

~750k, Pythia8, Phojet, Pythia6; ND, SD, DD

MC Empty

MinBias and RNDM stream for 1

st 7 TeV run:

Run 152166~

600k

L1_MBTS_2

&

BCID

=

1

Events

7719

mb

.

Noise in

TILE

show non-Gaussian profile (double Gaussian).

TILE is not required for hermetic coverage of ATLAS and is

currently excluded

from the analysis.

mc09_7TeV.105003.pythia_sdiff.recon.ESD.e514_s764_s767_r1307/

mc09_7TeV.105004.pythia_ddiff.recon.ESD.e514_s764_s767_r1307

mc09_7TeV.105001.pythia_minbias.recon.ESD.e517_s764_s767_r1307

data10_7TeV.00152166.physics_MinBias.recon.ESD.r1239/

data10_7TeV.00152166.physics_RNDM.recon.ESD.r1239/

mc09_7TeV.106096.PhojetNdiff.recon.ESD.e514_s764_s767_r1307

mc09_7TeV.106097.PhojetSdiff.recon.ESD.e514_s764_s767_r1307

mc09_7TeV.106098.PhojetDdiff.recon.ESD.e514_s764_s767_r1307mc09_7TeV.108316.Pythia8_minbias_ND.recon.ESD.e533_s764_s767_r1307mc09_7TeV.108317.Pythia8_minbias_SD.recon.ESD.e533_s764_s767_r1307

mc09_7TeV.108318.Pythia8_minbias_DD.recon.ESD.e533_s764_s767_r1307

Oldrich

&

Pavel

EM Barrel

Slide9

9

EM Barrel

EM

Endcap

Inner

EM

Endcap

Outer

FCAL

HEC

Tile

Calo

Cell

Significance

Data10 MBTS_2

Data10 RD0_EMPTY

MC Py8 Inelastic

MC Empty

Very Bad

Slide10

10

Cell Requirement

Using an

Energy VetoCurrently require 1 cell above 5.5 E/σ

.Can perhaps improve our sensitivity by combining cells?

No good for all cells in a slice.

E.G Barrel slice:

39,925 cells

RMS Noise

of all cells combined:

8.6 GeV!

Swamps any minimum bias energy deposition.

However, combination of cells with E/

σ

> 3 looks more promising.

Order 50

cells

per

η

slice, RMS noise of ~

100

MeV

–

manageable

.

Under current investigation, may improve systematics.

Slide11

Gap Finding

11

Largest

η Gap

E.G.

Single

Diffractive

Event

M

X

M

Y

η

•Trigger on

L1_

MBTS_2.

•

Negligible

pileup

, no

prescales, low beam backgrounds

.

•Divide the calorimeter into

10x

rings spanning 1 unit of η

.

•Ring

is

full if 1 or more calo cells satisfy E/σ

Cell

>

Threshold

•Look

for

largest continuous gap of empty rings

.

•Plot

Start of Gap

[|η|]

vs. Size of Gap [Δη].

•The

start of the gap is defined as the

|

η| of it’s farthest edge from η=0

•Provides

separation

of diffractive and

non-diffractive.

MBTS

Being Quantified

Systematic Check

Slide12

12

Overall Trigger Eff.

Trigger efficiency calculated from

MC on a bin-by-bin basis.Fractional breakdown of event types in each bin displayed below.

Overall

Trigger

Efficiency

MC

Bin-by-bin

Breakdown of

Event types

Key

Totally Full

Detector Bin

Totally Empty

Detector Bin

Slide13

13

Tr. Eff Correction

Trigger

efficiency corrections are applied bin-by-bin to the data histogram.Bins with a MC trigger efficiency of < 50 % are not used in the fit.These are the `here be dragons bins’

Before Correction

After Correction

Low Efficiency

Slide14

14

Formation of MC Templates

ND

SD

DD

Map

2D distribution onto 1D

plot for the purpose of fitting.

Pre-trigger MC distributions

.

Normalise

each

MC shape to a single event.

Bins with overall trigger efficiency < 50% are plotted last –

beyond the black line.

These bins are not included in the fit.

Slide15

15

Check of Fit

Form pseudo-data by applying bin-by-bin Poisson smear

to MC prediction.Fit relative SD, DD and ND templates to pseudo-data.Fit constrained to sum to 1

One Trial

Residuals for one trial

1000

Trials

Slide16

16

Fit to Data – Pythia8

Fit to

data normalization in fitted bins.Log likelihood fit.MC shapes do not fully describe the data. Poor χ

2

Pythia8:

ND:

73.6%

SD Like: 23.7%

DD Like: 2.7%

Fit:

ND:

74.2%

SD Like: 18.1%

DD Like: 7.74%

Slide17

17

Detector Acceptance

To convert to a

cross-section, need to correct back to the particle level.Run gap finding algorithm over truth level data.

Move from Calo Cells 

Truth ParticlesMove from E/

σ cut 

PT cut

Build

bin-by-bin correction histogram to apply to data

.

c.f

the trigger efficiency

correction.

Correction factors manageable

in

Single Diffractive Like

(

edge gap

) region.

Large correction factors

in

Double Diffractive Like

region (

central gaps

)

Slide18

18

Single-Diffractive

Comparison of algorithms at:

Truth (PT > 200

MeV

)

Reco

(E/σ > 5.5)

Hatched boxes denote located gaps.

Correct effects such as:

Low P

T

particles not reaching

callo

.

Multiple scattering

.

Slide19

19

Cross Section

This slide 0

MeV PT cut gap definition

Data is... Corrected for Trigger Efficiency &

for Detector Acceptance.

Scaled to Lumi of run

.Plotted systematics, `best and worse case scenarios’

E/

σ

> 5 noise

AND

+

10% Cell energy scale

AND

111%

Lumi

.

E/

σ

>

6

noise AND -10% Cell energy scale AND 89% Lumi

.

Last two bins have Trigger Efficiency < 50%

Suggestion from R. Barlow! Will check adding in quad as well.

Systematic Band

(systematics invert)

Default & fitted

fractional visible

Xsec

MC lines

Statistical error

negligible

Slide20

20

Cross Section

This slide 100

MeV PT cut gap definition

Interesting sensitivity to choice of PT

cut defining cross section.

Most effect at small gaps, explained as increasing pollution from non-diffractive events.

Flick back and forth...

Slide21

21

Cross Section

This slide 200

MeV PT cut gap definition

Flick back and forth...

Slide22

22

Cross Section

This slide 500

MeV PT cut gap definition

Flick back and forth...

Work ongoing to match a sensible P

T

cut with the E/σ cut.No particles in the gap (P

T

> 0 cut) looks sensible.

Can also publish differential in P

T

Slide23

23

Cross Section

Combination into a single result.

Pythia8 taken as trigger and acceptance correction MC

Maximum deviation per data point for Pythia6 and Phojet taken as systematic.Added in quad.

, last two (low trigg

. Eff.) bins not plotted.

0

MeV

P

T

Definition

100

MeV

P

T

Definition

Slide24

Conclusion

24

Single diffractive like cross section d

σ/dη = 1 mb η-1 Measurement of the single diffractive like cross section at 7 TeV possible with very little minimum bias data.

Many

cross checks, systematic checks, systematic quantification

is underway.

Still a bit more work to do with Calo Cell selection &

P

T

dependence on acceptance

correction

.

We are working to

compare with

Tevatron

results at 2 TeV

.

We are writing, detailed internal

note for the analysis being produced at

the moment.

Details theory, all steps & cross checks.