Recoil Analysis Page

Page maintainer: Caro

This is a working page which will continuously change. The content is not reviewed and is intended for people working on this specific topic.

Back to Recoil Detector main page

Introduction

This is page intended for people performing a physics analysis combining data from the Recoil Detector and the spectrometer. The page is intended to compile all kinds of useful information needed for such an analysis, and it is presently (status: February 2009) still in a developing phase. Please add missing pieces, or correct wrong information. This page is open for suggestions!

Meetings

Exclusive Meeting: every Tuesday 13:10 in 1b/5 (unless announced differently in an e-mail).
TRM: (used to be) every Tuesday 13:00, seminar room 55a
Recoil Meetings: (used to be) monthly
For more meetings, see DVCS Analysis Page.

Recoil Data Productions and Statistics

Information about available main and uDST data productions including Recoil data is compiled on the Recoil Data Productions page.

Recoil Monte Carlo Productions

2009/2010 productions: see [054]-[056] on MC Productions page
August 2008 MC productions: see [022]-[027] on MC Productions (these productions are compatible with 07c1 uDSTs)
Example shell script how to run your analysis code over hmc productions in a pipe, without producing huge hrc files or alike, and how to send the script to the PCfarm (no guarantee given, you might have to play around with the pathes). Advantage: you can give any flags to XTC you like.
old html page (not updated anymore)

Recoil Commissioning

Recoil Calibration

Some documentation.

SSD

SFT

Directory with Weilin's SFT calibration code, including a README file:

/group02/rcoilgrp/user/yuwl/SFT_Calibration

Weilin's mail May 2010

PD

SSD Flex Foil Corrections

Mail by Jonathan 7.10.2009

Recoil PID

Information about proton-pion separation with the Recoil detector is compiled on the Recoil PID page.

Heavy gas contamination

Internal Note 09-023, Jennifer Bowles October 2009

Recoil Data Quality

Run level

Time dependence of Recoil detector quantities. Code documentation
General
- SFT: no general run restrictions.
- PD: after run 14244 in 2006
- SSD: after run 27771 in 2006 on (SSD timing final setting, Sept. 2006)
- Data Quality run lists
  - SSD:

/group02/rcoilgrp/DATA/silicon/daqllist_06.txt
/group02/rcoilgrp/DATA/silicon/daqllist_07.txt

- - SFT:

/group02/rcoilgrp/Software/magic_stuff/epio2xy/SFT_DataQuality_2011Feb

- - PD run list:

/production/opa/udstprod/udst_07d1/runlists/pd/pd_goodlist.dat

Burst level

Bits 27 (SSD), 29 (SFT) and 31 (PD) in the 2006 and 2007 badburstlists are reserved for Recoil data quality on burst level. They will be activated in the uDST productions based on the final 06f and 07d productions. See Feb.2011 presentation. The bits are based on the run level lists.
SFT TDC information:
- Bit pattern in XTC-format data, also described here
- Preliminary summary of Irina's TDC studies, July 20, 2010
One strip of the PD was not connected during 2007 for HERMES run 5207 until 5371 and 33327 until 33359.

Track level

SSD data quality (using hrc format), also described here

Analysis Algorithms and Techniques

Important note: This page has not yet been reviewed by an expert! Use the information below with caution.

Other pages that have not been reviewed yet can be found in the category FORREVIEW.

This section has (not yet) been reviewed; still, it compiles the best knowledge of the group as of the end of 2010. (CR 17.12.2010)

Reading Recoil track information from uDST data

Docu

Contents of uDST Recoil tables
Organisation of Recoil data: schematic picture, flowchart of data structure

Tracks in the Recoil Detector

The recommendation is to constrain analysis to tracks with g1RDTrack.iMethod==7 (M7)
Recoil tracking (XTC)
Energy in SSD related to tracks: the energy in the SSD related to tracks in the RD is available via the g1SSDTr table. One has to navigate from g1RDTrack to this table. g1SSDTr contains a field 'iLayer' which is either 1 or 2 and a field 'rPuls' which is the energy. The information on untracked hits is stored in g1SSD.

Spacepoints

M7 tracks have spacepoints (SP) in either 4, 3 or 2 out of the layers (outer SFT, inner SFT, outer SSD, inner SSD).
The spacepoint hit pattern (xxxx) for each track is obtained from g1RDTrack.bHitP (to be read from right to left!).
- 4 SP: (1111)
- 3 SP: (0111) or (1011) or (1101) or (1110)
- 2 SP: (0011)

PID

If available, read the rdPID values g1RDPID.rPID[i], i=1...6, for 3 and 4 SP tracks (ADAMO navigation link from g1RDTrack)
According to the rdPID, decide if the track is a pion or a proton (see also Recoil PID).
- If no rdPID available, the track is considered to be
  - a negatively charged pion (then only the Pion link should exist and its momentum should have negative sign) OR
  - a proton if it is a 2 SP track (then at least one of the Proton links should exist)

Track parameters $\left.(p,\theta ,\phi ,\mathrm {ZVx} )\right.$

- If available, read g1RDTrack.Proton parameters in table g1RDTrackPar, unless the chi2 of the track fit is equal or greater 100
- If Proton not available, or if the chi2>=100, read g1RDTrack.StoppedProton parameters in table g1RDTrackPar
- Else, read Pion parameters
Note: the charge of the track can be accessed already in g1RDTrack via g1RDTrack.bHitP

Event selection

Choose your favourite spectrometer track configuration (e.g. 1 lepton and 2 oppositely charged hadrons for exclusive Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \left.\rho_0\right.} -analysis).
Kinematic fitting.
- A preliminary description by Caro is given here from the DVCS analyzer's point of view. More information is given in this section. Sergey, you might want to change this link to an official location if you decide to write a wiki or LaTexed docu.
- The kinematic fitting function should be called for every Recoil track in an exclusive event candidate, using as input a complete set of four-momenta of the event. A Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \chi^2} value is given back by the function, indicating how well the kinematics of the plugged in four-momenta match the hypothesis assumed during the event reconstruction. The analyzer can determine the Recoil track with the smallest Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \chi^2} in the event. For example, in the DVCS with Recoil group, the pure elastic sample is obtained by allowing only those events for which the smallest Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \chi^2} in the event is smaller than a certain value (as of end of 2010, this value is 13.7, which corresponds to a fit probability of larger than 1 percent). The complementary data sample, the DVCS background sample, is obtained by allowing only those events for which the smallest $\chi ^{2}$ in the event is larger than the certain cut value. Note that as of the end of 2010, no rdPID information is used in selecting the Recoil track with the smallest Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \chi^2} , i.e. the track is not required to be a proton.
An alternative and less advanced method (that was used to demonstrate first correlations between Recoil detector and front Spectrometer) is the method of one-dimensional cuts. It does, other than kinematic fitting, not take into account the complete set of kinematic parameters at the same time. Its advantage is that the analyzer does not have to call an external function.
The Recoil track sample can be cleaned up using certain cuts. In order to decrease the Recoil event multiplicity, one can try to reject tracks from the rdtrack buffer...
- whose z-vertex is not in correlation with the z-vertex of the scattered beam lepton.
  - this requires reading spectrometer data before Recoil data
  - this cut might not be useful for all kinds of physics analyses
- which have similar kinematics Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \left.(p,\theta,\phi)\right.} as another M7 track in the event (clone track).
  - Caro's algorithm to find clones
    - In the loop over all Recoil tracks, loop for each track over all tracks stored in g1RDTrack before this track (skipping the considered one) and calculate the difference in each P, Theta and Phi. Keep the smallest differences DeltaP, DeltaPhi and DeltaTheta.
    - If abs(DeltaP) < 0.4 GeV/c && abs(DeltaPhi) < 0.55 rad && abs(DeltaTheta) < 0.2 rad, mark as clone track. Otherwise, mark as no-clone. This algorithm was tested on MC (gmcDVCS). About 65% of the true clone M7 tracks (i.e. generated tracks that are multiple reconstructed) are found by these settings, while about 2.7% of all reconstructed clone M7 tracks are wrongly tagged as clones. Altogether, 14.3% of all reconstructed M7 tracks are tagged as clones. Various cuts were tried and these numbers were chosen as optimized ones. Numbers as of: 20.2.2011.
    - Some Delta-distributions are shown in Caro's "suspicious" track studies February 2009. The cut values indicated in these slides are less strict in theta and p (before tuning on MC).
- which have a bad Fehler beim Parsen (MathML mit SVG- oder PNG-Rückgriff (empfohlen für moderne Browser und Barrierefreiheitswerkzeuge): Ungültige Antwort („Math extension cannot connect to Restbase.“) von Server „https://wikimedia.org/api/rest_v1/“:): {\displaystyle \chi^2} of the track fit (the effectiveness of this cut is questionable, though).
- Remember that for uDST data, there is no spacepoint and cluster information available.
- Before the Recoil track multiplicity has not been not significantly reduced, it might be wiser to not require exactly one Recoil track. Rather look to kinematic correlations by comparing measured and expected kinematics (and/or geometry -> coplanarity!) of your favourite physics process, or in the advanced version, by kinematic fitting.

Kinematic Fitting

/group02/rcoilgrp/Software/KinFit/

Use this link to find details on the event selection.

Recoil Acceptance

Recoil Analysis Tools is a library with additional functions for the analysis of recoil detector information. It is written in c and can be linked to any Hanna based program. It includes a parameterization of the Recoil acceptance in (ZVx,Theta) and (phi,p) extracted from Monte Carlo data (background generator).
More detailed description of the Recoil acceptance by Andreas Mussgiller (available on request, C++ format)

Recoil Efficiency

The efficiency in the Recoil layers can be studied by running XTC with the flag --removelayer n (n=1,2,3,4). Spacepoints in layer n are then not considered in the tracking. In 2009, Xianguo Lu calculated efficiency maps on cluster level (SFT fiber number vs. log(beta*gamma)) from 06e and 07c data.

Plan: generate MC sample with efficiencies implemented (also provide detector efficiency to users (as part of Recoil Analysis Tools)?). Compare distributions and asymmetry amplitudes with and without inefficiency implemented.

Documentation by X. Lu on SFT efficiency
- Efficiency files located at:

/user/luxg/public/SFT_EFF/efffile_20090718

Analysis Topics

DVCS

Analysis

DVCS with Recoil Analysis Page NEW

Talks

[DVCS-004] Talk on 'Extended Snalysis Report' DVCS with Recoil, C. Riedl, Collaboration Meeting Oct.2010
[DVCS-003] Status of DVCS analysis, W. Yu, Recoil meeting talk 22.4.2009
[DVCS-002] Status of DVCS analysis, W. Yu, Recoil meeting talk 13.2.2009 (contains a review of the standard cuts)
[DVCS-001] Recoil Analysis with the 07c1 uDSTs, C.Riedl, Recoil meeting talk 13.2.2009

Exclusive Vector Meson Production

Exclusive pi0 Production

Spectator Proton Tagging