MonikaSharmaSandbox < Sandbox/SandboxArchive

Sandbox/SandboxArchive Web>TWikiUsers>MonikaSharma>MonikaSharmaSandbox (2020-08-20, TWikiAdminUser)

Response	As you rightly pointed out that this is a nice measurement and an important result. We definitely wouldn't want to be perceived as having cut any corners to rush into a journal publication and be proved wrong later on. Moreover, we would never push to have a publication which has a figure that would make CMS look bad. Quite the contrary, we believe that we are achieving a significant heavy ion physics result using a meson probe 1000 times less massive than the Higgs boson for which the ECAL system was primarily designed. People should be impressed at the extensive dynamic range being provided by the CMS detector. It has been suggested privately that we quantify the conversion correlation background level using HYDJET simulation events, essentially extending what we have shown in the analysis note. Then, the shape of this simulation background could be added to the shape of the data-driven mixed-event background. Please refer to Section 10 of version 11 of the AN, available at http://cms.cern.ch/iCMS/jsp/analysis/admin/analysismanagement.jsp?ancode=HIN-11-009. In this Section we demonstrate that the negative yield exists when converted photons are not rejected. We will also be making a presentation to the ARC. Please refer to https://twiki.cern.ch/twiki/pub/CMS/HIN-11-009/ARCpresentationJuly2012.pdf. Please refer to slides 6-8 in the presentation where results from HYDJET simulations are presented. On the basis of this we propose to change Fig. 1 (as suggested by Gunther Roland). We now propose to make a four panel figure instead of two panels. Two new panels will now show data from HYDJET simulations with conversion photons rejected. Shown as panel three (c) which demonstrates that the problem of negative yield goes away when such a requirement is made. In the fourth panel (d), we compare combinatoric background subtracted pi0-meson invariant mass distribution from real data and HYDJET simulations. Converted photons are not rejected in order to make an apples-to-apples comparison with the real data. One can see that within statistical uncertainties HYDJET simulations results agree very well with the real data. It is also being suggested that another technique for generating the background be used instead of mixed events. This technique is to randomly combine photon clusters in the same event, but with the second cluster member having a flipped azimuthal angle compared to its actual position. The authors also demonstrate in the AN as well as in the presentation to the ARC that this technique, as expected does not work. The problem of negative yield still persists which we understand arise due to real correlations in the same-event distribution from converted photons. Please refer to slides 3-5 in the presentation or Section 10 of the AN. Section 6 and 7 in the AN (http://cms.cern.ch/iCMS/jsp/analysis/admi/analysismanagement.jsp?ancode=HIN-11-009) demonstrate we did spent a substantial amount of effort understanding the origin of the negative yields in the higher mass regions which makes the subtracted mass distribution imperfect. We are reconstructing pi0 mesons in a complex heavy ion with the additional burden of having a very high photon conversion rate. This conversion rate leads to a correlated (same-event) background on a scale which was not faced by the RHIC experiments. Despite this, the results are numerically robust, and the systematic uncertainties not so large that these preclude us from making an important conclusion that the baryon-meson anomaly persists at the LHC. These results are an important counterpart to the unidentified charged particle v2 results being published by CMS. Now, we are not trying to sweep the conversion correlation effect under the blanket of systematic effects. Pages 88 - 92 lists various sources of systematic effects and their percentages in different pT ranges and centrality. http://cms.cern.ch/iCMS/jsp/analysis/admin/analysismanagement.jsp?ancode=HIN-11-009. Quoting one example: for the most central case, 20-30% and the highest pT interval 6.0 < pT < 8.0 GeV/c, systematic uncertainty from S4/S9 ratio is actually much larger than the asymmetric mass integration region which includes the effect in subject. In passing we point out that we have studied carefully the EGM-10-003 AN (AN2010_216_v1.pdf) which was prepared in support of the ICHEP 2010 presentation. It is our intention (see below) to learn more from the pi0 reconstruction in pp events to improve our analysis in the heavy ion events, and we will rely on this AN or its successor. The correct, data-driven procedure for quantifying the same-event conversion photon correlations, in both pp and PbPb collisions, is to identify the secondary electrons or positrons which are entering the ECAL and being paired with a presumably unconverted partner photon. This should be straightforward to accomplish in the pp collisions. In fact, one of us initiated a thread on the ECAL performance hypernews just before this CWR started with the goal of learning how to do exactly that study: https://hypernews.cern.ch/HyperNews/CMS/get/ecal-performance/575.html. The extension of the study to the heavy ion events will not be straightforward. First, the heavy ion tracking software sacrifices the tracking of displaced vertex tracks (e.g. the conversion electrons) because the density of particle tracks in central collision heavy ion events is so large. Second, the reverse tracking of electrons using seeds from the ECAL into the silicon similarly does not exist in the current heavy ion reconstruction software. At least this second method will have to be validated in the heavy ion reconstruction software in order to do a study of conversion electron correlations in the actual data. Once this is done, a re-reco pass will be necessary on the 2010 minimum bias heavy ion data set to get the necessary objects in the RECO file. That step will be actually the easiest part since the three authors are also the principle support group for the Heavy Ion Tier 2 at Vanderbilt. The above outline of work is clearly far longer than a few weeks. It could easily take into 2013, especially given the fact that the group has to prepare for the Quark Matter 2012 conference, and after that prepare for the next heavy ion (pPb) run. This means getting new hardware commissioned for the Tier 2 and taking care of our Tier 0 prompt reconstruction responsibilities. Two of the three authors will be going to CERN a month before pPb run just for that purpose. As we show below, all of this work will not actually lead to a major reduction in our systematic uncertainties. In fact, the major purpose of the work be to extend the range of the pi0 measurement beyond a pT = 8 GeV/c by doing gamma + e+ + e- triple coincidences. ALICE has shown that the four body channel (both photons convert) can lead to publishable physics results. CMS, with its much bigger ECAL system, should be able to do the three body study. We have done extensive investigations in order to understand the effect conversion correlations have on our analysis, and will do more as suggested. We adamantly do not intend to cut corners, but we the scale and the cost/benefit of any new studies to be well understood in advance.

Response

As you rightly pointed out that this is a nice measurement and an important result. We definitely wouldn't want to be perceived as having cut any corners to rush into a journal publication and be proved wrong later on. Moreover, we would never push to have a publication which has a figure that would make CMS look bad. Quite the contrary, we believe that we are achieving a significant heavy ion physics result using a meson probe 1000 times less massive than the Higgs boson for which the ECAL system was primarily designed. People should be impressed at the extensive dynamic range being provided by the CMS detector.

It has been suggested privately that we quantify the conversion correlation background level using HYDJET simulation events, essentially extending what we have shown in the analysis note. Then, the shape of this simulation background could be added to the shape of the data-driven mixed-event background. Please refer to Section 10 of version 11 of the AN, available at http://cms.cern.ch/iCMS/jsp/analysis/admin/analysismanagement.jsp?ancode=HIN-11-009

. In this Section we demonstrate that the negative yield exists when converted photons are not rejected. We will also be making a presentation to the ARC. Please refer to https://twiki.cern.ch/twiki/pub/CMS/HIN-11-009/ARCpresentationJuly2012.pdf. Please refer to slides 6-8 in the presentation where results from HYDJET simulations are presented. On the basis of this we propose to change Fig. 1 (as suggested by Gunther Roland). We now propose to make a four panel figure instead of two panels. Two new panels will now show data from HYDJET simulations with conversion photons rejected. Shown as panel three (c) which demonstrates that the problem of negative yield goes away when such a requirement is made. In the fourth panel (d), we compare combinatoric background subtracted pi0-meson invariant mass distribution from real data and HYDJET simulations. Converted photons are not rejected in order to make an apples-to-apples comparison with the real data. One can see that within statistical uncertainties HYDJET simulations results agree very well with the real data.

It is also being suggested that another technique for generating the background be used instead of mixed events. This technique is to randomly combine photon clusters in the same event, but with the second cluster member having a flipped azimuthal angle compared to its actual position. The authors also demonstrate in the AN as well as in the presentation to the ARC that this technique, as expected does not work. The problem of negative yield still persists which we understand arise due to real correlations in the same-event distribution from converted photons. Please refer to slides 3-5 in the presentation or Section 10 of the AN.

Section 6 and 7 in the AN (http://cms.cern.ch/iCMS/jsp/analysis/admi/analysismanagement.jsp?ancode=HIN-11-009

) demonstrate we did spent a substantial amount of effort understanding the origin of the negative yields in the higher mass regions which makes the subtracted mass distribution imperfect. We are reconstructing pi0 mesons in a complex heavy ion with the additional burden of having a very high photon conversion rate. This conversion rate leads to a correlated (same-event) background on a scale which was not faced by the RHIC experiments. Despite this, the results are numerically robust, and the systematic uncertainties not so large that these preclude us from making an important conclusion that the baryon-meson anomaly persists at the LHC. These results are an important counterpart to the unidentified charged particle v2 results being published by CMS.

Now, we are not trying to sweep the conversion correlation effect under the blanket of systematic effects. Pages 88 - 92 lists various sources of systematic effects and their percentages in different pT ranges and centrality.

http://cms.cern.ch/iCMS/jsp/analysis/admin/analysismanagement.jsp?ancode=HIN-11-009

. Quoting one example: for the most central case, 20-30% and the highest pT interval 6.0 < pT < 8.0 GeV/c, systematic uncertainty from S4/S9 ratio is actually much larger than the asymmetric mass integration region which includes the effect in subject.

In passing we point out that we have studied carefully the EGM-10-003 AN (AN2010_216_v1.pdf) which was prepared in support of the ICHEP 2010 presentation. It is our intention (see below) to learn more from the pi0 reconstruction in pp events to improve our analysis in the heavy ion events, and we will rely on this AN or its successor.

The correct, data-driven procedure for quantifying the same-event conversion photon correlations, in both pp and PbPb collisions, is to identify the secondary electrons or positrons which are entering the ECAL and being paired with a presumably unconverted partner photon. This should be straightforward to accomplish in the pp collisions. In fact, one of us initiated a thread on the ECAL performance hypernews just before this CWR started with the goal of learning how to do exactly that study: https://hypernews.cern.ch/HyperNews/CMS/get/ecal-performance/575.html

.

The extension of the study to the heavy ion events will not be straightforward. First, the heavy ion tracking software sacrifices the tracking of displaced vertex tracks (e.g. the conversion electrons) because the density of particle tracks in central collision heavy ion events is so large. Second, the reverse tracking of electrons using seeds from the ECAL into the silicon similarly does not exist in the current heavy ion reconstruction software. At least this second method will have to be validated in the heavy ion reconstruction software in order to do a study of conversion electron correlations in the actual data. Once this is done, a re-reco pass will be necessary on the 2010 minimum bias heavy ion data set to get the necessary objects in the RECO file. That step will be actually the easiest part since the three authors are also the principle support group for the Heavy Ion Tier 2 at Vanderbilt.

The above outline of work is clearly far longer than a few weeks. It could easily take into 2013, especially given the fact that the group has to prepare for the Quark Matter 2012 conference, and after that prepare for the next heavy ion (pPb) run. This means getting new hardware commissioned for the Tier 2 and taking care of our Tier 0 prompt reconstruction responsibilities. Two of the three authors will be going to CERN a month before pPb run just for that purpose.

As we show below, all of this work will not actually lead to a major reduction in our systematic uncertainties. In fact, the major purpose of the work be to extend the range of the pi0 measurement beyond a pT = 8 GeV/c by doing gamma + e+ + e- triple coincidences. ALICE has shown that the four body channel (both photons convert) can lead to publishable physics results. CMS, with its much bigger ECAL system, should be able to do the three body study.

We have done extensive investigations in order to understand the effect conversion correlations have on our analysis, and will do more as suggested. We adamantly do not intend to cut corners, but we the scale and the cost/benefit of any new studies to be well understood in advance.

-- MonikaSharma - 16-Jul-2012

Topic revision: r3 - 2020-08-20 - TWikiAdminUser

Sandbox/SandboxArchive

Public webs

- Cern Search
- TWiki Search
- Google Search
Sandbox/SandboxArchive All webs

Copyright &© 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback