Date: Fri, 28 Sep 2001 02:16:25 -0400 (EDT) From: esub-adm@aps.org To: prltex@ridge.aps.org Cc: slee@fnal.gov Subject: [Web] resub LF8481 Affolder (es2001jun18_063) -- Subject: LF8481 Manuscript code: LF8481 RECVD: Fri Sep 28 02:16:23 2001 Resubmission to: Physical Review Letters Resubmission type: resubmit Figures: all changed figures included New files with this resubmission: prl_v41_preprint.tex 09-28-2001 Deleted files from previous submission: prl_v31_preprint.tex 09-28-2001 Replaced files: lsdilepton_prl_fig1.eps 09-28-2001 lsdilepton_prl_fig2.eps 09-28-2001 Details of changes: Sep. 27, 2001 Dear PRL editor: Enclosed is response to reviewers' comments/questions on our paper. Please send this to each reviewer. Sincerely Sungwon Lee ---------- Forwarded message ---------- Date: Tue, 31 Jul 2001 09:11:31 -0400 (EDT) From: Physical Review Letters <prl@ridge.aps.org> To: slee@fnal.gov Cc: prl@ridge.aps.org Subject: lf8481 Dr. Sungwon Lee Texas A+M Univ. MS 318 Fermi National Accelerator Lab. P.O.B. 500 Batavia, IL 60510 slee@fnal.gov Re: LF8481 Search for gluinos and squarks using like-sign dileptons in ${p}$${\bar{p}}$ collisions at $\sqrt{s}$=1.8 TeV By: T. Affolder, H. Akimoto, A. Akopian, M.G. Albrow, P. Amaral, et al. Dear Dr. Lee, The above manuscript has been reviewed by our referee(s). We ask you to consider the enclosed comments from the report(s). While we cannot make a definite commitment, the probable course of action if you choose to resubmit is indicated below. (X) Acceptance, if the editors can judge that all or most of the criticism has been met. (X) Return to the previous referee(s) for review if available. ( ) Submittal to new referee(s) for review. Please accompany any resubmittal by a summary of the changes made, and a brief response to all recommendations and criticisms. Sincerely yours, Samindranath Mitra Assistant Editor Physical Review Letters Email: prl@aps.org Fax: 631-591-4141 -------------------------------------------------------------------- Report of Referee A (LF8481 Affolder,T) -------------------------------------------------------------------- This paper gives the first results on a search for like-sign dileptons in events with 2 jets and missing transverse energy from the Tevatron. Since the Tevatron is the world high energy frontier, and since this channel has especially small backgrounds from standard model processes, it is very interesting, and deserves publication in Physical Review Letters. However, before publication, I would like the authors to adress the following comments: ====================== 1) They list 4 ee, 10 emu, and 5 mu mu events passing the OS cuts. They list the backgrounds as: Drell-Yan 8.7, ttbar 4.0. other backgrounds are small. Assuming all 10 observed emu events are top and not DY, and assuming ee+mumu=emu from top, this leave no room for their Drell-Yan background? But, they list this as their dominant background? [A] Referee A's concern is a large size of Drell-Yan (DY) backgrounds and a connection with 10 emu events. With MET > 25 GeV, we have a total of 19 (4 ee, 10 emu, 5 mumu) events. The DY (including Z production) events are dominant (8.7 for ee+emu+mumu) in our analysis, because: a) Our Pt cuts for leptons are lower, compared to those in the top dilepton analysis (see Phys. Rev. Lett. 80, 2779 (1998) by the CDF collaboration). We have 11 and 5 GeV for 1st and 2nd leptons, while 20 GeV for both leptons in the top analysis; b) With our detector resolution for the MET measurement, there are many DY events passing our MET>25 GeV cut; c) A dominant source of emu events is Z->tau tau. With our lepton Pt cuts, we substantially accept OS ee/mumu events from DY gamma* production and OS ee/mumu/emu events from Z(->tau tau) production even after the MET cut. Based on our MC studies, we expect N(ee) ~ N(mumu) ~ N(emu) ~ 3 from DY gamma* and Z production. Since the final state (dilepton + jets + MET) is the same as in the top dilepton signature, we have compared our events with the CDF top dilepton events in Phys. Rev. Lett. 80, 2779 (1998). There are a total number of 9 events (1 ee, 7 emu, 1 mumu). Of 9 events, 6 events (1 ee, 5 emu, 0 mumu) are in our analysis (before the LS cut). The other three events (3 emu) are rejected because of difference in its lepton isolation criteria between our analysis and the top analysis. From our MC study, we expect 4.0+-0.3+-1.2 events (compared to 6 observed). This is a good agreement. Subtracting 6 observed top event candidates from 19 events, we are left with 13 events (3 ee, 5 emu, 5 mumu). Our MC predicts 10.1 non-top background events (see Table II of our paper). This is again a good agreement. We note the event acceptances for ee+mumu events and emu events are different for the ttbar events. This is because we impose the Z removal for e+e- and mu+mu- events. Thus, the reviewer's assumption (ee+mumu=emu) is not valid for this particular analysis. We also added sentences into a paragraph starting with "Given the large $\met$ signature from SUSY, we require ..." to provide information of the top dilepton event candidates (so that the readers can understand our emu events better). The added sentences are: "We note that the 19 event sample also contain six dilepton (1 $ee$, 5 $e\mu$) events out of nine $\ttbar \to W^+ b W^- \bar{b} \to (\ell^+ \nu b)(\ell^- \bar{\nu}\bar{b})$ event candidates (1 $ee$, 7 $e\mu$, 1 $\mu\mu$) from the CDF $\ttbar$ analysis in the dilepton channel \cite{cdf_topdilepton_prl}. The three top dilepton event candidates are not in our final sample because our lepton isolation requirement for the second lepton is stricter than the top analysis." ====================== 2) I had trouble understanding the description of the SUSY model. I think the sentance "We impose common scalar and gaugino masses..." should be "We impose common scalar and common gaugino...". the first time I read this, I thought they set the scalar and gaugino masses equal at the GUT scale. Also, I'm not sure what the phrase " slepton masses which require m(squark)>=0.9m(gluino)..." means. From their description, I guessed that they took M(squark), ran it up the the GUT scale, set the slepton mass equal to the squark mass at the GUT scale, and then ran the slepton mass back down. But, then I don't understand the "require M(squark...)" bit. [A] The text was not well written. We changed to "We impose common scalar and common gaugino masses at a GUT scale as in the minimal supergravity model~\cite{sugra}, and use the renormalization group equations~\cite{RGE} that relate the mass parameters, leading to a general prediction: $\msquark\ \gtsim\ 0.9 \mgluino$. As Referee A pointed out, we set the scalar masses equal at the GUT scale and ran down to the electroweak (EWK) scale using RGEs. RGEs give relations between sleptons, squarks, and gluinos at the EWK scale, and have no solution if the squark mass is lighter than about 90% of the gluino mass. ====================== 3) They quote the tracking resolution, but not the electron or jet resolutions. This seems strange. I think they should quote all three. [A] Very good suggestion. However, we intend to show only the momentum resolution because it provides an idea of how precisely the CDF detector can measure the sign of charge, which is appropriate in this analysis. Thus we gave an approximate momentum resolution. Actual numbers on tracking resolution and energy resolution of various calorimeters can be found in other references, so that we would like to keep the current sentence. ====================== 4) Their result would be of more general interest if they tried to present their results in as model-independent way as possible. For example, they should quote the kinematic acceptance and the total lepton identification efficiency separately for several different squark/gluion masses. A theorist could calculate a kinematic acceptance for their own favorite model, and use the total lepton identification efficiency quoted as a rough guide, to translate this result into rough limits on their own model. [A] We agree it would be nice if we show all numbers. We tried to do this in an earlier version during the preparation of this paper. However, because of the line counting, we removed details and referred to Ref. 14 that provided the lepton ID efficiencies: Tight Cuts Loose Cuts Electrons: 85% 89% Muons : 93% 94% (|Eta|<0.6), 93% (0.6<|Eta|<1.0) As for the kinematical/geometical acceptance for our dilepton+dijet+MET selection, we have given information of the total event acceptance for M(gluino)=200 GeV in the text: A = 3% for M(gluino)=M(quark)=200 GeV A = 1% for M(gluino)=200 GeV, M(squark) >> M(gluino) More details can be found in Ref. 14. This is not an ideal set of the model-independent information, but we believe it gives a complete information. ====================== 5) are the jet ET cuts before or after jet energy scale corrections? [A] In our paper, all Et or Pt cut values are after full correction(s). Otherwise, we stated. ====================== 6) they should update reference 11 to Phys. Rev. Lett 83 (1999) 4937 and update figure 2 to include the results in this reference. [A] We were aware of this nice paper by the D0 collaboration. We did not include this in our paper, because the D0 analysis include all SUSY production (gluino/squark pair production, associate production such as gluino+chargino, and chargino/nutralino pair production, but excluding stop production), making hard to extract the squark and gluino mass limits from the its DIRECT production. In constrast, we only used the production of gluinos and squarks (excluding stop) to set mass limits on gluino and squarks. Therefore, we would like not to add PRL 83, 4937 (1999) into our references as direct searches of gluino and squark production. With the same reason, we would like not to modify Figure 2. ====================== 7) this result was first presented in conference in 1996 (see fermilab-conf-96/181-E), so I'm not sure this result is really "timely". [A] Since the 1st presentation of the preliminary result in 1996, we have been working very carefully on evaluation of low Pt lepton trigger efficiencies, various systematics, understanding the SM backgrounds, and cross-check with the other CDF analyses such as the CDF top dilepton analysis. These are main reasons why it took several years. -------------------------------------------------------------------- Report of Referee B (LF8481 Affolder,T) -------------------------------------------------------------------- Using events containing like-sign di-leptons, jets and missing $E_T$ to search for gluinos and squarks in $p\overline{p}$ collisions represents a useful addition to the literature on direct experimental searches for SUSY. I judge the analysis to have been competently performed and the paper is, in general, clearly explained. This work certainly deserves to be published in PRL. ====================== 1) I do, however, have one question concerning a feature of the data presented that looked surprising. 19 OS data events survive all cuts, a number that is consistent with the 14.1 expected from Standard Model sources. However 10 of these events are $e\mu$. The number of $e\mu$ events expected is not given. (I think it would be a good idea to give the flavour breakdown of the expected SM events.) But I would guess this number is approximately half the number expected from sources other than Drell-Yan, i.e., $2.7\pm0.5$. In making this guess I am assuming roughly equal identification probability for electrons and muons, which is obviously an approximation. On the face of it this looks like a four sigma discrepancy. Do the properties of these events (lepton $p_T$, missing $E_T$, jet activity) show any unusual features? I would suggest that some comment on the number and properties of the observed $e\mu$ events should be made in the paper. [A] Referee B's concern is a large size of Drell-Yan (DY) backgrounds and a connection with 10 emu events. With MET > 25 GeV, we have a total of 19 (4 ee, 10 emu, 5 mumu) events. The DY (including Z production) events are dominant (8.7 for ee+emu+mumu) in our analysis, because: a) Our Pt cuts for leptons are lower, compared to those in the top dilepton analysis (see Phys. Rev. Lett. 80, 2779 (1998) by the CDF collaboration). We have 11 and 5 GeV for 1st and 2nd leptons, while 20 GeV for both leptons in the top analysis; b) With our detector resolution for the MET measurement, there are many DY events passing our MET>25 GeV cut; c) A dominant source of emu events is Z->tau tau. With our lepton Pt cuts, we substantially accept OS ee/mumu events from DY gamma* production and OS ee/mumu/emu events from Z(->tau tau) production even after the MET cut. Based on our MC studies, we expect N(ee) ~ N(mumu) ~ N(emu) ~ 3 from DY gamma* and Z production. Since the final state (dilepton + jets + MET) is the same as in the top dilepton signature, we have compared our events with the CDF top dilepton events in Phys. Rev. Lett. 80, 2779 (1998). There are a total number of 9 events (1 ee, 7 emu, 1 mumu). Of 9 events, 6 events (1 ee, 5 emu, 0 mumu) are in our analysis (before the LS cut). The other three events (3 emu) are rejected because of difference in its lepton isolation criteria between our analysis and the top analysis. From our MC study, we expect 4.0+-0.3+-1.2 events (compared to 6 observed). This is a good agreement. Subtracting 6 observed top event candidates from 19 events, we are left with 13 events (3 ee, 5 emu, 5 mumu). Our MC predicts 10.1 non-top background events (see Table II of our paper). This is again a good agreement. We note the event acceptances for ee+mumu events and emu events are different for the ttbar events. This is because we impose the Z removal for e+e- and mu+mu- events. Thus, the reviewer's assumption (ee+mumu=emu) is not valid for this particular analysis. We also added sentences into a paragraph starting with "Given the large $\met$ signature from SUSY, we require ..." to provide information of the top dilepton event candidates (so that the readers can understand our emu events better). The added sentences are: "We note that the 19 event sample also contain six dilepton (1 $ee$, 5 $e\mu$) events out of nine $\ttbar \to W^+ b W^- \bar{b} \to (\ell^+ \nu b)(\ell^- \bar{\nu}\bar{b})$ event candidates (1 $ee$, 7 $e\mu$, 1 $\mu\mu$) from the CDF $\ttbar$ analysis in the dilepton channel \cite{cdf_topdilepton_prl}. The three top dilepton event candidates are not in our final sample because our lepton isolation requirement for the second lepton is stricter than the top analysis." ====================== 2) There was one particular aspect which I thought could have been more clearly explained in the paper. Some of the SUSY signal channels discussed (those involving the decay $\chi^0_2 \rightarrow \ell^+\ell^- \chi^0_1$) have to lead to events containing three leptons if they are to produce two LS leptons. Experimentally, of course, all three leptons will not necessarily be identified as isolated and high $p_T$. I think this should be mentioned in the paper. This prompts the questions (which I think should be answered in the paper): how are events with three observed isolated leptons treated in the selection and are any such events observed in the data? [A] Thank you for pointing out that our paper does not describe the treatment of the 3rd lepton. In our analysis we search for dilepton events inclusively, and then check the charges for the two leading-Pt leptons. We modified a paragraph starting with "The analysis begins..": "The analysis begins with a sample of 515,699 loosely selected dilepton events~\cite{thesis,TRIL1A_1B} from which we select an initial dilepton plus dijet sample. To ensure that the trigger is fully efficient, we require each event to have a lepton with $\pt \ge 11$~\gevc\ and $|\eta|<1.0$ for electrons or $|\eta|<0.6$ for muons. A second electron or muon is required with $\pt \ge 5$~\gevc\ and $|\eta|<1.0$. If there are more than two isolated leptons, we take two leading-\pt\ leptons. Each lepton is required to be isolated ....." ====================== I had a few more minor questions/suggestions: 3) In the sentence at the bottom of page 9: ``We define the acceptance as the ratio of the number of di-lepton events \ldots\ which contain at least two leptons.' I assume the ``cuts' include the requirement of LS, but the number of generated events includes both OS and LS events? Perhaps this could be spelled out. [A] Thanks for pointing out this. We changed to ``We define the acceptance as the ratio of the number of LS dilepton events that pass our cuts to the total number of generated SUSY events which contain at least two leptons.' (We simply added "LS" in the text.) ====================== 4) In table~I I would have been interested to see the SUSY numbers for the first two rows. Is it not possible to give, also, the SM background number for the second row? [A] a) SM number for the 2nd row: In the 2nd row, the data still contains c-cbar/b-bbar resonance events, such as J/psi and Upsilion, which are not in our MC generator (ISAJET) unless we use a specific generator separately. Using the specific generator in addition to ISAJET will require additional (and substantual) effort to understand systematics between two generators. Unfortunately, we have not done such studies. Thus, we showed our comparion between data and MC events from the 3rd row (namely, after the resonance events are effectively removed by M_{ll} > 12 GeV). b) SUSY number for 1st and 2nd rows: We lost the original information. However, we can provide approximate number of dilepton events produced at 106 pb-1 for the SUSY events. Standard Model SUSY Dilepton produced --- about 200 events Dilepton Dataset --- --- Dilepton-Dijet --- --- M(ll) > 12 GeV 279+-9+79 27 +- 1 +-5 Z veto 158+-7+-45 27 +- 1 +-5 MET > 25 GeV 14.7+-1.3+-2.8 24 +- 1 +-5 LS Dilepton 0.55+-0.25+-0.08 5.9+-0.6+-1.4 We note that "Dilepton Dataset" is after the dilepton events pass our trigger requirement. ====================== 5) $\chi^0_2 \rightarrow \ell^\pm\ell^\mp \chi^0_1$ (middle of page~1) seems unnecessarily complicated when $\chi^0_2 \rightarrow \ell^+\ell^- \chi^0_1$ would do. [A] Yes, you're absolutely right. We changed to what you suggested.