U. M. Meller Senior Assistant Editor Physical Review D Email: prl@aps.org Dear Dr. Meller, We would like to thank the referee for their time, effort and thoughtful reading. We are glad to find a positive response to our 2nd draft. We have included a set of PACS numbers in the new manuscript. Enclosed you will find a revised draft of our manuscript. Below you can find detailed responses to the referees comments. Regards, P. Wagner and D. Toback =============================================== ------------------------------------------------------------------------- Report of the second Referee: ------------------------------------------------------------------------- > In their work the authors present a very useful study on signatures > which typically occur in gauge mediated supersymmetry breaking, but > which are more general in nature. Because I am not an experimentalists > I will not consider the technical details of the study, which look > solid to me. Instead, I am going focus on the phenomenological > aspects. The paper looks very sound, I have only very few comments > which I would like the authors to address: > > * Could the authors please elaborate on one of their main issues, > namely the use of EMTiming? For a phenomenlogist, a more detailed > discussion of this issue for example for Figs.6,7 would be > useful. Moreover, in their conclusions they could add a few more > sentences about the effect of EMTiming on the cross section limits > and on the coverage of the GMSB parameter space. A significant portion of the text describes the use of the EMTiming information and its impact on the results. Figures 6 and 7 show the 95% CL cross section limit results as a function of the neutralino mass and lifetime. The dashed curve shows the limit with kinematical cuts only, the solid line for cuts which use the timing information as well. For both with and without timing system usage the event selection requirements, the expected background and optimization procedure for the gg+MET analysis (for Fig. 6) and the g+MET+jets analysis (for Fig. 7) are shown in Tables II and III respectively. As discussed in sections IIb.3 ("In these plots we see four trends: ...") the analyses which include timing information lower the cross section limit throughout the whole neutralino mass and lifetime range. We explain that the improvement in cross section limit is largest at high neutralino lifetime and low mass as there the kinematical cut has the least effect compared to the timing cut and timing provides a good separation between prompt and signal photons. The improvement ranges from a factor of 1 to 7 (for GMSB production 1 to 13 respectively) for lifetimes up to 60 ns, further increasing at higher lifetimes (see Figs. 8 and 12). Efficiency questions are addressed in Section IIa and Appendix A. We are open to specific suggestions to make the text more clear about the usage of the EMTiming system, but we are not sure where more detail would be helpful. We have added more information in the Conclusion section about the effect of the EMTiming system on the coverage in GMSB parameter space. The text now reads: "For a given mass, as the lifetime increases, more of the neutralinos leave the detector and the overall sensitivity goes down. But timing provides additional rejection power and allows for significant exclusions even at large lifetimes and improves the cross section limits by a factor of 5-10 at lifetimes around 50 ns. [...]" > * p.4. Strictly speaking, the neutralino lifetime \tau is not > defined. More importantly, for the SPS points discussed in Section > III it is not clear which parameters are fixed, which are varied, > and how the values for the neutralino life time is obtained in > relation to the SPS parameter points. We agree that these aspects are kept short in our paper, since they are described in detail elsewhere. We refer to Ref. 5 for the definition of neutralino lifetime and Ref. 15 for the Snowmass parameters. As described there we follow Eur. Phys. J. C25, 113 (2002), and take the messenger mass scale M_{M}=2 * \Lambda, tan(\beta)=15, sgn(\mu)=1 and the number of messenger fields N_{M}=1. The parameters c_{Grav} (gravitino mass factor) and \Lambda (supersymmetry breaking scale) are varied to obtain the neutralino lifetime and mass which are a function of m_Gr (see Ref. 5), and \Lambda respectively, if all other parameters are left constant. This is made more explicit in Ref. 15. > * p.5. How critical are the assumptions on the systematics listed in > Tab.1? The authors briefly talk about that later, but it would be > very sueful to add a few sentences either here or later in the > discussion. For example: what would change if we knew the > backgrounds better/worse by a factor of -say- 50%? The values given in Table I are estimated using conservative assumptions from previous searches at the Tevatron (see Ref. 9, 10 and 11). So, the results shown in Fig. 13 are our best estimates of what we will likely exclude (assuming no signal). We believe that background errors of more than 50% might not be realistic estimates for coming analyses, so we have not mentioned this study explicitly. However, for better understanding we have added a forward reference to the end of section IIIc, which covers the dominant problems which would significantly affect our results, namely the cosmic ray background and the timing resolution. The uncertainty on the background, estimated to be 30% is already incorporated in the limit. > * p.5. The authors refer to some details of the analysis in > Ref.[19]. It is not clear at first sight what these details are and > where this reference would come in. For the calculation of the expected 95% CL cross section limits in the GMSB models we use the method described in Ref. 19, specifically Eq. 3. This is important as it yields an averaged expected cross section for a given set of cuts. This cross section limit is minimized by varying the cuts. We have added a remark to Ref. 19 to be more specific on this point. > * p.8. What do cross section limits mean for the general GMSB > analysis? Do the authors now consider cascade decays to neutralinos > from squarks and gluinos? If yes, how different is the acceptance > for direct production and for cascade decays? Basically: what is > \sigma_95 in Figs.10,11? For full GMSB model simulations we allow all production channels to contribute to the final state according to their predicted cross sections. The main contribution comes from gaugino pairs which cascade-decay to jets and neutralinos (see the introduction to section III). Squarks and gluinos contribute but are not dominant. Given all production channels this defines the total production cross section and allows a calculation of the acceptance for each set of cuts. We then calculate 95% C.L. cross section limits which are uniquely defined as a function of acceptance, background and luminosity. For instance for the g+MET+0 jets analysis we get at m_chi = 110 GeV and tau_chi = 60 ns the following values: acc = 2.44% bkg = 5.66 with a production cross section of 0.4158fb at a luminosity of 2fb-1 (compare Table V). This approach of setting limits is fairly common in SUSY. See for example Ref. 3, or hep-ex/0410053. > * p.9. `There is good reason to believe that...' is a litte > vague. Maybe the authors could make it clear what they refer to as > `advertized' and why there should be an improvement, if they would > like to include a comment like that. Our work group, Texas A&M, has lead the EMTiming system installation from the beginning. From our experience and from unpublished data from the EMTiming system, the resolution is estimated to be the 1ns given in the text. However, as we take more data, i.e. higher statistics, the system becomes better understood, thus possibly leading to even lower resolution. We have added a statement to Ref. 7 to make this point clearer. > * p.34. The mass and life time are the same as in Fig.15? It would be > useful to mention that in the caption to Fig.16 and following. We are not quite sure what this comment refers to ("p.34" does not exist). We will try to answer the question from our best understanding. Please let us know if this misses your point. The choice of neutralino mass and lifetime is consistently at m=110GeV and tau=40ns. In every Figure caption that is relevant, the mass and lifetime are explicitly given. Figure 16 refers to the Monte Carlo simulation in Appendix A which is independent of the neutralino mass. We have made this point clearer in the beginning of Appendix A. > * For all supersymmetric cross sections mentioned in this paper there > are NLO results available in public computer codes. By not assigning > a large theoretical error to the PYTHIA predictions, the authors > implicitely use these numbers. I would like them to briefly check > that the NLO cross section predictions do not alter significantly > the GMSB parameter space coverage. We have used PYTHIA, a simulation tool which generates SUSY events and calculates their production cross sections in LO, and have taken a systematic error of 10% on the acceptance (see Table I and Ref. 16). In recent GMSB analyses (see e.g. http://arxiv.org/abs/hep-ex/0410053) an averaged K-factor of 1.2 has been used in the same neutralino mass range. To be conservative we have not used the K-factor to augment the production cross section predictions. We agree that it is important for future analyses to incorporate this factor in the calculations to get a true limit. As this will increase the production cross section one can again interpret our result in Figure 13 as a conservative estimate. We have added a statement to Ref. 16 to clarify the usage of LO. Note that the K-factor does not affect the cross section limits, rather they affect the exclusion region in the neutralino mass vs. lifetime plane (see Figs. 13 and 14). > * Could the authors very briefly mention how the coverage of the > parameter space develops with luminosity, to give the reader an idea > what the statistics vs. systematics limits on this analysis are? Since we present the results as averaged 95% CL cross section limits, there are no statistical uncertainties in the calculation. We do take into account statistical fluctuations in the data. Figure 10 in Ref. 22 shows the mass limits as a function of luminosity for the gg+MET analysis at CDF. We preferred not to make predictions to even higher luminosity in our paper as one encounters increasing technical difficulties which depend sensitively on the specifications of accelerator and detector, as the first referee of our note already pointed out. We added another exclusion region for 1 fb-1 and separated Figures 13a and b to help address this issue. > * There are a few layout glitches in the text and in the size of the > figures. I suppose they will be taken care of in the final version? Unfortunately we cannot find any layout "glitches" in the text. If the referee will provide us with more details, we are happy to correct our draft for this. It is our understanding that these types of problems are taken care of by the PRD editors in the final layout and proofing process. > As mentioned above, I consider the paper interesting and > useful. If the authors properly adress all the points mentioned above > I recommend its publication. We thank the referee again, and we hope that he finds our responses satisfying. ======================================