1 00:00:02,310 --> 00:00:08,160 On another side of the dark matter and Dark Sector uh world that is searches for 2 00:00:08,160 --> 00:00:10,890 unconventional signatures and long lived particles. 3 00:00:11,850 --> 00:00:12,750 So, hi, can you hear me? Yes 4 00:00:14,070 --> 00:00:18,210 And we can also see your slides. Okay. So 5 00:00:19,200 --> 00:00:24,330 yeah, then I can get started, I guess. Thanks. Good afternoon, everyone um thanks 6 00:00:24,330 --> 00:00:28,260 to the organizers for this opportunity. So, I will give a talk on the searches for 7 00:00:28,260 --> 00:00:32,610 unconventional signature and long lived particles on behalf of CMS and ATLAS 8 00:00:32,610 --> 00:00:38,190 collaboration. So, our knowledge try to understand what is a long lived particle 9 00:00:38,250 --> 00:00:42,750 and what are its properties. So, from the experimental point of view, you can see 10 00:00:42,750 --> 00:00:46,860 the picture here, it is showing a different signature based on uh whether 11 00:00:46,860 --> 00:00:52,080 particle is neutral charge or any charge. So then if a long lived particle is a neutral 12 00:00:52,080 --> 00:00:57,420 particle, let's say, then uh that decays a macroscopically reconstructible distance 13 00:00:57,450 --> 00:01:02,010 from the P P interaction point. Example in the case of a displaced photon, a displaced jet 14 00:01:02,310 --> 00:01:07,890 etc. Or it can be a charged particle which will decay as a neutral particle or is 15 00:01:07,890 --> 00:01:12,210 quasi stable on the scale of the relevant detector. So, these quasi stable long lived 16 00:01:12,210 --> 00:01:17,250 particles will cross the detector. So, long lived particles have unconventional 17 00:01:17,250 --> 00:01:22,050 signature for example, displaced objects, disappearing objects etc and then these 18 00:01:22,050 --> 00:01:29,580 signatures depend on the lifetime of the particles that is C Tau. So um, then these are since 19 00:01:29,790 --> 00:01:33,780 the entries are unconventional, there are different challenges from the experimental 20 00:01:33,780 --> 00:01:38,340 point of view, first and foremost is a non standard reconstruction for the 21 00:01:38,340 --> 00:01:42,810 displacement and the timing and then sudden sudden and secondly, we have 22 00:01:42,810 --> 00:01:48,150 dedicated triggers for the topologies. Then the third challenge is non standard 23 00:01:48,150 --> 00:01:53,880 background which arise from the detector noise cosmic rays etc. And these should be 24 00:01:53,880 --> 00:02:01,050 estimated from the data. So then uh there are many long lived searches of course performed by 25 00:02:01,050 --> 00:02:05,430 both the CMS and ATLAS experiment. So in this particular talk, I'm showing some 26 00:02:05,430 --> 00:02:10,320 of the recent results, and which I present here. And more you can see uh in the 27 00:02:10,320 --> 00:02:14,640 talks, which happened, I believe already in the parallel session, one was by Allison uh which 28 00:02:14,640 --> 00:02:19,080 were talking about the long lived particles in CMS, and then about the Atlas 29 00:02:21,390 --> 00:02:23,220 get started. So obviously, 30 00:02:28,350 --> 00:02:32,100 it is about the displace vertex and muons with the large impact parameter. 31 00:02:33,810 --> 00:02:38,700 So as I said, it's a new result with the full run two. So this particular search 32 00:02:38,730 --> 00:02:44,250 basically targets a pair production of long lived stop squarks, that decay via a 33 00:02:44,250 --> 00:02:51,510 small RPV coupling into a quark and a muon. And so in the models where we have this sufficiently 34 00:02:51,510 --> 00:02:56,970 small lambda k coupling, and this top squark is the LSP this suppression of the 35 00:02:56,970 --> 00:03:01,860 decay basically causes it to appear a discernible distance from the P P interaction 36 00:03:01,860 --> 00:03:06,570 point where this top squark pair is produced. So this would give rise to a 37 00:03:06,720 --> 00:03:11,220 feature that we have a muon and a high mass vertices that are displaced from the 38 00:03:11,220 --> 00:03:16,650 interaction point yielding a distinctive signature in the collider experience. So, 39 00:03:16,650 --> 00:03:21,000 if we have such a signature, we need definitely a dedicated tracking and the 40 00:03:21,000 --> 00:03:26,670 second vertexing algorithm. And that is what we have. So a dedicated track reconstruction is 41 00:03:26,670 --> 00:03:31,410 performed using the large radius tracking to reconstruct the tracks from the decay of 42 00:03:31,410 --> 00:03:36,720 this long BSM particle, which basically have an impact on a meter that will fall 43 00:03:36,750 --> 00:03:40,350 outside the standard constraints like we're showing in this particular picture there. 44 00:03:40,920 --> 00:03:46,350 Then we also have a dedicated uh secondary vertex reconstruction to uh reconstruct 45 00:03:46,350 --> 00:03:51,150 displaced vertices, and in this particular case, none should be non pointing. So in 46 00:03:51,150 --> 00:03:55,020 this particular search, we have different backgrounds coming from cosmic rays, fake 47 00:03:55,020 --> 00:03:58,230 muons from random hits, semileptonic decays of standard model 48 00:03:58,260 --> 00:04:05,370 dijets. So then how do we select these kind of uh uh interesting uh final state. So we 49 00:04:05,370 --> 00:04:09,870 have two mutually exclusive exclusive trigger based selection. The first one is 50 00:04:09,870 --> 00:04:14,190 based on the cluster based missing transverse energy, and the other one is based on the 51 00:04:14,190 --> 00:04:20,910 muon P T. So in the bottom, I'm showing you the plot about the observed event yield in the 52 00:04:20,940 --> 00:04:27,210 control region, in the validation region, and the signal region. So uh in the signal region uh 53 00:04:27,270 --> 00:04:30,870 in the left is for the missing E T trigger selection and on the right is for the 54 00:04:30,870 --> 00:04:35,430 muon selection, you can see in the missing E T trigger selection we observe signal events, 55 00:04:35,490 --> 00:04:40,560 whereas in the uh muon selection we observe none event, but these are consistent with the expectation 56 00:04:40,560 --> 00:04:44,490 which we see from the backgrounds and where the background basically comes from 57 00:04:44,490 --> 00:04:47,820 the transfer factor in the region that will pass the full selection which we 58 00:04:47,820 --> 00:04:54,270 have. Since everything agrees nicely, so we move on and set limits on the top squark 59 00:04:54,270 --> 00:04:59,970 mass as a functional the lifetime. So this left plot shows that for the proper mean 60 00:05:00,000 --> 00:05:05,850 lifetime of uh let's say point one nanosecond masses below roughly of the top squark mass 61 00:05:05,910 --> 00:05:11,400 up to 171 point seven GeV are excluded. And then if you try to see this plot with the 62 00:05:11,400 --> 00:05:14,760 big try to compare this particular analysis with the bigger picture meaning 63 00:05:14,760 --> 00:05:19,620 comparing with the other analysis, this is where this particular analysis fit in and 64 00:05:19,620 --> 00:05:24,570 we can see basically that it is covering a major unexplored phase space, which was not 65 00:05:24,600 --> 00:05:28,410 covered merely by the other searches which was done earlier. So, this is a very nice 66 00:05:28,410 --> 00:05:34,350 result which came up recently. Then uh we move on uh to the other than analysis which is 67 00:05:34,350 --> 00:05:39,870 performed by CMS. So, its a displaced jet analysis now, this is also uh a full run two analysis uh 68 00:05:39,930 --> 00:05:46,500 on the 132 point one inverse of luminosity. So uh this particular analysis has a particular 69 00:05:46,500 --> 00:05:50,520 topology that we have a pair of objects which are originating at a secondary 70 00:05:50,520 --> 00:05:56,190 vertex. So, there are many different signal models that has been studied by CMS. I'll 71 00:05:56,190 --> 00:05:59,880 explain here the two models. One is the jet jet simplified model and the other 72 00:05:59,880 --> 00:06:04,590 one is the exotic higgs model. So in the jet jet model there we have a long lived 73 00:06:04,590 --> 00:06:08,880 scalar neutral particle x, which are produced by an off shell z boson 74 00:06:08,880 --> 00:06:14,670 mediator, then this each x particle will decay to a quark and an anti-quark pair. And then 75 00:06:14,670 --> 00:06:17,910 we choose a signature weill be having a displaced vertices which will be having a pair 76 00:06:17,910 --> 00:06:18,750 of associated jets to it. 77 00:06:20,070 --> 00:06:24,540 Then the other interesting model which has been added by this 17 and 18 search is 78 00:06:24,540 --> 00:06:28,980 that we consider long lived particles arising from the exotic higgs. In this 79 00:06:28,980 --> 00:06:34,080 particular case you can see that the Higgs boson will decay to uh two long lived scalars and then this 80 00:06:34,110 --> 00:06:39,450 each scalar will decay to a quark and an anti-quark pair. And then there are other models like 81 00:06:39,450 --> 00:06:43,560 the beyond standard model models which have been studied also the split susy, the R P V susy, 82 00:06:43,560 --> 00:06:50,070 etc. And the limits are the limits are also shown in the future slides. So then uh talking about 83 00:06:50,070 --> 00:06:54,690 this analysis, the pre selection is that of course, since it's uh a displaced uh uh object, 84 00:06:54,690 --> 00:06:58,770 we need a dedicated trigger for this. So we have two triggers. One is the displaced 85 00:06:58,800 --> 00:07:02,970 trigger and one is inclusive. And then we have our secondary vertex reconstruction 86 00:07:02,970 --> 00:07:06,900 also there. So the major background, uh which is there is the Q C D multijet 87 00:07:06,900 --> 00:07:12,840 process, and then uh to get the event selection basically be train a B D T on the different 88 00:07:12,840 --> 00:07:17,070 variables. So, let's try to understand these, our variables, a little bit. First 89 00:07:17,070 --> 00:07:20,970 one is the vertex track multiplicity, which is the number of tracks associated 90 00:07:20,970 --> 00:07:26,040 with a vertex, then we have a vertex L X Y significance. So what is that? If you 91 00:07:26,040 --> 00:07:30,570 look at this particular picture here, so uh let's try to understand this. So for each 92 00:07:30,570 --> 00:07:35,100 displaced track, which is associated with a dijet and expected to give one 93 00:07:35,160 --> 00:07:40,770 consistent displaced dijet hypothesis is determined by finding the crossing point 94 00:07:40,800 --> 00:07:44,610 between the track helix which is this one and the dijet direction, which is this 95 00:07:44,610 --> 00:07:48,420 one. So this is the crossing point. And this is what this black dot is there in the 96 00:07:48,420 --> 00:07:53,370 picture. So then this L X Y, which is the uh transverse distance, can be estimated based 97 00:07:53,370 --> 00:07:56,400 on the different quantities which is written here, which is the impact 98 00:07:56,400 --> 00:08:01,290 parameter in 2D plane, track direction and the dijet direction. So this is how we 99 00:08:01,290 --> 00:08:05,940 estimate this L X Y. Then the interesting other variable which we have here is a 100 00:08:05,940 --> 00:08:11,580 K. So this K is basically the sum of the significance of the impact parameter 101 00:08:11,970 --> 00:08:18,690 in the 2D of the uh six leading tracks of the secondary vertex. So you can see a picture 102 00:08:18,690 --> 00:08:22,020 here, which shows this variable, you can see that for the background processes, 103 00:08:22,020 --> 00:08:27,150 which are given by uh this blue distribution, this variable is uh peaking at zero, whereas for 104 00:08:27,150 --> 00:08:31,320 the signal, it is, it is basically having a larger value. So this is a great display 105 00:08:31,320 --> 00:08:34,440 of discrimination with a signal and the background to this variable, which we 106 00:08:34,440 --> 00:08:39,690 got. So we basically train B D T with these particular four variables. Then 107 00:08:39,720 --> 00:08:44,820 after doing that, there is the plot which shows the G B D T score for the data, simulated Q C D 108 00:08:44,820 --> 00:08:49,200 events, and the simulated signal events. So you can see here for the different GBDT 109 00:08:49,200 --> 00:08:53,790 score, we have the different things here. And the last thing, which is the G greater 110 00:08:53,790 --> 00:08:59,700 than 192 is the signal region for us. And this is where we observe one event which 111 00:08:59,700 --> 00:09:03,150 is in consistent with the predicted background expectation, as you can see here. 112 00:09:03,750 --> 00:09:09,090 So then what we do, we basically go ahead and look at the uh different signal models 113 00:09:09,090 --> 00:09:14,430 for this particular search. So we have the combined results for the full run two data. So 114 00:09:14,430 --> 00:09:17,820 on the left here, I'm showing the exclusion limit on the cross section on 115 00:09:17,820 --> 00:09:21,780 the neutral uh long lived particle decaying to the two jets. So this is the jet jet model, 116 00:09:21,930 --> 00:09:26,850 basically the simplified jet jet model. Then in the um middle, here, I'm showing the 117 00:09:26,850 --> 00:09:31,890 RPV model uh the RPV model where the top squark basically decays to a bottom quark and a 118 00:09:31,890 --> 00:09:36,300 charged lepton. So this is a result from CMS and you can see here that from the 119 00:09:36,300 --> 00:09:42,300 CMS side, we're excluding the top squark masses up to 1.6 TeV for the range of lifetime. And then 120 00:09:42,330 --> 00:09:45,810 I tried to compare here the results from the Atlas, which I presented a few slides 121 00:09:45,810 --> 00:09:50,490 before. And one can see that for the Atlas they exclude the top squark 122 00:09:50,490 --> 00:09:59,970 masses up to 1.4 TeV for the range of lifetimes. Do you see these screens now? Yes. So 123 00:10:00,000 --> 00:10:04,050 sorry about that now, let me start from this now, when I was with the slide thirteen, 124 00:10:04,410 --> 00:10:10,950 so, I was talking about the about the yeah, about the exotic decay of Santa Model higgs model. So in 125 00:10:10,950 --> 00:10:14,580 this particular case, I was saying that the branching fraction larger than one person 126 00:10:14,610 --> 00:10:18,330 was excluded for the mean proper decay length between one millimeter and one 127 00:10:18,330 --> 00:10:23,100 meter. And this is the first particular model, this is the first time we looked at 128 00:10:23,100 --> 00:10:26,250 this model and the best limits about setting this particular model. 129 00:10:27,930 --> 00:10:32,700 So uh that brings me to the next search from the Atlas which is a displaced hadronic 130 00:10:32,700 --> 00:10:39,000 jet analysis in the calorimeter. So, this analysis was done on the uh 2016 data. 131 00:10:39,390 --> 00:10:43,170 The particular uh feature for this analysis is that we have a neutral long lived 132 00:10:43,170 --> 00:10:47,280 particles which are decaying in the hadronic calorimeter or at the outer edge 133 00:10:47,280 --> 00:10:51,420 of the electromagnetic calorimeter. So the benchmark model which is considered here 134 00:10:51,450 --> 00:10:57,870 is a neutral long lived scalar which uh basically uh produced from the decay of this heavy boson 135 00:10:58,080 --> 00:11:02,820 and this then this scalar will decay further to the Standard Model fermions. So if 136 00:11:02,850 --> 00:11:06,840 these scalar decay occurs in the calorimeter, then these two resulting 137 00:11:06,840 --> 00:11:10,530 tracks, there will be no activity in the tracker, and these jets will have a high 138 00:11:10,530 --> 00:11:15,270 ratio of the E H over E. So, we will have an energy deposit in the H cal. This is called a 139 00:11:15,270 --> 00:11:21,180 cal ratio jet. And another property will be will be a narrow shower. And then uh at the end, 140 00:11:21,180 --> 00:11:25,500 we required these two cal ratio jets in the analysis. So there were two dedicated 141 00:11:25,500 --> 00:11:29,070 triggers that were developed to target this topology for the low mass and the 142 00:11:29,070 --> 00:11:34,740 high mass for the boson. And so if you look at the uh plot here, basically I'm showing you 143 00:11:34,740 --> 00:11:38,730 the upper image on the branching ratio phi, which is decaying to the two long lived 144 00:11:38,730 --> 00:11:42,930 scalar as a function of the proper decay length. So this pink line is the line from this 145 00:11:42,930 --> 00:11:47,490 particular analysis which is called C R which is the calorimeter region and this was 146 00:11:47,490 --> 00:11:53,730 compared with the previous search which was done in the muon spectrometer. And we 147 00:11:53,730 --> 00:11:56,940 also did the combination of this so you can see this is again is the combined 148 00:11:56,970 --> 00:12:03,900 limit. So at the larger proper decay length we can see that the uh the muon 149 00:12:03,900 --> 00:12:09,330 spectrometer analysis was better, but at the uh lower uh basically lower decay length we 150 00:12:09,330 --> 00:12:13,650 can see the combination is giving us a little bit of gain. So, that was this analysis. 151 00:12:13,860 --> 00:12:20,040 Then the same analysis was done, but uh the hadronic jets were decaying in the inner 152 00:12:20,040 --> 00:12:25,200 detector and the muon spectrometer, so, we use the same model uh from the Atlas side it 153 00:12:25,200 --> 00:12:30,000 is based on thirty six femtobarn inverse, but in this particular case we require that 154 00:12:30,000 --> 00:12:33,990 one long lived particle will decay in the inner detector and the other one in the muon 155 00:12:33,990 --> 00:12:37,200 spectrometer. So, that is the difference with respect to the previous search where we 156 00:12:37,200 --> 00:12:42,630 explicitly require that to decay in the C R. So uh, this uh analysis is particularly 157 00:12:42,960 --> 00:12:48,780 sensitive to the lifetime between the uh centimeters to meters and basically here I'm 158 00:12:48,780 --> 00:12:53,760 showing again the comparison. So, to see basically I'm showing two plots, the top plot 159 00:12:53,760 --> 00:12:59,760 is for the scalar mass of 125 GeV, the bottom plot is for the 200 GeV. So, we see that the 160 00:12:59,760 --> 00:13:05,100 sensitivity to the low mass scalars basically for the 125 GeV increases with 161 00:13:05,640 --> 00:13:09,450 shorter proper life as you see in this particular plot to the left, but this 162 00:13:09,450 --> 00:13:14,220 analysis related to the combined analysis which was done so, if you see this and ID 163 00:13:14,520 --> 00:13:19,410 is this particular analysis the C R plus M S analysis and once we combine these the 164 00:13:19,410 --> 00:13:23,970 sensitivity increases at this a lower proper life then for the lower scalars whereas, 165 00:13:23,970 --> 00:13:28,050 for the high scalars, high mass scalars, this is not the case. So, that was the analysis 166 00:13:28,050 --> 00:13:34,530 which was performed again recently with the 2016 data. So uh now I will talk 167 00:13:34,530 --> 00:13:38,160 about the reinterpretation of the displacement on each analysis using this 168 00:13:38,160 --> 00:13:43,560 recast framework, which came up written recently, just few months back. So, before 169 00:13:43,560 --> 00:13:47,820 talking about the analysis, let's try to understand what is the basics of this 170 00:13:47,820 --> 00:13:53,130 recast framework. So, it is really unfeasible if you try to construct the optimized 171 00:13:53,160 --> 00:13:57,420 analysis for the different variety of physics model. That is not possible. But 172 00:13:57,480 --> 00:14:01,890 if the analysis is optimized for the original model it can offer good sensitivity to 173 00:14:01,890 --> 00:14:07,410 the other model also provided the signatures are similar. So, event selection that can 174 00:14:07,410 --> 00:14:10,980 estimate and the observed data distribution do not change the contents of 175 00:14:10,980 --> 00:14:15,390 the interpretation that is what we have from the original analysis. What will be 176 00:14:15,390 --> 00:14:18,900 required here is only the signal distribution, a new signal model which we 177 00:14:18,900 --> 00:14:22,770 are trying to derive. So, this forms the basis of reinterpretation approach 178 00:14:22,770 --> 00:14:27,780 which basically recast the input to this new decay. So, in summary this 179 00:14:27,810 --> 00:14:31,620 particular recast framework is designed to be used the estimate of background 180 00:14:31,680 --> 00:14:36,630 uncertainty is an observation to test the alternative signal hypothesis. So, with 181 00:14:36,630 --> 00:14:41,340 this particular analysis Atlas has tried to search three model, stealth susy, higgs portal 182 00:14:41,370 --> 00:14:46,920 and dark photon. In this particular slides I'm talking here about the dark photon model and the 183 00:14:46,920 --> 00:14:51,240 interpretation from we see. So, it is another hidden sector model it is the FRVZ model 184 00:14:51,240 --> 00:14:56,070 and here you can see two Feynman diagrams. So, in first case either the dark 185 00:14:56,070 --> 00:15:01,170 fermion decays into a dark photon and the lighter darker fermion, which is a 186 00:15:01,260 --> 00:15:07,440 L S P. Or in the other cases, the dark fermion decays into an L S P and a dark scalar 187 00:15:07,680 --> 00:15:11,940 which will in turn decay into a pair of dark fermions. So, in this particular case, we 188 00:15:11,940 --> 00:15:15,390 will get four dark photons, whereas in this particular case we'll get two dark photons. 189 00:15:17,310 --> 00:15:22,290 So uh, then what we did we looked at the uh cross the limits on the cross section into the 190 00:15:22,290 --> 00:15:27,840 branching fraction as a function of the long lived particle decay near the end basically, 191 00:15:27,840 --> 00:15:32,490 the higgs can decay with two dark photons or four dark photons. So, the top plots 192 00:15:32,490 --> 00:15:39,420 are from the uh uh the particular case of uh higgs one twenty five GeV and the bottom plots are for the 800 GeV. So, 193 00:15:39,420 --> 00:15:43,740 for the higgs 125 GeV, these are the first constraints set uh by the Atlas in this 194 00:15:43,740 --> 00:15:49,800 particular model. And then for the 800 GeV, we have some uh uh uh basically results with the 195 00:15:49,830 --> 00:15:54,150 uh which would be published by the collimated leptons or the light 196 00:15:54,150 --> 00:15:58,290 hadron results. So that was completed. This is this particular analysis and this was 197 00:15:58,290 --> 00:16:03,840 the previous analysis and you you can see the complementarity between the two ends. So 198 00:16:03,900 --> 00:16:07,560 this is basically telling you how the importance of the recasting but that we 199 00:16:07,560 --> 00:16:10,770 really don't need to work through the entire analysis we save things and then we 200 00:16:10,770 --> 00:16:14,550 can just put down the new signal model signal model and obtain the first 201 00:16:14,550 --> 00:16:22,140 constraints, which is really nice. So, then uh moving along. So uh, now, I will talk 202 00:16:22,140 --> 00:16:26,400 about the another interesting search from the CMS which is again based on the full run two 203 00:16:26,670 --> 00:16:33,060 data set, it is about the disappearing tracks. So uh, this particular search 204 00:16:33,600 --> 00:16:37,230 uh basically targets a long lived particle, which will decay within the 205 00:16:37,259 --> 00:16:38,099 C M S tracker. 206 00:16:38,670 --> 00:16:44,100 So if the decay products of this particular particle is undetected, maybe because of 207 00:16:44,100 --> 00:16:48,480 their particular momentum to be reconstructed or because they interact really weakly. So what 208 00:16:48,480 --> 00:16:52,500 we have we will have a disappearing track kind of signature. So this particular 209 00:16:52,500 --> 00:16:57,030 signature is identified as an isolated particle track that we extend from the 210 00:16:57,030 --> 00:17:02,220 interaction region, but after the point of disappearance, it will leave no hit in the muon 211 00:17:02,280 --> 00:17:06,090 or tracking detector or the tracking detector and then it will have only an energy 212 00:17:06,090 --> 00:17:10,500 deposit in the calorimeter set. So, that will be the signature that we will have only 213 00:17:10,500 --> 00:17:14,040 energy deposit in the calorimeter there will be no hit in the muon or the tracking detector 214 00:17:14,040 --> 00:17:18,840 the test bar. So, the benchmark signal model which has been considered here is 215 00:17:18,840 --> 00:17:24,630 the A M S B model. So, the particle mass spectrum of a chargino and neutralino uh that 216 00:17:24,660 --> 00:17:29,790 are nearly degenerate in mass. So, then what will happen is that the chargino is 217 00:17:29,790 --> 00:17:34,020 long lived and then it can reach the CMS tracking detector before it will decay to a uh 218 00:17:34,110 --> 00:17:38,340 neutralino and a uh pion and then neutralino will interact weakly and the 219 00:17:38,340 --> 00:17:41,130 pion will be very soft to be deconstructed. So, that will result in a 220 00:17:41,610 --> 00:17:48,960 disappearing like signature. So uh, let me remind you that CMS has performed this 221 00:17:48,960 --> 00:17:54,150 search at eight and thirteen TeV, but at that point using a three layer pixel tracker, 222 00:17:54,600 --> 00:17:59,190 and the sensitivity was limited to the short tracks. This is the uh result of the 223 00:17:59,190 --> 00:18:03,570 uh previous result as published in J HEP. you can see that we could go only up to the point 224 00:18:03,570 --> 00:18:10,920 one nanosecond. But then in 2017, now we have a new four uh layer pixel tracker. So we can 225 00:18:10,920 --> 00:18:14,700 think about, we can basically take advantage of that. And then we do the 226 00:18:14,700 --> 00:18:18,570 new analysis being resulting in a number of layers with measurement. And the 227 00:18:18,570 --> 00:18:22,650 number of layers are the number of uh tracker layer, which you can see here 228 00:18:22,770 --> 00:18:27,000 is the four, five and greater than or equal to six. And this can result 229 00:18:27,000 --> 00:18:30,450 to the new sensitivity to the shorter particle lifetime, which I'll show in the 230 00:18:30,450 --> 00:18:36,420 next slides. So uh what are the backgrounds for this search? Since the Standard Model 231 00:18:36,420 --> 00:18:41,310 particles do not disappear, so they rarely produce a signature. So the background will be 232 00:18:41,370 --> 00:18:46,080 really I mean, entirely composed of failures of the particle reconstruction or 233 00:18:46,080 --> 00:18:51,960 track finding algorithm. So what can we uh do to control them? So there are two ways. First 234 00:18:51,960 --> 00:18:56,910 is that we explore and mitigate all the possible ways to lose the tracker hits, or the 235 00:18:56,910 --> 00:19:00,390 second one is that we estimate the remaining probability to fall into the signal region. You 236 00:19:01,860 --> 00:19:03,360 have about five minutes left. 237 00:19:03,900 --> 00:19:09,330 Okay, sure. So, then uh the selection which we have. So at the trigger level, since 238 00:19:09,330 --> 00:19:12,900 there is no tracking information available at the level one trigger, we trigger on the 239 00:19:12,900 --> 00:19:17,430 missing transverse momentum. And then finally, the criteria which will define the 240 00:19:17,430 --> 00:19:21,660 condition to have a disappearance, we have two conditions here that the track must 241 00:19:21,660 --> 00:19:25,530 have at least three missing outer hits, which reject basically most of the standard model 242 00:19:25,530 --> 00:19:32,400 tracks. And then the sum of the associated uh um calorimeter energy within this delta less than point 243 00:19:32,400 --> 00:19:36,360 five should be less in ten GeV, and what this condition will do that it will reject 244 00:19:36,630 --> 00:19:41,100 all the electrons and the charged hadrons. So, example electron with significant brem energy 245 00:19:41,100 --> 00:19:44,460 causing a track reconstruction failure will be removed through this particular 246 00:19:44,460 --> 00:19:50,190 condition. So, then uh there I'm showing you the results. So, this is a two 247 00:19:50,190 --> 00:19:55,080 dimensional constraint on the chargino mass and the mean prop as a function of 248 00:19:55,080 --> 00:20:01,050 the proper lifetime for the wino like neutrino in the left plot and higgsino like 249 00:20:01,080 --> 00:20:06,540 neutrino in the right hand side. So these basically results exclude uh charginos below 250 00:20:06,690 --> 00:20:12,270 for the wino like case 884 GeV for a lifetime of 3 nanoseconds, and up to 474 GeV for a 251 00:20:12,270 --> 00:20:17,310 lifetime of one to two nanoseconds. So you can see that using this uh upgraded uh pixel 252 00:20:17,640 --> 00:20:22,020 track pixel detector, we could reach to the this lower lifetime, which was not possible with 253 00:20:22,020 --> 00:20:26,460 our previous searches. And then we had a new interpretation for the 2017 and 18 254 00:20:26,460 --> 00:20:30,990 data for the higgsino like case. And there we exclude up to 750 GeV for a 255 00:20:30,990 --> 00:20:35,580 lifetime of three nanosecond and 175 GeV for a lifetime of point oh nine oh five 256 00:20:35,580 --> 00:20:39,480 nanoseconds which you can also see from these plots that we could go beyond this 10 to the 257 00:20:39,480 --> 00:20:45,960 minus one, which was not the case earlier in this published result. So uh then I'll 258 00:20:45,960 --> 00:20:49,980 just talk a little bit about this neural network based long lived search which we have 259 00:20:49,980 --> 00:20:57,660 performed since um I don't have a lot of time here. So this search was um basically done to 260 00:20:57,660 --> 00:21:02,550 uh enhance the sensitivity with degenerative displaced jet tagger jet tagger. So, basically, we 261 00:21:02,550 --> 00:21:05,940 had a deep neural network based multi class classifier to identify the 262 00:21:05,970 --> 00:21:10,950 displaced jet. So, this uh there is already an uh analysis done with this displaced jet. So what we have 263 00:21:10,980 --> 00:21:16,620 we created this analysis using this dnn classifier. So, what we found I think I 264 00:21:16,620 --> 00:21:21,330 can directly take you to the result. So, what we found there is uh this is the result 265 00:21:21,330 --> 00:21:25,500 here, so, there was a so, in this particular case, I'm showing here the lower 266 00:21:25,500 --> 00:21:30,330 limit on the gluino mass as a function of the lifetime. So, this uh dotted line is the 267 00:21:30,330 --> 00:21:36,690 previous result which we have observed and this uh uh blue line here is the result using this jet tagger. 268 00:21:37,050 --> 00:21:41,310 So, what you can see that there is a significant gain in excluding values of 269 00:21:41,310 --> 00:21:48,060 this gluino masses you can see here, gluino masses, we have gained here up to around 500 GeV that 270 00:21:48,090 --> 00:21:55,290 but then to for the uh C tau above uh one meter, one millimeter. So, this topic is also 271 00:21:55,290 --> 00:21:59,190 competitive with a special dedicated reconstruction that was reported earlier. 272 00:21:59,490 --> 00:22:04,260 But what happened in this region, which was less than one millimeter, we have a bit uh 273 00:22:04,260 --> 00:22:08,430 poor performances here, the tagger performance was degrading, because of our limited 274 00:22:08,430 --> 00:22:12,510 ability actually to tag our uh long lived particle jet in the vicinity of the 275 00:22:12,690 --> 00:22:17,250 primary P P interaction vertex. But the left hand search, which is this particular search is 276 00:22:17,250 --> 00:22:21,810 able to explore the distinguishing kinematic feature of this particular model. So, that is 277 00:22:21,810 --> 00:22:26,970 where this default search was better, but we could do better here with uh this jet tagger. 278 00:22:27,990 --> 00:22:33,060 So uh, okay, so that brings me to the overview of the all the CMS long lived 279 00:22:33,060 --> 00:22:37,320 searches which we have performed. So here we have divided the search into three different 280 00:22:37,320 --> 00:22:42,270 models, you see R P V, R P C, other and I'm highlighting here the two analyses which 281 00:22:42,270 --> 00:22:46,530 I've talked about today, the displaced jet and the disappearing track, which was done 282 00:22:46,530 --> 00:22:52,140 using the uh you can see here with the full run two dataset. And similarly, here are the 283 00:22:52,140 --> 00:22:57,480 results from the Atlas side. And again, I presented today the displaced vertex plus muon, 284 00:22:57,630 --> 00:23:02,070 the I D M S vertex search and two mu jets. And you can see all the luminosities which were used here. 285 00:23:03,240 --> 00:23:10,260 So uh that brings me to the summary. So I have highlighted several unconventional signatures and 286 00:23:10,260 --> 00:23:14,520 long lived searches today from the full run two data set. From the CMS side, I have shown the 287 00:23:14,520 --> 00:23:18,900 disappearing track search, which actually made use of this upgrade to extend the 288 00:23:18,900 --> 00:23:22,830 sensitivity to the lower particular lifetime. And in this particular case, we have added a new 289 00:23:22,830 --> 00:23:26,880 higgsino like interpretation. Again, for full run two, I 290 00:23:26,880 --> 00:23:31,260 showed the CMS displaced jet search, which again, I didn't do so it's not just a 291 00:23:31,260 --> 00:23:34,890 simple reload of the previous analysis, we again added an interpretation from the 292 00:23:34,890 --> 00:23:36,090 new model, which was the exotic decay of standard model higgs boson 293 00:23:37,980 --> 00:23:41,670 And then uh the third search from the full run two I showed from the Atlas, which was 294 00:23:41,670 --> 00:23:45,780 the displaced vertex and muon search and that covered the unexplored phase space, which 295 00:23:45,780 --> 00:23:50,490 was not done earlier. And then uh there are like ongoing effort from Atlas to use 296 00:23:50,490 --> 00:23:54,630 this recast framework, which is really nice to reinterpret their new results, because we don't 297 00:23:54,630 --> 00:23:58,860 have to get that we can just use this framework which is really a nice thing. 298 00:23:59,220 --> 00:24:02,550 And then from this CMS side, so I showed you something very, very quite out of the box 299 00:24:02,550 --> 00:24:07,230 thinking like if you have a neural network based search to target the long lived particle signature. 300 00:24:07,710 --> 00:24:11,700 So, this is what we have at present. So, there are more results which are coming 301 00:24:11,700 --> 00:24:16,800 out so, please stay tuned for that. And as far as moving ahead uh for the long lived 302 00:24:16,800 --> 00:24:20,820 particle searches towards run three, I would say we have new ideas for trigger 303 00:24:20,820 --> 00:24:24,660 reconstruction techniques and application of machine learning all this is going on, 304 00:24:24,780 --> 00:24:28,440 so that we can extend the phase space of the long lived particle and there was also a workshop 305 00:24:28,440 --> 00:24:31,650 recently there all this has been discussed. So, yeah, we are looking 306 00:24:31,650 --> 00:24:35,220 forward to all the new results which will come up and thanks for listening to me and 307 00:24:35,220 --> 00:24:37,650 sorry for this uh trouble in between. Thank you. 308 00:24:38,910 --> 00:24:40,260 Thank you very much for the talk. 309 00:24:42,060 --> 00:24:45,240 And uh I will ask the audience for questions, if there's any 310 00:25:00,630 --> 00:25:05,250 Not yet, so I'm intrigued, of course, by the new ideas for trigger reconstruction 311 00:25:05,250 --> 00:25:07,440 techniques. Could you expand a little on that? 312 00:25:08,519 --> 00:25:13,019 Yeah, so, there are like some signatures, I mean, which which we can see benefit, 313 00:25:13,019 --> 00:25:18,239 because there are some that we need to have this muon P T right where we have 314 00:25:18,239 --> 00:25:23,069 not really gone below 10 GeV and there are a few models which which requires that the 315 00:25:23,069 --> 00:25:27,809 muon P T to be very soft like 4 GeV 5 GeV. So, we need to have some triggers which are 316 00:25:27,809 --> 00:25:31,379 basically designed for these kinds of searches for targeting the long lived 317 00:25:31,379 --> 00:25:35,309 particles. Similarly, the reconstruction techniques, I mean, there are a few models 318 00:25:35,309 --> 00:25:42,059 where uh uh where we are not able to reconstruct the particles like uh in case of uh really uh the jet, the 319 00:25:42,059 --> 00:25:48,269 jet searches where we have this uh the particular jet, which is like narrow, and what we 320 00:25:48,359 --> 00:25:51,839 basically got from is the same for jet. So we need to have some techniques where we 321 00:25:51,839 --> 00:25:54,689 are targeting the particular model which we're interested in and be able to 322 00:25:54,689 --> 00:25:59,519 reconstruct this long lived particle. So all these uh new ideas are coming up for the 323 00:25:59,579 --> 00:26:03,239 trigger we added menus for the trigger. And that is where we are trying to see how we can 324 00:26:03,239 --> 00:26:07,769 improve the search. Because if we're able to go to lower PT for the objects, then 325 00:26:07,769 --> 00:26:12,629 that will definitely have the real impact on the final limits or or, or the objects 326 00:26:12,629 --> 00:26:13,439 which we're seeing at the end of the day. And 327 00:26:14,700 --> 00:26:17,340 that is where I think the new ideas for trigger is really helpful. 328 00:26:19,650 --> 00:26:20,610 Thank you very much. 329 00:26:23,400 --> 00:26:30,750 So last chance to ask any questions here for this session. There is a question and 330 00:26:30,750 --> 00:26:40,080 it's a Alison Elliot. And I'm not sure I can. After the glitch, I don't have the 331 00:26:40,110 --> 00:26:43,830 option to talk. Okay, good. Someone did this for me. So, 332 00:26:45,210 --> 00:26:48,180 yes, please unmute yourself. And you can you can ask the question. 333 00:26:49,470 --> 00:26:54,390 Hi, hi. Can you hear me? Yes we can. Yes. 334 00:26:56,190 --> 00:27:02,010 Okay, thank you. Sorry. Um yeah, I just wanted to follow up on the idea for 335 00:27:02,130 --> 00:27:06,000 triggers question. I'm wondering, and it kind of follows along from the the 336 00:27:06,000 --> 00:27:11,280 previous couple of talks, have you considered data parking for looking for long live 337 00:27:11,280 --> 00:27:19,260 particles especially since you'd you would benefit from low P T muons or low P T 338 00:27:19,530 --> 00:27:26,580 jets or muons you could park the data you could reconstruct it later has has 339 00:27:26,580 --> 00:27:29,040 this been considered at all for long live particle searches? 340 00:27:49,890 --> 00:27:51,990 So is Bhawna speaking 341 00:27:59,760 --> 00:28:03,000 yeah Do you hear me? Yes, no, no, it 342 00:28:03,180 --> 00:28:08,040 looks like I was muted. Okay. So I will say yes, this has been considered in CMS. 343 00:28:08,070 --> 00:28:12,030 I'm from CMS, so I can talk from the CMS side. So we have been really looking into 344 00:28:12,030 --> 00:28:15,900 this parking parking feature and they were some analysts also are done using parking 345 00:28:15,900 --> 00:28:20,160 but not in long lived. But as long lived we are we are, we're taking into account this 346 00:28:20,160 --> 00:28:24,300 feature. I know from CMS side. I do not know from ATLAS side. 347 00:28:26,730 --> 00:28:29,430 Thank you so much. That's really interesting. It'll be really interesting 348 00:28:29,430 --> 00:28:30,300 to see the results with that. 349 00:28:39,420 --> 00:28:41,220 Okay, so I think uh this 350 00:28:42,540 --> 00:28:44,760 possibly concludes our session. 351 00:28:47,160 --> 00:28:53,880 And thanks to all the speakers again, virtual clapping hands, don't know if we have 352 00:28:54,570 --> 00:29:02,130 the option and I probably let it to the 353 00:29:03,600 --> 00:29:04,950 Uber chairs to 354 00:29:06,450 --> 00:29:11,220 say what's going on next, I believe. Yes, thank you everyone because uh the next 355 00:29:11,220 --> 00:29:15,810 session will be starting in nine minutes in another webinar. So please connect there 356 00:29:15,810 --> 00:29:23,040 and we will broadcast uh soon uh the next session also on awebcast. See you there for 357 00:29:23,040 --> 00:29:27,750 the closing plenary session. Bye bye bye.