Set your browser to use document fonts to see all symbols correctly.

Daniel Denegri (CERN) "Higgs and Susy Higgs Studies at the LHC"

  1. Standard Model Higgs discovery potential:
    1. Please delineate the regions in which LHC can see only one, or two, channels of Higgs decay. How do the mass reaches vary with accumulated luminosity?

      For the variation on SM Higgs mass reaches with luminosity you have in my talk CMS and ATLAS reaches from 10**4 to 10**5pb-1 and there is not much to be added at present. Concerning the number of observable decay channels - again visible on ATLAS and CMS SM Higgs reach summaries. The following points merit to be mentioned:

      1. the H to gamma-gamma mode (~100 - 155 GeV) is always extremely demanding in terms of gamma-gamma mass resolution, ~1% is required, and the more so lighter is the H;
      2. the difficulties in observing H to bb (with H to gamma-gamma) in the m(H) < 120 GeV range: H to bb difficult due to very demanding requirements on the b-tagging performance for b-jets in the Et 20 to 50 GeV range, and of a jet-jet mass resolution of about 10%. The low Et b-jet range is essential to be able to observe a "shoulder" on the left side of a bb peak - and thus be able to claim that there is peak; ATLAS assuming a 60% b-tag eff and a 1% mistag rate obtains godd signal visibility; detailed CMS b-tag simulations do not give (yet) such good performance for Et below ~ 40 GeV - and thus no convincing "evidence" for signal, but work goes on both on final tracker (pixel) optimisation and on tagging algorithms development - so there is still hope.
      3. for 125 < m(H) < 155GeV we should have both H to gamma-gamma and H to ZZ* modes.
      4. the domain m(H) ~ 158/160 to 170 GeV is also difficult due to the collapse of the H to ZZ* branching ratio at WW threshold requiring thus at least 10**5 pb-1 of integrated Lumi (Lint) to see a peak in the 4-leptons mass; the alternative channel H to WW to two leptons+two neutrinos gives no peak - only an "excess of events" as evidence for an H - and will thus require a full understanding of bkgds (tt,WW,Wtb....) before the case can be made.
      5. for m(H) > 300 GeV both H to ZZ and H to WW modes are accessible (see ATLAS SM Higgs summary plots)

    2. In the case that Higgs is in the <150 GeV region, how confident are we that H-> gg can be distinguished? Are backgrounds and resolutions sufficiently understood to make a clear case (with what luminosity?) What other channels (gg +X, WW*, bbbar) or W+Higgs associated production can be added to increase the confidence in this region? What interplay is expected between Tevatron and LHC to more fully understand the nature of a Higgs in the low mass region?

      Concerning H to gamma-gamma: Several experimental channels are available:

      1. inclusive H to gamma-gamma (dominated by gg to H): as S/B ~ 1/20 this is most demanding in terms of gamma-gamma mass resolution; CMS expects (hopes) to achieve ~1%, ATLAS about 1.2 to 1.4%; L_int required > 10**4pb-1, signal sample from few hundred to ~1500 events (depends on L_int and mass);
      2. WH and ttH channels: here S/B is ~O(1) thus significantly less demanding in terms of mass resolution, however small cross section requires at least 10**5pb-1 for signal samples of 10-20 events/10**5pb-1. More phenomenological work/understanding of the various (small) bkgds would be welcome.
      3. H + jet(s) i.e. H production at large pt (say > 30-40 GeV): incorporating ttH, WH, ZH, qqH and QCD radiative corrections to gg to H is intermediate in terms of requirements on both mass and Lint between the two preceeding cases, S/B again of order ~1, event samples about 100 signal events/10**pb-1; significant uncertainty still exists on the (main) signal contribution from gg to H contribution, more phenomenological work to have good understanding of Pt of Higgs at production ( and recoil jet multiplicity...) would be most welcome;
      4. qqH (H to gamma-gamma) : too small cross section to be really useful by itself; qqH with H to tau-tau is promising - under detailed investigation now in CMS, but in any case requires at least 10**5pb-1.

  2. Measurement of Higgs properties:
    1. What is the prospect for determining Higgs BRs into fermion pairs (bbbar, ccbar, mm, tt), gluon pairs and WW, ZZ? How do these measurements differ for low, intermediate and high mass Higgs. Can these be done in a model independent way, or is theoretical input necessary?

      1. SM Higgs BR for H to bb relative to H to gamma-gamma accessible only for m(H) < 120 GeV (with the Q1 proviso on H to bb observability), uncertainty on the ratio of BR's is ~ 30%, stat. limited.
      2. I am not aware of any promising H to cc studies;
      3. H to mu-mu not possible for SM Higgs - overwhelmed by nearby Z to mu-mu; this is not the case for SUSY h,H,A to mu-mu thanks to large tanbeta enhancement.
      4. H(SM) to tau-tau (for m<150 GeV): good chance to be detectable in the qqH channel - more work needed to be fully affirmatory- underway right now in CMS; in any case requires at least 10**5pb-1 as sample small, about 10 signal events for 10**5 (see my talk for exact event numbers); triggering difficulties in the tau-tau to h+h- channel but not so in l+l- or lepton-hadron channels.
      5. the H to WW vs H to ZZ branching ratios can be measured in the 160 300 GeV range both H to ZZ (to 4l,llneutneut,lljj) and WW (to l neut jj)are measurable - see ATLAS SM Higgs summary plot in talk;
      6. the ratio of BR's for H to gamma-gamma and H to ZZ can be determined in the 125 < m < 155 GeV range, should be determined to ~ 15% precision; ratio of ttH to WWH couplings can be determined in the ~100 < m < 130 GeV range with 25% precision, statistics limited. No work nor thinking has yet been done on the H to glue-glue BR, but I think that it could be obtained from the measured sigma*BR for H to gamma-gamma - knowing Lint, provided BR for H to gamma-gamma is determined/known/calculated from for ex. WH from qqbar to W* to WH with well known/controllable qqbar partonic luminosity determined from W and WW production. I think theoretical input will be necessary at least to unfold the gluon-gluon differential partonic luminosity part.

    2. What are the prospects for determinations of Higgs spin and CP? (What luminosity is required for different mass regions?)

      to my knowledge no study has yet been made in CMS or ATLAS on ways to determine the spin and CP nature of the Higgs... - the very observation of a signal in gamma-gamma already limits spin to 0 or 2; - H to ZZ to 4 leptons should be sensitive through the alignment of Z to l+l- to parent Higgs spin, sensitivity not yet investigated, in fact cuts favoring longitudinal vs transverse Z decays have been sometimes "used" up to now to enhance signal visibility over ZZ continuum. Once the signal is established arguments shouls be reversed and used to ascertain Z alignements and parent H spin - not tested yet.

  3. Unraveling Susy Higgs:
    1. In a MSSM Higgs decoupling (heavy A,H) scenario, how would LHC determine that we are in the Susy sector rather than the SM Higgs world? Do Higgs measurements play a role in this?

      For SUSY Higgs searches three channels play a dominant role: h to gamma-gamma and to bb and A,H to tau-tau. The instrumental problems/requirements for gamma-gamma and bb have been mentioned already in the context of SM-H, they are the same. For h to gamma-gamma there are new possible difficulties ie suppression of production rate in a scenario with large stop mixing with a light stop;in gg to h production top-stop loop interference can suppress the cross section - partially compensated by stop-W interference in the h to gamma-gamma decay loop, but all together in case of m(stop) ~ m(stop) the sigma*BR can fall below the observability level (~30fb). Fortunately the enhanced (factor <2) h to gamma-gamma BR allows slightly easier observation in the Wh, tth channels, but still requiring ~ 10**5pb-1. These points are well illustrated in my talk. In general h to gamma-gamma is limited to m(A) > 200 - 230 GeV. This is why the h to bb channel - instrumentally difficult and not yet fully understood (at least in CMS) is particularly important as it would allow to cover essentially the entire MSSM tanbeta vs M(A) plot.....(there are plots in my talk).

      The role of A,H to tau-tau is particularly important as:

      1. it gives access to A,H for masses of order 500 GeV, up to 800 GeV if tanbeta > 35 - see ATLAS and CMS plots- decays to sparticles have yet to be investigated, and
      2. it allows to explore the difficult region m(A) < 200-250 GeV, 7-8 < tanbeta < 10-15 in fact it is the only channel allowing it - with h to bb which has its own and instrumentally very different problems/limitations. For tau-tau the problem in tau-tau to h+h- whish has the best mass resolution (~10%) is triggering.

      on Q3a : The decoupling limit. SM vs SUSY discrimination in h(H) to gamma-gamma based on i) absolute rate measurement and ii) ratio of BR's gamma-gamma/bb. Absolute rates will be limited by lumi measurements to ~10% at best to 5% precision. There should be also a ~ 10% uncertainty on bkgd subtraction in h to gamma-gamma. If there are no complications due to light stop (ie mstop ~ 1TeV) the rates for SM and SUSY-h to gamma-gamma differ by less than 10% for m(A) > 550 GeV (according to Gunion et al.). The ratio of gamma-gamma to bb BR's (not dependent on lumi) between SM and MSSM do not differ by more than 15% for m(a) > 500 GeV, and 15% is a reasonable estimate of the experimental uncertainty on bb signal(bkgd shape,subtraction). Thus for m(A) > 500-550 GeV - if nothing else is observed but gamma-gamma - hard to distinguish SM from SUSY.

    2. What are the regions in which A and H masses can be determined reliably? If A and H are unresolved, can one deduce that both are present? We would like to understand how these issues vary as tan b becomes large or for MA > 500 GeV.

      the regions where A+H can be determined reliably are clearly visible for ex on CMS MA,tanbeta MSSM summary plots in my talk for 3*10**4pb-1. At large tanbeta gg to H and bbH available; if gg to H were suppressed by top-stop interference, would not be the case for bbH. At given m(A) larger tanbeta easier the observation of A,H in both tau-tau and mu-mu decays. In tau-tau channel A and H are unresolved - presence of two objects should come from observed rate; A possibility to resolve A from H in mu-mu channel where the mass resolution is better or ~ 1% - plot in my talk suggesting this possibility but has not yet been seriously studied.

    3. In what regions can quantum numbers of heavy MSSM higgs be determined? stop mixing angle?
    4. Can the LHC experiments access H ->tt decays? These seem particularly well suited to extending the MSSM Susy reach. How about H -> t tbar?

      A,H to tau-tau OK as already explained, with mant plots illustating it in the talk;A,H to tt accessible at low tanbeta (<3), 350 < m(A) < 450 GeV visible in ATLAS MSSM summary plots; charged H to tau+neutrino and tb also accessible for masses of order 500 GeV -plots available in talk - at present under active study.

    5. How would LHC experiments detect Higgs if H-> neutralino pairs? (A particular case of interest is that for which H-> c10 c10 is allowed, and missing energy signatures are required.) What difficulties or gains result if there is significant Higgs production from sparticle decays?

      observation of h to bb in squark/gluino decays in fact much easier - if kinematically allowed - than in Wh and tth channels thanks to additional ways to suppress bkgd, essentially hard Etmiss cuts; several plots in my talk with rather conservative reach estimates from CMS and more optimistic ones from ATLAS. H to neutralino-neutralino is at present under investigation - too early to say something definitive.

    6. Are there modifications to the Susy Higgs searches if one goes from mSugra to gauge mediated or R parity violating Susy?

  4. Experimental issues
    1. What capability do the experiments have for separated vertex triggers? What is the sensitivity to multiple interactions? What are the efficiencies and purities for b tagging and c tagging?

      in CMS no separated vertex triggers are foreseen at first (hardware) level, neither (yet) at second (software) level, but should be present at level 3; in fact only first level is well understood at present, higher levels are now under intense study (there is a "continuum of levels" ie smooth transition between on and off line, will really depend on estimated/extrapolated performance of switches and processors....);the situation is similar for ATLAS - possibly at level two, certainly at level 3. Sensitivity to multiple interactions will be understood in this context, it is too early now. Efficiencies and purities for b-tagging (off-line) well understood at 10**33 and 10**34 cm-2s-1 and illustrated by several plots in talk. For abs(rap)< 1.5 efficiencies/purities are OK (Et b-jet dependent), performance less satisfactory at 1.5 < abs(rap)< 2.5 a region very important for bbH(SUSY) final states. c-tagging not explicitely studied in CMS, but as a byproduct or rather as a bkgd to b-tagging, seems to be at 5-10%level; could probably be optimized separately but not done yet.

    2. Studies at LHC have tended to use no K factors for signals or backgrounds; this is conservative if the K factors for signal turn out to be larger than for background. What is the probability that this will be true? What are the areas where one might expect the most difficulty from enhanced background K factors?

      ATLAS has evaluated Higgs visibilities without k-factors - a conservative approach; in CMS k factors are included in H to gamma-gamma studies and indeed they turn out to be slightly larger for signal than for bkgd and even if same you gain as significance is S/sqrt(B). What would be very interesting is to calculate higher order corrections to bbH final states, not only as "k-factors" but also to see whether there are modifications to the accompanying b-jets kinematics - these jets are soft and thus difficult to tag efficiently - but are essential to extract A,H signals in tau-tau and mu-mu. The main bkgd to bbH is tt and bbZ ; danger is not in K factor for bkgd but rather the hope that bbH could be significantly enhanced (S/sqrtB gain) and b-jets kinematics maybe somewhat more favorable.

    3. How are the QCD jet backgrounds to photons and leptons to be determined experimentally? What accuracy is expected and what needed?

      QCD jet bkgd to leptons ie pi+- +pizero to e+- should be no problem for e+- of Et > 5 GeV with E/p,ECAL granularity,Hcal/ECAL samplings/cuts and preshower. For soft e+- for ex. psi to e+e- samples could allow understand contamination with/without preshower cuts etc Note that for H studies we always require Et(e) > 10 GeV. For QCD jets (pizeros) bkgd to gamma - this is a big topic in itself - but again granularity of ECAL, presampling/preshower,evaluation of samples versus isolation in tracker and/or calorimetry versus isolation (pt,Et) threshold and domain (Del(eta),Del(phi),cone size) should allow to monitor contamination. For Higgs studies we rely on gammas of Et > 20 -25 GeV. It is worth reminding that for H to gamma-gamma we are entirely dominated by real (irreducible) gamma-gamma bkgd, both ATLAS and CMS studies expect a QCD (jet-jet and jet+gamma) bkgd not exceeding 15% of overall bkgd. The problem of gamma-gamma vertex selection in CMS in the absence of a preshower at high luminosity and competition between hardness of min bias events/tracks and Higgs pt recoil system is well documented in talk.

    4. Can you reassure us that ttbar(+jets), diboson production (WW, WZ etc.) backgrounds are under control? triple boson production?

      The bulk of tt, WZ,WW,ZZ production is well understood - the higher order QCD corrections are precisely the subject of the CERN "SM at LHC workshop" studies (yellow report expected in March 2000); there is no doubt that improved calculations,phenomenological understanding,MC implementation for these channels , but also for W+jets, Z+jets... would be most usefull as these are our main bkgds in Higgs searches.

  5. Other issues
    1. Can Higgs pair production be observed?

      H to hh to bbgamma-gamma accessible in a limited region for tanbeta < 3.5 in 170 < m(A) < 350 GeV - see ATLAS MSSM summary plots, more studies under way.

    2. Is discovery in general 2 higgs doublet models understood?

      Systematic SUSY Higgs searches in CMS and ATLAS done only in MSSM framework - more general than mSUGRA - but even this is not yet fully investigated; right now work goes on H to sparticle decays and on qqH final states (SM and MSSM) potentially significantly constraining the couplings.

    3. Where can measurements from other machines (linear collider, Tevatron) be most usefully added to LHC studies for better understanding? We are interested in understanding the possible complementarity of lepton colliders and the LHC.