1. 21 March 2005LCWS 2005 – Chris Damerell 1 Vertex Detectors and the Linear Collider Chris Damerell Rutherford Appleton Lab What is their purpose? (still somewhat contentious, long after LCWS 1991!) General principles of vertex detector design, and prospects for ILC What specific technology to use? How not to be blown away by 10 9 pixels – electromagnetic interference, signal sampling, and other issues How can we get to ILC physics, from where we are now?

2. 21 March 2005LCWS 2005 – Chris Damerell 2 Dave Burke, LCWS 1991 Particle flow almost reveals the underlying Feynman diagram … What are vertex detectors for?

3. 21 March 2005LCWS 2005 – Chris Damerell 3 Vertex detector will tag b and c jets and tau leptons with high efficiency (reducing background, both combinatoric and from other multi-jet processes) Will also efficiently perform heavy quark sign selection, via measurement of the vertex charge (net charge of displaced vertex) – new from SLD

4. 21 March 2005LCWS 2005 – Chris Damerell 4 Particle flow plus efficient b and c tag plus heavy quark sign selection permits fully reconstructed production and decay distributions for a sizable fraction of Z 0 and W decays: Z 0 : 15% b-bbar + 12% c-cbar = 27% W + : 31% c-sbar leading to: enhanced information and increased statistics wrt leptonic channels different physics reach (eg for Z 0, coupling asymmetries A b =0.94, A c = 0.67, where A l = 0.15) For top quark decays, a powerful tool to measure the top quark polarisation

5. 21 March 2005LCWS 2005 – Chris Damerell 5 Not a case of hadronic versus leptonic channels: both are valuable Gudi Moortgat-Pick: We have about 100 parameters in the MSSM and need all observables … we need to explore the angular distributions in all decay channels – not only in the leptonic case This is largely uncharted territory, since the potential of vertex charge has not been widely recognised, but has recently become established by successful physics analyses in SLD Look at a few physics examples, starting with the simplest SM process, having the largest cross-section of all

6. 21 March 2005LCWS 2005 – Chris Damerell 6 e+e- q qbar differential cross-sections for L and R-polarised electrons at max sqrt(s), as probe of BSM processes Sabine Riemann, Fermion-pair production at a linear collider – a sensitive tool for new physics searches [LC-TH-2001-007] Sensitive to Z, leptoquarks, R-parity violating scalar particles, and extra spatial dimensions Requires efficient quark charge sign-selection out to large |cos theta|

7. 21 March 2005LCWS 2005 – Chris Damerell 7 For these channels, sensitivity extends to M D around 5 TeV For muon pairs, effects are much weaker (not measurable even with 1 ab -1 of data) Sabine Riemann: one example for large ED, sqrt(s) = 0.5 TeV

8. 21 March 2005LCWS 2005 – Chris Damerell 8 Chargino and neutralino angular correlations G Moortgat-Pick and H Fraas PRD 59 (1999) 015016 SY Choi, HS Song and WY Song PRD 61 (2000) 075004 G Moortgat-Pick et al hep-ph/0002253 (2000) TESLA TDR (2001) III-64 Publications so far have focused on leptonic channels eg Quark pairs were disregarded since the possibility of quark sign selection was not yet apparent Adding b-bbar and c-cbar will increases statistics and provide additional physics sensitivity (soon to be studied by G Moortgat-Pick)

9. 21 March 2005LCWS 2005 – Chris Damerell 9 Momenta in e+e- rest frame for associated neutralino production and in the rest frame of the decaying neutralino Following analysis from G Moortgat-Pick, A Bartl, H Fraas and W Majerotto hep-ph/0002253 (2000)

10. 21 March 2005LCWS 2005 – Chris Damerell 10 Production process Fwd-Bkwd asymmetries plotted for the decay electron in the process as fn of the gaugino parameter M1. Left and right plots correspond to different masses of the -/0/+ means left-/un-/right-polarised beam, electrons listed first, positrons second These asymmetries strongly constrain the selectron masses as well as the mixing properties of

11. 21 March 2005LCWS 2005 – Chris Damerell 11 Top quark polarisation – a general tool for physics top quark decays before hadronization, and long before its spin can flip So, polarisation of top quark at production is reflected in its decay not true for any other quark, where depolarisation effects during fragmentation wash out the quark helicities In t b W+, W+ c sbar, the top polarisation is best measured from the angular distn of the s-quark jet [E Ruiz Morales and M Peskin Proc LCWS 1999 (Sitges) and hep-ph/9909383] To distinguish between t and tbar, need to tag b and c jet, and sign-select at least one of them (so quark sign selection doesnt need to be super-efficient)

12. 21 March 2005LCWS 2005 – Chris Damerell 12 c sbar t W+W+ b sbar jet has 1-cos theta angular distribution wrt top polarisation direction, as does l+ in the leptonic W decays In complex events, expect the fully reconstructed hadronic decays to have lower background. In general, aim to use both leptonic and hadronic W decays Measurement of top polarization in scalar top and scalar bottom decays can be used to determine fundamental SUSY parameters: tan and the trilinear couplings A t and A b [E Boos et al, hep-ph/0303110]

13. 21 March 2005LCWS 2005 – Chris Damerell 13 Using all observables, the ILC makes available: Production angular distributions Decay distributions Beam polarisation asymmetries can then search for anomalous top production and decay form factors, including effects of t anomalous magnetic moment t b R W + decay (forbidden in V-A theory) anomalous t-tbar-Z couplings Michael Peskin LC lectures, March 2002 and TESLA TDR III-146

14. 21 March 2005LCWS 2005 – Chris Damerell 14 If no light Higgs, provides a means to probe strong EWSB at the ILC, via the process Tim Barklow, Proc Snowmass 1996 p 819 E Ruiz Morales and M Peskin, Proc LCWS 1999 (Sitges) 461 Need fully reconstructed top quarks, for suppression of competing processes Sensitive to new resonances at large sqrt(s); vector resonance (techni-rho) coupling to helicity conserving amplitudes and scalar resonance (techni-sigma) to helicity- flip amplitudes Spin of these resonances can be probed by measurement of top quark polarisation (again, by angular distributions of s-quark jets)

15. 21 March 2005LCWS 2005 – Chris Damerell 15 Other physics examples Higgs branching ratios [very important but not very challenging technically] Higgs decay angle distributions (where H decays to b-bbar, c-cbar or W-W*) Higgs self-coupling SUSY CP asymmetries Form T-odd triple product, [needs fermion sign selection) A Bartl et al, hep-ph/0306304 Sensitivity of b-bbar compared to mu pairs: needs event sample of only 1000 staus compared to 10 5 (benefiting from the larger branching fraction, and fact that A b = 6.4 x A e ) ----------------------------------------------

16. 21 March 2005LCWS 2005 – Chris Damerell 16 benchmark for charm tag (A Sopczak, A Finch, H Nowak LCWS 2004) Particularly challenging for small m between neutralino and stop masses; theoretically motivated: talk in SUSY session by Caroline Milstene, and Jonathan Fengs colloquium The totally unexpected [what we are all hoping for] The universe is not only queerer than we suppose, but it is queerer than we can suppose. JBS Haldane

17. 21 March 2005LCWS 2005 – Chris Damerell 17 Sincere thanks to many experts who kindly helped with this part of the talk: Juan Antonio Aguilar Saavedra, Dave Bailey, Alfred Bartl, Nicolo de Groot, Stefan Hesselbach, Olaf Kittel, Akiya Miyamoto, Gudi Moortgat-Pick, Michael Peskin, Sabine Riemann, Andre Sopczak, Su Dong, Peter Zerwas

18. 21 March 2005LCWS 2005 – Chris Damerell 18 General design principles Fixed target experiments were much easier! For the collider environment, the adequate vertex detector has yet to be built – hence numerous options and constant upgrades NA32, 1985

19. 21 March 2005LCWS 2005 – Chris Damerell 19 SLD, thanks to Su Dong

20. 21 March 2005LCWS 2005 – Chris Damerell 20

21. 21 March 2005LCWS 2005 – Chris Damerell 21 For the ILC, we can certainly do much better: R bp 12-15 mm, cf 25 mm at SLD Layer thickness 0.05-0.1% X 0, cf 0.4% X 0 at SLD [20 m of Si is 0.02% X 0 ] Point measurement precision at least as good as at SLD (approx 3 m) Resulting impact parameter precision will be m Compared with, at SLD: m [really helped by more open geometry, with longer lever arm provided by 5 layers compared to 3]

22. 21 March 2005LCWS 2005 – Chris Damerell 22 Generic long-barrel detector (TESLA TDR)

23. 21 March 2005LCWS 2005 – Chris Damerell 23 SiD vertex detector design concept – Norm Graf

24. 21 March 2005LCWS 2005 – Chris Damerell 24

25. 21 March 2005LCWS 2005 – Chris Damerell 25 TESLA TDR Purity means for quarks from Z 0 decays – its only one benchmark Case of charm tag with low light-quark background is interesting, eg for adding flavour tag to reconstructed Ws, to enable top polarization measurement Measuring flavour ID at ILC

26. 21 March 2005LCWS 2005 – Chris Damerell 26 Population in bins with charge +/- 2 is very sensitive to tiny tracking inefficiencies – a valuable monitor, important for particle- flow measurement of jet energies [In SLD, this study established that 1.5% apparent tracking inefficiency was real, not an error in the MC track multiplicity] SLD: Tom Wrights thesis SLAC-R-302 (2002) Measuring vertex charge At SLD

27. 21 March 2005LCWS 2005 – Chris Damerell 27 SLD –measurement of parity violation parameters A b and A c PRL 94 (7 March 2005) 091801 eLeL eLeL eReR eReR

28. 21 March 2005LCWS 2005 – Chris Damerell 28 Measurement of the branching ratio of the Z 0 into heavy quarks (PRD, to be published) 1/40 of LEP sample, yet measurement of R c is by itself more precise than the current world average ----------------------------------- Hope to do much better at ILC. Challenge is to associate each PV track with the PV, and each SV/TV track with the decay vertex Lowest momentum track in b decay chain is ~ 1 GeV/c, so multiple scattering is important – need small- radius, thin-walled beampipe and thin detector layers Full Geant 4 description of vertex detector, with correct treatment of cluster finding and realistic backgrounds is essential. This forms part of LCFIs MIPs to Physics project. Hoping to broaden the study beyond our collaboration – discussions Wednesday morning at SLAC Meanwhile, we have some indications using fast Monte Carlo SGV

29. 21 March 2005LCWS 2005 – Chris Damerell 29 Prospects for vertex charge at ILC Sonja Hillert, LCFI: preliminary study – could get better or worse - dont yet believe these figures!

30. 21 March 2005LCWS 2005 – Chris Damerell 30 Primary requirement – minimise leakage of neutral Bs into charged sample Charged sample (~40%) and part of neutral sample (~20%) will give clean quark sign selection Similar for charm, but in many cases (low background from light quarks) one can push tag efficiency by including 1-prongs, which are ideal for quark sign selection Sign-selection effic per b or c jet ~60%. For individual W decays, you have one b and possibly one c: for Z decays, use of both jets leads to ~84% overall effic. For t -> b c sbar, both b and c jets can be used to sign select, again leading to ~84% overall effic. Fot t-tbar, up to 4 jets available for sign selection: need only one good one, ~97% overall efficiency may be possible Use of charged kaons provided a further boost in SLD to sign select charm quarks in c jets or in the b decay chain. Any chance for ILC? [Bob Wilson has done some thinking about this. Also, advice from Jerry VaVra would be great.]

31. 21 March 2005LCWS 2005 – Chris Damerell 31 In 1981, LEP was hell-bent on a 10 cm RADIUS beampipe … (Villars workshop 1-7 June 1981) Disappointed, I followed other examples and turned to the New World SLC? Whats the beampipe radius? About the size of a drinking straw! A vital parameter – the beampipe radius

32. 21 March 2005LCWS 2005 – Chris Damerell 32 Such a time dependence is not inevitable: at LEP it went the other way! Their R_bp was reduced from 10.6 cm in 1991 to 5.6 cm in 1995 Maybe the ILC machine design will be a balance between European conservatism, American optimism and Asian realism, hence more stable In Europe, Nick Walker (ECFA workshop, Obernai, 1999) promised us a radius of 1.5 cm – but not a millimetre less! Dont let him forget this promise!

33. 21 March 2005LCWS 2005 – Chris Damerell 33 Really important for the machine people to design the FF system for minimal beampipe radius (though they cant of course reduce the pair background) What if 10 mm R bp had been possible at LEP and SLC? SLD would very probably have measured the B_s mixing parameter – and we are still waiting for that … At LEP, had they been able to flavour-tag every jet cleanly, would have reached a definitive conclusion about a Higgs boson in their mass range, with only a handful of events, on ZERO background As with any microscope, getting close helps. The physics potential has still to be investigated

34. 21 March 2005LCWS 2005 – Chris Damerell 34 What vertex detector technology for the ILC? Technologies Groups CAPBirmingham UPNSensor (Munich) CPCCD Bonn URAL DEPFETBristol USLAC FAPS Brunel UStrasbourg U FPCCD DESYTohoku Gakuin U HAPS Glasgow UToyama College of ISIS – edge readout Insubria U Maritime Tech ISIS – distributed readout KEK U of Mining and MAPS – transverse readout Lancaster U Metallurgy, Krakow MAPS-digital LBNLWarsaw U SoI Liverpool UYale U Macro-pixel/Micro-pixelMannheim U U of Hawaii sandwichMPI Munich (Halbleiterlabor) Nijmegen U NIKHEF (Amsterdam) Oregon U Oxford U Both lists are probably incomplete – apologies!

35. 21 March 2005LCWS 2005 – Chris Damerell 35 3.2 M-shell plasmons per m (17 eV) 0.6 L-shell plasmons per m (120 eV) 0.01 K-shell plasmons per m (1.5 keV) ~4 primary collisions m with wildly fluctuationg energy loss Final thermalisation yields one e-h pair per 3.6 eV deposited Sensor operating principles

36. 21 March 2005LCWS 2005 – Chris Damerell 36

37. 21 March 2005LCWS 2005 – Chris Damerell 37 Imagine p and p+material brought into contact at same potential Holes pour from p+, leaving a negative space-charge layer (depletion) and forming a positive space charge layer in the p material (accumulation) This space-charge must of course sum to zero, but it creates a potential difference, which inhibits further diffusion of majority carriers from p+ to p and incidentally inhibits diffusion of minority carriers (electrons) from p to p+ This barrier is thermally generated, but the penetration coefficient is temperature independent, and is simply the ratio of dopant concentrations. eg 0.1/1000, so 10 -4 - this interface is an almost perfect mirror! Minority carrier diffusion length ~ 200 m ------------------------------ ~ 0.1 m What epi-layer thickness? Prefer it thin, to avoid losing precision for angled tracks But not too thin, or lose tracking efficiency 20 m is about right

38. 21 March 2005LCWS 2005 – Chris Damerell 38 We can repeat this on the top surface – here the p-well can be used to implant structures (notably n-channel transistors), monolithic with respect to the detector layer below Positively biased n implants (reverse-biased diodes) serve to collect the signal charges, partly by diffusion, partly by drift in depleted regions created in the p-type epi layer Overlaying dielectric layers, and photolithographically patterned metal layers complete the toolkit for interconnecting the circuit Here you have the essentials of a MAPS pixel detector

39. 21 March 2005LCWS 2005 – Chris Damerell 39 Another variant – same below, but positively biased n-layer above, creating a full-area depletion region. Signal charge is pulled to a potential energy minimum buried channel about 1 m below the surface Metal or polysilicon gates control the location of stored charge in 1-D. Channel stop implants create orthogonal potential barriers Signal charge packets can be transported in turn to an output node, similar to that already seen on the MAPS device. This is a charge-coupled device or CCD To learn about all the beautiful options for ILC vertex detectors, read the slides from Session H last Saturday morning (12 talks)

40. 21 March 2005LCWS 2005 – Chris Damerell 40 How not to be blown away by 10 9 pixels (a) Electromagnetic interference Electron/positron beams traversing the IR radiate massive RF power (wakefields) Numerous imperfections (thin walled sintered Be beampipe, non-welded flanges, ports for BPMs,kicker magnet circuits, beam-size monitors, vac pumps, …) provide leakage paths for RF A linear collider is intrinsically more hostile in terms of beam-induced RF than storage rings The vertex detector (in which ~10 9 unamplified signal charges of ~1000 e- are transformed to voltage on the gates of tiny transistors within ~1 mm of the beampipe) is more liable to disturbance by this RF than most detector systems Beam-induced pickup disrupted the SLD vertex detector electronics for several s after each bunch Dangerous to assume it will be quieter at a machine with 10 times the energy and 10 4 times the luminosity of SLC, needing far more instrumentation to preserve its performance Problems may not be primarily related to RF from the beam – control systems, such as kicker magnet pulsing circuits active during the train may also be dangerous

41. 21 March 2005LCWS 2005 – Chris Damerell 41 Typical example: ideal CCD

42. 21 March 2005LCWS 2005 – Chris Damerell 42 Reality, during the bunch train: From SLD experience, signal charges stored in buried channel are virtually immune to disturbance by pickup. They were transferred in turn to the output node and sensed as voltages between bunches, when the RF had completely died away Could this also be done at ILC?

43. 21 March 2005LCWS 2005 – Chris Damerell 43 Bunch train: 2820 bunches @ 337 ns, every 200 ms If signals were accumulated throughout a bunch train, background would be totally unacceptable (except maybe for FPCCD of GLC Group) Seemingly need to read the detector ~20 times during train, at ~50 s intervals This may be like trying to hear someone whisper on a railway platform while an express train is roaring by. Why not simply wait? All detector options considered till recently suffered from this problem After ECFA workshop in Montpellier, November 2003, we came up with a new idea David Burt of e2V told us Its been done! Even better!

44. 21 March 2005LCWS 2005 – Chris Damerell 44 Pioneered by W F Kosonocky et al IEEE SSCC 1996, Digest of Technical Papers, p 182 Current status: T Goji Etoh et al, IEEE ED 50 (2003) 144 Frame-burst camera operating up to 1 Mfps, seen here cruising along at a mere 100 kfps – dart bursting a balloon Evolution from 4500 fps sensor developed in 1991, which became the de facto standard high speed camera (Kodak HS4540 and Photron FASTCAM) International ISIS collaboration now considering evolution to 10 7 – 10 8 fps version! ISIS: Imaging Sensor with In-situ Storage

45. 21 March 2005LCWS 2005 – Chris Damerell 45 charge collection to photogate from ~20 m silicon, as in a conventional CCD signal charge shifted into storage register every 50 s, to provide required time slicing string of signal charges is stored during bunch train in a buried channel, avoiding charge-voltage conversion totally noise-free charge storage, ready for readout in 200 ms of calm conditions between trains The literature is littered with failed attempts …

46. 21 March 2005LCWS 2005 – Chris Damerell 46 (b) Correlated Double Sampling CDS is a vital part of the strategy to avoid a data deluge, one which most technologies for the ILC vertex detector claim to employ Even if reading between trains, fluctuations in transistor noise and detector- related pickup sources will be present, as seen at SLD CDS - term given in early 70s to pedestal subtraction in CCDs used in astronomy and elsewhere to sense very small signals, where reset noise and 1/f noise would otherwise have imposed severe performance restrictions Same functionality was achieved in late 70s in CCD-based particle detectors, where the sparse data permitted resetting only between rows, hence faster sampling ERF - Extended Row Filter, was an important refinement added in SLD Beware of imitations

47. 21 March 2005LCWS 2005 – Chris Damerell 47

48. 21 March 2005LCWS 2005 – Chris Damerell 48 Extended Row Filter (ERF) suppresses residual noise and pickup:

49. 21 March 2005LCWS 2005 – Chris Damerell 49 SLD experience: Read out at 5 MHz, during quiet inter-bunch periods of 8 ms duration Origin of the pickup spikes? We have no idea, but not surprising given the electronic activity, reading out other detectors, etc Without ERF, rate of trigger pixels would have deluged the DAQ system

50. 21 March 2005LCWS 2005 – Chris Damerell 50 How can we get to ILC physics, from where we are now? Currently many groups are pursuing an expanding range of options for the ILC vertex detector All use silicon pixels, but there the similarity ends How to converge on (hopefully) two technologies for two large detectors? The ILC vertex detector community has informally undertaken to provide working ladders in test beams, circa 2010 (+ Overall detector collaborations should evaluate their results carefully, considering all performance criteria, including efficiency, spatial resolution, material budget at small angles due to mechanical supports and electronics at ends of ladders, robustness wrt EMI, etc Dont believe what any of us claim we can deliver!

51. 21 March 2005LCWS 2005 – Chris Damerell 51 Only then should they make their technology choices, deciding between long barrels, short barrels plus endcaps, etc etc Everyone who has participated in the R&D should be welcome to join one of the winning technologies for detailed design, construction, commissioning, and extracting the physics (ideally, a rerun in miniature of the ITRP process) [Several losers will change direction and find wonderful applications for their technologies, and may gain more than they lose!] In the meantime, we should continue building a world-wide community who would all enjoy working together I suggest we increase the frequency of our inexpensive world-wide phone conferences to maybe twice a year [Arlington TX 8 Jan 2003, Mumbai 14 Dec 2003] Overall detector collaborations should design for convenient access to the vertex and other small-radius equipment for several reasons, including future upgrades. A losing technology, could prove to be a long-term winner (as happened with CCDs at SLC) The SLD vertex detector was considered a jewel in the crown. This may also be true at ILC – the physics potential, combined with particle flow for jet energy measurement, goes far beyond what can be done at LHC

52. 21 March 2005LCWS 2005 – Chris Damerell 52 Can we be sure that silicon pixels will provide the best solution? Do you remember the front running technology for SLC in 1982? CCDs were regarded as a risky outsider. If you come up with a revolutionary new idea, please do follow it up! Dont be discouraged by the so-called experts! In the 70s, when most expertise in silicon detectors was in the hands of nuclear science people with string-and-sealing wax production facilities, the construction of semiconductor detectors was more like cookery than science … To these experts, the concept of CCDs as particle detectors seemed outrageous

53. 21 March 2005LCWS 2005 – Chris Damerell 53 SLC Experiments Workshop 1982

54. 21 March 2005LCWS 2005 – Chris Damerell 54 SOME EXPERT OPINIONS IN 1979 "Put such a delicate device in a beam and you will ruin it". "Will work if you collect holes, not electrons". "Far too slow to be useful in an experiment". "It's already been tried; didn't work". "It will work but only with 50% efficiency". "To succeed, you will have to learn to custom-build your own CCDs: investment millions". "At room temp it would be easy, but given the need to run cold, the cryogenic problems will be insurmountable". "May work in a lab, but the tiny signals will be lost in the noise (RF pickup etc) in an accelerator environment". However, Wrangy Kandiah from AERE, Veljko Radeka and Pavel Rehak from BNL, Joe Killiany from NRL, Herb Gursky from Harvard Smithsonian, Emilio Gatti from Milano and a few others were supportive