1

QUARTIC, A TOF for ATLAS/CMS Forward Protons

• You didn’t know that ATLAS+CMS had forward protons?

• FP420 = Forward Protons 420m downstream of CMS & ATLAS (joint R&D project)

• Physics is

p H/WW/ZZ p• Measure p’s with very high precision within 3 mm

of beam• Gives mass and quantum numbers of central state.

Andrew Brandt (UTA) Jim Pinfold (Alberta), Mike Albrow (FNAL)

2

Tracker, perhaps 10 layers per station, 3 stations, 3D silicon(edgeless, 10um res., fast, rad hard) .... Hawaii/SLAC + Brunel

TOF counter, quartz fiber Cerenkov + MCP-PMTs

Vacuum mechanics: Detectors in secondary vacuum, space constraints between beam pipes, not room for roman pots. High precision movement (microns), reproducibility, fail-safe.

Funding, approvals, etc.

FP420 Challenges

3

n=1 n>>1

Cerenkov Effect

Use this property of prompt radiation to develop a fasttiming counter

particle

4

It’s been done!

Can’t put our PMT in 7 TeV beam!

5

Preliminary UTA drawing of Mike Albrow’s concept for a fast time resolution Cerenkov counter:

proton

Microchannel plate PMT

Initial design used 2 mm2 rods, but not enough light, this drawing shows

6mm2 rods

QUARTIC

z=c(TR-TL)/2 z (mm) =0.21 t (psec)(2.1 mm for t=10 psec)

6

Where do Protons go at 420m

3mm

x<2.5 cm

120 GeV Higgs courtesy Peter Bussey, Manchester (irony of particle physics—easier to get plot of protons from Higgs than single diffraction)

xx

y

x

7

GEANT4 Simulation (Alberta)GEANT4 Simulation (Alberta)

• A GEANT4 simulation is well underway (see GEANT4 produced graphics above.

• The detector simulation includes:• Tracking of Cerenkov photons to the MCP-PMT through the medium.• Wavelength dependent refractive index of the medium • Wavelength dependent attenuation of the photons• Wavelength dependent reflectivity of the aluminium reflector• Timing of photons from generation to the MCP-PMT • The effects of coupling grease (if necessary)

8

QUARTIC Background Rejection (UTA)1) 2 single diffractive protons overlayed with a hard scatter (1% of interactions have a proton at 420m)

97.4% of events primary vertex and fake vertex from combining proton times more than 2.1mm (1) apart ; 94.8% if 20 psec

2) double pomeron overlayed with a hard scatter

97.8% of time vertices more than 2.1mm apart; 95.6% if 20 psec

3) hard SD overlayed with a soft SD

95.5% of time primary vertex and fake vertex more than 2.1mm apart; 91.0% if 20 psec

9

Cerenkov Light in Fused Silica (UTA):

UV is important! 640-650 total pe’s : 130 pe/6mm rod

λ #PE QE #p*Q E Θc n

180-250 1652.6 15.70% 259.5 49.6 1.544

250-350 1148.7 18.00% 206.8 47.8 1.490

350-450 624.7 19.90% 124.3 47.2 1.471

450-550 394.3 11% 43.4 46.9 1.464

550-650 271.1 1.50% 4.1 46.7 1.458

total 638.0

#P ave θc ave L Q E ave #P.ave*QE ave

4161.6 48.8 3.99cm 15.57% 648.0

maybe we should call it Fusstic2

2 2

1

# 2 sin( ) 1/pe L c d

10

0.01

0.01

Single n=1.52 c=49;7.4% of pe’s in 10 psec21.3% in 50 psec

over including QE1.9% of pe’s in 10 psec19.1% in 50 psec

Preliminary Time Distributions:

50 psec

50 psec

red = totally internally reflected lightgreen = extra light if aluminized

11

Next Steps• Preliminary design studies are promising

• Need to learn more about MCP’s; preliminary indications are Burle tubes are currently only option due to size of active area

• Burle 85021 600 has 1.5 mm pixels could give very useful x-segmentation for measuring multiple protons in same detector, but it seems

that best timing using single MCP_OUT channel (1/tube) (Wilma Raso)

• Starting to think about electronics (Alberta)

• Test beam at Fermilab summer 2006