1. Forward Shower Counters FSC in CMS
What: We will install a set of scintillation counters around both
outgoing warm, accessible beam pipes at CMS, ~ 60 m – 90 m
They detect showers produced by v. forward particles in pipe +
Very simple, low tech: 8 PMTs on each side in 4 pairs.
35mm PMT
Between 2 MBX magnets
25 cm x 25 cm scint.
Schematic of
counter pair.
Cut-out depends
on location
Four locations: z (m) = 59.2, 79.1, 88.5 m + (maybe) 22 m

2. tagging source of PKAM* events
Simulations by
Steffen Muller
FSC can be very good at
tagging source of PKAM* events
*Previously known as Monster.
Beam Gas Inelastic
Beam Halo
Beam Gas Elastic
Origin and contributions of beam-
induced backgrounds at BCM1F,
mostly BGI (vacuum quality!)
Peaks ~ residual gas or TCT:
BH + BGE lost on TCT (150m)
Possible initial location
before Q1 (z ~ 22 m) to help
this issue.

3. Physics, especially diffractive in no-PileUp interactions
(a) As veto in Level 1 diff. triggers to reduce useless pile-up events.
(b) To detect rapidity gaps in diffractive events.
(c) Measure low mass diffraction and double pomeron exchange.
(d) Measure σINEL (if luminosity known, e.g. by Van der Meer)
(e) Help establish exclusivity in central exclusive channels
(f) To monitor beam halo on incoming and outgoing beams.
(g) To test forward flux simulations (MARS etc.)
(h) Additional Luminosity monitor.
(i) Info on radiation environment for future (?) proton spectrometers
Beam monitoring etc, parallel uses:

4. 2009 JINST 4 P10001
FSC
Do not see primary particles, but showers in pipe and other material.
Simulations of 10 stations (in z) … we propose 4 most effective.

5. IP5 1A,B 2A,B 3A,B Incoming protons Outgoing protons
LOCATION, LOCATION, LOCATION!
59.32m
71.90 m after 3rd MBX/6
4.250 m
IP5
MBXW
1A,B
2A,B
SIDE VIEW
6th MBX magnet
88.55 m
50 m straight section
25 cm diameter pipe
Incoming protons
Same pipe
3A,B
Outgoing protons
MBX magnets separate
incoming and outgoing beams
& ANYWHERE
127.5 or m ?
or 22 m before TAS

6. Bunches spaced by 50 ns = 15 m dt = 3.33 ns x z(mod 15m) e.g.
z = 59.3 m dt = 2.3 ns
z = 71.9 m dt = 10.3 ns
z = 88.6 m dt = 4.6 ns
Z1, bunches
Simultaneous
As at z = 0
time
Z2, bunches
separated in t z

7. Some CDF Beam Shower Counters BSC experience
Zero-bias data: No tracks or Central Calo
cf Events with tracks + vertex
CDF Preliminary
CDF Preliminary
For no-interaction. (no tracks, calos clean)
max ADC count in 4 PMTs
of BSC < |η| < 5.9.
Interaction plots require also hit
in adjacent PMTs.
BSC3 behind separation dipole,
6.7 < |η| < 7.4
When one of a pair has signal,
so does the other (showers!)
4 CDF PRL’s depended on these counters (so far)

8. CMS NOTE-2010/015 Approved by CMS MB for Jan-Feb 2011 installation.
“Limited approval” :
Go ahead without detracting from
necessary shutdown work.
Most value is 2011 running
& when <n/x> < ~ 5
(Do not expect to use > 2012)

9. Using HF as Rapidity Gap detector / noise levels.
Nicolas Schul
Signal-Noise separation
in calorimeters:
Hottest PMT in regions:
0-bias with tracks
0-bias with no tracks
Unpaired bunches
Zoom in
(later data)
Single collisions
(no-PU)
with HF gap
HF veto very effective at killing
PU and most inelastic collisions.
99% of 0-bias no-track events
have no HF channel > 4.2 GeV

10. We already have exclusive μ+μ- and π+π- candidates

11. Efficiency for FSC detecting forward particles, GEANT (Lamsa, Orava):
Low : > 20% for > 9.0, > 60% for 9.5 < < 11.5
N.B. Single particle efficiencies. For p*  p π π (e.g.) combine the effiencies for ε(p*) π+ π0
FSC efficiencies vs η at pT = 0.5 GeV/c
(p & n dominate)
CMS (& ATLAS) currently blind between η = 6.4 (CASTOR)
and beam rapidity (yp = 7 TeV) except ZDC (neutrals).
Cannot distinguish most diffractive/non-diffractive events.

12. Efficiency SDE vs Mass:
FSC & others
FSC alone
ZDC alone
10 GeV
Generated diffractive mass (PYTHIA/PHOJET)
as log(MX), MX in GeV/c2,
cf to calculated from rapidity gap edge:
(a) full η coverage
(b) η < 4.7 (no FSC)
Below 10 GeV/c2 FSC contain most particles
>4 hits in FSC or > 1 track in HF
or CASTOR or ZDC(min)

13. Efficiency for detecting low mass single diffraction
No pile-up, rest of CMS empty (= noise), with signals in
ZDC, or [HF or CASTOR], or FSC (>4 hits)
Unique to CMS+FSC

14. Central events (0-bias trigger) with forward rap-gaps
(FSC, ZDC, CASTOR, HF)
studied for generic Double Pomeron Exchange processes (~ 10 μb)
Low mass DPE:
HF VETO
HF VETO
FSC hits
FSC hits y
CENTRAL STATE
Even without seeing quasi-elastic protons,
gaps >~ 4 units select D IP E

15. What about total inelastic cross section σINEL? Direct measurement.
σTOT if you know σEL ? (& vice versa!)
Can measure rate of totally empty events, P(0), in 0-bias trigger
Without FSC we miss the low mass diffraction that give hits
only with |η| >~ 6, or M <~ 5 GeV/c2 This is several mb!
With FSC, P(0) only faked by events with all particles in cracks
(can study with fake cracks) or inefficient regions (small);
and inefficient because of noise (we can study with data).
FSC fills 2 large cracks: 7.5 <~ |η| < ~ 9.5
Check independence
on beam conditions,
luminosity, …
η a poor variable here!

16. What is the Effective Lumi (No-PU, single interactions) vs Luminosity?
How small cross sections can we study without pile-up?
For simplicity, take L flat for 10 hours (~ true day)
Integrated L=1 per day
A possible 2011
scenario:
1 month = 25 stores
averaging 10 hours
flat L,
σinel = 80 mb
105 s-1
1 kHz
Total delivered 5.6/fb
Total Leff (no PU) 180/pb
pb σ OK
Typical GXG σ ~ 10 μb ~ 10-4 σinel
 L1 Rate ~ 100 Hz [1,2] Hz [8,9]
L = 2.5 x 1033 cm-2s-1

17. Other scenarios (e.g. from Trigger Review):
cf recent situation
Say 4 months of each of these:
Total delivered luminosity = 4 x x 900 = 5.4/fb
Single interaction (no PU) luminosity = 4 x x 28 = 200/pb
This is same as previous slide “guess”.
Note : When <n/x> <~ 5 it is still useful.
When <n/x> >~ 6 even at the end of a store it is not.

18. χb0 Does FSC have anything to do with possible future HPS ?
(High Precision Spectrometers for p + p  p + X + p; X = H, W+W-, etc.)?
Indirectly YES, improving measurements of related processes t b q γ
Jet H
χb0
SMH(130)
σ ~ 3 fb
Needs high L
and HPS
(not FSC!)
See next slide
Level 1 triggers:
GAP-μμ-GAP
GAP-2EM-GAP
GAP-2JET-GAP

19. Also measured excl: μ+μ-, J/ψ, ψ(2S), Υ, χc
Central Exclusive Studies in CDF : requiring all detectors empty, incl. BSC
except signal itself.
IP + IP → γ + γ
M(γγ) up to 20 GeV
Also measured excl: μ+μ-, J/ψ, ψ(2S), Υ, χc
QED γγ→e+e-
M(ee) up to 75 GeV/c2
IP + IP → Jet + Jet

20. Measurement of Exclusive DiJets
Prediction of ExHuME:
14 TeV, |η| <~ 3
Scaling down by ~ 10 for
7 TeV, |η| < 2.5 : pb
Integral 110 – 140 GeV
= 40 pb {gg}, 3.5 fb {bb}
= {gg} events in 500/pb
& ~ 2 {bb} events
Thanks A.Pilkington, V.A.Khoze
We can check these predictions, spectra and light quark suppression,
angular distributions, b-tagging …
Exclusive dijets are only gg or 10-4) ; b+g is forbidden.
~ All b’s come from gluon splitting. Hence small background to
Tevatron observations:
CDF and D0 each have a few exclusive JJ events > 100 GeV

21. γγ → e+e-, μ+μ-,τ+τ- (?), W+W-
Two Photon Processes
γγ → e+e-, μ+μ-,τ+τ- (?), W+W-
Observed in CDF (using BSC)
γγ → μμ cross section very well-known, proposed as luminosity calibration
But did one or both protons dissociate? Different cross section.
FSC can measure dissociation probability and check factorisation.
How does pT(μμ) depend on dissociation or not?
Undissociated p’s : momenta very well known (beam spread):
Calibration of forward proton spectrometers (HPS).

22. CDF search for exclusive Z (BSM if seen)
12 candidates M(l+l-) > 40 GeV/c2 without using BSC.
4 events have BSC hits (all >= 2) killing Z candidate.
(Still possible Z photoproduction + p-dissociation, but only 1 event!)
pT(e+e-) = 2.44 GeV/c
M(e+e-) = 92.1 GeV/c2
CDF II Preliminary
π – Δφ (rad) 1o
pT(l+l-) < 1.14 GeV/c
We have exclusive dilepton
candidates in CMS,
but are they “elastic” (no p-dissociation)?

23. Schedule:
By end Winter shutdown, mid Feb, have all infrastructure installed
(esp. Stands, cables, DAQ) and either all counters, or many …
any remaining can then be installed in short (day) accesses.
Commissioning & add to triggers only when all good

24. Physics, especially diffractive in no-PileUp interactions
SUMMARY
We propose to install a set of scintillation counters around both
outgoing beam pipes at CMS, ~ 60m – 100 m
Physics, especially diffractive in no-PileUp interactions
(a) As veto in Level 1 diff. triggers to reduce useless pile-up events.
(b) To detect rapidity gaps in diffractive events (p or no-p).
(c) Measure low mass diffraction and double pomeron exchange.
(d) Measure σINEL (if luminosity known, e.g. by Van der Meer)
(e) Help establish exclusivity in central exclusive channels
(f) To monitor beam halo on incoming and outgoing beams.
(g) To test forward flux simulations (MARS etc.)
(h) Additional Luminosity monitor.
(i) Info on radiation environment for future (?) proton spectrometers
Beam monitoring etc, parallel uses:
MORE PHYSICS
LOW COST
*Subject to support approval by LHC
ZERO RISK*

25. Back Ups

26. Technical Implementation
Scintillator paddles + light guides from INFN,Trieste (Penzo), cut, wrapped, tested
PMTs + bases, shields available from Fermilab.
Counter assembly + tests (light leaks, noise, source) at CERN, Dec+Jan.
Stands (standard LHC) + cables (16 HV + signal) : CERN- EN-MEF (A.Ball), Jan.
Signals split: (a)  BSC clone  LHC ops (b) ADC +{ Disc  L1 trigger}
4 LVDS inputs to L1 trigger
(each is OR of 4 PMTs)
(ZDC)
~ Adding 6 channels to
ZDC / HF readout As
& HF
Commission with collisions
Simulate L1 triggers, Rates(L,n)
Test with collisions
Test HLT offline
Implement
[VETO+]*[VETO-]*[VETO-HF] [2EM.OR.2μ.OR.JJ]
[+]*[-]*[VETO-HF] … etc.
HLT: Track count or central calo limits etc.
E.g.

27. Beam line IP5  ZDC (< 140 m) … used in MARS simulations
z-locations (not sizes!) FSC 1-4

28. Implementation Only 16 PMT channels! People : Hardware:
Martin Gastal , Gabor Veres… CERN
M.G.Albrow (contact person) Fermilab
A.Penzo INFN Trieste
P.Debbins, E.Norbeck, Y.Onel, I.Schmidt Univ.Iowa
O.Grachov, M.Murray Kansas Univ.
Arno Heister
A.N. Others showed interest (co-authors of note)
Hardware:
Scintillation counter paddles, light guides, cut, polished, wrapped, tested.
Several already tested in test beam and at CERN. INFN Trieste
Photomultipliers, housings, bases Fermilab
Stands, installation in tunnel CERN (EN-MEF)
Cables (HV+Signal), purchase, laid in tunnel to USC CERN (EN-MEF)
Signals split  BRM + CMS and BRM electronics CERN (EM-MEF)
DAQ (CMS): Trigger and ADCs (As ZDC, HF) U.Iowa + Kansas U.
Software and monitoring
CMS DAQ side: U.Iowa + Kansas +
BRM side: CERN
Only 16 PMT channels!

29. Summary