Selected Topics on Central Exclusive Production
V.A. Khoze (IPPP, Durham)
(based on works with A. Kaidalov, M. Ryskin, A.D. Martin amd W.J. Stirling)
aims: to list & expose the main uncertainties in the theoretical expectations for CEP rates,
to propose the measurements which will allow to restrict the predictions
By popular demand (ADR, Brian..)
Higgs sector study- one of the central targets of FP420 physics menu
1
revisited
refer. purposes
ExHume tuning
2
This is what matters for the CEP rates !
KMRS-04
3
 How reliable are the calculations ?
 Are they well tested experimentally ?
● How well we understand/model soft physics ?
● How well we understand hard diffraction ?
● Is ‘hard-soft factorization’ well justified ?
 What
else could/should be done in order to
improve the accuracy of the calculations ?
So far the Tevatron diffractive data have been Durham-friendly)
clouds on the horizon ?
Theory side
or
- •Hard rescattering corrections to CEP, (BBKM-06)
(agreement with KMR at the TPF-level)
• New papers/talks by GLM-07 ( Strikman et al, 06-07 )
4
5
Absorptive effects and soft survival factors
(J.D. Bjorken, 1992)
6
15 years on
Available data on soft diffraction at high energies are still rather fragmentary.
Theoretical models contain various assumptions and many parameters.
Durham models are tuned to describe available ‘soft’ diffractive data at
high energies and predict the total, elastic, SD and DD dissociation cross sections
which can be tested at the LHC.
Durham models allowed to make predictions for the CEP jj and diphotons
at the Tevatron which are broadly confirmed by the data , more tests to come.
A way to compare the models : 
S 2 (s, b)  / b2
with the same exponential slope b in ME
(an agreement within a factor of 2 is still a miracle! )
MC model predictions should be confronted with the CDF data ( e.g. proton spectra in SD)
At the moment- no need to revise the Exhume default numbers,
but we have to be opened-eyed.
( note, on the theory side –downward tendency (stronger absorption effects), but
CDF data rather favour upward )
Recall: in reality survival factor is not universal (depends on the nature of the hard process, 7
kinematics, selection criteria, acceptances, pt- spread….)
PDF’s DEMOCRACY
 KKMR -04 – factor of 1.5-2 difference between CTEQ6M and MRST02
 very recently, CLP-07, factor of 4 difference between CTEQ6L1 and MRST2002NLO,
CTEQL1 7.38 fb for SM Higgs.
Here we are on the conservative side, but further studies and tests are needed
Higher-Order QCD effects
Uncomfortably large higher-order QCD effects
in the case of exclusive
processes, exemplified by the Sudakov effect.
Seen now in the dijet exclusive data.
(Thanks to CDF


)
Further serious theoretical studies needed, NNLO Sudakov ?
Self-consistent combined treatment of higher order effects in unintegrated
struct. functs and in the hard cross-section – requires detailed studies
8
Semi-enhanced hard rescattering and softhard factorization
“enhanced”
correction
to sH(excl)?
enhanced absorption,
discussed first KKMR-01
in the diffractive dijet
context
Bartels,Bondarenko,Kutak,Motyka-06
KMR-00(07): use 2(3)-channel eikonal
+ ‘soft’ enhanced contributions
(A. Martin’s talk)
used pert.thy.corrn could be
large and s H(excl) modified ?
KMR-06  arguments for small effect
9
On top of KMR theoretical
arguments
New ZEUS data
g
g
Leading neutron
prod. at HERA, Zeus
g
KKMR ‘06
gap due to p exchange
~ exclusive Higgs
eikonal
g
yi > 2 – 3
enhanced
g
g
g
correction prop. to rap. interval
prop. to g energy
(negative)
Prob. to observe leading neutron
must decrease with g energy
But expt.  flat
 small enhanced correction
sSD may
change (flat) behaviour at the LHC if enh . contr. is large
10
EXPERIMENTAL CHECKS
(Yesterday and Today)
Up to now the diffractive production data are consistent with
Still more work to be done to constrain the uncertainties
• CED high-Et dijets
(CDF: Run I, Run II)
K(KMR)S
results
data up to (Et)min>50 GeV
• ‘Factorization breaking’ between the effective diffractive structure functions measured
at the Tevatron and HERA.
(KKMR-01 ,a quantitative description of the results, both in normalization and the shape of the distribution)
•The ratio of high Et dijets in production with one and two rapidity gaps
• Preliminary CDF results on exclusive charmonium CEDP.
•Energy dependence of the RG survival (D0, CDF).
•
CDP of γγ (….pp,) , PRL paper
(in line with the KMRS calculations)
 Leading neutrons at HERA
11
12
Available CDF data on
proton spectra are well
described by KMR model
Governs the rate of the pile-up backgr.
MCs should be compared with the CDF data
13
Higher sensitivity to the
parameters of models
for Soft Diffraction
y=-ln , =(1-x)
( also for calculations of the pile-up backgrounds)
14
Exclusive dijet monitor & Interferometer
CEP of diphotons (rate permitting) would provide an excellent
combined test at M>10-20 GeV (better accuracy!)
Dijet rate- combined effect of all basic ingredients (Surviv, Sudakov, pdfs, Enhanc. Absp)
( ET > 10 GeV)
ET-dependence -dominantly Sudakov (+anom dimens),
weaker dependence on Surviv.
At low ET- higher sensitivity to the Enhanced Absorption
Correlations between proton transverse momenta, azim. distribts
Practically insensitive to pdfs and Sudakov effects.
High sensitivity to soft model parameters.
Proton opacity scanner
(KMR-02, also Kupco et al-05, Petrov et al -05)
Advantages
• Comparatively high rate (3 orders of magnitude higher than for the Higgs at the same ET).
• Possibility to separate different effects and to restrict different uncertainties by
studying the same process
15
16
KMR-02
Correlations
17
KMR-02
High ET central jets are not required (in principle)
18
weak dependence on (t) – integrated effect
19
20
rich diffractive
structures
21
22
Uncertainties in the non-PU background calculation
Myths
For the
bb
channel bgds are well known and incorporated in the MCs:
Exclusive LO -
bb
production (mass-suppressed) + gg misident+ soft & hard PP collisions.
Reality
The complete background calculations are still in progress
(uncomfortably & unusually large high-order QCD and b-quark mass effects).
About a dozen various sources (studied by Durham group)
 admixture of |Jz|=2 production.
 NLO radiative contributions (hard blob and screened gluons)
On top of MC studies
(Andy, Marek et al. )
 NNLO one-loop box diagram (mass- unsuppressed, cut-non-reconstructible)’
 ‘Central inelastic’ backgrounds
 b-quark mass effects in dijet events – still incomplete
potentially, the largest source of theoretical uncertainties!
23
24
25
(+ n soft gluons)
26
Should be (strongly) reduced by
the existing cuts : mass matching,
azimuth. correlations etc
More MC studies needed
27
Conclusion
We are now at the qualitatively new stage when the theoretical predictions
for the CEP cross sections have reached the level of a factor of 3 accuracy.
So far Durham group has been able to describe/predict the diffractive data.
Essential improvement of the accuracy will require a lot of work and may not happen
until the LHC experiments come FORWARD and produce the data (already) in the early runs.
This will not be easy. It is not like a walk in the park
(J.D. Bjorken 1992)
LET THE DATA TALK !
28
BACKUP
29
In reality KMR –calculational procedure is (much) more
complicated
To account for the effects of the screening corrections the calculations are performed
in impact parameter bt –space
For illustration in a single-channel eikonal approx. hor the process
is the Fourier transform to impact parameter space
The soft rescattering effects are denoted only symbolically by a factor
in the
effective PP –luminosity (even for the integrated over proton transverse momenta quantities)
(KMR-hep-ph/0111078)
In practice, there is no factorized form, and
should be viewed as the soft
survival factor appropriately averaged over the diffractive eigenstates (2,3 channel eikonals..)
30
In the topical case of
the introduction of the
from  0.02 to 0.026
production
and angular correlations raise effective
( KKMR hep-ph/0307064)
There are other assumptions.
for example, factorization of the unintegrated distributions
still schematically
+
correlation effects +…
31
model B(2) as compared to the elastic data
secondary Regge poles can
contribute
32
Difference on the level of only a factor of 2 is still a miracle
!
(LL)
 S ( BBKM TPF
2
0.024
33
34