Page 1
Inclusive Diffraction at HERA From The ZEUS Experiment
New data from the running period 1999-2000.
The last period with the ZEUS Forward Plug Colorimeter (FPC) installed and the Leading Proton Spectrometer (LPS) installed.
Results from all three methods to extract inclusive diffractive events:
• Leading proton spectrometer
• The MX – method
• Large rapidity gap method
Bernd Löhr, DESY Page 1Diffraction 2006 , 5-10 September , Milos Greece
using the same dataset.
Presented by B.Loehr on behalf of ZEUS
Page 2
Bernd Löhr, DESY Page 2Diffraction 2006 , 5-10 September , Milos Greece
Extraction of diffractive events (I)
1.) Forward proton detection
1x L -> diffractively scattered proton ;
2p
L
2L
L
2T M
x
)x1(
x
p- t
the only method to measure the t-distribution
IPL x -1 x
Forward proton tagged events are not free of proton dissociation background at small xL values, respectivly at large xIP
ZEUS preliminary
Forward proton tagged events contain contributionsfrom Reggeon exchange at high xIP.
2.) The large (pseudo)rapidity (max) method
max minη = -ln tan(Θ /2)
No tracks or energy deposits in calorimeter for rapidities greater than max or at angles less than min.
Events tagged by a large rapidity are dominatedby diffraction but they contain contributions from proton dissociation and from Reggeon exchange.
Page 3
3.) The Mx-method
(i) Nondiffractive events :
Rapidity z
z
E+p1y = ln
2 E-p
Property of a produced particle
Uncorrelated particle emission between incoming p-direction and scattered quark.
partdN = λ = const.
dy
2Xb ln Mnondiff
2X
dN = c e
d ln M
(ii) Diffractive events :
diff2 2X X
dN 1
dM (M )n
At high energies and not too low MX
n 1
(iii) Nondiffractive + diffractive contributions
2Xb ln M
2X
dN = D + c e
d ln M For
2 2X 0ln M ln W
Fit slope b and c -> subtract nondiffractive events.
diff2X
dN const.
d lnM
Poisson distr. for y innondiffractive events
-λΔyP(0)=e
max miny -y20W =c e limit miny -y2
X 0M =c e
Bernd Löhr, DESY Page 3Diffraction 2006 , 5-10 September , Milos Greece
Extraction of diffractive events (II)
Page 4
The ZEUS MX-Analysis I
Bernd Löhr, DESY Page 4Diffraction 2006 , 5-10 September , Milos Greece
MC-simulation :
nondiffractive : DJANGOH
diffractive : SATRAP
proton diss.: SANG
SANG adjusted to fit data which aredominated by proton dissociation
Proton dissociation can bereliably calculated forMN > 2.3 GeV and has beensubtracted from data
The ZEUS MX-resultscontain contributions fromproton dissociation for masses MN < 2.3 GeV.
Example of lnMX-distributions forfour kinematical bins :
fit to data in the fit region fitted slope of of nondiffractive events
Data are presented for 9 Q2-bins, 7 W-bins, and 6 MX-bins at reference values Q2ref, Wref, MXref
Diffractive data selected by the MX-method contain proton dissociative eventsbut no contributions from Regge exchange
Page 5
ZEUS MX – data from 1998 – 2000 (I)
Mx 98-99 : *|Published data from1998-1999 period
Prel. Mx 99-00: |Preliminary results from1999-2000 period.Extension of Mx 98-99 analysisto higher Q2.
Mx 98-99 and Mx 99-00 analyses have common bin at Q2 = 25 GeV2
(ZEUS Coll., S.Chekanov et al. Nucl. Phys B 713, 3 (2005) )
Bernd Löhr, DESY Page 5Diffraction 2006 , 5-10 September , Milos Greece
Mx 98-99, Mx 99-00 (prel.)
Page 6
Mx 98-99: | *
Prel.Mx 99-00: |
Within syst. Errors goodagreement between Mx 98-99 and Mx 99-00 results
Mx 98-99 and Mx 99-00 analyses have common bin at Q2 = 55 GeV2
ZEUS MX– data from 1998 – 2000 : II
Bernd Löhr, DESY
ZEUS Mx 98-99, ZEUS Mx 00 (prel.)
Page 6Diffraction 2006 , 5-10 September , Milos Greece
Page 7
Fit W-dependence of inclusive DIS and inclusive diffractive DIS cross sections
Inclusive DIS:
Bernd Löhr, DESY Page 7Diffraction 2006 , 5-10 September , Milos Greece
For small x, F2 rises rapidly as x-> 0
-2 xc F
1)0( IP x
1 W
Inclusive diffractive DIS:
diffaX
0X
Np
W
Wh
dM
d *
4
a 1
diff
IP averaged over t
2
0IP W
Wln)1)(2(
ef(t) dt
d
t
tIP )0( )t( IPIP
tAe dt
d
for small t.
take 29.0
5.0- GeV syst.)( 0.5(stat.) 7.9 A
as measure by ZEUS LPS
Inclusive DIS and inclusive diffractive DIS are not describedby the same ‘Pomeron’.
Page 8
MX – dependence of the diffractive cross-section at constant W
Q2d/dMX shows a rapid variation with Q2
at low MX .
Q2d/dMX tends towards a common valueat higher MX .
This kind of scaling behaviour is reachedonly at higher MX for higher Q2 whenMX
2 becomes of the order of Q2 .
Bernd Löhr, DESY Page 8Diffraction 2006 , 5-10 September , Milos Greece
Mx 89-99
Mx 00
Page 9
Ratio of total diffractive cross-section to total DIS cross-section
Ratio plotted at W=220 GeV because only there the full MX range is covered by measurments
ZEUS Mx 00 (prel.)
FPC I
FPC II
Within the errors of the measurementsr is independent of W.
At W=220 GeV, r can be fitted by
r = diff(0.28<MX<35 GeV)/tot at W=220 GeV
r = 0.22 – 0.034.ln(1+Q2)
This logarithmic dependence of the ratioof total diffractive cross-section to the total DIS cross section indicates thatdiffraction is a leading twist process for not too low Q2.
Bernd Löhr, DESY Page 9Diffraction 2006 , 5-10 September , Milos Greece
Page 10
Diffractive Cross-Section and Diffractive Structure Functions
Bernd Löhr, DESY
42 D(4) 2em
2 IP2 2IP
2παd σ = [1-(1-y) ] F (Q ,t,x ,β)
dQ dtdx dβ βQ
ZEUS neglects the contribution from longitudinal structure function
H1 defines :
2D D Dr 2 L2
yσ = F F
1+(1-y)
sizable only at high y
If t is not measured : (LRG and MX-method )
222
IP
N ,)1(1x
QIPD(3)242
diffXp
3
x,FyQ
2
dddQ
d *
2
IPD(3)2IP2
X22
2
X
2X
diffXp
X
QxFxMQQ
4
dM
QW,Md
M
1 *
,,,
2N
If is interpreted in terms of quark densities, it specifies the probability to find in a proton which undergoes a diffractive interaction a quark carrying a fraction x = xIP
of the proton momentum.
2IP
D(3)2 QxF ,,
Page 10Diffraction 2006 , 5-10 September , Milos Greece
Page 11
xIPF2D(3) Results from the MxII-Analysis
Bernd Löhr, DESY Page 11Diffraction 2006 , 5-10 September , Milos Greece
Page 12
The ZEUS data support taking nT(Q2) = ng(Q2) = n1`ln(1+Q2/Q20) and nL= 0
Taking x0= 0.01 and Q20= 0.4 GeV2 results in the modified BEKW model
(BEKW(mod))with the 5 free papameters :
cT , cL , cg , n1T,g ,
ZEUS modified BEKW Fit
Bernd Löhr, DESY Page 12Diffraction 2006 , 5-10 September , Milos Greece
Page 13
xIPF2D(3) Results from the MxI and MxII-Analyses with BEKW(mod) Fit (I)
BEKW-fit (prel.): transverse qq contribution
longitudinal qq contribution transverse qqg contribution
sum of all contributions> 400 points, 5 parameters/nD = 0.71
Bernd Löhr, DESY Page 13Diffraction 2006 , 5-10 September , Milos Greece
Page 14
Bernd Löhr, DESY Page 14Diffraction 2006 , 5-10 September , Milos Greece
xIPF2D(3) Results from the FPC I +II MX-Analyses with BEKW(mod) Fit: II
xIPF2D(3) shows considerablescaling violations:
from positive scaling violations over near constancy tonegative scaling violations.
Result of the BEKW(mod) fit
Page 15
xIPF2D(3) Results from the FPC I +II MX-Analyses with BEKW(mod) Fit ( III)
Bernd Löhr, DESY Page 15Diffraction 2006 , 5-10 September , Milos Greece
xIPF2D(3) exhibits :
• a broad maximum around =0.5 which is due to the transverse qq-contribution,
• a steep rise towards small which is generated by the qqg-contribution,
• a longitudinal qqg-contribution which is sizeable only at very high and causes the structure function not to vanish at = 1.
-
--
-
Page 21Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Page 16
xIPF2D(3) Results from the FPC I +II MX-Analyses with BEKW(mod) Fit (IV)
Fixed xIP = 0.01
25 < Q2 < 190 GeV2
in one plot
The 3 contributions from BEKW(mod)fit for the above Q2 values plotted
qqg-contributions longitudinal qq-contributions- -
The BEKW model has an effective QCD-type Q2-evolution incorporated.
Page 16Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Page 17
),((),(,( IP2IP
2IR2IPIRIR
2IP
D(4)2 QFt),xfn QFt)xf F Regge fit :
ZEUS Results from the LPS I
Page 17Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Diffractive structure functions from ZEUS LPS mesurements
ZEUS LPS 00 (prel.) ZEUS LPS 00 (prel.)
Page 18
ZEUS Results from the LPS (II)
Page 18Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Comparison of LPS results with recent H1 FPS results
Not shown are the normalization Uncertainties of +12/-10 % for the ZEUS LPS data and +/-10% for the H1 FPS data.
The agreement is good
XpT
XpL
D **
R pp /
RD = 0 -> D(3)rIP
D(3)2IP x Fx
ZEUS LPS 00 (prel.), H1 FPS 06
Page 19
ZEUS Results from the LRG Method (I)
Page 19Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Events selected by : max < 3.0 and energy in the Forward Plug Calorimeter (FPC) < 1 GeV
Results of a Regge-fit Input parameters to the Regge -fit:
IR(0) = 0.75, BIR= 2.0 GeV-2
’IP = 0.0 GeV-2 , BIP = 7.2 GeV-2
Fit results :
IP(0) = 1.117 +/-0.005 +0.024 -0.007
ZEUS LRG 00 (prel.) ZEUS LRG 00 (prel.)
Page 20
ZEUS Results from the LRG Method (II)
Page 20Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
The Regge-fit gives a good description of the ZEUS LRG data with 2/ndf = 159/185
Comparison of ZEUS LRG data with LPS data
The ratio LPS/LRG = 0.82 +/- 0.01(stat.) +/- 0.03(syst)
Not shown is the normalization uncertainty of the LPS measurement of about 10%.
ZEUS LRG 00 (prel.)
ZEUS LRG 00 (prel.), ZEUS LPS 00 (prel.)
Page 21
ZEUS : Comparison of Results from the MX-, and LRG- Method (I)
In general good agreementfor xIP < 0.01
For xIP > 0.01 one can expect some differences from Reggeon contributions to the LRGdata.
Page 21Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
ZEUS MX 98-99, ZEUS MX 99-00 (prel.), ZEUS LRG 00 (prel.)
Page 22
ZEUS : Comparison of Results from the MX-, and LRG- Method (II)
Page 22Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
Comparison of ZEUS LRG with ZEUS MX results:
xIPF2D(3)
as a function of Q2
ZEUS Mx 90-99
ZEUS Mx 99-00 (prel.)
ZEUS LRG 00 (prel.)
ZEUS MX 99-00, ZEUS MX 99-00 (prel.), ZEUS LRG 00 (prel.)
Reasonable agreement,maybe there is a normalizationdifference.
Work is continuing to understand remaining differences
Page 23
Additional Material
Page 24
Comparison of Results from ZEUS LRG 00 with H1 LRG
Page 22Diffraction 2006 , 5-10 September , Milos GreeceBernd Löhr, DESY
xIP = 0.003 xIP = 0.01ZEUS LRG 00 (prel.), H1 LRG ZEUS LRG 00 (prel.), H1 LRG
The ZEUS LRG 00 data are normalized to the H1 LRG data