The ratio of SDD to SD production is measured for pbar p interactions which produce a leading antiproton in the Roman pots and a central rapidity gap (overlapping eta=0) with width Delta eta > 3 at sqrt(s)=1800 and 630 GeV. Our results are compared with central pseudorapidity gap fractions measured in minimum bias pbar-p collisions and with theoretical expectations.

This analysis closely follows our double diffraction (DD) study (web page, PRL 87, 141802 (2001)). We look for Roman-pot-triggered events with a central rapidity gap (overlapping eta=0). We do this as opposed to looking for the largest rapidity gap in an event (other than the forward gap associated with the leading antiproton) because the latter method would be more likely to be biased by detector effects such as cracks or inefficiencies in the calorimeter. In the method we choose, we look for the first particle (track or hit calorimeter tower) on each side of eta=0, and define (eta_max) eta_min as the eta of the first particle in the (anti)proton direction. The width of the rapidity gap is then Delta eta = eta_max - eta_min.

Note that we require the Beam-Beam Counters (BBC's) (3.2<|eta|<5.9) to be hit on the outgoing proton side of the detector in order to exclude double pomeron exchange (DPE) events. (If there are no particles found in the detector with |eta|<3.2, then |eta_max(min)| is defined to be 3.3, an overflow bin. If there are also no hits in the BBC on the outgoing pbar side, eta_min is taken to be -3.4.)

We look only at events with no more than one reconstructed vertex. Additional pbar-p interactions are likely to spoil any rapidity gaps. Many events with a central gap will not have a vertex found.

The MC distributions are normalized as described in the next plot.
Note that the SD events alone cannot account for the events with large |eta_max(min)| observed in the data. When the SDD events are added, the MC distribution reproduces the data well.
1800 GeV ps fig including caption eps (no caption)
630 GeV ps fig including caption eps (no caption)

For events with a gap overlapping eta=0, this is essentially the gap width; note however that included in the bins for Delta eta^0 > 3.2 are events in which only the BBC (nothing) is hit on the outgoing p(pbar) side of the detector and |eta_max(min)| is defined to be 3.3(3.4)
The normalization of the SD and SDD MC distributions were fit to this distribution.
1800 GeV ps fig including caption eps (no caption)
630 GeV ps fig including caption eps (no caption)

Preliminary SDD/SD ratios for central rapidity gaps (overlapping eta=0) of width Delta eta > 3 are (corrections made include the efficiency for SDD events triggering the BBC-p array (~75%) calculated from a MC simulation):

experimental gap: no tracks or hit calorimeter towers above set eta-dependent E_T thresholds within a region which includes eta=0 and has width Delta eta > 3

exp SDD/SD(1800 GeV, Delta eta^0 > 3) = 0.192 +/- 0.001(stat) +/- 0.032(syst)

exp SDD/SD( 630 GeV, Delta eta^0 > 3) = 0.184 +/- 0.001(stat) +/- 0.042(syst)

nominal gap Delta eta = ln(s/M_1^2M_2^2) (units of GeV) corrected from experimental gap using MC

nom SDD/SD(1800 GeV, Delta eta^0 > 3) = 0.157 +/- 0.001(stat) +/- 0.028(syst)

nom SDD/SD( 630 GeV, Delta eta^0 > 3) = 0.134 +/- 0.001(stat) +/- 0.030(syst)

and for all gaps of width Delta eta > 3 are:

all SDD/SD(1800 GeV, Delta eta > 3) = 0.226 +/- 0.001(stat) +/- 0.038(syst)

all SDD/SD( 630 GeV, Delta eta > 3) = 0.187 +/- 0.001(stat) +/- 0.041(syst)

We extrapolate from central to all gaps using the differential cross section from Regge theory based on the triple Pomeron amplitude and factorization.

ps fig with caption eps (no caption)

Mary Convery, Dino Goulianos
The Rockefeller University
Date blessed: 3/7/02

last updated 3/28/02 convery@rock16.rockefeller.edu