Minutes of the CALCOM meeting, Friday August 21, 1998 ===================================================== Agenda: ======= C. Loukachine - Time-of-Flight offsets H. Aznauryan - Some problems when using the beam rf for particle id V. Burkert - Kinematics corrections and spa corrections for acceptance holes A. Skabelin - Beam position and vertex reconstruction K. Egiyan - Empty target analysis A. Stavinsky - More on empty target C. Smith - ep -> ep-pi0 analysis in Delta region K. Joo - Comparison of various acceptance methods for ep -> ep-pi0 Excecutive summary: ------------------ This meeting showed progress on several fronts: (1) Time dependent shifts in the TOF/RF calibration need to be carefully watched. More frequent calibrations of global RF shifts may be necessary (C. Loukachine) (2) The shift in the vertex reconstruction from sector to sector is now understood in terms of an offset of the beam in CLAS by about 5-6 mm. (A. Skabelin) This is in accordance with survey data of the target position. (3) H. Aznauryan showed that the accidental rate in CLAS is very small (10**-4). (4) We are beginning to better unnderstand the empty target rate in terms of contributions from the downstream end caps (K. Egiyan, A. Stavinsky). (5) The missing mass resolution in CLAS was shown to be better than one obtains from simply using RECSIS results (V. Burkert). (6) The analysis of ep->ep-pi0 data in the Delta region is approaching publicly presentable results (C. Smith, K. Joo) Volker Burkert ----------------------------------------------------------------------- Individual reports: C. Loukachine: ------------- RF calibration: It was found that RF offset calibration was changed with the fixes in tracking code and geometry maps. The largest deviation was about a half of ns. The new calibration of RF offset was done in about 20 time points of e1 period, the constants are put in the map. The RF offset will be monitored and archived during the cooking run by run. The map should be updated after each cooking pass ( my opinion ). H. Aznauryan: ------------- Accidental events in CLAS - problems when using the beam rf for particle id. Level of accidental coincidences in the CLAS were studied using 4 GeV e1 data. Events with at least 2 particles in the final state were chosen, where one of the particles was an electron and the second - a charge pion (negative or positive). Pions were identified independent of time-of-flight measurements using only dE/dX in the scintillator counters. Then difference of time-of- flights calculated using momentum and pion mass and measured by scintillator counter was studied. This difference was investigated for negative and positive pions and for cases when start time (electron time) was - or was not - corrected by RF time. All these studies showed that accidental coincidences at the luminosity about of 3-4x10**33 is on the level of 10**(-4). While doing these studies few miss calibrated scintillators were found. Also it was found that at 4 GeV in 10% of all cases RF correction leads to peaking at the wrong bunch for electron. That effects on the timing on the level of 2.004ns. This effect needs more studies. V. Burkert: Missing mass resolutions, single particle acceptances ---------- (1) An update was given on corrections of kinematical quantities which are necessary because of remaining inaccuracies in the RECSIS event reconstruction. Upon plotting invariant quuantities such as missing mass and invariant hadronic mass versus various measured quantities such as momentum and angles, it becomes clear that there are still problems on the event reconstruction which need to be resolved. For example, at angles below about 22deg the electron elastic peak shows a clear theta and phi dependence. A possible explanation is that the reconstructed momentum is incorrect due to a misrepresentation of the magnetic field. This could be due to either a larger than requiredgrid size in the map, or that the calculated field deviates significantly from the real field. Assuming that the angles are reconstructed correctly one can adjust the momentum to yield the correct position of the elastic peak independent of azimuthal and polar angles. From this I obtain the following results in terms of missing mass resolution (MeV, sigma) before/after correction: process 2.4GeV /0.6B0 4.0GeV/0.6B0 4.0/0.87B0 =================================================================== W | 19.4 11.2 | >40 18.5 | 33.3 12.7 ep -> ep | | | -------------------------------------------------------------------- M(pi0) | 30.0 26.6 | 58.2 40.5 | 44.0 29.8 ep->eppi0 | | | -------------------------------------------------------------------- M(eta) | 15.8 13.5 | 38.0 19.9 | 23.2 17.5 ep->epeta | | | -------------------------------------------------------------------- M(neutron) | 15.6 13.8 | 29.1 19.4 | 20.9 15.6 ep->epi+n | | | -------------------------------------------------------------------- M(Lambda) | 10.0 10.0 | 19.6 17.8 | 15.7 12.9 ep->K+Lambda| | | -------------------------------------------------------------------- I believe the "corrected" numbers can ultimately achieved when the detector is well understood. I would be interested in finding out if anyone has achieved better resolution, and if so how this was achieved? 2) Several people have started to use the single particle acceptance (spa) functions that I showed a few weeks ago. However, in order to make more efficient use of the spa when extracting cross sections one needs to account for inefficiencies in the reconstruction due to missing wires, tof counters, etc.. To first order, these effects show up as massive localized inefficiencies in the event reconstruction, and can be taken into account by simple parametrizations that eliminate (gives efficiency = 0) entire bands in the kinematic plane. In the angle vs momentum plane these bands may be seen easily. While this procedure eliminates more than a detailed simulation with GSIM would find, it may be used in very fast simulations, at least until GSIM provides a more exact representation of all detector components. A potentially beneficial use of these parametrizations may also be in conjunction with GSIM simulations. A. Skabelin: Vertex corrections due to beamline offset ------------ Empty target data shows variation of the vertex z position from sector to sector. Variation is of the order 1-2 cm. Study of the Phi and Theta angular dependence of the vertex and study of its dependence on the particle type have been done. These studies show that sector dependence can be explained by the fact that target is shifted in the radial direction. The fit of the vertex position gives dx=2.6mm dy=-5.8 mm for target shift. K. Egiyan: Empty Target Data Analysis (Continuation) ---------- The "Empty Target" (ET) data from the Run 8938 were analyzed. Data from the 8937 "Full Target" (FT) Run were used also. 1. Vertex analysis show that there are significant Z-Vertex shifts for all sectors. These shifts are different for different sectors. Shifts have angular dependences but do not depend on particle sign of charge. Data obtained can be used to verify the target position, which is, probably shifted to the 6-th sector. 2. The thickness and origin of additional source of (e,e'p) data was studied. If assume that it is a ice molecules, then to fit data (e.g., missing momentum distributions), the 5.8*10**21 H_2O molecules (180 micM thick) should be exist at downstream cell-wall. This means, that Oxygen Luminosity L_O=0.54*L_H(FullTarget Luminosity), which is in contradiction with experimental data: L_O should be less then 0.2*L_H. So, what nucleus is "working" at downstream cell-wall? We have to find out. 3. It was shown, that there are some physics data, which can be considered as "publishable", e.g., nucleon momentum distribution, missing energy-missing momentum correlation, deuteron electroproduction, etc. A. Stavinsky: More on empty target analysis ------------- 1. Z-coordinate of vertex position in standard RECSIS proved to be systematically shifted from its nominal position. This shift depends on sector number and emission angle and could be explained by target shift in transversal direction of the order of 5 mm. When corrected in accordance with this hypothesis, Z-distributions proved to be the same for different sectors and runs. 2. The ratio r=N_out/N_in, where N_out = the number of interactions at outcoming window and N_in = at incoming one is systematically larger than 1 in accordance with Egiyan's report at the previous CALCOM meeting. This ratio changes from run to run from 20 to 1. There are at least two runs at the same initial energy where this ratio proved to be completely different (9183,r~7; and 9208,r~1; both at 4 GeV). The ratio N_in/fcup_live for both runs is the same. 3. Suppose Z<0 part of vertex position distribution arise from Z_Al and rest Z_h, while Z>0 part apart Al and rest H arise from unknown substance with Z_A nuclear protons and Z_H hydrogen protons. We can extract from relative importance of left and right peaks (Z_AL - left, and Z_Al+Z_A+Z_H -right) and smooth background (Z_h) two ratios R1=Z_Al/(0.5*Z_h) and R2=(Z_Al+0.5*Z_h)/(Z_Al+0.5*Z_h+Z_A+Z_H). There are two classes of events which could not arise from interaction with hydrogen: a) for x_b=Q^2/(2*[n]*m_p)>1, where [n]=E_0-E b) for cos A < 0, where A-angle between proton and virtual gamma. When apply one of these cuts we can get R3=Z_Al/(Z_Al+Z_A). From these three relations one can get Z_H/Z_A=0+-0.16 for run 8657(2.4 GeV, 11 Feb 98). It could be compared with H2O(0.25) and CH(0.16). Both within statistic and other runs must to be studied to make preference for one of these substances. 4. There are some nuclear physics which could be studied before solving this problem. An example of this physics is angular correlations between protons when its emission angle cos(A1)<0,cos(A2)<0 (so called cumulative particles). To study correlations between these protons we used mixing procedure with special selection criteria Q<0.3(GeV/c), where Q^2=(p_x1-p_x2)^2+(p_y1-p_y2)^2+(p_z1-p_z2)^2+(E_1-E_2)^2, and index 1,2 are for virtual gamma from different events, used for calculation of uncorrelated background. Our preliminary results show the strong dependence of correlation function on the angle between protons B:(dR/dcos(B)~1-1.4). It could be compared with the same for hA interactions (~0.5 for hC and ~0.3 for hTi) and also with model estimations for single intranuclear interactions (1.8+-0.4). Our preliminary results means that both initial and secondary particles cascade inside nuclear seems to be not important for cumulative particle production in eA interactions. L.C. Smith: ----------- Very preliminary differential cross sections for pi-zero electroproduction at Eb = 1.645 GeV were presented, covering a Q**2 range from 0.3 to 1.0 GeV**2. Acceptance in Q**2,W,costh(c.m.) and phi(c.m.) was calculated using a simple inelastic event generator combined with Volker Burkert's fiducial parameterization. Distributions generally follow the trends seen in previous measurements, with cross sections systematically 20-30% lower than previous data. However, no corrections for acceptance holes, inelastic radiative tails or target empty backgrounds were made. Some distortions in the c.m. distibutions are evident, suggesting the acceptance losses are underestimated in some regions. Further refinement of the acceptance calculation using GSIM is in progress. K. Joo: ------- I've studied the CLAS acceptance for single pi0 production using the 1.6 GeV high field runs. For this acceptance study, I generated 560k single pi0 events in phase space for Q2=0.35-0.6 and W=1.15-1.30. I used three different approaches. 1) Use the only fiducial cut (made by Volker) 2) Use the GSIM and RECSIS w/ fiducial cut 3) Use the GSIM and RECSIS w/o fiducial cut The following is the summary of this analysis. - Total number of events generated: 560 K - reconstructed events using method 1): 221 K (39%) - reconstructed events using method 2): 190 K (34%) - reconstructed events using method 3): 286 K (51%) >From the e1 data (26M trigger events) -w/o fiducial cut : 231 K -w/ fiducial cut : 173 K Based on this result, the following is the cross section at the W=1.225 and Q2=1.135-1.475 - cross section from method 1) : 198 microbarn - cross section from method 2) : 230 microbarn - cross section from method 3) : 185 microbarn The Latham's fitting based on their measurement at NINA at Q2=0.5 GeV^2 shows 240 microbarn at this kinematics. This is very preliminary result. This study does not include radiative effects, target window effects and dead wires of the chamber.