Measurement of D+- production and the charm contribution to (3)

and W DA).

5Monte Carlo simulation

A GEANT3.13-based[26]Monte Carlo(MC)simulation program which incorporates the best

current knowledge of the ZEUS detector and trigger was used to correct the data for detector and acceptance e?ects.The event generator used for the simulation of the QED radiation from the leptonic vertex was RAPGAP[27]interfaced to HERACLES4.1[28].The charm quarks were produced in the BGF process calculated at leading order(LO).The charm mass was set to1.5GeV.The GRV94HO[29]parton distribution functions(pdf’s)were used for the proton.Fragmentation was carried out using the Lund model[30],as implemented in JETSET 7.4[31],with the full parton shower option.The fraction of the original c-quark momentum which is carried by the D?±is determined from the‘SLAC’fragmentation function,which is equivalent to the Peterson model[32],with the fragmentation parameter?set to0.035[33]. The HERWIG5.9[34]event generator was also used,with the same pdf’s and c-quark mass as used in RAPGAP,to investigate the e?ects of fragmentation.

Generated events with at least one D?+→D0π+s→(K?π+or K?π+π?π+)π+s(or c.c.) were selected.These events were then processed through the detector and trigger simulation and through the same reconstruction program as was used for the data.

6Event selection

9

6.1Trigger

Events were selected online with a three-level trigger[15].At the?rst level(FLT),inclusive DIS events are triggered by the presence of an isolated electromagnetic cluster in the RCAL or any energy deposition in excess of3GeV in any electromagnetic section of the CAL[35]. During high-luminosity periods,when the rate was high,a coincidence with an FLT track was also required.Tracks at the FLT are de?ned as a series of CTD hits pointing to the nominal interaction point.The e?ciency of this trigger,with respect to the calorimeter-only trigger, was greater than99.5%and in good agreement with the MC simulation.

At the second level,algorithms are applied to reduce the non-e+p background.The full event information is available at the third level trigger(TLT).At this level,events are accepted as DIS candidates if a high-energy scattered positron candidate is found within the CAL(‘in-clusive DIS trigger’).Because of the high rate of low-Q2events,this trigger was turned o?in the region around the RCAL beampipe during high-luminosity operation.For the K2πdecay channel,a D?-?nder(based on computing the Kππmass using tracking information and se-lecting loosely around the D?±mass)was available in the TLT.Events at low Q2were then kept by requiring a coincidence of an identi?ed scattered positron anywhere in the CAL and a tagged D?±candidate.This will be referred to as the‘D?trigger’.

Using data selected from periods when both triggers were in use,the relative e?ciency of the D?trigger with respect to the inclusive DIS trigger is found to be about80%and independent of Q2,x,p T(D?)andη(D?)within the measured kinematic regions.The MC simulations reproduce this e?ciency to an accuracy better than the statistical accuracy of the data(≈2%).

6.2O?ine selection

The DIS event selection was similar to that described in an earlier publication[36];namely,the selection required:

?a positron,as identi?ed by a neural network algorithm,with a corrected energy above10 GeV;

?the impact point of the scattered positron on the RCAL was required to lie outside the region26×14cm2centered on the RCAL beamline;

?40<δ<65GeV;and

?a Z-vertex position|Z vtx|<50cm.

The DIS events were restricted to the kinematic region

?1

D?±candidates were reconstructed from CTD tracks which were assigned to the recon-structed event vertex.Only tracks with at least one hit in the third superlayer of the CTD were considered.This corresponds to an implicit requirement that p T>0.075GeV.Tracks were also required to have|η|<1.75,where the pseudorapidity is de?ned asη=?ln(tanθ

D0→K?π+π?π+(+c.c.).The branching ratio for D0to Kπ(K3π)is0.0385±0.0009(0.076±0.004)[37].The remaining selection criteria were di?erent for the two?nal states and are discussed separately.

6.3Selection for the K2π?nal state

Pairs of oppositely-charged tracks were?rst combined to form a D0candidate.Since no particle identi?cation was performed,the tracks were alternatively assigned the masses of a charged kaon and a charged pion.An additional slow track,with charge opposite to that of the kaon track and assigned the pion mass(πs),was combined with the D0candidate to form a D?±candidate.

The combinatorial background for the K2πdecay channel was further reduced by requiring ?that the transverse momenta of the K and theπwere greater than0.4GeV,and that of theπs was greater than0.12GeV.

In addition,the momentum ratio requirement

?p(D0)/p(πs)>8.0

was imposed.This requirement was used in the D?trigger to reduce the rate of candidate events with large?M≡(M Kππs?M Kπ),far from the signal region.

The D?±kinematic region of the present analysis was de?ned as

?1.5

Finally,the signal regions for the mass of the D0candidate,M(D0),and?M were ?1.80

?143

6.4Selection for the K4π?nal state

Permutations of two negatively-and two positively-charged tracks were?rst combined to form a D0candidate.As for the K2πchannel,the tracks were alternatively assigned the masses of a charged kaon and a charged pion.An additional track,with charge opposite to that of the kaon track and assigned the pion mass(πs),was combined with the D0candidate to form a D?±candidate.The combinatorial background for the K4πdecay channel was reduced by requiring

?that the transverse momentum of the K and eachπwas greater than0.5and0.2GeV, respectively,and that of theπs was greater than0.15GeV.

In addition,the momentum ratio requirement

?p(D0)/p(πs)>9.5

was imposed.

The D?±kinematic region was de?ned as

?2.5

The signal regions for M(D0)and?M≡(M Kππππs?M Kπππ)were

?1.81

?143

11

6.5Mass distributions

Figures2(a)and(c)show the distributions of M(D0)for candidates with?M in the signal region for the two?nal states,while Figs.2(b)and(d)show the distributions of?M for candidates with M(D0)in the signal region.Clear signals are observed around the expected mass values.

For the K2π?nal state,a?t to the M(D0)distribution of two Gaussians plus an exponen-tially falling background gives a peak at M(D0)=1863.2±0.8MeV and a width of23±2MeV. The second Gaussian around1.6GeV in Fig.2(a)originates primarily from D0decays to K?π+π0in which the neutral pion is not reconstructed.For the K4πchannel,the background level was determined by using the side-bands outside the?M signal region(see the dashed histogram in Fig.2(c))to make a M(D0)distribution.The?t to the M(D0)distribution,made by adding a Gaussian to the background distribution,yielded M(D0)=1862.7±1.5MeV and a width of20±2MeV.The deviations of the data from the?t,visible in the region just above the signal in Fig.2(c),are mostly due to the mass misassignments of the K andπcandidates with the same charge from D0decay.This was veri?ed by MC studies[38].The mass values found for the D0are consistent with the PDG[37]value of1864.6±0.5MeV.

The solid curve in Fig.2(b)shows a binned maximum-likelihood?t to the?M distri-bution from the K2πchannel using a Gaussian plus a background of the form A(?M?mπ)B exp[C(?M?mπ)],where A,B and C are free parameters and mπis the pion mass. The?t to the K2πplot gives a peak at?M=145.44±0.05MeV,in good agreement with the PDG value of145.397±0.030MeV,and a width of0.79±0.05MeV,in agreement with the experimental resolution.The multiplicative exponential term is needed to describe the background suppression at large?M,which comes from the requirement on the momentum ratio p(D0)/p(πs).

The solid curve in Fig.2(d)shows a?t of the?M distribution from the K4πchannel using a Gaussian plus a background distribution obtained from the side-bands outside the M(D0) signal region(see the dashed histogram in Fig.2(d)).The?t yields a peak at?M=145.61±0.05MeV and a width of0.78±0.07MeV.

For the K2πchannel,the number of D?±events obtained from a?t to the?M distribution in the restricted region of Q2,y,p T(D?)andη(D?)is2064±72.The number of events in the K4πchannel is determined from the?M distribution using the side band method[14],which properly accounts for the combinatorial background and the background arising from the mass misassignments.The side bands,1.74

7Data characteristics

The properties of the selected events are compared with those of the RAPGAP Monte Carlo simulation.All distributions shown are background-subtracted since they represent the number of signal events obtained by?tting the various mass distributions in a given bin.The data for the K2πchannel are shown in Figs.3and4as solid points and the RAPGAP simulation as shaded histograms.All MC plots are normalized to have the same area as the data distributions.

Figure3shows histograms of E′e,θ′e,γh andδand Fig.4(a-c)displays the distributions of Q2e,x eΣand W eΣ.In general,reasonable agreement is observed between data and the MC

12

simulation.Figure4(d-f)shows the transverse momentum,p T(D?),the pseudorapidity,η(D?), and the energy fraction carried by the D?±in theγ?p center-of-mass frame,x(D?).Although the p T(D?)spectrum of the data is well described,the MC pseudorapidity spectrum is shifted to lowerηcompared to the data and the x(D?)spectrum for the MC is shifted to slightly larger values.These discrepancies are examined in more detail below.

搜索“diyifanwen.net”或“第一范文网”即可找到本站免费阅读全部范文。收藏本站方便下次阅读，第一范文网，提供最新实用文档Measurement of D+- production and the charm contribution to (3)全文阅读和word下载服务。