This document describes the present status of STAR-EMC simulations and analysis software (SAS). There are six branches of software to be considered: simulations, analysis, calibration, evaluation, physics, and trigger. A consensus priority schedule for the presently incomplete efforts in these branches must be defined promptly. Those efforts that will provide the information needed by the STAR-EMC group to arrive at a defensible design of the overall system must be done first.

The EMC-SAS group recently grew in size, which is expected to continue as several open positions at EMC institutions are filled. However, quite a bit of information is needed on open questions on all levels, including the mechanical design (e.g., effects due to cracks), the electronics (e.g. jet trigger efficiencies using sector data), the integration with the STAR-SAS and SOFI groups, and the precision of particular off-line analyses.

The situation is further complicated by the fact that the present STAR GEANT software is giving way to GSTAR, and the analysis shell TAS is giving way to MOAST. It is therefore important to determine which efforts are so critical that they should be done using the present framework, and which topics would better be studied following the development of the EMC infrastructure for the GSTAR and MOAST environments.

EMC-SAS Topics

A survey probing the present status of, and plans for, EMC SAS efforts was distributed amongst the EMC institutions. The remainder of this document includes a summary of the responses to this survey. It is broken into sections by topic. Each includes a short description, as well as an delineation of what has and has not been done to date. There is sometimes overlap between what is included under each topic. In the following, SPEMC stands for the Small Prototype EMC, and LPEMC stands for the Large Prototype EMC, which is the full module prototype(s) that will be built soon.

A table is included at the end of this document that lists the EMC SAS institutions. For each institution, the number of people working on EMC SAS efforts, the available computer systems, and the SAS topics of interest are listed.

Direct-[[gamma]] measurements

The goal is the measurement of structure functions - for gluon distributions in nucleons and nuclei, and the spin distributions in nucleons in the spin physics program. We have claimed that these can be measured using the STAR-EMC for 8 GeV/c <= PT <= 20 GeV/c. A somewhat wider range of sensitivity from 5 GeV/c <= PT <= 25 GeV/c is suggested by scaling the available CDF data. The STAR note SN0077, "Estimates of Rates and Errors for Measurements of Direct-[[gamma]] and Direct-[[gamma]]+Jet Production by Polarized Protons at RHIC", by M.E. Beddo, H. Spinka, and D.G. Underwood, gives detailed information on this topic.

There are two kinds of backgrounds to direct-[[gamma]] studies, both arising from high energy [[pi]]0`s in jets. A fake direct-[[gamma]] can occur from a high energy [[pi]]0 leading to two [[gamma]]`s that are less than ~5 cm apart at the EMC, or in asymmetric [[pi]]0 decays. Simulation results have already been obtained on the following issues: the SMD occupancy, the effects due to cracks, and the low-energy cutoff.

These simulations need to be refined, especially concerning the importance of the low-energy cutoff for jets with multiple [[pi]]0's. There are also several open questions.

The background estimates that were recently done at ANL disagree with those previously done using Hijing and GEANT. The new simulations demonstrate that much of the discrepancy arises from the "steepness" of the spectra. Studies of the backgrounds should continue. Also, during the recent ANL spin review, CDF members stressed the importance of multiple [[pi]]0's in a jet as a background to the direct-[[gamma]] searches. In addition, bugs have been discovered in Hijing, and the opinion was that these were only partially fixed.

At the spin physics presentations earlier this year at BNL, all of the error calculations were done assuming a 50% background, and the results were still favorable. These simulations nonetheless need to be repeated, and compared to the CDF data. To do such comparisons, the CDF data can either be scaled in X, or scaled using parton scattering cross sections at 1800 and 200 GeV.

For the gluon distributions in nuclei, the simulations that are needed involve events with a fairly low PT direct-[[gamma]] (~10 GeV) that goes into the EEMC. The effects that the materials that are in front of the EEMC, i.e. the TPC wheel and electronics cards, have on this measurement must be investigated. Information on this material is in SN0206, "Time Projection Chamber Wheel Assembly", by G. Koehler and R. Wells, and the drawings referenced therein.

Jet measurements

The goals are searches for jet-quenching, and also measurements of the gluon distributions in nuclei in p+A reactions.

In SN0183, "Comparison of Standard Jets Reconstruction Algorithms in pp, pAu, and AA Collisions for STAR", by A. Pavlinov, the UA1 and CDF jet finding algorithms were compared for p+p, p+Au, C+C, and Si+Si reactions leading to PT ~ 10 GeV/c jets. Another study is described in SN0196, "Efficiency of Modified UA1 Jet Reconstruction Algorithm in PP, PA, and AA Collisions at STAR", by W.B. Christie and K. Shestermanov. Here, jet reconstruction was simulated for p+p, p+A, and A+A reactions, where A = C, O, Si, and Fe. For p+A reactions, jets with quark scatters in the range 10 GeV/c <= PT <= 30 GeV/c were studied. In p+p and A+A, all of the simulations were performed for PT =30 GeV/c. A modified version of the UA1 jet reconstruction software was used.

However, this study considered only those scatters in which the rapidity of each quark was in the range |[[eta]]| < 0.3. Given that jet radii are ~0.7 in ([[eta]],[[phi]]), these quark scatters are only those for which each jet is entirely contained in the BEMC. The simulations of this topic therefore need to be extended to predict the efficiencies and acceptance for reconstructed jets over the wider range of [[eta]] covered by the BEMC plus EEMC, i.e. for scatters at all rapidities between -0.3 <= [[eta]] <= 1.3. This necessarily involves studies of the importance of losses at the [[eta]]=1 crack between the BEMC and EEMC.

The CDF Collaboration uses edge corrections to jet measurements that are reputed to work well. These corrections should be studied for the present purposes; a presumably larger jet acceptance would then be open for study. Considering just the BEMC, it is possible that parton scatters in the range |[[eta]]| <= 0.8 might be measurable.

The simulations should also be extended to give more systematic information versus PT, in all of the symmetric entrance channel reactions.

The efficiencies and resolution for jet reconstruction in Au+Au collisions have never been quoted. These simulations should be done, even if they are costly in terms of CPU time or disk space.

The study of [[gamma]]-Jet processes is incomplete. These processes need to be studied to a level of detail similar to that for the Jet-Jet processes. The analysis routines for the [[gamma]]-Jet events are not simply variants of those being used for the jet-jet analyses. What is needed are direct-[[gamma]] specific algorithms that use all of the available information and return a 2, for example, that a particular hit is a direct-[[gamma]] or a [[pi]]0.

There are several important issues concerning jets and sector data that have been raised in the recent STAR workshops and preliminary reviews. First, the efficiencies of jet triggers based on sector data are not known. Second, it is not known what level of processing of the TPC data can be read out a 100 Hz that gives ~10% momentum resolution on the tracks. That is, are the fitted tracks sufficient, or is it necessary to record the clusters, or even the raw data? This is important given the needed reduction in the data rate so that ~100 events/second can be written to tape. Third, there are concerns about the recoverability of the tracks which cross sector boundaries. These tracks are needed to obtain the correct jet energy and direction. It might be possible to use the EMC data to flag certain tracks, particularly if only a limited amount of TPC data is to be read out (i.e. at high rates in the p+p running). On this point, see SN0080, "Update on the Study of p+p Collisions in STAR", by P. Jones, and SN0169, "Associating the Correct Tracks to the Trigger Event in STAR for High Luminosity Proton on Proton Running", by W.B. Christie.

e+/- measurements

Backgrounds to W+/-,Z measurements

Of interest is the efficiencies for the identification of the ~40-50 GeV electrons that result from the decays of W's and Z's. For these high momentum tracks, the comparison of the track momentum and the EMC energy allows a rather efficient separation between charged pions and electrons. The important questions here relate to the backgrounds.

A major background is provided by jets in a narrow part of the fragmentation function leading to a high energy leading-[[pi]]0. However, other sources of backgrounds come from jets with multiple [[pi]]0's, and from charged pions in the jets as well. It is therefore also important to understand the backgrounds from these sources. Work already done on this topic is summarized in SN0168, "W and Z Event Rates and Background Estimates for the STAR Detector at RHIC in pp Collisions", by V.L. Rykov, and K. Shestermanov; and in SN0210, "Drell-Yan Pairs, W and Z Event Rates and Background at RHIC", by A. Derevschikov, V. Rykov, K. Shestermanov, and A. Yokosawa.

Here, one starts with comparisons of the jet rate versus PT to the rate of electrons from W's and Z's versus PT. The reduction to the relative background that is obtained at each stage of the analysis should then be tabulated.

Other e+/- related measurements

The two goals here are, one, the identification of low energy (below ~2 GeV/c) electrons from the, e.g., J/[[psi]], formed in the QGP and, two, the study of high mass Drell-Yan processes in the spin physics program.

Regarding the first, the major issue regarding the EMC/SMD performance is related to the multiplicity of hits in each tower. The SMD may not be useful when there are more than ~4 hits per 5 towers, as a SMD cell covers 5 towers. The studies of the optimal SMD segmentation that are important in other topics are important here as well.

There are also other, albiet relatively minor, issues here. The comparison of the track momenta and the EMC energy is not efficient at separating pions from electrons for momenta below ~2 GeV/c. The use of other EMC information, such as the SMD summed-PHs, SMD PH-widths, and PH centroids compared to the predicted centroids, will provide some improvements to the e/[[pi]] discrimination. This information is clearly less useful if the occupancy is large, as noted above. The use of the TPC dE/dx and TOF information should also improve the low energy e/[[pi]] separation. The discrimination and backgrounds that are achieved using a judicious set of cuts on the EMC, SMD, TPC dE/dx, and TOF information should be simulated. The SPEMC and prototype SMD data is providing useful information on the lateral and longitudinal shower profiles for e's and [[pi]]'s with momenta at and below 8 GeV/c.

Regarding the second, the difficulty in the Drell-Yan physics in the spin program is caused by the low cross section for large masses. Large masses are generally needed to get to low enough X to see the interesting effects (i.e., X1 and X2 near 0.05-0.1). See SN0210, "Drell-Yan Pairs, W and Z Event Rates and Background at RHIC", by A. Derevschikov, V. Rykov, K. Shestermanov, and A. Yokosawa.

Errors involved in spin-asymmetry measurements

A spin asymmetry is in general defined as (X++-X+-)/(X+++X+-), where X is some experimental observable, such as the W yield, and +/- refers to the polarization of the colliding beam particles. More information is needed in general on the importance of the backgrounds to the error bars on particular spin asymmetry measurements. So far, known formulas have been used to determine the errors in the presence of backgrounds.

Relevant information is contained in SN0077, "Estimates of Rates and Errors for Measurements of Direct-[[gamma]] and Direct-[[gamma]]+Jet Production by Polarized Protons at RHIC", by M.E. Beddo, H. Spinka, and D.G. Underwood, as well as SN210, "Drell-Yan Pairs, W and Z Event Rates and Background at RHIC", by A. Derevschikov, V. Rykov, K. Shestermanov, and A. Yokosawa.

The performance of the luminosity monitors needs to be simulated. Examples of these monitors include low-PT [[pi]]0's in the EMC, and the number of tracks in the CTB. Other observbles might prove useful for this task. All of the monitors also need to be studied when a vertex cut on the VPD is applied. A relevant reference is SN0100, "Luminosity Monitor based on Proton Scattering", by S. Nurushev.

Neutral energy measurements

It is necessary to complete the simulations started by Christie on the extraction of the total neutral and total charged calorimeter energy over particular samples of events.

These studies have were done using only the information available at early stages of the trigger, for which there is no tracking. Analyses of this topic should be extended to include that information available off-line, for which the momentum of each (charged) track is known. Some improvement in the sensitivity of these observables is expected.

Calorimeter/SMD Design Issues

The studies listed here may quite obviously impact the design of the device. While these may already be included in the listings above, they are also included here as these questions might need to be answered definitively while the mechanical design of the device is still in progress (i.e. soon).

It is possible that only ~1/4 of the BEMC modules will be installed by Day 1 of RHIC running. It is therefore important to understand the physics capabilities of such a fraction of the full detector system.

The best BSMD and ESMD segmentation needs further study.

The study of the performance of the BEMC for low energy particles at grazing angles to the EMC is important. SPEMC test beam data was collected for non-normally incident particles, but at only one angle (~15 degrees).

The importance of the cracks, e.g. for the direct-[[gamma]] studies, needs further study..

The optimal depth-segmentations of the BEMC and EEMC are also somewhat of an open question. A significant amount of information in this topic has been provided by the SPEMC test beam data for depth segmentations of 10/10 and 5/15 (expressed in layers). These two segmentations result in different momentum-dependent efficiencies and backgrounds for e-identification. The choice of the best depth segmentation depends on the momentum of the particles one is interested in identifying via the shower depth information. Therefore, simulations of the efficiencies and backgrounds for particular STAR analyses assuming the performance of the depth-segmented device predicted by the SPEMC data should be performed.

Studies of the extent to which the energy resolution can be improved via the separate read out of the first scintillator layer are either incomplete or not sufficiently publicized within the group.

The EMC Fast Simulator

An EMC Fast Simulator (EFS), possibly done entirely in TAS and/or MOAST, is needed. For this tracking option, GEANT/GSTAR is asked only to propagate particles to the first active layer of the EMC, at which point the track information is saved to tables and the particle is stopped. The EFS software then generates the lateral and longitudinal shower profiles for each incident track using Monte Carlo methods, including fluctuations. The main motivations for the EFS are in it's speed, and in the fact that it will exactly reproduce the available test beam data.

A rather obscene amount of data (~1.3 Billion events on tape) has been collected on the performance of the SPEMC to [[pi]], u, and e over the momentum range from 0.3 to 8 GeV/c, which will serve as input to the EFS development. The EFS will also attempt to describe the calorimeter response for more extreme incident momenta (e.g. up to ~50 GeV), using extrapolations of the available data that are consistent with (slow) GEANT simulations of the SPEMC for these momenta. The SPEMC data and the EFS need to be compared to the test-beam results collected from the LPEMC, as the stack design of the SPEMC differs from that planned for the LPEMC and the STAR-EMC.

The EFS was schematically described at the recent STAR-SAS workshop, and it was felt that the EFS does not cause significant of software integration problems. Work will begin on this code when the SPEMC analyses are closer to completion, and the GSTAR/MOAST framework is more fully developed.

Software infrastructure

A significant fraction of the actual work that remains involves the development of the software needed to address the questions outlined above. Such software development efforts, and a few others, are now listed.

The EMC software must stay current with the changing framework for STAR simulations and analysis, which is moving towards the GSTAR and MOAST environments. This involves new philosophies for the geometry files, the table definitions, the data flow, and perhaps, the analysis routines as well.

Recently the software mocking up the interface between the EMC Level-0 trigger and the STAR Trigger system was developed.

The simulated output from the SMDs is now available to TAS.

It is important to understand the flow of EMC data through the entire EMC and STAR trigger and DAQ systems. The EMC group should therefore perform MODSIM simulations to obtain a detailed understanding of the EMC data flow. The MODSIM package is expensive, but it is already installed at BNL and is available for our use.

The [[gamma]]-Jet analysis routines need to be written. These are not simply revisions to the present Jet-Jet analysis routines.

The analysis routines for [[pi]]0 and reconstruction need to be written. While the physics justification for [[pi]]0 reconstruction is not obvious, it may be useful to reconstruct `s for calibration purposes.

Inst.  No.   Hardware              SAS Topics                                                
                                                                                             
 ANL   6     Alpha/VMS,            Direct-[[gamma]]'s, Jets, Spin physics, Global            
             Sun/Solaris           Observables, EMC/SMD Geometry                             
                                                                                             
 BNL   1     RHIC Cluster          Direct-[[gamma]]'s, Jets, Global Observables, Software    
             (SGI/IBM)             Infrastructure                                            
                                                                                             
IHEP   5     Alpha/VMS and VAX's   Direct-[[gamma]]'s, Jets, Spin physics, SMD Geometry      
                                                                                             
JINR   11    IBM-PCs/LINUX         Direct-[[gamma]]'s, Jets, High Pt, EEMC/ESMD Geometry     
                                                                                             
Rice   3     IBM RS6000/AIX        Global Observables, Spin physics, Software                
                                   Infrastructure, Low energy e/[[pi]]  discrimination                     
                                                                                             
UCLA   2     Sun                   Global Observables, Direct-[[gamma]]'s                    
             SPARC-20's/Solaris                                                              
                                                                                             
 WSU   1     SGI's                 Spin physics