This page contains some early results from low statistics simulations for initial 'ballpark estimates'. First some general numbers from a Glauber calculation, followed by Fritiof simulation results, then some rate estimates based on hopefully realistic detector simulations using HIJING and RQMD thru PISA, the response chain and the Level 2 test framework.

Want to jump to the Summary? Go to the correlation section?

I used the PHENIX Glauber Monte Carlo Code, in CVS at offline/analysis/glauber_mc.
A deuteron is, I assume, just approximated as A=2, Z=1, a light nuclei, i.e. without any special treatment. The plot below shows the distribution of the number of participants from the Au nucleus, when both the deuteron nucleons participated in the collision. The mean number of participants is about 10, with a maximum a little above 20. The expectation from colliding two spherically homogenous nuclei A and B (A>=B), with the same density at impact parameter b = 0 is
Npart = A + B - ( A^(2/3) - B^(2/3) )^(3/2)

Check:
B -> 0 gives 0 part.
B -> A gives A+B part.
A = Au, B = d gives ~16 part.

The columns are: Npart and Ncoll (3 columns each: mean, sigma, rms)

  0-  5%:   14.3    0.3    3.6    12.9    0.3    3.5       
  5- 10%:   11.4    0.2    2.6    10.1    0.2    2.5       
 10- 15%:   10.5    0.2    2.7     9.4    0.2    2.6       
 15- 20%:    9.3    0.2    2.1     8.2    0.2    2.0       
 20- 25%:    8.8    0.2    2.3     7.7    0.2    2.3       
 25- 30%:    8.1    0.2    2.1     7.0    0.2    2.0       
 30- 35%:    7.8    0.2    2.2     6.7    0.2    2.1       
 35- 40%:    7.3    0.2    2.1     6.2    0.2    2.1       
 40- 45%:    6.8    0.2    2.3     5.7    0.2    2.3       
 45- 50%:    6.3    0.2    2.1     5.2    0.2    2.1       
 50- 55%:    5.5    0.2    2.3     4.5    0.2    2.3       
 55- 60%:    5.3    0.2    2.4     4.3    0.2    2.5       
 60- 65%:    4.6    0.2    2.0     3.6    0.2    2.0       
 65- 70%:    4.1    0.2    2.0     3.1    0.1    1.9       
 70- 75%:    3.9    0.1    1.9     2.9    0.1    1.9       
 75- 80%:    3.4    0.1    1.5     2.4    0.1    1.5       
 80- 85%:    3.4    0.1    1.6     2.4    0.1    1.5       
 85- 90%:    3.1    0.1    1.5     2.1    0.1    1.4     

FRITIOF 7.02 is a simple Lund MC model, basically a shell around JETSET and PYTHIA, similar in part to HIJING and RQMD.
The Au nuclei is going South in all these simulations. 10000 events with b < 3 fm give a dNch/deta distribution as depicted in the plot below. A maximum of dNch/deta (note: not at midrapidity..) of about 30 with 10 participant pairs gives a dNch/deta/(0.5*Npart) ~ 3, which seems like the right order of magnitude.

HIJING and RQMD events, 1000 of each were processed thru PISA. Only the HIJING events have been studied so far. For this and the rest of the things shown here, with the exclusion of the response chain, the pro.27 version of the PHENIX libraries were used. For the response chain, new.2 from Aug. 15 was used.
For comparison with the d-Au files, I also used small samples of Au-Au,p-p and J/Psi events.

To be able to run the Level-2 test framework on simulated PRDFs, only a subset of the triggers (L2Mu*) were used. Only the South arm was included. As a first test, a pure J/psi sample (2 muons within ~10-30 degrees) was processed; 97 out of 100 gave rise to a single muon trigger and 69 to a dimuon trigger. (This corresponds to fairly ideal detector response).

For further 'calibration', I ran some 'pure vanilla' Hijing Au+Au and saw that 22 out of 100 events gave a L2MuSingleMuonTrigger, and 4 out of 100 gave a L2MuDiMuonTrigger. This can be compared with what we saw in Run-2: 1/7 (0.15) for L2MuSingleMuonTrigger and 1/45 (0.022) for L2MuDiMuonTrigger. I'm within a factor of 2 with this very limited statistics, and could be on the high/conservative side for the d-Au numbers below.

For the d-Au Hijing events, 19 out of 1000 gave a L2MuSingleMuonTrigger and 0 out of 1000 gave a L2MuDiMuonTrigger. If the basic trigger requirement is BBCLL1 * SingleMuon, then the acceptance fraction would be a factor of 10 less for d-Au than for Au-Au. On the other hand, the incoming event rate is expected to be a factor of ~10 (average) - 20 (peak) higher for d-Au than for Au-Au and even if the rate of accepted events should then be about the same, the crucial rejection needs to be done already at LVL1 (the incoming event rate is otherwise simply too high). As a first approximation, we would need to scale down single muon triggers but not dimuon triggers.

The question is what kind of rejection one can expect for d-Au. As a first test, I ran preco on the produced PRDFs that were previously studied with the level 2 test framework. On the resulting DSTs there is a dMuiPseudoTriggerOutNode, and I looked whether the SingleDeep and DoubleDeep bits were set. Rather high values, compared to the L2TestFramework results, were obtained. 237 out of the 1000 events gave a SingleDeep bit and 57 gave a DoubleDeep bit, i.e more than a factor 10 for the single muons and a factor > 50 for the double muons. The standard LuxorLinker files for shallow and deep were used (mui_pseudotrigmap*dat).

L2SingleMuonTrigger may be a somewhat different requirement from BLT SingleDeep, so I am perhaps not doing a fair comparison. Also, if one would use DeepShallow as a BLT requirement, this gives 146 triggers for the 1000 events; still pretty high.

To possibly verify whether these numbers were reliable or not, I also ran 1000 Hijing pp events through PISA and the response chain and preco, all in the same way as for the dAu events. The results were: 3 SingleDeep and 0 DoubleDeep out of 1000 events. If one includes all events where any bit (SingleShallow || DoubleShallow || SingleDeep || DeepShallow || DoubleDeep) were set, one gets 20 triggers out of 1000 events. These numbers are in rough agreement with the expected SingleDeep (=SingleDeep||DeepShallow||DoubleDeep) rate of 1/100. (again see Hiroki's note: the rates seem to have fluctuated somewhat over time: 0.004 to ~0.03))
I also ran these pp events through L2TestFramework and got one L2SingleMuon trigger for the 1000 events.

By looking at the number of bits that are on in the south vs north packets from the response chain, a difference of approx. a factor of 5 is observed. More in South, where the Au is headed; average of 11 bits/per event (~5 GEANT panel hits).
A natural first approximation is that the number of triggers, hit occupancies and bits that are on, should scale in the same way for the two arms. Hence, a factor of ~5 difference in the trigger rate from North relative to South is expected, for a perfect trigger.

For a trigger such as BLT, which is subject to combinatorial background, it should work a lot better on the North side than on the South side.

Does HIJING describe d-Au collisions well? Need more statistics also. (S. Johnson et al. are working on generating events.)
Include North arm in LVL2 (Jason has prepared the code so that this should be easy).
Make a LVL2 version more close to the expected MUID LL1 (quite a few things are in place for this)
Rerun simulations when new/more realistic MUID shielding code (by Vasily) is in place.
Does adding the muTr to LVL2 help? Will a STPP for muTr be available?

Glauber ntuple and Fritiof histograms: 
/phenix/u/silvermy/dAu/glauber_d_au_calc00.root
/phenix/u/silvermy/dAu/f7.root

PISA working dir:
/phenix/data19/silvermy/dAu

Response chain dir:
/phenix/data25/silvermy/work/presp

L2TestFramework - libraries, macros etc:
on rcas2068: /home/silvermy/lvl2test
/phenix/u/silvermy/lvl2test.tar.gz

All the numbers for accepted triggers in this table are for South, for both L2TestFramework and BLT. (It seems important to have MUID LL1 for South.)

The same 1000 dAu events analyzed with L2MuiPseudoTrigger/BLT: Arm 0 (South)

Nacc = Naccepted, Rej. = Rejection factor, mhX = obtained with mui_pseudotrigmap_deep_mhX.dat (X=1,2 - 1 was used for the early part of the Run-2 pp run and 2 were used for the later part [run>=38189]).
Note that e.g. Deep + Deep also fulfill Deep + Shallow and the rejection factor for Deep + Shallow is calc. including this. In other words, the number of accepted events is exclusive, while the rejection factor is inclusive.

The same 1000 dAu events analyzed with L2MuiPseudoTrigger: Arm 1 (North)

For more realistic numbers :
make LVL2 more like MUID LL1 and use that for the South arm
+ Have the latest and greatest implementation of MUID shielding in PISA (not yet committed)
+ Use a lot more statistics

* = number within () shows the corresponding number when requiring a 'matching' MUTR Station 3 hit.

  
0: Single Shallow South
1: Double Shallow South
2: Single Deep South
3: Deep + Shallow South
4: Deep + Deep South

5: Single Shallow North
6: Double Shallow North
7: Single Deep North
8: Deep + Shallow North
9: Deep + Deep North

10: Single Deep LL1 South
11: Double Deep LL1 South
12: Single Deep LL1 North
13: Double Deep LL1 North

14: Single Deep L2Muon South
15: Double Deep L2Muon South
16: Single Deep L2Muon North
17: Double Deep L2Muon North