Options for E906/Drell-Yan with a Solid Iron Magnet (Fe906?) Paul E. Reimer 20 June 2008

Options for E906/Drell-Yan with a

Solid Iron Magnet(Fe906?)

Paul E. Reimer20 June 2008

1. Solid Iron magnet

2. Target dump separation

3. Chamber rates

4. Resolution—Mass, xF, x1, x2

5. Acceptance

220 June 2008 Paul E. Reimer

Why a Solid Iron Magnet?

E906 budget is very tight. – We are expecting ≈$2M for DOE/Nuclear Physics. Of this $1.6M will be used to produce

coils for an open aperture magnet. The remainder is for spectrometer upgrades.– NSF, Taiwan and Japan are planning contributions to the spectrometer upgrades

Fermilab has requested that we contribute an additional $1.5M for M&S that they need to spend for E906

– Some tasks e.g. flammable gas system (Illinois) and cryotargets may be done by collaborators.

– RIKEN may to contribute ≈$600k to “common fund” items– Still short ≈$500k

Solid Iron Magnet– Essentially free—coils and iron exist, Iron must be cut ($100k)– Reprogram DOE/ONP funds to Fermilab (Under discussion with Brad Tippens)

– Increased acceptance--larger “aperture” (if you can call it that)– Decreased resolution


Magnet Description

Use existing coils from SM3

– 3 coil packs:

191, 163 or 135 inch coils

– Iron blocks from SM12 (E866) magnet

– Simulations shown for 191” Suggested by Chuck BrownSuggested by Chuck Brown 135”

191”

SM3 coils which were constructed in 1981 for E605 by Sumitomo and supervised by KEK and Kyoto. Supervised by Prof. Miyake, Dr. Maki and Dr. Sakai


Magnet Description

Use existing coils from SM3

– 3 coil packs

– 191, 163 or 135 inch coils

– Simulations shown for 189” Issues:

– Target/dump separation

– Chamber rates

– Mass Resolution

– x2 resolution

– Field simulation • no measurements• Accurately determine saturation curve

– Acceptance (i.e. statistical uncertainty at the end of the day)

Field simulation for 189” solid Fe magnet (DFG)


Dump/Target Separation

Resolution significantly worse

Trigger and reconstruction cuts allow for clean separation

– Remove events with tracks at dump face within 2.25” of beam axis

– Remove events with tracks greater than 10” from beam axis at targets

Recall dump starts at 0”. These cuts effectively remove all events from the front of the dump!!

Target

Open aperture in redSolid Fe in Blue

Dump

DumpTarg

et

Slight problem from dump J/’s


Dump/Target Sep., Solid Fe Magnet Only Target—solid histograms

-70 < z < -50 Dump—dashed histo.

0<z

Magenta, all events Green: Magneta and

– xF>0 and

– M>4.5 GeV

– and pz>20 GeV

Blue: Green and

– |ytrack|>2.25 in at z=0 (dump face)

Red: Blue and

– |ytrack|<10.0 in at z=-60 (target)

Reconstructed

Generated


Station 1 Chamber Rates

Occasionally a muon showers in the absorber If this happens in the center of the absorber,

no effect is seen as shower is also absorbed If this happens in the last few inches of the

absorber, shower can create extremely large rates in Station 1 (of low momentum particles)

Solution is to have an absorber-free region at the end of the field volume and use field as a sweeper

In Solid Iron magnet, there is no absorber-free sweeper region! (Can we find a wide gap sweeper magnet?)

Requires GEANT MC to see magnitude of effect

Absorber and B Field

Sta. 1



B Field only


Mass Resolution

Reconstruction of Muon tracks yields virtual photon properties:

– M2, xF, pT

Mass resolution dictates J/ cut Critical for x2 (and x1) resolution

Drell-Yan

J/450 MeV(and don’t forget about the 0)

Drell-Yan

J/470 MeV(and don’t forget about the 0)

“Standard” E906 Analysis Mass cut of 4.5 GeV


Mass Calibration

Track reconstruction apparently skews virtual photon mass

Events with tracks which significantly miss target have poor mass reconstruction

yrtp and yrtn are the reconstructed y offsets of the two tracks at a plane perpendicular to the beam in the center of the target. P and n denote positive and negative.

Mass is the difference between generate and reconstructed mass in the Monte Carlo

Drell-Yan (M>4.5 GeV) MC in blackJ/ MC in red

Use known mass (J/) to calibrate this correction

– Effect is slightly different, but similar to higher mass Drell-Yan


Calibrated mass resolution

Introduces mass offset in Drell-Yan mass range (4.5 < M<8.5 GeV) but better than uncorrected

Drell-Yan

J/298 MeV

Significant improvement near J/ General improvement at all masses


xF Resolution

No Known xF point to use for calibration.


x1 Resolution

x1 resolution worseworse with correction

Resolutio

n

for all x 1

Resolutio

n

x 1≥ 0.75


x2 Resolution

x2 resolution improved by correction.

Resolution significantly worse for high x2—is this a problem?

Resolutio

n

for all x 2

Resolutio

n

x 2≥ 0.35

Recall D-Y falls exponentially with x2.

Events may be reconstructed into neighboring bins.


x2 resolution—solid iron and miss reconstruction

Can we live with this Can we live with this misidentification? Or find an misidentification? Or find an additional correctionadditional correction

Red events have reconstructed to the correct x2 bin

For a significant fraction of the events, x2

true is

one (blue) or

two (green)

bins lower than x2

recon.

Ess

entia

lly n

o da

ta in

thi

s bi

n


x2 resolution— Open ApertureOpen Aperture and miss reconstruction Red events have

reconstructed to the correct x2 bin

Open aperture design also suffered from misreconstruction, but not as much as solid iron design E

ssen

tially

no

data

in

the

se b

ins

Does anyone want to look at this in E866 data?


Statistical Precision

Increased statistical precision w/Solid Fe Magnet

Larger transverse “aperture” in Solid Fe Magnet (34” vs. 25”)

Larger Station 3 wire chamber

– Needed to take advantage of larger aperture

– Note: open aperture would also benefit from larger station 3

– Part of Japanese contribution

Significant gain in large-x2 region

Open Ap. Design may also gain in Stat. as well from larger Station 2 & 3 combination


Drell-Yan Ratio Statistical Uncertainty

Better Statistical Uncertainty at high x2

Crude trigger matrix applied

– not used in previous plots

– Accounts for fall-off at low x2 in solid iron magnet


126 inch Magnet

Resolution in Mass and x2 not better

Lower acceptance at high-x2

– See relative normalization

Mass All x2 x2 for x2 > 0.35


Magnet Polarity:which events do we want?

We really want the high-x2 events

All the other events come for free


Magnet polarity 8 GeV


Magnet Polarity 6 GeV


Magent Polarity 4 GeV


Magnet Polarity 3 GeV

Conclusion Running in opposite polarity may allow smaller chambers

– Important for Station 3 which is LARGE

– Acceptance must be verified with Monte Carlo

Preliminary Monte Carlo shows resolution is the same (expected)


Things left to do

Independent verification of results

1. Background rate simulations See GEANT MC talk—this seems to be in hand

2. Opposite Polarity MC with reduced Sta. 3 size I’ve run the MC and could have results next week

3. Resolution improvements with target retrace?

4. Angular resolution for cos 2 distributions

5. Partonic Energy Loss—is x1 resolution good enough?

6. What is the saturation curve of the SM12 iron—this clearly effects the field Some data from Chuck et al. that must be analyzed


Resolution is poorer with Solid iron magnet, but acceptable?

– x2 bin resolution

Dump/Target separation is achievable

Conclusion: Solid Fe magnet will work (always listen to Chuck)

Ess

en

tially

no

dat

a in

thi

s b

in

Larger aperture provides better statistical precision at high x2

Opposite polarity configuration may allow for smaller Station 3 and 4 chambers

Documents

Options for E906/Drell-Yan with a Solid Iron Magnet (Fe906?) Paul E. Reimer 20 June 2008