Kaon Aerogel Cherenkov Detector

Brief Overview of a new experiment for JLab Hall C

Nathaniel HlavinStephen Rowe

Catholic University of America

Catholic University of America, Washington, D.C. August 5,

2010

OutlinePhysical MotivationDetector OverviewCalculations for Refractive Indices of AerogelComputer Simulations Results and OptimizationOutlook for Detector

Physical Motivation: Proton Substructure

To understand further what makes up a proton, we use Generalized Parton Distributions (GPD).

They combine the spatial and momentum distribution of the quarks into one concise chart. Below are some examples based on models. We seek experimental confirmation. “x” is the fraction of the momentum of the quark over the momentum of the proton

Meson Electroproduction A high-energy electron beam releases a

photon of a very high frequency. The high-frequency allows us to observe objects with greater detail.

The interaction of the photon and proton (pictured “GPD” here) results in a neutron (pictured here with momentum p’) and a meson (for example a pion or a kaon). For our proposed experiment, the meson would be a kaon.

We are able to detect/calculate the portions of the diagram above to dotted line, and from them we can contsruct the GPD. However, the meson released must conform with the Quantum Chromo Dynamic (QCD) theory of quarks.

K+ or

Above dotted Line: Hard Scattering Process (QCD) We can calculate this.

Below Dotted Line: GPD. This is what we want.

Problematic Pions We know how to calculate the

hard or perturbative QCD model, but our data for the pion spatial quark distribution does not match the QCD prediction.

Further experimentation with the kaon could provide data that matches the QCD prediction. The dotted prediction seems to doubt it.

Matching data would enable GPD construction to reveal the proton substructure.

The 12 GeV upgrade allows for kaon production above a Q2 value of 1, and even higher into the hard QCD range.

T. Horn et al., Phys. Rev. Lett. 97 (2006) 192001.T. Horn et al., arXiv:0707.1794 (2007).

A.P. Bakulev et al, Phys. Rev. D70 (2004)]

Kaon Aerogel Detector•As a charged particle (kaon) passes through a substance (aerogel) faster than light passes through that substance, light is emitted.•This light is called Cherenkov light and can be thought of as a shock wave in the electromagnetic field. It is analogous to a sonic boom.•The light is collected by Photomultiplier Tubes (PMT) and converted into an electron signal by the photoelectric effect.•This signal is then amplified into a signal that can be analyzed. •An example image of a detector is pictured at right.

Kaon

Cherenkov Light

PMTs

Aerogel Panels

Front View

Side View

Note on Efficiencies The efficiency of the detector is a good deal less than 100% percent for a

number of reasons. Below is a list of contributing inefficiencies Detection efficiency/Quantum Efficiency: 26% Diffusion wall reflectivity: 90% (This figure could increase to 96% if

millipore material is used) PMT transmission: 60% Aerogel reflectivity: 80% Total efficiency: 11.2% These numbers have been taken account of in our calculations to get an

accurate idea of how large the signal from the PMTs will really be.

Calculations Below is an example graph of the number of photoelectrons for the range

of momentum we are interested in (2-8 GeV/c). The refractive index used here was 1.031. The upper curve is for pions, the middle for kaons, and the lower for protons.

Since a detector already adequately distinguishes pions and kaons, we are interested in distinguishing protons and kaons, which becomes a little more difficult, due to the proton interference for a wide range of momenta.

0 1 2 3 4 5 6 7 8 90

5

10

15

20

25

30

Photoelectrons emitted by Cherenkov Radiation of Kaons, Pions and Protons

Momentum (GeV/c)

Num

ber

of P

hoto

elec

tron

s

Pions

Kaons

Protons

Finding Likely Candidates The refractive index of our aerogel must

be between 1.007 and 1.13, approximately, which is due to limits the material itself has.

The index must also cut down on proton signal and maximize kaon signal.

We could have a few indices for our one detector to cover the full range in momentum.

To find likely candidates, we graphed the number of photoelectrons emitted over a range of indices for each momentum value in the range we are interested in (2-8 GeV). At right is the graph for 3 GeV. (Kaons above, protons below)

After analyzing the 7 graphs, we ascertained several candidate values: 1.01, 1.015, 1.02, and 1.03

1.000 1.020 1.040 1.060 1.080 1.100 1.120 1.1400

10

20

30

40

50

60

70

80

90

Refractive Indices

Num

ber

of P

hoto

elec

tron

s

Kaon

Proton

Testing the Candidates

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.000

1

2

3

4

5

6

7

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.000

2

4

6

8

10

12

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.000

2

4

6

8

10

12

14

16

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.000

5

10

15

20

25

Index: 1.01 Index: 1.015

Index: 1.02

Index: 1.03

•Below are the graphs of number of photoelectrons vs. momentum (GeV/c) for the four index candidates.

KaonKaon

KaonKaon

Proton

Proton

Proton

Proton

Np.e.

Np.e.

Np.e.

Np.e.

Momentum (GeV/c)

Momentum (GeV/c)Momentum (GeV/c)

Momentum (GeV/c)

Choosing Candidates As we examined the data several things became apparent to us: 1.01 would not work due to its weak signal for the kaon and its

lack of signal for the kaon in the early momenta (up to appx. 3 GeV)

Since we need a strong signal in the early momenta of 2-4 GeV, 1.03 seemed to be a good choice.

Particles in momenta 6-8 GeV can be distinguishing using kinematic checks, time of flight, etc.

The last range we needed to cover then, was 4-6 GeV This range is covered with minimal proton interference and

decent kaon signal by the 1.015 index. 1.015 and 1.03 thus seemed like desirable choices.

Choices of Optimum Parameters for Different Experiments

For certain momenta and indices of refraction, kaons emit a detectable signal where protons do not.

We have a lot of data. In order to have convenient access to this information I created 4 Fortran programs.

In each of these programs, the user can input a desired momentum and then receive optimized values from the data.

These programs would allow us and other researchers to have easy access to the optimized data and configure these parameters for a specific experiment.

Since these programs is easy to use and only requires the user to input momentum, they would be easy to edit and update with new data.

Program Types

There are two ways of presenting the optimized parameters.

3 of the Fortran programs, mom1015.f, mom1020.f, mom1030.f, operate within one index of refraction. The refractive index is fixed, but the user can see the optimized values for a particular momentum.

The user is also able to interpolate between values (the program works in increments of 0.5)

The second approach is even more simple. The user does not have to know anything about refractive indices and simply inputs a momentum and the program returns an optimized refractive index.

Program Examples

jlabs1> mom1015.x momentum=4.5 p=4.5, beta=0.9940, theta=0.1472, Npe=9

jlabs1> mom_ind.x momentum=7 beta=0.9975, n=1.007, theta=0.0947, Npe=4

Program name.xPrompts momentumInput desired momentum

p=momentumbeta=relativistic particle velocitytheta=Cherenkov Angle (radians)Npe=Number of Photoelectorns

beta=relativistic particle velocityn=refractive indextheta=Cherenkov Angle (radians)Npe=Number of Photoelectorns

Accessing Programs

These programs were compiled and tested on jlabs1, one of the computers on a JLab computation cluster. Its architecture is SUN.

These programs are located in /u/home/stephenr/stephendir

To compile these programs type f77 program.f –o program.x

To run these programs type program.x on the command line.

Notes on Programs

Momentum values range from 2 GeV/c to 8 Gev/c in increments of 0.5.

mom1015.f, mom1020.f, and mom1030.f give optimized values regardless of proton signals. The output PNpe=the number of proton photoelectrons.

For some values, there are no kaon signals or proton signals. In these cases, the output will say n/a.

For 8 GeV/c, there is no kaon signal without a proton signal as well. For 8 GeV/c, we have a solution that's based on kinematic separation of protons and kaons. No aerogel is used for 8 GeV/c.

These are my first formal programming works and they have proven to be a valuable learning experience.

How the Programs Work

The programs begin with a prompt allowing the user to input their desired momentum.

That input is then matched to a variable. A series of IF statements follow. IF the variable (inputted momentum)

equals ___ then output _____. write (*,*) 'momentum=' read (*,*) m

if (m.eq.2) print*, 'beta=0.9709, n=1.102, theta=0.3634, Npe=52' if (m.eq.3) print*, 'beta=0.9867, n=1.047, theta=0.2538, Npe=26' if (m.eq.4) print*, 'beta=0.9925, n=1.027, theta=0.1947, Npe=16' if (m.eq.5) print*, 'beta=0.9952, n=1.017, theta=0.1547, Npe=10' if (m.eq.6) print*, 'beta=0.9966, n=1.012, theta=0.1306, Npe=7' if (m.eq.7) print*, 'beta=0.9975, n=1.007, theta=0.0947, Npe=4' if (m.eq.8) print*, 'n/a, see program introduction' if (m.eq.2.5) print*, 'beta=0.9811,n=1.067,theta=0.3001,Npe=36' if (m.eq.3.5) print*, 'beta=0.9902,n=1.032,theta=0.2073,Npe=18' if (m.eq.4.5) print*, 'beta=0.9940,n=1.017,theta=0.1472,Npe=9' if (m.eq.5.5) print*, 'beta=0.9960,n=1.012,theta=0.1257,Npe=7' if (m.eq.6.5) print*, 'beta=0.9971,n=1.007,theta=0.0905,Npe=3' if (m.eq.7.5) print*, 'beta=0.9978,n=1.007,theta=0.0980,Npe=4'

Computer Simulations Testing the detector’s photoelectron output for different

parameters with a computer simulation is the next step for the design of the detector.

SimCherenkov, a FORTRAN program written by D. W. Higinbotham, will provide this simulation. I have edited it considerably to allow further testing of more varied parameters. Nuclear Instruments and Methods in Physics Research A 414 (1998) 332-339

It allows inputs of the many variables such as dimensions of the detector, width of aerogel, various reflectivities and efficiencies and so on.

By experimenting with this program, we were able to find the more efficient design characteristics of our detector

These include but are not limited to the number, radius and placement of the PMTs and the dimensions of the box.

How many PMTs?

Top and

bottom?

Left and

right?What size box?

What kind of

PMT?

FORTRAN Programming

If (WALL .eq. 'NORTH') Then CPMT = ((WIDTH)/(PMTT)) ZPMT = (HEIGHT/2) Do COUNTER = 1,PMTT If (COUNTER .eq. ONE) Then YPMT = CPMT/2

Else YPMT = YPMT+CPMT

End If If ((COUNTER .ge. PF1) .and. (COUNTER .le. PF2)) Then

RING = SQRT ((NOWY-YPMT)**2+(NOWZ-ZPMT)**2) If (RING .le. RPMT) Then STAT = 'HIT!'

WHICHPMT=COUNTER Write(5,*),COUNTER End If Else If ((COUNTER .le. PF1) .or. (COUNTER .ge. PF2)) Then RING = SQRT ((NOWY-YPMT)**2+(NOWZ-ZPMT)**2)

If (RING .le. PR2) Then STAT = 'HIT!'

WHICHPMT=COUNTER Write(5,*),COUNTER End If End If End Do End If End If

• The original SimCherenkov Simulation was written in FORTRAN, so that was language I used as well.• Input parameters are written into a “variable.file” and read by the main simulation code.• At right is an example of a small piece of code that I wrote.• And by “small piece” I do really mean small: I added ~400 lines to an already ~1900 line program!• This piece is part the a large Do loop and If conditional that is testing what happens when a photon hits the NORTH wall when the box is set to the three-walled PMT configuration.

N

EW Bottom

Top

S

Hourglass Distribution At the spot where our

detector would be placed within the stack of the SHMS, the particle distribution is as pictured

The hourglass distribution is created by the arrangement of focusing and defocusing magnets that guide the scattered particles from the target to the detector stack.

The large dotted box is 110 x 100 cm and encloses the full spectrum.

The smaller box is 90 x 60 and encloses the area of largest density.

Distribution Location

SHMS

Length and Width of the Box

Here are three of the box optimizations. The blue lines are kaons and green are protons.

The most efficient was the one the most closely enclosed the hourglass distribution: 90 x 60

Other slightly larger aerogel panels resulted in some loss of photoelectron signal.

1 2 3 4 5 6 7 8 90.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

90 x 60

110 x 80

110 x 100

Momentum (GeV)

Phot

oele

ctro

ns

n = 1.015

Aerogel Thickness

In this case a considerable gain in signal was achieved by thicker aerogel values

10 cm seems to be far and away the best choice.

1 2 3 4 5 6 7 8 90.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

10 cm thick

8 cm thick

5 cm thick

Phot

oele

ctro

ns

Momentum (GeV)

n = 1.03

Photomultiplier Tube Placement

As far as PMT placement goes, lots of configurations had to be tried out.

The low values were for PMTs placed on one wall.

The next curves up is for PMTs on either the E/W walls or N/S, or three walls with mixed PMT radii

The cluster of three lines next up is PMTs on four walls, and PMTs on three walls with 90x90 and 110x100 panels

The top curve is for PMTs on three walls with the 90x60 panel.

1 2 3 4 5 6 7 8 90.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

Momentum (GeV)

Phot

oele

ctro

ns

90x90 110x100

90x60 (6x6x4)

110x100

110x100

n = 1.015

N

EW Bottom

Top

S

Light Box Depth/PMT Coverage %

Changing the Depth of the light box does not change the photoelectron output much

A smaller box outputs a slightly higher amount of photoelectrons because the PMT surface area coverage percentage is then slightly higher.

The higher the coverage percentage, the higher the output, generally.

4 4.5 5 5.5 6 6.5 7 7.5 83.00

3.50

4.00

4.50

5.00

5.50

6.00

6.50

7.00

Phot

oele

ctro

ns

Momentum (GeV)

22.5 cm 24.5 cm 26.5 cm

n = 1.015

Results of Optimization

The studies with the SimCherenkov Program result in several conclusions.

90 x 60 is the optimal size of the aerogel panel

10 cm is the optimal thicknessA three-walled 6 x 6 x 4 5” radius PMT

configuration is the optimal PMT placementThere will be an option left for future

expansion to a 110 x 100 panel. The aerogel thickness remains the same but the PMT configuration goes to 6 x 6 x 6 5” radius.

The Future of the Detector: JLab Hall C

Dr. Muller made it in the picture!

Trip to Jlab Summer 2010

The Future of the Detector: JLab Hall C

Hall C

The Future of the Detector: JLab Hall C

Our detector would be placed between the gas Cherenkov detector (for kaons and pions) and the second plane of the scintillator hodoscope. This also gives us physical constraints to work with for our detector.

SHMSGreen Box: Our proposed detector

Conclusion

Detector will operate with the new 12 GeV upgrade that the JLab is receiving.

This summer we found optimal solutions for various components of the detector using computational means such as calculations and simulations.

Future steps include making technical drawings to communicate the design to engineers, as well as constructing a test setup to test the performance of the aerogel and PMTs.

Research with our detector could bring us one step closer to understanding the substructure of the proton!