zukai wang university of virginia 1 aps april meeting 04/14/2013

1

Search of Magnetic Monopoles with the NOnA Detector

Zukai WangUniversity of Virginia

For the NOnA Collaboration

APS April Meeting04/14/2013

2

Dirac Quantization Condition

Assume an electron is transported along a closed path enclosing the Dirac String, the phase transition of its wave function should be:

Dirac Quantization Condition

Dirac ChargeMonopoles In NOvAZukai Wang

3

Generalized Maxwell Equations

To accommodate magnetic monopoles in classic electromagnetism, let’s rewrite the Maxwell Equations in a symmetric way:

Scalar

pseudoscalar

vector

pseudovector

Monopoles In NOvAZukai Wang

NOnA Monopole Search Strategy1. Look for highly-ionizing, penetrating particles• Covers the high-b range: b > 10-3

2. Look for slow, penetrating particles• Covers the low-b range: b < 10-3

dE/dx dt/dx

4Monopoles In NOvAZukai Wang

5

Geant4 simulation of monopole transportation in Silicon.And a linear interpolation was implemented to the unknown region.


D.E. Groom, PHYSICS REPORTS (Review Section of Physics Letters) 140. No.6(1986) 323-373

Minimum Ionizing

Simulation: Energy Loss

6

NOnA Detector


Weight: 14 kTon

Size: 15.6m x 15.6m x 59m

Please listen to other NOnA talks for details.

1. Total Surface Area: 4290 m2

2. It is a surface detector.

1 Far Detector = Could Fill 6,812 Kiddie Pools

7

NOnA Far Detector Technology


3.9 cm x 6.0cm x 15.6 cm

8

Simulation: Detector Response


b = 10-1

b = 10-2

b = 10-3

b = 10-4Minimum IonizingAm

plitu

de o

f Sig

nal f

rom

cel

l/AD

C

time since monopole entered cell /ns

9

Simulation: Event Display of High Energy Muon

Zukai Wang Monopoles In NOvA

10

Simulation: Event Display of Single Monopole


11

Trigger Algorithm Illustration


• The NOVA far detector keeps all of the zero-suppressed data for 1s in a online computer farm.

• Only a fraction of that data can be permanently stored.

• Hence an online software trigger is needed to reconstruct and select monopole-like events from the thousands of cosmic-ray events going through the detector each second.

• In what follows we illustrate a fast tracking finding algorithm that can be used to identify slow-moving tracks.

12

Illustration: Cosmic Raw Hits


Here is an example of simulated cosmic-ray events in 500 ms, containing ~10,000 cell hits.

13

3D Hough Space of Cosmic Tracks

A third parameter: reciprocal velocity, has been introduced to the conventional Hough Transform.


Signature of a Slow Monopole track Re

cipr

ocal

vel

ocity

(ns/

cm)

14

Test Result• A test using a cosmic simulation of 50 ms live

time has been done: containing ~5,000 cosmic tracks with 1,004,344 hits in FD.

• Timing & Overall Performance: Finds all tracks that hit more than 2 planes Working on improving the speed of algorithm

# of total tracks # of tracks longer than 2 planesMC Generated Muon Tracks 4840 2987

Reconstructed Tracks 3272 2987(100% reconstructed!)


15

Illustration: Reconstructed 3D Monopole Track


The plot above is a 3D event display of a simulated magnetic monopole with mass = 1016GeV/c2 and b = 10-3, shooting from the top of the NOnA detector and leaving a long track behind.

16

NOnA Reach

The NOnA reach curve has been generated with a toy MC.


NOnA: 4290 m2

MACRO: 3482 m2

SLIM: 427 m2

OHYA: 2000 m2

Surface Area Comparison:

17

Summary• The NOnA far detector will be able to search for magnetic

monopoles with excellent sensitivity over a large range of monopole masses and velocities.

• Fabrication of the far detector is well underway and will be completed in 2014.

• The tools needed to simulate the monopoles and the cosmic-ray background are in place, allowing reconstruction algorithms to be tested.

• Good progress has been made in developing fast track-reconstruction algorithms needed to enable a dead-time-less software trigger.

• Sensitivity estimates are in progress.


18

Acknowledgement

Zukai Wang Monopoles In NOvA

NOnA Collaboration: Huge progress have been made by working along with the awesome colleagues. I especially appreciate the help from Exotics group and DDT group.

Computing Division of Fermilab: These “artists” contributed a lot in setting up the software framework in both offline and data driven trigger. They also helped a lot in improving the performance of pattern recognition module, which is used in monopole triggering and reconstruction.

19

No Slow Monopoles on this floor

Reconstruction Algorithm: “Ground Floor” of Hough Space

19Monopoles In NOvAZukai Wang

20

Illustration: Reconstructed 2D Tracks


21

Back Up: Problem 1: Split Tracks


22

Back Up: Problem: Split Tracks

Monopoles In NOvA

No matter how loose the binning is, you always have a chance to split the Hough peak.To prevent looping over all hits again, the binning is pre-determined.

Binning boundary

Zukai Wang

23

Potential Solution: Combining Grids

Corp The entire time slice

Division Division Division

Monopoles In NOvA

Partitioned by DCM boundaries

Hough Transform

Looping over each combination of cells in the division, and combining all the ballots from all divisions.

CompanyCompany Company

Each Hough result is put into the cubic grids:

✗✓✓Above significance threshold

regiment Combining only adjacent (let me explain this in the next slide) grids

Below significance threshold

Zukai Wang

24

Defining Adjacent Grids

• Now we have grids in the cube (v bins in vpro, etc), and each grid can be labeled as:

• The distance of the two grids and is defined as following:

• Two grids are adjacent to each other if their distance is below 4.


25

Algorithm: General Organizing

Monopoles In NOvA

Hough Transform

Looping over each combination of cells in the division, and combining all the ballots from all divisions.

CompanyCompany Company

Each Hough result is put into the cubic grids:

✗✓✓Above significance threshold

Below significance threshold✓✓

regiment regiment

Aristophanes’ ProcessLet me explain later..

platoonplatoon

Platoon: hit list, which contains all the hits in a track (if perfectly done).

Zukai Wang

26

Back Up: Problem 2: Fake Tracks

noise

Hits associated with a shower

noiseA hit caused by a cosmic Muon


27

Back Up: Problem 2: Fake Tracks

noise

Hits associated with a shower

noiseA hit caused by a cosmic Muon

Accidentally, these hits meet the timing requirements to be filled into the same velocity grid.


28

Back Up: Path Length Inside Cell

200,000 Isotropic generated monopole’s distribution.Monopoles In NOvAZukai Wang

29

Back Up: Number of Cells Hits per Monopole in FD


30

Back Up: Number of Saturated Cells Hits per Monopole in FD

Note: assuming hits with PE > 1500 will be saturated, without considering attenuation. Monopoles In NOvAZukai Wang

31

Back Up: Energy Deposit per Monopole in FD


32

Back Up: Path Length Inside FD


33

Back Up: Monopole Physics: Ahlen Formula

1 2 3 4 5 6

B 0.248 0.672 1.022 1.243 1.464 1.685

K 0.406 0.346 0.346 0.346 0.346 0.346

Mean Ionization Potential:

Shifting Parameters:


34

Back Up: Monopole Physics: Lindhard Technique

Assuming the monopole passes through a degenerate Fermi gas of non-interacting electrons (this assumption is applicable when the monopole is slow enough: ):

as the monopole’s velocity decreases, fewer electrons of the Fermi sea are “available” for ionizing.


• Sensitivity roughly proportional to detector area• Very high-mass monopoles come isotropically from

all sides, unlike cosmic rays, lower mass monopoles from above

• The observed isotropic rate is: R = pFAe• F is the flux of monopoles (cm-2sr-1)• A is the total detector area (cm2)• e is the detector efficiency, livetime, etc.

• What we are after is not R, but the flux F = R/pAe• If we see no monopoles assume R = 2.3 to get the

90% CL limit:• F(90% CL) = 2.3 / pAe

Monopole Sensitivity

Zukai WangMonopoles In NOvA 35

Some areasNOnA: 4290 m2

MACRO: 3482 m2

SLIM: 427 m2

OHYA: 2000 m2

36

Calibration: Single Cell Hit of a slow Monopole

Illustration of APD response of a monopole with passing through a cell horizontally described by an analytical expression.

F = Fall Time = 7000nsR = Rise Time= 380ns


DDT: PatRec Algorithm• “3D” Hough Transform • Take all pairs of hits and find three voting

parameters for each.• DOCA

• cosq

• 1/v

• In this 3D Hough space monopoles are identified as clusters of points, “noise” is randomly spread out

•Ordinary straight track reconstruction algorithm

•This additional parameter implies a timing cut in recognizing a track with certain velocity

Through-going track.

Background pair.

37

Monopoles In NOvA37

Zukai Wang

38

DDT: Structure of Trigger With PatRec

Global Pattern Recognition

Time Slicer

Remove Noise

Space Slicer

Nue Calibration

TDC Sorter

Monopoles Supernova ……


39

DDT: Algorithm: Logic Flow

Making Pairs• Looping over all hits combinations• Calculating the voting parameters for each

pair

Partitioning• Partition by DCM and Time Slice: this step reduces the

number of combinations

Peak Identification

• Transform results of each pair are put into corresponding containers

• Making selections of each container to register a peak(track)


40

DDT: Algorithm: Why Partition?


41

DDT: Algorithm: Why Partition?


DCM Boundary

Limit: NOnA Potential: Overburden

Vertical Overburden:6” barite: 68.3 g/cm2

55” concrete: 347.9 g/cm2

atmosphere: 1030.0 g/cm2

Zukai Wang Monopoles In NOvA 42

43

Simulation: Energy Loss of Monopole: In All Related Material


44

Sensitivity: NOnA Potential


• Sensitivity goes as surface area: pFA, where F is the flux• Our acceptance is not yet known: we hope we can do better for

80% for high-mass monopoles and perhaps half that for low-mass • Eventually, if the acceptance is large enough, we can beat MACRO • Should be able to beat SLIM for intermediate-mass monopoles

45

NOnA Far Detector: First Data


46

NOnA Far Detector: First Data


47

Monopole Physics: Interaction

Rutherford Scattering

Correction by considering the electron spin (Y. Kazama, C. N. Yang, and A. S. Goldhaber, Phys. Rev. D 15, 2287 (1977) )


48

NOnA Far Detector TechnologyASIC: Application Specific Integrated Circuit

R = 380 nsF = 7000 ns


49

Simulation: Event Display of Single Monopole


zukai wang university of virginia 1 aps april meeting 04/14/2013

Documents