1. Imaging Maya Pyramids with Cosmic Ray Muons
An Application of the Tools of High Energy Physics

2. The Maya: Extraordinary American Culture

3. Some Background
1839-ff: John Lloyd Stephens with Frederick Catherwood, artist
Incidents of Travel in Central America, Chiapas, and Yucatan (1841)
Incidents of Travel in Yucatan (1843)
Linda Schele (1942 – 1998) UT Austin
The Code of Kings (1998) with Peter Mathews

4. What is the internal structure?
Underground Muon Detectors
Measure Spatial Distribution of Material Inside
by Muon Tomography

5. This is Proven Technology
Luis Alvarez* invented muon tomography in 1960’s to study the 2nd Pyramid of Chephren
Spark chambers used to track muons from Belzoni Chamber
System worked well—could see structures of caps
Main discovery: No other chambers exist
* L.W. Alvarez, et al, Search for Hidden Chambers
in the Pyramids Using Cosmic Rays, Science 167,
, 1970.

6. Cosmic Rays
Very high energy “primary” cosmic rays – typically protons – interact in upper atmosphere
Shower of unstable sub-nuclear particles created: typically pions, kaons
Muons and neutrinos are decay products of pions and kaons

7. Muons (and Neutrinos) are the Main Component of Cosmic Rays at the Earth’s Surface and Below
Primary cosmic rays interact in upper atmosphere
Mainly high energy protons
Showers of p’s/K’s created
Decay within 10’s – 1000’s meters or collide with nuclei in air
Muons are produced in decays of p/K
Do not have nuclear interactions
Lifetime much longer than p/K and dilated by relativity
Approximate muon rate at Earth’s surface:

8. Muon Flux at Surface
q = 0°
q = 75°

9. Muon Interactions in Matter
Energy loss: predominately by ionization
Multiple-Coulomb Scattering Ei L Ef
dirt, rock:

10. Basic Scheme: Muon enters nominal surface Void along underground path
Some muons will “range out” before reaching detector depending on incident energy and total material along path to detector
Trajectories having voids or depressions below nominal surface will have higher rates of muons hitting detector
Standard tomographic techniques can be used to reconstruct density of material in sensitive volume above detector
Detector measures
muon trajectory

11. Feasibility Depends on Underground Flux
Compute flux through a vertical surface buried in homogeneous material at a depth D under a flat plane
Zenith-angle shape approx. same at all depths
Relative Contrast describes change in flux relative to change in depth
D = 40 m d D

12. Arrangement Involving Cylindrical Detectors
Use 2 or more detectors
Compensates for “blind cone” inherent in cylindrical detectors
Improved sampling of target volume
Minimizes excavation

13. Basic Rates/Exposure Times
Extreme Case: 1 m3 void at 50 m
Solid angle DW = 1/502
Active area A = 3 m2
Resulting rate per DW bin: ~ 100/day/detector
Contrast ~ 1.5 x 1/50
1 s measurement needs 1000 events per DW bin
Second detector may see higher rates for same void
Bottom line:
1 s survey requires ~10 days
3 s survey requires ~ 3 months
Contours of Equal Counting Rate
for 2 Detectors (counts/day/pixel)
z (meters)
x (meters)

14. Detector Requirements
Efficiently trigger/record trajectories of underground muon tracks
Projected tracking errors
Angle: ±10 mrad
Location: ± 1 50 m
Provide for momentum threshold to reduce scattering errors
Large acceptance:
Active Area ~ few m2
0 < F < 2p
20° < Q < 90°
Operate in jungle conditions
robust/waterproof/simple
Projected Track
Error
Active Area
Track of Muon with Momentum p
Solid Angle
Acceptance

15. Detectors Cylindrical structure Muon tracking
1.5 m diameter
4.5 m long
Muon tracking
3 stereo layers
WLS-scintillator technology
Threshold energy selection
To improve pointing accuracy
Use inner volume as a Cherenkov radiator
Other systems
Electronics
Mechanical
Power/communications

16. Tracking Implementation
MINOS Scintillation Detectors:
Use scintillator strips with wavelength-shifting (WLS) fibers
Well-developed method, e.g. MINOS
Cost-effective
Robust
Geometry:
Three layers of strips on cylindrical form (Cherenkov radiator tank)
Axial, 0° stereo
± 30° stereo helices
30 mm strips Þ ~1 cm s

17. Tracking Resolution Measurement errors Multiple Scattering
Vertical Plane
Horizontal Plane
Distance through Rock (m)
s (m) ®
¬ s (m)
Measurement
Scattering
Q = 60°
Measurement errors
Determined by strip width and stereo angle
s < 1m in both planes
Multiple Scattering
Will dominate tracking errors for R > 40 m
Higher Cherenkov threshold can mitigate
GEANT simulations in progress

18. Discrimination Against Low-energy Muons
Multiple-Coulomb scattering is large for muons near the end of their range
Require minimum value for detected Ef ; options:
Energy loss in iron absorber:
Heavy ~ 1m thickness
Requires additional tracking layers
Detect Cherenkov radiation Þ v > c / ngas
Cherenkov threshold can be imposed/adjusted at data analysis stage
Cherenkov threshold curves:

19. Cherenkov Threshold Implementation
High energy muon
Fill central cylindrical volume of detector with Cherenkov radiator gas: C4F10
Muon threshold ~ 2 GeV
b = 1 p.e. yield: ~ 35/meter of radiator
C4F10 is a freon used for fire suppression
Make inner surface of cylinder optically reflecting
Place array of photon detectors on bottom of cylinder
C4F10 Radiator
Cherenkov Photons
Photo-detectors

20. Detector Electronics Systems
Data from detector
Tracking: 2X448 “hit” bits
Cherenkov: Analog out
Trigger
Based on tracking information only
Programmable logic
DAQ
All tracking bits
Cherenkov hits above pedestal
Control
Trigger/DAQ control
Monitor all detector systems

21. Tracking PMT Front-end Boards
Based on Harvard design for Atlas Muon System (LHC)
Harvard will design/procure boards for us

22. Trigger Requirements Use only tracking information Require:
>/= 2 Hit “Triplets”
Chord c > cmin
Direction ?
Flexible definition of Triplet
Coincidence gate: 20—50 ns
Number/pattern of hits to balance:
Noise – singles rates
Inefficiencies
Typical rates:
True events ~ 100 Hz
CR singles:
~ 4 KHz full detector
~ 25 Hz per strip

23. Detector Summary 2 units to be built (for 3-D reconstruction) Tracking
1.5 m diameter
4.5 m long (for convenience in fabrication/operations)
Tracking
3 stereo (0°, ±30°) layers of scintillator/WLS fibers
30 mm strip width; 448 strips per detector unit, readout both ends
PMT readout; 28 Hamamatsu M-64’s total (no multiplex)
Cherenkov threshold discrimination
Polished inner cylinder with C4F10 radiator
PMT detection on bottom face of cylinder
Typical thresholds ~ 5 – 10 photoelectrons
Data acquisition/trigger: ~100 bytes/event, 10—100 Hz rates
Commercial DAQ: e.g. “Lab View-like”
Simple majority-logic/event-topology-based trigger
Local data storage with periodic readout to base station

24. The Real World Funding Logistics
Sandia National Lab has provided detector development funds
UT funds supporting initial studies, students
National Instruments providing assistance with DAQ
Private donors expected to fund “jungle-related” costs
Logistics
First establish feasibility at UT
Support in Belize: transportation, excavation, operations

25. People & Things UT Physics UT Electrical & Computer Eng.
Brian Drell, Cesar Rodriquez
Anandi Salinas, Onise Sharia
Mark Selover, H. Adam Stevens
Austin Gleeson, RFS
UT Electrical & Computer Eng.
Bill Bard, Lizy John
National Instruments
Hugo Andrade, Joe Peck
UT Mesoamerican Archaeological Research Laboratory (MARL)
Prof. Fred Valdez, Director
Fermilab—Scintillator Production
Anna Pla-Dalmau
Harvard HEPL—Front-end Electronics
John Oliver
Other physicists who contributed in the early stages
Prof. Rich Muller, UC Berkeley
Dr. Dick Mischke, LANL

26. UT Mesoamerican Archaeological Research Laboratory

27. Potential Target Structure
La Milpa “Structure 1”
La Milpa site has relatively good access/infrastructure
Developing simulation tools to optimize detector design and placement
Plan excavations for deployment

28. Timeline Fall 2004 Winter 2004/05 Summer 2005 Fall 2005 Spring 2006
Complete detector design
Complete funding arrangements
Develop collaboration assignments
Winter 2004/05
Begin construction
Simulation tools available
Summer 2005
Detector integration and checkout at UT
Look for hidden structures in ENS or RLM!
Fall 2005
Complete second detector
Complete checkout at UT
Spring 2006
Transport to Belize

29. Other Potential Applications
Muon Tomography is good for monitoring large underground volumes (~100 m)3, provided:
You are interested in structures of scale 1 m – 10 m
You can afford to wait for weeks to months to acquire the data
The volume of interest is between your detector and the surface
Geological studies of aquifers
Shapes of underground cavities
Time-dependence of water levels
Monitoring of geology surrounding underground sites, e.g. underground nuclear waste storage

30. Summary Muon tomography is feasible
Proven in Alvarez experiment
New technologies enable simplified detector design
WLS/scintillator tracking well-developed/good match
Cherenkov threshold detector is indicated
New approach to problem of low-energy multiple-scattering
Well-understood physics/technology
Simplifies system design
Excellent project for engaging students
Other applications are possible
Maybe we can help to learn more about the Maya!