1. Status of Muon Tomography with SRS at FIT and some early beam results with SRS Michael Staib, Marcus Hohlmann Florida Institute of Technology Kondo Gnanvo University of Virginia RD51 Mini Week WG5 June 13, 2012

2. Muon Tomography for Homeland Security Photo: Sci. Am., 04/2008 Radiation Detectors But it is easy to avoid passive detection of nuclear threat material that is looking for radiological signatures. Just add shielding! There may be a solution... More than 23 million cargo containers and 93 million privately-owned vehicles were processed by U.S. Customs and Border Protection in the last year. New requirement: 100% scanning of all U.S.-bound containers using non-intrusive imaging equipment and radiation detection equipment to search for a nuclear threat.

3. Muon Tomography Concept μ μ FeU Large Scattering Small Scattering μ Iron Small Scattering Uranium μ Large Scattering μ μ Incoming muons (from natural cosmic rays) Note: Angles Exaggerated! Tracking detectors Multiple Coulomb scattering to 1 st order produces Gaussian distribution of scattering angles θ with width σ = Θ 0 :

4. Muon Tomography with Drift Tubes Brass Cu PbW Fe Al Original idea from Los Alamos (2003): Muon Tomography with Drift Tubes Reconstruction of 1 inch thick Pb letters J.A. Green, et al., “Optimizing the Tracking Efficiency for Cosmic Ray Muon Tomography”, LA-UR-06-8497, IEEE NSS 2006. INFN : Muon Tomography with spare CMS Muon Barrel Chambers (Drift Tubes) S. Presente, et al., Nucl. Inst. and Meth. A 604 (2009) 738-746. 1.4 m 4.3 m 1 2 3 Decision Sciences Corp.: Multi-Mode Passive Detection System, MMPDS TM INFN CMS Decision Sciences prototype using drift tubes large enough to scan a vehicle. C. Milner, et al., “Non-Invasive Imaging of Reactor Cores Using Cosmic Ray Muons”, SMU Physics Department Seminar, March 2012.

5. Compact Cubic-Foot Muon Tomography Station with GEMs Plastic Scintillator Triple-GEM Detector ~ 1 ft 3 Discriminator and coincidence card Two of the GEM detectors used were assembled at Florida Tech

6. SRS for Muon Tomography Current station configuration with 8 detectors: 96 APV Hybrid (48 M/S pairs) 6 ADC/FEC cards 2 Gigabit network switches Six 25 ns frames of data recorded for each APV per trigger yields event size of ~200kb @ 30 Hz. DATE for data acquisition. AMORE for data decoding, event monitoring and data analysis. 6

7. Tomographic POCA Reconstructions of Target Scenarios Tomographic Reconstructions Presented Material discrimination performance using five targets. Depleted uranium shielded with medium-Z shielding. POCA Reconstruction Limitations Assumes multiple scattering is well approximated by a single point. Not valid in the case of large amounts of material! Statistical methods must be employed. Does not take into account the momentum of the muon. The POCA can be found using the fact that the shortest line segment joining the incoming and outgoing vectors will be orthogonal to both. Reconstruction Algorithm Point of Closest Approach (POCA) Object

8. Five-Target Scenario Lead Z = 82 ρ = 11.4 g/cm 3 X 0 = 0.56 cm Depleted Uranium Z = 92 ρ = 19.0 g/cm 3 X 0 = 0.32 cm Tungsten Z = 74 ρ = 19.3 g/cm 3 X 0 = 0.35 cm Tin Z = 50 ρ = 7.3 g/cm 3 X 0 = 1.21 cm Iron Z = 26 ρ = 7.9 g/cm 3 X 0 = 1.76 cm 6mm Al shielding Five 75 cm 3 targets were placed inside the imaging volume at three different Z locations.

9. Five-Target Scenario Single cluster track selection 155,104 reconstructed tracks NNP cut = 5 2 mm x 2 mm x 40 mm voxels Pb W Fe Sn U Results are good! Can discriminate between high/low Z as well as high/medium Z. Tungsten vs. Uranium not so easy... Results match 1/X 0 dependence quite well

10. Five-Target Scenario XZ Slices YZ Slices Sn +X +Y Fe U U Sn PbW W W Fe -70 mm < Y < -30 mm-20 mm < Y < 20 mm30 mm < Y < 70 mm 30 mm < X < 70 mm-20 mm < X < 20 mm-70 mm < X < -30 mm Single cluster track selection 155,104 reconstructed tracks NNP cut = 5 2 mm x 2 mm x 40 mm voxels

11. Stacked Five-Target Scenario Stacks of each of the five materials were imaged using the MTS Targets vary in size from 27 cm 3 to 150 cm 3 175,022 tracks reconstructed using single cluster selection

12. Stacked Five-Target Scenario Single cluster track selection 175,022 reconstructed tracks NNP cut = 6 2 mm x 2 mm x 40 mm voxels UraniumTungstenLeadTinIron 152.8140.6112.272.964.1 Simple Scattering Density [deg/cm3] We are able to discriminate between the low/medium/high-Z materials! 40 mm slice 40 mm XY slice descending in Z by 5 mm per frame

13. Stacked Five-Target Scenario WPbWFe UU Sn Fe SnPb Single cluster track selection 175,022 reconstructed tracks NNP cut = 6 2 mm x 2 mm x 40 mm voxels XZ Slices YZ Slices

14. Depleted Uranium with Bronze Shielding Mixed track selection 187,731 reconstructed tracks 2 mm x 2 mm x 40 mm voxels 40 mm XY slice with NNP cut increasing by 1 per frame The shielded uranium can be discriminated from the bronze shielding using POCA reconstruction

15. Depleted Uranium with Bronze Shielding Mixed track selection 187,731 reconstructed tracks 2 mm x 2 mm x 40 mm voxels XZ slice with NNP cut increasing by 1 per frame The shielded uranium can be discriminated from the bronze shielding using POCA reconstruction YZ slice with NNP cut increasing by 1 per frame What about the side views?

16. Status of Muon Tomography We have shown the ability of the MTS using GEMs to discriminate between materials of similar volume with different Z, even as shielding. SRS is working very well! Still unresolved issue of network switch requirements and exact cause of missing triggers. Will try to implement the zero suppression firmware soon and work on clock synchronization. Plans to possibly scale to ~1 m 3 active volume in the future. Many thanks to Sorin, Hans, Filippo and Leszek for their help throughout the process.

17. CAD Design by C. Pancake, Stony Brook 2 mm Beam Test 2012 Zig-Zag (Chevron) strips to reduce readout channels while maintaining spatial resolution Preliminary Results

18. Zero Suppression Raw DataPedestal Subtracted Data Zero Suppressed (5σ RMS) DataTime Evolution of Zero Suppressed Signal Floating channels (not connected) 48 Zig-Zag RO Strips Strip Number

19. Mean = 2.7 Strips Mean = 1.1 Clusters Pedestal Noise Cluster Charge Distribution Cluster Size DistributionCluster Multiplicity

20. Beam Test 2012 New RD51 tracker with resistive strip MicroMegas and SRS APV readout Readout Strips parallel to resistive strips Readout perpendicular to resistive strips

21. Thanks! Questions?

22. Backup Slides

23. Post-Processing to Remove Noise A “number of neighboring POCA ” (NNP) cut is made in order to improve the quality of the reconstructions. There is also a cut removing all voxels with mean scattering angle less than 2 degrees. V Add up the total number of POCA points in the blue voxels surrounding voxel V, this is the NNP. If the NNP is less than a threshold, remove the contents of voxel V. Repeat for all voxels in the histogram. MethodResults